Reservoir Stimulation

elgoajiroblanco@hotmail.
com
© 2010 COPYRIGHT MERCADO NEGRO, LAS PLAYITAS. MARACAIBO-EDO. ZULIA, VENEZUELA.

PARA COMPRAR AL DETAL O AL MAYOR, ESTE Y OTROS PRODUCTOS, FAVOR PREGUNTAR POR EL GÖAJIRO BLANCO, EN EL MERCADO LAS PLAYITAS.
ADVERTENCIA: "EL DERECHO DE AUTOR NO ES UNA FORMA DE PROPIEDAD SINO UN DERECHO CULTURAL. EXIGE TU DERECHO"
elgoajiroblanco@hotmail.com
MICHAEL J. ECONOMIDES is Professor of Chemical Engineering at

the University of Houston. Until the summer of 1998, he was the Samuel
R. Noble Professor of Petroleum Engineering at Texas A & M University
and served as Chief Scientist of the Global Petroleum Research Institute
(GPRI).Prior to joining the faculty at Texas A & M University, Professor
Economides was the Director of the Institute of Drilling and Production
at the Institute of Drilling and Production at the Leoben Mining Institute
in Austria (1989-1993). From 1984 to 1989, he worked in a variety of
senior positions with the Schlumberger companies, including Europe
Region Reservoir Engineering and Stimulation Manager and Senior
Staff Engineer, North America. Publications include authoring or co-
authoring of 7 textbooks and more that 150 journal papers and articles
ranges of industrial consulting, including major retainers by national oil
companies at the country level and by Fortune 500 companies. He is the
founder and a major shareholder in OTEK (Australia), a petroleum
service and consulting firm with offices in five Australian cities. In
addition to his technical interests, he has written extensively in wide
circulation media on a broad range of topics associated with energy and
geopolitical issues.
KENNETH G. NOLTE has held various senior technical and marketing

positions with Schlumberger since 1986. From 1984 to 1986, he was
with Nolte-Smith, Inc. (now NSI Technologies, Inc.). Prior to 1984, Dr.
Nolte was a research associate with Amoco Production Company,
where he worked for 16 years in the areas of offshore/arctic technology
and hydraulic fracturing. Dr. Nolte holds a BS degree from the University
of Illinois and received and MS and PhD from Brown University. He has
authored numerous journal publications and has authored numerous
journal publications and has various patents relating to material
behavior, drilling, offshore technology and fracturing. Dr. Nolte was
1986-1987 SPE Distinguished Lecturer and received the Lester C. Uren
Award in 1992.

© Schlumberger 2000
Schlumberger Dowell
300 Schlumberger Drive
Sugar Land, Texas 77478
Technical editor: Elsa Kapitan-White

Graphic design and production: Martha Dutton
Production manager: Robert Thrasher
Published by John Wiley & Sons Ltd,

Baffins Lane, Chichester,
West Sussex PO19 lUD, England
National 01243 779777

International (+44) 1243 779777
e-mail (for orders and customer service enquires): es-books@wiley.co.uk
Visit our Home Page on http://www.wiley.co.uk

or
Visit our Home Page on http://www.wiley.com
All right reserved. No part of this publication may be reproduced, stored

in a retrieval system, or transmitted, in any form or by any means, electronic,
mechanical, photocopying, recording, scanning or otherwise, except under the terms
of the Copyright, Designs and Patents Act 1988 or under the terms of a license issued
by the Copyright Licensing Agency, 90 Tottenham Court Road, London UK WIP)HE;
without the permission in writing of the publisher.
Other Wiley Editorial Offices
John Wiley & Sons, Inc., 605 Third Avenue,

New York, NY 10158-0012, USA
Wiley-VCH Verlag GmbH, Pappelalee 3,

D-69469 Weinheim, Germany
Jacaranda Wiley Ldt, 33 Park Road, Milton,

Queensland 4064, Australia
John Wiley & Sons (Asia) Pte Ltd, 2 Clementi Loop #02-01,
Jin Xing Distripark, Singapore 0512
John Wiley & Sons (Canada) Ltd, 22 Worcester Road,

Rexdale, Ontario M9W 1L1, Canada
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library
ISBN 0 471 49192 6
Produced from camera-ready copy supplied by Schlumberger Dowell

Printed and bound in Great Britain by Bookcraft (Bath) Ltd
This book is printed on acid-free paper responsibly manufactured from sustainable
forestry, in which at least two trees are planted for each one used for paper production.

◆
Reservoir Stimulation
Contents
Preface: Hydraulic Fracturing, A Technology for All Time

Ahmed S. Abou-Sayed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . P-1
Chapter 1 Reservoir Stimulation in Petroleum Production

Michael J. Economides and Curtis Boney
1-1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
1-1.1. Petroleum production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
1-1.2. Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-3
1-2. Inflow performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-3
1-2.1. IPR for steady state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-4
1-2.2. IPR for pseudosteady state. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-5
1-2.3. IPR for transient (or infinite-acting) flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-5
1-2.4. Horizontal well production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-6
1-2.5. Permeability anisotropy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-10
1-3. Alterations in the near-wellbore zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-11
1-3.1. Skin analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-11
1-3.2. Components of the skin effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-12
1-3.3. Skin effect caused by partial completion and slant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-12
1-3.4. Perforation skin effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-13
1-3.5. Hydraulic fracturing in production engineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-16
1-4. Tubing performance and NODAL* analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-18
1-5. Decision process for well stimulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-20
1-5.1. Stimulation economics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-21
1-5.2. Physical limits to stimulation treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-22
1-6. Reservoir engineering considerations for optimal production enhancement strategies . . . . . . . 1-22
1-6.1. Geometry of the well drainage volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-23
1-6.2. Well drainage volume characterizations and production optimization strategies . . . . 1-24
1-7. Stimulation execution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-28
1-7.1. Matrix stimulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-28
1-7.2. Hydraulic fracturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-18
Chapter 2 Formation Characterization: Well and Reservoir Testing

Christine A. Ehlig-Economides and Michael J. Economides
2-1. Evolution of a technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1
2-1.1. Horner semilogarithmic analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1
2-1.2. Log-log plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-2
2-2. Pressure derivative in well test diagnosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-3
2-3. Parameter estimation from pressure transient data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-7
2-3.1. Radial flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-7
2-3.2. Linear flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-9
Reservoir Stimulation v

2-3.3. Spherical flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-10

2-3.4. Dual porosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-11
2-3.5. Wellbore storage and pseudosteady state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-11
2-4. Test interpretation methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-12
2-5. Analysis with measurement of layer rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-14
2-6. Layered reservoir testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-15
2-6.1. Selective inflow performance analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-15
2-6.2. Analysis of multilayer transient test data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-16
2-7. Testing multilateral and multibranch wells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-16
2-8. Permeability determination from a fracture injection test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-17
2-8.1. Pressure decline analysis with the Carter leakoff model . . . . . . . . . . . . . . . . . . . . . . . . . . 2-17
2-8.2. Filter-cake plus reservoir pressure drop leakoff model
(according to Mayerhofer et al., 1993). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-21
Chapter 3 Formation Characterization: Rock Mechanics

M. C. Thiercelin and J.-C. Roegiers
3-1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1
Sidebar 3A. Mechanics of hydraulic fracturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-2
3-2. Basic concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4
3-2.1. Stresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4
3-2.2. Strains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-5
Sidebar 3B. Mohr circle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-5
3-3. Rock behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-6
3-3.1. Linear elasticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-6
Sidebar 3C. Elastic constants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-8
3-3.2. Influence of pore pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-8
3-3.3. Fracture mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-9
3-3.4. Nonelastic deformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-11
3-3.5. Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-11
3-4. Rock mechanical property measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-12
3-4.1. Importance of rock properties in stimulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-12
3-4.2. Laboratory testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-13
3-4.3. Stress-strain curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-14
3-4.4. Elastic parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-15
3-4.5. Rock strength, yield criterion and failure envelope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-19
3-4.6. Fracture toughness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-19
Sidebar 3D. Fracture toughness testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-20
3-5. State of stress in the earth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-21
3-5.1. Rock at rest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-22
3-5.2. Tectonic strains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-23
3-5.3. Rock at failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-23
3-5.4. Influence of pore pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-25
3-5.5. Influence of temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-26
3-5.6. Principal stress direction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-26
3-5.7. Stress around the wellbore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-26
3-5.8. Stress change from hydraulic fracturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-27
3-6. In-situ stress management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-28
3-6.1. Importance of stress measurement in stimulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-28
3-6.2. Micro-hydraulic fracturing techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-28
vi Contents

3-6.3. Fracture calibration techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-34

3-6.4. Laboratory techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-34
Chapter 4 Formation Characterization: Well Logs

Jean Desroches and Tom Bratton
4-1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1
4-2. Depth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-2
4-3. Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-2
4-4. Properties related to the diffusion of fluids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-3
4-4.1. Porosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-3
4-4.2. Lithology and saturation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-5
4-4.3. Permeability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-6
Sidebar 4A. Permeability-porosity correlations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-8
4-4.4. Pore pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-10
4-4.5. Skin effect and damage radius . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-11
4-4.6. Composition of fluids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-12
4-5. Properties related to the deformation and fracturing of rock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-13
4-5.1. Mechanical properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-13
4-5.2. Stresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-15
4-6. Zoning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-24
Chapter 5 Basics of Hydraulic Fracturing

M. B. Smith and J. W. Shlyapobersky
5-1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1
5-1.1. What is fracturing?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1
5-1.2. Why fracture? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-4
5-1.3. Design considerations and primary variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-6
Sidebar 5A. Design goals and variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-7
5-1.4. Variable interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-9
5-2. In-situ stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-9
5-3. Reservoir engineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-10
5-3.1. Design goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-11
Sidebar 5B. Highway analogy for dimensionless fracture conductivity . . . . . . . . . . . 5-11
5-3.2. Complicating factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-12
5-3.3. Reservoir effects on fluid loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-13
5-4. Rock and fluid mechanics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-13
5-4.1. Material balance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-13
5-4.2. Fracture height . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-14
5-4.3. Fracture width . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-15
5-4.4. Fluid mechanics and fluid flow. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-15
5-4.5. Fracture mechanics and fracture tip effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-16
5-4.6. Fluid loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-17
5-4.7. Variable sensitivities and interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-18
5-5. Treatment pump scheduling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-20
5-5.1. Fluid and proppant selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-21
5-5.2. Pad volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-21
5-5.3. Proppant transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-23
5-5.4. Proppant admittance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-24
5-5.5. Fracture models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-25
Reservoir Stimulation vii

5-6. Economics and operational considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-26

5-6.1. Economics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-26
5-6.2. Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-27
Appendix: Evolution of hydraulic fracturing design and evaluation

K. G. Nolte . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A5-1
Chapter 6 Mechanics of Hydraulic Fracturing

Mark G. Mack and Norman R. Warpinski
6-1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1
6-2. History of early hydraulic fracture modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-2
6-2.1. Basic fracture modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-2
6-2.2. Hydraulic fracture modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-3
Sidebar 6A. Approximation to the Carter equation for leakoff . . . . . . . . . . . . . . . . . . . 6-6
Sidebar 6B. Approximations to Nordgren’s equations . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-6
Sidebar 6C. Radial fracture geometry models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-8
6-3. Three-dimensional and pseudo-three-dimensional models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-8
Sidebar 6D. Field determination of fracture geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-10
6-3.1. Planar three-dimensional models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-11
Sidebar 6E. Lateral coupling in pseudo-three-dimensional models . . . . . . . . . . . . . . . 6-12
Sidebar 6F. Momentum conservation equation for hydraulic fracturing . . . . . . . . . . 6-13
Sidebar 6G. Momentum balance and constitutive equation for
non-Newtonian fluids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-14
6-3.2. Cell-based pseudo-three-dimensional models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-16
Sidebar 6H. Stretching coordinate system and stability analysis . . . . . . . . . . . . . . . . . 6-20
6-3.3. Lumped pseudo-three-dimensional models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-23
6-4. Leakoff. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-25
6-4.1. Filter cake . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-25
6-4.2. Filtrate zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-26
6-4.3. Reservoir zone. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-26
6-4.4. Combined mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-26
6-4.5. General model of leakoff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-27
6-4.6. Other effects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-27
6-5. Proppant placement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-28
6-5.1. Effect of proppant on fracturing fluid rheology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-28
6-5.2. Convection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-28
6-5.3. Proppant transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-29
6-6. Heat transfer models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-29
6-6.1. Historical heat transfer models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-30
6-6.2. Improved heat transfer models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-30
6-7. Fracture tip effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-30
Sidebar 6I. Efficient heat transfer algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-31
Sidebar 6J. Verification of efficient thermal calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-32
6-7.1. Linear elastic fracture mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-32
Sidebar 6K. Crack tip stresses and the Rice equation . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-33
6-7.2. Extensions to LEFM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-34
6-7.3. Field calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-35
6-8. Tortuosity and other near-well effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-36
6-8.1. Fracture geometry around a wellbore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-36
6-8.2. Perforation and deviation effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-36
viii Contents

6-8.3. Perforation friction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-37

6-8.4. Tortuosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-37
6-8.5. Phasing misalignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-38
6-9. Acid fracturing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-40
6-9.1. Historical acid fracturing models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-40
6-9.2. Reaction stoichiometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-40
6-9.3. Acid fracture conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-41
6-9.4. Energy balance during acid fracturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-42
6-9.5. Reaction kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-42
6-9.6. Mass transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-42
6-9.7. Acid reaction model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-43
6-9.8. Acid fracturing: fracture geometry model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-43
6-10. Multilayer fracturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-44
6-11. Pump schedule generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-46
Sidebar 6L. Approximate proppant schedules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-47
6-12. Pressure history matching. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-48
Sidebar 6M. Theory and method of pressure inversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-48
Chapter 7 Fracturing Fluid Chemistry and Proppants

Janet Gulbis and Richard M. Hodge
7-1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1
7-2. Water-base fluids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1
7-3. Oil-base fluids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-6
7-4. Acid-based fluids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-7
7-4.1. Materials and techniques for acid fluid-loss control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-7
7-4.2. Materials and techniques for acid reaction-rate control . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-8
7-5. Multiphase fluids. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-8
7-5.1. Foams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-9
7-5.2. Emulsions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-9
7-6. Additives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-10
7-6.1. Crosslinkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-10
Sidebar 7A. Ensuring optimum crosslinker performance . . . . . . . . . . . . . . . . . . . . . . . . 7-13
7-6.2. Breakers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-14
Sidebar 7B. Breaker selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-16
7-6.3. Fluid-loss additives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-16
7-6.4. Bactericides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-18
7-6.5. Stabilizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-18
7-6.6. Surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-19
7-6.7. Clay stabilizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-19
7-7. Proppants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-19
7-7.1. Physical properties of proppants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-19
7-7.2. Classes of proppants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-21
Sidebar 7C. Minimizing the effects of resin-coated proppants . . . . . . . . . . . . . . . . . . . . 7-22
7-8. Execution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-22
7-8.1. Mixing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-22
7-8.2. Quality assurance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-23
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-23
Reservoir Stimulation ix

Chapter 8 Performance of Fracturing Materials

Vernon G. Constien, George W. Hawkins, R. K. Prud’homme and Reinaldo Navarrete
8-1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-1
8-2. Fracturing fluid characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-1
8-3. Characterization basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-2
8-4. Translation of field conditions to a laboratory environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-2
8-5. Molecular characterization of gelling agents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-2
8-5.1. Correlations of molecular weight and viscosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-2
8-5.2. Concentration and chain overlap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-3
8-5.3. Molecular weight distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-4
8-5.4. Characterization of insoluble components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-5
8-5.5. Reaction sites and kinetics of crosslinking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-5
8-6. Rheology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-6
8-6.1. Basic flow relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-7
8-6.2. Power law model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-7
8-6.3. Models that more fully describe fluid behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-8
8-6.4. Determination of fracturing fluid rheology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-10
8-6.5. Rheology of foam and emulsion fluids. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-12
8-6.6. Effect of viscometer geometry on fluid viscosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-15
8-6.7. Characterization of fluid microstructure using dynamic oscillatory measurements . . 8-16
8-6.8. Relaxation time and slip . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-17
8-6.9. Slurry rheology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-17
8-7. Proppant effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-19
8-7.1. Characterization of proppant transport properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-19
8-7.2. Particle migration and concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-21
8-8. Fluid loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-22
8-8.1. Fluid loss under static conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-23
8-8.2. Fluid loss under dynamic conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-24
8-8.3. Shear rate in the fracture and its influence on fluid loss . . . . . . . . . . . . . . . . . . . . . . . . . . 8-25
8-8.4. Influence of permeability and core length . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-26
8-8.5. Differential pressure effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-26
Chapter 9 Fracture Evaluation Using Pressure Diagnostics
Sunil N. Gulrajani and K. G. Nolte
9-1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-1
9-2. Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-2
9-3. Fundamental principles of hydraulic fracturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-3
9-3.1. Fluid flow in the fracture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-3
9-3.2. Material balance or conservation of mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-4
9-3.3. Rock elastic deformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-4
Sidebar 9A. What is closure pressure? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-6
Sidebar 9B. Pressure response of toughness-dominated fractures . . . . . . . . . . . . . . . . . 9-9
9-4. Pressure during pumping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-10
9-4.1. Time variation for limiting fluid efficiencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-12
9-4.2. Inference of fracture geometry from pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-12
9-4.3. Diagnosis of periods of controlled fracture height growth . . . . . . . . . . . . . . . . . . . . . . . . 9-14
9-4.4. Examples of injection pressure analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-15
Sidebar 9C. Pressure derivative analysis for diagnosing pumping pressure . . . . . . . . 9-16
9-4.5. Diagnostics for nonideal fracture propagation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-18
x Contents

Sidebar 9D. Fluid leakoff in natural fissures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-23

9-4.6. Formation pressure capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-24
9-4.7. Pressure response after a screenout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-27
9-4.8. Fracture diagnostics from log-log plot slopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-28
9-4.9. Near-wellbore effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-30
Sidebar 9E. Rate step-down test analysis—a diagnostic for fracture entry. . . . . . . . . 9-32
9-5. Analysis during fracture closure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-34
9-5.1. Fluid efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-34
9-5.2. Basic pressure decline analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-37
9-5.3. Decline analysis during nonideal conditions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-38
9-5.4. Generalized pressure decline analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-42
Sidebar 9F. G-function derivative analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-43
9-6. Pressure interpretation after fracture closure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-45
9-6.1. Why linear and radial flow after fracture closure? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-46
9-6.2. Linear, transitional and radial flow pressure responses . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-48
Sidebar 9G. Impulse testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-49
9-6.3. Mini-falloff test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-50
9-6.4. Integration of after-closure and preclosure analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-50
9-6.5. Physical and mathematical descriptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-51
9-6.6. Influence of spurt loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-53
9-6.7. Consistent after-closure diagnostic framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-54
9-6.8. Application of after-closure analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-56
9-6.9. Field example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-57
9-7. Numerical simulation of pressure: combined analysis of pumping and closing . . . . . . . . . . . . . 9-59
9-7.1. Pressure matching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-60
9-7.2. Nonuniqueness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-60
9-8. Comprehensive calibration test sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-61
Appendix: Background for hydraulic fracturing pressure analysis techniques

Sunil N. Gulrajani and K. G. Nolte . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A9-1
Chapter 10 Fracture Treatment Design

Jack Elbel and Larry Britt
10-1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-1
Sidebar 10A. NPV for fixed costs or designated proppant mass . . . . . . . . . . . . . . . . . . . . . . . . . . 10-2
10-2. Design considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-3
10-2.1. Economic optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-3
10-2.2. Treatment optimization design procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-3
10-2.3. Fracture conductivity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-4
10-2.4. Dimensionless fracture conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-6
10-2.5. Non-Darcy effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-8
10-2.6. Proppant selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-8
10-2.7. Treatment size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-9
10-2.8. Fluid loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-10
10-2.9. Viscosity effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-11
Sidebar 10B. Fluid exposure time. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-12
10-2.10. Injection rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-13
10-3. Geometry modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-14
Sidebar 10C. Geometry models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-14
10-3.1. Model selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-15
Reservoir Stimulation xi

10-3.2. Sources of formation parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-16

Sidebar 10D. In-situ stress correlation with lithology . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-16
10-4. Treatment schedule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-17
Sidebar 10E. Fracturing economics sensitivity to formation permeability and skin effect . . . 10-17
10-4.1. Normal proppant scheduling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-18
10-4.2. Tip screenout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-21
10-5. Multilayer fracturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-24
10-5.1. Limited entry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-24
10-5.2. Interval grouping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-25
10-5.3. Single fracture across multilayers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-25
10-5.4. Two fractures in a multilayer reservoir . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-26
10-5.5. Field example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-28
Sidebar 10F. Fracture evaluation in multilayer zones . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-28
10-6. Acid fracturing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-30
10-6.1. Acid-etched fracture conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-31
Sidebar 10G. Acid-etched conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-32
10-6.2. Acid fluid loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-33
Sidebar 10H. Fluid-loss control in wormholes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-34
10-6.3. Acid reaction rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-35
10-6.4. Acid fracturing models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-36
10-6.5. Parameter sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-36
10-6.6. Formation reactivity properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-41
10-6.7. Propped or acid fracture decision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-41
10-7. Deviated wellbore fracturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-42
10-7.1. Reservoir considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-43
10-7.2. Fracture spacing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-45
10-7.3. Convergent flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-45
10-7.4. Fracturing execution in deviated and horizontal wells . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-47
10-7.5. Horizontal well example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-49
Chapter 11 Fracturing Operations

J. E. Brown, R. W. Thrasher and L. A. Behrmann
11-1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-1
11-2. Completions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-1
11-2.1. Deviated and S-shaped completions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-1
11-2.2. Horizontal and multilateral completions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-2
11-2.3. Slimhole and monobore completions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-2
11-2.4. Zonal isolation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-2
Sidebar 11A. Factors influencing cement bond integrity . . . . . . . . . . . . . . . . . . . . . . . . . 11-3
Sidebar 11B. Coiled tubing–conveyed fracture treatments . . . . . . . . . . . . . . . . . . . . . . . 11-6
11-3. Perforating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-8
11-3.1. Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-8
Sidebar 11C. Estimating multizone injection profiles during hydraulic fracturing . . 11-9
Sidebar 11D. Propagating a microannulus during formation breakdown. . . . . . . . . . . 11-11
11-3.2. Perforation phasing for hard-rock hydraulic fracturing . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-11
11-3.3. Other perforating considerations for fracturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-14
11-3.4. Frac and packs and high-rate water packs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-16
11-3.5. Fracturing for sand control without gravel-pack screens. . . . . . . . . . . . . . . . . . . . . . . . . . 11-16
Sidebar 11E. Calculation of minimum shot density for fracture stimulation . . . . . . . 11-17
11-3.6. Extreme overbalance stimulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-18
xii Contents

11-3.7. Well and fracture connectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-18

11-4. Surface equipment for fracturing operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-19
11-4.1. Wellhead isolation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-19
11-4.2. Treating iron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-19
11-4.3. High-pressure pumps. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-22
11-4.4. Blending equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-23
11-4.5. Proppant storage and delivery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-23
11-4.6. Vital signs from sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-24
11-4.7. Equipment placement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-26
11-5. Bottomhole pressure measurement and analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-26
11-6. Proppant flowback control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-29
11-6.1. Forced closure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-30
11-6.2. Resin flush . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-30
11-6.3. Resin-coated proppants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-30
11-6.4. Fiber technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-30
11-7. Flowback strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-30
Sidebar 11F. Fiber technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-31
11-8. Quality assurance and quality control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-32
11-9. Health, safety and environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-32
11-9.1. Safety considerations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-32
11-9.2. Environmental considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-33
Appendix: Understanding perforator penetration and flow performance
Phillip M. Halleck . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A11-1
Chapter 12 Post-Treatment Evaluation and Fractured Well Performance

B. D. Poe, Jr., and Michael J. Economides
12-1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-1
12-1.1. Fracture mapping techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-1
12-1.2. Pressure transient analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-6
12-2. Post-treatment fracture evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-10
12-2.1. Wellbore storage dominated flow regime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-11
12-2.2. Fracture storage linear flow regime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-11
12-2.3. Bilinear flow regime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-11
12-2.4. Formation linear flow regime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-13
12-2.5. Pseudoradial flow regime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-14
12-2.6. Pseudosteady-state flow regime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-15
12-3. Factors affecting fractured well performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-16
12-3.1. Non-Darcy flow behavior. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-16
12-3.2. Nonlinear fluid properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-20
12-3.3. Fracture damage and spatially varying fracture properties . . . . . . . . . . . . . . . . . . . . . . . . 12-21
12-3.4. Damage in high-permeability fracturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-25
12-3.5. Heterogeneous systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-26
12-4. Well test analysis of vertically fractured wells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-27
12-4.1. Wellbore storage dominated flow analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-28
12-4.2. Fracture storage linear flow analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-28
12-4.3. Bilinear flow analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-29
12-4.4. Formation linear flow analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-29
12-4.5. Pseudoradial flow analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-29
12-4.6. Well test design considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-30
Reservoir Stimulation xiii

12-4.7. Example well test analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-31

12-5. Prediction of fractured well performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-39
Chapter 13 Introduction to Matrix Treatments

R. L. Thomas and L. N. Morgenthaler
13-1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-1
13-1.1. Candidate selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-1
Sidebar 13A. The history of matrix stimulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-2
13-1.2. Formation damage characterization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-3
13-1.3. Stimulation technique determination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-3
13-1.4. Fluid and additive selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-3
13-1.5. Pumping schedule generation and simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-3
13-1.6. Economic evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-4
13-1.7. Execution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-4
13-1.8. Evaluation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-4
13-2. Candidate selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-4
13-2.1. Identifying low-productivity wells and stimulation candidates . . . . . . . . . . . . . . . . . . . . 13-4
Sidebar 13B. Candidate selection field case history . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-6
13-2.2. Impact of formation damage on productivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-6
13-2.3. Preliminary economic evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-7
13-3. Formation damage characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-8
Sidebar 13C. Formation damage characterization field case history . . . . . . . . . . . . . . . . . . . . . . . 13-9
Sidebar 13D. Fluid and additive selection field case history . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-10
13-4. Stimulation technique determination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-10
13-5. Treatment design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-11
13-5.1. Matrix stimulation techniques. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-11
13-5.2. Treatment fluid selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-12
13-5.3. Pumping schedule generation and simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-19
Sidebar 13E. Placement study case histories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-29
13-6. Final economic evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-32
13-7. Execution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-32
13-7.1. Quality control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-32
13-7.2. Data collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-34
13-8. Treatment evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-35
13-8.1. Pretreatment evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-35
13-8.2. Real-time evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-35
13-8.3. Post-treatment evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-37
Chapter 14 Formation Damage: Origin, Diagnosis and Treatment Strategy

Donald G. Hill, Olivier M. Liétard, Bernard M. Piot and George E. King
14-1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-1
14-2. Damage characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-1
14-2.1. Pseudodamage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-2
14-2.2. Pseudoskin effects and well completion and configuration . . . . . . . . . . . . . . . . . . . . . . . 14-3
14-3. Formation damage descriptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-4
14-3.1. Fines migration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-4
14-3.2. Swelling clays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-6
14-3.3. Scales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-6
14-3.4. Organic deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-7
14-3.5. Mixed deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-8
xiv Contents

14-3.6. Emulsions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-9

14-3.7. Induced particle plugging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-9
14-3.8. Wettability alteration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-10
14-3.9. Acid reactions and acid reaction by-products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-11
14-3.10. Bacteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-11
14-3.11. Water blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-12
14-3.12. Oil-base drilling fluids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-13
14-4. Origins of formation damage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-13
14-4.1. Drilling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-13
14-4.2. Cementing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-21
14-4.3. Perforating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-21
14-4.4. Gravel packing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-22
14-4.5. Workovers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-22
14-4.6. Stimulation and remedial treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-23
14-4.7. Normal production or injection operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-24
14-5. Laboratory identification and treatment selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-26
14-5.1. Damage identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-26
14-5.2. Treatment selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-28
14-6. Treatment strategies and concerns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-31
14-6.1. Fines and clays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-33
14-6.2. Scales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-34
14-6.3. Organic deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-35
14-6.4. Mixed deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-35
14-6.5. Emulsions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-36
14-6.6. Bacteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-36
14-6.7. Induced particle plugging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-36
14-6.8. Oil-base drilling fluids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-37
14-6.9. Water blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-37
14-6.10. Wettability alteration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-38
14-6.11. Wellbore damage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-38
14-7. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-39
Chapter 15 Additives in Acidizing Fluids
Syed A. Ali and Jerald J. Hinkel
15-1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-1
15-2. Corrosion inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-2
15-2.1. Corrosion of metals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-2
15-2.2. Acid corrosion on steel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-2
15-2.3. Pitting types of acid corrosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-3
15-2.4. Hydrogen embrittlement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-3
15-2.5. Corrosion by different acid types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-3
15-2.6. Inhibitor types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-4
15-2.7. Compatibility with other additives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-4
15-2.8. Laboratory evaluation of inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-5
15-2.9. Suggestions for inhibitor selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-5
15-3. Surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-5
15-3.1. Anionic surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-6
15-3.2. Cationic surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-6
15-3.3. Nonionic surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-6
15-3.4. Amphoteric surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-7
Reservoir Stimulation xv

15-3.5. Fluorocarbon surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-7

15-3.6. Properties affected by surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-7
15-3.7. Applications and types of surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-9
15-4. Clay stabilizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-11
15-4.1. Highly charged cations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-11
15-4.2. Quaternary surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-12
15-4.3. Polyamines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-12
15-4.4. Polyquaternary amines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-12
15-4.5. Organosilane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-13
15-5. Mutual solvents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-13
15-5.1. Adsorption of mutual solvents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-14
15-5.2. Chlorination of mutual solvents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-14
15-6. Iron control additives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-14
15-6.1. Sources of iron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-14
15-6.2. Methods of iron control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-15
15-7. Alcohols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-16
15-8. Acetic acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-18
15-9. Organic dispersants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-18
15-10. Organic solvents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-18
15-11. Diversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-18
15-12. Additive compatibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-19
15-13. Facility upsets following acid stimulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-19
15-13.1. Discharge requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-19
15-13.2. Prevention of facility upsets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-20
Chapter 16 Fundamentals of Acid Stimulation

A. Daniel Hill and Robert S. Schechter
16-1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-1
16-2. Acid-mineral interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-2
16-2.1. Acid-mineral reaction stoichiometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-2
16-2.2. Acid-mineral reaction kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-4
Sidebar 16A. Calculating minimum acid volume using dissolving power . . . . . . . . . 16-5
Sidebar 16B. Relative reaction rates of sandstone minerals . . . . . . . . . . . . . . . . . . . . . . 16-8
16-2.3. Precipitation of reaction products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-10
Sidebar 16C. Geochemical model predictions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-11
16-3. Sandstone acidizing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-13
16-3.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-13
16-3.2. Acid selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-13
16-3.3. Sandstone acidizing models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-13
Sidebar 16D. Comparison of acid volumes for radial and perforation flow . . . . . . . . 16-16
16-3.4. Permeability response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-19
16-4. Carbonate acidizing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-19
16-4.1. Distinctive features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-19
16-4.2. Wormholes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-20
16-4.3. Initiation of wormholes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-21
16-4.4. Acidizing experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-23
Sidebar 16E. Optimum injection rate for initiating carbonate treatment . . . . . . . . . . . 16-26
16-4.5. Propagation of wormholes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-27
xvi Contents

Appendix: Advances in understanding and predicting wormhole formation

Christopher N. Fredd. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A16-1
Chapter 17 Carbonate Acidizing Design

J. A. Robert and C. W. Crowe
17-1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-1
17-2. Rock and damage characteristics in carbonate formations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-1
17-2.1. Rock characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-1
17-2.2. Damage characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-2
17-3. Carbonate acidizing with hydrochloric acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-2
17-3.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-2
17-3.2. Historical background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-2
17-3.3. Reactivity of carbonate minerals with hydrochloric acid . . . . . . . . . . . . . . . . . . . . . . . . . 17-3
17-3.4. Acidizing physics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-4
Sidebar 17A. Wormhole initiation and propagation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-6
17-3.5. Application to field design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-7
Sidebar 17B. Acidizing case study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-8
17-4. Other formulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-9
17-4.1. Organic acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-9
17-4.2. Gelled acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-10
17-4.3. Emulsions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-11
17-4.4. Microemulsions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-11
17-4.5. Special treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-12
17-4.6. Self-diverting acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-12
Sidebar 17C. Examples of special treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-13
Sidebar 17D. Placement using self-diverting acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-13
17-5. Treatment design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-14
17-5.1. Candidate selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-14
17-5.2. Pumping schedule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-14
17-5.3. Additives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-14
17-5.4. Placement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-14
17-6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-14
Chapter 18 Sandstone Acidizing

Harry O. McLeod and William David Norman
18-1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-1
18-2. Treating fluids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-1
18-2.1. Hydrochloric acid chemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-2
18-2.2. Chemistry of hydrofluoric acid systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-2
18-3. Solubility of by-products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-4
18-3.1. Calcium fluoride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-5
18-3.2. Alkali fluosilicates and fluoaluminates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-5
18-3.3. Aluminum fluoride and hydroxide. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-5
18-3.4. Ferric complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-5
18-4. Kinetics: factors affecting reaction rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-6
18-4.1. Hydrofluoric acid concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-6
18-4.2. Hydrochloric acid concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-6
18-4.3. Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-7
Reservoir Stimulation xvii

18-4.4. Mineralogical composition and accessible surface area . . . . . . . . . . . . . . . . . . . . . . . . . . 18-7

18-4.5. Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-7
18-5. Hydrofluoric acid reaction modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-7
18-6. Other acidizing formulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-8
18-6.1. Fluoboric acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-8
18-6.2. Sequential mud acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-10
18-6.3. Alcoholic mud acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-11
18-6.4. Mud acid plus aluminum chloride for retardation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-11
18-6.5. Organic mud acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-11
18-6.6. Self-generating mud acid systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-12
18-6.7. Buffer-regulated hydrofluoric acid systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-12
18-7. Damage removal mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-12
18-7.1. Formation response to acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-13
18-7.2. Formation properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-13
18-7.3. Formation brine compatibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-13
18-7.4. Crude oil compatibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-14
18-7.5. Formation mineral compatibility with fluid systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-14
18-7.6. Acid type and concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-16
18-8. Methods of controlling precipitates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-18
18-8.1. Preflush . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-18
18-8.2. Mud acid volume and concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-18
18-8.3. Postflush or overflush . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-18
18-9. Acid treatment design considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-19
18-9.1. Selection of fluid sequence stages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-20
18-9.2. Typical sandstone acid job stages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-20
18-9.3. Tubing pickle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-20
18-9.4. Preflushes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-20
18-9.5. Main fluid stage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-21
18-9.6. Overflush stage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-21
18-9.7. Diversion techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-22
18-9.8. Typical sandstone acid job stages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-22
18-10. Matrix acidizing design guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-23
18-10.1. Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-24
18-10.2. Flowback and cleanup techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-25
18-11. Acid treatment evaluation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-26
18-12. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-27
Chapter 19 Fluid Placement and Pumping Strategy

J. A. Robert and W. R. Rossen
19-1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-1
19-2. Choice of pumping strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-1
19-2.1. Importance of proper placement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-1
19-2.2. Comparison of diversion methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-2
19-2.3. Fluid placement versus injection rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-3
19-2.4. MAPDIR method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-3
19-3. Chemical diverter techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-4
19-3.2. Diverting agent properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-4
19-3.3. Classification of diverting agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-4
19-3.4. Potential problems during diversion treatment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-5
xviii Contents

19-3.5. Laboratory characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-6

19-3.6. Modeling diverter effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-7
19-3.7. Field design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-9
19-4. Foam diversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-10
19-4.2. Foam mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-10
19-4.3. Foam behavior in porous media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-12
19-4.4. Foam diversion experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-14
19-4.5. Modeling and predicting foam diversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-15
19-4.6. Application to field design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-16
19-5. Ball sealers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-18
19-6. Mechanical tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-19
19-7. Horizontal wells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-20
19-7.1. Optimal treatment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-20
19-7.2. Placement techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-22
19-8. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-23
Chapter 20 Matrix Stimulation Treatment Evaluation

Carl T. Montgomery and Michael J. Economides
20-1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-1
20-2. Derivation of bottomhole parameters from wellhead measurements . . . . . . . . . . . . . . . . . . . . . . . 20-1
20-3. Monitoring skin effect evolution during treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-1
20-3.1. McLeod and Coulter technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-1
20-3.2. Paccaloni technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-2
20-4. Prouvost and Economides method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-4
20-4.1. Deriving skin effect during treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-4
20-4.2. Determining reservoir characteristics before treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-4
20-5. Behenna method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-5
20-6. Inverse injectivity diagnostic plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-5
20-7. Limitations of matrix treatment evaluation techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-5
Sidebar 20A. Example calculation of the Prouvost and Economides method . . . . . . . . . . . . . . 20-6
Sidebar 20B. Example application of the Hill and Zhu method . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-7
20-8. Treatment response diagnosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-8
Sidebar 20C. Production indications for matrix stimulation requirements . . . . . . . . . . . . . . . . . 20-10
20-9. Post-treatment evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-11
20-9.1. Return fluid analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-11
20-9.2. Tracer surveys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-11
20-10. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-12
References
Chapters 1–12 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . R-1
Chapters 13–20 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . R-45
Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . N-1
Reservoir Stimulation xix


◆
Preface: Hydraulic Fracturing,
A Technology For All Time
Ahmed S. Abou-Sayed, ADVANTEK International
I was quite pleased when my friend Joseph Ayoub by managing the various chapters from the vast array
at Schlumberger Dowell approached me to write of contributors.
the preface for the third edition of Reservoir A leading contributor, however, to this publica-
Stimulation. It is indeed a pleasure and a compliment tion’s success is Michael Economides, who, over the
to be associated with the distinguished list of indi- last two decades, has contributed substantially to the
viduals contributing to this volume. As an active integration of reservoir performance into well stimu-
member of this close-knit community for the past lation technology and design. He has proficiently
25 years, I have enjoyed working with most of the filled this gap in practice with his thorough work
47 contributing authors. These outstanding scientists related to performance prediction and evaluation.
and engineers have carried the technology of Michael provides the continuous thread that gives
hydraulic fracturing forward to its current high state. the volume its integrated form.
This third edition is an updated classic reference The other leading contributor is Ken Nolte, who
for well stimulation—or in today’s lingo, well per- presents a compelling story that puts forward the
formance enhancement technology—that includes history of hydraulic fracturing technology in the
not only hydraulic fracturing but also an expanded Appendix to Chapter 5. He describes its evolution
treatment of well acidizing and chemical treatment from the late 1940s from his vista, easily scoring a
as well as formation damage migration. Reservoir true bull’s-eye. His towering work since the mid-
Stimulation covers the topics necessary for under- 1970s affords him a unique view of the technological
standing the basis and practical aspects of treatment progress that he helped shape.
design and execution. It addresses the scientific fun- What further insight can I add to the views of
damentals, engineering considerations and opera- these two? I guess you can call it the maverick’s
tional procedures of a job. Pre- and post-treatment view. I will be informal and hope my anecdotal style
analyses, job monitoring and economic elements of will not offend any serious student of the subject.
the various injectivity and productivity enhancement What follows is my view of this fascinating technol-
processes are well presented. ogy, which has renewed itself many times since its
Before I get into a technical discussion of the vol- inception and has contributed substantial financial
ume’s contents, let me share with the reader a bit of benefits to the oil and gas industry.
history and my personal point of view of the future. During the late 1970s, considered the banner years
I am not trying to preempt the excellent contents of fracturing technology advances, there was a say-
compiled by the volume’s editors, Michael Econo- ing often used in jest by most of us working on frac-
mides and Ken Nolte. The two editors have suc- turing:
ceeded in bringing to the reader an integrated
“When everything else fails, frac it.”
account of the objectives, mechanics and implemen-
tation of the well and reservoir aspects of productiv- How true this has been; a lot of “fraccing” was
ity enhancement. Other significant contributions that done for well stimulation in those days and since. We
helped bring Reservoir Stimulation to the reader now speak more appropriately about improved well
came from Joseph Ayoub and Eric Nelson, who pro- performance, downhole flow integrity and enhanced
vided continual technical advice and reviewed the productivity or injectivity. How did we get here?
contents for technical accuracy, and Bob Thrasher, During the late 1940s, fracturing was a timid tech-
who with utter competence and, I must say, sheer nique. In the 1950s, its proliferation took place. In the
patience and persistence pulled this treatise together 1960s, we aimed at understanding what we were
Reservoir Stimulation P-1

doing. The 1970s was a time to expand, quantify and amount of gas trapped in tight sands and unconven-
optimize our procedures. In the 1980s, we worked to tional reservoirs gave impetus to the development of
influence the outcome through monitoring and real- sophisticated fluids (e.g., crosslinked gels), tougher
time engineering. Now in the 1990s, it has been the proppants (e.g., bauxite and lower density ceramics)
time for new frontiers. And in the middle of all that, and large-volume pumping equipment and proppant
we have managed environmental cleanup via injection. handling capacity. More in-depth analysis of fractur-
Since its inception in 1946, hydraulic fracturing ing processes and vastly improved monitoring and
technology has been tapped by the oil and gas indus- analysis techniques were necessary during this period
try to solve a variety of problems and provide answers for optimizing treatments.
to a multitude of difficult issues. These issues cover During the last decade of the twentieth century, the
the areas of enhanced well productivity (by far the technology has pushed forward in more uncharted
most significant utilization), improved well injectivity, waters. Examples of the new directions are horizontal
storage and environmental remediation. and complex well fracturing for reservoir manage-
Three watershed phases have contributed to the ment (part of what Ken Nolte refers to as reservoir
widespread technological advances. The first phase plumbing). Others applications include refracturing,
coincided with the quick development of well stimu- frac and pack treatments for sand control and, finally,
lation techniques in the late 1940s and early 1950s by a class of operations termed environmental fracturing,
the pioneering operating and service companies. The which includes produced water disposal, drilling cut-
second phase of prolific advances occurred in the mid- tings injection and soil reclamation by crush and
1970s, when the national energy policy makers inject technology.
directed their attention and crucial U.S. government The new generation of hydraulic fracturing applica-
funding to developing tight gas sands and unconven- tions is associated with additional operational costs,
tional energy resources. In the early to mid-1980s, the in contrast to revenue-enhancing well stimulation
most significant advances better adapted the technol- work. However, in the final analysis these new appli-
ogy for use in stimulating medium- to high-permeabil- cations are extremely beneficial and essential to
ity (i.e., prolific) reservoirs. The economic payout to achieving industry’s goals of protecting the environ-
the industry has increased rapidly at each phase of ment and moving toward zero emission of explo-
development and with every successful treatment. ration and production waste. They have been easy
Ken Nolte provides the details in the Appendix to extensions of the know-how. That the underlying
Chapter 5. understanding of conventional hydraulic fracturing
In each phase, the producers and service companies operations, and the tools and techniques developed,
have collaborated to provide innovative and cost- has been instantly applicable to these areas is a trib-
effective approaches to problems. During the infancy ute to the innovation and robustness of the hydraulic
of development, slow but consistent progress im- fracturing technologists. They saw the opportunities
proved low-rate well productivity by pumping better as ways to expand their horizons and, let us not for-
materials and using progressively more reliable get, sometimes preserve their jobs.
equipment. Now, what about the present volume? Reservoir
During the 1960s, the emphasis shifted to under- Stimulation consists of 20 chapters with numerous
standing the fracturing process and increasing its sidebars and appendixes prepared by authors of indis-
effectiveness by enhancing the quality of the pumped putably high reputation and respected as experts in
materials (i.e., fluids and proppants) as well as by their fields. The list contains operating company
developing chemical additives, including acid researchers, practitioners, scientists from academic
fracturing. and national laboratory institutions as well as field
In the mid-1970s, massive hydraulic fracturing operations staff. The latter group offers unique insight
replaced an ill-fated attempt to use a nuclear device on how the technologies are operationally imple-
to fracture a tight gas reservoir (the Gas Buggy mented in the real world. As such, the contents of this
experiment, part of Project Plowshare, which was volume provide a balanced view and offer something
designed to develop peaceful uses of nuclear explo- for everyone. This integrated knowledge approach
sive technology). The need for creating the massive makes Reservoir Stimulation a must read for engi-
hydraulic fractures required to unlock the vast neers and geoscientists working with well stimulation
P-2 Preface

technology, from mechanical stimulation (fracturing) the well tubulars, only to become viscous after turning
through chemical treatments (acidizing). the corner of well perforations into the formation.
The reader must view this volume as a confirma- What comes next in this ever-changing world of
tion and accurate account of the larger context of the well stimulation and performance enhancement?
exciting progress that has been made in the field of Current emphasis by the service industry in fluid
hydraulic fracturing and well stimulation. Recent development is on providing cleaner fluids to the user
emphasis has focused on fluid and proppant develop- community. Such fluids maintain the designed frac-
ment, field equipment for mixing and pumping mate- ture conductivity, improve the treatment economics
rials, highly sophisticated (but simple to use) inter- and extend fracturing applications to higher perme-
pretation techniques or monitoring treatment ability reservoirs.
parameters, and computers that monitor, provide Intermediate-density ceramic proppants are stronger
feedback and control the fracture. The available hard- and lighter, so they can be carried farther into the frac-
ware enables real-time redesign during pumping. ture at greater depths. Extensive efforts are directed at
Efforts also have been made, prior to job design obtaining a more thorough understanding of proppant
and execution, to thoroughly characterize reservoir transport mechanisms. Monitoring techniques and
qualities and properties for the optimization of stimu- proppant placement and distribution are conducted
lation treatment design and better economic results. using multiple-isotope radioactive tagging.
Logging tools are used for lithology, permeability, More sophisticated logging tools and interpretation
stress and natural fracture detection. Detection of the algorithms are adding the ability to track the location
created fracture azimuth and length received attention of several pumped stages. This development has
with the development of techniques such as passive improved the understanding of how to design more
borehole seismic methods, crosswell tomography, tilt- effective fracture treatments and has prompted an
meters and hydraulic impedance tests. emphasis on fracture containment.
The myriad techniques available for in-situ stress Pumping and surface handling equipment have
magnitude and azimuth determination include core progressed substantially ahead of the other technolo-
relaxation, differential strain curve analysis, micro- gies, and more advances are under way. The avail-
fracturing and wellbore breakouts. Results of well ability of new-generation blenders, offshore gelling
tests and mini-fracture treatments are used readily and crosslinking of fluids on the fly, and high-pres-
in fracture treatment designs. sure–high-flow-rate pumps and intensifiers provides
The development of accurate downhole pressure the industry with the capacity to execute and control
gauges with digital memory provides a detailed the most complicated fracture. Emphasis must also be
account of fluid pressure at the fracture inlet and directed toward zone isolation techniques and the
assists on-site redesign and post-treatment analysis. hardware to conduct large stimulation jobs in long,
Recent efforts are directed at the development of complex wells.
downhole gauges that transmit pressure, flow rate and As the hardware side of the technology (materials
fluid rheology data in real time. Such gauges are now and equipment) developed at a rapid pace over the
in service in well monitoring and testing applications. last two decades, the software side (modeling, moni-
Simpler techniques, such as using the annulus or a toring and interpretation) also moved forward. The
tubing-based manometer, have been highly success- U.S. government, Gas Research Institute (GRI) and
ful. These applications are limited operationally to academic communities with consulting company sup-
wells with large-diameter casing and tubing and by port are delivering design codes with varying degrees
rig cost. Coiled tubing operations may reduce this of sophistication to the industry. Some of the codes
limitation and expand the application of real-time are field based and used extensively for the optimiza-
downhole pressure monitoring. tion and redesign of fracture treatments. Computer
Fluids now are available with excellent shear sensi- hardware advances and experience-based intelligence
tivity and high-proppant carrying capacities for use at software must provide a window of opportunity for
high temperatures and flow rates. Additives such as broader and more effective use of modeling develop-
borates make it possible to design fluids that have low ments.
frictional or viscosity properties while traveling down

The results of the current software and hardware issues and arrive at satisfactory design and optimiza-
advances are manifested in a growing area of application criteria.
tion, refracturing and simultaneous multiple-job exe- Criteria for the completion, perforation, zonal iso-
cution. In an economically successful effort, the lation and design of job stages for extended reach,
Prudhoe Bay Unit Operating Partners took advantage horizontal, multilateral, openhole and high-rate wells
of these advances in technology to refracture many are needed to help produce a material difference in
wells in Alaska North Slope fields such as Kuparuk. the productivity of fractured or acidized horizontal
Previous jobs, performed in the mid-1980s, were wells. Innovative techniques are required for execut-
based on the standard approaches and supplies of that ing multiple jobs within a horizontal well without
time. The refracturing process systematically and excessive cost while minimizing interference of the
consistently used the latest technology and materials, created fractures. Monitoring techniques for proppant
including progressively bigger proppants, more placement, production logging and downhole pres-
aggressive designs and a better knowledge base of sure profiling also are desirable, if not necessary, for
the reservoir characteristics. The success of refractur- job optimization.
ing treatments between 1988 and 1995 was most sig- A new chapter was opened with the application
nificant in wells where the original design parame- of hydraulic fracturing technology for sand control
ters, materials or execution was below the current in producing soft and unconsolidated formations.
standard. The result has been an additional 50 million This technology involves sequential, uninterrupted
barrels of reserves and a much needed 50,000 B/D of fracturing and gravel-pack operations through down-
oil production from the North Slope. The industry hole gravel-pack hardware. The first application of
will see more refracturing efforts emerge worldwide. the frac and pack technique was implemented in 1985
Fracturing technology has also been applied to hor- for well stimulation at west Sak, a heavy-oil reservoir
izontal and highly deviated wells. The advent of this on the North Slope. The treatment was successful,
type of fracturing posed challenges to operators. but the heavy-oil development was halted owing to
Well-to-fracture conductivity affects treatment pres- the price collapse and drop in global oil economics.
sure, premature screenouts, proppant flowback and Finally, operators recently began development of
well productivity. The pumping of sand slugs, high- deepwater reservoirs using frac and pack technology
overpressure perforating, dynamic formation break- in the Gulf of Mexico. The technique has almost
down and fracture initiation are among the most suc- eliminated the chronic high positive skin factor
cessful methods for improving well connection and (20–30) normally associated with gravel-pack opera-
reducing fracture cornering and tortuosity in the near- tions. Well productivity has increased 2 to 3 times
wellbore region. and skin factors are reduced to almost zero or below,
In economically sensitive fields, the use of S-shaped while complete sand control is maintained. These
wells instead of highly deviated wells is gaining treatments will grow in number and magnitude as
increased acceptance when fracture stimulations are deeper water is conquered and higher well produc-
planned. This approach, used successfully by BP tivity is a must.
Amoco in the North Sea Ravenspurn field, avoids In this context, let us review the contents of this
many of the difficulties of fracturing and producing volume.
deviated wells. It also minimizes the effect of any The third edition begins with an introductory chap-
near-wellbore tortuosity on treatment pressure, prop- ter by Michael Economides and Curtis Boney. They
pant flowback and production efficiency. discuss the inflow performance of regular, horizontal
More complex issues face the horizontal well stim- and complex wells in combination with various reser-
ulation industry. Multiple fracturing perpendicular to voirs. This chapter lays the groundwork for assessing
the borehole provides the most attractive hydrocarbon the need for well stimulation, its applicability and
sweep and reserve recovery. However, fracturing expected rewards. Finally, the authors address reser-
along the borehole is operationally easier but may voir engineering considerations for optimal produc-
have to be executed using multiple treatments tion enhancement strategies.
through perforated intervals in long horizontal wells. In Chapter 2, the other Dr. Economides joins
The industry must sort through these competing Michael. Christine Ehlig-Economides and Michael
P-4 Preface

Economides present the well testing methodology control materials, crosslinking agents, breakers, sur-
and pressure transient analysis used to characterize factants, clay stabilizers and bactericides, and
formations and describe the status of well damage. describe their appropriate uses.
The well-recognized rock mechanics engineers The performance of fracturing materials, a subject
Mark Thiercelin and Jean-Claude Roegiers (known that has seen tremendous advances in the last 20
as “JC”) authored Chapter 3. They present a well- years, is presented in Chapter 8 by Vern Constien,
thought-out treatment of rock mechanics—the char- George Hawkins, Bob Prud’homme and Reinaldo
acterization of the box containing recoverable Navarrete. The chapter outlines techniques for mea-
hydrocarbons. Their work details the theoretical com- suring and designing the necessary rheology for frac-
ponents describing rock behavior and reactions under turing fluids and treatment chemicals. The authors
the loads and stresses generated by E&P operations. also discuss the important topic of propped fracture
The presentation is thorough and on a high funda- conductivity and proppant flowback and the impact
mental level, while providing insight into the practi- of fluid rheology on both. Damage resulting from
cal application of this specialty in a useful and polymer loading is also covered in this chapter.
tractable fashion. Sunil Gulrajani and Ken Nolte discuss the latter’s
Jean Desroches and Tom Bratton describe in favorite topic of fracture evaluation using pressure
Chapter 4 how to use well logs and other geophysical diagnostics in Chapter 9. These techniques, when first
information to obtain pertinent properties of the rock introduced in 1978, provided quantitative tools for
formation for effective treatment design. In addition assessing the nature, extent and containment of the
to the conventional, routine properties such as poros- hydraulic fracture. They subsequently established the
ity, permeability and saturation, they cover the esti- basis for efforts toward real-time diagnostics and con-
mation of pore pressure, formation tests, skin effect trol of the well treatment progress. The authors exam-
and damage extent, in-situ stress and other mechani- ine the mathematical foundation of the diagnostic
cal properties. An interesting treatment of predicting technique, including an accompanying Appendix,
in-situ rock stress and strength from logs is presented. provide field verification examples and present means
In Chapter 5, Mike Smith (the pipe-smoking half of of integrating this approach with other evaluation
the well-known Nolte-Smith duo) and my dear late tools, well measurements and field observations.
friend Jacob Shlyapobersky collaborated to lay down Jack Elbel and Larry Britt collaborated in Chapter
for the reader the basics of hydraulic fracturing. This 10 to present the art and science of fracture treatment
is a pragmatic chapter that serves well as a primer for design. The inclusion of economic analysis as the first
new engineers searching for a quick appreciation of step in the design optimization process, along with the
the factors with an impact on fracture design. Its value authors’ vast experience with treatment design and
is further enhanced by the historical perspective writ- field implementation, offers a unique glimpse of this
ten as the aforementioned Appendix by Ken Nolte. essential process. Staff from the operating divisions
Mark Mack joins Norm Warpinski of Sandia (or asset teams, in today’s lingo) will find this material
National Laboratories in Chapter 6 to provide a com- readily applicable for both hydraulic fracturing and
prehensive treatment of the mechanics of hydraulic acidizing treatments. The subject matter is well orga-
fracturing and discuss the science behind the technol- nized with simple recommendations that can be fol-
ogy. The chapter reflects their massive contributions lowed without great effort.
to the understanding, through extensive field observa- Ernie Brown, Bob Thrasher and Larry Behrmann
tion efforts, of the phenomena and processes involved use Chapter 11 to introduce the reader to the opera-
in hydraulic fracturing. The theoretical and practical tional procedures and completion considerations nec-
knowledge collected throughout their illustrative essary for successful field execution of well stimula-
careers is well represented. tion treatments. Their discussion includes vertical,
Chapter 7 exposes the reader to the materials deviated and S-shaped wells, in addition to wells
pumped in the well during stimulation treatments. with more complex geometries. Factors that have an
Janet Gulbis and Richard Hodge have written a rigor- impact on quality assurance, technologies for treat-
ous, but easily read, discussion of the chemical and ment monitoring and operational integrity during job
rheological aspects of stimulation fluids and prop- execution are all addressed in detail. Field instrumen-
pants. They cover fluid additives, including fluid-loss tation, real-time analysis and recommended remedi-

ation actions during execution are presented in a logi- scientifically rigorous treatment of the subject sum-
cal sequence. The Appendix by Phillip Halleck thor- marizes extensive research results with great clarity.
oughly reviews perforator penetration and its relation The implications of pumping procedures and fluid
to permeability damage. chemistry on well stimulation results are thoroughly
In Chapter 12, Bob Poe and Michael Economides presented for the reader. Acidizing of both sandstone
discuss post-treatment evaluation and well perfor- and carbonate rock is covered. The subjects of worm-
mance prediction. The authors also present applica- hole formation and permeability enhancement in the
tions of fracture-mapping techniques, production log- acid injection path are discussed, with an additional
ging and pressure transient analysis. The chapter treatment of wormhole formation in the Appendix by
concludes this half of the volume on the salient Chris Fredd.
aspects of hydraulic fracturing technology for well The subject of carbonate acidizing is well served
stimulation. in Chapter 17 by Joel Robert and Curtis Crowe. The
The second set of chapters is dedicated to the tech- chapter includes a detailed discussion of the reaction
nology of chemical stimulation and formation damage of hydrochloric acid with carbonate rocks. Placement
mitigation using chemical treatments. It begins with and diversion techniques are highlighted along with
the introduction of matrix treatment (acidizing) by case studies and field illustrations.
chemical means by Ron Thomas and Lee Morgen- Harry McLeod and W. David Norman present an
haler. Both authors are highly experienced in their authoritative treatment of sandstone acidizing in
topic, as reflected by the thoroughness of the informa- Chapter 18. The authors share their thorough knowl-
tion in Chapter 13. The initial overview of using a edge of the theoretical and practical aspects of this
candidate selection advisor walks the reader through subject. Treatment fluids, reaction kinetics and mod-
the process in a simple and able manner. They also eling, and damage removal mechanisms are covered.
provide an interesting historical progression of the The effective discussion of acid staging, diverting
technology. In addition to candidate selection, the and operational procedures will guide the practicing
authors cover fluid selection, treatment placement and engineer in successfully planning job execution.
operational processes as well as treatment economics. Joel Robert and Bill Rossen continue the opera-
The discussion is well supported by the introduction tional theme in Chapter 19. They discuss pumping
of case histories and process flow charts, and the strategies, diversion techniques, foam treatments, ball
theme of advice is clear throughout the chapter. sealers and mechanical tools. Horizontal well acidiz-
Don Hill, George King, Olivier Liétard and Bernard ing is specifically addressed.
Piot discuss formation damage in Chapter 14. Covered In the final chapter, Carl Montgomery and Michael
thoroughly are the origin and diagnosis of formation Economides address matrix stimulation treatment
damage along with suitable treatment strategies for its evaluation. Several analysis methods are presented
mitigation and elimination. Various mechanisms and and critiqued. The authors also provide sharp insights
stages of the formation damage process are presented into the comparative analytical tools used for treat-
in a logical sequence. Damage causes are identified for ment monitoring and diagnostics. Chapter 20 ends
drilling, completion (i.e., casing, cementing, perforat- with a process diagram of job treatment evaluation
ing and gravel packing), stimulation, and production and an assessment of production logging and tracer
and injection operations. methods as evaluation tools.
Chapter 15 covers damage removal materials and So here you have it, a technology for all time and
their impact on well performance and integrity. Syed all places, a money-making endeavor and a challeng-
Ali and Jerry Hinkel explain the association of treating engineering effort that is made easier with refer-
ment materials and damage types for recommended ences such as this volume.
stimulation processes. The discussion is both compre- The third edition of Reservoir Stimulation will be,
hensive and concise, providing the practicing engineer like its two predecessors, on the desk of every prac-
with a useful guide for assessing the impact of chemi- ticing engineer and technologist working on well
cal treatment additives and production chemistry on stimulation, whether the concern is hydraulic fractur-
formation deliverability and well productivity. ing, acidizing or chemical treatment. The tradition
Professors Dan Hill and Bob Schechter explain the established by the previous editions is continued and
fundamentals of acid stimulation in Chapter 16. This further expanded in the present version. The pleasure
P-6 Preface

I had in reading through the vast amount of knowl-

edge imbedded in the 20 chapters more than makes
up for the strange hour at which I am working on
these final thoughts. I hope the reader will find this
volume as stimulating (no pun intended), educational
and useful as I believe it to be and will recognize and
utilize the contributions and know-how of its authors
to achieve his or her goals.
Good reading.


◆
Reservoir Stimulation
in Petroleum Production
Michael J. Economides, University of Houston
Curtis Boney, Schlumberger Dowell
1-1. Introduction reservoir is known as saturated or two phase. Gas

reservoirs exist below the dewpoint pressure.
Reservoir stimulation and artificial lift are the two Petroleum reservoirs also always contain water.
main activities of the production engineer in the The water appears in two forms: within the hydro-
petroleum and related industries. The main purpose carbon zone, comprising the interstitial or connate
of stimulation is to enhance the property value by water saturation Swc, and in underlying water zones,
the faster delivery of the petroleum fluid and/or to which have different magnitudes in different reser-
increase ultimate economic recovery. voirs. The connate water saturation is always pre-
Matrix stimulation and hydraulic fracturing are sent because of surface tension and other adhesion
intended to remedy, or even improve, the natural affinities between water and rock and cannot be
connection of the wellbore with the reservoir, which reduced.
could delay the need for artificial lift. This chapter The underlying water, segregated from hydrocar-
outlines stimulation techniques as tools to help man- bons by gravity, forms a gas-water or oil-water
age and optimize reservoir development. contact that is not sharp and may traverse several
Understanding stimulation requires understanding feet of formation because of capillary pressure
the fundamental issues of petroleum production and effects. The water may intrude into the hydrocarbon
the position and applicability of the process. zone as a result of perturbations made during petro-
leum production.
The ideas of porosity and connate water satura-
1-1.1. Petroleum production tion are coupled with the areal extent of a reservoir
Petroleum reservoirs are found in geologic forma- A and the reservoir net thickness h to provide the
tions containing porous rock. Porosity φ is the frac- hydrocarbon volume, referred to as initial-oil-in-
tion of the rock volume describing the maximum place or initial-gas-in-place:
possible fluid volume that can be stored. VHC = Ahφ(1 – Swc). (1-1)
Petroleum, often referred to in the vernacular
as oil or gas depending on the in-situ conditions of Because oil and gas production rates in the petro-
pressure and temperature, is a mixture of hydrocar- leum industry are accounted in standard-condition
bons ranging from the simplest, methane, to long- volumes (e.g., standard pressure psc = 14.7 psi or
chain hydrocarbons or complex aromatics of consid- 1 atm [1 × 105 Pa] and standard temperature Tsc =
erable molecular weights. Crude oils are frequently 60°F [15.6°C]), the right-hand side of Eq. 1-1 is
referred to as paraffinic or asphaltenic, depending on divided by the formation volume factor for oil Bo
the dominant presence of compounds within those or for gas Bg.
hydrocarbon families. Wells drilled to access petroleum formations
The phase behavior of petroleum hydrocarbons is cause a pressure gradient between the reservoir
usually greatly simplified, separating compounds in pressure and that at the bottom of the well. During
the gaseous phase from those in the liquid phase into production or injection the pressure gradient forces
two pseudocompounds (oil and gas). The bubble- fluids to flow through the porous medium. Central
point pressure pb of the mixture becomes important. to this flow is the permeability k, a concept first
If the reservoir pressure is greater than this value, introduced by Darcy (1856) that led to the well-
the fluid is referred to as undersaturated. If the reser- known Darcy’s law. This law suggests that the flow
voir pressure is below pb, free gas will form, and the rate q is proportional to the pressure gradient ∆p:
Reservoir Stimulation 1-1

q ∝ k∆p. (1-2) The skin effect s, which is analogous to the film

coefficient in heat transmission, was introduced by
The fluid viscosity µ also enters the relationship,
Van Everdingen and Hurst (1949) to account for
and for radial flow through an area 2πrh, Eq. 1-2
these phenomena. Fundamentally a dimensionless
becomes
number, it describes a zone of infinitesimal extent
qµ r
p – pwf = ln , (1-3) that causes a steady-state pressure difference, conve-
2 πkh rw niently defined as
qµ
where pwf and rw are the bottomhole flowing pressure ∆ps = s. (1-5)
and wellbore radius, respectively. 2 πkh
Equation 1-3 is also well known and forms the
basis to quantify the production (or injection) of flu- Adding Eqs. 1-3 and 1-5 results in
ids through vertical wells from porous media. It is
qµ  r 
perhaps the most important relationship in petroleum p – pwf =  ln + s , (1-6)
engineering. 2 πkh  rw 
The permeability k used in Eq. 1-3 is absolute,
where the pwf in Eq. 1-6 is different from that in Eq.
implying only one fluid inhabiting and the same fluid
1-3. A positive skin effect requires a lower pwf,
flowing through the porous medium. This is, of
whereas a negative skin effect allows a higher value
course, never true for oil or gas flow. In the presence
for a constant rate q. For production or injection, a
of another fluid, such as connate water, an effective
large positive skin effect is detrimental; a negative
permeability is in force, which is usually symbolized
skin effect is beneficial.
by a subscript (e.g., ko) and always implied. The
Two extensions of Eq. 1-6 are the concepts of
effective permeability in a reservoir is smaller than
effective wellbore radius and the important produc-
the absolute permeability, which may be measured
tivity (or injectivity) index.
in a laboratory on cores extracted from the reservoir.
With simple rearrangement and employing a basic
If more than one fluid flows, relative permeabili-
property of logarithms, Eq. 1-6 yields
ties that are functions of the fluid saturations are in
effect: qµ  r 
p – pwf = ln – s  . (1-7)
k k k 2 πkh  rw e 
kro = o , krw = w and krg = g , (1-4) The
k k k
expression rw e–s is the effective wellbore radius,
where kro, krw and krg are the relative permeabilities denoted as rw´. A positive skin effect causes the
and ko, kw and kg are the effective permeabilities of effective wellbore radius to be smaller than the actu-
oil, water and gas, respectively. al, whereas a negative skin effect has the opposite
Equation 1-3, in conjunction with appropriate dif- result.
ferential equations and initial and boundary condi- A second rearrangement yields
tions, is used to construct models describing petro- q 2πkh
=
[ ]
. (1-8)
leum production for different radial geometries. p – pwf µ ln(re / rw ) + s
These include steady state, where the outer reservoir
pressure pe is constant at the reservoir radius re; The left-hand side of Eq. 1-8 is the well productiv-
pseudosteady state, where no flow is allowed at the ity (or injectivity for pwf > p) index.
outer boundary (q = 0 at re); and infinite acting, The entire edifice of petroleum production engi-
where no boundary effects are felt. Well-known neering can be understood with this relationship.
expressions for these production modes are pre- First, a higher kh product, which is characteristic
sented in the next section. of particular reservoirs, has a profound impact. The
Regardless of the mode of reservoir flow, the near- current state of worldwide petroleum production and
well zone may be subjected to an additional pressure the relative contributions from various petroleum-
difference caused by a variety of reasons, which producing provinces and countries relate intimately
alters the radial (and horizontal) flow converging with the kh products of the reservoirs under exploita-
into the well. tion. They can range by several orders of magnitude.
1-2 Reservoir Stimulation in Petroleum Production

There is virtually nothing that a petroleum engineer Table 1-1. Unit conversions for petroleum
can do to substantially alter this situation. Mature production engineering.
petroleum provinces imply that following the
Variable Oilfield SI Units Conversion
exploitation of far more prolific zones is the Units (multiply
exploitation of increasingly lackluster zones with oilfield
small kh values, which characterize more recently units)
discovered formations. Area, A ft2 m2 9.29 × 10–2

A second element of mature fields is reservoir Compressibility, ct psi–1 Pa–1 1.45 × 10–4
pressure depletion, the effect of which can be seen Length ft m 3.05 × 10–1
readily from Eq. 1-8. Although the right-hand side Permeability, k md m2 9.9 × 10–16
of the equation may be constant, even with a high Pressure, p psi Pa 6.9 × 103
kh, the production rate q will diminish if p – pwf is Rate (oil), q STB/D m3/s 1.84 × 10–6
reduced. For a constant pwf, reduction in the reservoir Rate (gas), q Mscf/D m3/s 3.28 × 10–4
Viscosity, µ cp Pa-s 1 × 10–3
pressure p has this effect.
The role of the petroleum production engineer,
who must deal with the unalterable kh and pressure
of a given reservoir, is to maximize the productivity 6000
index by reducing the skin effect and/or the required
bottomhole flowing pressure to lift the fluids to the 5000
top. Maximizing the productivity index by reducing
the skin effect is central to the purpose of this vol- Bottomhole flowing pressure, pwf
4000
ume and constitutes the notion of stimulation; reduc-
ing the bottomhole flowing pressure leads to artificial
lift (both gas and pump assisted). Finally, the bot- 3000
tomhole flowing pressure may have an allowable
lower limit to prevent or retard undesirable phenom-
2000
ena such as sand production and gas or water coning.
1000
1-1.2. Units
The traditional petroleum engineering oilfield units 0
0 1000 2000 3000 4000 5000 6000
are not consistent, and thus, most equations that are
Flow rate, q
cast in these units require conversion constants. For
example, 1/(2π) in Eq. 1-3 is appropriate if SI units Figure 1-1. The inflow performance relationship relates
are used, but must be replaced by the familiar value the production rate to the bottomhole flowing pressure.
of 141.2 if q is in STB/D (which must be multiplied
also by the formation volume factor B in RB/STB);
µ is in cp; h, r and rw are in ft; and p and pwf are in 1-2. Inflow performance
psi. Table 1-1 contains unit conversion factors for the
The well production or injection rate is related to the
typical production engineering variables.
bottomhole flowing pressure by the inflow perfor-
For unit conversions there are two possibilities.
mance relationship (IPR). A standard in petroleum
Either all variables are converted and then two ver-
production, IPR is plotted always as shown in Fig. 1-1.
sions of the equation exist (one in oilfield and a sec-
Depending on the boundary effects of the well
ond in SI units), or one equation is provided and the
drainage, IPR values for steady-state, pseudosteady-
result is converted. In this volume the second option
state and transient conditions can be developed read-
is adopted. Generally, the equations are in the tradi-
ily. In the following sections, the relationships for
tional oilfield units predominant in the literature.
the three main flow mechanisms are presented first
for vertical and then for horizontal wells. The

expressions, almost all of which are in wide use, are The usefulness of the Vogel approximation is that
in oilfield units. A complete outline of their develop- it can be used to predict the oil production rate when
ment is in Economides et al. (1994). free gas flows (or is present) although only oil prop-
erties are employed. For steady state, Eqs. 1-12 and
1-13 can be combined with Eq. 1-9:
1-2.1. IPR for steady state
 pwf  pwf  
2
Equation 1-6 can be converted readily to a steady- ko hpe 1 − 0.2 − 0.8  

state expression by simply substituting p with pe and  pe  pe  
qo = (1-14)
[ ]
.
r with re. Thus, in oilfield units and with simple 254.2 Bo µ o ln(re rw ) + s
rearrangements, the IPR for oil is
q=
(
kh pe – pwf ) The subscript o is added here to emphasize the
[ ]
. (1-9) point that oil properties are used. The subscript is
141.2 Bµ ln(re / rw ) + s frequently omitted, although it is implied. Although
A plot of pwf versus q forms a straight line, the ver- neither Eq. 1-11 (for gas) nor Eq. 1-14 (for two-
tical intercept is pe, and the flow rate at the horizontal phase flow) provides a straight-line IPR, all steady-
intercept (i.e., at pwf = 0) is known as the absolute state IPRs provide a stationary picture of well deliv-
open-flow potential. The slope is, of course, constant erability. An interesting group of IPR curves for oil
throughout the production history of the well, assum- is derived from a parametric study for different skin
ing single-phase flow, and its value is exactly equal effects, as shown in Fig. 1-2.
to the reciprocal of the productivity index.
For gas, the analogous expression is approximately 6000
q=
(
kh pe2 – pwf2 )
[ ]
, (1-10) 5000
1424µ ZT ln(re / rw ) + s
Bottomhole flowing pressure, pwf
—
where Z is the average gas deviation factor (from 4000
–
ideality), T is the absolute temperature in °R, and µ
is the average viscosity.
3000
Equation 1-10 has a more appropriate form using
the Al-Hussainy and Ramey (1966) real-gas pseudo-
pressure function, which eliminates the need to aver- 2000
age µ and Z:
q=
[ ( )] .
kh m( pe ) – m pwf
1000
s = –5
1424T [ln(r / r ) + s]
(1-11)
e w
0 s = 20 s = 10 s=0
0 1000 2000 3000 4000 5000 6000
For two-phase flow, production engineers have
Flow rate, q
used several approximations, one of which is the
Vogel (1968) correlation, which generally can be Figure 1-2. Variation of the steady-state IPR of an oil well
written as for different skin effects.
2
qo p p 
= 1 − 0.2 wf − 0.8 wf  (1-12)
qo,max p  p  • Example of steady-state IPR: skin effect variation
Suppose that k = 5 md, h = 75 ft, pe = 5000 psi,
AOFP
qo,max = , (1-13) B = 1.1 RB/STB, µ = 0.7 cp, re = 1500 ft and
1.8 rw = 0.328 ft. Develop a family of IPR curves
where qo is the oil production rate, qo,max is the maxi- for an undersaturated oil reservoir for skin
mum possible oil rate with two-phase flow, and AOFP effects from –5 to 20.
is the absolute open-flow potential of single-phase oil
flow.

Solution
6000
Using Eq. 1-9 and substituting for the given
variables:
5000
 5000 − pwf 
q = 3.45

.
 8.43 + s  4000
Figure 1-2 is a plot of the family of IPR

3000
curves. For a reasonable pwf = 2000, the flow p
=
p
=
p
=
p
=
p
=
30 3 4 45 5
rates at s = 20, 0 and –5 are approximately 365, 00 500 000 00 000
1230 and 3000 STB/D, respectively, showing the 2000

extraordinary impact of a negative skin effect.
1000
1-2.2. IPR for pseudosteady state

0
At first glance, the expression for pseudosteady-state 0 500 1000 1500 2000 2500
flow for oil, Flow rate, q
q=
(
kh p − pwf ) Figure 1-3. Variation of the pseudosteady-state IPR for
[ ]
,
141.2 Bµ ln(0.472 re rw ) + s (1-15) an oil well for declining reservoir pressure.
appears to have little difference from the expression • Example of pseudosteady-state IPR: effect of
for steady state (Eq. 1-9). However, the difference is average reservoir pressure
significant. Equation 1-15 is given in terms of the This example repeats the preceding “Example
–
average reservoir pressure p, which is not constant of steady-state IPR: skin effect variation” (page
but, instead, integrally connected with reservoir –
1-4) for s = 0 but allows p to vary from 5000 to
depletion. 3000 in increments of 500 psi.
Material-balance calculations such as the ones Solution
introduced by Havlena and Odeh (1963) are required
to relate the average reservoir pressure with time and Using Eq. 1-15 and substituting for the given
the underground withdrawal of fluids. variables (including s = 0):
–
Interestingly, the productivity index for a given q = 0.45( p – pwf).
skin effect is constant although the production rate
–
declines because p declines. To stem the decline, the In the Fig. 1-3 family of IPR curves for differ-
–
production engineer can adjust the pwf, and thus, arti- ent values of p, the curves are parallel, reflecting
the constant productivity index. (This type of
ficial lift becomes an important present and future
construction assumes that oil remains undersatu-
consideration in well management. Successive IPR
rated throughout; i.e., above the bubblepoint
curves for a well producing at pseudosteady state at
pressure.)
different times in the life of the well and the result-
–
ing different values of p are shown in Fig. 1-3.
The analogous pseudosteady-state expressions for
1-2.3. IPR for transient (or infinite-
gas and two-phase production are
acting) flow
q=
[
kh m( p ) − m pwf( )] The convection-diffusion partial differential equa-
[
1424T ln(0.472 re rw ) + s ] (1-16) tion, describing radial flow in a porous medium, is
 p p  
2
∂ 2 p 1 ∂p φµct ∂p
khp 1 − 0.2 wf − 0.8 wf   + = , (1-18)
 p  p   ∂r 2 r ∂r k ∂t
q=
[ ]
.
254.2 Bµ ln(0.472 re rw ) + s
(1-17)

where ct is the total system compressibility, p is pres- Transient IPR curves can be generated for each
sure, t is time, and r is radial distance. This equation, instant in time as shown in Fig. 1-4.
in wide use in many other engineering fields, pro- • Example of transient IPR
vides a well-known solution for an infinite-acting
Using the variables of the previous two exam-
reservoir producing at constant rate at the well.
ples and φ = 0.25, ct = 10–5 psi–1 and pi = 5000
Using dimensionless variables (for oil, in oilfield
psi, develop transient IPR curves for t = 3, 6 and
units) for pressure and time, respectively:
36 months. The time in Eq. 1-22 must be
kh∆p entered in hours. Assume s = 0.
pD = (1-19)
141.2 qBµ Solution
Using Eq. 1-22 and substituting for the given
0.000264 kt variables:
tD = . (1-20)
φµct rw2
q=
(
3 5000 − pwf ).
For r = rw (i.e., at the well) a useful approximate log t + 4.19
form of the solution in dimensionless form is simply Figure 1-4 is a graph of the three transient
IPRs. The expected flow rate declines for a con-
(ln t D + 0.8091) .
1
pD = (1-21) stant pwf = 2000. The flow rates at 3, 6 and 36
2
months are 1200, 1150 and 1050 STB/D, respec-
Equation 1-21 provided the basis of both the fore- tively. The 36-month calculation is unrealistic
cast of transient well performance and the Horner because it is unlikely that a well would remain
(1951) analysis, which is one of the mainstays of infinite acting for such long period of time.
pressure transient analysis presented in Chapter 2. –
Thus, a pseudosteady-state IPR with a p inter-
Although Eq. 1-21 describes the pressure transients section at a point below pi is most likely in
under constant rate, an exact analog for constant pres- effect at that time.
sure exists. In that solution, pD is replaced simply by
the reciprocal of the dimensionless rate 1⁄qD.
The dimensioned and rearranged form of Eq. 1-21, 1-2.4. Horizontal well production
after substitution of the natural log by the log base
10, is Since the mid-1980s horizontal wells have prolifer-
ated, and although estimates vary, their share in the
q=
(
kh pi − pwf  ) k 
−1
 log t + log − 3.23 ,

162.6 Bµ  φµct rw
2
 6000
t = 36 months
t = 6 months
(1-22) t = 3 months
5000
where pi is the initial reservoir pressure. The skin
effect can be added inside the second set of paren-

4000
theses as 0.87s.
As previously done for the pseudosteady-state
IPR, gas and two-phase analogs can be written: 3000
q=
[ ( )]  log t + log
kh m( pi ) − m pwf

k 
− 3.23
−1
2000
1638T  φµct rw
2

1000
(1-23)
 pi  pi  
2
khpi 1 − −  
0
 pwf  pwf   0 500 1000 1500 2000 2500
q= . (1-24) Flow rate, q
 k 
254.2 Bµ log t + log − 3.23 Figure 1-4. Transient IPR curves for an oil well.
 φµct rw2 

production of hydrocarbons will probably reach 50% parentheses in the denominator and multiplied by the
or more. scaled aspect ratio Ianih/L.
A model for a constant-pressure ellipse at steady- For gas and two-phase flow, Eq. 1-25 can be
state production was introduced by Joshi (1988) and adjusted readily by the transformations (compared
augmented by Economides et al. (1991): with Eq. 1-9) shown in Eqs. 1-11 and 1-14.
For pseudosteady state, a generalized horizontal
q=
(
kH h pe − pwf ) ,
well production model, accounting for any position-
  
a + a 2 − ( L 2)   I h ing of a well laterally and vertically within a
2
     ani Iani h 
141.2 Bµ ln  + drainage, was presented by Economides et al.
[ ]
ln 
  L2  L rw ( Iani + 1)  (1996). The basic model in Fig. 1-5 has reservoir
   
dimensions xe, ye and h, horizontal well length L and
(1-25) an angle ϕ between the well projection on the hori-
zontal plane and xe.
where L is the horizontal well length and kH is the
horizontal permeability. The latter is the same as that
used in all vertical well deliverability relationships.
The subscript distinguishes it from the vertical per- z
meability kV, which is related to the index of the hor- θ
izontal-to-vertical permeability anisotropy Iani: h
L
kH
Iani = . (1-26) ϕ y
kV
x
The large half-axis a of the horizontal drainage xe
ellipse formed around a horizontal well within an ye
equivalent radius reH is
1/ 2
L   reH   
4 1/ 2
 Figure 1-5. Generalized well model for production from an
a = 0.5 + 0.25 +     , (1-27) arbitrarily oriented well in an arbitrarily shaped reservoir
2 
  L 2    (Economides et al., 1996).
 
where reH is the equivalent radius in a presumed cir-
cular shape of a given drainage area. Equation 1-27 The solution is general. First, the pseudosteady-
transforms it into an elliptical shape. state productivity index J is used:
Equation 1-25 can be used readily to develop a
q kxe
horizontal well IPR and a horizontal well produc- J= = , (1-28)
tivity index. p − pwf  xe 
887.22 Bµ pD +
A comparison between horizontal (Eq. 1-25) and  2 πL ∑ s 
vertical (Eq. 1-9) productivity indexes in the same
formation is an essential step to evaluate the attrac- where the reservoir permeability k is assumed to be
tiveness or lack thereof of a horizontal well of a isotropic throughout (it is adjusted later) and xe is the
given length over a vertical well. Such comparison well drainage dimension. The constant allows the
generally suggests that in thick reservoirs (e.g., h > use of oilfield units; the productivity index is in
100 ft) the index of anisotropy becomes important. STB/D/psi. The summation of the skin effects Σs
The smaller its value (i.e., the larger the vertical per- accounts for all damage and mechanical skin effects.
meability), the more attractive a horizontal well is Finally, the dimensionless pressure is
relative to a vertical well. For thinner formations
(e.g., h < 50 ft), the requirements for good vertical xe CH x
permeability relax. pD = + e sx . (1-29)
4 πh 2 π L
A skin effect can also be added to the horizontal
well deliverability of Eq. 1-25, inside the large

Equation 1-29 decomposes a three-dimensional

(3D) problem into one two-dimensional term and 5
rw = 0.4 ft
L = 4000 ft
one one-dimensional term. The first term on the Well is centered
right-hand side accounts for horizontal positioning 4

effects, with CH as a shape factor. The second term L = 200 ft
accounts for both the reservoir thickness (causing a
distortion of the flow lines) and the additional effects 3
sx
from vertical eccentricity in the case that the well is
not positioned at the vertical middle of the reservoir.
2
The vertical effects skin effect sx is (after Kuchuk
et al., 1988)
1
 h  h 0 50 100 150 200 250 300
sx = ln – + se , (1-30) h (ft)
 2 πrw  6 L
Figure 1-6. Vertical effects skin effect for a horizontal well
where se is the vertical eccentricity skin: (Economides et al., 1996).
h  2 zw 1  2 zw  1  πz 
2
se =  − −  − ln sin w   , 1.2
L h 2 h  2   h 
1.0 zw /h = 0.1
(1-31) 0.8 h
zw
where zw is the elevation from the bottom of the
0.6
se
reservoir. For a well at the vertical middle, se = 0.

zw /h = 0.2
• Example calculation of sx for two thicknesses 0.4
Assume that L = 2000 ft and rw = 0.328 ft.

0.2 zw /h = 0.3
Calculate sx for h = 50 ft and h = 200 ft.
zw /h = 0.4
Solution 0
0 0.2 0.4 0.6 0.8 1.0 1.2
Using Eq. 1-30 for h = 50 ft: h/L
50 50
sx = ln − = 3.2 . Figure 1-7. Vertical eccentricity skin effect (Economides
(2)(3.14)(0.328) (6)(2000) et al., 1996).
sented by Economides et al. (1996) is general

For h = 200 ft, sx = 4.6. This calculation sug-
and a computer program is available, the library
gests that for thicker reservoirs the distortion of
of shape factors in Table 1-2 is useful for quick
the flowlines has relatively more severe detri-
approximations (in the style of the classic Dietz
mental effects.
[1965] factors for vertical wells). Multiple hori-
Figure 1-6 provides values for sx for a range
zontal well configurations are also included.
of reservoir thicknesses and a centered well
(rw = 0.4 ft). • Example calculation of horizontal well produc-
For the case of a vertically eccentered well, tivity index: comparison with a vertical well
Fig. 1-7 provides values for se for various levels Assume that L = 2000 ft, xe = ye = 2700 ft,
of eccentricity. The values in Fig. 1-7 are the h = 200 ft, rw = 0.328 ft, B = 1 RB/STB and
same for symmetrical eccentricity; i.e., se is the µ = 1 cp. The well is in the vertical middle (i.e.,
same for zw /h = 0.1 and 0.9. At zw /h = 0.5, se = 0). Permeability k = 10 md. For this exam-
se = 0, as expected. ple, the productivity index is calculated for an
To account for the position of the well in the isotropic reservoir. However, the permeability of
horizontal plane, a series of shape factors is pre- most reservoirs is not isotropic between the ver-
sented in Table 1-2. Although the solution pre- tical and horizontal planes, resulting in a consid-

Table 1-2. Shape factors for well productivity (Economides et al., 1996).
L /xe CH CH
xe = 4ye 0.25 3.77 xe = ye Ly = 2Lx 1.10
0.5 2.09 Lx /xe = 0.4 Ly = Lx 1.88
0.75 1.00 Ly = 0.5Lx 2.52
1 0.26
xe = 2ye 0.25 3.19 xe = ye Ly = 2Lx 0.79

0.5 1.80 Lx /xe = 0.4 Ly = Lx 1.51
0.75 1.02 Ly = 0.5Lx 2.04
1 0.52
xe = ye 0.25 3.55 xe = ye Ly = 2Lx 0.66

0.4 2.64 Lx /xe = 0.4 Ly = Lx 1.33
0.5 2.21 Ly = 0.5Lx 1.89
0.75 1.49
1 1.04
2xe = ye 0.25 4.59 xe = ye Ly = 2Lx 0.59

0.5 3.26 Lx /xe = 0.4 Ly = Lx 1.22
0.75 2.53 Ly = 0.5Lx 1.79
1 2.09
4xe = ye 0.25 6.69

0.5 5.35
0.75 4.63
1 4.18
xe = ye 0.25 2.77
0.5 1.47
0.75 0.81
1 0.46
xe = ye 0.25 2.66
0.5 1.36
0.75 0.69
1 0.32
xe = ye 0 1.49
L /xe = 0.75 30 1.48
ϕ 45 1.48
75 1.49
90 1.49
erable reduction in the productivity index, as The productivity index of a vertical well in the
shown in the next section. same formation, under pseudosteady-state condi-
Solution tions and assuming that the well is in the center
of the square reservoir, is
From the “Example calculation of sx for two
kh
thicknesses” (page 1-8), sx = 4.6, and from Table JV = .
1-2 for xe = ye and L/xe = 2000/2700 ≈ 0.75, 141.2 Bµ ln(0.472 re / rw )
CH = 1.49.
The drainage area is 2700 × 2700 ft, resulting
Using Eq. 1-29:
in re = 1520 ft. Thus,
(2700)(1.49) (2700)(4.6)
pD = + = 2.59 , (10)(200)
(4)(3.14)(200) (2)(3.14)(2000) JV =
(141.2)(1)(1) ln[(0.472)(1520) / (0.328)]
and using Eq. 1-28: = 1.84 STB / D / psi .
The productivity index ratio between a hori-

(10)(2700)
JH = = 11.7 STB / D / psi . zontal and a vertical well in this permeability-
( .22)(1)(1)(2.59)
887 isotropic formation is 11.7/1.84 = 6.4.

1-2.5. Permeability anisotropy

(k k )
1/ 2 (1-35)
α=
x y
From dealing with vertical wells, petroleum engi- kz
neers learned to generally ignore the concept of per-
meability anisotropy and refer to reservoirs as having 1/ 2
permeabilities equal to 0.1, 3, 100 md, etc., as if per-  k k  (1-36)
β =  y cos 2 ϕ + x sin 2 ϕ
meability were a scalar quantity.  kx ky 
Although it has been known for a long time that
permeability has different values in different direc- (1-37)
tions (i.e., it is a vector) and although the impact of k y kz
x′ = x
such anisotropy is recognized in waterflooding and k
even in the spacing of wells, for production from a
(1-38)
single vertical well it is of little concern. Muskat kk
(1937), in one of his many early contributions, sug- y′ = y x z
k
gested that the permeability affecting vertical well
production is (1-39)
kx ky
z′ = z
k = kH = kx ky , (1-32) k
(1-40)
–
where k is the average permeability, which for a ver- k = 3 k x k y kz .
tical well is equal to the average horizontal perme-
–
ability k H, and kx and ky are the permeabilities in the
x and y directions, respectively. • Example of horizontal well productivity index in
Although the “average” permeability in Eq. 1-32 an anisotropic reservoir
could equal 10 md, this value could result because Repeat the calculations in “Example calculation
the permeabilities in the x direction and y direction of horizontal well productivity index: compari-
are both equal to 10 md or because kx = 100 md and son with a vertical well” (page 1-8) but with
ky = 1 md. Horizontal-to-horizontal permeability kx = 20 md, ky = 5 md (the average horizontal
anisotropy of such magnitude is rare. However, per- permeability is still 10 md) and kz = 1 md.
meability anisotropies in the horizontal plane of 3:1 Assume that the well is parallel to the xe bound-
and higher are common (Warpinski, 1991). Logi- ary; i.e., the angle ϕ = 0.
cally, a horizontal well drilled normal to the maxi- Solution
mum rather than the minimum permeability should From Eqs. 1-35 and 1-36, α = 3.16 and β =
be a better producer. 0.707, respectively. The horizontal well length
Suppose all permeabilities are known. Then the is then adjusted using Eq. 1-33 and becomes
horizontal well length, wellbore radius and reservoir 964 ft. The wellbore radius is adjusted using
dimensions can be adjusted. These adjusted vari- Eq. 1-34 and becomes 0.511 ft. The reservoir
ables, presented by Besson (1990), can be used dimensions xe, ye and h are adjusted using Eqs.
instead of the true variables in predicting well per- 1-37 through 1-39 and become 1304, 2608 and
formance with the model in Section 1-2.4: 432 ft, respectively.
The vertical effect skin effect from Eq. 1-30 is
Length: L ′ = Lα –1/ 3β (1-33) 4.85. The adjusted reservoir dimensions become
2xe = ye. The adjusted penetration ratio L/xe
remains the same (0.75). Thus, from Table 1-2
the shape factor is 2.53.
α 2/3  1 
Wellbore radius: rw′ = rw  + 1 , (1-34) Using Eq. 1-29 for dimensionless pressure and
2  αβ  substituting with the adjusted variables:
where (1304)(2.53) (1304)(4.85)
pD = + = 1.65 ,
(4)(3.14)(432) (2)(3.14)(964)

and using Eq. 1-28, the productivity index

becomes rs
(4.63)(1304)
J= = 4.12 STB / D / psi ,
(887.22)(1)(1)(1.65)
k
ks pe
representing a 65% reduction from the value h rw
of 11.7 STB/D/psi calculated in the preceding

“Example calculation of horizontal well produc-
tivity index: comparison with a vertical well” for
the isotropic case.
re
Zone of altered
1-3. Alterations in the near- permeability
wellbore zone Figure 1-8. Zone of altered permeability k s near a well.
The skin effect s is intended to describe alterations
in the near-wellbore zone. One of the most common that ps is the pressure at the outer boundary of the
problems is damage to the permeability that can be altered zone, from Eq. 1-9 the undamaged relation is
caused by practically any petroleum engineering
activity, from drilling to well completions to stimula-
tion itself. As mentioned in Section 1-1.1, the skin q=
(
kh ps − pwf , ideal
,
) (1-41)
r
effect is a dimensionless number that can be 141.2 Bµ ln s
rw
obtained from a well test, as explained in detail
in Chapter 2. and if damaged,
The nature of radial flow is that the pressure dif-
ference in the reservoir increases with the logarithm
q=
(
ks h ps − pwf , real
,
) (1-42)
of distance; i.e., the same pressure is consumed with- r
in the first foot as within the next ten, hundred, etc. 141.2 Bµ ln s
rw
If the permeability of the near-wellbore zone is
reduced significantly it is entirely conceivable that using the respective values of undamaged ideal and
the largest portion of the total pressure gradient may damaged real bottomhole flowing pressure.
be consumed within the very near wellbore zone. Equations 1-41 and 1-42 may be combined with the
Similarly, recovering or even improving this perme- definition of skin effect and the obvious relationship
ability may lead to a considerable improvement in
the well production or injection. This is the role of ∆ps = pwf , ideal − pwf , real (1-43)
matrix stimulation.
to obtain
1-3.1. Skin analysis 141.2 qBµ  rs   1 1 

∆ps = ln   −  . (1-44)
Figure 1-8 describes the areas of interest in a well h  rw   ks k 
with an altered zone near the wellbore. Whereas k
is the “undisturbed” reservoir permeability, ks is the
permeability of this altered zone. Equations 1-44 and 1-5 can then be combined:
The Van Everdingen and Hurst (1949) skin effect
has been defined as causing a steady-state pressure k  r
difference (Eq. 1-5). Skin effect is mathematically s =  − 1 ln s , (1-45)
 ks  rw
dimensionless. However, as shown in Fig. 1-8, it
reflects the permeability ks at a distance rs. A rela- which is the sought relationship. This is the well-
tionship among the skin effect, reduced permeability known Hawkins (1956) formula.
and altered zone radius may be extracted. Assuming

Equation 1-45 leads to one of the best known con- could be altered by hydraulic fracturing treatments.
cepts in production engineering. If ks < k, the well is The other three terms are the common skin factors.
damaged and s > 0; conversely, if ks > k, then s < 0 The first is the skin effect caused by partial comple-
and the well is stimulated. For s = 0, the near-well- tion and slant. It has been well documented by
bore permeability is equal to the original reservoir Cinco-Ley et al. (1975a). The second term represents
permeability. the skin effect resulting from perforations, as
Certain well logs may enable calculation of the described by Harris (1966) and expounded upon by
damaged radius, whereas pressure transient analysis Karakas and Tariq (1988). The third term refers to
may provide the skin effect and reservoir permeabil- the damage skin effect.
ity. Equation 1-45 may then provide the value of the Obviously, it is of extreme importance to quantify
altered permeability ks. the components of the skin effect to evaluate the
Frick and Economides (1993) postulated that effectiveness of stimulation treatments. In fact, the
in the absence of production log measurements, an pseudoskin effects can overwhelm the skin effect
elliptical cone is a more plausible shape of damage caused by damage. It is not inconceivable to obtain
distribution along a horizontal well. A skin effect skin effects after matrix stimulation that are extreme-
expression, analogous to the Hawkins formula, was ly large. This may be attributed to the usually irre-
developed: ducible configuration skin factors.
k   1 4  asH , max asH , max  

2
seq =  − 1 ln   2 + r + 1  , 1-3.3. Skin effect caused by partial

 ks   Iani + 1 3  rw 
 w
 completion and slant
(1-46) Figure 1-9 is relevant to Cinco-Ley et al.’s (1975a)
development. Table 1-3 presents the pseudoskin fac-
where Iani is the index of anisotropy and asH,max is tors caused by partial penetration and slant. To use
the horizontal axis of the maximum ellipse, normal them, it is necessary to evaluate several dimension-
to the well trajectory. The maximum penetration of less groups:
damage is near the vertical section of the well. The
shape of the elliptical cross section depends greatly Completion thickness hwD = hw /rw (1-48)
on the index of anisotropy. Elevation zwD = zw /rw (1-49)
The skin effect seq is added to the second logarith-
mic term in the denominator of the horizontal well Reservoir thickness hD = h /rw (1-50)
production expression (Eq. 1-25) and must be multi- Penetration ratio hwD´ = hw /h. (1-51)
plied by Ianih /L. One obvious, although not neces-
sarily desirable, way to offset the effects of damage The terms hD, hwD, zwD /hD and hwDcosθ/hD must
is to drill a longer horizontal well. be known to evaluate the skin effect.
As an example, assume hD = 100, zwD /hD = 0.5
(midpoint of the reservoir) and hwDcosθ/hD = 0.25
1-3.2. Components of the skin effect (θ = 60°, hw /h = 0.5). For this case, sc + θ = 5.6 (from
Table 1-3). If the penetration ratio is reduced to 0.1,
Matrix stimulation has proved to be effective in the skin effect increases to 15.5.
reducing the skin effect caused by most forms of It is apparent that this skin effect alone could
damage. However, the total skin effect is a compos- dwarf the skin effect caused by damage. The skin
ite of a number of factors, most of which usually effect resulting from the partial penetration length
cannot be altered by conventional matrix treatments. hwD´ may be unavoidable because it typically results
The total skin effect may be written as from other operational considerations (such as the
prevention of gas coning).
st = sc+θ + s p + sd + ∑ pseudoskins . (1-47) From Table 1-3 and for full penetration it can be
seen readily that a deviated well, without damage,
The last term in the right-hand side of Eq. 1-47 should have a negative skin effect. Thus, a small
represents an array of pseudoskin factors, such as skin effect or even one equal to zero obtained from
phase-dependent and rate-dependent effects that

rw
θ
rw
hw hw
h h
zw
zw
Vertical well Slanted well
Figure 1-9. Geometry for partial and off-centered completions and slant skin effects (Cinco-Ley et al., 1975a).
a well test in a highly deviated well may mean con- 

rperf kV 
siderable damage. Removal of this damage with rpD = 1 + k  , (1-56)
appropriate stimulation could increase the deviated 2h  H 
well production (or injection) considerably. where rperf is the perforation radius; and
rw
rwD = . (1-57)
1-3.4. Perforation skin effect l p + rw
Karakas and Tariq (1988) developed a procedure to
The vertical pseudoskin effect is then
calculate the skin effect caused by perforations. This
skin effect is a composite involving the plane-flow sV = 10 a hDb−1rpD
b
, (1-58)
effect sH, vertical converging effect sV and wellbore
effect swb: where a and b are
s p = sH + sV + swb . (1-52) a = a1 log rpD + a2 (1-59)
The pseudoskin factor sH is given by b = b1rpD + b2 . (1-60)

r
sH = ln w , (1-53)
rw′ (θ) The values of the constants a1, a2, b1 and b2 are
where rw´(θ) is the effective wellbore radius and is a given in Table 1-5 as functions of the phasing angle θ.
function of the perforation phasing angle θ: Finally, the wellbore skin effect swb can be approx-
imated by
 lp 4 when θ = 0
rw′ (θ) = 
(
α θ rw + l p ) when θ ≠ 0
, (1-54) swb = c1e c r . 2 wD
(1-61)
where lp is the length of the perforation and αθ is a The constants c1 and c2 can be obtained from
phase-dependent variable and can be obtained from Table 1-6.
Table 1-4. As an example, assume rw = 0.406 ft, lp = 0.667 ft,
The vertical pseudoskin factor sV can be calculated h = 0.333 ft (3 shots per foot [spf]), kH/kv = 3, rperf =
after certain dimensionless variables are determined: 0.0208 ft [0.25 in.] and θ = 90°.
From Eq. 1-54 and Table 1-4, rw´(θ) = 0.779 ft,
h kH and thus from Eq. 1-53, sH = –0.65. From Eqs. 1-55,
hD = , (1-55)
lp kV 1-56 and 1-57, the dimensionless variables hD, rpD
and rwD are equal to 0.86, 0.05 and 0.38, respect-
where h is the distance between perforations and is ively. From Eq. 1-59 and Table 1-5, a = 2.58, and
exactly inversely proportional to the shot density; from Eq. 1-60 and Table 1-5, b = 1.73. Then, from

Table 1-3. Pseudoskin factors for partially penetrating slanted wells (Cinco-Ley et al., 1975).
θw zwD /hD θw
hwD cosθ sθ + c sc sθ θw zwD /hD θw
hwD cosθ sθ + c sc sθ
(°) hD (°) hD
hD = 100 hD = 100 continued
0 0.95 0.1 20.810 20.810 0 0 0.6 0.5 2.430 2.430 0
15 20.385 20.810 –0.425 15 2.254 2.430 –0.176
30 18.948 20.810 –1.861 30 1.730 2.430 –0.700
45 16.510 20.810 –4.299 45 0.838 2.430 –1.592
60 12.662 20.810 –8.147 60 –0.466 2.430 –2.897
75 6.735 20.810 –14.074 75 –2.341 2.430 –4.772
0 0.8 0.1 15.809 15.809 0 0 0.5 0.5 2.369 2.369 0

15 15.449 15.809 –0.36 15 2.149 2.369 –0.175
30 14.185 15.809 –1.623 30 1.672 2.369 –0.697
45 12.127 15.809 –3.682 45 0.785 2.369 –1.584
60 8.944 15.809 –6.864 60 –0.509 2.369 –2.879
75 4.214 15.809 –11.594 75 –2.368 2.369 –4.738
0 0.6 0.1 15.257 15.257 0 0 0.625 0.75 0.924 0.924 0

15 14.898 15.257 –0.359 15 0.778 0.924 –0.145
30 13.636 15.257 –1.621 30 0.337 0.924 –0.587
45 11.583 15.257 –3.674 45 –0.411 0.924 –1.336
60 8.415 15.257 –6.842 60 –1.507 0.924 –2.432
75 3.739 15.257 –11.517 75 –3.099 0.924 –4.024
0 0.5 0.1 15.213 15.213 0 0 0.5 0.75 0.694 0.694 0

15 14.854 15.213 –0.359 15 0.554 0.694 –0.139
30 13.592 15.213 –1.620 30 0.134 0.694 –0.560
45 11.540 15.213 –3.673 45 –0.581 0.694 –1.275
60 8.372 15.213 –6.841 60 –1.632 0.694 –2.326
75 3.699 15.213 –11.514 75 –3.170 0.694 –3.864
0 0.875 0.25 8.641 8.641 0 0 0.5 1 0 0 0

15 8.359 8.641 –0.282 15 –0.128 0 –0.128
30 7.487 8.641 –1.154 30 –0.517 0 –0.517
45 5.968 8.641 –2.673 45 –1.178 0 –1.178
60 3.717 8.641 –4.924 60 –2.149 0 –2.149
75 0.464 8.641 –8.177 75 –3.577 0 –3.577
0 0.75 0.25 7.002 7.002 0 hD = 1000

15 6.750 7.002 –0.251
30 5.969 7.002 –1.032 0 0.95 0.1 41.521 41.521 0
45 4.613 7.002 –2.388 15 40.343 41.521 –1.178
60 2.629 7.002 –4.372 30 36.798 41.521 –4.722
75 –0.203 7.002 –7.206 45 30.844 41.521 –10.677
60 22.334 41.521 –19.187
0 0.6 0.25 6.658 6.658 0 75 10.755 41.521 –30.766
15 6.403 6.658 –0.249
30 5.633 6.658 –1.024 0 0.8 0.1 35.840 35.840 0
45 4.290 6.658 –2.447 15 34.744 35.840 –1.095
60 2.337 6.658 –4.32 30 31.457 35.840 –4.382
75 –0.418 6.658 –7.076 45 25.973 35.840 –9.867
60 18.241 35.840 –17.599
0 0.5 0.25 6.611 6.611 0 75 8.003 35.840 –27.837
15 6.361 6.611 –0.249
30 5.587 6.611 –1.023 0 0.6 0.1 35.290 35.290 0
45 4.245 6.611 –2.365 15 34.195 35.290 –1.095
60 2.295 6.611 –4.315 30 30.910 35.290 –4.380
75 –0.451 6.611 –7.062 45 25.430 35.290 –9.860
60 17.710 35.290 –17.580
0 0.75 0.5 3.067 3.067 0 75 7.522 35.290 –27.768
15 2.878 3.067 –0.189
30 2.308 3.067 –0.759 0 0.5 0.1 35.246 35.246 0
45 1.338 3.067 –1.729 15 34.151 35.246 –1.095
60 –0.082 3.067 –3.150 30 30.866 35.246 –4.380
75 –2.119 3.067 –5.187 45 25.386 35.246 –9.860
60 17.667 35.246 –17.579
75 7.481 35.246 –27.765

Table 1-3. Pseudoskin factors for partially penetrating slanted wells (Cinco-Ley et al., 1975) continued.
θw zwD /hD θw
hwD cosθ sθ + c sc sθ θw zwD /hD θw
hwD cosθ sθ + c sc sθ
(°) hD (°) hD
hD = 1000 continued hD = 1000 continued
0 0.875 0.25 15.733 15.733 0 0 0.6 0.5 4.837 4.837 0
15 15.136 15.733 –0.597 15 4.502 4.837 –0.335
30 13.344 15.733 –2.389 30 3.503 4.837 –1.334
45 10.366 15.733 –5.367 45 1.858 4.837 –2.979
60 6.183 15.733 –9.550 60 –0.424 4.837 –5.261
75 0.632 15.733 –15.101 75 –3.431 4.837 –8.268
0 0.75 0.25 14.040 14.040 0 0 0.5 0.5 4.777 4.777 0

15 13.471 14.040 –0.569 15 4.443 4.777 –0.334
30 11.770 14.040 –2.270 30 3.446 4.777 –1.331
45 8.959 14.040 –5.081 45 1.806 4.777 –2.971
60 5.047 14.040 –8.993 60 –0.467 4.777 –5.244
75 –0.069 14.040 –14.109 75 –3.458 4.777 –8.235
0 0.6 0.25 13.701 13.701 0 0 0.625 0.75 1.735 1.735 0

15 13.133 13.701 –0.568 15 1.483 1.735 –0.252
30 11.437 13.701 –2.264 30 0.731 1.735 –1.004
45 8.638 13.701 –5.063 45 –0.512 1.735 –2.247
60 4.753 13.701 –8.948 60 –2.253 1.735 –3.988
75 –0.288 13.701 –13.989 75 –4.595 1.735 –6.330
0 0.5 0.25 13.655 13.655 0 0 0.5 0.75 1.508 1.508 0

15 13.087 13.655 –0.568 15 1.262 1.508 –0.246
30 11.391 13.655 –2.264 30 0.528 1.508 –0.980
45 8.593 13.655 –5.062 45 –0.683 1.508 –2.191
60 4.711 13.655 –8.944 60 –2.380 1.508 –3.888|
75 –0.321 13.655 –13.976 75 –4.665 1.508 –6.173
0 0.75 0.5 5.467 5.467 0 0 0.5 1 0 0 0

15 5.119 5.467 –0.348 15 –0.206 0 –0.206|
30 4.080 5.467 –1.387 30 –0.824 0 –0.824
45 2.363 5.467 –3.104 45 –1.850 0 –1.850
60 –0.031 5.467 –5.498 60 –3.298 0 –3.298
5 –3.203 5.467 –8.670 75 –5.282 0 –5.282
Table 1-4. Dependence of αθ on phasing. Table 1-5. Vertical skin correlation coefficients.
Perforating Phasing (°) αθ Perforating

Phasing (°) a1 a2 b1 b2
0 (360) 0.250
0 (360) –2.091 0.0453 5.1313 1.8672
180 0.500
180 –2.025 0.0943 3.0373 1.8115
120 0.648
120 –2.018 0.0634 1.6136 1.7770
90 0.726
90 –1.905 0.1038 1.5674 1.6935
60 0.813
60 –1.898 0.1023 1.3654 1.6490
45 0.860
45 –1.788 0.2398 1.1915 1.6392
Eq. 1-58, sV = 1.9, and from Eq. 1-61 and Table 1-6,
swb = 0.02. k  r 
(s ) =  − 1 ln s + s p  = ( sd )o + s p ,
k
The total perforation skin effect obtained with
 ks   rw
d p
Eq. 1-52 is equal to 1.3 for this example.  ks
• Combination of damage and perforation skin (1-62)

effect where the perforations terminate inside the dam-
Karakas and Tariq (1988) showed that the damage age zone (lp < ld), rs is the damage zone radius,
and perforation skin effect can be approximated by and (sd)o is the equivalent openhole skin effect

Table 1-6. Variables c1 and c 2. from hydraulic fracturing is illustrated by revisiting

“Example of steady-state IPR: skin effect variation”
Perforating Phasing (°) c1 c2 (page 1-4). With permeability equal to 5 md, the
0 (360) 1.6E–1 2.675 reduction in the skin effect from 10 to 0 (e.g., pwf =
180 2.6E–2 4.532
2000 psi) results in production rates of 560 and 1230
STB/D, respectively, and this difference of 670
120 6.6E–3 5.320
STB/D is clearly an attractive target for matrix stim-
90 1.9E–3 6.155
ulation. However, at a permeability of 0.05 md, all
60 3.0E–4 7.509 rates would be divided by 100, resulting in an incre-
45 4.6E–5 8.791 mental production of only 6.7 STB/D.
Interestingly, for k = 0.05 md, reducing the skin
effect to –5 leads to a poststimulation production rate
(Eq. 1-45). If, for example, lp = 1.2 ft (rs = 1.606 equal to 30 STB/D and an incremental production
ft) and the permeability reduction ratio k/ks = 5, rate (over the s = 10 case and k = 0.05 md) of about
from Eq. 1-62 and the perforation skin effect cal- 25 STB/D. Such an equivalent skin effect can be the
culated in the previous section, (sd)p = 12. result of hydraulic fracturing.
Karakas and Tariq (1988) also showed that A great portion of this volume is devoted to this
the damage skin effect for perforations terminat- type of stimulation, its fundamental background and
ing outside the damaged zone can be approxi- the manner with which it is applied in petroleum
mated by engineering. Here, hydraulic fractures are presented
as well production or injection enhancers.
(s )
d p
= s p − s ′p , (1-63) Prats (1961), in a widely cited publication, pre-
sented hydraulic fractures as causing an effective
where sp´ is the perforation skin effect evaluated
wellbore radius and, thus, an equivalent skin effect
at the modified perforation length lp´ and modi-
once the well enters (pseudo)radial flow. In other
fied radius rw´:
words, the reservoir flows into a fractured well as
if the latter has an enlarged wellbore. Figure 1-10
l p′ = l p − 1 − s  ld
k is Prats’ development graphed as the dimensionless
(1-64)
 k effective wellbore radius rwD´ versus the relative
capacity parameter a.
rw′ = rw + 1 − s  ld .
k
(1-65) The dimensionless relative capacity parameter has
 k
been defined as
The quantities lp´ and rw´ are used instead of lp
and rw, respectively, to calculate sp as presented πkx f
a= , (1-66)
in Section 1-3.4. 2k f w
Assume that in the previous example ld = 0.4 ft,
which makes the modified length lp´ and modi- where k is the reservoir permeability, xf is the frac-
fied radius rw´ equal to 0.347 and 0.726 ft, re- ture half-length, kf is the fracture permeability, and
spectively. From Eq. 1-63, (sd)p = 1, which is a w is the fracture width.
marked decrease from the value calculated for The dimensionless effective wellbore radius is
the length of the damage larger than the length simply
of the perforations. rw′ rw e
−sf
′ =
rwD = . (1-67)
xf xf
1-3.5. Hydraulic fracturing in production
engineering Thus, if xf and kf w are known (as shown later in
this volume, this is the essence of hydraulic fractur-
If removal of the skin effect by matrix stimulation ing), then Fig. 1-10 enables calculation of the equiv-
and good completion practices does not lead to an alent skin effect sf that the well will appear to have
economically attractive well, the potential benefit while it flows under pseudoradial conditions. Cinco-

1 3.0
0.5
Dimensionless effective
wellbore radius, rwD
'
0.1 2.5
0.01 2.0
sf + ln(xf /rw)
0.001 1.5
0.01 0.1 k 1f w 10 100 1000
C fD =
Relative capacity parameter, a
kf x
Figure 1-10. Dimensionless effective wellbore radius of a 1.0
hydraulically fractured well (Prats, 1961).
Ley and Samaniego-V. (1981b) later introduced a 0.5

0.1 1 10 100 1000
direct correlation for sf (Fig. 1-11).
Dimensionless fracture conductivity, CfD
Graphed on the x-axis of Fig. 1-11 is the dimen-
sionless fracture conductivity CfD, which is simply Figure 1-11. Equivalent fracture skin effect (Cinco-Ley and
Samaniego-V., 1981b).
kf w
C fD = (1-68)
kf x xf
s f + ln = 0.9 ,
and is related to Prats’ relative capacity by rw
π which for xf = 500 ft and rw = 0.328 ft gives

C fD = . (1-69) sf = –6.4.
2a
The production rates at pretreatment (s = 10),
The following example illustrates the impact of a after matrix stimulation (s = 0) and after fracturing
hydraulic fracture on well production. (s = –6.4) are 56, 123 and 518 STB/D, respectively.
• Example calculation of production from • General requirements for hydraulic fractures
a hydraulically fractured well
What general requirements should be expected
Using the variables in “Example of steady-state from the design of hydraulic fractures? As dis-
IPR: skin effect variation” (page 1-4) but with k = cussed in later chapters of this volume, the execu-
0.5 md, demonstrate the production improvement tion of a hydraulic fracture should provide a frac-
from a hydraulic fracture at CfD = 5 and xf = 500 ture length and propped width, and selection of the
ft. Also, compare this result with the pretreatment proppant and fracturing fluid is crucial for fracture
production if s = 10 and after a matrix stimulation, permeability. Because of physical constraints the
assuming that all skin effect is eliminated (s = 0). resulting values may not be exactly the desired
Use pwf = 2000 psi. ideal values, but certain general guidelines should
Solution permeate the design.
The IPR for this well is simply The dimensionless fracture conductivity CfD
is a measure of the relative ease with which the
 5000 − pwf  reservoir (or injected) fluid flows among the well,
q = 0.345 . fracture and reservoir. Obviously, in a low-perme-
 8.43 + s 
ability reservoir even a fracture of narrow width
Using Fig. 1-11 (Fig. 1-10 can also be used) and and relatively low permeability results de facto in
CfD = 5: a high-conductivity fracture. The limiting value is
an infinite-conductivity fracture, which mathemati-

cally implies that once fluid enters the fracture it is Obviously, the aim is to minimize the denomi-
instantaneously transported into the well. Thus, in nator.
low-permeability reservoirs, the length of the frac- This optimization problem was solved by Prats
ture is critical and the design must consider this (1961) for steady-state flow. He found the maxi-
requirement. The longer the fracture, subject to the mum production rate occurs at a = 1.25 (CfD =
economic constraints of its execution, the more 1.26 from Eq. 1-69). For this value of a, rw´/xf =
desirable it is. 0.22 from Fig. 1-10 and Eq. 1-66 gives
Conversely, for high-permeability reservoirs, as
shown by Eq. 1-68, to increase CfD requires increas- x f w = 0.8 k f k = 0.8 × 10, 000 0.4 = 20, 000 .
ing the kf w product. Thus, maximizing conductivity
must be the major result from the design. Arresting Using Vf = 2whxf with this xf /w ratio,
the length growth and inflating the fracture are
means to accomplish this purpose. A process x 2f = 20, 000Vf 2 h = (20, 000 × 3500) (2 × 65) ,
involving tip screenout (TSO) has been developed,
exactly to effect such a fracture geometry. and xf = 730 ft; hence, w = xf /2000 = 0.037 ft =
• Optimal fracture conductivity 0.44 in. From rw´ = 0.22 xf and rw = 0.33 ft, Eq.
1-7 gives rw´/rw = e–s = 490, and s = –(ln 490) =
With advent of the TSO technique especially in
–6.1.
high-permeability, soft formations (called frac and
For pe = 5000 psi and pwf = 3000 psi, the opti-
pack), it is possible to create short fractures with
mized production rate is
unusually wide propped widths. In this context a
strictly technical optimization problem can be for-
(0.4)(65)(2000)
mulated: how to select the length and width if the q= = 204 STB / D .
propped fracture volume is given. The following (141.2)(1)(1)(ln3000 − 6.2)
example from Valkó and Economides (1995)
It is necessary to check if the resulting half-
addresses this problem, using the method derived
length is less than re (otherwise xf must be selected
by Prats (1961).
to be equal to re). Similarly, the resulting optimal
• Example of optimal fracture conductivity width must be realistic; e.g., it is greater than 3
Consider the following reservoir and well data: times the proppant diameter (otherwise a threshold
k = 0.4 md, h = 65 ft, re /rw = 1000, µ = 1 cp, pe = value must be selected as the optimal width). In
5000 psi and pwf = 3000 psi. Determine the opti- this example both conditions are satisfied.
mal fracture half-length xf, optimal propped width This example provides an insight into the real
w and optimal steady-state production rate if the meaning of dimensionless fracture conductivity.
volume of the propped fracture is Vf = 3500 ft3. The reservoir and the fracture can be considered a
Use a value of 10,000 md for the fracture perme- system working in series. The reservoir can deliver
ability kf, taking into account possible damage to more hydrocarbons if the fracture is longer, but
the proppant, and assume that the created fracture with a narrow fracture, the resistance to flow may
height equals the formation thickness. Use the be significant inside the fracture itself. The optimal
Cinco-Ley and Samaniego-V. (1981b) graph (Fig. dimensionless fracture conductivity CfD,opt = 1.26
1-11), which assumes pseudoradial flow. in this example corresponds to the best compro-
Solution mise between the requirements of the two subsys-
The same propped volume can be established by tems.
creating a narrow, elongated fracture or a wide but
short one. In all cases the production rate can be 1-4. Tubing performance and
obtained from Eq. 1-9, which with the incorpora-
tion of sf takes the form NODAL* analysis
kh∆p The inflow performance relationships described in
q= . Section 1-2 provide a picture of the pressure and
 re 
141.2 Bµ ln + s f  rates that a reservoir with certain characteristics (per-
 rw  meability, thickness, etc.), operating under certain

conditions (pressure, mode of flow), can deliver into in Fig. 1-13 for two wellhead flowing pressures. As
the bottomhole of a well. The fluid must traverse a the flow rate increases (on the right side of the
path from the bottom of the well to the top and then curves) the required bottomhole flowing pressure
into surface equipment such as a separator. Figure increases, reflecting higher friction pressures at the
1-12 describes such a path, which consists of several higher rates. On the left side of the curves, the pecu-
segments, joints and valves, all of which cause a liar shape is due to liquid holdup; lower rates do not
pressure drop. NODAL analysis considers the reser- have sufficient momentum to purge liquid accumula-
voir/wellbore system and uses calculations of the tion in the well, resulting in an unavoidable increase
pressure loss across each segment to predict the pro- in the hydrostatic pressure.
duction rate and identify any restrictions that may The correlations to calculate the required pressure
reduce the hydrocarbon flow rate. drops take full account of the phase behavior of the,
At its simplest manifestation, for a given wellhead almost always, two-phase oil and gas mixture. An
pressure, tubing performance allows calculation of increase in the wellhead pressure ordinarily results
the required bottomhole flowing pressure to lift a in a disproportionate increase in the bottomhole pres-
range of flow rates to the top. The total pressure drop sure because the higher pressure level in the tubing
in the well consists of the hydrostatic and friction causes a more liquid-like fluid and a larger hydrosta-
pressure drops. tic pressure component (density is higher).
Several correlations for tubing performance are in Combining the tubing performance curve, often
use in the petroleum industry (Beggs and Brill, 1973; known in vertical wells as the vertical lift perfor-
Hagedorn and Brown, 1965). Brown (1977), in a mance (VLP), with an IPR provides the well deliver-
widely used work, outlined the procedure for pres- ability at the determined bottomhole flowing pres-
sure drop calculations in production strings as shown sure (Fig. 1-14).
∆p8 = ptf – psep

Sales line
ptf ∆p6 = pdsc – psep Gas
Separator
pdsc psep
Surface choke Liquid
Stock
∆p5 = ptf – pdsc tank
pdsv
∆p4 = pusv – pdsv
∆p 7 = p wf – ptf pusv ∆p1 = p – pwfs = Loss in porous medium
∆p 2 = pwfs – pwf = Loss across completion
Bottomhole ∆p 3 = pur – pdr = Loss across restriction
restriction
pdr ∆p4 = pusv – pdsv = Loss across safety valve
∆p 5 = ptf – pdsc = Loss across surface choke
∆p3 = pur – pdr ∆p 6 = pdsc – psep = Loss in flowline
∆p 7 = pwf – ptf = Total loss in tubing
∆p 8 = ptf – psep = Total loss in flowline
pur
pwf p wfs p pe
∆p2 = p wfs – pwf ∆p1 = p – p wfs
Figure 1-12. Well hydraulic system. pdr = downstream restriction pressure, pdsc = pressure downstream of the surface
choke, pdsv = pressure downstream of the safety valve, psep = separator pressure, ptf = tubing flowing pressure,
pur = upstream restriction pressure, pusv = pressure upstream of the safety valve, pwfs = wellbore sandface pressure.

Bottomhole flowing pressure (psi) 6000 6000
5000
5000
4000

pwh = 1200 psi
4000
3000
pwh = 800 psi VLP
2000 3000 pwf
1000 q
2000
0 IPR
0 200 400 600 800 1000
1000
Producing rate (STB/D)
Figure 1-13. Vertical lift performance (also known as tub- 0

ing intake) curves for two values of wellhead flowing pres- 0 1000 2000 3000 4000 5000 6000
sure pwh. Flow rate, q
Figure 1-14. IPR and VLP curves combined for the predic-
NODAL analysis is one of the most powerful tools tion of well deliverability.
in production engineering. It can be used as an aid in
both the design and optimization of well hydraulics
and IPR modification. Figure 1-15 shows one of the 4000
most common uses of NODAL analysis. The well
IPR
IPR is plotted with three VLP curves (e.g., each cor- 3000 VLP
d tbg,1
responding to a different wellhead pressure—and
perhaps a different artificial lift mechanism—in the q1 d tbg,2
case of an oil well or a different tubing diameter in 2000
q2 d tbg,3
a gas well). The three different production rates over
time can be balanced against the incremental eco- 1000 q3
nomics of the various well completion options.

Figure 1-16 demonstrates a single VLP but three
0
different IPRs (e.g., each corresponding to a different 0 1000 2000 3000 4000 5000 6000
hydraulic fracture design). Again, the incremental Flow rate, q
benefits over time must be balanced against the
incremental costs of the various fracture designs. Figure 1-15. VLP curve variation for different tubing diam-
eters ( dtbg ) and the effect on well deliverability.
The use of NODAL analysis as an engineering
investigative tool is shown in Fig. 1-17. Suppose that
several perforations are suspected of being closed. A activities in the petroleum or related industries use
calculation allowing several different scenarios of such a wide spectrum of sciences and technologies
the number of open perforations and comparison as well stimulation, both matrix and fracturing. This
with the actual flow rate can provide a convincing volume is intended to present these technologies and
answer to the problem. their interconnections.
As with many engineering processes, stimulation
must culminate in the design, selection of the spe-
1-5. Decision process for well cific treatment and, of course, selection of candidate
stimulation wells. To choose among the various options, of
To be done properly, the engineering exercise of the which one is to do nothing, a means for an economic
decision process for well stimulation requires consid- comparison of the incremental benefits weighted
erable knowledge of many diverse processes. Few against the costs is necessary.

4000 4000

3000 3000
VLP VLP
2000 2000
q3
q2 q3
q1 q2
q1
1000 1000
IPR1 IPR2 IPR3 Nperf,1 Nperf,2 Nperf,3

qactual
0 0
0 1000 2000 3000 4000 5000 6000 0 1000 2000 3000 4000 5000 6000
Flow rate, q Flow rate, q
Figure 1-17. Well diagnosis (e.g., for unknown number of

Figure 1-16. IPR curve variation (e.g., for different skins) open perforations Nperf) using a comparison of predicted
and the effect on well deliverability. versus actual IPRs.
1-5.1. Stimulation economics or the net value (profit) for the operator; rather, it
is a measure of liquidity or how fast the investment
Because the whole purpose of stimulation is to
will be recovered.
increase the value of the producing property through
The indicator can be adjusted to show the time
an accelerated production rate or increased recovery,
value of money (discounted payout), the hurdle rate
economics should be the driver in deciding whether
necessary for the company to invest or both factors.
to conduct the stimulation, what type of stimulation
The hurdle rate is the annualized percentage of
to do and which various aspects of the treatment to
return that must be achieved to make the project as
include.
good an investment as the average company invest-
Several economic indicators can be used to show
ment. The discounted payout is
the value of stimulation. Because of the wide variety
of operating conditions, companies may not have a n
∆$ n
single indicator for the “answer” in all stimulation ∑ (1 + i)
n =1
n − cost = 0 . (1-71)
investments. Although the common ground in eco-
nomics is profit, in many petroleum activities liquid- The interest (hurdle) rate i is the indicator that sug-
ity, risk and corporate goals may make it necessary gests when the investment will be returned without
to choose investments that differ from the ultimate lowering the corporate investment returns and
maximum value of a project. accounting for inflation (time value of money).
The oldest indicator used in oil production is pay- When the full stream of cash flows for the pro-
out time, which is the amount of time necessary to jected relative life of the project is used, an indicator
recoup the money invested. If the actual time is less called net present value (NPV) is defined as
than the required time, the investment is considered n
∆$ n
attractive: NPV = ∑ − cost . (1-72)
n =1 (1 + i)n
n
∑ ∆$
n =1
n − cost = 0 , (1-70) NPV gives a dollar value added to the property at
present time. If it is positive, the investment is attrac-
where ∆$n is the incremental revenue (minus the tive; if it is negative, it means an undesirable invest-
incremental expenses and taxes that are due to opera- ment. NPV is the most widely used indicator show-
tions), n is the time period increments (e.g., years) in ing a dollar amount of net return.
which it is received, and cost consists of the total To get an indicator on relative profitability against
expenses associated with the stimulation. This indi- more global investments such as stocks, bonds and
cator does not provide for the time value of money corporate profits, the rate of return (ROR) is used.

ROR is simply varying i to get an NPV equal to • Well location size limits the equipment and mate-
zero. That value of i is the ROR. The limitation in rials that can be used.
using the ROR indicator is that it does not provide a • Tubular integrity prevents or limits the type of
mechanism of how the cash comes in (cash flow ver- treatments that can be employed without compro-
sus time). mise.
In Fig. 1-18 there are two investment possibilities.
• Completion tools and their location limit where
A has the highest NPV for small interest rates but
the treatment is placed and the magnitude of the
falls off quickly with increasing rates, whereas B has
rates and volumes.
a smaller NPV with low rates but remains flatter as
rates rise. This means that A makes more money, but • Zonal isolation is whether the zone can be isolated
as interest rates rise its return is hurt more than that from other intervals through perforating and/or
for B. B pays the money back with a better ROR, pipe integrity limitations.
even if it has a smaller NPV at low interest rates. Typical reservoir constraints are
Another indicator of investment profitability is the
benefits to cost ratio (BCR): • production failures: water or gas coning or influx,
formation sanding
NPV • physical location of the zones and their thicknesses:
BCR = , (1-73) pay zone qualities limit or dictate treatments.
cost
which shows the relationship of relative return for a

given investment (cost) size. BCR is a good indica- 1-6. Reservoir engineering
tor if there are more investment opportunities than considerations for optimal
money to invest.
production enhancement
250
strategies†
200
Cost-effective production improvement has been the
industry focus for the past several years. Fracturing,
150
A stimulating, reperforating and recompleting existing
NPV ($1000)
100 wells are all widely used methods with proven

B ROR = 14%
50 results in increasing the NPV of old fields. Now
ROR = 18%
0
reentry drilling is generating high interest for the
potential it offers to improve recovery from damaged
–50
or depleted zones or to tap into new zones at a gen-
–100 erally low cost. Applied to mature reservoirs, all
0 5 10 15 20 25 30
these strategies have the advantage of starting with
i
a fair to good reservoir description as well as a
Figure 1-18. Determination of the rate of return for pro- working trajectory to the target formation. Even
jects A and B. when a new well is drilled, the decision whether to
drill a vertical, slanted or horizontal well and how to
complete the productive interval can profoundly
effect the well’s productivity and the size of the vol-
1-5.2. Physical limits to stimulation ume drained by the well. Today’s technology also
treatments entertains multiple branches from a main trunk,
Physical limits are dominant aspects for stimulation which may be a newly drilled or existing well.
treatment decisions as often as economic indicators.
For the well, these include the following:
• Maximum allowable treating pressure limits injec-
tion rates and the type of treating fluids.
• Tubular size limits rates and pipe erosion. †
This section by Christine Ehlig-Economides, Schlumberger GeoQuest.

1-6.1. Geometry of the well drainage with the well. This forms a circle for a vertical well;
volume an approximate ellipse is formed for hydraulically
fractured and horizontal wells.
The geometry of the well drainage volume depends Wells drilled in a square pattern impose a square
on the well trajectory within the productive zone, drainage area. For vertical wells, this is similar to the
neighboring wells, geometry of hydraulic fractures, circular effective drainage shape (Fig. 1-19c), but for
nearby reservoir limits and reservoir flow character- horizontal wells, the equivalent drainage efficiency
istics. Areas drained by an isolated well in an effec- corresponds to an elongated area. As a rule of
tively infinite reservoir are diagrammed in Figs. thumb, the length of the horizontal well drainage
1-19a and 1-19b. A vertical well creates a circular area can be as long as the length of the horizontal
cylinder pressure sink whereas a hydraulically frac- well plus one diameter of the comparable vertical
tured well creates a pressure sink in the shape of a well drainage area. For the case in Fig. 1-19d, one-
finite slab with dimensions defined by the formation half as many horizontal wells of the length shown
thickness and the total fracture length. With adequate could be used to drain the same pattern, as shown in
vertical permeability the horizontal well drainage Fig. 1-20a. With longer horizontal wells, even fewer
area is similar to that of a vertical fracture, with the are required.
total fracture length equal to that of the horizontal Figure 1-20b shows another consideration. If the
well. The extent of the effective drainage area is vertical well pattern does not take the direction of
approximately defined by the locus of points equidis- maximum horizontal stress σΗ,max into account,
tant from the surface of the pressure sink associated
(a) Isolated vertical well (b) Isolated horizontal or hydraulically fractured well
(c) Pattern of vertical wells (d) Pattern of horizontal wells
Figure 1-19. Drainage areas for single and multiple vertical and horizontal wells.

σH,max
(a) Pattern of horizontal wells (b) Pattern of hydraulically fractured wells
Figure 1-20. Drainage areas resulting from (a) longer horizontal wells draining more area per well and (b) hydraulically
fractured wells in a square pattern that is not in line with the direction of maximum stress.
hydraulically fracturing the wells may result in an option in this case because the productivity would
unplanned drainage geometries. be severely penalized by the low vertical permeabil-
ity, and in a thick formation, a horizontal well may
not even produce the entire formation thickness. A
1-6.2. Well drainage volume vertical well—barefoot, perforated and gravel
characterizations and production packed, or gravel packed—can provide excellent
optimization strategies productivity in formations with moderate mobility. A
Figures 1-19 and 1-20 assume that the reservoir is slanted well can produce a marginal increase in pro-
homogeneous and isotropic over vast areas. In reali- ductivity over a vertical well.
ty, typical reservoir geology is much more complex. In very high mobility laminated reservoirs (such as
Formation flow characteristics may favor one well turbidites), a frac and pack may provide sand control
geometry over others. The chart in Fig. 1-21 summa- and the means to bypass near-wellbore damage.
rizes production optimization strategies for a series However, in a low-mobility reservoir, hydraulically
of 10 common well drainage volume characteriza- fracturing the well is preferred over any other option
tions. The chart addresses five potential well paths: because it provides an effective planar sink, greatly
conventional vertical, hydraulically fractured verti- increasing the well productivity. For thin and lami-
cal, slanted, horizontal and hydraulically fractured nated reservoirs, hydraulic fractures in a horizontal
horizontal. For any one of the drainage volume char- well may be the optimal choice because the longer
acterizations, well path options are shown in block well provides greater reach that increases the drain-
diagrams. age volume of the well and the hydraulic fractures
Laminated reservoirs (chart row 4 on Fig. 1-21) enable horizontal flow to the well through the entire
are a good starting point to understanding the infor- formation thickness. Hydraulic fractures in a hori-
mation in the chart. The chart distinguishes layered zontal well can be planned either as longitudinal, by
from laminated by defining a reservoir as layered if drilling the well in the direction of maximum hori-
the recognized sands are thick enough to be targeted zontal stress, or as transverse, by drilling the well in
by a horizontal well. If not, the reservoir is classed the direction of minimum stress.
as laminated. In general, laminated reservoirs have Horizontal wells offer particular advantages in nat-
poor vertical permeability. A horizontal well is not urally fractured reservoirs (chart row 5 on Fig. 1-21)

Figure 1-21. Production optimization strategies. Completion options include perforating, gravel packing and stimulation in
combination with an applicable strategy.

when they are drilled normal to the fracture planes. layer, but stacked reentry laterals are a highly effec-
Locating natural fractures and determining their ori- tive strategy. In the latter design, the length of the
entation is crucial to developing the best well design lateral can be roughly inversely proportional to the
in these formations. Hydraulic fracturing places layer’s flow capacity. A slanted well offers a less
proppant in a series of natural fractures, which typi- expensive strategy for boosting productivity in a lay-
cally results in a propped fracture that is parallel to ered reservoir. By designing the trajectory with more
the natural fractures. A horizontal well normal to drilled length in less productive layers, some confor-
natural fractures usually provides better productivity mance control can be achieved. However, if early
than hydraulic fracturing. water breakthrough occurs in the higher productivity
Although natural fractures usually are subvertical layer, it is generally much easier to shut off produc-
(nearly vertical), shallower reservoirs and overpres- tion in one of the stacked laterals than in a midlength
sured zones may have subhorizontal (nearly horizon- portion of the slanted well.
tal) fractures open to flow. Vertical and slanted wells Hydraulic fracturing in slanted wells is performed
are a reasonable choice in these cases. Injection of typically in offshore wells that commonly follow the
proppant into horizontal fractures in overpressured same deviation used to reach the reservoir location
zones keeps them open after production lowers the from a platform. These fractures are typically frac
pore pressure. Otherwise, the weight of the over- and pack treatments designed for sand control.
burden tends to close horizontal natural fractures. Because the deviated trajectory may be detrimental
Likewise, high-pressure injection can reopen natural to the fracture treatment design, some operators
fractures in depleted zones or natural fractures that direct the trajectory downward to nearly vertical
were plugged during drilling. before passing through the productive formation
Moving up the chart to the layered reservoirs in if hole stability problems do not preclude this
row 3 offers an opportunity to address the impor- approach.
tance of conformance control. The conventional ver- At the top row of the chart in Fig. 1-21 are thick,
tical well commingles production from multiple lay- homogeneous formations. Any of the well path
ers. Productivity and storage capacity contrasts can options may be applied for these reservoirs. Mobility
result in the differential depletion of layers that are extremes may favor hydraulic fracturing, whereas
not in hydraulic communication vertically other than moderate mobility allows using less expensive, con-
at the well. In this case, when the production rate is ventional vertical well completions. A slanted well
reduced or the well is shut in, crossflow occurs in the may be more cost effective than hydraulic fracturing
wellbore as the higher pressure layers recharge the or a horizontal well, provided that the ratio of verti-
depleted zones. Another risk of commingled produc- cal to horizontal permeability is not too small.
tion is that downdip water or updip gas will advance Hydraulic fractures along a horizontal well can com-
to the well, resulting in early breakthrough of pensate for a productivity reduction caused by low
unwanted fluids in the most productive layer or lay- vertical permeability in a thick reservoir.
ers. In this case the oil in the lower productivity lay- Thick reservoirs with overlying gas or underlying
ers is bypassed. Reentry drilling offers a modern water pose special production problems for which
solution by targeting the bypassed oil with a horizon- chart row 2 on Fig. 1-21 illustrates some important
tal well. points. In vertical wells, a strategy to delay bottom-
Strategies for conformance control begin with per- water breakthrough is to perforate near the top of the
forating with a higher shot density in the lower pro- productive interval. However, the pressure gradient
ductivity layers. Hydraulic fracturing in layered resulting from radial flow toward the well is suffi-
reservoirs can be useful for conformance control, cient to draw the water upward in the shape of a
especially if the treatment is phased to target con- cone. Once the water reaches the deepest perfora-
trasting zones separately. Unphased, ill-designed tions, water may be preferentially produced because
hydraulic fracture treatments can be detrimental the water mobility may be greater than oil mobility
to production by opening up the high-productivity for low-gravity crudes (owing to the higher oil vis-
zones and aggravating the productivity imbalance. cosity) and/or there may be considerable energy to
A single horizontal well is not an option for a lay- support water production because of a strong bot-
ered reservoir because it produces from only one tomwater drive. Once water breakthrough occurs,

there may be little further rise of the cone, and addi- Interestingly, the same conditions that penalize a
tional oil production will be at an increasing water cut horizontal well in a reservoir without overlying gas
and marginal. One strategy to produce additional oil is or underlying water (thick zone, low vertical perme-
to plug back the well above the top of the cone and ability) favor the horizontal well if overlying gas or
reperforate. Another is to try to inject gel radially underlying water is present. This also illustrates
below the perforations. At times, water breakthrough designing the well spacing to be close enough to
is delayed or avoided with gel injection, and the shape cause interwell interference. The interwell or inter-
of the cone is widened in any case so that a greater lateral interference is beneficial in this case because
volume of oil is displaced toward the perforations. it both accelerates production and enhances recovery.
A horizontal well drilled near the top of the oil Another case that may favor close parallel lateral
zone above bottomwater produces a pressure gradi- spacing is in chart row 6 on Fig. 1-21. Although ori-
ent normal to the well, and the bottomwater will rise enting a horizontal well normal to natural fractures
in the shape of a crest instead of a cone. The crest- boosts well productivity, this approach may risk
shaped water advance displaces oil in its path, lead- early water breakthrough, especially in reservoirs
ing to greater oil recovery than with a vertical well under waterflood. Injecting water opposite of a bank
by virtue of the flow geometry. Ehlig-Economides of parallel laterals drilled at sufficiently close spac-
et al. (1996) discussed strategies for production ing may allow them to withdraw oil from the matrix
enhancement under a strong bottomwater drive. rock before the injected water front advances to the
Previous work cited from the literature has analytical production wells. Water may be injected above frac-
estimates for breakthrough time and indicates that turing pressure to boost injectivity. When horizontal
recovery efficiency is independent of the production or multilateral wells are not economically justified,
rate under a strong bottomwater drive. Ehlig- the likely short-circuiting of water between vertical
Economides et al. showed that the relationship well injector/producer pairs may be plugged by gel,
between recovery and the spacing of parallel hori- thereby forcing the displacement process into the
zontal wells is matrix rock.
The remaining rows 7 through 10 on the chart are
3zw h − 0.5 
kH kV 
h
rv = 0.5236 , (1-74) reminiscent of 3D reservoir geometries. Although
xe  2.5  conventional vertical wells do not address a 3D
reservoir geometry, hydraulically fractured and hori-
that recovery efficiency is a simple function of the zontal wells do, and knowledge of structural and
half-spacing between wells: stratigraphic reservoir heterogeneities can greatly
improve the design of these wells.
πrN h
qt BT = kH kV (1-75) Structural compartmentalization (chart row 7 on
3 2 xe Fig. 1-21) results from faults that may not be visible
in seismic data interpretations. Even if faults are
and that the optimal half-spacing between wells is clearly indicated in the seismic data, only dynamic
data derived from formation or well tests or longer
xe,opt = h kH kV . (1-76) term production history matching can establish
whether the faults are sealing or conductive.
In these three equations, rv is the fraction of the Stratigraphic compartmentalization (chart row 8) is
well drainage volume occupied by the crest at the a result of depositional processes. Facies with con-
time of water breakthrough. For the optimal well siderable contrasts in flow characteristics may serve
spacing from Eq. 1-76 and a well standoff from the as buffers or flow conduits that act as first-order con-
oil-water contact zw approximately equal to the trols on well productivity and ultimate hydrocarbon
thickness of the oil column h, the maximum water- recovery. Both structural and stratigraphic hetero-
free oil recovery (assuming piston-like displacement) geneities may be complicated by diagenetic proc-
is π/6 = 0.5236. In this case, the optimal interwell esses occurring at a later time.
spacing is most likely too close for conventional well Horizontal wells can target one or more reservoir
drilling but may be economical if the laterals can be compartments, and multibranch wells enable shut-off
drilled from a common main trunk. of a branch that produces unwanted gas or water. In

tight reservoirs with considerable faulting, the faults cant disadvantage in that flow is commingled among
may be associated with natural fractures that can be the layers, and when gas or water breakthrough occurs
targeted with horizontal wells, or they may provide it interferes with production from other layers. The
reliable information on the maximum stress direction better strategy may be to drill multiple horizontal
that is essential for planning hydraulic fractures in wells, each on strike and staying in a specific bed.
vertical or horizontal wells. The advantage to this strategy is that each of the wells
Stratigraphic limits (chart row 8 on Fig. 1-21) may is optimal in its standoff from the gas-oil or oil-water
account for additional reservoir compartmentaliza- contact, thus delaying multiphase production as long
tion, both vertically and areally. In some cases the as possible, and in its productive length within the for-
reservoir sands may be too thin to be individually mation, thus maximizing productivity.
identified in a seismic data cross section, but they
may have sufficient areal extent to be visible in seis-
mic attribute maps for a structural horizon. In that 1-7. Stimulation execution
case, horizontal wells may be an ideal strategy for A good understanding of job execution is necessary
producing thin formations and for reaching multiple for making decisions on the applicability and risk of
sands. various treatments. As with any well work, basic
Chart row 9 on Fig. 1-21 refers to elongated com- safety procedures must be developed and followed
partmentalization. Although these diagrams depict to prevent catastrophic failure of the treatment,
fluvial reservoir geology, elongated reservoirs can which could result in damage to or loss of the well,
also occur in heavily faulted formations. In either personnel and equipment. Specific standards and
case, the apparent drilling strategies depend on the operating procedures have been developed for stimu-
objective for the well. For example, the well direc- lation treatments, which if followed can lead to a
tion can be planned to stay in an elongated reservoir safe, smooth and predictable operation. Chapters 11
body or to intersect as many reservoir bodies as pos- and 19 fully detail execution concerns.
sible. The latter case implies drilling in the direction
normal to the elongation, which for a fluvial reser-
voir means drilling normal to the downslope direc- 1-7.1. Matrix stimulation
tion at the time of deposition. Another approach may
Matrix stimulation, mainly acidizing, is the original
be a multibranch well designed to target channels
and simplest stimulation treatment. More than
identified with borehole seismic measurements in the
40,000 acid treatments are pumped each year in oil
horizontal trunk well.
and gas wells. These treatments (Fig. 1-22) typically
Hydraulic fracturing offers different challenges
involve small crews and minimal equipment. The
and possibilities. First, unlike a well trajectory plan,
equipment usually consists of one low-horsepower,
the direction of the hydraulic fracture is not a design
single-action reciprocating pump, a supply centrifu-
choice. Rather, the fracture propagates normal to the
gal and storage tanks for the acid and flush fluids.
direction of minimum stress. A hydraulic fracture
Blending equipment is used when solids are added
may propagate into isolated sand bodies not con-
to the treatment.
tacted by the drilled well trajectory, but in other
The most common process is for the fluids to be
cases the fracture propagation may be inhibited by
preblended at the service company facility and then
facies changes or structural discontinuities, and a
transported to the location. This allows blending
screenout may occur. In general, drilling solutions
small volumes accurately, controlling environmental
may be more flexible in elongated reservoir systems.
hazards. The fluids are then pumped with little effort
The last chart row on Fig. 1-21 is for the special
or quality risk.
geometry of the attic compartment. In this case,
steeply dipping beds may be in contact with an updip
gas cap, downdip aquifer or both. One strategy is to 1-7.2. Hydraulic fracturing
drill a horizontal well that passes through several of
the beds and stays sufficiently below the updip gas Unlike matrix stimulation, fracturing can be one of
and above the downdip water. Although this seems the more complex procedures performed on a well
to be an efficient approach, it suffers from a signifi- (Fig. 1-23). This is due in part to the high rates and

Figure 1-22. Matrix stimulation treatment using a coiled tubing unit, pump truck and fluid transport.
pressures, large volume of materials injected, contin-

uous blending of materials and large amount of
unknown variables for sound engineering design.
The fracturing pressure is generated by single-
action reciprocating pumping units that have
between 700 and 2000 hydraulic horsepower (Fig.
1-24). These units are powered by diesel, turbine or
electric engines. The pumps are purpose-built and
have not only horsepower limits but job specification
limits. These limits are normally known (e.g., smaller
plungers provide a higher working pressure and
lower rates). Because of the erosive nature of the
materials (i.e., proppant) high pump efficiency must
be maintained or pump failure may occur. The limits
are typically met when using high fluid velocities
and high proppant concentrations (+18 ppg). There
may be numerous pumps on a job, depending on the
design.
Mixing equipment blends the fracturing fluid sys- Figure 1-23. This large fracturing treatment used 25,000
tem, adds the proppant and supplies this mixture to hydraulic horsepower and 1.54 million gal of fracturing fluid
to place 6.3 million lbm of propping agent. The job lasted
the high-pressure pumps. The slurry can be continu- 11 hours.
ously mixed by the equipment (Fig. 1-25) or batch
mixed in the fluid storage tanks. The batch-mixed
fluid is then blended with proppant in a continuous
stream and fed to the pumps.

Figure 1-24. One thousand hydraulic horespower pumping unit.
Figure 1-25. For this fracturing treatment, propping agent

was introduced into the fracturing fluid via conveyors to the
blender. The blender added the propping agent to the con-
tinuously mixed fracturing fluid (creating a slurry) and dis-
charged the slurry to the high-pressure pumping equip-
ment.

◆
Formation Characterization:
Well and Reservoir Testing
Christine A. Ehlig-Economides, Schlumberger GeoQuest
2-1. Evolution of a technology 6000
The evolution of well and reservoir testing has gone

through three important milestones, each of which Apparent line 1
5000
shaped both the manner with which well tests are
Apparent line 2
interpreted and the information that can be extracted
pws (psi)
4000
from them. These three major developments are the
semilogarithmic straight line (Horner analysis), log-
Apparent line 3
log diagnostic plot and log-based derivative. They 3000
are briefly outlined in the following text. Apparent line 4
2000
100 101 102 103 104 105 106 107
2-1.1. Horner semilogarithmic analysis (tp + ∆t)/∆t
Using the semilogarithmic approximation of the
Figure 2-1. Analysis of pressure buildup data on a semilog
solution of the partial differential equation (Eq. 1-18) plot. Arrows denote beginning and end of semilog linear
shown in Chapter 1 and employing the superposition trends.
principle, Horner (1951) presented the mainstay for
buildup analysis, which, appropriately, was named
after him. From the extension of the correct straight line to
Presuming infinite-acting radial flow, the expres- t = 1 hr, the value of the pressure p1hr can be extracted,
sion for the shut-in pressure pws in psi is and the Horner analysis suggests that the skin effect
s can be calculated by
162.6qBµ t + ∆t
pws = pi − log p , (2-1)  p1hr − pwf ( ∆t =0 ) k 
kh ∆t s = 1.151 − log + 3.23 . (2-3)
 m φµct rw
2

where pi is the initial reservoir pressure in psi, q is
the rate during the flowing period in STB/D, B is The value of pwf (∆t = 0) is the last value of the bot-
the formation volume factor in RB/STB, µ is the tomhole flowing pressure, m is the slope of the line,
viscosity in cp, k is the permeability in md, h is the φ is the porosity (unitless), ct is the total compress-
reservoir thickness in ft, tp is the producing (flow- ibility in psi–1, rw is the wellbore radius in ft, and the
ing) time in hr, and ∆t is the time since shut-in in hr. constant 3.23 is to account for oilfield units and the
A semilogarithmic plot of log([tp + ∆t]/∆t) versus conversion from ln to log.
pws should form a straight line (Fig. 2-1) with the For a drawdown (flowing) test the analogous
slope equal to semilogarithmic straight-line equation is
162.6qBµ 162.6qBµ  k 
m=− , (2-2) pwf = pi −  logt + log − 3.23 + 0.87s
kh kh  φµct rw
2

from which the unknown k, or kh if h is also not
known, can be determined. Although from a visual (2-4)
observation several straight lines through the data are and for skin effect
generally plausible, the question of which of them is  p − p1hr k 
the correct one is resolved in the next section. s = 1.151 i − log + 3.23 . (2-5)
 m φµct rw
2


Two reasons make buildup tests far more popular At infinite shut-in time (i.e., at (tp + ∆t)/∆t = 1),
(and reliable) than drawdown tests: the straight line on the Horner plot should intercept
the pressure at pi (for a new reservoir).
• Both solutions imply constant rate q. Although
The problem with semilogarithmic constructions
this is difficult to accomplish for drawdown, for
from both drawdown and buildup data is that the
buildup the rate is constant and, simply, equal to
correct straight line (i.e., which data fall on it) is
zero. Rate fluctuations before the buildup can be
often difficult to identify, as can be seen readily in
“smoothed” by defining an equivalent production
Fig. 2-1. This dilemma was resolved by the second
time as
major development in modern well testing.
Np
t *p = , (2-6)
qlast
2-1.2. Log-log plot
where Np is the cumulative production and qlast is
the last flow rate. Equation 2-6 can be shown to H. J. Ramey and his coworkers introduced the log-
be a reasonable approximation, fundamentally log plot as a means to diagnose the well pressure
based on the superposition principle. transient response. The first of these landmark papers
• The initial reservoir pressure pi required for the is by Agarwal et al. (1970).
drawdown analysis (Eq. 2-5) is rarely known with Figure 2-2 presents some of the common pressure
certainty, especially in a new formation. Buildup response patterns for a well test. Early-time wellbore
analysis not only does not require pi, it can deter- storage effects manifest themselves with a unit slope
mine its value. on the log-log plot. Figure 2-2 contains two sets of
curves. The first set, to the left, represents reduced
10 4
10 3
∆p and derivative (psi)
10 2
10 1
Downhole buildup pressure change

Downhole buildup derivative
Surface buildup pressure change
Surface buildup derivative
10 0
10 –4 10 –3 10 –2 10 –1 10 0 10 1 10 2
∆t (hr)
Figure 2-2. Log-log plot of pressure buildup data.
2-2 Formation Characterization: Well and Reservoir Testing

wellbore storage that can be accomplished with a analysis would not be appropriate for the interpre-
downhole shut-in for pressure buildup. The second tation of even a long test.
set is the response with surface shut-in or a draw- • Different reservoir features may result in practically
down test. The upper curve in both sets is the pres- indistinguishable pressure responses, especially in
sure response; the bottom is the pressure derivative. reasonably short well tests. Therefore, there is the
The latter is explained and justified in the next sec- issue of uniqueness in the interpretation.
tion. Although the reader is not yet familiar with
other pressure/pressure derivative responses, the Type-curve matching by superimposing relatively
minimization of wellbore storage effects may reveal short-duration field data over the mathematical
certain early-time patterns that are otherwise dis- model solutions has been attempted but with fre-
torted or totally masked by uncontrolled, lengthy quent problems of uniqueness. The technique
wellbore storage effects. involves plotting the mathematical (dimensionless)
Mathematically, the relationship of dimensionless solution to a problem and the real data in identical
pressure pD (which is exactly proportional to the real log-log formats. Keeping the axes parallel, the data
∆p) versus dimensionless time tD during dominant are matched with a portion of the solution and the
wellbore storage effects is overlying coordinates are determined. From the rela-
tionships between the mathematical and real vari-
tD ables, missing parameters such as permeability,
pD = , (2-7)
CD porosity or fracture length are calculated. Type-curve
matching has not been proved to be a particularly
where CD is the dimensionless wellbore storage coef- successful exercise, especially because it is not sen-
ficient (defined in Section 2-3.5). sitive to changes in pressure. These changes can
Agarwal et al. (1970) also suggested a rule of denote important phenomena—but with subtly dif-
thumb according to which infinite-acting radial flow ferent responses.
would be separated from the end of wellbore storage It is in this environment that the pressure deriva-
effects by 11⁄2 log cycles of time. Data from after this tive emerged.
transition period can be plotted on the semilogarith-
mic plot and analyzed as suggested in the previous
section.
Thus, well test analysis became a technology con-
2-2. Pressure derivative in well
sisting of diagnosis using the log-log plot in a pattern test diagnosis
recognition exercise to find the beginning of the cor- When the dimensionless pressure pD is differentiated
rect straight line followed by the semilogarithmic with respect to the natural logarithm of dimension-
plot for permeability and skin effect determination. less time tD, then
For the drawdown log-log diagnostic plot, the
dpD dp
appropriate variables to plot are pi – pwf versus t = t D D = t D pD′ , ( 2-8)
(again, pi is most likely unknown) and for buildup d ( lnt D ) dt D
the far more convenient pws – pwf (∆t = 0) versus ∆t.
where pD′ is the dimensionless pressure derivative
There are three problems with the log-log plot of
with respect to dimensionless time tD.
pressure difference versus time, and they affect both
The use of this particular form of pressure deriva-
the likelihood of Horner analysis and the unique
tive represents a major advancement in pressure tran-
determination of other reservoir and well variables:
sient analysis. It was first presented to the petroleum
• Well tests are usually shorter than required to enter literature by Bourdet et al. (1983). Figure 2-3 repre-
fully developed infinite-acting radial flow. This is sents the complete solution of Gringarten et al.’s
particularly true for low-permeability reservoirs, (1979) work for an infinite-acting reservoir, comple-
and in such cases the use of Horner analysis would mented by the pressure derivative as developed by
be inappropriate. Bourdet et al. (1983).
• Other geometries or reservoir and well features During wellbore storage effects, the dimensionless
such as fractures and dual-porosity systems may pressure is related to dimensionless time and dimen-
affect the test response. In such cases, Horner

100
CDe2s
1030
20
10
1015
1010
108
10 106
pD and p′D (t D /CD) 104
103
102
10 3
0.3 0.1
0.1
0.1 1 10 100 1000 10,000
Dimensionless time, t D /CD
Figure 2-3. Dimensionless type curves for pressure drawdown and derivative for an infinite-acting reservoir with wellbore
storage and skin effect (see discussion of type-curve use in Bourdet et al., 1983).
sionless wellbore storage by Eq. 2-7, which, when analytical solution. More importantly, the derivative
differentiated and combined with Eq. 2-8, yields is invaluable for definitive diagnosis of the test
response. Although pressure trends can be confusing
dpD d (lnt D ) = t D pD′ = t D CD . (2-9) at “middle” and “late” times, and thus subject to
multiple interpretations, the pressure derivative val-
On log-log paper, this shows a unit straight line ues are much more definitive. (The terms early, mid-
exactly as does the dimensionless pressure. dle and late time are pejorative expressions for early-,
During the radial flow period and when the semi- midway- and late-appearing phenomena. For exam-
logarithmic approximation is in effect (Eq. 1-19), ple, wellbore storage effects are early, fracture
behavior is middle, and infinite-acting radial flow
dpD d (lnt D ) = t D pD′ = 0.5 , (2-10) or boundary effects are late.) Many analysts have
come to rely on the log-log pressure/pressure deriva-
and, thus, the dimensionless derivative curve at late
tive plot for diagnosing what reservoir model is rep-
time approaches a constant value equal to 0.5.
resented in a given pressure transient data set. To
In general, if
apply this method of analysis, the derivative of the
pD ~ t Dm , (2-11) actual pressure data must be calculated. A variety of
algorithms is available. The simplest is to calculate
where m is equal to 1.0 for wellbore storage, 0.5 for the slope for each segment, using at least three time
linear flow and 0.25 for bilinear flow, then intervals. More sophisticated techniques also may be
contemplated.
dpD d (lnt D ) = t D ( dpD dt D ) ~ mt Dm , (2-12) Patterns visible in the log-log and semilog plots
for several common reservoir systems are shown in
which on log-log coordinates implies that the deriva- Fig. 2-4. The simulated curves in Fig. 2-4 were gen-
tive curve is parallel to the pressure curve departed erated from analytical models. In each case, the
vertically by log m. buildup response was computed using superposition.
The derivative is useful in pressure transient The curves on the left represent buildup responses,
analysis, because not only the pressure curve but and the derivatives were computed with respect to
also the pressure derivative curve must match the the Horner time function.

Log-Log
Diagnostic Horner Plot Specialized Plot
Wellbore storage
Infinite-acting
A radial flow
C From specialized
plot
Wellbore storage
B Partial penetration
Infinite-acting
radial flow
Linear flow to an
infinite-conductivity
C vertical fracture
2
kxf From specialized
plot
Bilinear flow to a
finite-conductivity
D vertical fracture
kfw From specialized
plot
Wellbore storage
E Infinite-acting
radial flow
Sealing fault
Wellbore storage
F No-flow boundary
Wellbore storage
Linear channel
G flow
kb2 From specialized
plot
Wellbore storage
H Dual-porosity
matrix to fissure
flow (pseudosteady
state)
Figure 2-4. Log-log and semilog plots for common reservoir systems.

Patterns in the pressure derivative that are charac- effect of a closed drainage volume. Unlike the draw-
teristic of a particular reservoir model are shown down pressure transient, which sees the unit slope in
with a different line, which is also reproduced on the late time as indicative of pseudosteady-state flow, the
Horner plot. In cases where the diagnosed behavior buildup pressure derivative drops to zero (Proano
can be analyzed as a straight line with a suitable and Lilley, 1986).
change in the time axis, the curves are shown as spe- When the pressure and its derivative are parallel
cialized plots in the third column. Determination of with a slope of 1⁄2 in late time, the response may be
the lines drawn on the Horner plots for each example that of a well in a channel-shaped reservoir (Ehlig-
was based on the diagnosis of radial flow using the Economides and Economides, 1985), as in example
derivative. G. The specialized plot of pressure versus the square
Example A illustrates the most common response, root of time is proportional to kb2, where b is the
a homogeneous reservoir with wellbore storage and width of the channel.
skin effect. The derivative of wellbore storage tran- Finally, in example H the valley in the pressure
sients is recognized as a hump in early time (Bourdet derivative is an indication of reservoir heterogeneity.
et al., 1983). The flat derivative portion in late time In this case, the feature is due to dual-porosity
is easily analyzed as the Horner semilog straight behavior (Bourdet et al., 1984).
line. In example B, the wellbore storage hump leads Figure 2-4 clearly shows the value of the pressure/
into a near plateau in the derivative, followed by a pressure derivative presentation. An important
drop in the derivative curve to a final flat portion. A advantage of the log-log presentation is that the tran-
plateau followed by a transition to a lower plateau is sient patterns have a standard appearance as long as
an indication of partial penetration (Bilhartz and the data are plotted with square log cycles. The vis-
Ramey, 1977). The early-time plateau represents ual patterns in semilog plots are enabled by adjusting
radial flow in an effective thickness equal to that of the range of the vertical axis. Without adjustment,
the interval open to flow into the partially penetrat- much or all of the data may appear to lie on one line,
ing wellbore. Later, radial flow streamlines emanate and subtle changes can be overlooked.
from the entire formation thickness. The effects of Some of the pressure derivative patterns shown are
partial penetration are rarely seen, except in tests that similar to those characteristic of other models. For
employ a downhole shut-in device or the convolu- example, the pressure derivative doubling associated
tion of measured downhole flow rates with pressure with a fault (example E) can also be an indication of
(Ehlig-Economides et al., 1986). transient interporosity flow in a dual-porosity system
Examples C and D show the behavior of vertical (Bourdet et al., 1984). The abrupt drop in the pressure
fractures (see Chapter 12). The half-slope in both the derivative in the buildup data can indicate either a
pressure change and its derivative results in two par- closed outer boundary or a constant-pressure outer
allel lines during the flow regime representing linear boundary resulting from a gas cap, aquifer or pattern
flow to the fracture. The quarter-sloping parallel lines injection wells (Proano and Lilley, 1986). The valley
in example D are indicative of bilinear flow. During in the pressure derivative (example H) could be an
linear flow, the data can be plotted as pressure versus indication of a layered system instead of dual porosity
the square root of ∆t, and the slope of the line in the (Bourdet, 1985). For these cases and others, the ana-
specialized plot is inversely proportional to √kxf2, lyst should consult geological, seismic or core analy-
where xf is the vertical fracture half-length in ft. sis data to decide which model to use for interpreta-
During bilinear flow, a plot of pressure versus the tion. With additional data, there may be a more con-
fourth root of ∆t gives a line with the slope inversely clusive interpretation for a given transient data set.
proportional to 4√k(kfw), where kf is the fracture per- An important use for pressure/pressure derivative
meability in md and w is the fracture width in ft. diagnosis is at the wellsite. The log-log plot drawn
Example E shows a homogeneous reservoir with during transient data acquisition can be used to
a single vertical planar barrier to flow or a fault. The determine when sufficient data have been collected
level of the second derivative plateau is twice the to adequately define the infinite-acting radial flow
value of the level of the first derivative plateau, and trend. If the objective of the test is to determine per-
the Horner plot shows the familiar slope-doubling meability and skin effect, the test can be terminated
effect (Horner, 1951). Example F illustrates the once the derivative plateau is identified. If hetero-

geneities or boundary effects are detected in the tran- ability can be determined from a rearrangement of
sient, the test can be run longer to record the entire Eq. 2-2:
pressure/pressure derivative response pattern
1.151α p qBµ
required for analysis. k= , (2-13)
mh
where m is the absolute value of the slope of a semi-
2-3. Parameter estimation from log line. The unit conversion constants in this and
other equations are provided in Table 2-1.
pressure transient data For vertical wells, horizontal (bedding plane) per-
The patterns identified in pressure transient data are meability is determined from radial flow. Natural
easily recognized either by their shape or their deriv- fractures and depositional features such as cross-
ative slope on the log-log diagnostic plot. Each of bedding give rise to a preferential flow direction
these patterns reflects a flow geometry in the reser- or permeability anisotropy in the bedding plane. In
voir, which, when identified, enables the computa- these cases, the horizontal permeability determined
tion of well and/or reservoir parameters. Ehlig- from analysis of the radial flow regime is actually
Economides (1995) summarized various computa- the geometric mean of the maximum permeability
tions based on flow regime equations. kx directed parallel to the principal permeability axis
and of the minimum permeability ky directed perpen-
dicular to it, as given by √kxky.
2-3.1. Radial flow Figure 2-5a shows radial flow toward a portion of
Radial flow is illustrated by the flow streamline the vertical wellbore. This occurs initially when the
geometries shown in Fig. 2-5. In each case, radial well only partially penetrates the formation, when
flow is characterized by flow converging to a line the well is only partially completed in the formation
at the center of a circular cylinder. The parameters thickness or when drilling or completion damage
sensed from radial flow are permeability, skin effect restricts flow to the well. In time, the radial flow
and reservoir pressure. In addition, the onset time for expands through the entire thickness, as in Fig. 2-5b.
radial flow indicates the effective radius of the cylin- For horizontal wells, the radial flow regime about
der to which the flow converges, and the departure the well (Fig. 2-5c) represents the geometric mean of
time from radial flow indicates the distance to what- the horizontal permeability kH (or √kxky) and the ver-
ever feature serves as an obstacle to continued radial tical permeability kV (or kz), given by √kHkV, or, more
propagation of the pressure signal in the formation. precisely, √kykz (if the well is oriented parallel to the
Analysis of the radial flow regime quantifies the principal permeability axis). For this case, the equa-
permeability in the plane of convergent flow (normal tion for the slope of the semilog line is
to the line source or sink). On a semilogarithmic plot 1.151α p qBµ
of pressure versus elapsed time, the reservoir perme- k y kz = , (2-14)
mepr Lp
(a) Partial radial flow (b) Complete radial flow (c) Radial flow to horizontal well
(d) Pseudoradial flow to fracture (e) Pseudoradial flow to horizontal well (f) Pseudoradial flow to sealing fault
Figure 2-5. Radial flow geometries.

Table 2-1. Unit conversion factors and constants. The shaded zones in each of the flow regime dia-
grams depict the approximate volume traversed by
Quantity Oilfield SI the expanding pressure disturbance. The time of
Unit Unit
departure from a flow regime trend on the pressure
Production rate, q STB/D m3/s derivative tdep corresponds to the distance d to what-
Formation thickness, h ft m ever flow barrier inhibits continued radial expansion
Permeability, k md m2 of the pressure perturbation according to the follow-
Viscosity, µ cp Pa⋅s
ing equation:
Pressure difference, ∆p psi Pa α t ktdep
d=2 . (2-15)
Pressure, p psi Pa φµct
Radius, r ft m
When the transient response ends with a level
Fracture half-length, xf ft m
derivative, the test radius of investigation is com-
Fracture width, w in. cm puted with Eq. 2-15, with tdep equal to the elapsed
Fracture stiffness, s f psi/in. Pa/m time associated with the last data point. An upward
Distance to sealing fault, L ft m departure from a level pressure derivative trend
Channel width, b ft m corresponds to a flow barrier at distance d. A down-
Productive length, Lp ft m
ward departure corresponds to an increase (outside
a radius of d) in k, kh, k/µ or kh/µ or to a constant-
Time, t hr s
pressure boundary.
Porosity, φ (fraction) Unitless Unitless The permeability in Eq. 2-13 is in the direction of
Total system the reservoir feature found at distance d. For a hori-
compressibility, c t psi–1 Pa–1
zontal well, departure from early radial flow occurs
Porosity-compressibility- when the pressure signal reaches a bed boundary. A
thickness product, φc t h ft/psi m/Pa
deflection for the bed boundary more distant from the
Wellbore storage, C bbl/psi m3/Pa
well will occur if the borehole is much closer to one
Skin effect, s Unitless Unitless bed boundary. When the position of the well is known
αp 141.2 1/(2π) = 0.1592 between bed boundaries, the departure time(s) can be
αt 0.000264 1 used to compute vertical permeability (Eq. 2-15).
αc 24 1 The onset of pseudoradial flow corresponds to
αlf 4.06 1/(2√π) = 0.2821
a distance computed with Eq. 2-15 of 10 times the
effective radial flow radius, or 10rw′, for hydraulically
αhl 4.06 0.2821
fractured and horizontal wells. For massive hydraulic
αbf 44.1 0.3896
fractures and long horizontal wells, the onset of radial
αpp 2453 0.0049 flow does not appear for a considerable length of
αf 0.000148 0.7493 time. Similarly, the onset of radial flow following the
αcf 8.168 1/(√π) = 0.5642 identification of slope doubling resulting from a bar-
rier corresponds to a radius 10 times the distance
where the subscript epr refers to early pseudoradial. between the well and the barrier. The onset of radial
Pseudoradial flow refers to radial flow converging flow after evidence of intersecting barriers to flow
to an effective wellbore radius rw′ larger than the corresponds to about 10 times the distance from the
well, such as to a vertical (hydraulic) infinite-con- well to the barrier intersection, and the level of the
ductivity fracture (rw′ = xf /2) or to a horizontal well derivative above that observed before evidence of the
(rw′ = Lp/4, where Lp is the productive length). These nearest flow barrier increases inversely with the sine
cases are illustrated in Figs. 2-5d and 2-5e, respec- of the angle between the barriers.
tively. Pseudoradial flow can also occur after the Logically, the permeability computed from the
pressure signal has propagated beyond one or more radial flow regime (Eq. 2-13) corresponds to the aver-
sealing boundaries (faults or stratigraphic limits), as age permeability between the radii corresponding to
illustrated in Fig. 2-5f. the onset to and the departure from radial flow.

The skin effect computed from radial flow (Eq. 2-3) The portion of the data exhibiting a linear flow
depends on the effective radius of the flow. For exam- trend (half-slope derivative) can be analyzed by plot-
ple, the skin effect computed from radial flow to a ting pressure against the square root of time. The
partial penetration or completion corresponds to the equation for the slope of the straight-line portion of
mechanical skin effect along the flowing interval, but pressure versus the square root of time is as follows
the skin effect computed later based on radial flow in for linear flow to a fracture:
the entire formation thickness includes the sum of skin
 α qB   µ 
1/ 2
effect components corresponding to mechanical skin
xf k =  lf    (2-16)
effect and an apparent skin effect caused by vertical  mlf h   φct 
flow convergence. Similarly, the skin effect computed
for linear flow to a horizontal well:
from pseudoradial flow corresponds to an apparent
1/ 2
skin effect dominated by the stimulation effect associ-  α qB   µ 
ated with a larger effective wellbore. Lp k =  hl    (2-17)
 mhl h   φct 
The average reservoir pressure is determined from
pressure buildup test data by extrapolating the radial and for linear flow inside an elongated reservoir:
flow trend to a Horner time of 1 on a Horner plot. The  α qB   µ 
1/ 2
extrapolated pressure p* is used to determine average b k =  cf    . (2-18)

pressure when the approximate drainage shape is  mcf h   φct 
known using the Matthews, Brons, Hazebroek (MBH) Each of these equations relates permeability and the
analysis described in Earlougher (1977). width of the linear flow (xf for the vertical fracture, Lp
for the horizontal well and b for the elongated reser-
voir). The permeability affecting linear flow analysis
2-3.2. Linear flow is in the direction of the linear flow streamlines.
Linear flow is the second most commonly observed In addition to Eqs. 2-16 through 2-18, departure
flow regime. It is characterized by entirely parallel from linear flow corresponds to the following equa-
flow in the formation and can result either because tions for a vertical fracture:
of the well completion or trajectory geometry or
φµct x 2f
because of outer reservoir boundaries. Figure 2-6a t= (2-19)
illustrates linear flow to a vertical fracture plane, 16α t k x
and Fig. 2-6b shows linear flow to a horizontal well. and for a horizontal well:
Both of these flow regimes occur before pseudo-
radial flow, and their duration is dependent on the φµct L2p
t= . (2-20)
fracture half-length, or the productive length in the 64α t k x
case of a horizontal well. Figure 2-6c shows linear Each of these equations also relates permeability
flow resulting from the elongated shape of the reser- and the width of the linear flow. Thus, when the linear
voir. This can be observed in wells located between flow regime and a departure from it are identified in
parallel faults or in elongated sands such as fluvial or the transient data, both parameters can be estimated.
deep marine channels. Linear flow can also occur
to a shallow horizontal fracture or a thin high-
permeability bed.
(a) Linear flow to fracture (b) Linear flow to horizontal well (c) Linear flow in elongated reservoir
Figure 2-6. Linear flow geometries.

For a well in an elongated reservoir, 2-3.3. Spherical flow

πφµct d 2
t= , (2-21) Spherical and hemispherical flow regimes are illus-
16α t k x trated in Fig. 2-8 as streamlines converging to a
where d corresponds to the distance between the well point. These flow regimes appear in partially pene-
and a barrier or a constant-pressure boundary normal trated and partially completed wells. The parameters
to the reservoir elongation. Equation 2-21 enables sensed by spherical flow are the spherical permeabil-
calculation of the distance d. ity, given by ksph = √kHkV, and the distances between
For finite-conductivity hydraulic fractures, bilinear the flowing interval and the bed boundaries.
flow, as shown in Fig. 2-7, may occur before or instead The spherical flow plot is pressure versus the reci-
of linear flow. The slope of the straight line on a plot of procal of the square root of elapsed time. The equa-
pressure versus the fourth root of elapsed time relates to tion for determining the spherical permeability from
fracture parameters as the following: the straight-line portion of a spherical flow plot is
2/3
 α qBµ 
2
 α qBµ   1 
1/ 2
k f w k =  bf (2-22) ksph =  pp φµct  . (2-24)

   .  m pp 
 mbf h   φµct 
The time of departure from bilinear flow is depen- In addition, the departure from spherical flow
dent on the fracture half-length, as for linear flow caused by the nearest bed boundary to the flowing
according to Eq. 2-19, when the departure is concave interval occurs at the time satisfying the following:
down toward radial flow. If linear flow follows bilin- φµct zw2
ear flow, Eq. 2-19 applies only for the departure t= , (2-25)
36α t kV
from linear flow. Bilinear flow involves the fracture
conductivity kfw, fracture half-length and reservoir where zw is the elevation of the midpoint of the per-
permeability ky. After bilinear flow, if linear flow forations from the bottom of the reservoir.
appears and a departure from it, then fracture con- For hemispherical flow that occurs after the near-
ductivity, fracture half-length and reservoir perme- est bed boundary has been sensed, the constant in
ability can all be determined by Eqs. 2-16, 2-19 and Eq. 2-21 is 18 in place of 36.
the following: A test can be designed to observe spherical flow
in a pilot hole prior to drilling a horizontal well. This
kf w  π m  1  strategy enables determination of the formation verti-
=  , (2-23)
kx f  3  1.151  ∆pint  cal permeability, which, in turn, enables much more
accurate forecasts of the horizontal well productivity.
where ∆pint is the difference in the pressure intercept.
In some cases such a test may discourage drilling the
Unfortunately, for horizontal wells and hydrauli-
horizontal well if the expected productivity is not suf-
cally fractured vertical wells, sufficient flow regime
ficient to justify it. Further, conducting the pilot hole
variation for complete analysis is typically lacking.
test enables a more meaningful interpretation of a
Many well tests exhibit only linear or bilinear flow
subsequent test in the horizontal well because para-
transients in fractured wells. Horizontal well tests
meters determined from the pilot hole test do not
usually do not exhibit linear flow, rendering ambigu-
have to be redetermined from the horizontal well test.
ous the distinction between early radial flow regimes
and pseudoradial flow.
(a) Spherical flow (b) Hemispherical flow
Figure 2-7. Bilinear flow to a hydraulic fracture. Figure 2-8. Spherical and hemispherical flow geometries.

2-3.4. Dual porosity Dual-porosity reservoirs are likely to exhibit

highly anisotropic flow behavior. Highly laminated
Dual-porosity flow behavior results when reservoir systems usually have considerable contrast in hori-
rocks contain distributed internal heterogeneities that zontal and vertical permeability, whereas naturally
have highly contrasting flow characteristics such that fractured systems usually have a preferential flow
flow occurs mainly in a high-permeability formation direction oriented parallel to the natural fractures.
feature that accounts for a small fraction of the for- Laminated formations favor vertical wells; horizontal
mation storativity. Examples are naturally fractured wells are particularly attractive in naturally fractured
reservoirs and laminated systems with thin high-per- reservoirs. Because natural fractures are usually sub-
meability layers. vertical and oriented with the stress field, the maxi-
The two commonly observed dual-porosity tran- mum permeability and maximum stress directions
sient trends are shown in examples E and H in Fig. are usually closely aligned, resulting in enormous
2-4. Example E illustrates transient dual-porosity implications for hydraulic fracture design and trajec-
behavior typical of highly laminated systems, which tory planning in deviated wells.
is virtually impossible to distinguish from a sealing Observance of dual-porosity transients should trig-
fault without additional information. In example H, ger additional measurements or analysis to establish
which shows pseudosteady-state dual-porosity the direction and magnitude of the implicit perme-
behavior typical of naturally fractured formations, a ability anisotropies.
valley-shaped drop in the pressure derivative signals
recharge from matrix rock into the natural fractures.
Dual-porosity behavior can appear during any of the 2-3.5. Wellbore storage and
flow regimes and complicates transient analysis. pseudosteady state
Recognizing and characterizing dual-porosity flow
behavior is extremely important to reserves estima- Wellbore storage and pseudosteady state both result
tion, trajectory planning for deviated and horizontal from fluid compression or expansion in a limited
wells, and stimulation design and post-treatment volume. For wellbore storage the control volume is
evaluation. the wellbore; for pseudosteady state it is the reser-
The two parameters that characterize dual-porosity voir drainage volume. Both these cases are sensitive
systems are the storativity ratio ω and the interporos- mainly to two parameters, fluid compressibility and
ity flow parameter λ. For pseudosteady-state dual the control volume, but other factors determine their
porosity, the time of the onset of dual-porosity onset and duration. Both are recognized by a unit-
behavior ton is a function of both parameters: slope trend in the pressure derivative.
For wellbore storage, the pressure change and the
φµct rw2 7ω 2  1 
2
derivative coincide with a unit-slope trend on the
ton = ln . (2-26)
α t k λ(1 − ω )  ω  log-log diagnostic plot in early time. The derivative
typically departs below the pressure change and
Similarly, the time of the end of dual-porosity appears as a hump during the transition to reservoir-
behavior tend is dominated flow regimes. This is typically the first
flow regime observed in any test, and it usually dom-
φµct rw2 (1 − ω ) inates the transient response. The wellbore storage
tend = , (2-27)
αt k 7λ coefficient can be computed from the straight-line
and the time of the valley minimum tmin is given by portion of the plot of pressure change versus elapsed
Bourdet et al. (1983) as time using the following equation (the line must pass
through the origin on the plot):
1  φηct   
1 ω
tmin =  1 + ω −ω .
1− ω 1− ω
 (2-28) qB
2  αt k    C= . (2-29)
α c mc
When two of these three times can be identified
in the pressure derivative response, ω and λ can be The duration of wellbore storage can be reduced by
estimated. designing buildup tests with downhole shut-in. Other
factors affecting wellbore storage duration are perme-

ability (duration increases for low permeability) and and derivative data for each transient by the rate
skin effect (duration increases for a higher positive change that initiated the transient. Reservoir-domi-
skin effect). Because heterogeneities or flow geome- nated flow regime derivative transients plotted in this
tries located inside the radius of investigation of the manner should overlie each other, and differences in
pressure signal at the time of the end of wellbore stor- the pressure change trends are a sign of rate-depen-
age are masked by wellbore storage, it is important to dent skin effects. It is also possible to confirm the
minimize this phenomenon. accuracy of the surface rate sequence.
Pseudosteady state is observed only in drawdown In flow regime analysis, the parameters associated
tests. In buildup tests, pseudosteady state behavior is with each flow regime identified on the log-log diag-
distorted by superposition, which causes a downturn nostic plot are computed using the techniques and
in the derivative that is indistinguishable from the equations outlined in Section 2-3.
effect of a constant-pressure boundary. The time of the The parameters estimated from the flow regimes
onset of pseudosteady state is a function of the shape serve as a starting point for the nonlinear regression
and magnitude of the drainage area and the position of step. The objective of this step is to find a match for the
the well within it. Modern flow regime analysis of the entire transient response. To do this, a model must be
transient behavior observed in the pressure derivative selected that accounts for all identified flow regimes.
before the onset of pseudosteady state enables charac-
• Example post-treatment test in a hydraulically
terization of the well and drainage boundary geome-
fractured well
try, which, in turn, enables quantification of the aver-
age reservoir pressure. This example illustrates a transient test response
following a hydraulic fracture treatment in a high-
permeability reservoir. Figure 2-9 is a log-log diag-
nostic plot of the drawdown data, obtained with a
2-4. Test interpretation methodology downhole pressure-measuring device. The flow
Interpretation of transient test data involves the fol- regimes identified in the transient response are well-
lowing steps: bore storage, bilinear flow and radial flow. The flow
• data processing rate from this well was more than 3600 STB/D.
From the unit slope in the early-time portion,
• model diagnosis the wellbore storage constant C is computed as
• flow regime analysis 0.0065 bbl/psi. From the radial flow portion, the
• nonlinear regression. reservoir permeability k is computed as 12.5 md
and the radial skin effect is equal to –4.3. Then,
In the data processing step, the data for analysis
from the bilinear flow portion, the product kf w is
are extracted from the complete data set as a series
estimated as 1900 md-ft. From Fig. 1-11 a trial-and-
of one or more transients, each of which is a
error procedure is indicated. Assuming xf = 100 ft
response to a single step change in the surface rate.
(the designed length), dimensionless fracture conduc-
The transient data are reduced, usually with a routine
tivity CfD = (1900)/(12.5)(100) = 1.5 and from Fig.
designed to sample or filter the data logarithmically
1-11, sf + ln(xf /rw) = 1.3. Because rw = 0.328 ft, the
in elapsed time since the surface rate change. Poor-
calculated fracture half-length is 90 ft, which is
quality data may be excluded in this step.
close enough. With these as starting parameters,
For model diagnosis, the pressure change and its
an initial simulation for the transient response is
derivative are computed from the data for a single
compared to the data in Fig. 2-10. Nonlinear regres-
transient, incorporating all recent surface rate
sion allows the well and reservoir parameters to
changes in the superposition time used for data dif-
vary until an optimized match is found for the entire
ferentiation. For multirate tests, the data from more
response, as shown in Fig. 2-11. The final parameter
than one transient can be plotted together on the log-
estimates are C = 0.005 bbl/psi, k = 12 md,
log diagnostic plot by dividing the pressure change
xf = 106 ft and CfD = 1.6.

104
103
∆p, d∆p/dlnt (psi)
102
101
10 –3 10 –2 10 –1 100 101 102 103
∆t (hr)
Figure 2-9. Log-log diagnostic plot of post-treatment test after hydraulic fracturing.
104
103
102
101
10 –3 10 –2 10 –1 100 101 102 103
∆t (hr)
Figure 2-10. Initial match with post-treatment test data.

104
103
102
101
10 –3 10 –2 10 –1 100 101 102 103
∆t (hr)
Figure 2-11. Final match with post-treatment test data.
2-5. Analysis with measurement each can be used for model diagnosis in the same
fashion as the pressure change and derivative are
of layer rate used. This data acquisition and processing technique
When downhole shut-in is not an option, a buildup reduces the duration of wellbore storage in a draw-
test with surface shut-in can be dominated for much down test by the same amount as downhole shut-in
of the test duration by wellbore storage. An alterna- does in a buildup test.
tive in Fig. 2-12 to the conventional buildup test is Some wells with commingled flow from several
designed to acquire downhole measurements of both layers are equipped with sliding sleeves. This enables
flow rate and pressure using a production logging
tool. The best data acquired with such tests are dur-
ing drawdown, but additional data processing is
Electric line
required for model diagnosis. In this case an analog
for the pressure change is the rate-normalized pres-
sure (RNP), computed as the ratio of the pressure Flowmeter Pressure gauge
change to the flow rate change for data acquired at
the same instant in time. The pressure change (flow
rate change) is the difference between the bottom-
hole pressure (flow rate) measured at any elapsed
time t since the start of the test transient and the bot-
tomhole pressure (flow rate) measured at the start
of that transient. The analog for the pressure deriva-
tive is the deconvolution derivative, computed as a
derivative of the RNP, or the convolution derivative,
which accounts for superposition effects caused by
each change in the continuously acquired downhole
rate. Both computations account for superposition
resulting from recent changes in the surface rate, and Figure 2-12. Acquisition of transient flow rate and pres-
sure data.

flow from a particular layer to be shut off by shutting test may show which layer skins have been lowered
the sleeve. Such completions allow a more direct test by a recent stimulation treatment.
of the layer, as diagrammed in Fig. 2-13. The figure The sequence of the multilayer transient test is the
shows two flowmeters spaced above and below the key to its success. This test merges stabilized and tran-
flow ports of the sliding sleeve. The flowmeter above sient flow rate and pressure measurements using a
the sleeve opening measures the flow rate q2 from lay- production logging tool. A typical test sequence is
ers 2 and 3, and the lower flowmeter measures the illustrated in Fig. 2-14. Beginning with the well shut-
flow rate q1 from layer 3 only. A simple subtraction, in, a flow rate survey is acquired, and then the tool is
q2(t) – q1(t), enables direct measurement of the flow positioned above the lowest layer. After a brief pause
from the layer, which can be used in the RNP analysis while the sensors equilibrate to the wellbore condi-
described in the previous paragraph. tions at this depth, the flow rate is increased at the sur-
face while leaving the tool stationary at this depth.
When flow and pressure have stabilized at radial,
pseudosteady-state or steady-state conditions, or when
Sliding sleeve in the desired behavior has been observed, another flow
open position
Layer 1 rate survey is acquired. Then the tool is positioned
above the next lowest layer to repeat the same proce-
dure at this depth, taking care to acquire the stabilized
Pressure gauge
profile data after the same length of time following the
Layer 2 rate change as for the first measurement sequence.
Paired flowmeters
above and below
When stationary measurements have been acquired
sleeve opening above each of the intervals of interest and stabilized
flow surveys have been acquired for at least three sur-
Layer 2 face flow rates, the tool is positioned for an optional
final buildup test. The final tool position is optional;
it is worthwhile to position the tool where wellbore
crossflow could occur while the well is shut in.
Figure 2-13. Acquisition of transient flow rate and pres-

sure data for a single layer in a commingled completion. Layer 1
3
Layer 2
Flow rate (B/D)
3
2-6. Layered reservoir testing Layer 3
Layered formations pose special problems for reser-

2
voir management that can be best addressed with a
layer-by-layer characterization of reservoir parame-
1
ters. The multilayer transient test is designed to pro-
vide the average pressure, productivity index, and Time (hr)
well and reservoir parameters for two or more layers
commingled in a common wellbore. When a contrast Figure 2-14. Test sequence for a multilayer transient test.
in performance is apparent in commingled layers,
this test can determine whether the contrast is due to
large variations in layer kh values or to large varia-
tions in skin effect. In the former case, there may be 2-6.1. Selective inflow performance
implications for waterflood vertical displacement analysis
efficiency. In the latter case, there may be a work- Selective inflow performance (SIP) analysis extends
over treatment that would improve the performance the concepts of flow-after-flow or isochronal tests to
of layers with higher skin factors. Alternatively, the multiple layers. Each time a flow profile is acquired,

subtracting the flow rate measured above a layer The analysis then proceeds to the next lowest
from that measured below it gives the flow rate to reservoir layer. Diagnosis of the model for this and
or from that layer. For SIP analysis, an inflow perfor- shallower reservoir layers is more difficult than for
mance plot of datum pressure versus layer flow rate the lowest layer. Again, the RNP and its derivative
provides the layer average pressures and productivity are computed, but this time the behavior observed
indices, as shown in Fig. 2-15. For oil wells, the is that of the lowest two layers combined. To see the
slope of each line is the reciprocal of the layer pro- behavior of the next lowest layer by itself requires
ductivity index J. For gas wells, a quadratic fit of the additional processing. However, by assuming that
data provides, for each layer tested, the productivity the model for the second layer is similar to that of
index and an estimate of the coefficient D associated the lowest layer, parameters for the second layer
with the rate-dependent skin effect caused by turbu- model can be estimated. Again, nonlinear regression
lent flow near the wellbore. In water injection wells, is used to refine the estimates, this time using a two-
SIP analysis can be useful for estimating the forma- layer model with the parameters for the lowest layer
tion parting pressure in each layer. held fixed in the model.
The analysis of succeeding layers continues in this
bottom-to-top fashion. Each time, the model for the
p1
behavior includes an additional layer. This procedure
1/J1
p2 has been labeled the sequential interpretation method.
Datum pressure (psi)
p3 Once all the layer parameters have been estimated, a

1/J2
simultaneous interpretation can be performed for all
1/J3
layers. In this case, the nonlinear regression uses all
the transients and a model for all the layers. This step
refines the parameter estimates and, because it uses
data covering a longer period of time than any single
transient, may account for external boundary effects
Flow rate (B/D) that are too distant to be apparent in a single transient.
Figure 2-15. Selective inflow performance analysis.

2-7. Testing multilateral and
multibranch wells
2-6.2. Analysis of multilayer transient Multilateral wells and, more generally, multibranch
test data wells have two or more well paths branching from a
Multilayer transient data analysis begins with the common main trunk as in Fig. 2-16. Dual completion
transient flow rate and pressure data acquired above strings may segregate production from the well paths,
the lowest reservoir layer. The RNP is computed, as but this limits the number of branches. Otherwise, the
indicated previously, as the ratio of the pressure flow from the branches is commingled in the main
change to the flow rate change for data acquired at trunk. If the branch departures from the main trunk
the same instant in elapsed time since the start of the are separated, then the flow rate can be measured as
transient. The log-log plot of RNP and its derivative
can be used to select a model for the transient behav-
ior of the lowest reservoir layer. Then, analogous Stacked lateral wells
procedures to those used for single-layer drawdown
tests with measured bottomhole pressure and flow
rate are applied to estimate parameters for observed
flow regimes and to refine these estimates using non-
linear regression. Once a suitable match between the
measured data and a model is found, the parameters
used to generate the match provide an interpretation Figure 2-16. Multibranch well with stacked horizontal
for the lowest reservoir layer. branches commingled in an inclined trunk section.

previously discussed, by the difference above and each other, relaxing the elastic force exerted on the
below the branch or inside a sliding sleeve in the main formation. The decrease of the induced stress results
trunk, where applicable. Alternatively, a test sequence in decline of the wellbore pressure. Because the
analogous to the multilayer transient test with data whole process is controlled by fluid leakoff, pressure
acquired in the main trunk above each branch (Fig. decline analysis has been a primary source of obtain-
2-17) enables the analysis of each branch, with SIP ing the parameters of the assumed leakoff model.
analysis providing the branch productivities and tran- The polymer content of the fracturing fluid is partly
sient analysis providing a set of model parameters for intended to impede fluid loss into the reservoir. The
each branch. If the branch departures are not sepa- phenomenon is envisioned as a continuous buildup
rated, the flow rate measurement must be acquired in of a thin layer (the filter cake) that manifests the ever
the branch. Permanent pressure and flow rate sensors increasing resistance to flow through the fracture face.
installed in the branch could also provide a means to In reality, leakoff is determined by a coupled system,
test the branch. of which the filter cake is only one element. In the fol-
Model selection for a branch depends on the trajec- lowing, pressure decline analysis is introduced as it
tory geometry of the branch, which can be vertical, pertains to the modeling of fluid loss (see Chapters 6
slanted or horizontal. Karakas et al. (1991) published and 9) along with another method coupling filter-cake
an interpretation for a series of tests in a bilateral well. resistance and transient reservoir flow.
Pressure gauge
2-8.1. Pressure decline analysis with the
Carter leakoff model
A fruitful formalism dating back to Howard and Fast
(1957) is to consider the combined effect of the dif-
ferent fluid-loss mechanisms as a material property.
According to this concept, the leakoff velocity uL is
given by the Carter equation:
CL
uL = , (2-30)
t
Paired flowmeters
above and below branch
where CL is the leakoff coefficient in ft/min1/2 and
Figure 2-17. Transient data acquisition in a multibranch t is the time elapsed since the start of the leakoff
well. process. The integrated form of Eq. 2-30 is
VL
= 2 CL t + S p , (2-31)
AL
2-8. Permeability determination from where VL is the fluid volume that passes through the
a fracture injection test† surface AL during the time period from time zero to
time t. The integration constant Sp is called the spurt-
Fracture injection tests, called also calibration treat-
loss coefficient. It can be considered the width
ments, consist of injecting a known amount of the
(extent) of the fluid flowing through the surface
fracturing fluid into the formation, shutting down the
instantaneously at the beginning of the leakoff
pumps and observing the decline of the pressure in
process. The two coefficients, CL and Sp, can be
the wellbore. It is assumed that up to the end of
determined from laboratory tests.
injection time te, the injection rate i into one wing is
Application of Eqs. 2-30 and 2-31 during fracturing
constant. After injection, the pressure in the wellbore
can be envisioned assuming that the given surface ele-
declines because the fluid is leaking off from the cre-
ment “remembers” when it has been opened to fluid
ated fracture and the fracture faces are approaching
loss and has its own zero time, which may be different
from location to location on a fracture surface.
†
This section by Professor Peter Valkó, Texas A&M University.

α
A hydraulic fracture injection may last from tens A t
of minutes up to several hours. Points at the fracture =  , (2-33)
Ae  te 
face near the well are opened at the beginning of
pumping, whereas points at the fracture tip are with the exponent α constant during the injection
younger. Application of Eq. 2-31 necessitates track- period.
ing the opening time of the different fracture face Considering the opening-time distribution factor,
elements. If only the overall material balance is con- Nolte realized that it is a function of the exponent α,
sidered, then Carter’s concept is used: only
κ = g0 (α ) .
(
Vi = V + 2 Arp κCL t + Sp , ) (2-32) (2-34)
The function g0(α) can be determined by an exact

where Vi = it is the total injected volume for one wing mathematical method and is given by Meyer and
with a fracture volume V, A is the surface area of one Hagel (1989):
face of one wing, and rp is the ratio of permeable area
to total fracture area (see Figs. 2-18 and 2-19 for detail). αΓ (α ) π
g0 (α ) = , (2-35)
The variable κ is the opening-time distribution factor.
Γ  + α
3
Clearly, the maximum possible value of κ is 2. The 2 
maximum is reached if all the surface opens at the first
where Γ(α) is the Euler gamma function. A remark-
moment of pumping.
able fact concerning the g0(α) function is that its val-
Nolte (1979, 1986a) postulated a basic assumption
ues for two extremely departing values of the expo-
leading to a remarkably simple form of material
nent α, namely at one-half and unity, differ only
balance. Assuming that the fracture surface evolves
slightly: g0(1⁄2) = π/2 ≅ 1.57 and g0(1) = 4⁄3 ≅ 1.33.
according to a power law, then
2i
qL/2
hp
rp =
hf
hp
A = hfxf
A = hfxf
hf
xf
qL/2
Figure 2-18. Basic notation for PKN and KGD geometries. hp = permeable height, hf = fracture height and qL = rate of fluid
loss.

2i
qL/2
1/2
2 hp hp hp hp
rp = 1– + arcsin
π 2Rf 2Rf 2Rf 2Rf
hp
i
Rf2π
A=
2
Rf
qL/2
Figure 2-19. Basic notation for radial geometry.
If the fracture area is assumed to remain constant

− 2 rp Sp − 2 rp CL te g(α, ∆t D ) .
Vi
after the pumps are stopped, at the time te + ∆t the wt + ∆t = (2-39)
e
Ae
volume of the fracture is
Hence, time variation of the width is determined
Vt + ∆t = Vi − 2 Ae rp Sp − 2 Ae rp g(α, ∆t D )CL te ,
e
(2-36) by the g(α, ∆tD) function, length of the injection
period and leakoff coefficient but is not affected
where dimensionless time is defined as
by the fracture area.
∆t The fracture closure process (i.e., decrease of aver-
∆t D = (2-37)
te age width) cannot be observed directly. However,
from linear elastic theory the net pressure is known
and the two-variable function g0(α, ∆tD) is the fol- to be directly proportional to the average width as
lowing mathematical expression (Valkó and
Economides, 1995): pnet = s f w , (2-40)
4α ∆t D + 2 1 + ∆t D × F  , α; 1 + α; 1 + ∆t D ( ) 
−1
1 where pnet = p – pc and pc is the closure pressure. The
(
g α, ∆t D ) =
2

1 + 2α
 . significance of the closure pressure is described in
Chapters 5, 6 and 9. The fracture stiffness sf is a pro-
(2-38) portionality constant for the fracture geometry mea-
The function F[a, b; c; z] is the hypergeometric sured in psi/ft, and it plays a similar role as the con-
function, available in tabular form (Abramowitz and stant in Hooke’s law. Its form depends on the frac-
Stegun, 1989) or computing algorithms. ture geometry, which may be PKN, KGD or radial
The average fracture width at time ∆t after the end (fracture geometries are described in Chapter 6). In
of pumping is petroleum engineering literature, its inverse 1⁄sf is

also called the fracture compliance. Expressions of

CL =
(− m ) N
. (2-42)
sf for common fracture geometries are in Table 2-2.
2 rp te s f
Table 2-2. Proportionality constant sf for different
fracture geometries. The intercept bN of the straight line at zero shut-in
time provides an expression for the spurt-loss coeffi-
PKN KGD Radial
cient:
2E ′ E′ 3πE ′ Vi b − pc
πhf πx f 16Rf
rp Sp = − N . (2-43)
2 Ae 2s f
The first term in Eq. 2-43 can be interpreted as the
The combination of Eqs. 2-39 and 2-40 yields gross width that would have been created without
any fluid loss minus the apparent leakoff width wL.
 
( )
sV Depending on the fracture geometry, expressions for
p =  pc + f i − 2 rp s f Sp  − 2 rp s f CL te × g(α, ∆t D ) sf can be substituted into Eqs. 2-42 and 2-43, result-
 Ae 
ing in the expressions for CL in Table 2-3 and for sf
= bN + mN × g(α, ∆t D ) , (2-41) and wL in Table 2-4. Tables 2-3 and 2-4 show that
calculation of the leakoff coefficient depends on the
so that a plot of p versus g(α, ∆tD) has a slope mN fracture geometry.
and intercept bN at g = 0.
Equation 2-41 suggests that the pressure variation Table 2-3. Leakoff coefficient CL for different
in the shut-in period is governed mainly by the leak- fracture geometries.
off coefficient, and a plot of wellbore pressure versus PKN KGD Radial
g(α, ∆tD) values results in a straight line provided
πx f
(−m )
πhf 8Rf
that the fracture area Ae, proportionality constant sf
(−m ) (−m )
N
3πrp t e E ′
N
and leakoff coefficient CL do not vary with time. Under 4rp t e E ′

N
2rp t e E ′
these assumptions the pressure behavior will depart
from the linear trend only when the fracture finally
closes. • PKN fracture geometry
The expression in Eq. 2-41 is the basis of Nolte’s For PKN geometry, the leakoff coefficient can be
pressure decline analysis. The technique requires determined from Eq. 2-42 because sf is dependent
plotting the wellbore pressure versus the values of on the fracture height, which is a known quantity.
the g-function, as first suggested by Castillo (1987). Similarly, the spurt-loss coefficient and the appar-
The g-function values should be generated with the ent leakoff width can be computed from Eq. 2-43
exponent α considered valid for the given model and using the expressions in Table 2-4.
rheology. Other choices for α (e.g., involving the • KGD and radial fracture geometries
estimated efficiency of the fracture) are discussed
in Chapter 9. A straight line is fitted to the observed For KGD geometry, sf is dependent on the fracture
points. half-length xf; for radial geometry, it is dependent
For a plot of pressure falloff data from a fracture on the fracture radius Rf. If the spurt loss is negli-
injection test versus the g-function, Eq. 2-41 implies gible, then xf or Rf can be determined from the
that the closure pressure pc must lie on the line fitted expressions in Table 2-5, and, in turn, the leakoff
through the data. Hence, independent knowledge of coefficient can be computed from the appropriate
the closure pressure, which can be determined from expression in Table 2-3. This analysis procedure
the step rate test described in Chapter 9, helps to was introduced by Shlyapobersky et al. (1988a).
identify which part of the falloff data to use for the If the spurt loss cannot be neglected, the more
straight-line fit. detailed analysis procedures in Chapter 9 must
The slope of the straight line is denoted by mN and be used.
the intercept by bN. From Eq. 2-41, the slope is re- If the closure pressure is not determined inde-
lated to the unknown leakoff coefficient by pendently, straightforward analysis with the

Table 2-4. Spurt-loss coefficient and apparent leakoff width for different fracture geometries.
PKN KGD Radial
Vi πhf (bN − p c ) Vi πx f (bN − p c ) Vi 8Rf (bN − p c )

Sp − − −
2rp x f hf 4rpE ′ 2rp x f hf 2rpE ′ πRf2 3πrpE ′
πhf (bN − p c ) πx f (bN − p c ) 8Rf (bN − p c )

wL 4rpE ′ 2rpE ′ 3πrpE ′
Table 2-5. Fracture exent from the reservoir is shown with its components. Thus, the
no-spurt-loss assumption. total pressure drop is
PKN KGD Radial ∆p(t ) = ∆p face (t ) + ∆p piz (t ) + ∆pr (t ) , (2-44)
2E ′Vi xf =
E ′Vi
Rf = 3
3E ′Vi where ∆pface is the pressure drop across the fracture
xf =
πh (b N − p c )
2
πhf (b N − p c ) 8(b N − p c ) face dominated by the filter cake, ∆ppiz is the pres-
f
sure drop across a polymer-invaded zone, and ∆pr is

the pressure drop in the reservoir.
g-function plot relies on correct identification In a series of experimental works using typical
of the portion of the data through which the line hydraulic fracturing fluids (e.g., borate- and zirconate-
should be fitted. As with Horner analysis of pres- crosslinked fluids) and cores with permeability less
sure buildup data, the success of this plot is under- than 5 md, no appreciable polymer-invaded zone
mined if—for whatever reason—identification is was detected. At least for crosslinked fluids, the sec-
not straightforward. More details on this analysis ond term on the right side of Eq. 2-44 can be
are provided in Chapter 9. ignored:
∆p(t ) = ∆p face (t ) + ∆pr (t ) . (2-45)
2-8.2. Filter-cake plus reservoir pressure
drop leakoff model (according to Using the Kelvin-Voigt viscoelastic model for
Mayerhofer et al., 1993) description of the flow through a continuously
depositing fracture filter cake, Mayerhofer et al.
Carter’s bulk leakoff model is not the only possible
(1993) gave the filter-cake pressure term as
interpretation of the leakoff process. Other models
have been suggested, but one reason why such models
have not been used widely is that it is difficult to p face
design a calibration test interpretation procedure that
is standardized (i.e., the results of which do not
Pressure
depend too much on subjective factors of the inter-

preter). The Mayerhofer et al. (1993) method over-
comes this difficulty. It describes the leakoff rate using pr
two parameters that are physically more discernible

than the leakoff coefficient to the petroleum engineer: Distance from fracture center
the reference resistance R0 of the filter cake at a refer-
ence time t0 and the reservoir permeability kr. To
obtain these parameters from an injection test, the
reservoir pressure, reservoir fluid viscosity, formation Filter Reservoir
cake Polymer-invaded
porosity and total compressibility must be known. zone
Figure 2-20 is a schematic of the Mayerhofer et al.
model in which the total pressure difference between Figure 2-20. Filter-cake plus reservoir pressure drop in the
the inside of a created fracture and a far point in the Mayerhofer et al. (1993) model.

R0 t q L Equation 2-49 can be used both in hydraulic frac-

∆p face = , (2-46) ture propagation and during fracture closure. It allows
rp A t0 2
determination of the leakoff rate at the time instant tn
where R0 is the characteristic resistance of the filter if the total pressure difference between the fracture
cake, which is reached during reference time t0. The and the reservoir is known, as well as the history of
flow rate qL is the leakoff rate from one wing of the the leakoff process. The dimensionless pressure solu-
fracture. In Eq. 2-46, it is divided by 2, because only tion pD[(tj – tj – 1)D] has to be determined with respect
one-half of it flows through area A. to a dimensionless time that takes into account the
The pressure drop in the reservoir can be tracked actual fracture length at tn (not at tj).
readily by employing a pressure transient model for The injection test interpretation processes data
injection into a porous medium from an infinite-con- given as (tn, pn) pairs with n > ne, where ne is the
ductivity fracture. For this purpose, known solutions index of the first time point after shut-in. For the
such as the one by Gringarten et al. (1974) can be consideration of dimensionless pressure, the early-
used. The only additional problem is that the surface time approximation can be used for an infinite-
area increases during fracture propagation. conductivity fracture:
Therefore, every time instant has a different fracture
length, which, in turn, affects the computation of
dimensionless time.
[(
PD tn − t j −1 ) D ]= πk
ct µ r φx f2n
tn − t j . ( ) (2-50)
The transient pressure drop in the reservoir is

The leakoff rates are strongly connected to the
µr
∑ (2 q ) [( ) ],
n
observed pressure changes according to
∆pr (tn ) = − 2 q j −1 pD tn − t j −1
2 πkr h f
( )
j D
j =1
Ae p j −1 − p j
(2-47) qj = (2-51)
s f ∆t j
where hf is the ratio of the leakoff area to the charac-
teristic length (given as rphf for PKN and KGD for j ≥ ne + 2.
geometries and as rpπRf /2 for radial geometry), µr is Combining Eqs. 2-50 and 2-51 with Eq. 2-48
the reservoir fluid viscosity, and pD is the dimension- obtains for n > ne + 2:
less pressure function describing the behavior of the pn −1 − pn
1/ 2
R0 t n
p n − pr =
reservoir (unit response). s f rp ∆t n
1/ 2
2te
The factor 2 must be used in Eq. 2-47 in front  µ   p −p p −p  
1/ 2
 (t )
n
  ∑
1

1/ 2
+ − − t j −1
j −1 j −2 j −1

r j
of qj because q is defined as the leakoff rate from k s r  πc φ    ∆t ∆t  

1/ 2 n
r f p t j = ne + 3 j j −1
one wing. In petroleum engineering literature, how- 1  µ   
1/ 2
ne + 2
  q t + ∑ ( q − q )(t − t )
1/ 2
+
1/ 2
 .
r
ever, dimensionless pressure is defined using the
k A r  πc φ  
1/ 2 1 n j j −1 n j −1
j =2
total flow into (from) the formation. r e p t
Substituting Eqs. 2-46 and 2-47 into Eq. 2-45 obtains (2-52)
µr
) [( ) ],
During fracture propagation, the leakoff rates qj
∑ (q
n
∆p(tn ) =
R0 tn
qn + − q j −1 pD tn − t j −1 for j = 1, . . ., ne + 1 are not known exactly (nor are
πkr h f
j D
2 rp An te j =1
any values for tj in this period). Therefore, some kind
(2-48) of assumption is required to proceed.
where the end of pumping is selected as the charac- The key assumption is that for these purposes the
teristic time for the filter-cake resistance. A simple first ne + 1 leakoff rates can be considered equal:
rearrangement yields
q j = qapp (2-53)
µ − q ) p [(t ) ]
n −1
∆p(t )− r

p (t − t ) + ∑ (q −q −t

n −1 Dn − 1 j −1 j −1
πk h
n D Dn j D n D
j =1 for j = 1, . . ., ne + 1.
q = .
r f
n
R 0
t n
+
µ p r D
[(t n
−t n −1
)]
D
In fact, it is more convenient to work not with the
2r A p n
t e
πk h r f
average leakoff rate but with the apparent leakoff
width defined by
(2-49)

wL =
(1 − η)Vi =
qapp
te , (2-54)
The Mayerhofer et al. method is based on the fact
Ae Ae that Eq. 2-57 can be written in straight-line form as
where η is the fluid efficiency. yn = bM + mM (c1 x1,n + c2 x2 ,n ) (2-58)

The apparent leakoff width can be estimated from
the Nolte-Shlyapobersky method as shown in Table for n > ne + 2, where
2-4. Then Eq. 2-52 leads to
 dn +2  tn − tn +2  1/ 2 n
d j − d j −1  tn − t j −1  
1/ 2
x1,n =    + ∑   
e e
pn −1 − pn
1/ 2
 dn  te tn  dn  te tn  
R0 t n
p n − pr = j = n +3 e
s f rp ∆t n
1/ 2
2te
p − p 
(t − t ) (2-59)
ne +1 ne + 2 1/ 2
 ne +1

1  µ   t −t
n
1/ 2
n ne +1
 1/ 2
+  
r
 tn +1 
k s r  πc φ   p −p p −p   1 − 1 −
1/ 2

e
 (t − t )
n
+ ∑    tn 
r f p t 1/ 2
−
j −1 j j −2 j −1
  ∆t  
j −1
∆t =
n
j = ne + 3 j j −1 x 2 ,n . (2-60)
te3/ 2 dn
 
[q ].
1/ 2
µr
1
( )
1/ 2
+   − q app t n − t n +1
1/ 2
t
1/ 2
kr Ae rp  πc φ  app n e
The coefficients c1 and c2 are geometry dependent

t
(2-55)
and discussed later. Once the x and y coordinates are
In terms of the apparent leakoff width and after known, the (x,y) pairs can be plotted. The corre-
rearrangement, Eq. 2-55 becomes sponding plot is referred to as the Mayerhofer plot.
( pn − pr )∆tn =
R
0 +
1  µ 
r
1/ 2
∆t
n
A straight line determined from the Mayerhofer plot
results in the estimate of the two parameters bM and
 
( n − 1 − pn )t1e / 2t1n / 2
p 2s r t
f pe k
1
r
/ 2
s r
f p
 πc φ 
t p
n −1
− p
n e
t
1/ 2 1/ 2
t
n ( ) mM. Those parameters are then interpreted in terms
p −
 n +1 n + 2 p 1/ 2 
 e e    of the reservoir permeability and the reference filter-
  tn − tn + 1 
∆t
n +2
 e  cake resistance. For the specific geometries, the coef-
 e 
×  
p  ficients c1 and c2 as well as the interpretation of the
( )
 n j −1
−p
j
p
j−2
−p
j −1 1/ 2
+ ∑  −  t −t  straight-line parameters are as follows.
 j = n + 3  ∆t ∆t
j −1


n j −1 
 e j 
1/ 2  
1/ 2 • PKN geometry
1/ 2 t − t − t 
 µ   n ne + 1
 R   πh f 
w n
[ ]
L r
+
1 / 2  πc φ 
∆t . hf
k r  t  p −p t ( n −1 n) e
3/ 2 1/ 2
t
n
yn =  0    + 1/ 2 c1 x1,n + c2 x2 ,n
 rp   4 E ′te  kr rp
r p n
(2-56)
1/ 2
 πµ r 
1/ 2
Introducing the notation w  µ 
c1 =   c2 = L  r 
p − pj  4( E ′ ) φct 
2
h f  πφct 
d j = j −1
∆t j 2
pn − pr  h  4 E ′te
yn = 1/ 2 1/ 2 kr , M = f  R0 ,M = rp bM
dn te t n  rp mM  πh f
Eq. 2-56 takes the form

• KGD geometry
  
1/ 2 
d  t −t  
 R   πx f 
[ ]
 ne + 2  n ne + 1  xf
 
1/ 2  yn =  0    + 1/ 2 c1 x1,n + c2 x2 ,n
 rp   2 E ′te  kr rp
µ 1/ 2 1/ 2
R
0 1 r  d t t 
y = +   ne n
πc φ   
( )( )
n 1/ 2
2s r t
f pe k s r  t  1 / 2 
r f p 1/ 2
− −
 πµ r 
1/ 2
  w  µ 
n d d t t
j j −1 n j −1
 + ∑  c1 =   c2 = L  r 
 ( E ′ ) φct 
1/ 2 1/ 2
 = +  x f  πφct 
j n 3 d t t 2
e ne n
1/ 2
 tn + 1  2
1 − 1 − e



 x  2 E ′te
= f  R0 ,M =
1/ 2
w  µ  
t
n  kr , M rp bM
+
k
L
1/ 2 
r

r  πct φ  t
3/2
d
. (2-57)  rp mM  πx f
r p e n

• Radial geometry Table 2-7. Example reservoir and well information.

 R   8Rf 
[ ]
Rf
yn =  0    + 1/ 2 c1 x1,n + c2 x2 ,n
Permeable height, hp 42 ft
 rp   3πE ′te  kr rp Reservoir fluid viscosity, µr 1 cp
1/ 2 Porosity, φ 0.23
 
1/ 2
16 2 µ r w  µ 
c1 =  2 3  c2 = L  r  Total compressibility, ct 2 × 10–5 psi–1
 3 π ( E ′ ) φct  R f  πφct 
2
Reservoir pressure, pr 1790 psi

2
 R  3πE ′te Plane strain modulus, E´ 8 × 105 psi
kr , M =  f  R0 ,M = rp bM
 rp mM  8Rf Pumping time, te 21.75 min
Injected volume (two wing), 2Vi 9009 gal
• Example interpretation of fracture injection test Closure pressure, pc 5850 psi
Table 2-6 presents pressure decline data, and Table Geometry Radial
2-7 presents reservoir and well information for this
example. The closure pressure pc determined inde- Figure 2-21 is a plot of the data in Cartesian
pendently is 5850 psi. coordinates and also shows the closure pressure.
The g-function plot in Fig. 2-22 is created using
Table 2-6. Example pressure decline data.
α = 8⁄9, which is considered characteristic for the
radial model.
∆t (min) pws (psi) From the intercept of the straight line is
0.2 7550.62 obtained the radial fracture radius Rf = 27.5 ft.
0.4 7330.59 (The straight-line fit also provides the bulk leakoff
0.6 7122.36 coefficient CL = 0.033 ft/min1/2 and fluid efficiency
0.8 6963.21 η = 17.9%.) The ratio of permeable to total area is
1.0 6833.39 rp = 0.76.
1.2 6711.23 Figure 2-23 is the Mayerhofer plot. From the
1.4 6595.02 slope of the straight line (mM = 9.30 × 107) is
1.6 6493.47 obtained the apparent reservoir permeability
1.8 6411.85
kr,app = 8.2 md and the true reservoir permeability
2.0 6347.12
kr = 14.2 md. The resistance of the filter cake at
2.2 6291.51
2.4 6238.43
the end of pumping (te = 21.2 min) is calculated
2.6 6185.85
from the intercept (bM = 2.5 × 10–2) as the apparent
2.8 6135.61 resistance R0,app = 1.8 × 104 psi/(ft/min) and the
3.0 6090.61 true resistance R0 = 1.4 × 104 psi/(ft/min).
3.2 6052.06
3.4 6018.61
8000
3.6 5987.45
3.8 5956.42 7500
4.0 5925.45
4.2 5896.77 7000
pws (psi)
4.4 5873.54
4.6 5857.85 6500
4.8 5849.29
6000
5.0 5844.81
5.2 5839.97 5500
5.4 5830.98 0 1 2 3 4 5
5.6 5816.30 ∆t (min)
5.8 5797.01
6.0 5775.67 Figure 2-21. Example of bottomhole pressure versus shut-
in time.

8000 1.2
7500 1.0
7000 0.8
p (psi)
0.6
y
6500
0.4
6000
0.2
5500
1.35 1.40 1.45 1.50 1.55 1.60 1.65 1.70 0
g 0 2 × 10–9 6 × 10–9 1 × 10–8
x
Figure 2-22. Example g-function plot.
Figure 2-23. Example Mayerhofer plot with radial geometry.


◆
Rock Mechanics
M. C. Thiercelin, Schlumberger Dowell
J.-C. Roegiers, University of Oklahoma
3-1. Introduction As deeper completions were attempted, borehole

collapses and instabilities became more common and
The National Academy of Sciences defines rock often led to expensive remedial measures. The pri-
mechanics as “the theoretical and applied science mary cause of these problems is instabilities caused
of the mechanical behavior of rock; it is that branch by large tectonic forces. The concepts developed by
of mechanics concerned with the response of the rock mining engineers that rocks are far from being inert
to the force fields of its physical environment.” From were found applicable (Cook, 1967; Hodgson and
this definition, the importance of rock mechanics in Joughin, 1967). Rocks are quite receptive to distur-
several aspects of the oil and gas industry can easily bances, provided that some energy limits are not
be understood. The fragmentation of rock governs its exceeded. If critical energies are exceeded, dynamic
drillability, whereas its mechanical behavior influences failure is likely to occur, such as rock burst or casing
all aspects of completion, stimulation and production. collapse. In addition, the importance of inherent
However, not until recently has this particular aspect discontinuities (e.g., faults, fissures) was realized
of earth sciences started to play a predominant role in (Goodman, 1976), especially as highly conductive
energy extraction. The impetus was to explain, quali- conduits. From this broader understanding of the role
tatively and quantitatively, the orientation of fractures of rock deformation, research focused on definition
(Hubbert and Willis, 1957), some unexpected reser- of the pertinent parameters required to properly char-
voir responses or catastrophic failures (e.g., less pro- acterize the targeted formations. Cores were taken not
duction after stimulation and pressure decline in wells only for the determination of permeability, porosity
surrounding an injection well; Murphy, 1982), casing and lithology, but also to run mechanical tests under
shear failure (Nester et al., 1956; Cheatham and simulated downhole conditions. Downhole tools were
McEver, 1964), sand production (Bratli and Risnes, developed to better characterize the formation in situ.
1981; Perkins and Weingarten, 1988; Morita et al., There will always remain some uncertainties on the
1987; Veeken et al., 1991; Kooijman et al., 1992; relevance of laboratory-determined parameters to the
Cook et al., 1994; Moricca et al., 1994; Geilikman field situation, either because of the disturbance a core
et al., 1994; Ramos et al., 1994), rock matrix collapse sample suffers during the coring and handling process
during production (Risnes et al., 1982; Pattillo and or because of scale effects. There are also limitations
Smith, 1985; Smits et al., 1988; Abdulraheem et al., on the use of simple constitutive laws to predict rock
1992) and borehole stability problems (Gnirk, 1972; behavior for heterogeneous, discontinuous, time-
Bradley, 1979; Guenot, 1989; Santarelli et al., 1992; dependent and/or weak formations. Studies are cur-
Ong and Roegiers, 1993; Maury, 1994; Last et al., 1995). rently being conducted on these issues, and our under-
The significant contribution as far as the orientation standing and predictive capability of downhole rock
of fractures is concerned was provided by the work behavior can be expected to continue to progress.
of Hubbert and Willis (1957; see Sidebar 3A), which This chapter briefly summarizes some of the most
indicates ever-increasing differences between vertical important aspects of rock mechanics to characterize
and horizontal stresses within the earth’s crust. Until the mechanical behavior of reservoirs and adjacent
then, all design considerations were based on the layers, as applied to the stimulation process.
assumption that an isostatic state of stress prevailed
everywhere.

3A. Mechanics of hydraulic fracturing
Hubbert and Willis (1957) introduced several key concepts

that explain the state of stress underground and its influence
on the orientation of hydraulic fractures. Reviewed here are
σ1 σ3
the fundamental experiments that Hubbert and Willis per-
formed to validate these concepts. σ3 σ1
Sand
State of stress underground
The general state of stress underground is that in which the
three principal stresses are unequal. For tectonically relaxed
areas characterized by normal faulting, the minimum stress
should be horizontal; the hydraulic fractures produced should Figure 3A-3. Approximate stress conditions in the sand-
be vertical with the injection pressure less than that of the box experiment.
overburden. In areas of active tectonic compression and
thrust faulting, the minimum stress should be vertical and tion is adopted of designating the maximum, intermediate and
equal to the pressure of the overburden (Fig. 3-20). The minimum principal effective stresses by σ1´, σ2´ and σ3´,
hydraulic fractures should be horizontal with injection pres- respectively (here taken as compressive). In the left-hand
sures equal to or greater than the pressure of the overburden. compartment, σ3´ is the horizontal effective stress, which is
To demonstrate these faulting conditions, Hubbert and reduced as the partition is moved to the right, and σ1´ is the
Willis performed a sandbox experiment that reproduces both vertical effective stress, which is equal to the pressure of the
the normal fault regime and the thrust fault regime. Figures overlying material minus the pore pressure. In the right-hand
3A-1 and 3A-2 show the box with its glass front and contain- compartment, however, σ1´ is horizontal, increasing as the
ing ordinary sand. The partition in the middle can be moved partition is moved, and σ3´ is vertical and equal to the pres-
from left to right by turning a hand screw. The white lines are sure of the overlying material minus the pore pressure. The
plaster of paris markers that have no mechanical significance. third type of failure, strike-slip faulting, is not demonstrated
As the partition is moved to the right, a normal fault with a dip in the sandbox experiment.
of about 60° develops in the left-hand compartment, as Next, the combination of shear and normal stresses that
shown in Fig. 3A-1. With further movement, a series of thrust induce failure must be determined. These critical effective
faults with dips of about 30° develops in the right-hand com- stress values can be plotted on a Mohr diagram, as shown
partment, as shown in Fig. 3A-2. in Fig. 3A-4. The two diagonal lines form the Mohr envelopes
The general nature of the stresses that accompany the of the material, and the area between them represents stable
failure of the sand is shown in Fig. 3A-3. The usual conven- combinations of shear stress and normal effective stress,
whereas the area exterior to the envelopes represents unsta-
ble conditions. Figure 3A-4 thus indicates the stability region
within which the permissible values of σn´ and τ are clearly
defined. The stress circles can then be plotted in conjunction
with the Mohr envelopes to determine the conditions of fault-
ing. This is illustrated in Fig. 3A-4 for both normal and thrust
faulting. In both cases, one of the principal effective stresses
Points of fracture
φ
φ 2a
2a
φ
Figure 3A-1. Sandbox experiment showing a normal fault. 0 φ σ
σ3´
σ v́
σ1´
Mohr
envelope
Figure 3A-4. Mohr diagram of the possible range of hori-

zontal stress for a given vertical stress σv´. The horizontal
stress can have any value ranging from approximately
one-third of the normal stress, corresponding to normal
faulting, to approximately 3 times the vertical stress, cor-
Figure 3A-2. Sandbox experiment showing a thrust fault. responding to reverse faulting.
3-2 Formation Characterization: Rock Mechanics

3A. Mechanics of hydraulic fracturing (continued)

is equal to the overburden effective stress σv´. In the case of normal
faulting, the horizontal principal stress is progressively reduced, there-
by increasing the radius of the stress circle until the circle touches the
Mohr envelopes. At this point, unstable conditions of shear and nor-
mal effective stress are reached and faulting occurs on a plane mak-
ing an angle of 45° + φ/2 with the minimum stress. For sand with an
angle of internal friction of 30°, the normal fault would have a dip of
60°, which agrees with the previous experiments. The minimum princi-
pal effective stress would reach a value at about one-third of the value
of the overburden effective stress (Eq. 3-58). For the case of thrust
faulting, the minimum principal stress would be vertical and remain
equal to the overburden pressure while the horizontal stress is pro-
gressively increased until unstable conditions occur and faulting takes
place on a plane making an angle of 45° + φ/2 with the minimum prin-
cipal stress or 45° – φ/2 with the horizontal. For sand, this would be a
dip of about 30°, which again agrees with the experiment. Failure
occurs when the maximum horizontal effective principal stress Favored
reaches a value that is about 3 times the value of the overburden fracture
effective stress (Eq. 3-59). The intermediate stress, which is the mini- direction
mum horizontal stress, is not defined by this process.
From these limiting cases and for a fixed effective vertical stress
σv ′, the effective horizontal stress may have any value between the
extreme limits of 1⁄3 and 3 times σv ′.
Least
principal
Orientation of hydraulic fractures stress
The second important contribution of Hubbert and Willis’ work con-
cerns the orientation of hydraulic fractures. When their paper was pre-
sented, technical debate was occurring on the orientation of hydraulic
fractures. A theoretical examination of the mechanisms of hydraulic
fracturing of rocks led them to the conclusion that, regardless of Figure 3A-5. Experimental arrangement for pro-
whether the fracturing fluid was penetrating, the fractures produced ducing the least stress in a horizontal direction.
should be approximately perpendicular to the axis of minimum princi-
pal stress.
To verify the inferences obtained theoretically, a series of simple
laboratory experiments was performed. The general procedure was to
produce fractures on a small scale by injecting a “fracturing fluid” into
a weak elastic solid that had previously been stressed. Ordinary gel-
atin (12% solution) was used for the solid, as it is sufficiently weak to
fracture easily, molds readily in a simulated wellbore and is almost
perfectly elastic under a short-time application of stresses. A plaster of
paris slurry was used as the fracturing fluid because it could be made
thin enough to flow easily and once set provided a permanent record
of the fractures produced. The experimental arrangement consisted of
a 2-gal polyethylene bottle, with its top cut off, used to contain a glass
tubing assembly consisting of an inner mold and concentric outer cas-
ings. The container was sufficiently flexible to transmit externally
applied stresses to the gelatin. The procedure was to place the glass
tubing assembly in the liquid gelatin and after solidification to withdraw
the inner mold leaving a “wellbore” cased above and below an open-
hole section.
Stresses were then applied to the gelatin in two ways. The first
way (Fig. 3A-5) was to squeeze the polyethylene container laterally,
thereby forcing it into an elliptical cross section and producing a com-
pression in one horizontal direction and an extension at right angles in
the other. The minimum principal stress was therefore horizontal, and
vertical fractures should be expected, as observed in Fig. 3A-6. In
other experiments, the container was wrapped with rubber tubing
stretched in tension, thus producing radial compression and vertical
extension. In this case, the minimum principal stress was vertical, and
a horizontal fracture was obtained.
From these analyses and experiments, Hubbert and Willis con-
cluded that
• the state of stress, and hence the fracture orientations, is gov-
erned by “incipient failure” (i.e., faulting) of the rock mass
• in areas subject to active normal faulting, fractures should be Figure 3A-6. Vertical fracture produced under
approximately vertical stress conditions illustrated in Fig. 3A-5.
• in areas subject to active thrust faulting, fractures should be
approximately horizontal.

3-2. Basic concepts y

3-2.1. Stresses
In considering a randomly oriented plane of area ∆A
B
centered on a point P within a body across which a
resultant force ∆F acts (Fig. 3-1), the stress vector σ
at that point is defined as σn
θ
∆F
σ = lim   . (3-1) σx
∆A→ 0  ∆A 
τ
Therefore, this quantity is expressed as a force per τxy
unit area. In geomechanics, by convention, compres-
sion is taken to be positive because the forces prevail-
ing in the earth are usually compressive in nature.
This resultant stress σ can be decomposed into a nor- O τyx A x
mal component σn and a shear component τ. The
shear component tends to “shear” the material in the
plane ∆A. It should be realized that an infinite amount σy
of planes can be drawn through a given point varying,
by the same token, the values of σn and τ. The stress Figure 3-2. Two-dimensional decomposition of normal and
condition, therefore, depends on the inclination. Con- shear stresses.
sequently, a complete description of a stress must
specify not only its magnitude, direction and sense, These expressions are obtained by writing equilib-
but also the direction of the surface upon which it rium equations of the forces along the σn and τ direc-
acts. Quantities described by two directions, such tions, respectively. The moment equilibrium implies
as stresses, are known as second-order tensors. that τxy is equal to τyx. There always exist two perpen-
In a two-dimensional (2D) situation, if σx, σy and dicular orientations of ∆A for which the shear stress
τxy are known (Fig. 3-2), the stress state on any plane components vanish; these are referred to as the princi-
with normal orientation at an angle θ from Ox can be pal planes. The normal stresses associated with these
derived using the following expressions: planes are referred to as the principal stresses. In two
dimensions, expressions for these principal stresses
σ n = σ x cos 2 θ + 2 τ xysinθcosθ + σ ysin 2 θ (3-2)
can be found by setting τ = 0 in Eq. 3-3 or, because
they are the minimum and maximum values of the
τ=
1
( )
σ y − σ x sin2θ + τ xy cos2θ . (3-3) normal stresses, by taking the derivative of Eq. 3-2
2 with respect to the angle θ and setting it equal to zero.
Either case obtains the following expression for the
∆F
value of θ for which the shear stress vanishes:
1  2 τ xy 
θ= arctan  (3-4)
2  σx − σy 
and the two principal stress components σ1 and σ2 are
1/ 2
1
( ) (
σ x + σ y + τ 2xy + σ x − σ y 
1
)
2
σ1 = (3-5)
2  4 
1/ 2
( ) ( )
P
σ x + σ y − τ 2xy + σ x − σ y  .
1 1 2
∆A σ2 = (3-6)
2  4 
Figure 3-1. Force on a point P.

If this concept is generalized to three dimensions, it

3B. Mohr circle
can be shown that six independent components of the
stress (three normal and three shear components) are Equations 3-2 and 3-3 can be used to derive σn and τ as a
needed to define the stress unambiguously. The stress function of σ1 and σ2 (effective stresses are considered):
vector for any direction of ∆A can generally be found
(σ 1 + σ 2 ) + 2 (σ 1 − σ 2 )cos2θ
1 1
σn = (3B-1)
by writing equilibrium of force equations in various 2
directions. Three principal planes for which the shear
(σ 1 − σ 2 )sin2θ .
1
stress components vanish—and, therefore, the three τ=− (3B-2)
2
principal stresses—exist. The angle θ is the angle at which the normal to the plane
It is convenient to represent the state of stress at a of interest is inclined to σ1. These expressions provide the
equation of a circle in a (σn, τ) plane, with its center located
given point using graphical methods. The most widely on the axis at 1⁄2 (σ1 + σ2) and of diameter (σ1 – σ2) (Fig. 3B-1).
used method is the Mohr representation described in This circle is known as the Mohr circle and contains all the
Sidebar 3B. Other useful quantities are stress invari- information necessary to determine the two-dimensional
stress state at any orientation in the sample. The intersection
ants (i.e., quantities that do not depend on the choice of this circle with the horizontal axis determines the maximum
of axes). For example, the mean stress σm: and minimum values of the normal stresses at a point in the
material. The apex represents the maximum value of the
( )
shear stress. For a three-dimensional state of stress, similar
σ x + σ y + σ z = (σ1 + σ 2 + σ 3 )
1 1
σm = (3-7) circles can be constructed for any two orthogonal directions.
3 3
and the octahedral shear stress τoct:
τ
1
τ oct =
3
[ 2
]
(σ1 − σ 2 ) 2 + (σ1 − σ 3 ) + (σ 2 − σ 3 )
2 1/ 2
(3-8)
are two stress invariants typically used in failure criteria.
3-2.2. Strains τ M
When a body is subjected to a stress field, the relative

position of points within it is altered; the body deforms. 2θ
If these new positions of the points are such that their
initial and final locations cannot be made to corre- σ2 σn σ1 σn
spond by a translation and/or rotation (i.e., by rigid
body motion), the body is strained. Straining along an Figure 3B-1. The coordinates of point M on the Mohr
circle are the values of normal stress and shear stress
arbitrary direction can be decomposed into two com- across a plane with the normal oriented at θ to the
ponents, as shown in Fig. 3-3: direction of maximum principal stress.
• elongation, defined as
l −l* perpendicular initially. Just as in the case of stresses,
ε = lim (3-9) principal strains can be defined as longitudinal strain
l→0 l
components acting on planes where the shear strains
• shear strain, defined as have vanished. It should be pointed out that the analogy
γ = tan( ψ ) , (3-10) between stress and strain analyses is not completely
valid and that equilibrium equations and compatibility
where ψ is the change of angle between two direc- equations have to be satisfied respectively for the
tions that were perpendicular prior to straining. stresses and for the strains. These relations put some
Consequently, strain (which is either a ratio of restrictions on the local variation of stress and strain in
lengths or a change of angle) is dimensionless. the neighborhood of a point. For example, compatibili-
Because stresses are taken as positive in compression, ty equations ensure that the strained body remains con-
a positive longitudinal strain ε corresponds to a decrease tinuous and that no cracks or material overlaps will
in length, and a positive shear strain γ reflects an occur. For further details on stresses and strains, the
increase in the angle between two directions that were reader is referred to the classic works by Love (1927),

P
Q
l
Original
π/2
O O P
P∗ Q∗
ψ
l∗ Deformed
O∗ P∗
O∗
Figure 3-3. Normal and shear strain components.
Timoshenko and Goodier (1970) and Muskhelishvili ticity is particularly useful for predicting the stress
(1953). concentration around a wellbore or the behavior of
soft materials during reservoir depletion.
3-3. Rock behavior

3-3.1. Linear elasticity
When a rock specimen or an element of the earth
is submitted to load, it deforms; the higher the stress To introduce the theory of linear elasticity, let us con-
level, the more strain the rock experiences. It is an sider a cylindrical sample of initial length l and diam-
important aspect of rock mechanics, and solid eter d. The sample shortens along the loading direc-
mechanics in general, to determine the relationship tion when a force F is applied to its ends (Fig. 3-4).
between stress and strain (i.e., the constitutive equa- According to the definitions in the previous section,
tions of the material under consideration). Various the- the axial stress applied to the sample is
ories have been developed to describe, in a simplified 4F
way, this relationship. The simplest one is the theory σ1 = (3-11)
πd 2
of elasticity, which assumes that there is a one-to-one
correspondence between stress and strain (and, conse- and the axial strain is
quently, that the behavior is reversible). Because this l −l*
is usually the assumed case in hydraulic fracturing, ε1 = , (3-12)
l
most of the simulation models use the theory of elas-
ticity. Other theories have been developed to better where l* is the resultant length.
take into account the complex behavior of rock, espe- Linear elasticity assumes a linear and unique rela-
cially in compression. For example, the theory of plas- tionship between stress and strain. The consequence
of uniqueness is that all strain recovers when the

Poisson’s ratio, by definition, is a positive quantity.

F
These stress-strain relations can be generalized to full
three-dimensional (3D) space by
σx ν
εx =
E E
(
− σy + σz )
σ ν
ε y = y − (σ x + σ z )
E E
σ ν
(
εz = z − σx + σy
E E
)
1 1 1
l γ xy = τ xy ; γ yz = τ yz ; γ xz = τ xz , (3-16)
d G G G
l∗
where the shear modulus G is
E
G= . (3-17)
2(1 + ν)
Another coefficient that is commonly used is the

bulk modulus K, which is the coefficient of propor-
d∗ tionality between the mean stress σm and volumetric
strain εV during a hydrostatic test. In such a test, all
Figure 3-4. Sample deformation under uniaxial loading. three normal stresses are equal and, consequently, all
directions are principal. For this case:
∆V E
material is unloaded. In the case of a uniaxial com- σ m = Kε V ; ε V = ;K= , (3-18)
pression test, this means that V 3(1 − 2 ν)
σ1 = Eε1 . (3-13) where V is the rock volume and ∆V is its variation.

In isotropic linear elasticity, only two elastic con-
The coefficient of proportionality E is Young’s mod- stants are independent. For example, and as discussed
ulus. previously, the shear modulus G and the bulk modulus
When a rock specimen is compressed in one direc- K can be written as functions of E and ν. The most
tion, not only does it shorten along the loading direc- commonly used constants in reservoir applications are
tion, but it also expands in the lateral directions. This defined in Sidebar 3C.
effect is quantified by the introduction of an additional Elasticity theory can be extended to nonlinear and
constant (Poisson’s ratio ν), defined as the ratio of lat- anisotropic materials. A nonlinear elastic material
eral expansion to longitudinal contraction: does not have a linear relationship between stress and
ε2 strain, but recovers all strain when unloaded. An
ν=− , (3-14) anisotropic material has properties that differ in differ-
ε1
ent directions. A common type is transverse anisot-
where ropy, which applies to materials that have a plane and
d−d* an axis of symmetry (the axis of symmetry is the nor-
ε2 = , (3-15) mal to the plane of symmetry). This is particularly
d
suited for bedded formations where the bedding plane
where d* is the new diameter. is the plane of symmetry. These materials, which
The negative sign is included in Eq. 3-14 because, exhibit the simplest type of anisotropy, are character-
by convention, expansion is considered negative and ized by five elastic constants.

3C. Elastic constants Force

Pores
Two independent constants characterize isotropic linear elas-
tic materials. Several different conditions can be considered
and specific equations can be derived from the three-dimen- Grains
sional elasticity relations (Eq. 3-16):
Unconfined axial loading

Specified: σx or εx, with σy = σz = 0
E = σx εx (3C-1)
ε y = ε z = − νε x (3C-2)
E = 2(1+ ν)G , (3C-3)

Force
where E is Young’s modulus, ν is Poisson’s ratio, and G is
the shear modulus. Figure 3-5. Load sharing by pore pressure. Total stress =
pore pressure + effective stress carried by the grains.
Hydrostatic (isotropic) loading
Specified: σx = σy = σz and εx = εy = εz
The volumetric strain εV is equal to εx + εy + εz.
• An increase of pore pressure induces rock dilation.
K = σ x εV = E 3(1− 2 ν) , (3C-4) • Compression of the rock produces a pore pressure
increase if the fluid is prevented from escaping
where K is the bulk modulus.
from the porous network.
Plane strain loading (all x-y planes remain parallel)
When the fluid is free to move, pore pressure diffu-
Specified: εz = 0, with the added constraint that σy = 0
sion introduces a time-dependent character to the
E ′ = σx εx (3C-5) mechanical response of a rock: the rock reacts differ-
ently, depending on whether the rate of loading is
E ′ = E (1− ν 2 ) = 2G (1− ν) , (3C-6) slow or fast compared with a characteristic time that
where E´ is the plane strain modulus used in fracture width
governs the transient pore pressure in the reservoir
models. (itself governed by the rock deformation).
Hence, two limiting behaviors must be introduced:
Uniaxial (laterally constrained) strain
drained and undrained responses. One limiting case is
Specified: εy = εz = 0
realized when a load is instantaneously applied to a
C = σ x ε x = E (1− ν) [(1+ ν)(1− 2ν)] = K + 43 G (3C-7) porous rock. In that case the excess fluid pressure has
no time to diffuse and the medium reacts as if it were
[ ]
σ z = σ y = ν / (1− ν) σ x , (3C-8) undrained and behaves in a “stiff” manner. On the
other extreme, if the pressurization rate is sufficiently
where C is called the constrained modulus and is used for
earth stresses and plane compressive seismic waves.
slow and excess pressure areas have ample time to
drain by diffusion, the rock is “softer.” The stiffening
effect is more important if the pores are filled with a
relatively incompressible liquid rather than a relatively
3-3.2. Influence of pore pressure compressible gas.
Pore fluids in the reservoir rock play an important role In 1923, Terzaghi first introduced the effective
because they support a portion of the total applied stress concept for one-dimensional consolidation and
stress. Hence, only a portion of the total stress, namely, proposed the following relationship:
the effective stress component, is carried by the rock σ′ = σ − p ,
matrix (Fig. 3-5). Obviously, this effective stress (3-19)
changes over the life of the reservoir. In addition, the where σ is the total applied stress, σ´ is the effective
mechanical behavior of the porous rock modifies the stress governing consolidation of the material, and p
fluid response. Two basic mechanisms highlight this is the pore pressure. However, Biot (1941, 1956a) pro-
coupled behavior (e.g., Detournay and Cheng, 1993): posed a consistent theory to account for the coupled

diffusion/deformation processes that are observed in which describes the in-situ stress change caused by
elastic materials. Such a strong coupling is due to the injection and/or production. This is addressed later
fact that any change in pore pressure is accompanied in the chapter.
by variation in the pore volume; hence, it affects the
overall mechanical response of the rock. This poro-
elastic material behavior is similar to that of an elastic 3-3.3. Fracture mechanics
solid when the stresses in Eq. 3-16 are replaced by the Fracture mechanics studies the stability of preexisting
following effective stresses: defects that are assumed to pervade a continuum.
σ ′ = σ − αp. (3-20) These inclusions induce high stress concentrations in
their vicinity and become the nucleus for crack initia-
This relation rigorously governs the deformation tion and/or propagation. Historically, Griffith (1921,
of a porous medium, whereas failure is controlled by 1924) established the foundation of fracture mechan-
Terzaghi’s (1923) effective stresses in Eq. 3-19 (Rice, ics; he studied propagation by considering the energy
1977; Rudnicki, 1985). The poroelastic constant α used in various parts of the fracturing process.
varies between 0 and 1 as it describes the efficiency Griffith’s original treatment expressed the condition
of the fluid pressure in counteracting the total applied that the total energy is unchanged by small variations
stress. Its value depends on the pore geometry and the in the crack length. The different approach presented
physical properties of the constituents of the solid sys- here states that the energy that is consumed by the
tem and, hence, on the applied load. It is related to the creation of new surfaces should be balanced by the
undrained Poisson’s ratio νu, drained Poisson’s ratio ν change in the potential energy of the system:
and Skempton (1960) pore pressure coefficient B,
defined as dWelas + dWext + dWs + dWkin = 0, (3-25)
∆p
B= , (3-21) where dWelas represents the change in elastic energy
∆σ stored in the solid, dWext is the change in potential
where ∆p represents the variation in pore pressure energy of exterior forces, dWs is the energy dissipated
resulting from a change in the confining stress ∆σ during the propagation of a crack, and dWkin is the
under undrained conditions. From these variables: change in kinetic energy. Energy dissipated as heat
is neglected. To proceed further, it is assumed that the
3( νu − ν) energy dWs required to create the new elementary
α= . (3-22)
B(1 − 2 ν)(1 + νu ) fracture surfaces 2dA is proportional to the area created:
Only in the ideal case, where no porosity change dWs = 2γ F dA , (3-26)

occurs under equal variation of pore and confining where γF is the fracture surface energy of the solid,
pressure, can the preceding expression be simplified to which is the energy per unit area required to create
new fracture surfaces (similar to the surface tension
K
α = 1− , (3-23) of a fluid). The factor 2 arises from the consideration
Ks that two new surfaces are created during the separa-
where K is the bulk modulus of the material and Ks is tion process. The propagation is unstable if the kinetic
the bulk modulus of the solid constituents. Typically, energy increases; thus, dWkin > 0 gives
for petroleum reservoirs, α is about 0.7, but its value Ge > 2γ F , (3-27)
changes over the life of the reservoir. The poroelastic
constant α is a scalar only for isotropic materials. It is where the strain energy release rate Ge is defined as
a tensor in anisotropic rocks (Thompson and Willis, d (Welas + Wext )
1991). Another important poroelastic parameter is the Ge = − . (3-28)
dA
poroelastic stress coefficient η, defined as
The onset of crack propagation, which is referred to
(1 − 2 ν) as the Griffith criterion, is
η=α , (3-24)
2(1 − ν) Ge = 2γ F . (3-29)

Another approach was developed by Irwin (1957). tance of the rock to crack propagation. It must not
He demonstrated that, for a linear elastic material, the be confused with the tensile strength of the rock To,
magnitude of the stresses in the vicinity of a stress- although these two properties can be related by the
free crack follows an r–1/2 relationship (Fig. 3-6): following formula:
KI K Ic
σ ij = − fij (θ) + ... , (3-30) To = , (3-32)
2 πr πac
where KI is referred to as the stress intensity factor for where ac is a length scale (e.g., flaws or grain size)
the opening mode of deformation of the fracture, fij(θ) characteristic of the rock under consideration.
represents a bounded function depending only on the Irwin’s (1957) approach is similar to Griffith’s
angle θ referenced to the plane of the crack, and r is (1921, 1924). It can be demonstrated that, for an
the distance from the point of interest to the tip of the isotropic and linear elastic material, the stress intensity
fracture. The negative sign is included because, by factor is related to the strain energy release rate by
convention, tensile stresses are negative.
1 − ν2 2
Ge = KI . (3-33)
E
y
As an example of this application to hydraulic frac-
turing, the stress intensity factor for a uniformly pres-
Circumferential
surized crack subjected to a far-field minimum stress

stress, σθθ
σ3 is
(
KI = p f − σ3 ) πL , (3-34)
where pf is the pressure in the crack, L is the half-length

of the crack, and plane strain is assumed. During prop-
Distance from the tip, r agation, the net pressure (pf – σ3) is therefore
Figure 3-6. Stress concentration near the tip of a crack. (p f − σ3 = ) K Ic

πL
. (3-35)
The width w near the tip of a stress-free crack is Using Sneddon’s (1946) solution, the width at the
also a function of the stress intensity factor: wellbore ww is
K Ic 4(1 − ν ) L
2
8(1 − ν 2 ) r ww = . (3-36)
w= KI . (3-31) πL E
E 2π
A propagation criterion based on the stress intensity
In Eq. 3-31, plane strain is assumed. factor is easily implemented in fracture propagation
The stress intensity factor is a function of the load- codes. However, the concept of fracture surface energy
ing parameters and of the geometry of the body. does not imply linear elasticity and can be used for
Hence, length is included in the unit to express KI. fracture propagation in nonlinear materials where the
A fracture propagates when KI reaches a critical value, strain energy release rate is replaced with the J-inte-
known as the critical stress intensity factor KIc or frac- gral (Rice, 1968).
ture toughness. For a perfectly elastic material, KIc is Stress intensity factors are not limited to opening
a material property. It must be evaluated experimen- modes. Other modes exist (Irwin, 1957) to analyze
tally. Experimental results show that for short crack 3D fracture propagation in complex stress fields
lengths, KIc increases with crack length. When this (e.g., propagation from inclined wellbores) where the
scale effect is observed, KIc cannot be considered a fracture changes direction during propagation. Finally,
material property. This behavior is discussed in more fracture mechanics has also been used to explain brit-
detail in Section 3-4.6. tle rock fracture in compression (Germanovich et al.,
The unit for KIc is pressure times the square root of 1994).
length. Fracture toughness is a measure of the resis-

For further details on fracture mechanics, the reader but is perfectly linear with a slope E. At zero stress,
is referred to Cherapanov (1979), Kanninen and part of the deformation has not been recovered. This
Popelar (1985) and Atkinson (1987). represents the plastic strain component in the theory
of elasto-plasticity. Point A is actually the initial yield
stress of the rock. During reloading, the sample
3-3.4. Nonelastic deformation behaves as a perfectly elastic solid up to point B,
As discussed in the next section, most rocks exhibit which is the new yield stress. The increase of yield
nonreversible deformations after unloading, or at least stress with an increase of plastic strain is called strain
a nonunique relationship between stress and strain. hardening, and the decrease of yield stress with an
This means that rocks are not perfectly elastic materi- increase of plastic strain is called strain softening. A
als, and a number of theories have been developed perfectly plastic material is a material with no strain
to model such behaviors. They include the theory of hardening or softening. As shown in this example, the
plasticity, damage mechanics and time-dependent yield stress is a function of the loading history of the
analysis (creep). As an example, the theory of elasto- rock if the rock hardens or softens.
plasticity is briefly described. In elasto-plasticity, part of the strain is predicted by
Figure 3-7 shows the stress-strain relationship of a the theory of elasticity; i.e., any strain increment asso-
cylindrical ideal sample. From O to point A, the rela- ciated with a stress increment is the sum of an elastic
tion between stress and strain is linear, and the slope component and a nonelastic component:
of the curve is Young’s modulus E. The stress-strain dε = dεe + dεp , (3-37)
relation does not change if the sample is unloaded in
this region. This is the region where the theory of where dε is the total strain increment, dεe is the elastic
elasticity applies. Beyond point A, the slope of the strain increment, and dεp is the plastic strain incre-
curve decreases. Moreover, if the sample is unloaded ment. Contrary to the elastic strain component, the
in this region, say at point B, the unloading portion plastic strain component cannot be recovered during
does not follow the same path as the loading portion unloading. Predicting the plastic strain increment
requires a yield criterion that indicates whether plastic
deformation occurs, a flow rule that describes how the
plastic strain develops and a hardening law.
σ The yield criterion is a relationship between stresses
that is used to define conditions under which plastic
deformation occurs. In three dimensions, this is repre-
sented by a yield function that is a function of the
state of stress and a hardening parameter:
B
f (σ1 , σ 2 , σ 3 , h ) = 0 . (3-38)
The hardening parameter h determines the evolution

of the yield curve with the amount of plastic deforma-
A tion of the material. Elasto-plastic deformation with
hardening is important in the study of the stability of
formations prone to sanding. Weak sandstones usually
show hardening behavior, which can be close to linear
hardening. For further details on elasto-plasticity, the
reader is referred to Hill (1951) and Chen and Han
(1988).
E E
O ε 3-3.5. Failure
A failure criterion is usually a relationship between
Figure 3-7. Stress-strain relationship for an elasto-plastic
material with strain hardening. OA = elastic, AB = plastic.
the principal effective stresses, representing a limit

beyond which instability or failure occurs. The

Terzaghi effective stress is used in failure criteria. τ
Several types of criteria have been proposed in the lit-
erature and have been used for various applications. rion
rite
The more popular criteria include the following: mbc
o
oul
hr-C
• Maximum tensile stress criterion maintains that fail- Mo
M
ure initiates as soon as the minimum effective prin-
cipal stress component reaches the tensile strength Co
To of the material: φ 2θ
σ 3 − p = − To . (3-39) σ3′ σ1′ σn
• Tresca criterion expresses that failure occurs when Figure 3-8. Graphical representation of a state of stress at
the shear stress (σ1 – σ3)/2 reaches the characteristic failure.
cohesion value Co:
σ1 − σ 3 = 2Co . (3-40) π φ
θ= + . (3-45)
4 2
• Mohr-Coulomb criterion expresses that the shear
stress tending to cause failure is restricted by the • Mohr failure envelope is a generalization of the lin-
cohesion of the material and by a constant analo- ear Mohr-Coulomb criterion. An example of a more
gous to the coefficient of friction times the effective general model is the following nonlinear model:
normal stress acting across the failure plane:
σ1 − p = σ c + A(σ 3 − p) ,
n
(3-46)
τ = Co + tan(φ)(σ n − p) , (3-41) where A and n are obtained experimentally. The
where φ is the angle of internal friction and Co is failure envelope can also be constructed graphically
the cohesion. The Mohr-Coulomb failure criterion (see Section 3-4.5).
can be rewritten in terms of the principal stresses to As shown here, the Tresca and Mohr-Coulomb cri-
give σ1 at failure in terms of σ3: teria do not include the influence of the intermediate
σ1 − p = σ c + N ϕ (σ 3 − p) , (3-42) stress σ2. Experimental evidence shows they are, in
many cases, good approximations. However, there are
where the coefficient of passive stress Nϕ is other criteria that include the effect of σ2.
π φ
Nϕ = tan 2  +  . (3-43)
 4 2
3-4. Rock mechanical property
The uniaxial compressive strength then becomes
measurement
σ c = 2Co Nϕ . (3-44)
3-4.1. Importance of rock properties
In a ((σn – p), τ) plane, this criterion is a straight in stimulation
line of slope tanφ and intercept Co. A rock fails as
soon as the state of stress is such that the criterion Most of the hydraulic fracture propagation models
is met along one plane, which is also the failure assume linear elasticity. The most important rock
plane. Using the Mohr circle graphical representa- parameter for these models is the plane strain modulus
tion described in Sidebar 3B, this means that the E´, which controls the fracture width and the value of
state of stress at failure is represented by a Mohr the net pressure. In multilayered formations, E´ must
circle that touches the failure envelope. The point be determined in each layer, as the variation of elastic
of intersection can be used to determine the angle properties influences the fracture geometry. Elastic
θ between the normal to the failure plane and the and failure parameters are also used in stress models
direction of σ1, as shown in Fig. 3-8. It can be to obtain a stress profile as a function of depth and
shown that rock properties. These profiles are important for esti-

mating the stress variation between layers and, conse- (Papanastasiou and Thiercelin, 1993) and rock damage
quently, the geometry of hydraulically induced frac- (Valkó and Economides, 1994). In-situ determination
tures. The parameters involved are Young’s modulus, is, however, difficult to achieve.
Poisson’s ratio, the poroelastic coefficient and the fric- Finally, rock failure must be considered in evaluat-
tion angle. The poroelastic stress coefficient η controls ing the long-term stability of the rock around the frac-
the value of stress changes induced by pore pressure ture or at the wellbore. In particular, in weak forma-
changes that result from depletion, injection or fractions (chalk or weak sandstones) part of the rock col-
ture fluid loss. lapses if the drawdown pressure is too high.
The role of fracture toughness in hydraulic fractur-
ing has been the subject of investigation in recent years
(Shlyapobersky, 1985; Thiercelin, 1989; Cleary et al., 3-4.2. Laboratory testing
1991; Johnson and Cleary, 1991; Advani et al., 1992; Uniaxial and triaxial tests are considered the most use-
SCR Geomechanics Group, 1993; Valkó and Econ- ful tests in the study of mechanical rock properties.
omides, 1994). Laboratory measurements give values The difference between them resides in the presence
of KIc of the order of 1000 psi/in.1/2 (at least in the or absence of confining pressure applied to the speci-
absence of confining pressure), whereas fracture propa- men. A typical triaxial testing system is shown sche-
gation models indicate that KIc must be at least 1 order matically in Fig. 3-9. It subjects a circular cylinder
of magnitude larger to influence fracture geometry. of rock to an axisymmetric confining pressure and a
These results are, however, a function of fracture longitudinal or axial load. Generally, these loads are
geometry and pumping parameters. Shlyapobersky similar to the in-situ state of stress. Relationships
(1985) suggested that in-situ fracture toughness, often between the mechanical properties of the rock and the
referred to as apparent fracture toughness, can be much degree of confinement are obtained by performing a
greater than laboratory values because of scale effects. series of tests using different stress and pore pressure
These effects include the influence of heterogeneities, conditions. Also, if the rock is anisotropic, an addi-
discontinuities (Thiercelin, 1989), large-scale plasticity tional series of tests should be performed using differ-
Loading ram
Pore pressure
inlet
Jacketing
material
Sample with
strain gauge
affixed
Pore pressure
outlet
Spherical
seat
Temperature Electrical feed-throughs

Loading frame
controller
Confining
pressure Triaxial cell
system
Stress
Data acquisition
Strain
Pore
pressure
system
Figure 3-9. Triaxial testing configuration.

ent orientations of the cylinder axis with respect to the 3-4.3. Stress-strain curve
plane of anisotropy. During the course of the test, the
primary information recorded is the deformation ver- Figure 3-10 presents a typical stress-strain relationship
sus load relationships (Fig. 3-10), from which both for rocks. The test is conducted under constant con-
Young’s modulus and Poisson’s ratio can be found. fining pressure pc and constant axial strain rate. Mea-
Because these primary elastic constants depend on surements include the values of axial stress, axial strain
confining stress, temperature, pore saturation and and radial strain. When confining pressure is applied
pressure, it is extremely important that the laboratory to the sample, the origin of the stress-strain plot is
environment encompass the representative field situa- usually translated to remove the influence of hydro-
tion to obtain representative data. static loading on the stress and strain (i.e., the axial
stress is actually the differential σa – pc).
During the initial stages of loading, from O to point
σ
A, the rock stiffens. This nonlinear regime is probably
due to the closing of preexisting microcracks pervad-
D
σc ing the specimen. This particular region of the stress-
C strain curve is a signature of the stress history under-
B
gone by the rock specimen during past geologic time,
including the coring process. This characteristic is dis-
cussed later as applied to in-situ stress determinations.
As the load increases further, the stress-strain curve
becomes linear (from point A to point B); this is the
A
portion of the stress-strain curve where the rock
behavior is nearly elastic. If unloading occurs in this
region, the strain returns almost to zero, usually along
εr O εa
a different path. This effect is called hysteresis and
indicates that some energy dissipates during a cycle
Figure 3-10. Stress-strain curves.
of loading and unloading.
When the rock specimen is loaded beyond point B,
The importance of good specimen preparation can- irreversible damage sets in. It is shown by a decrease
not be overemphasized, and the International Society of the slope of the stress versus radial strain curve. At
of Rock Mechanics (ISRM) recommended procedures this stage, the damage is not seen on the axial strain.
that must be followed (Rock Characterization Testing At point C, the axial strain also becomes nonlinear,
and Monitoring; Brown, 1981). The end faces must and large deformations eventually occur. If the rock
be parallel; otherwise, extraneous bending moments is unloaded in this region, permanent strains at zero
are introduced, making correct interpretation of the stress are observed. Point D is the maximum load that
results more difficult. In addition, because of the mis- the rock can sustain under a given confining pressure.
match between the rock properties and those of the Rock failure (i.e., when the sample loses its integrity)
testing platens, shear stresses that develop at the rock/ occurs at about this point. Some rocks, especially
platen interfaces create an additional confinement those with high porosity, may not exhibit a maximum
immediately adjacent to the specimen ends. This dic- peak stress but continue to carry increasing stress (i.e.,
tates the use of specimens with a length:diameter ratio continue to harden).
of at least 2 or the use of appropriate rock inserts or Another interesting rock characteristic is revealed
adaptive lubricant. The loading rate should also be by the volumetric strain, defined as the change in vol-
maintained between 70 and 140 psi/s to avoid ume with respect to the original specimen volume.
dynamic effects. Finally, some rock types (such as For a triaxial test, the volumetric strain of the cylindri-
shales) are sensitive to the dehydration of natural pore cal specimen is εa + 2εr , where εa is the axial strain
fluids; care must be taken to preserve their integrity by and εr is the radial strain. As seen on Fig. 3-11, the
avoiding drying cycles and contact with air during volumetric strain versus axial stress can reverse its
specimen recovery, storage and test preparation. trend upon reaching point E; i.e., the rock specimen

of rock type but also confining pressure, loading rate

σ and temperature, with a general transition from brittle
to ductile behavior with an increase in confining pres-
sure, increase in temperature and decrease in loading
rate. Moreover, in porous rocks (e.g., sandstones,
shales), a transition from dilatant to compactant
E behavior with confining pressure is also observed.
3-4.4. Elastic parameters

As discussed previously, rocks are not perfectly elas-
tic. Especially in soft rocks, it could well be difficult
to find a portion of the stress-strain curves that exhib-
its nearly elastic behavior. On the other hand, the
knowledge of elastic parameters is of great impor-
tance for engineering applications, and assuming, as
εv
a first approximation, that the rock behaves as an elas-
O
tic material has significant advantages.
Figure 3-11. Axial stress versus volumetric strain. There are two main approaches to elastic parameter
determination. The first one is to find elastic param-
eters that can be used to predict as close as possible
starts to increase in volume under additional compres- the behavior of the rock along an expected loading
sive load. This is referred to as dilatancy. path. These parameters do not measure the real elastic
Dilatancy is responsible for the nonlinearity that is component of the rock but approximate the rock
observed in the radial strain and consequently in the behavior. This is the approach used in engineering
variation of volume. It is due either to the creation of design, although the assumption underlying the mea-
tensile cracks that propagate in a direction parallel to surement must be kept in mind. The other approach
the axis of loading (in that case, point B shown in is to develop a test procedure that measures, as close
Fig. 3-10 is distinct from point C) or to frictional as possible, the elastic component of the strain. This
sliding along rough surfaces and grains (in that case, approach is useful if a correlation is sought between
point B is close to point C). Soft rocks under confin- downhole measurements made using sonic tools and
ing pressure could show a decrease in volume instead core measurements. Because of the variety of
of an increase, because of compaction. This is typical approaches that can be used, it is essential to always
of chalk and weak sandstones. Compaction in cohe- mention how elastic properties have been measured.
sive rocks requires the destruction of cohesion, which Elastic property measurement can be made under
could create a sanding problem during production. static conditions or under dynamic conditions. The
Finally, if the framework of elasto-plasticity is used, ratio of the dynamic to static moduli may vary from
point B is the initial yield point. If the nonelastic com- 0.8 to about 3 and is a function of rock type and con-
ponent of the variation of volume is negative, the rock fining stress. In most cases, this ratio is higher than 1
is dilatant; otherwise, the rock is compactant. (e.g., Simmons and Brace, 1965; King, 1983; Cheng
Brittle rocks and ductile rocks must also be differ- and Johnston, 1981; Yale et al., 1995). Possible expla-
entiated. Brittle rocks are characterized by failure nations for these differences are discussed in the fol-
prior to large nonelastic deformation. Low-porosity lowing. Static elastic properties, as measured in the
sandstones and hard limestones are typical brittle laboratory during sample loading (see the following
rocks. Ductile rocks are characterized by the absence section), are generally assumed more appropriate than
of macroscopic failure (i.e., theoretically, the rock will dynamic ones for estimating the width of hydraulic
yield indefinitely). Salt, young shales and very high fractures. Knowledge of dynamic elastic properties is,
permeability sandstones are typical ductile rocks. however, required to establish a calibration procedure
These behaviors, however, are functions not only to estimate static downhole properties from downhole

measurements, which are obtained essentially from Poisson’s ratio is determined using similar meth-
sonic tools (see Chapter 4). ods and the axial strain–radial strain curve.
These elastic constants must be adjusted to the
• Static elastic properties
proper reservoir conditions for design purposes.
Static elastic properties are usually measured using Moreover, for stimulation purposes an approach
the equipment described in Section 3-4.2. For clas- that requires failure of the sample is not necessary.
sification purposes, the ISRM proposed the follow- The best method is either to use the average modu-
ing recommended procedures that use, for the mea- lus or to measure a tangent modulus at a state of
surement of Young’s modulus, the axial stress–axial stress near the expected downhole effective state
strain curve measured during the loading of the of stress. The average value simulates the effect of
sample (Brown, 1981) (Fig. 3-12): the width, causing stresses that are maximum at the
– tangent Young’s modulus Et —the slope at a fracture face and decay to zero away from the face.
stress level that is usually some fixed percentage Ideally, the sample must be tested in a direction nor-
of the ultimate strength (generally 50%) mal to the expected hydraulic fracture plane (i.e., in
– average Young’s modulus Eav —determined from the horizontal direction if the fracture is expected
the average slopes of the generally straight-line to be vertical). The best way to reproduce downhole
portion of the curve conditions is probably to apply a confining pressure
equal to the effective downhole mean pressure
– secant Young’s modulus Es —usually the slope (σh + σv + σH)/3 – p, where σh, σv and σH are the
from zero stress to some fixed percentage of the minimum horizontal stress, vertical stress and maxi-
ultimate strength (generally 50%). mum horizontal stress, respectively. The tangent
properties are then measured using an incremental
increase of the axial load. Terzaghi’s effective stress
σ
is used here rather than the Biot effective stress con-
C cept because the tangent properties are essentially
controlled by this effective stress (Zimmerman et
σfailure al., 1986).
The second approach utilizes small unloading-
loading cycles that are conducted during the main
loading phase. If the cycle is small enough, the
slope of the unloading stress-strain curve is close
to that of the reloading stress-strain curve (Fig. 3-13,
Hilbert et al., 1994; Plona and Cook, 1995). This
leads to the measurement of elastic properties that
0.5σfailure A are close to the actual ones and also close to the
value determined using ultrasonic techniques. It is
also important to perform these measurements at
the relevant confining pressure and axial stress.
• Elastic properties determined using sonic
measurements
B Sonic measurements are conveniently used to deter-
O εa mine the elastic properties under dynamic conditions
in the laboratory. These properties are also called
Figure 3-12. ISRM-recommended methods to measure dynamic elastic properties. To obtain them, a mech-
Young’s modulus: anical pulse is imparted to the rock specimen, and
• derivative of the stress-strain curve at point A is Et, the time required for the pulse to traverse the length
measured at 50% of the ultimate strength of the specimen is determined. Then, the velocity
• slope of the straight line BC is Eav
• slope of the straight line OA is Es. of the wave can be easily calculated. Again, these
measurements should be performed under simulated

Applied load
σ
Transmitter transducer
B
σB
Platen
A Environmental Specimen
chamber
Platen
O εa
Receiver transducer
Figure 3-13. Young’s modulus measured using small cycles
(Hilbert et al., 1994). Young’s modulus at σB is the slope of Applied load
line AB.
Figure 3-14. Ultrasonic pulse measurement.
downhole conditions and can be conducted during
triaxial compression tests (Fig. 3-14).
As also discussed in Chapter 4, two types of elas- elastic component of the rock (Hilbert et al., 1994;
tic body waves can be generated: compressional Plona and Cook, 1995). Because of poroelastic
(also called P-waves) and shear (S-waves). Elastic effects and rock heterogeneity, the acoustic velocity
wave theory shows that the velocities of P- and is also a function of wave frequency. But in dry
S-waves (uP and uS, respectively) are related to the rocks, the influence of the frequency appears to
elastic constants through the following relationships be of second order compared with that of the strain
(in dry rocks): amplitude (Winkler and Murphy, 1995). Conse-
quently, when the dynamic and static small-ampli-
1/ 2
1/ 2 K + 4 G  tude loading/unloading measurements are compared,
C   dyn 3 dyn  their values agree quite well (Fig. 3-15; Plona and
uP =  dyn  = 
 ρ   ρ  Cook, 1995).
 
Correlations can be established between static and
( )
1/ 2
 Edyn 1 − ν dyn  dynamic moduli (Coon, 1968; van Heerden, 1987;
=  (3-47)
 ρ 1 (
+ )(
ν dyn 1 − 2 ν dyn )  Jizba and Nur, 1990). Coon demonstrated that the
coefficient of correlation can be improved if consid-
1/ 2
eration of the lithology is included. These correla-
G 
1/ 2
 Edyn  tions allow an estimation of large-amplitude static
uS =  dyn  =  ,
(
(3-48)
 ρ   2ρ 1 + ν dyn ) 
in-situ values from log data where core data are not
available (see Chapter 4). Figure 3-15 suggests
where ρ refers to the mass density of the rock speci- another procedure in which a corrective factor is
men and the relationship between the various elastic found by the ratio of the loading to unloading tan-
moduli is as in Sidebar 3C. The subscript dyn refers gent moduli for low-amplitude static tests.
to dynamic, as the values of the elastic constants • Scale effects in elastic properties
obtained by dynamic techniques are in general higher
The elastic properties of rock are scale dependent,
than those obtained by static methods. This differ-
as are any rock properties. This means that the
ence is now believed to be due mainly to the ampli-
value of an elastic parameter that is determined on
tude of the strain, with the very low amplitude
a laboratory sample may be quite different of that
dynamic measurements representing the actual
of a rock mass, mainly because of the presence of

11
a fracture propagation model to invert the pressure
Dynamic E response in terms of elastic properties. The geometry
10 Small-amplitude static E and mechanical assumptions of the fracture propa-
Large-amplitude static E
9 gation model must be as close as possible to the
8
actual situation. If the fracture propagates radially,
this technique can extract an estimate of the plane
Young's modulus (GPa)
7
strain modulus E´ (Desroches and Thiercelin, 1993).
6
• Poroelastic properties
5
For isotropic rocks, it is generally recommended to
4
conduct tests that measure the volumetric response
3 of the sample, as poroelastic effects are volumetric
2 ones. Three tests are usually made to measure the
five properties that characterize an isotropic poro-
1
elastic material. All three tests involve hydrostatic
0 loading but differ on the boundary conditions
–2 0 2 4 6 8 10 12
applied to the pore fluid. For the drained test, the
Stress (MPa)
fluid in the rock is maintained at constant pressure;
Figure 3-15. Dynamic versus static Young’s modulus mea- for the undrained test, the fluid is prevented from
surements (after Plona and Cook, 1995). escaping the sample; and for the unjacketed test, the
pore pressure is maintained equal to the confining
pressure. The reader is referred to Detournay and
discontinuities in the rock mass. Various approaches Cheng (1993) for further information. Presented
are being developed to take this phenomenon into here is the determination of α, which, with knowl-
consideration (Schatz et al., 1993). An alternative is edge of the drained Poisson’s ratio, allows determi-
to determine the properties downhole, as described nation of the poroelastic stress coefficient η, which
in the next section. However, downhole measure- is probably the most important poroelastic parame-
ments are usually limited to a scale on the order of ter for hydraulic fracturing applications. This mea-
3 ft, whereas a large fracture involves a scale on the surement is conducted using the drained test, in
order of 100 ft. Rock imperfection on this scale can which the volume change of the sample ∆V and the
be mapped by a combination of wellbore seismic volume change of the pore fluid ∆Vf are measured
and sonic measurements. as a function of an incremental increase of the con-
• Elastic properties determined using downhole fining pressure. The value of α is then given by the
measurements following relation:
Downhole measurements are made to estimate the ∆Vf
elastic properties. Dynamic log measurements are α= . (3-49)
∆V
described in detail in Chapter 4. Other techniques
include direct downhole static measurements and As for the elastic properties, the test must be con-
inversion of the pressure response obtained during ducted with a confining pressure close to the down-
a micro-hydraulic fracturing test. A direct downhole hole mean stress. These properties must be tangent
static measurement requires measuring the deforma- properties and, for practical purposes, are a function
tion of a small portion of the wellbore during pres- of the Terzaghi effective stress. Mathematical con-
surization. This can be done by using downhole sideration and experimental results confirm that
extensiometers (Kulhman et al., 1993). Usually this poroelastic properties are controlled by the Terzaghi
technique yields only the shear modulus G. Pressure effective stress (Zimmerman et al., 1986; Boutéca
inversion techniques (Piggott et al., 1992) require et al., 1994).

3-4.5. Rock strength, yield criterion and A specific case is the study of pore collapse. Pore
failure envelope collapse is usually not associated with a sudden loss
of integrity and therefore has to be detected from the
The strength of a rock is the stress at which the rock initial yield envelope rather than the failure envelope.
fails (i.e., the rock loses its integrity). This strength In some instances, yield can be initiated under iso-
obtained with a uniaxial test is called the uniaxial tropic loading (Fig. 3-17). The portion of the yield
compressive strength σc (UCS). The overall strength curve that shows a decrease of the shear stress at yield
of rocks is a relationship between the principal effec- as a function of the confining pressure is characteristic
tive stress components (in the sense of Terzaghi, see of compactant materials. This usually occurs with
Section 3-3.5). This relationship is called the failure poorly consolidated rocks. In cohesive materials
criterion, and its graphical representation is called the compaction is associated with pore collapse and
failure envelope. consequently with cohesion loss. This is a potential
To obtain the failure envelope of a particular rock failure mechanism of the matrix that could lead to
type, a series of triaxial tests should be performed the production of formation particles (e.g., sanding).
under different confining pressures until failure of the
specimen occurs for each condition. There are various
ways to represent the failure envelope. The classic τ
approach in rock mechanics is to plot the effective
stresses at failure for each test using a Mohr circle rep-
resentation (see Sidebar 3B) of diameter (σfailure´ – σ3´),
where σfailure´ represents the ultimate strength of the Failure line
specimen measured under confinement σ3´ (Fig. 3-16).
The envelope of these circles is a locus separating sta- Zone of pore
ble from unstable conditions. It should be emphasized collapse
that the failure of rocks occurs when the matrix Yield envelope
stresses reach a critical level; hence, the failure enve-
lope represents a relationship between the “effective”
stress levels. Therefore, the knowledge of such a char- σn
acteristic can also be used to put some limits on the
allowable variation of the reservoir pore pressure dur- Figure 3-17. Failure and initial yield envelopes for poorly
ing production. Indeed, a change in pore pressure cor- consolidated sandstones.
responds to a translation of the pertinent Mohr circle
along the normal stress axis.
3-4.6. Fracture toughness
τ To = Tensile strength Determining the value of fracture toughness requires
σc = Uniaxial compressive strength using a sample that contains a crack of known length.
The stress intensity factor, which is a function of the
load and sample geometry, including the length of the
Failure envelope
preexisting crack, is then determined. Testing mea-
2 sures the critical load and, therefore, the critical stress
3 intensity factor KIc at which the preexisting crack is
reinitiated. Another approach is to measure the frac-
1
ture surface energy and use Eq. 3-33. An example
using a simple geometry is discussed in Sidebar 3D.
To 0 σ´3 σć σ´failure σn Testing the sample under downhole conditions is also
required because fracture toughness increases with
Figure 3-16. Failure envelope. 1 = Mohr circle correspond- effective confining pressure and is affected by temper-
ing to uniaxial tensile test; 2 = Mohr circle corresponding to
uniaxial compressive test; and 3 = Mohr circle correspond- ature. Various sample geometries have been proposed,
ing to triaxial test with effective confining stress σc´ and fail- but the most practical ones from an engineering point
ure stress (i.e., ultimate strength) σfailure´.

−1/ 2
3D. Fracture toughness testing Fc  L  L 
K Ic ≈ πL 1 −  1−  , (3D-2)
2b  2b   b 
To illustrate the measurement of fracture toughness and the
where Fc is the critical load.
influence of a crack on material behavior, a bar of unit thick-
It can also be demonstrated that the area OAB in Fig. 3D-1
ness containing a central crack of length 2L is considered
corresponds to the energy dWs that was dissipated to propa-
(Fig. 3D-1). Although this is a simple geometry, in practice
gate the crack from 2L to (2L + 2∆L). The strain energy
such an extension test is difficult to conduct. The crack length
release rate is, therefore, the dissipated energy divided by the
is supposed to be small compared with the bar width and the
created surface area 2∆L:
width small compared with the bar length. The stress intensity
factor for this geometry is given by dW s
Ge = . (3D-3)
−1/ 2
2∆L
 L  L 
K I ≈ σ πL 1−  1−  , (3D-1) A similar approach can be used if the crack is propagated
 2b   b 
to the sample end; in that case:
where σ is the stress applied to the sample (i.e., F/2b). Ge =

dW s
, (3D-4)
Figure 3D-1 also shows a plot of the load versus displace- 2b
ment curve. The load increases to the point where the crack
starts to propagate. During stable crack propagation, the load where dWs corresponds to the area under the load-displace-
decreases. If the sample is unloaded at this stage, the load- ment curve and the initial crack length is assumed to be small
displacement curve exhibits a slope different from the one compared with the sample width (i.e., ∆L ≅ b). Using this
obtained during initial loading. However, the displacement is approach, there is no need to measure crack length. For lin-
recovered upon complete unloading. This behavior is funda- ear elastic behavior:
mentally different from that of elasto-plasticity, and a perfectly
1− ν 2 2
brittle behavior is exhibited. The change of slope is not a Ge = KI . (3D-5)
material property but is due to the increased length of the E
crack. It can, therefore, be used to estimate the crack length.
The load-displacement curve shown in Fig. 3D-1 can also be
The critical stress intensity factor is the value of KIc when
used to determine the process zone behavior (Labuz et al.,
the crack starts to propagate:
1985).
F
2b A
Fc
2L
B
2L + 2∆L
O δ
Figure 3D-1. Fracture toughness measurement. The shaded area on the left of the plot represents the energy required
to propagate the crack from 2 L to (2 L + 2∆L).

of view are those based on core geometries (Ouchter-

lony, 1982; Thiercelin and Roegiers, 1986; Zhao and 6200
Mudstone
Roegiers, 1990; ISRM Commission on Testing
Sandstone
Methods, 1988, 1995).
However, the existence of very large stress values
near the tip of the crack makes it difficult to develop
6400
a rigorous test configuration because a cloud of micro-
cracks is created ahead of the crack tip. This is com-
monly referred to as the process zone (Swanson and
Spetzler, 1984; Labuz et al., 1985). The extent of this
nonlinear region must be limited so that it does not
6600
reach the edge of the laboratory sample. Also, this
process zone must be relatively small compared with
the size of the crack if linear elastic calculations are
to be valid (Schmidt, 1976; Schmidt and Lutz, 1979;
Boone et al., 1986). 6800
The development of the process zone is one of the
causes of the scale effects that are observed in fracture
Depth (ft)
toughness testing; i.e., the determined value of frac-
ture toughness increases with sample size. Modeling
of process zone behavior can be conducted using the 7000
information obtained during tensile failure of a speci-
men (see Sidebar 3D). Modeling can also give some
insight on the tip behavior of large-scale hydraulic
fractures (Papanastasiou and Thiercelin, 1993).
7200
3-5. State of stress in the earth

The propagation and geometry of hydraulic fractures
are strongly controlled by the downhole state of stress.
7400
In particular, it is generally accepted that the degree
of fracture containment is determined primarily by the
in-situ stress differences existing between layers. In the
absence of a meaningful stress contrast, other mecha-
nisms such as slip on bedding planes (Warpinski et al., 7600
1993) and fracture toughness contrast (Thiercelin et al., 5000 6000 7000 8000
1989) can have a role. Moreover, hydraulic fractures Stress (psi)
propagate, in most cases, normal to the minimum stress
direction. Consequently, knowledge of the minimum Figure 3-18. Stress profile for Well MWX-3 (Warpinski and
Teufel, 1989).
stress direction allows prediction of the expected direc-
tion of the hydraulic fracture away from the wellbore.
Stresses in the earth are functions of various para- tion between regional and local considerations is
meters that include depth, lithology, pore pressure, important as it controls the stress variation between
structure and tectonic setting. A typical example from layers. In some stress regimes the adjacent layers are
the Piceance basin in Colorado (Warpinski and Teufel, under higher stress than the pay zone, enhancing frac-
1989) is shown in Fig. 3-18. The stress regime in a ture height containment; in others, the adjacent layers
given environment depends, therefore, on regional are under lower stress than the pay zone, and fracture
considerations (such as tectonics) and local considera- propagation out of the zone is likely, limiting lateral
tions (such as lithology). Understanding the interac- fracture penetration. Key regional stress regimes and

the consequences of these regimes on the local state mate values of Ko. However, stress predictions using
of stress in a reservoir are reviewed in the following. these assumptions must be used with great caution
These regimes lead to the introduction of simple stress and may not be applicable in lenticular formations
models that allow making rough estimates of the (Warpinski and Teufel, 1989). Nevertheless, they are
stress profile as a function of depth and rock properties. useful for understanding the state of stress in the earth
These models can also be used to obtain a calibrated and can be used as a reference state (Engelder, 1993).
stress profile from log and stress measurement infor- With the assumption of elasticity and for the bound-
mation, as shown in Chapter 4. The influence of the ary conditions outlined previously, Ko is
variation of temperature and pore pressure on the state ν
of stress is also analyzed. Finally, the influence of Ko = , (3-52)
1− ν
industrial intervention on the state of stress is pre-
sented. Intervention includes drilling a hole and and the relationship between the total minimum hori-
depleting or cooling a formation. zontal stress σh and the overburden σv is, after re-
arranging and using the Biot effective stress for σ´,
3-5.1. Rock at rest ν

σh = σ v + 2 ηp . (3-53)
1− ν
One stress regime is when the rock is under uniaxial
strain conditions (i.e., there is no horizontal strain any- The dependence of horizontal stress on rock lithol-
where). To estimate the state of stress that is generated ogy results from the dependence of Poisson’s ratio ν
under this regime, it is assumed that the rock is a on rock lithology. In most cases, the model predicts
semi-infinite isotropic medium subjected to gravita- that sandstones are under lower stress than shales as
tional loading and no horizontal strain. Ko in sandstones and shales is about equal to 1⁄3 and 1⁄2,
Under these conditions, the vertical stress is gener- respectively. The use of Eq. 3-53 to obtain stress pro-
ated by the weight of the overburden and is the maxi- files in relaxed basins is presented in Section 4-5.2.
mum principal stress. Its magnitude, at a specific More complex elastic models that are associated with
depth H, is estimated by this stress regime have been developed to consider
H
rock anisotropy (Amadéi et al., 1988) and topography
σ v = ∫ ρ( H )gdH , (3-50) (Savage et al., 1985).
0
For purely frictional materials, Ko can be approxi-
mated by (1 – sinφ) (Wroth, 1975), which gives the
where ρ is the density of the overlying rock masses
following relationship for the total stresses:
and g is the acceleration of gravity. The value of this
stress component is obtained from the integration of σ h ≈ (1 − sinφ)σ v + sin(φ) p , (3-54)
a density log. The overburden gradient varies from
about 0.8 psi/ft in young, shallow formations (e.g., where φ is the angle of internal friction of the rock
Gulf Coast) to about 1.25 psi/ft in high-density (Eq. 3-41), of the order of 20° for shales and 30° for
formations. Assuming that quartz has a density of sandstones. In this expression, the Terzaghi effective
165 lbm/ft3, the overburden gradient ranges between stress concept prevails because this case involves fric-
the well-known values of 1.0 and 1.1 psi/ft for brine- tional behavior.
saturated sandstone with porosity ranging between This equation implies that rocks with a high value
20% and 7%, respectively. of friction angle are under lower stress than rocks with
With uniaxial strain assumed, the other two princi- low value of friction angle; i.e., in general, sandstones
pal stresses are equal and lie in the horizontal plane. are under lower stress than shales. The observation
If they are written in terms of effective stress, they are that models based on elasticity and models based on
a function of only the overburden: frictional behavior give the same trend of stress contrast
always occurs, although the fundamental assumptions
σ ′h = Ko σ ′v , (3-51) for these models have nothing in common.
where Ko is the coefficient of earth pressure at rest For purely viscous materials (salt), Ko is simply
and σh´ is the minimum effective horizontal stress. equal to 1 and the state of stress is lithostatic (Talobre,
Assumptions about rock behavior can be used to esti- 1957, 1958):

σh ≈ σv (3-55)
Formation A
(a lithostatic state of stress as such does not require
the uniaxial strain condition, and therefore, it defines Stiff
a stress regime by itself). fixed plate Formation B Stiff plate
Over geologic time, rock experiences, in various
combinations and degrees, diverse mechanical behav- Formation C
iors and various events. Behaviors include elastic,
frictional and viscous behaviors, and events include
the occurrence of tectonic strain, variation of pore
pressure and temperature, erosion and uplift. As Constant displacement
reviewed by Prats (1981), these mechanisms lead to

Figure 3-19. By analogy, the stiffer the spring, the more
deviations from these simple reference states, some load it will carry.
of which are briefly reviewed here.
is a principal stress but not necessarily the maximum.

3-5.2. Tectonic strains The state of stress described in this section cannot
be considered to define a particular stress regime
Tectonic stresses and strains arise from tectonic plate
(although one could speak of compressional stress
movement. In this section, the notion of tectonic strain
regime) as it does not define a reference state. Only if
is introduced, which is a quantity added to or sub-
the strains are high enough for the rock to fail are ref-
tracted from the horizontal strain components. If
erence states obtained, as discussed in the next section.
incremental tectonic strains are applied to rock forma-
tions, these strains add a stress component in an elas-
tic rock as follows: 3-5.3. Rock at failure
E Eν
dσ h ≈ dε h + dε H (3-56) If the strains are high enough, the rock fails either
1− ν 2
1 − ν2 in shear or in tension. Three stress regimes can be
defined if the rock fails in shear. These stress regimes
E Eν
dσ H ≈ dε H + dε h , (3-57) are associated with the three classic fault regimes
1− ν 2
1 − ν2 (Anderson, 1951): normal, thrust and strike-slip fault
where dεH and dεh are the (tectonic) strains with dεH > regimes (Fig. 3-20). Stresses can be estimated by the
dεh. The resulting stress increments are not equal, with adapted shear failure model. The simplest shear fail-
dσH > dσh, where dσH is the stress increment gener- ure model that applies to rocks is the Mohr-Coulomb
ated in the dεH direction and dσh is the stress increment failure criterion. A stress model based on this criterion
generated in the dεh direction. These relations are assumes that the maximum in-situ shear stress is gov-
obtained by assuming no variation of the overburden erned by the shear strength of the formation (Fenner,
weight and provide a dependence of stress on Young’s 1938). Hubbert and Willis (1957) used this criterion
modulus E. This means that the greater the Young’s and sandbox experiments in their classic paper on
modulus, the lower the horizontal stress if the strains rock stresses and fracture orientation (see Sidebar
are extensive and the higher the horizontal stress if the 3A). As presented in Eq. 3-42, the Mohr-Coulomb
strains are compressive. To understand this mecha- failure criterion can be written to give σ1 at failure
nism, the different layers can be compared to a series in terms of σ3. In sandstones and shales, Nϕ is about
of parallel springs, the stiffness of which is propor- equal to 3 and 2, respectively.
tional to Young’s modulus as depicted in Fig. 3-19. If failure is controlled by slip along preexisting sur-
This model is actually a good qualitative description faces, the compressive strength σc can be assumed
of the state of stress measured in areas in which com- negligible. However, a residual strength may still
pressive tectonic stresses occur. The model can exist. The angle of internal friction φ is usually mea-
account for situations where sandstones are under sured by using ultimate strength data as a function of
higher horizontal stress than adjacent shales (Plumb the confining pressure obtained during triaxial testing.
et al., 1991; see also Chapter 4). The overburden stress This angle can also be measured by using residual

Normal fault regime

σv = σ1
σH = σ2 σh = σ3
Thrust fault regime
σv = σ3
σh = σ2
σH = σ1
Strike-slip fault regime
σv = σ2
σh = σ3
σH = σ1
Figure 3-20. The three fault regimes (Anderson, 1951).
strength data as a function of the confining pressure If the formation is in extension (i.e., normal fault
obtained during triaxial testing once the sample has regime, Fig. 3-20), the vertical stress is the maximum
failed. Using the residual angle of friction rather than principal stress. The minimum principal stress is in
the angle of internal friction in a failure stress model the horizontal plane and is therefore σh. Equation 3-42
should be more consistent with the assumption that becomes
the minimum stress is controlled by friction along pre-
(σ v − p) ,
1
existing planes. Generally, the residual angle of fric- σh − p ≈ (3-58)
Nϕ
tion is smaller than or equal to the internal angle of
friction.

in which the effect of strength is neglected. An equa- stress component) or when the in-situ stress ratio is
tion similar to Eq. 3-51 can be retrieved. However, if too large. As the rock is close to a uniaxial state of
the rock is at failure, the coefficient of proportionality stress, this regime can occur only in rocks with a com-
cannot be considered as a coefficient of earth stress at pressive strength high enough to avoid normal fault-
rest. The most surprising and confusing result is that, ing (as a rule of thumb, the uniaxial compressive
in practice, Eqs. 3-53 and 3-58 give similar predic- stress must be equal to or greater than the effective
tions, especially if, in the elastic model, α is assumed overburden stress). This condition is achieved for tight
equal to 1. The coefficient of proportionality in sand- gas sandstones in some areas of the Western United
stones and shales is, whether elasticity or failure is States and East Texas.
assumed, about equal to 1⁄3 and 1⁄2, respectively. This Failure models also have an important role in provid-
similarity has been demonstrated in more detail for ing bounds for the in-situ stress. They represent a limit
one area of East Texas by Thiercelin and Plumb state above which the rock is unstable in the long term.
(1994b). In the extension regime in particular, it is unlikely that
If the formation fails under compressive tectonic a minimum stress value below the value predicted by
strain, the maximum principal stress is in the horizon- the failure model can be obtained.
tal plane and is therefore σH. In the thrust fault
regime, the minimum principal stress is the vertical
stress (Fig. 3-20): 3-5.4. Influence of pore pressure
σ H − p ≈ Nϕ (σ v − p) . (3-59) It is of interest to understand what happens when
depleting or injecting into a reservoir. Elastic models
In this case, σh is the intermediate principal stress with uniaxial strain conditions can be applied with
and is equal to or greater than the vertical stress. some confidence, as the variation of stresses occurs
Horizontal hydraulic fractures could be achieved. over a short period of geologic time, although it is
Thus, the principal stresses can be estimated and always necessary to double check the assumptions
ordered by looking at the fault regime. In practice, because failure models could well be the real physical
these considerations must be checked with downhole mechanism, as shown in the following.
measurements, as the state of stress may deviate from If the material behaves elastically, and assuming
the expected ordering of stresses because of stress his- uniaxial strain conditions, Eq. 3-53 gives
tory. These models assume that the fault plane was
created under the current tectonic setting; i.e., the nor- dσ h = 2 ηdp . (3-60)
mal to the fault plane makes an angle (π/4 + φ/2) with The range of 2η is approximately between 0.5 and
the direction of the maximum principal stress. Pre- 0.7. Geertsma (1985) demonstrated the applicability
existing faults can be reactivated under a state of stress of this model to stress decrease during depletion.
that differs from the one that created them. A Mohr- A failure model can also be applied. For example,
Coulomb stability criterion can still be applied, but Eq. 3-58 gives
Eq. 3-42 must be modified to take into consideration
that the fault plane orientation was not induced by the Nϕ − 1
dσ h = dp . (3-61)
current state of stress. Nϕ
Another stress regime is associated with tensile fail-
If the coefficient of friction is 30°, the coefficient
ure. Tensile failure is sometimes observed downhole,
of proportionality is 0.67. As previously, a strong sim-
although it appears to contradict the general compres-
ilarity exists between the predictions from the elastic
sional regime of the earth. This mode of failure sim-
and failure models. To use a failure model, however,
ply states that σ3 – p = 0 (by neglecting the tensile
requires checking that the effective state of stress sat-
strength of the rock) and may be suspected if it is
isfies the failure criterion prior to and during the varia-
observed from downhole images that the normal to
tion of pore pressure. The effective stresses increase
the plane of the preexisting fractures is the direction
during depletion, although the total minimum stress
of minimum stress. This condition can occur in exten-
σh decreases.
sional regions with overpressured zones (where the
Field data generally support the predictions of these
pore pressure tends to be the value of the minimum
models and show that variation in the minimum stress

ranges from 46% to 80% of the change in pore pres- far-field stress and the stress value. An example is the
sure (Salz, 1977; Breckels and van Eekelen, 1982; stress field at the top of the Ekofisk formation, where
Teufel and Rhett, 1991). the maximum principal horizontal stress is oriented
perpendicular to the structure contour around Ekofisk
dome (Teufel and Farrell, 1990).
3-5.5. Influence of temperature
Temperature variation also changes the state of stress
(Prats, 1981). Cooling happens during uplift or the
3-5.7. Stress around the wellbore
injection of a cool fluid. This induces an additional So far, only the far-field stress components resulting
stress component in the horizontal plane, which using from geologic contributions or reservoir production
the uniaxial strain assumption again is have been considered. In addition, the magnitude and
Eα T orientation of the in-situ stress field can be altered
dσ = dT , (3-62) locally, as a result of excavation. These induced
1− ν stresses usually result in large stress concentrations,
where dT is the temperature variation and αT is the differing significantly from the original values.
linear thermal expansion coefficient. In this case, an Drilling a borehole, for example, distorts the preexist-
influence of Young’s modulus on the state of stress is ing stress field. The following expressions can be
also obtained. Cooling the formation reduces the nor- obtained for the stresses around the wellbore, where
mal stress; hence, cool-water injection could lead to σx and σy are principal stresses in the x-y plane, pw is
tensile fracturing of the formation in the long term. the wellbore pressure, rw is the wellbore radius, and r
is the distance from the center of the well (Fig. 3-21):
3-5.6. Principal stress direction  r2  1  4r 2 3r 4 

σr =
1
2
( ) ( ) r2
σ x + σ y 1 − w2  + σ x − σ y 1 − 2w + 4w cos2θ + pw w2
 r  2  r r  r
Figure 3-20 indicates the expected direction of the
( )1 + rr  1
( )1 + 3rr 
2 4
minimum stress as a function of the fault regime 1 r2
σθ = σx + σy w
2  − σx − σy
w
4
2θ − pw w2
cos2
(Anderson, 1951). In practice, it is observed that at 2  2  r
( )1 + 2rr 3rw4 

2
shallow depths the minimum principal stress is the 1
τ rθ = − σx − σy w
2
− sin 2θ .
vertical stress; i.e., a hydraulic fracture is most likely 2 r4 
to occur in a horizontal plane. The transition between (3-63)
a vertical minimum principal stress and a horizontal
minimum principal stress depends on the regional sit-
uation. In an extension regime, however, the mini- σy
mum stress direction can be expected to be always in σx
the horizontal plane, even at shallow depths. This is
usually not observed, probably because of the exis- σθ = f(r)
3σy – σx
tence of residual stresses and because vertical stress 3σx – σy
is usually the minimum principal stress at shallow σr = f(r) σθ = f(r)
depths. In normally pressured sedimentary basins, the σx σy
0
minimum stress is most probably in the horizontal σx 0 σx
plane at depths greater than 3300 ft (Plumb, 1994b).
Stress rotation may also occur because of topology.
However, at great depths, rotation is induced mainly σr = f(r)
by fault movement. In some situations, overpressur-
ization has been observed to generate a change in the σy
σy
ordering of stress, with the value of the minimum hor-
izontal stress higher than that of the vertical stress. Figure 3-21. Stress concentration around a circular hole in
Finally, changes in structural or stratigraphic position the absence of wellbore pressure.
can locally affect the stress direction dictated by the

To derive these expressions, it is assumed that the ular, σθ at the wellbore becomes a function of time if
rock remains linear elastic, the borehole is drilled par- σx is not equal to σy (Detournay and Cheng, 1988).
allel to one of the principal stress directions, and the The long-time solutions are
wellbore fluid pressure pw does not penetrate the rock
(e.g., because of the presence of mudcake). At the (σ )
θ θ =0 = 3σ y − σ x − pw + 2 η( pw − p) (3-68)
borehole wall (i.e., r = rw), the following expressions
are obtained: (σ )
θ θ= π = 3σ x − σ y − pw + 2 η( pw − p) . (3-69)
2
σ r = pw If the wellbore fluid pressure is higher than the far-

τ rθ = 0 field pore pressure, poroelastic effects increase the
( ) ( )
σ θ = σ x + σ y − 2 σ x − σ y cos2θ − pw . (3-64)
stress concentration at the wellbore. The initiation
pressure (Haimson and Fairhust, 1969) is obtained
from Eq. 3-69, with (σθ – pw) = –To:
Considering only the directions parallel and per-
pendicular to the minimum stress direction (i.e., 3σ h − σ H + To − 2 ηp
θ = 0 and θ = π/2, respectively), these expressions pif = . (3-70)
2(1 − η)
further simplify:
A typical value of η is 0.25.
(σ )
θ θ =0 = 3σ y − σ x − pw (3-65)
These equations are used in Section 3-6.2. Finally,
(σ )
θ θ= π = 3σ x − σ y − pw . (3-66)
plasticity effects reduce the stress concentration at the
wellbore. Particularly in highly plastic rocks, the tan-
2
As an example, consider the case of 3000-psi well- gential stress at the wellbore never becomes tensile. In
bore pressure in equilibrium with the pore pressure of this case, fracture could initiate in shear
the reservoir and values of 3500 psi for σx and 5000 psi (Papanastasiou et al., 1995).
for σy. The equations lead to maximum values for the
effective tangential stress (σθ – p) of 5500 psi in com-
pression (θ = 0°) and 500 psi in tension (θ = 90°). The 3-5.8. Stress change from hydraulic
latter result indicates the possibility for the occurrence fracturing
of tensile failure in a direction perpendicular to the min- Two effects are considered in this section. The first
imum stress, solely as a result of drilling the borehole. one addresses the increase of minimum stress because
A hydraulic fracture is induced by increasing the of the poroelastic effect. During the fracturing pro-
wellbore pressure pw up to the point where the effec- cess, fracturing fluid leaks into the formation. This
tive tangential stress (σθ – p) becomes equal to –To. leakage induces a pore pressure increase around the
If σx = σh, this happens at θ = 90° (where the stress fracture that results in dilation of the formation and,
concentration induced by the far-field state of stress therefore, an increase of the minimum stress in this
is minimum), which means that fracture initiates in region. For a 2D crack in an infinite sheet, the
a direction perpendicular to the minimum horizontal increase of minimum stress as a function of time
stress direction. Fracture initiation at the breakdown is (Detournay and Cheng, 1991)
pressure pif is, therefore, obtained when (Hubbert and
Willis, 1957) ( )
∆σ 3 = η p f − p f ( τ c ) , (3-71)
pif = 3σ h − σ H + To − p . (3-67) where pf is the fracturing fluid pressure and τc is a

characteristic time given by
2tkG(1 − ν)( νu − ν)
These induced stresses diminish rapidly to zero
away from the wellbore. Consequently, they affect the τc = , (3-72)
α 2 µ(1 − 2 ν) (1 − νu ) L2
2
pressure to induce a fracture, but not the propagation
of the fracture away from the wellbore wall. where G is the shear modulus, k is the permeability, t
If the wellbore fluid penetrates the formation, poro- is the time, µ is the viscosity, and L is the fracture
elastic effects must be taken into account to calculate half-length.
the stress concentration around the wellbore. In partic-

The function f varies between 0 and 1 as τc increas- It is also generally accepted that the degree of frac-
es from 0 to infinity. The value of the characteristic ture containment is determined mainly by the in-situ
time at which poroelastic effects can start to influence stress differences between layers, although, as men-
the state of stress around the fracture is about 10–3. In tioned in the previous section, other mechanisms can
reservoir rocks where η is equal to about 0.25, poro- play a role in the absence of a meaningful stress con-
elastic effects are important when the net pressure and trast. Knowledge of the stress variation between the
characteristic time are high. This effect decreases if a pay zone and the adjacent layers is therefore essential
high-pressure drop occurs at the fracture face. High- for predicting the extent that the fracture grows out of
pressure drops happen, for example, when a good the zone.
fluid-loss control agent is used. Poroelastic effects Fracturing from inclined or horizontal wellbores has
eventually disappear when injection is stopped and the also brought new requirements for determination of
excess pore pressure dissipates into the formation. the far-field in-situ stress because it controls the com-
The second effect is the stress increase caused by plex stress field that is generated near the wellbore.
opening of the fracture. This effect remains if the frac- The efficiency of a fracturing treatment is a function
ture is held open by proppant (Warpinski and of wellbore inclination with respect to the principal
Branagan, 1988; Palmer, 1993). An upper bound for stress direction and magnitude (Martins et al., 1992a;
the stress increase once injection has been stopped is Pearson et al., 1992).
the value of the net pressure at shut-in, but in practice, Finally, the amount of stress confinement influences
because the in-situ proppant concentration at shut-in is the rock properties (elasticity, strength and permeability).
about one-half of its compacted concentration, it could Several methods are regularly used in the petroleum
be about 50% of this value. This effect can be signifi- industry to estimate the magnitude and orientation of
cant for short and wide fractures, where crack-tip in-situ stress at depth. Some of them rely on field data
screenout induces a large net pressure (of the order of interpretations, whereas others rely on core measure-
1000 psi) at the end of the job. It has been proposed ments. Techniques based on wellbore images and
that this mechanism strengthens weak formations and, sonic logs are presented in Chapter 4.
therefore, decreases the risk of sanding.
3-6.2. Micro-hydraulic fracturing techniques

3-6. In-situ stress measurement Fracturing techniques are commonly used to measure
the minimum stress. The micro-hydraulic fracturing
3-6.1. Importance of stress measurement technique is certainly the most reliable technique if
in stimulation conducted properly, although it could be used in con-
The value of the minimum stress is one of the most junction with other methods for added completeness.
important parameters in hydraulic fracturing. At typi- This technique uses the pressure response obtained
cal reservoir depths, the fracturing pressure is a strong during initiation, propagation and closure of a
function of the minimum stress (or closure pressure). hydraulically induced fracture to accurately determine
With some pumping regimes, the value of the net the state of stress. Because stresses are functions of
pressure, which is the fracturing pressure minus the rock properties, it is quite important to ensure that the
closure pressure, could be quite small compared with test provides a measure that is representative of a
the closure pressure. The net pressure is the most given lithology. Small-scale hydraulic fractures are
robust and usually the only parameter that is available usually required, especially if the measurements will
for obtaining information on fracture geometry. An be correlated with log or core information. However,
error in closure stress measurement can lead to a sig- the fracture must be large compared with the wellbore
nificant error in the estimation of the net pressure and, radius to measure the far-field minimum stress com-
consequently, the fracture geometry. Because of the ponent, and a fracture with a size of 5–15 ft is a good
small value of the net pressure compared with the compromise. At this scale, a tool that includes a
minimum stress, knowledge of the in-situ state of gamma ray sonde is recommended for accurate place-
stress at depth also gives insight into the expected ment with regard to lithology. Analysis of the sonic
treatment pressures. and gamma ray logs should be made prior to testing

to decide on the location of the most appropriate

lithologies for the tests. It is recommended to select
locations that span lithologies with different values of
Poisson’s ratio and Young’s modulus if the objective
Wireline
of the measurement is to establish a complete stress
profile (see also Chapter 4).
To perform a micro-hydraulic fracture in an open-
hole, the selected test interval is isolated from the sur-
rounding well using a packer arrangement (Fig. 3-22). Gamma ray
sonde
Fluid is then injected in the interval at a constant flow
rate. During injection, the wellbore is pressurized up
to the initiation of a tensile fracture. Initiation is usu-
ally recognized by a breakdown on the pressure Pumpout module
versus time record, hence the name breakdown pres-
sure (Fig. 3-23). In practice, breakdown is not always Pressure gauge P
obtained. Initiation could also occur prior to break-
down. After the initial breakdown, injection should
continue until the pressure stabilizes. Injection is then Inflate seal
stopped and the pressure allowed to decay. The frac- valve
turing fluid is usually a low-viscosity fluid for low-
permeability zones or a drilling mud for zones with
higher ranges of permeability. Usually less than 100 Packer
gal is injected into the formation at flow rates ranging
from 0.25 to 25 gal/min. The amount of fluid and the
injection rate used during fluid injection are preferably
selected to achieve a predetermined fracture size at the
P
end of the test. This approach, however, requires the
3.25 ft (minimum)
use of a fracture propagation model to estimate the Interval seal
fracture geometry during propagation and closure. valve
Several injection/fall-off cycles are performed until
repeatable results are obtained (Fig. 3-24; Evans et al., Packer
1989). The most accurate stress measurements are
made using downhole shut-off devices, downhole P
pumps and downhole pressure gauges. A downhole Sliding
shut-off tool is used to shut in the straddle interval and coupling
minimize any wellbore storage effects (Warpinski et
al., 1985). This is required because careful monitoring
of the shut-in behavior is used to determine the mini-
mum stress. Downhole pumps have the advantage of
minimizing wellbore storage during pumping and
shut-in (Thiercelin et al., 1993). Low-storage tools
enable the effective control of fracture propagation
Figure 3-22. Wireline stress tool.
and analysis of the pressure response in great detail.
• Estimating minimum stress more may result when using this approach, espe-
Various techniques are used to estimate the magni- cially for zones that develop significant net pressure
tude of the least principal stress. The simplest one is or in a porous formation. Recently, techniques to
to take the instantaneous shut-in pressure (ISIP) as determine the closure pressure have replaced the
an approximation of the minimum stress (Fig. 3-23). ISIP as a measure of the minimum principal stress.
However, errors of the order of hundreds of psi or Conceptually, the closure pressure is the pressure at

which the fracture would close completely in the

Breakdown pressure absence of fracture face irregularities (e.g., ww = 0
in Eq. 3-36).
Instantaneous shut-in In the broadest sense, the techniques involve plot-
Bottomhole pressure
pressure ting the pressure decline after shut-in on specialized

plots that accentuate a slope change when closure
Closure stress
occurs. The pressure decline, after creating an
unpropped fracture, can exhibit different and identi-
Flow rate fiable behaviors, as discussed in Chapter 9. These
Pore pressure pressure behaviors are the result of various events
that include height growth closure, fracture exten-
sion after shut-in, fracture recession, transition
Time through closure pressure, consolidation of filter-
cake and face irregularities, reservoir linear flow
Figure 3-23. Downhole pressure during a micro-hydraulic and pseudoradial flow. It is generally difficult to
fracturing test.
analyze pressure decline because the transitional
pressure response through most of these various
Breakdown Reopen 1 Reopen 2 Reopen 3

250
Permeability
test
Downhole pressure (bar)
200
150
Wilkins
sand 100
712.5 m
50
0 5 10 15 20 25 30 35
10
Flow rate
(L/min)
0
Vin = 5.5 L Vin = 10 L Vin = 15 L Vin = 40 L
Vout = 3.7 L Vout = 6.5 L Vout = 5.5 L
200 Breakdown Reopen 1 Reopen 2 Reopen 3

Permeability
Downhole pressure (bar)
test
150
Wilkins 100
shale
724 m 50
0 5 10 15 20 25
10
Flow rate
(L/min)
0
Vin = 4 L Vin = 10 L Vin = 15 L Vin = 25 L
Vout = 2.6 L Vout = 2.8 L Vout = 2.6 L
Figure 3-24. Pressure and injection rate record obtained into a sand and immediately underlying shale (Evans et al., 1989).

behaviors can be featureless. Therefore, lacking a To overcome the ambiguity of decline analysis,
robust, comprehensive prediction model for these the pump-in/flowback (PI/FB) test was developed
behaviors, no combination of specialized plots pro- (Nolte, 1979). This test, illustrated in Fig. 3-26, pro-
vides a reliable tool for extracting closure pressure vides a robust, unique signature when correctly exe-
from decline data. The exception appears to be spe- cuted. The flowback period essentially involves
cialized plots for closure governed by reservoir- flowing fluid out of the fracture at a constant rate,
controlled fluid loss. These plots can provide usually between one-sixth and one-quarter of the
detectable slope changes during several of the tran- injection rate. The initial interpretation assumed that
sition periods. In permeable formations, where the closure occurred at the inflection point above the
fracturing fluid leaks off from the fracture face, clo- intersection point shown on Fig. 3-26b. Subse-
sure pressure is inferred when the pressure decline quently, Shlyapobersky et al. (1988b) suggested
deviates from a linear dependence on the square that closure occurred at the onset of the linear
root of shut-in time or the square root of the sum of response, which is below the intersection point
shut-in time and injection time ti (Fig. 3-25; Nolte, shown in Fig. 3-26b. Their basis was that linear
1982, 1988a). response corresponds to wellbore storage only (i.e.,
An improved representation of fluid loss is pro- when the fracture is closed). Plahn et al. (1997) pro-
vided by the G-plot, which is discussed in Section
9-5. Castillo (1987) introduced use of the G-plot
(a)
for closure inference along with specialized func-
tions for the pressure (i.e., linear pressure for wall Pump-in Flowback
cake-controlled fluid loss, square root for filtrate-
controlled fluid loss and logarithm if the resistance Bottomhole pressure
of the movement of the reservoir fluid controls Rate too low

fluid loss). Additional examples of the G-plot-
based method are given by Shlyapobersky et al. Correct rate
(1988b). Although the G-plot provides a firmer
foundation than the square-root plot, its derivation
does not consider the well-established additional Rate too high
fracture extension and recession after shut-in (see
Section 9-5). Time
(b)
Flowback Pressure rebound

Bottomhole pressure
Bottomhole pressure
Intersection of tangents
(closure pressure estimate)
Closure stress
es
liti
sibi
s
Po
Time
∆t or ∆t + ti Figure 3-26. Flowback and pressure rebound: (a) influence

of flowback rate on pressure response (Nolte, 1982, 1988a)
and (b) recommended approach for closure pressure esti-
Figure 3-25. Pressure decline analysis (Nolte, 1982, 1988a). mation (Plahn et al., 1997).

vided a study of the PI/FB test using a comprehen- exceed the closure pressure; however, for cases with
sive model that couples the wellbore, fracture meaningful fluid loss to the formation, the rebound
growth and recession, and the reservoir. They rec- pressure is generally below the closure pressure.
ommended the intersection of tangents illustrated For the preferred arrangement of straddle packers
on Fig. 3-26b and demonstrated that this was pre- and downhole shut-in, wellbore storage becomes
cisely the case for low fluid loss and an ideal fric- small and the second straight line approaches a ver-
tionless fluid; i.e., the curved parts of the pressure tical line. In low-permeability formations, the pres-
response during flowback result from pressure gra- sure rebound tends toward the closure pressure
dients that develop as a result of fluid flow within (Fig. 3-27) (Thiercelin et al., 1994). The observa-
the fracture either to the fracture tip or to the well- tion of pressure rebound is also a quality control
bore. test as it demonstrates that a fracture was created
The PI/FB test has another feature, the pressure without bypassing the packers.
rebound that is observed once the well has been A companion test to estimate the closure pressure
shut in to end the flowback period. During rebound, is the step rate test (Fig. 3-28; see the Appendix to
the fluid in the fracture flows into the well until Chapter 9). During a step rate test, the injection is
equalization between the well pressure and the pres- increased by steps up to the point that the pressure
sure within the fracture occurs. Nolte (1982) sug- response indicates that a fracture is extending (i.e.,
gested that the maximum pressure value of the the extension pressure). The indication is a slope
rebound was a lower bound for the closure pressure change, with the decrease in dp/dq reflecting
(i.e., the pressure equilibrated into the closed but
conductive fracture).
Plahn et al. (1997) also investigated the rebound Depressurization
of the PI/FB test. They found that generally most
Bottomhole pressure
Pressure rebound
of the fracture remained open during flowback and
Minimum
the initial rebound phase and that the characterizing stress
behavior resulted from pinching of the fracture
width at the wellbore because of the reversed flow.
After shut-in of the reversed flow, reopening of the
pinched fracture permitted pressure equalization
between the wellbore and the fracture (i.e., pressure
rebound). They concluded that even when the Time
flowback phase is continued to below the second
straight dashed line on Fig. 3-26b and is dominated Figure 3-27. Pressure rebound in a low-permeability
by wellbore storage, the rebound pressure can formation.
Bottomhole pressure
Extension pressure
Injection rate
at step end
Time Injection rate
Figure 3-28. Step rate test (Nolte, 1982, 1988a).

increased injectivity as the extending fracture is less accurate because it depends strongly on
exposes an increasing fluid-loss area and storage assumed rock behavior near the wellbore. In addi-
volume. Successful execution of the test ensures tion, the breakdown pressure magnitude is highly
that a fracture was created, which is a necessary dependent on the pressurization rate. For example,
quality control for permeable formations where the the initiation pressure depends upon whether the for-
pump-in/shut-in response for a nonfractured reser- mation behaves elastically or inelastically. Initiation
voir can be mistaken for fracture propagation and pressure depends also on the diffusion of fracturing
closure. Obviously, a clear indication that a fracture fluid into the formation, leading to a dependence on
was created is a necessary condition for closure the pressurization rate (Haimson and Zhongliang,
pressure determination. The extension pressure also 1991). However, bounds on the initiation pressure
provides an upper bound on the closure pressure to value can be obtained (Detournay and Cheng, 1992)
assist in planning and analyzing subsequent tests. using Eqs. 3-67 and 3-70.
The step rate and rebound tests are discussed again The subsequent analysis presented in this chapter
in Section 3-6.3. assumes that the material behaves in a linearly elas-
In low-permeability formations, Lee and Haimson tic manner and that the pressurization rate is low
(1989) proposed using statistical analysis proce- enough to use a long-time solution for the predic-
dures with an exponential pressure-time decay tion of initiation pressure. It is also assumed that the
method. wellbore is vertical, the overburden pressure is a
Modeling shows that the stress estimated by frac- principal stress, and the fracture initiates vertically.
turing can be in error by η(pf – p) because of poro- If the fluid is nonpenetrating, the initiation pressure
elastic effects (Detournay and Cheng, 1991), where is given by Eq. 3-67.
pf is the fracture propagation pressure. An estimation For the injection cycles that follow the first injec-
of the characteristic time of the process allows esti- tion cycle, the initiation pressure corresponds to
mating whether these poroelastic effects are impor- reopening the fracture, and so To in Eq. 3-67 is
tant. In practice, they are negligible in low-perme- effectively equal to zero. As σh was determined
ability formations (i.e., shales) but may become sig- from the closure, this formula can be used to esti-
nificant in millidarcy-permeability formations unless mate the intermediate stress σH.
a high flow rate and/or high-viscosity fluids are used Equation 3-70 applies when the fracture fluid dif-
(Boone et al., 1991) or fluid-loss control agents are fuses in the rock. This is the preferred equation for
added. predicting or interpreting the initiation pressure for
Finally, inversion of the pressure response (Piggott low-pressurization rates. For low-porosity rocks
et al., 1992) is probably the most powerful tech- (such as hard limestones), η = 0 and the value of
nique in situations where the fracture geometry is the initiation pressure is
not too complex and the fluid and rock are well
3σ h − σ H + To
characterized. This method uses a fracture propaga- pif = (3-73)
tion model to invert the pressure response obtained 2
during propagation and closure of the fracture. To if the fracturing fluid fully penetrates the micro-
be successful, the model must use assumptions on cracks. In particular, if σh + To < σH, the initiation
fracture propagation and closure that represent fairly pressure is less than the minimum stress. A break-
well the governing in-situ conditions. For example, down pressure may not appear, and care must be
radial geometry is usually appropriate for a micro- taken to differentiate between the initiation pressure
fracturing interpretation. The inversion of micro- and the maximum pressure. If η is equal to 0.5,
hydraulic fracturing data is, however, more complex Eq. 3-70 gives a prediction similar to Eq. 3-67. For
if conducted for an inclined or horizontal borehole, reservoir rocks, η is equal to about 0.25. It can be
in which case planar fractures are an exception seen from these equations that using a nonpenetrat-
rather than a rule. ing fluid increases the breakdown pressure, and if
• Estimating maximum horizontal stress the fluid penetrates preexisting microcracks, it is
easier to break nonporous formations (where
Attempts have also been made to determine the η = 0) than porous formations (for rocks exhibiting
maximum horizontal stress component from the the same tensile strength).
fracture initiation pressure. This stress measurement

However, the estimate of σH using Eq. 3-67 axes for the in-situ stresses. The recovered strains
or Eq. 3-70 could be quite poor, especially if the generally include an instantaneous, elastic part that
assumption of linear isotropic elasticity does not is impossible to detect (because it occurs as soon as
apply or the wellbore is not aligned with a principal the drill bit passes the particular depth) and a time-
stress direction. dependent, inelastic component. A typical ASR
curve is shown in Fig. 3-29.
The ASR method relies on strain measurements
3-6.3. Fracture calibration techniques made on cores retrieved from their downhole envi-
Whereas the micro-hydraulic fracturing method is ronment by conventional procedures. Therefore,
intended for measuring almost at a point in the forma- strains corresponding to the initial elastic recovery,
tion, large-scale fracture analysis requires characteriz- as well as part of the inelastic portion, are lost
ing the “average” stress over larger sections (e.g., 30 ft). because of the finite time it takes to bring the ori-
The average stress is termed closure pressure. ented core to the surface. The interpretation requires
For this purpose, fluid volumes greater than that for an assumption regarding the relationship existing
the microfracture are used to determine closure pres- between the time-dependent strain and the total
sure, especially when a precise stress profile is not strain. The suggestion of direct proportionality
available. Consequently, the fracture must be large made by Voight (1968) is typically employed. The
enough for making the measurement of closure pres- relative magnitude of strain recovery in the different
sure on a rock volume that is representative of the directions is used as an indication of the relative
complete zone. The methods used for determining the stress magnitudes, and the absolute magnitudes are
closure pressure with these larger fractures are similar related to the known overburden stress.
to those for micro-hydraulic fracturing. However, the • Differential strain curve analysis (DSCA)
closure pressure determination becomes more com- Differential strain curve analysis (DSCA) relies on
plex, with the possibility of multiple closures resulting strain relaxation as an imprint of the stress history
from stress variations within the reservoir. In particu- and considers the consequence of this relaxation
lar, shut-in decline tests could be quite difficult to (Siegfried and Simmons, 1978; Strickland and Ren,
interpret and a combination of the step rate and flow- 1980). This approach assumes that the density and
back tests is recommended. The procedures also dif- distribution of the resulting microfracturing are
fer slightly because of the larger amounts of fluid directly proportional to the stress reduction the core
involved and the resulting higher net pressure. With
these larger fractures, an apparent breakdown pres-
sure could be observed but must not be interpreted for
maximum horizontal stress determination, even in an
openhole situation. One variation presented by Wright ε0
et al. (1995) consists of a multiple, discrete injection
t0 = Sample cored
during the shut-in decline phase. Large-scale tests are t1 = Sample strain gauged
also discussed in Sidebar 9A. ε0 – ε1 t2 = End of measurements
Strain
ε0 – ε1 = Component of elastic strain relaxation

ε1 – ε2 = Component of anelastic strain relaxation
3-6.4. Laboratory techniques

• Anelastic strain recovery ε1
The anelastic strain recovery (ASR) method requires ε1 – ε2

access to oriented core samples. The method is
based on the relaxation that a rock core undergoes ε2
following its physical detachment from the stressed
rock mass (Teufel, 1983). The recovered strains are
measured in various directions, and the principal t0 t1 t2 Time
strain axes are determined. These principal direc-
tions are assumed to be the same as the principal Figure 3-29. ASR curve.

sustained. Hence, if the microcrack-distribution core. However, as can be seen from Fig. 3-31, stan-
ellipsoid could be delineated, it may reveal the pre- dard tests are run using twice as many strain
existing stress condition. The existence of pervad- gauges. This duplication allows several combina-
ing microdiscontinuities plays an important role in tions of solutions; hence, statistical data analysis
the beginning of the loading cycle, as it introduces methods can be used. The resulting standard devia-
a “softening” element. Consequently, accurate strain tion is a good measure of the confidence possible in
measurements in various directions should make the results.
possible anisotropic characterization of the micro- Although ASR and DSCA seem to be based on
cracking, which can be related to measurement of the same fundamental phenomenon, differences
the preexisting stress state. A typical DSCA curve may occur in the results. This is usually the case
is shown in Fig. 3-30. when a particular rock formation has been subjected
The results of DSCA yield the orientation of the during its geologic history to a stress field large
three principal stresses. The stress magnitudes can enough to induce a microcracking pattern that over-
only be indirectly calculated by knowing the elastic shadows the one resulting from the present unload-
constants and the overburden as for ASR. To deter- ing. DSCA reflects the sum of whatever happened
mine the in-situ stress state, only six strain measure- in the stress history of the rock, whereas ASR is
ments are theoretically required on any oriented limited to its present state of stress.
13
Curves correspond to specific
gauges on cubic sample
11
9
Strain (0.01%)
1 3 5 7 9 11 13 15 17
Pressure (1000 psi)
Figure 3-30. DSCA plots.
N
(East)
y
10
(North) 2
x 3
1 45˚ 4
9
0
12 5 W E
8
11
7 6
z
S
Figure 3-31. Gauge pattern and typical results plotted on a polar stereonet for DSCA.


◆
Well Logs
Jean Desroches, Schlumberger Dowell
Tom Bratton, Schlumberger Sugar Land Product Center
4-1. Introduction tion. Sections 4-4 and 4-5, respectively, discuss the
various means of determining these properties.
The purpose of this chapter is to describe, strictly in All that is then necessary is to transform the data
the context of reservoir stimulation, the use of geo- gathered into a consistent model. The first step for
physical information (commonly referred to as logs) this process is to fill any data gaps by relying on
to obtain a description of the formation affected by geologic information. Each depositional basin has its
a stimulation treatment. The entire process of classic own “style,” characterized by unique properties.
formation evaluation for the determination of hydro- Moreover, the succession of layers is not arbitrary,
carbon reserves is out of the scope of this volume. but obeys worldwide a logic that is captured by the
A fundamental difference between these two pro- notion of sequence (e.g., sand/shale sequence, lime-
cesses is that all properties are required not only for stone/marl sequence) (Wagoner et al., 1930; Fried-
the hydrocarbon-bearing formations (pay zones) but man and Sanders, 1978). Combining information
also for the adjacent formations (bounding layers). characterizing the basin with information from the
The output of the process described in this chapter relevant sequence makes it possible to estimate miss-
is a model of the formation to be stimulated. The ing data or interpolate between data points with an
model consists of a series of planar, parallel layers educated guess or correlation. This process is partic-
or beds, with known properties for each layer. Each ularly crucial for the determination of lithology, pore
property in a layer is either a constant, averaged pressure and stress profiles. The need for a model
property or a linear function of depth (e.g., pore based on a basin and field perspective is emphasized
pressure) (Fig. 4-1). throughout this chapter. This approach also clearly
To construct the model, estimates first must be indicates the need for bringing geological expertise
obtained of the relevant properties in each layer. into the picture.
These properties are in two broad classes: properties Once a complete description versus depth is
relevant to the diffusion of fluid in the formation and achieved, the boundaries of the layers to be consid-
properties relevant to the deformation of the forma- ered are defined (i.e., zoning), which is also guided
by geologic information. Section 4-6 discusses the
Layer 1
zoning process.
Layer 2
An important aspect stressed throughout this chap-
True vertical depth, zTVD
ter is the calibration of geophysical data. Apart from

Layer 3
the most basic properties, such as porosity and water
Layer 4 saturation, geophysical information requires calibra-
Layer 5 tion with laboratory data. Again, the notion of basin
and sequence is useful: the calibration procedure is
Layer 6 usually portable within the same sequence in the
same basin. In other words, if no core samples have
Profile of a Profile of Profile of pore been tested for a particular well, results from other
property with horizontal pressure—
a constant stress—linearly linearly varying cores taken in the same formation and basin may be
value within varying within within each used, albeit with some caution.
each layer each layer layer
(e.g., porosity) In this chapter, it is assumed that the geophysical
information comes from the well to be stimulated.
Figure 4-1. Model of the earth in the vicinity of the forma- The porting of geophysical information from one
tion to be stimulated.
well to another is not addressed.

4-2. Depth where θ is the deviation of the well from the vertical
at that point. Integration of Eq. 4-1 along the well-
The several types of “depths” used in the oil field bore trajectory allows the conversion between MD
impact the building of the final representation of the and TVD.
reservoir to be stimulated. True vertical depth (TVD) In the context of this chapter, formation beds have
is, as its name indicates, the depth measured along a constant thickness, but their boundaries may make
the vertical. Given a zero reference point, it is unique. an angle δ with the horizontal. The angle δ is called
Measured depth (MD) is the distance measured along the dip of the formation. It is convenient to present
the wellbore trajectory (which is never exactly ver- logs versus true bed thickness (TBT) as an imaginary
tical). It is not unique, because it depends on the way line perpendicular to the bed boundaries (Fig. 4-2).
it is measured. MD during drilling is measured by the This presentation is useful for comparing logs
length of pipe that has gone into the ground. MD at acquired in deviated wells in a dipping reservoir. An
wireline logging time is measured by the length of increase of the true bed thickness ∆TBT is related to
cable that has gone into the ground. Finally, logging- an increase of measured depth by
while-drilling (LWD) data are logged versus time,
which complicates the issue because several measure-
∆TBT =
[
∆MD cos(θ + δ ) ]. (4-2)
ments can be assigned to the same depth. cos δ
Special care must be taken to ensure that all log
measurements are referenced to a single measured Layer thicknesses, therefore, relate to properties in
depth. The reference depth is usually from a gamma TBT. For horizontal beds, TVD and TBT coincide. A
ray log or that of an imaging tool for complex reser- schematic of the effect of a deviated well path on the
voirs. However, the conversion of MD to TVD must presentation of a log in MD, TVD and TBT is pre-
be reliable because MD is required to compute the sented in Fig. 4-2.
fluid friction and fluid displacement, but TVD is
required to design the treatment placement. For
example, fracture height is related to properties in 4-3. Temperature
TVD, not in MD. At any point along the wellbore The temperature of the formation is critical for the
trajectory, an increase of the true vertical depth ∆TVD performance of both matrix stimulation products and
is related to an increase of measured depth ∆MD by hydraulic fracturing fluids.
∆TVD = ∆MD cos θ, (4-1) The mud temperature acquired at wireline logging
time is typically used for an estimate of the formation
δ
θ
Layer 1
We
ll tr
aje
Layer 2
cto
ry
Layer 3
MD
Original log versus MD TVD TBT Log versus TVD Log versus TBT
Figure 4-2. The effect of wellbore deviation and dipping beds on the presentation of logs.
4-2 Formation Characterization: Well Logs

temperature. It yields only a lower bound of the for- space present within sedimentary particles at the
mation temperature and can underestimate the forma- time of deposition.
tion temperature by as much as 30°F [15°C]. Wire- • Secondary porosity consists of the space that was
line temperature is, however, the only continuous created by tectonic forces creating microcracks and
temperature measurement. water dissolution creating cavities.
Discrete point measurements of temperature can be
obtained during fluid sampling with formation testers. The distinction between primary and secondary
The temperature of the sampled fluid is continuously porosity is important insofar that porosity is used in
monitored during sampling to yield the most accurate many correlations to develop a first estimate of other
temperature measurement currently available. properties (e.g., rock strength or permeability). The
Another good estimate of reservoir temperature is correlations rely mostly on the primary porosity, not
the bottomhole temperature recorded by a slickline- the secondary porosity.
conveyed gauge after perforation and initial flow of Another important distinction is that of the total
the formation. porosity φtotal and effective porosity φeff. The total
porosity is the volume not occupied by solid rock.
However, part of the volume of total porosity is occu-
pied by fluid that cannot move (i.e., bound water).
4-4. Properties related to the The effective porosity is the volume occupied by
diffusion of fluids moveable fluids, and it is the porosity of interest for
The diffusion of fluids is governed by porosity, per- most oilfield applications. A notable exception is the
meability, pore pressure and the fluid types in the use of φtotal for all reservoir calculations involving
formation. This section introduces means for deter- transient flow.
mining these parameters from logs. Lithology and A final concern is that no open- or cased hole log
saturation are discussed because that information measures porosity directly. Rather, a property related
is used to infer permeability, but fluid viscosity and to porosity is what is measured. This is why a combi-
compressibility cannot be obtained from logs. nation of porosity measurements is preferred for esti-
mating φeff. Which measurements are used varies
depending on the lithologic sequence and possibly
4-4.1. Porosity the geologic province.
Porosity is the fraction of the total formation volume • Porosity from density
that is not occupied by solid rock (i.e., filled with for- Density tools measure the electron density of a for-
mation fluid). The common symbol for porosity is φ mation, which is extremely close to its bulk den-
and the measurement unit is either the volume frac- sity ρb (Tittman and Wahl, 1965). Density is a
tion (range from 0 to 1) or percent (from 0 to 100). shallow measurement (i.e., the depth of investiga-
The porosity of formations can vary from nearly zero tion is typically 0.5 ft), and the volume of rock
for evaporites to up to 40% for unconsolidated for- sampled is usually within the flushed zone (i.e., the
mations (e.g., shales or sandstones) and even higher zone in the vicinity of the wellbore where the for-
for chalk or diatomite. mation fluids have been displaced by mud filtrate).
Porosity is a cornerstone of formation analysis. It If the density of the matrix components ρma and
is measured through several geophysical methods, for that of the pore fluid ρf are known, the total poros-
which the principles and assumptions of each are pre- ity from density can be found by volume balance:
sented. The conveyance method of the measuring ρ − ρb
tool is irrelevant: the principles apply to both wireline φ D = ma , (4-3)
ρma − ρ f
and LWD tools. The subsequent processing of the
data is not addressed in detail in this chapter (see where ρma is determined from the lithology and
Quirein et al., 1986). ρf is taken as that of the mud filtrate, which is
Porosity is classically divided into two groups: obtained from charts as a function of temperature,
• Primary porosity consists of the original space pressure and salinity. If there is gas present, φD
between the grains that form the rock matrix or the overestimates the total porosity.

• Porosity from neutron • Porosity from nuclear magnetic resonance

Neutron tools measure an index of how much Nuclear magnetic resonance (NMR) tools measure
hydrogen is present in the formation (Allen et the relaxation time of protons. The porosity mea-
al., 1967). It is also a shallow measurement (i.e., sured by NMR tools is similar to φN but is influ-
the depth of investigation is typically 1 ft). If no enced primarily by moveable fluids.
hydrogen is in the rock matrix and the hydrogen Extracting porosity from NMR measurements
index of the fluid is known, the neutron porosity requires complex processing. However, NMR-
φN obtains a measure of the total porosity. measured porosity has two major advantages.
Water and liquid hydrocarbons have similar First, because it is influenced primarily by move-
hydrogen indices; therefore, φN is insensitive to the able fluids, it is extremely close to φeff, and second,
presence of oil in the sampled volume. However, if it yields excellent estimates of porosity in shaly
gas is present, φN underestimates the total porosity. (i.e., clay-bearing) formations, which are typically
The deviation of φN and that of φD from the total challenging for estimating porosity (Minh et al.,
porosity tend to be in opposite directions. One 1998).
example, mentioned previously, is their opposite • Porosity from resistivity
behavior in the presence of gas. A simple average
Porosity can also be estimated from resistivity
of φN and φD yields a good estimate of the effective
measurements. If the resistivity of the rock matrix
porosity:
is assumed to be infinite compared with that of the
(φ N + φ D ) .
1 fluid, the conductivity of the formation is propor-
φ eff ≅ (4-4)
2 tional to the porosity.
The formation factor F is introduced as a ratio:
• Porosity from sonic
Ro
The presence of pore fluid increases the rigidity F= , (4-6)
of the rock over the case where porosity is truly Rw
empty (i.e., “moon dry samples”). Moreover, most where Ro is the resistivity of the formation 100%
of the attenuation of the sonic waves in a porous saturated with brine of resistivity Rw.
formation comes from the fluid (Biot, 1956b, Archie (1942) postulated that the formation fac-
1956c). If the elastic properties of the solid and the tor is related to the total porosity by the relation
fluid are known, and therefore the traveltimes in
a
these media, an estimate of the total porosity can F= , (4-7)
be computed: φm
∆t − ∆tma where a and m are constants depending on the type
φS = A , (4-5)
∆t f − ∆tma of formation. For example, a = 0.62 and m = 2.15
for clean sandstones.
where A is a constant and ∆t denotes the measured Therefore, if Rw is known (e.g., from water “cat-
transit time of a sonic wave in the formation. The alogs,” samples or spontaneous potential [SP]
transit time in the matrix ∆tma is known from the measurements), an estimate of the total porosity of
lithology. Because sonic is a shallow measurement the formation can be obtained by equating Eqs. 4-6
(i.e., the depth of investigation is typically 1 ft), and 4-7:
the fluid in the pore space is usually approximated 1m
by the mud filtrate, the properties of which are  aR 
φ= w . (4-8)
known functions of the temperature, pressure and  Ro 
salinity. However, the sonic porosity φS is a strong
This technique is not recommended, because it is
function of the properties of the fluid in the pore
greatly affected by the fluid saturation and conduc-
space. Therefore, if the fluid in the sampled pore
tive minerals in the matrix.
space contains hydrocarbons, φS can deviate signif-
icantly from the total porosity. • Final estimate of porosity
The value of φS is of particular interest because As mentioned previously, the best estimate of
it is sensitive primarily to the primary porosity, not porosity is obtained from a combination of logs,
to the secondary porosity. using synergistic processing that accounts for the

response of each tool and is tailored to the geo- where A, m and n are constants that are functions
logic environment. If such a processed porosity of the formation.
estimate is not available, the simple average of Because resistivity is usually measured at differ-
φD and φN in Eq. 4-4 is a good first-order estimate ent depths of investigation, it is possible to mea-
of φeff in noncomplex environments, and the poros- sure both the resistivity of the flushed zone Rxo,
ity from NMR yields a good estimate in more where all movable fluid has been displaced from
complex environments. the formation by mud filtrate, and the resistivity of
the virgin formation Rt, far from the wellbore. The
following expression is thus useful because it is
4-4.2. Lithology and saturation independent of the porosity:
α
Lithology and saturation are primary inputs for  R R 
designing matrix acidizing treatments. Lithology is Sw =  xo t  , (4-10)
also part of the more general information that must be  Rmf Rw 
considered to build a stress model (see Section 4-5.2). where Rmf is the known resistivity of the mud fil-
For hydraulic fracturing, saturation is used to esti- trate and α is typically assigned a value of 5⁄8. Rw
mate the compressibility of the formation fluid for can be determined from water catalogs, samples
computing the compressibility-controlled leakoff. or SP measurements.
If conductive minerals (e.g., clay minerals) are
• Saturation
in the formation, their effect can be accounted for
Water saturation Sw is the fraction of the pore vol- by several weighted-average techniques, especially
ume occupied by water. By definition, 1 – Sw is the for shaly formations (Poupon et al., 1970). For
fraction of the pore volume occupied by hydrocar- complex cases, the best estimate of Sw is deter-
bons. Also of interest is the irreducible water satu- mined by lithological analysis, as discussed in the
ration Swi, which is the fraction of the pore volume following.
occupied by water bound to the formation.
• Lithology
Special attention is devoted in formation evalua-
tion to both porosity and saturation because their The goal of a lithological analysis is to obtain a
product defines the volume of hydrocarbons in the volumetric distribution of the minerals and fluids
reservoir. Similarly for reservoir stimulation, these in the formation as a function of depth. The concept
are important quantities because they indirectly is first explained in this section with the example
govern the flow of water-base fluids in a porous analysis of a sand/shale sequence. The concept
medium filled with both water and hydrocarbons. is then generalized to any kind of formation.
The relative permeability to water is linked to the Historically, much work was devoted to sand/
saturation (see Section 4-4.3). The total compress- shale sequences (e.g., 40 papers in Shaly Sands,
ibility of the formation fluids, which is used for the 1982). The simplest form of lithological analysis
determination of leakoff during hydraulic fracturing in such formations is a shaliness indicator—i.e., an
treatments, is also computed using the saturation. indicator of the volumetric ratio of clay minerals to
The value of Sw is obtained mainly through clean sand.
resistivity measurements. In a clean formation The crudest shaliness indicator is based on
(i.e., a formation without shales or other conduc- gamma ray measurement. If all natural radio-
tive minerals), all the conductivity of the formation activity in a sand/shale sequence is assumed to
is associated with the brine in the pore space. Con- come from clay minerals, high gamma ray values
ductivity is the reciprocal of resistivity. (>90 API units) indicate a shale (i.e., almost exclu-
For the true resistivity of the formation Rt, sively clay minerals), low gamma ray values
which is beyond the disturbances associated with (<20 API units) indicate a clean sandstone, and
the wellbore, Archie (1942) derived experimen- values in between indicate a mixture of clay and
tally that sand.
This scale can be further refined by introducing
Rw A
Rt = , (4-9) a simple response equation of a gamma ray tool.
φ m Swn For example,

GRAPI = 15[K 2 O]. (4-11)
This linear equation relates quantitatively the

gamma ray count to the concentration of potas-
sium oxide (K2O) in percent. For a simple miner-
alogical model, such as quartz and a potassium-
bearing clay mineral, the ratio of quartz to clay can
be determined with the help of this equation.
Tools based on different physical principles are
sensitive to different combinations of atoms or
minerals. Each tool response TR can be character-
ized by an equation, not necessarily as simple as
Eq. 4-11, but which can always be cast in a form
involving the volumetric fraction of n minerals and
at least one fluid volumetric fraction Vf:
(
TR = f Vmineral
1 2
, Vmineral n
, . . ., Vmineral )
, Vf , (4-12)
i
where VmineraI is the volumetric fraction of mineral
i. The equations can be linear or nonlinear. Each
tool output yields one equation similar to Eq. 4-12
with the volume fractions of n minerals and the
formation fluids to form a system of usually non-
linear equations solved on a depth level by depth
level basis (usually every 6 in.) for the various
volume fractions (Quirein et al., 1986).
Obviously, assumptions must be made about
which minerals are in the formation. First, more
minerals than there are equations cannot be solved
for, so the smaller the number of logs, the more
limited the number of minerals. Second, no unique
set of minerals explains the response of a set of
tools to a formation. Several sets of minerals can
equally well fit the set of logs with very different
results. A mineralogical model for the formation
must be identified before lithological analysis is
conducted. The mineralogical model must be vali-
dated by a geologist familiar with the area before
any credit may be given to the lithological output.
Although a lithological analysis output (see the
example in Fig. 4-3) is an extremely useful tool for
building a pressure profile or stress model, it con-
sists of only volumetric outputs. The analysis out-
put cannot be interpreted without extreme caution
for mechanical properties, zones that are hydrauli-
cally connected, etc.
Figure 4-3. Example of lithological analysis.

4-4.3. Permeability
1 1
Permeability is a measure of the ease with which flu-
ids can flow through a formation. The S.I. permeabil-
ity unit is m2, but the customary oilfield unit is the End-point kro
millidarcy (md).
Permeability is a tensor, and its value depends on
the orientation of the flow. In the absence of exten-
krw
kro
End-point krw
sive natural fractures or fissures or large horizontal
stress anisotropy, the permeability kH parallel to the
bedding of the formation, which is horizontal in the
model of the earth assumed here, can be considered
isotropic, and the horizontal flow has no preferred
direction. In the presence of natural fractures or large 0
horizontal stress anisotropy, however, permeability 0 Swi 1 – Sor 1
can vary horizontally and the horizontal flow will Water saturation, Sw
have a preferred direction. The permeability perpen-
dicular (i.e., vertical) to the bedding kV is usually at Figure 4-4. Typical variation of relative permeability as
least 1 order of magnitude smaller for sandstone a function of water saturation (Dake, 1982). The curves
simply and practically describe the simultaneous flow of
reservoirs. For laminated sandstone reservoirs, it can oil and water through a porous medium. Sor = residual oil
be more than 2 orders of magnitude smaller. In some saturation.
carbonate reservoirs, however, kV can be equal to or
larger than kH. If no precision is given, the permeabil-
kro =
(1 − S )
w
a
, (4-16)
ity of interest is kH. For specific problems, kH and kV
and even the horizontal components of kH may be
(1 − S )
wi
b
required. For vertical wells, kH is sufficient to where m is typically 3, whereas a and b must be
describe practically all production or injection phe- determined by calibration (e.g., on core data). The
nomena. For horizontal or multibranched wells, any permeabilities to water and to oil are then calculated
contrast in kH will also have an impact on production simply by multiplying the total permeability k by the
or injection. relative permeabilities of the sequence of interest to
When several fluids are present, it is customary to water and oil, respectively.
introduce relative permeabilities. If k is the absolute The irreducible water saturation Swi can be obtained
(or “total”) permeability of the formation for a single for hydrocarbon-bearing formations either from labo-
fluid, the relative permeability to water krw is ratory measurements on cores or NMR measurements
(The Log Analyst, 1996, 1997); it is assigned a value
kw
krw = , (4-13) of 1 in non-hydrocarbon-bearing formations.
k In the context of reservoir stimulation, fluid dis-
where kw is the effective permeability of the forma- placement is usually treated like piston displacement.
tion to water. Similarly, the relative permeability to Therefore, the so-called end-point permeabilities are
oil kro relates the effective permeability to oil ko to the important (i.e., permeability of the formation to oil at
total permeability k by irreducible water saturation and permeability of the
ko formation to water at maximum water saturation).
kro = . (4-14) The end-point permeabilities can be very different.
k
For candidate selection and subsequent treatment
Relative permeabilities typically depend on the water evaluation, the permeability of interest is that of the
saturation of the formation Sw (Fig. 4-4) (Amyx et al., formation to native hydrocarbons. For the design and
1960). For example, the Park-Jones relations yield execution of a stimulation treatment, however, inter-
m est is in the permeability of the formation to water
 S − Swi 
krw =  w  (4-15) injection because water-base fluids are usually
 1 − Swi  injected into the formation.

Permeability to the native mobile reservoir fluids

4A. Permeability-porosity correlations
can be estimated either by indirect or direct measure-
ments. Direct measurements (formation testing) pro- Three well-know relations of permeability and porosity are
vide discrete measurements of permeability. Indirect 3
k i  φi 
measurements provide a continuous log of permeabil- Labrid (1975) =  (4A-1)
k  φ
ity that must be calibrated with the direct individual
measurements (e.g., on cores or through formation ki 
= exp 7.5 ×
(φ i − φ) 
Lund and Fogler (1976)  (4A-2)
testing). k  0.08 
• Indirect measurements
Lambert (1981)
ki
k
[ ]
= exp 45.7 × (φ i − φ) . (4A-3)
The following indirect measurements related to
permeability require calibration, usually on cores, These equations correlate a change in permeability to a
change in porosity in sandstones. For an assumed initial for-
to a water zone. However, they are the best means mation porosity φi of 0.2 and permeability ki of 20 md, if the
for interpolating permeability between direct per- porosity φ after an acid job is 0.28, then the three correla-
tions give values for the new permeability k of 55, 36,000
meability measurements. and 774 md, respectively.
– Correlation to porosity and water saturation Clearly, proper choice and calibration of a permeability-
porosity correlation for the sequence of interest are necessary.
Porosity is an obvious quantity to relate to per-
meability because interconnected porosity is
required to have permeability. Several equations where the constants C and m again depend on
that relate permeability to porosity (e.g., Car- the type of formation. For lower permeability
men, 1983; Wylie and Rose, 1950; Timur, 1968) formations, corrections must be made to obtain
are of the following form: a first-order estimate of permeability. The k-
lambda model in Eq. 4-18, therefore, is best
φx
k=C , (4-17) used for high-permeability formations.
Swiy Λ is a length scale that is difficult to measure.
where C is a constant that is a function of the However, it is inversely proportional to the sur-
rock type and grain size of the granular material. face area of the pores S divided by the volume
The exponents x and y are a function of the of the pores V, which is a ratio measured by
sequence of interest, and porosity estimation NMR tools. This makes k-lambda permeability
is as covered in Section 4-4.1. a good method for estimating permeability from
It must be emphasized that calibration is NMR measurements in high-permeability for-
absolutely necessary and that the calibration is mations (Herron et al., 1998).
specific to both the sequence and the lithology. Also, the values of both S and V, and there-
Once expressions have been calibrated for a par- fore Λ, can be measured for single minerals. If a
ticular sequence, however, they also provide mineralogical analysis of the formation has been
good continuous estimates of permeability made (see “Lithology” in Section 4-4.2), Λ can
within the field. Sidebar 4A is a practical exam- be estimated using a volumetric average of Λ for
ple that illustrates the caution necessary when each constituent mineral. The better the quality
using a permeability-porosity correlation. of the lithology estimate, the better the resulting
permeability estimate. In particular, this
– Permeability from lithology and porosity approach performs best when geochemical log-
Permeability is governed by a length scale Λ ging is used to determine the lithology (Herron
that corresponds to the size of the dynamically et al., 1998).
connected pores (i.e., pores that permit fluid
flow) (Herron et al., 1998). – Permeability from the Stoneley wave
For high-permeability formations (i.e., >100 md), When a pressure pulse is emitted within a well-
a good estimate of permeability is provided by bore, a guided wave called a Stoneley wave is
the k-lambda model: readily propagated along the wellbore. The wave
travels along the wellbore and pushes fluid
k = CΛ2 φ m , (4-18) through the mudcake into the formation. As the
fluid is mobilized in the formation, it alters the

attenuation and frequency response of the for- • Direct measurements

mation. This effect was modeled by Biot – Formation testers
(1956b, 1956c). If the mudcake is considered an
To measure permeability, formation testers
elastic membrane, an estimate of the permeabil-
hydraulically isolate the part of the formation to
ity of the formation to water can be computed
be tested. During drawdown, fluid is withdrawn
(Brie et al., 1998).
from the formation at a controlled rate. Pressure
The formation must have a minimum of 12%
is then allowed to rebound to an equilibrium
porosity and low to medium permeability (1 to
value (i.e., buildup). Both the drawdown and
100 md) for the best results. This is a shallow
buildup are analyzed to estimate the permeabil-
permeability measurement, typically limited to
ity of the formation.
the invaded zone. If gas is in the invaded zone,
Formation testers can be divided into two
the results are erratic.
groups on the basis of the resulting permeability
Although permeability from the Stoneley
measurement: tools with a single probe and
wave is an indirect measurement, it is closely
tools with multiple probes or a packer and probe
related to a true permeability measurement.
assembly (Fig. 4-5).
Again, calibration is important because the mea-
Single-probe tools press a probe against the
surement is sensitive to the elastic properties of
wellbore wall to achieve a hydraulic seal. A
the mudcake, which are unknown.
small amount of fluid, usually mud filtrate, is
Probe Tools Packer and Probe Tools
8 ft
6.6 ft
2.3 ft
> 3 ft
Sampling Sampling Sampling Sampling

Sampling VIT VIT VIT In-situ stress VIT VIT
kH2 kV, kH kH, kV kH, kV, φct kH, kV kH, (kV) kH, kV kH, kV
Figure 4-5. Probe tools and packer and probe tools. VIT = vertical interference testing.

withdrawn during drawdown. The steady-state of both kH and kV (Zimmerman et al., 1990; Pop
pressure drop at the sink probe ∆p is given by et al., 1993).
In the packer and probe configuration, a strad-
µq C  k 
∆p = × K  1 – V  for kV < kH dle packer is used to isolate part of the forma-
2 πkH rp  kH  tion to withdraw what can be large volumes of
(4-19) fluids. The response of the formation is moni-
tored by a probe several feet above the packer.
µq C
∆p = × for kV = kH This configuration also enables the determina-
4 kH rp tion of both kH and kV, especially in high-perme-
ability formations that require a large-volume
(4-20)
withdrawal to create a pressure response mea-
µq C  k  surable at the probe.
∆p = × K 1 − H  for kV > kH ,
2 π kV kH rp  kV  The obvious advantage of the multiple-probe
tools and packer and probe tools is that they pro-
(4-21) duce kH and kV estimates. However, the radius of
where C is a shape factor accounting for the cur- investigation into the formation is also much
vature of the wellbore, µ is the fluid viscosity larger than with a single-probe tool (i.e., at least
(usually that of the mud filtrate), q is the rate 3 times the spacing between the probes), which
of fluid withdrawal, rp is the radius of the probe, departs from a discrete point measurement. In
and K(η) is the complete elliptical integral of the particular, the viscosity and compressibility
first kind of modulus η (Wilkinson and Ham- terms in the equations governing the flow corre-
mond, 1990). K(η) can be approximated by spond to those of the formation fluid. These
π/2(1 + 1/4η2 + η4). The expression for the extra parameters must be determined to estimate
isotropic case kH = kV is typically used to the permeability. Thus, the design of formation
compute an estimate of kH. tests using multiple-probe and packer and probe
Once drawdown is stopped, the pressure tools is critical for obtaining representative data
bounces back to the formation pressure (see (Goode et al., 1991).
Section 4-4.2). In addition to µ, the expression – Well tests
governing the pressure behavior now also
The same procedure as that used for formation
includes the formation porosity φ and the total
testing is used during well testing. The well is
compressibility ct of the fluid in the formation.
flowed at a constant rate during drawdown
Drawdown and buildup permeability esti-
before it is shut in and pressure buildup
mates are usually different. For small drawdown
observed (see Chapter 2).
volumes (i.e., of the order of 20 cm3), use of the
Well testing has a much larger volume of
drawdown portion of the test is preferred for
investigation and produces an estimate of a com-
estimating the permeability of the formation.
posite kH. This composite value is crucial for the
Then, a measure of the horizontal permeability
economic evaluation of stimulation treatments.
kH to water can be obtained.
However, for matrix stimulation, especially fluid
Permeability tests conducted with single-
placement and diversion, knowledge of the per-
probe tools are commonly performed because
meability within each layer is essential, hence
they are simple, quick and reliable. Moreover,
the need for continuous measurements of perme-
these tests provide a good relative indicator of
ability by the methods outlined here.
formation permeability. The volume of investi-
gation is of the order of a few feet.
In the multiple-probe configuration of a for-
mation tester tool, several probes are used, sepa- 4-4.4. Pore pressure
rated by several meters. A pressure pulse is sent Pore pressure is the pressure of the fluid in the forma-
through one probe, and the response of the for- tion. After production, its value can differ signifi-
mation is monitored at the location of the other cantly from one layer to the next within a sequence.
probes. This method enables the determination

It is a necessary input for designing matrix stimula- • Pressure profiles

tion treatments of multiple layers, if only to account Because pore pressure is a point measurement,
for and potentially take measures to control crossflow there are gaps between the measurements. The
between zones. The pore pressure also strongly influ- gaps can be filled by building pressure profiles.
ences the state of stress in a formation (by as much as If there are several fluids in the formation (e.g.,
50%) and is therefore a critical piece of information a mixture of water, oil and gas), only one of them
for designing hydraulic fracturing treatments. is mobile and continuous across the pore space.
• Pore pressure measurement This fluid is called the continuous fluid. Under
static conditions (i.e., no flow), pressure changes
In addition to being measured by well tests, pore
as a function of depth result from gravity only. In
pressure is measured by formation testers. Once
other words, the pressure changes are a function of
a seal has been achieved between the formation
the density of the continuous fluid in the formation:
tester and the formation, fluid is withdrawn from
the formation to lower the pressure in the tester to dp
= ρ f g, (4-22)
below the far-field pore pressure. Flow is then dzTVD
stopped, and if the formation has sufficient perme-
where p is the pore pressure, zTVD is the true verti-
ability, the formation will flow to equilibrate the
cal depth, ρf is the density of the continuous fluid,
tester with the bulk of the formation. The pressure
and g is the acceleration of gravity.
in the tester will rise until a plateau is reached. The
A normally pressured reservoir consists of for-
pressure corresponding to the plateau is taken as
mations where p is close to a reference pressure
an estimate of the pore pressure.
given by a continuous column of brine up to the
In low-permeability formations (<1 md), two
surface, which corresponds to a pressure gradient
effects can have an impact on the pressure mea-
between 0.43 and 0.45 psi/ft in oilfield units:
surement. First, the formation may not supply
zTVD
enough fluid in a short enough time to equilibrate
with the tester. In this case, the measurement may p( zTVD ) ≈ ∫ρ brine dz , (4-23)
0
be aborted before a meaningful plateau is reached.
Second, even if a plateau is reached, it may indi- where ρbrine is the density of the brine. An over-
cate equilibrium with the higher local pore pres- pressured reservoir consists of formations where p
sure induced by overbalance drilling (Fig. 4-6). is much larger than the reference pressure, and an
This effect is commonly referred to as supercharg- underpressured reservoir exhibits values of p that
ing (Phelps et al., 1984; Lane, 1993). It is there- are much smaller than the reference pressure.
fore a reasonable practice to question pore pressure In simple virgin (i.e., unproduced) reservoirs, all
measurements associated with a mobility that is the permeable zones are equilibrated with each
less than 1 md/cp. other, and the pore pressure measurements plotted
versus TVD should form a straight line with a
slope given by the density of the continuous fluid
pbh pbh
in the formation. The straight line can be used to
interpolate the pore pressure values between the
pf pf point measurements. The slope will change
depending on the fluid, which enables the determi-
nation of fluid contacts (Fig. 4-7) (Stewart et al.,
High-permeability formation Low-permeability formation 1982; Desbrandes and Gualdron, 1988).
In developed reservoirs or those with complex
Figure 4-6. In high-permeability formations, the fluid leak- structures, building pressure profiles requires many
ing through the mudcake dissipates easily with no signifi- pore pressure measurements. The identification of
cant increase in formation pressure. In low-permeability gradients is a powerful tool for detecting disconti-
zones, the fluid leaking through the mudcake cannot dissi-
pate easily. In this latter case, the pressure measured nuities in pore pressure that most likely correspond
immediately behind the mudcake is higher than the true to discontinuities in permeability and discontinu-
formation pressure. pbh = borehole pressure, pf = far-field ities in stress magnitudes.
pore pressure.

• Skin effect
There are several means for measuring the produc-
tion skin effect. At logging time, the production
X500
skin effect is measured by formation testing. It is,
however, extremely dependent on the geometry of
Gas
the test (Frimann-Dahl et al., 1998). Selective well
testing measures an integrated value of the produc-
tion skin effect over the zones of interest at the
X600 time the well test is conducted. There is currently
Depth (ft)
no means to continuously measure production skin

effect.
Oil It is even more difficult to estimate the injection
skin effect. The injection skin effect can be
obtained at logging time by conducting injection
X700
testing with a packer and probe configuration of
the tester. Using Stoneley wave measurements
Water
together with permeability estimates is a potential
means for determining the injection skin effect
continuously. However, with current technology,
X450 X550 X650 the injection skin effect is a parameter that requires
Pressure (psi)
calibration.
Figure 4-7. A vertical pressure profile defines fluid types • Damage radius
and fluid contact levels.
For a vertical well drilled with conductive mud, an
invasion profile can be computed from resistivity
4-4.5. Skin effect and damage radius logs (Souhaite et al., 1975). Resistivity logs pro-
vide resistivity measurements at several depths of
Skin effect is a measure of the damage inflicted to the investigation (e.g., 10, 20, 30, 60 and 90 in.).
formation permeability in the vicinity of the wellbore. Provided that there is a resistivity contrast between
Damage may result from drilling and completion the mud filtrate and virgin formation fluid, each
processes or the production of formation fluids. The resistivity measurement reflects how much the for-
skin effect can therefore vary during the life of a well. mation fluid was displaced by mud filtrate. The
For candidate selection and subsequent treatment resulting invasion profile does not strictly correlate
evaluation, the skin effect of interest is that related to to a drilling-induced damage profile, but it is a rea-
flow from the formation (i.e., the production skin sonable first-order estimate. This technique can be
effect). For the design and execution of a stimulation extended to nonvertical wells by including the
treatment, however, the skin effect of interest is that effect of dip between the wellbore and the forma-
related to the injection of treatment fluids into the for- tion (Anderson et al., 1996).
mation. Production skin and injection skin effects can There are currently no geophysical means to
be very different. The damage radius is an estimate of measure the location of damage resulting from the
how far the damage extends into the formation. Skin production of formation fluids, although time-lapse
effect and damage radius are important inputs for the gamma ray may correlate with the buildup of
design of matrix stimulation because they affect the radioactive scale.
type and volume of matrix fluids required and the
diversion techniques for fluid placement (see Chapter
19). However, these parameters are difficult to obtain,
and no current method is fully satisfactory. The fol-
lowing methods provide crude estimates.

4-4.6. Composition of fluids formation (Paillet and Chang, 1991). The two
types of waves of interest for estimating the elastic
Although there are indirect logging means for deter- constants of a medium are compressional waves
mining the composition of formation fluids, they are (P-waves) and shear waves (S-waves). The sonic
limited basically to distinguishing among water tool measures the characteristic propagation speed
(brine), gas and oil (e.g., Schiuma et al., 1997). Some of the P- and S-waves.
conclusions can be made about the type of oil by In an isotropic medium, only the two elastic
measuring its specific density through pressure profil- constants of the shear modulus G and Poisson’s
ing (see Section 4-4.4). ratio ν are independent. They are related to the
The only reliable method for determining the velocity of propagation of a P-wave uP and that
composition of fluids downhole is to collect unconta- of an S-wave uS by
minated fluid samples for laboratory analysis.
Sampling is best achieved with formation testers (see G = ρb uS2 (4-24)
“Formation testers” in Section 4-4.3). For example,
the MDT* Modular Formation Dynamics Tester can 2uS2 − uP2
ν= . (4-25)
measure both the resistivity and optical properties of 2(uS2 − uP2 )
the flowing fluid to distinguish between uncontami-
nated formation fluid and the invading water- or oil- The propagation velocity can be replaced with
base drilling mud or mud filtrate. For more informa- the time ∆t it takes a wave to travel a fixed dis-
tion on fluid sampling, see Hashem et al. (1997) and tance d (e.g., between a source and a receiver):
Akram et al. (1998). d2
G = ρb (4-26)
∆tS2
4-5. Properties related to the 2

1  ∆tS 
deformation and fracturing   −1
2  ∆tC 
of rock ν= 2 , (4-27)
 ∆tS 
As presented in Chapters 3 and 5, the geometry of   −1
 ∆tC 
a hydraulic fracture depends strongly on both the
mechanical properties of the rock formation, which where ∆tS and ∆tC are the S-wave and P-wave
describe how the formation will deform under a traveltimes, respectively.
change in stresses, and the stresses acting on the for- Young’s modulus E is related to the two con-
mation. The mechanical properties of the formation stants by
and the in-situ stresses are not independent. A differ- E = 2 G(1 + ν) . (4-28)
ence in mechanical properties between two adjacent
These equations deserve a few comments. First,
formations usually leads to a difference in the in-situ
a measurement of ρb is required. Second, good-
stresses. Except for the simplest situations, at least a
quality compressional and shear wave velocity
qualitative model must be used to explain the rela-
measurements are crucial for the determination of
tionship between the formation mechanical properties
E and ν. In particular, both the compressional and
and stresses.
shear waves must travel in the undisturbed forma-
tion to yield a realistic estimate of ν—in other
4-5.1. Mechanical properties words, “deep” measurements of uP and uS are
essential. A combination of sonic logs recorded
• Elastic properties with different sources (e.g., monopole and dipole)
The only tool that responds to the elastic properties may therefore be necessary to ensure the quality of
of the formation is the sonic. When a pressure the compressional and shear wave velocity esti-
pulse is created in a wellbore filled with fluid, the mates. For zones where the shear velocity is not
complex phenomena that occur at the boundary known, synthetic traveltimes can be generated
between the wellbore and the formation result in from lithological analysis, but these values must
the propagation of several types of waves into the be used with caution. Third, because G is obtained

more directly, it should be considered the primary Thus, values determined for Poisson’s ratio are
elastic parameter in correlations, as discussed sub- usually valid. The data for Young’s modulus
sequently. should be reviewed.
The result of sonic processing—in consideration Several correlations are used to estimate static
of other effects, such as that from fluids in the elastic properties from dynamic elastic properties
vicinity of the wellbore—is a continuous estimate (e.g., Coon, 1968; van Heerden, 1987; Jizba and
versus depth of the elastic properties of an iso- Nur, 1990; Morales, 1993). The correlations were
tropic linear elastic material. However, wave prop- developed by comparing elastic properties from
agation is a phenomenon of small strain with a laboratory tests on core samples to elastic proper-
large strain rate. Rock formations appear stiffer in ties determined from sonic logs run in the cored
response to an elastic wave than in a rock mechanics wells. From a plot of dynamic versus static data
laboratory test, where much larger strains are for several wells in a specific formation, a cloud
applied at a lower strain rate. In fact, the weaker of points was obtained and a curve fitted to it.
the rock, the larger the difference between the elas- Most of the correlations are a variant of
static = ( A × dynamic α ) + B ,
tic properties derived from acoustic measurements
(4-29)
(i.e., the so-called dynamic properties) and those
derived from laboratory experiments (i.e., static where static is the static elastic property and
properties). It is believed that the difference results dynamic the corresponding dynamic elastic prop-
not so much from the applied strain rate as the erty. The constants A, α and B are determined; α
amount of strain applied to the rock (Hilbert et al., is commonly assigned a value of 1.
1994; Plona and Cook, 1995). Nevertheless, there For clean sandstones, correlations have been
is not a one-to-one correspondence between elastic developed between the porosity and the ratio of
properties measured by sonic logs and those mea- dynamic Young’s modulus and static loading
sured by laboratory experiments, which more Young’s modulus. At very low porosity (<5%),
closely approximate the behavior during fracturing. the ratio is close to 1, whereas it is close to 10 at
The sonic and rock mechanics communities 40% porosity.
debate which set of measurements represents the The correlations can be used to obtain first esti-
appropriate elastic properties of materials, and a mates of the static Young’s modulus and Poisson’s
similar concern applies to which measurements ratio of a formation from dynamic measurements.
of Young’s modulus and Poisson’s ratio should If the correlation used was developed from data
be used to model hydraulic fracturing. from rocks belonging to a sequence (e.g., sand/
Work by Papanastasiou (1997), together with shale turbidite) similar to that of the formation—
experimental evidence from van Dam et al. or, even better, from rocks belonging to the same
(1998), shows that the relevant Young’s modulus basin as the formation—the resulting elastic prop-
for hydraulic fracturing is the unloading modulus erties can be considered representative. Otherwise,
measured during laboratory experiments. The a calibration point should be identified (e.g., by
value of the unloading modulus lies between that conducting a fracture calibration test; see Section
of the classic tangent modulus measured with lab- 9-7). Even with a reliable correlation, calibration is
oratory tests of cores and that of the modulus critical for determining Young’s modulus, and the
determined with sonic measurements (Hilbert results may completely rescale the Young’s modu-
et al., 1994; Plona and Cook, 1995). lus profile. Poisson’s ratio in this situation usually
For Poisson’s ratio, there is not much difference is unchanged.
between the loading and unloading values of The preferred method for obtaining a continuous
Poisson’s ratio determined on cores. Moreover, the log of elastic properties for hydraulic fracturing is
difference between Poisson’s ratio measured in the to calibrate the sonic processing output with good-
laboratory and that determined by sonic tools, with quality results of core tests (i.e., to determine a
the provisions listed previously, is not large, pro- specific correlation) (see Section 3-4.4). The cores
vided that no compaction occurred during the lab- should be from the reservoir section and its bound-
oratory test. If compaction occurs, a purely elastic ing layers to fully represent the sequence. The tests
model is no longer valid. should be conducted at a confining pressure simi-

lar to that in the formation downhole and with a Indeed, the key is the lithology, not a particular
controlled pore pressure similar to that in the reser- physical property.
voir. Both the loading and the unloading Young’s
• Soft formations
moduli and Poisson’s ratios should be measured.
The dynamic versus static properties are then plot- The deformation of soft formations is difficult to
ted for each formation to select the appropriate describe with linear elasticity. Several approaches
scaling function; e.g., is the ratio between the have been proposed, such as modified elasticity
dynamic and static properties constant throughout (Franquet and Economides, 1999), for which
the sequence, or is it a function of porosity, of ver- Young’s modulus and Poisson’s ratio are functions
tical depth, of lithology, etc.? The resulting calibra- of the state of stress, and plasticity (e.g., Papanas-
tion is usually applicable within the field (e.g., tasiou, 1997; Nikolaevskiy and Economides,
Edimann et al., 1998). 2000). These approaches require additional para-
meters that cannot be determined by current geo-
• Fracture toughness, strength and friction angle
physical methods.
The three parameters of fracture toughness, strength
and friction angle ϕ are of interest because they are
all related to the failure of rocks. However, no geo- 4-5.2. Stresses
physical measurements are directly sensitive to any
The state of stress in the earth is one of the major fac-
of the three. Strength and friction angle are rou-
tors influencing the geometry of a hydraulic fracture.
tinely measured on cores, but fracture toughness
The state of stress can be described by three principal
is much more difficult to measure in the laboratory
stresses that are perpendicular to each other: maxi-
(see Section 3-4.6).
mum principal stress σ1, intermediate principal stress
Fracture toughness and friction angle are fairly
σ2 and minimum principal stress σ3. Because the
consistent within a particular rock type, provided
principal stress directions are orthogonal, the direc-
that it is consolidated. Tables can be used (Senseny
tion of two principal stresses automatically describes
and Pfiefle, 1984; Atkinson and Meredith, 1987;
the direction of all of them. However, the description
Plumb, 1994a) to provide first estimates of their
is complete only when the order is known (Fig. 4-8).
values.
A complete description of the state of stress is best
A number of factors affect the fracturing resis-
illustrated by an example of a nominal reservoir at
tance or toughness and the scale of the processes
depth (>2000 ft). The weight of the sediments, or the
(see Section 6-7). If fracture toughness is defined
overburden stress σv, is usually one of the principal
as a material parameter, it is not scale dependent
stresses. The two other principal stresses are therefore
(Vliet, 1999), and its effect on hydraulic fracturing
horizontal. The azimuth of the minimum horizontal
is small (Desroches et al., 1993), which in turn
stress σh completes the description of the orientation
allows some error in its determination. However,
of the stresses, because the maximum horizontal
an apparent toughness can be used to include sev-
stress σH is horizontal and orthogonal to σh. What is
eral processes occurring in the vicinity of the tip of
missing is the order of the stresses. Is σv the maxi-
a fracture propagating in soft materials, especially
mum principal stress σ1, in which case σh is the mini-
in unconsolidated sandstones (Ayoub et al., 1992b).
mum principal stress σ3 and σH is the intermediate
In that case, the so-called fracture toughness is
principal stress σ2? Or is σv the minimum principal
purely a calibrated parameter (see Section 5-4.5 and
stress, or even the intermediate principal stress? Each
Sidebar 9B) that does not have much in common
of these cases corresponds to a different stress
with the true fracture toughness of the material.
regime, as subsequently discussed. A complete
For sandstones, the unconfined compressive
description of the state of stress is of particular
strength (UCS) of the rocks can also be estimated
importance because hydraulic fractures propagate
from correlation with sonic data (e.g., Kowalski,
perpendicular to the minimum principal stress. If σ3
1975).
is horizontal, a vertical fracture will be created; if σ3
The friction angle can also be estimated using
is vertical, a horizontal fracture will be created; if σ3
correlations with physical properties such as
is inclined, an inclined fracture normal to it will be
porosity (e.g., Plumb, 1994a) but with less success.
created.

most likely a principal stress, provided that the loca-

Lithostatic State of Stress tion is not near a major fault.
σv = σo
• The value of σv is determined first.
• Second, the orientation of σh is determined. This
could be a consideration in planning the drainage
pattern of the reservoir. But more importantly, the
σH = σo σh = σo orientation of the principal stresses can vary signif-
icantly from one layer to the next and thereby
enhance hydraulic fracture containment.
Equal Horizontal Stresses • Finally, the value of the minimum stress as a func-
σv = σ1 tion of depth is determined and compared with that
of the overburden to determine which stress is σ3.
For some cases in a thrust-faulted basin, especially
in carbonates with strongly contrasting elastic proper-
ties, the order of the stresses can differ from one layer
σh = σH = σ2 = σ3
to the next, strongly affecting the treatment.
In other words, not only is the magnitude of σh of
States of Stress and Fault Regimes
concern, but also the orientation and order of the prin-
cipal stresses. For shallower formations (≤2000 ft),
σv = σ1
Normal fault regime the situation can be more complex because the over-
burden is not necessarily a principal stress.
• Determination of overburden
The overburden is the weight of the column of
σH = σ2 σh = σ3
sediments. Although it is not measured directly, it
can be easily computed as the integral over depth
of the bulk density:
Thrust fault regime σv = σ3
zTVD
σ v ( zTVD ) = ∫ ρ dz .
b
(4-30)
0
σh = σ2 However, ρb is rarely measured up to the surface

σH = σ1
more than once in the lifetime of a field (i.e., for
seismic profiling). Research may be necessary to
Strike-slip fault regime
σv = σ2 obtain an estimate of ρb between the top of the log
and the surface. Furthermore, for deepwater pro-
jects, the significant affect of the water column on
σv must be included.
σh = σ3 • Stress orientation
σH = σ1
The orientation of the principal stresses can be
determined using diverse resources. The popular
Figure 4-8. Stress regimes and order of stresses approaches are discussed here.
σo = equal-stress constant.
– World stress map
If no information is available for the well or
For hydraulic fracturing applications, the recom- field, the world compilation of stress orienta-
mended methodology for determining the stresses is tions can be consulted (e.g., see the World
as follows. This approach is restricted to reservoirs at Stress Map site on the World Wide Web at
depths greater than 2000 ft because the overburden is http://www-wsm.physik.uni-karlsruhe.de/).

However, most of the data are from the analysis A dipole sonic tool can be run in the wellbore
of deep seismicity, which occurs at much greater to measure the velocity of shear waves polarized
depths than typical oilfield wells. in an orthogonal arrangement. From these data,
– Geologic maps the azimuth of the fastest shear wave can be
extracted. In a well close to vertical, the azimuth
A geologic map of the field should be examined
corresponds to the direction of σH (Endo et al.,
to see if major faults have been identified. In the
1996).
vicinity of the faults, the principal stresses have
One question that often arises is whether the
most likely rotated to align themselves parallel
anisotropy measured from the slow and fast
to the fault surface (e.g., Barton and Zoback,
shear directions is induced by the stress contrast
1994).
or the intrinsic properties of the medium. In a
– Shear anisotropy vertical well, most of the anisotropy is due to
Shear waves are polarized waves and are there- stress effects and can readily be related to the
fore sensitive to differences in the elastic proper- direction of the far-field stresses (Endo et al.,
ties of the material in the direction perpendicular 1996). This is not necessarily the case in hori-
to their travel path. If the properties are suffi- zontal or deviated wells, and the use of shear
ciently different, shear waves polarized in one anisotropy to infer stress directions is not recom-
direction will travel significantly faster than mended in these situations.
those polarized in the orthogonal direction (Fig. Processing dipole sonic data for anisotropy
4-9). If the rock formation contains a population analysis is challenging. As a rule, the azimuth
of defects that have a uniform distribution of of the fast shear wave that is reported on the
orientation, the defects that are normal to σH are logs should be fairly stable regardless of the
under more stress than those normal to σh. A rotation of the tool. If this is not the case, the
shear wave traveling vertically and polarized rotation of the tool—not a formation property—
parallel to σH will therefore travel faster than was measured. In addition, the uncertainty win-
a shear wave polarized perpendicularly to σH. dow reported around the azimuth should be tight
(i.e., less than 5° on each side). These simple
quality control measures identify most of the
poor-quality data. The fast shear azimuth com-
puted from good-quality shear anisotropy data
provides a robust estimate of the azimuth of the
maximum horizontal stress σH in a vertical well
(Fig. 4-10).
Fast
S-wave
– Wellbore breakouts
Dipole
receivers The presence of a wellbore in the formation
Slow generates a local change in the stresses, a
S-wave process referred to as stress concentration (see
Section 3-5.7). When the pressure in the well-
bore falls below a certain level, the formation
may fail in compression or shear, and wellbore
breakouts can form (Zoback et al., 1985; Plumb
and Cox, 1987). The breakout is symmetrical
and bi-wing (Fig. 4-11). In a vertical well, inter-
Source
Dipole pulse pretation is straightforward because the azimuth
source of the breakout corresponds to the azimuth of σh
if σh is a principal stress direction. In deviated or
horizontal wells, the location of breakouts is a
function of the ratio of stresses (σ2 – σ3)/(σ1 – σ3)
Figure 4-9. Relation between the polarization of shear and the orientation of the wellbore with respect
waves and stresses.

Figure 4-10. Stress direction log obtained by the analysis

of shear wave anisotropy. The orientation on track 3
changes between the lower section (X425 and below)
and the upper section.
to the stresses. Processing is required (Cesaro et

al., 1997), and in some cases it can identify both
the orientation of the stresses and the ratio
between them.
The simplest way to detect breakouts is to
examine the log track recorded by a four-arm
caliper (Fig. 4-12) (Plumb and Hickman, 1985).
More sophisticated imaging logs can also be Figure 4-11. Reflectance image showing breakouts (A and
B) and drill collar wear (C) in an interbedded sandstone
used (Fig. 4-13). Special care must be taken not and shale. The breakout width varies with the rock
to misidentify features such as key seating and strength and increases with depth below 7565 ft.
overreaming as breakouts. For more information
on the identification of breakouts, see Bratton et
al. (1999) and Maury et al. (1999). when the pressure in the wellbore rises above
Simple interpretation assumptions are used to a certain level. If the resistance of the rock is
infer stress magnitudes from breakouts (Cesaro reached, drilling-induced fractures will be cre-
et al., 1997). However, currently no technique ated. These readily identifiable fractures occur
reliably infers the stress magnitudes from the parallel to the wellbore axis in two sets, 180°
presence of breakouts alone. apart (Fig. 4-14). They can be continuous or
discontinuous; the latter are called en echelon
– Drilling-induced fractures
fractures.
Because of the stress concentration resulting Drilling-induced fractures are shallow and
from the presence of the wellbore, tensile conductive, which makes them obvious on elec-
stresses may be generated at the borehole wall trical imaging logs, such as that recorded by the

σθθ = 3σ2 – σ1 – pm
σ1 σ1
σθθ = –σ1 σθθ = –pm
pm = 0 pm pm σ2
σθθ = 3σ1 σθθ = 3σ1 – σ2 – pm
Interpreted Borehole Azimuth Pad 1

Cross Section N E S W N
In gauge
Pad 1 Calipers
Direction 1 and 3
of σh
Breakout
1
Mechanical 1 Bit size

wear Well deviation
azimuth
Severe
breakout
1
Calipers
2 and 4
Figure 4-12. Interpretation of calipers for breakouts and Figure 4-13. Three-dimensional (3D) computer model of
the stress direction. σθθ = circumferential stress, pm = mud a 2-ft section of an 81⁄2-in. diameter borehole constructed
pressure in the wellbore. from UBI* Ultrasonic Borehole Imager transit-time data.
The ridges on the right and left sides are breakout elong-
ations resulting from stress-induced wellbore failure.
FMI* Fullbore Formation MicroImager. They
are distinguished from natural fractures because
the fracture sets parallel the wellbore, in some Strickland and Ren, 1980). In the case of hard
cases for hundreds of feet. rocks, fast unloading of the core creates a popu-
In a vertical well, drilling-induced subvertical lation of defects (i.e., microcracks) related to the
fractures occur in the direction of the maximum stress field at the time of coring (i.e., the current
horizontal stress. In deviated and horizontal state of stress). Provided that the population of
wells, en echelon fractures are observed. An defects is dominant in the sample, which is usu-
analysis similar to that conducted for breakouts ally the case, the DSA technique will yield infor-
can yield stress ratios and stress directions mation on the current state of stress. This tech-
(Wiprut and Zoback, 1998). nique also provides the order of the principal
A final check consists of verifying that any stresses, making it extremely useful when it is
breakouts occur 90° from drilling-induced suspected that the vertical stress is not a principal
fractures. stress or not the maximum principal stress.
– Tests on cores – Population of natural fractures
The recovery of cores introduces nonlinear This technique does not provide information
effects because of the brutal change of stress about the current state of stress and therefore is
(i.e., unloading) to which the cores were sub- not useful unless the fractures were created by
jected. Anelastic strain recovery (ASR) or differ- the current state of stress. The state of stress
ential strain analysis (DSA) may yield the direc- changes often and significantly at the geologic
tion of the principal stresses (see Section 3-6.4) time scale. The current state of stress, which is
(Voigt, 1968; Siegfried and Simmons, 1978; of interest for hydraulic fracturing, is typically

Laboratory analysis of cores (see the previous

mention of DSA and ASR) can provide from strain
measurements the ratio of effective stresses acting
on the formation at the time the core was taken.
These techniques are discussed in detail in Section
3-6.4.
Stress measurements using the micro-hydraulic
fracturing technique (Haimson, 1993) can be con-
ducted in open- or cased holes. As also discussed
in Section 3-6.2, the tests measure the pressure at
which a 3- to 10-ft fracture opens or closes to pro-
duce an estimate of the minimum total stress act-
ing on the formation (Desroches and Kurkjian,
1999) averaged over a height of several feet.
Another advantage of the technique is that bound-
ing layers can also be tested.
Closure pressure determined from a hydraulic
fracturing calibration treatment is also an estimate
of the minimum total stress acting on the forma-
tion. The difference from the previous technique
is that the measurement is averaged over a much
larger volume of rock and usually corresponds to
the pay zone. Provided that the averaging is taken
into account, the minimum principal stress values
yielded by both the micro-hydraulic fracturing
technique and hydraulic fracturing calibration
agree (e.g., Desroches and Woods, 1998) with the
exception of complex situations. Pressure deriva-
tive analysis of the calibration treatment can also
provide the average minimum total stress for the
bounding layer, allowing the primary height
Figure 4-14. Drilling-induced extension fractures in sand- growth of the fracture (see Sidebar 9C).
stone in the Travis Peak formation, Wascom field, East
Texas. • Measurement of intermediate principal stress
The intermediate principal stress is always inferred
different from that at the time of creation of the from a relation that involves at least the minimum
fractures in the formation. principal stress (Haimson and Huang, 1989; Ito
– Hydraulic fracture diagnostics and Hayashi, 1991; Guo et al., 1993). Obtaining
a good estimate of the intermediate principal stress
Microseismicity or tiltmeters (e.g., Warpinski,
is difficult. It is mentioned only because it was
1994) (see Section 12-1.1) can provide an esti-
once popular to use the breakdown pressure for the
mate of the azimuth of a hydraulic fracture.
calculation (Hubbert and Willis, 1957). However,
Because a hydraulic fracture propagates perpen-
more recent work on the breakdown process
dicular to the minimum principal stress, its diag-
(Detournay and Carbonell, 1994) shows that the
nostics provide information about the direction
breakdown pressure provides unreliable estimates
of the minimum principal stress, albeit after the
and should not be used in the analysis of stresses.
fact.
• Continuous estimate of minimum principal stress
• Measurement of the minimum principal stress versus depth
The minimum principal stress acting on a forma- As for the other properties mentioned in this chap-
tion can be measured only discretely and indirectly. ter, the goal is to obtain a stress profile (i.e., a con-

tinuous description) of the value and orientation of Depending on the ordering of the stresses,
the minimum principal stress over the zone of three cases are defined:
interest. To generate the best continuous descrip- – σv is the maximum principal stress σ1: normal
tion from the information available, geologic infor- fault regime
mation must also be used. – σv is the minimum principal stress σ3: thrust
The three-step procedure proceeds from the gen- fault regime
eral to a particular case. First the global basin per- – σv is the intermediate principal stress σ2:
spective is used to determine the most likely stress strike-slip fault regime.
regime acting on the formation. A mathematical
Stress measurements from around the world
model (also called a stress model) is chosen in the
also indicate that pore pressure, lithology and
second step to represent the variation of the
the position of the layer within the structure
stresses across the sequence of interest comprising
(e.g., Whitehead et al., 1987; Thiercelin and
the pay zone and bounding layers. The model is
Plumb, 1991; Aleksandrowski et al., 1992) are
selected and calibrated using stress measurements
factors influencing variation of the stresses in
and formation properties. It provides stress data for
sedimentary rocks. Typically, pore pressure
the design of hydraulic fracturing treatments for
changes induce stress changes that are greater
the sequence of interest. Finally, after a calibration
than those associated with lithology, tectonic
treatment, further refinement based on lithology
setting and structural or stratigraphic position.
yields the stress variations in the layers to be
The stress regime is extremely important.
stimulated.
Depending on the stress regime, identical litho-
– Stress regimes logic structures can induce very different stress
An assessment of the stress regime establishes contrasts. For example, in a typical sand/shale
the expected cause for bed-to-bed stress varia- sequence (Fig. 4-15A), the shales bear higher
tions through the formation and helps to identify stresses than the sandstones in a relaxed tectonic
a mathematical model for defining the stress setting, whereas the sandstones bear higher
profile. stresses than the shales in a compressive tectonic
Five states of stress may be defined in terms of setting (Fig. 4-15B).
the principal stresses (Fig. 4-8) (Engelder, 1993). – Mathematical stress models
The simplest case consists of all three princi-
For the tectonic setting of the basin, the most
pal stresses being equal (i.e., lithostatic state of
likely of the five states of stress described previ-
stress). This can occur in materials with little of
ously is identified. The probable state of stress
no shear strength, such as poorly consolidated
guides the ordering of the stresses and selecting
shales, or in materials that flow, such as salt. It
may also be approached in extremely overpres-
sured sediments. This state of stress is not Sedimentary Sequence Relaxed State Compressed State
widely documented because stress is usually not
measured in these materials. Shale σh = σ3
σv = σ2
The second simplest case occurs in regions
σh = σ2
where the two horizontal stresses are equal and Sandstone σh = σ3
σv = σ3
less than the vertical stress of the overburden.
This regime may be expected in basins located
in quiet, intraplate settings.
p σv = σ1 p σh σv
For the remaining cases, the three principal
stress magnitudes differ significantly. In-situ
stress measurements and borehole image analy- Figure 4-15A. Schematic stress profiles through a bedded
sand/shale sequence show the two end-member stress
sis concur to indicate that this is the most likely profiles observed in sedimentary rocks. The relaxed-state
situation. Unequal horizontal stresses can be profile is most frequently observed. The compressed-state
attributed to tectonic forces or effects that result profile is expected wherever significant horizontal strain is
from the presence of a geologic feature (e.g., present.
a fold or fault).

For the remaining three cases of unequal hori-

zontal stresses, it is appropriate to use either the
poroelastic stress model or a friction failure
model, with failure computed between the mini-
mum and maximum principal stresses and the
ordering coming from the faulting regime.
To select the stress model, measurements of
the magnitude of the minimum principal stress
are required for several points in the considered
sequence. Two measurements are the absolute
minimum.
The methodology defines the coefficient of
earth stress K as the ratio between the effective
vertical stress σv′ and the effective minimum
horizontal stress σh′:
σ ′h
K= , (4-32)
σ ′v
which is approximated, to the first order, by
σh − p
K* = , (4-33)
σv − p
where p is the pore pressure (Fig. 4-16).
K* is plotted with respect to the variables ν,
E, true vertical depth zTVD, ϕ and p. Strong corre-
lation among K*, ν and p favors using the uni-
0.50
0.45
Figure 4-15B. Lithology-controlled breakouts, indicating 0.40

a compressed state.
0.35
the stress model that provides a continuous
K factor
description of the minimum horizontal stress

as a function of depth over the zone of interest. 0.30
For the simplest, lithostatic case:

zTVD 0.25
σ h ( zTVD ) = σ H ( zTVD ) = σ v ( zTVD ) = ∫ ρb ( z )dz .
0
0.20
(4-31)
For the case in which the equal horizontal 0.15
stresses are less than the vertical stress, the 3600 3800 4000 4200 4400 4600 4800 5000
model can be either the uniaxial model or the Vertical effective stress, σv´ (psi)
friction failure model, with failure computed
between the horizontal stress and the vertical Figure 4-16. K factor (= (σh – p)/(σv – p)) as a function of
vertical effective stress (equivalent to zTVD) (Morita et al.,
stress. 1998).

axial model. A continuous profile of σh is σ h = (1 − sin ϕ )σ v + (sin ϕ ) p. (4-38)

obtained by using the following equation with
the Biot constant α as an ad hoc potentially This model is appropriate in high-porosity, high-
lithology-dependent calibration parameter: permeability sandstones, chalks and shales. It
corresponds to rock at rest if the rock compacts.
ν
σh =
1− ν
(σ v − αp) + αp. (4-34) There are no calibration parameters in the mod-
els governed by friction except for the friction
Use of this model is reasonable for low-porosity, angle. Calibration is conducted by modifying the
low-permeability sandstones, shales and carbonates. friction angle.
Strong correlation among K*, ν, E and p The effect of pore pressure variations (through
favors using the poroelastic model. Indeed, a either depletion or water flooding; see Section
significant effect of Young’s modulus suggests 3-5.7) and temperature variations can also be
tectonic activity. A continuous profile of σh is added to the selected model (e.g., Perkins and
obtained by using the following equation: Gonzales, 1984). Another effect that can be con-
sidered is stress relaxation; it reduces the differ-
ν νE
( σ v − αp) + αp +
E ence between the minimum and maximum
σh = εh + εH ,
1− ν 1− ν 2
1 − ν2 stresses (e.g., Prats, 1981), especially in shales.
Once the stress model is built, it usually is
(4-35) portable within the same sequence in the same
basin, provided that significance differences in
where the minimum tectonic strain εh and maxi- pore pressure and possibly in temperature are
mum tectonic strain εH are the primary calibra- taken into account.
tion factors (i.e., adjusted first with α = 1) and The methodology advocated in this chapter
α is the secondary calibration parameter adjusted relies on physical processes underlying the
once the tectonic strains are adjusted. The mathematical form of the model selected to
amount and sign of the tectonic strains must be build a stress profile. Other mathematical forms
compatible with the geologic setting. For exam- can be selected, but without geologic confirma-
ple, negative strains are not possible in a com- tion, the portability of the stress profile to
pressive environment. This model applies in another well is not likely.
lithologies similar to those applicable to Eq. 4-34.
Strong correlation among K*, the friction – Stress profile adjustment after a calibration
angle ϕ and p (but not ε or ν) favors use of the treatment
incipient failure model: As mentioned previously, the analysis of a cali-
bration treatment yields an average of the mini-
(σ v − p) + p
1
σh = (4-36) mum stress acting on the pay zone (called the
Nϕ closure pressure) and possibly a value of the
average stress contrast between the pay zone
π ϕ
Nϕ = tan 2  +  , (4-37) and the bounding layer in which the fracture
 2 4 may have grown.
where Nϕ is the coefficient of passive stress. If only the stress acting on the pay σPZ is
This model is appropriate for medium- to high- available, the model-based stress profile can be
porosity sandstones (15% and greater) and clay- scaled or shifted to honor its value at depth. If
supported rocks, particularly in regions of nor- the stress acting on a bounding layer σBL is also
mal (or growth) faulting. The friction angle of available, further calibration can be performed.
these rocks does not vary greatly, so this setting It is preferable to modify the parameters of the
yields only small variations of the stresses as a mathematical model to honor these stress values,
function of lithology. but the parameters may not be available at the
Correlation among K*, ϕ and p with depar- time of the calibration treatment. If this is the
ture from Eq. 4-37 suggests use of the purely case, a lithology-based stress interpolation can
frictional model: be used as follows.

One of the formation properties can usually A simple solution is to consider as many zones as
be identified as representative of the difference there are significant changes. But this is not practical
in stresses (e.g., the percentage of dolomite in because more than 100 zones could be defined and
a carbonate sequence or the volume of clay in only a limited number of zones can be considered.
a sand/shale sequence). For this property A, APZ Therefore, broader zones must first be defined. The
and ABL are its values in the pay zone and the first step is to decide which property, or set of proper-
bounding layers, respectively, and the stresses in ties, should guide the definition of zones. In sand/
the vicinity of the pay zone may be estimated as shale sequences, for example, the volume of clay is
a linear function of A: a primary guide, whereas in carbonate sequences, the
σ( zTVD ) = m × A( zTVD ) + b, (4-39) percentage of dolomite is a good primary guide.
However, permeability barriers (usually associated
where the linear regression constants are with pore pressure and hence stress changes), changes
σ BL − σ PZ in elastic properties and changes in the stresses should
m= (4-40) also be considered. The set of properties selected to
ABL − APZ
guide the zoning process must also reflect the geo-
logic understanding that guided choosing the model
b = σ PZ − m × APZ . (4-41)
for the stresses. Determination of the leakoff height
This approach has been successfully applied is also part of the zoning process. The zones must be
for both sand/shale (Miller and Smith, 1989; defined in a consistent manner so that the leakoff
Smith et al., 1989) and carbonate sequences. coefficient determined from a calibration treatment
is meaningful for the remainder of the sequence.
A zoning example is presented in Fig. 4-17. This
4-6. Zoning sand/shale sequence reservoir has strong variations
in permeability and pore pressure because of water
Most of the models used in reservoir stimulation flooding. The main properties selected to define zones
require defining zones with constant, or possibly lin- were the separation between sands and shales and the
early varying, properties within each zone. Once all profile of pore pressure. More zones were added to
the required properties have been collected, the logs consider sharp changes in Young’s modulus, which
must be zoned into distinct zones that will be the final resulted in the definition of 25 zones. Zoning for for-
input to the stimulation process. If parallel, horizontal mation evaluation would have been significantly dif-
beds are assumed, the zones must ultimately be ferent, reflecting the analyst’s different point of view.
defined in TVD. If the beds are not parallel or not Once the zones have been defined, they must be
horizontal, more complicated and ad hoc transforma- populated with a constant value (or possibly linearly
tions must be made, but these adjustments are outside varying, for pore pressure and stresses). For thin beds
the scope of this chapter. with significant contrast, taking the peak value is pre-
Zoning can be done subjectively by hand or objec- ferred, whereas for thick beds, taking an average
tively in a manner called log squaring, which was away from the bed boundaries is preferred.
developed for formation evaluation (Trouiller et al., Finally, most log measurements sample the forma-
1989; Serra and Andreani, 1991). Log squaring uses tion to only a few feet from the wellbore. The zoning
the response functions of the tools introduced in process described here, however, populates zones
“Lithology” in Section 4-4.2. However, it is not with a constant value along the bed extension, ignor-
directly applicable to reservoir stimulation because it ing any heterogeneity along the bed. The constancy
considers only the classic measurements used in for- may be arbitrary, especially for geologic environ-
mation evaluation (e.g., porosity, density, resistivity). ments known to typically have lateral variations of
It ignores pore pressure, permeability, mechanical properties (e.g., channels, bar sands). Crosswell
properties and stresses, which are integral to the zon- tomography (Warpinski et al., 1998b) shows that for-
ing process considered here. Furthermore, because no mation properties can vary significantly along a bed
response equation exists yet for these latter proper- in less than 50 ft of horizontal displacement. Yet
ties, the zones output by log squaring must be manu- knowledge of the geologic environment is the only
ally refined. qualifier of the information developed from the
zoning process.

Figure 4-17A. Log information for zoning.

Formation Pressure Poisson’s Ratio Young’s Modulus Minimum Stress Porosity

PR .ASCII_Load [A5817406] YME .ASCII_Load [A5817403 AVG_EFF_FAR_PRIN_STRESS NPHI .ASCII_Load [A59006]
0.25 0.5 0.5 (Mpsi) 1 2000 (psi) 3000 0.6 (ft3/ft3) 0
MD PA_PAQP_FORM_PRESS .MLIS PR .MLIS_2 [A5877057] YME .MLIS_3 [A5877064] AVG_EFF_FAR_PRIN_STRESS NPHI .MLIS_1 [A5900984]
1:500
(ft) 1000 (psi) 1500 0.25 0.5 0.5 (Mpsi) 1 2000 (psi) 3000 0.6 (ft3/ft3) 0
2900
2950
3000
3050
3100
Figure 14-17B. Corresponding zoned formation model.

◆
Basics of
Hydraulic Fracturing
M. B. Smith, NSI Technologies, Inc.
J. W. Shlyapobersky,† Shell E&P Technology Co.
5-1. Introduction treatment (see Chapters 7 and 8). However, the

design must also be consistent with the physical
Since its introduction, hydraulic fracturing has been, limits set by actual field and well environments.
and will remain, one of the primary engineering Also, treatments must be conducted as designed to
tools for improving well productivity. This is achieve a desired result (i.e., full circle to the critical
achieved by role of operations). Proper treatment design is thus
• placing a conductive channel through near- tied to several disciplines:
wellbore damage, bypassing this crucial zone • production engineering
• extending the channel to a significant depth into • rock mechanics
the reservoir to further increase productivity
• fluid mechanics
• placing the channel such that fluid flow in the
• selection of optimum materials
reservoir is altered.
• operations.
In this last instance, the fracture becomes a tool
for true reservoir management including sand Because of this absolutely essential multidiscipli-
deconsolidation management and long-term nary approach, there is only one rule of thumb in frac-
exploitation strategies. As first visualized (see the turing: that there are no rules of thumb in fracturing.
Appendix to this chapter), the concept of hydraulic The multidisciplinary nature, along with the diffi-
fracturing was quite straightforward. This visual- culty in firmly establishing many of the design vari-
ization is described in the following, and in general, ables, lends an element of art to hydraulic fracturing.
for reasonably simple geology, the basic physics of This is not to say that the process is a mystery nor
fracturing is straightforward and well established. is it to say that for most cases the basic physics con-
Complexity arises from two directions: geologic trolling the process is not defined (see Chapter 6). It
reality and the inherent multidisciplinary nature of simply says that the multitude of variables involved,
the fracturing process. along with some uncertainty in the absolute values
Historically, the control of fracturing has rested of these variables, makes sound engineering judg-
with drilling and operations groups owing to the ment important.
nature of field procedures using pumps, packers,
pressure limits, etc. However, the final results (and
thus design) are dominantly a production engineer-
5-1.1. What is fracturing?
ing exercise, and fracturing cannot be removed from If fluid is pumped into a well faster than the fluid
intimate contact with reservoir engineering. At the can escape into the formation, inevitably pressure
same time, designing a treatment to achieve the rises, and at some point something breaks. Because
desired results is also intimately connected with rock rock is generally weaker than steel, what breaks is
mechanics (which controls fracture geometry; see usually the formation, resulting in the wellbore
Chapters 3 and 4), fluid mechanics (which controls splitting along its axis as a result of tensile hoop
fluid flow and proppant placement inside a fracture; stresses generated by the internal pressure. The
see Chapter 6) and the chemistry that governs the mechanics of this process are described in Section
performance of the materials used to conduct the 3-5.7, and the simple idea of the wellbore splitting
like a pipe (shown as a cartoon in Fig. 5-1) becomes
†
Deceased more complex for cased and/or perforated wells and

Figure 5-1. Internal pressure breaking a vertical wellbore. Figure 5-2. Cross-sectional view of a propagating fracture.
nonvertical wells. However, in general, the wellbore open and maintain a conductive flow path for the
breaks—i.e., the rock fractures—owing to the action increased formation flow area during production.
of the hydraulic fluid pressure, and a “hydraulic” The propping agent is generally sand or a high-
fracture is created. Because most wells are vertical strength, granular substitute for sand (see Section 7-7).
and the smallest stress is the minimum horizontal Alternatively, for carbonate rocks, the hydraulic fluid
stress, the initial splitting (or breakdown) results in may consist of acid that dissolves some of the forma-
a vertical, planar parting in the earth. tion, leaving behind acid-etched channels extending
The breakdown and early fracture growth expose into the reservoir.
new formation area to the injected fluid, and thus the After the breakdown, the fracture propagation rate
rate of fluid leaking off into the formation starts to and fluid flow rate inside the fracture become impor-
increase. However, if the pumping rate is maintained tant. They are dominated by fluid-loss behavior. As
at a rate higher than the fluid-loss rate, then the introduced by Carter (1957) and discussed in the fol-
newly created fracture must continue to propagate lowing (and in Chapters 6 and 9), the fluid-loss rate
and grow (Fig. 5-2). This growth continues to open qL from a fracture can be expressed as
more formation area. However, although the
hydraulic fracture tremendously increases the forma- 2 CL A
qL ≈ , (5-1)
tion flow area while pumping, once pumping stops t−τ
and the injected fluids leak off, the fracture will close
and the new formation area will not be available for where CL is the fluid-loss coefficient, A is an element
production. To prevent this, measures must be taken of the fracture area (i.e., increased inflow area), t is
to maintain the conductive channel. This normally time measured from the start of pumping, and τ is
involves adding a propping agent to the hydraulic the time when each small area element of a fracture
fluid to be transported into the fracture. When pump- is created or opened. As a direct consequence of this
ing stops and fluid flows back from the well, the relation, the highest rate of fluid loss is always at the
propping agent remains in place to keep the fracture fracture tip. Newly created fracture area exists at that
5-2 Basics of Hydraulic Fracturing

point (t – τ = 0 in the denominator), making qL in- Thus, proppant concentration (i.e., volume fraction
stantly infinite. of solid proppant) increases as the slurry stages dehy-
Initially, fracture penetration is limited, and hence drate. The pump schedule, or proppant addition
fluid loss is high near the wellbore. For that reason, schedule, must be engineered much like handicapping
the first part of a hydraulic fracture treatment con- horse races, but with no single winner. Rather, all
sists of fluid only (no proppant); this is termed the stages should finish at the right place, at the right
pad. The purpose of a pad is to break down the well- time, with the right final proppant concentration. The
bore and initiate the fracture. Also, the pad provides pad should be completely lost to the formation, and
fluid to produce sufficient penetration and width to the first proppant stage should be right at the fracture
allow proppant-laden fluid stages to later enter the tip (which should be at the design length).
fracture and thus avoid high fluid loss near the frac- As the proppant slurry stages move down the frac-
ture tip. After the pad, proppant-laden stages are ture, they dehydrate and concentrate. Slurry stages
pumped to transport propping agent into the fracture. pumped later in the treatment are pumped at a higher
This chapter describes the process for propped frac- concentration. These stages are not in the fracture for
ture treatments; acid fracture treatments are dis- long prior to the treatment end (i.e., prior to shut-
cussed in Section 10-6. down) and are thus exposed to less fluid loss and less
However, because fluid loss to the formation is still dehydration. Ideally, the first proppant stage pumped
occurring, even near the well, the first proppant is reaches the fracture tip just as the last of the pad
added to the fluid at low concentrations. The prop- fluid is lost into the formation (a correctly handi-
pant-laden slurry enters the fracture at the well and capped race), and this first stage has concentrated
flows toward the fracture tip (Fig. 5-3). At this point, from its low concentration to some preselected,
two phenomena begin. First, because of the higher higher final design concentration. Meanwhile, the
fluid loss at the fracture tip, slurry flows through the slurry concentration being pumped is steadily
fracture faster than the tip propagates, and the prop- increased to the same final design concentration.
;
pant-laden slurry eventually overtakes the fracture tip. At treatment end, the entire fracture is filled with the
Next, because of fluid loss, the proppant-laden slurry design concentration slurry. Design considerations
stages lose fluid (but not proppant) to the formation. for the final concentration are discussed later in this
section and in detail in Section 10-4.
The preceding description might be termed a “nor-
;; ;;
; ;
mal” design, where the entire fracture is filled with a
uniform, preselected, design proppant concentration
just as the treatment ends. If pumping continues past
that point, there would be little additional fracture
extension because the pad is 100% depleted. Con-
tinued pumping forces the fracture to become wider
;;;;;;
(and forces the pressure to increase) because the
increased volume simply acts like blowing up a bal-
loon. In some cases the additional propped width
that results may be desirable, and this procedure is
used purposely. This is termed tip-screenout (TSO)
fracturing.
At the conclusion of the treatment, the final flush
stage is pumped. This segment of a treatment con-
sists of one wellbore volume of fluid only and is
intended to sweep the wellbore clean of proppant
(Fig. 5-4). The well is generally then shut-in for
some period to allow fluid to leak off such that the
fracture closes on and stresses the proppant pack.
Shut-in also allows temperature (and chemical
Figure 5-3. Introducing proppant into the fracture. breakers added to the fluid while pumping) to reduce

are addressed separately in this section, they fre-

quently overlap.
;;;
;
• Damage bypass
Near-wellbore damage reduces well productivity.
This damage can occur from several sources,
including drilling-induced damage resulting from
fines invasion into the formation while drilling
;;;
;
and chemical incompatibility between drilling flu-
ids and the formation. The damage can also be
due to natural reservoir processes such as satura-
tion changes resulting from low reservoir pressure
near a well, formation fines movement or scale
deposition. Whatever the cause, the result is unde-
sirable. Matrix treatments (discussed in Chapters
13 through 20) are usually used to remove the
damage chemically, restoring a well to its natural
productivity. In some instances, chemical proce-
dures may not be effective or appropriate, and
hydraulic fracture operations are used to bypass
the damage. This is achieved by producing a high-
conductivity path through the damage region to
Figure 5-4. Flushing the wellbore to leave a propped frac- restore wellbore contact with undamaged rock.
ture. • Improved productivity
the viscosity of the fracturing fluid (see Section 7-6.2). Unlike matrix stimulation procedures, hydraulic
Ideally, this process leaves a proppant-filled fracture fracturing operations can extend a conductive
with a productive fracture length (or half-length xf), channel deep into the reservoir and actually stimu-
propped fracture height and propped fracture width late productivity beyond the natural level.
(which determines the fracture conductivity kfw). All reservoir exploitation practices are subject
Here, xf is the productive fracture half-length, which to Darcy’s law:
may be less than the created half-length L or less kh ∆p  A 
than the propped length. q≈ , (5-2)
µ ∆x  h 
5-1.2. Why fracture? where the all-important production rate q is relat-

ed to formation permeability k, pay thickness h,
Hydraulic fracture operations may be performed on a
reservoir fluid viscosity µ, pressure drop ∆p and
well for one (or more) of three reasons:
formation flow area A. Reservoir exploitation
• to bypass near-wellbore damage and return a well revolves around manipulating this equation. For
to its “natural” productivity example, pressure drop may be increased by using
• to extend a conductive path deep into a formation artificial lift to reduce bottomhole flowing pres-
and thus increase productivity beyond the natural sure, water injection to increase or maintain reser-
level voir pressure, or both. For other cases, in-situ
• to alter fluid flow in the formation. combustion or steam injection is used to reduce
reservoir fluid viscosity and thus increase produc-
In the third case, fracture design may affect and tivity. For fracturing, as pictured in Fig. 5-5, oper-
be affected by considerations for other wells (e.g., ations are on the formation area in the equation,
where to place other wells and how many additional with the increased formation flow area giving the
wells to drill). The fracture becomes a tool for reser- increased production rate and increased present
voir management. Although these three motivations value for the reserves. (Strictly speaking, it is the

accelerating production from a single well to reser-

voir management, occurred with the application of
;;;
;;
;;
; ;
massive stimulation treatments in tight gas forma-
tions (see Appendix to this chapter). Although out-
wardly a traditional application of fracturing to
poorer quality reservoirs, these treatments repre-
sented the first engineering attempts to alter reser-
;;;
;;
voir flow in the horizontal plane and the methodol-
ogy for well placement (e.g., Smith, 1979).
Fracturing for vertical inflow conformance (i.e.,
reservoir management) was successfully used in
the Gullfaks field (Bale et al., 1994), where selec-
tive perforating and fracturing were used to opti-
mize reserve recovery and control sand production
while maintaining (but not necessarily increasing)
the required production rates. This is illustrated in
Fig. 5-6, where the bottom, low-permeability
Rannoch-1 zone was perforated to create a
propped fracture that extends up and into the
high-permeability (>1000-md) Rannoch-3 zone.
Without fracturing, the entire zone can be perfo-
rated, and a low drawdown allows a significant
production rate on the order of 20,000 STB/D,
Figure 5-5. Increased flow area resulting from a fracture. sand free. However, sand production is triggered
by water breakthrough in the high-permeability
flow shape that is altered, as discussed in detail in zone (from downdip water injection). The result-
Chapter 1.) ing wellbore enlargement caused by sand produc-
tion acts to stimulate production from the high-
This is the classic use of fracturing, to increase permeability zone. To stop sand production, draw-
the producing rate by bypassing near-wellbore
formation damage or by increasing exposure of
the formation area and thus stimulating well per- Stress (psi)
formance beyond that for no damage. For a single 4500 5500
well, treatment design concentrates on creating Rannoch-3
1820
the required formation flow area to yield increased
production at minimal cost. More formally, the Rannoch-3
design should optimize economic return on the
1840
basis of increased productivity and treatment cost. Rannoch-3
• Reservoir management
Along with improving well productivity, fractures Rannoch-2
1860
also provide a powerful tool for altering reservoir
flow. In combination with the other parts of field Rannoch-1
development, the fracture becomes a reservoir
management tool. For example, creating long 1880
fractures in tight rock (k < 0.1 md) enables field
development with fewer wells. However, even 0 20 40 60 80
fewer wells are required if the fracture azimuth TVD Fracture penetration (m)
(m below
is known and the wells are located appropriately sea level)
(e.g., not on a regulatory-required square pattern).
The actual philosophy shift for fracturing, from Figure 5-6. Fracturing for vertical inflow conformance.

down must be reduced even more. The production (as compared with a simple completion in which
is then essentially 100% water coming from the the entire zone is perforated).
stimulated high-permeability zone, and the well Another example of reservoir management is
must be abandoned. This further diminishes pro- waterflood development utilizing fractures and a
duction from the large reserves found in the “line drive” flood pattern (i.e., one-dimensional
deeper zones with lower permeability. [1D] or linear flow from injection fractures to pro-
Open- or cased hole gravel packing could be duction fractures). Knowledge of the fracture
used to eliminate the sand production. However, azimuth, combined with conductive fractures (or
such completions are less than satisfactory for two correctly controlled injection greater than the frac-
reasons. First, the deeper, lower permeability ture pressure) results in improved sweep efficiency
zones can significantly benefit from stimulation. and enables more efficient field development.
Second, significant scaling occurs with water
breakthrough and quickly plugs the gravel pack.
The fracturing tool selected to manage the 5-1.3. Design considerations and
Gullfaks field is termed an indirect vertical frac- primary variables
ture completion (IVFC). The IVFC accomplishes This section introduces the primary variables for frac-
several goals: ture design. Sidebar 5A summarizes how the design
– Some (although choked) production is achieved variables originate from treatment design goals.
from the main zone to enable the well to reach As mentioned previously, fracturing was con-
minimum productivity standards. trolled historically by operational considerations.
– Production from the lower, moderate-perme- This limited its application because fracturing is
ability zone is stimulated, maximizing reserves dominantly a reservoir process, and hence why a
from this zone. reservoir is fractured and what type of fracture is
– Greater drawdown is allowed because the weak required should be dominated by reservoir engineer-
high-permeability rock is separated from the ing considerations. The permeability k becomes the
perforations, and greater drawdown increases primary reservoir variable for fracturing and all
the total rate and significantly increases recov- reservoir considerations. Other, so-called normal
ery from the lower zones. reservoir parameters such as net pay and porosity
dominate the economics and control the ultimate
– If the upper high-permeability zone has sand viability of a project but do not directly impact how
production tendencies (as is typically the case), the fracturing tool is employed. As discussed in
then producing this zone via the fracture totally Chapter 12, postfracture productivity is also gov-
avoids the need for sand control. erned by a combination of the fracture conductivity
– Any potential for water breakthrough in the kf w and xf, where kf is the permeability of the prop-
high-permeability zone is retarded, and post- pant in the fracture, w is the propped fracture width,
water-breakthrough oil production is signifi- and xf is the fracture penetration or half-length.
cantly increased. These variables are controlled by fracturing and
To achieve these goals, fracture conductivity therefore identify the goals for treatment design.
must be tailored by synergy between the reservoir The productive fracture half-length xf may be less
and fracture models. Too much conductivity than the created (or the created and propped) half-
accelerates production and the time to water length L because of many factors (see Section 12-3).
breakthrough from the high-permeability main For example, the fracture width near the tip of a
zone. Also, too much conductivity, because of fracture may be too narrow to allow adequate
surface or tubular limits for the production rate, propped width. As another example, vertical varia-
restricts drawdown on the lower zones, and the tions in formation permeability, or layering, can
desired, more uniform vertical production profile cause the apparent productive length xf to be less
is not achieved. The fracture design goal is not to than the actual propped length (Bennett et al., 1986).
simply accelerate the rate but to achieve maxi- Similarly, this also makes the fracture height hf
mum reserves recovery with no sacrifice of rate important in several ways (Fig. 5-7):

5A. Design goals and variables
This discussion briefly summarizes the design goals of hydraulic fracturing that provide a road map for the major design variables.
Design goals
Design goals result from Darcy’s law (Eq. 5-2), in which the dimensionless term A/(∆xh) is defined by flow conditions and equals
ln(re /rw´ ) for steady-state flow (as discussed in Chapter 1). For steady-state flow, Prats (1961) showed that a fracture affects produc-
tivity through the equivalent wellbore radius rw´ and that rw´ is related to the fracture half-length or penetration xf by the dimension-
less fracture conductivity (CfD = kfw/kxf). Cinco-Ley et al. (1978) extended these concepts for transient flow with the relation among
xf, rw´ and CfD shown in Fig. 5-11 for pseudoradial flow (where the pressure-depletion region >> xf but is not affected by external
boundaries). Thus, the primary design goals are fracture half-length or penetration and the fracture conductivity kfw, with their rela-
tive values defined by CfD.
Design variables
Design variables result from material balance, rock mechanics and fluid mechanics considerations.
The material balance is (Eqs. 5-10 through 5-12)
Vf = Vi − VLp ; Vf = 2Lhf w , Vi = q i t p and VLp ≈ 6CLhLL t p + 4LhLS p , (5A-1)
where CL and Sp are fluid-loss parameters that can be determined by the results of a fluid-loss test (Fig. 5A-1) for which the filtrate
volume divided by the exposed area VL /A = Sp + 2CL√t . Combining the relations in Eq. 5A-1 gives Eq. 5-13:
q it p
L ≈ ,
6CLhL t p + 4hLS p + 2whf
where fracture penetration L is related to pump rate, fluid loss, height, width, etc.
Next is the elasticity equation (Eq. 5-14):
2p net d
w max = ,
E'
where pnet = pf – σc, and width is related to net pressure as a function of modulus and geometry and the pressure required to propa-
gate the fracture (Eq. 5-21):
p tip = (p − σ c ) at tip ∝ K Ic −apparent 1 / d , (5A-2)
where d is the characteristic fracture dimension and generally is the smaller dimension between hf and L.
Third is the fluid flow equation (Eqs. 5-15 through 5-19), in which Eq. 5-15 (dpnet /dx = 12µq/hfw3) is combined with the width equation:
1/ 4
 E' 3 
p net ≈  4 {κµq i L} + p net
4
tip 
, (5A-3)
 hf 
where the pressure drop down the fracture is related to viscosity, pump rate, fracture length (and thus to fluid loss), etc. The net
pressure distribution gives the fracture width distribution and thus the final propped fracture width (i.e., kfw). Hence the primary
design variables are CL, hL, Sp, hf, E ,́ KIc-apparent, qi , µ and σc .
Optimum design
The optimum design results from maximizing
Volume lost/area, VL /A
revenue $(rw´) minus the costs $(xf, kfw) by

using the preferred economic criteria.
2Cw
Sp
Figure 5A-1. Ideal laboratory fluid-loss data for spurt

loss Sp and the wall-building or filter-cake fluid-loss
coefficient Cw. If the total fluid loss is dominated by the √time
filter cake, then the total fluid-loss coefficient CL = Cw.

(a) rock mechanics considerations are related to the net

pressure pnet:
pnet = p f − σ c , (5-3)
Oil where pf is the pressure inside the fracture and σc is the
minimum in-situ stress (or fracture closure pressure).
For an ideal, homogeneous zone, closure pressure
is synonymous with the minimum in-situ stress.
However, such ideal conditions do not exist. Stress
(b) is a point value, and stress varies from point to point.
For realistic in-situ conditions, closure pressure
reflects the pressure where the fracture is grossly
closed, although the pressure may still be greater
Oil than the minimum in-situ stress at some points. For
zones that are only slightly nonhomogeneous, the
closure pressure represents a zone-averaged stress
over the fracture. However, other conditions may be
more complex. Consider the three-layer case of two
low-stress sandstone intervals with a thick interbed-
(c) ded shale. The correct closure pressure may be the
zone-averaged stress over the two low-stress zones,
without including the higher stress interbedded zone.
The fracture width is also of major importance for
Oil
achieving the desired design goals. Typically, this is
expressed as the product of fracture permeability
times fracture width; i.e., kf w is the dimensional con-
ductivity of the fracture. Figure 5-8 is an ideal well-
Water
bore/fracture connection for a propped fracture that
is intended to bypass near-wellbore formation dam-
age. To achieve the desired production goals, a nar-
Figure 5-7. The importance of fracture height.
row fracture must, at a minimum, carry the flow that
• In Fig. 5-7a, the fracture is initiated near the top

of the interval, and hf is not large enough to con-
tact the entire zone, which is clearly an important
reservoir concern.
• In Fig. 5-7b, the fracture grew out of the zone and
contacted mostly nonreservoir rock, diminishing xf rw
relative to the treatment volume pumped.
• In Fig. 5-7c, the fracture grew downward past the
oil/water contact and if propped would possibly
result in unacceptable water production.
In all these cases, as discussed in Section 5-4.2,
fracture height growth is controlled by rock mechan-
ics considerations such as in-situ stress, stress gradi-
ents, stress magnitude differences between different
geologic layers and differences in strength or frac- Figure 5-8. An ideal wellbore/fracture connection for a
ture toughness between different layers. All these propped fracture that is intended to bypass near-wellbore
formation damage.

would have been produced through the entire well- that many of these interactions will be contradictory
bore circumference (had there been no damage). The or incompatible. This is discussed later, but an exam-
fracture conductivity kf w must be greater than 2πrwk, ple is as follows. Consider a case where reservoir
where rw is the wellbore radius. For higher perme- goals require a long fracture. With deep penetration
ability formations that can deliver high rates with into the pay zone, getting good proppant transport
sufficient fracture permeability, fracture width and down a long fracture clearly requires high fluid vis-
any variables that affect width become important. As cosity. However, high viscosity increases the net
discussed in the following and in Section 6-2, width pressure inside the fracture. This reacts with the
is controlled by the fracture dimensions (hf and L), stress difference between the pay and the overlying
net pressure inside the fracture acting to open and and underlying shales and causes height growth,
propagate the fracture, and another property, the resulting in less penetration than desired, and thus
modulus or stiffness of the rock. less viscosity is required.
As implied by the term hydraulic fracturing, fluid Inherent contradictions controlling fluid selection
mechanics is an important element in fracturing. The abound:
two dominant fluid mechanics variables, injection
• Good viscosity is required to provide good prop-
(pump) rate qi and fluid viscosity µ, affect net pres-
pant transport, but minimal pipe friction is also
sure inside the fracture (and thus width) and largely
desirable to reduce surface pump pressure.
control transport and the final placement of proppant
in the fracture. These variables also have a role in • The fluid system is expected to control fluid loss,
controlling the volume of fluid lost to the formation but without damage to the formation or fracture
during pumping. For example, high pump rates permeability.
reduce the total fluid loss because for a given volume • Performance at high temperature, for long periods
pumped there is less time for fluid loss to occur. of time, is required from a fluid system that does
Another key factor of a good design is selection not cost much.
of the fluid and proppant systems with performance
characteristics (e.g., µ, CL, kf) that best meet the
requirements for the fracture treatment (i.e., material 5-2. In-situ stress
selection). In addition, the performance variables for In-situ stress, in particular the minimum in-situ stress
the materials must be properly characterized. Fluids (termed the fracture closure pressure for nonhomoge-
and proppants are addressed in Chapter 7, and their neous zones, as discussed earlier) is the dominant
performance is discussed in Chapter 8. parameter controlling fracture geometry. It is dis-
Finally, all the design parameters must be molded cussed in detail in Chapter 3. For relaxed geologic
to be compatible with existing well conditions (i.e., environments, the minimum in-situ stress is gener-
operational considerations). For example, it does lit- ally horizontal; thus a vertical fracture that formed
tle good to complain that the detailed design and when a vertical wellbore broke remains vertical and
analysis done in planning a treatment for an existing is perpendicular to this minimum stress. Hydraulic
well call for a high pump rate of 60 bbl/min when fractures are always perpendicular to the minimum
the wellbore conditions limit the maximum allow- stress, except in some complex cases, and even for
able pump rate to one-half that rate. Clearly, for new those cases any significant departure is only at the
wells the operational considerations (detailed in well. This occurs simply because that is the least
Chapter 11) should be an integral part of planning resistant path. Opening a fracture in any other direc-
for the drilling and completion process (e.g., well tion requires higher pressure and more energy.
trajectory for extended reach wells) (Martins et al., The minimum stress controls many aspects of
1992c). fracturing:
• At very shallow depths or under unusual condi-
5-1.4. Variable interaction tions of tectonic stress and/or high reservoir pres-
sure, the weight of the overburden may be the
It is clear that with major design considerations com-
minimum stress and the orientation of the
ing from multiple disciplines, the variables will
hydraulic fractures will be horizontal; for more
react, interact and interconnect in multiple ways and

normal cases, the minimum stress is generally hor- sure, the closure stress is typically between 0.6 and
izontal and the maximum horizontal stress direc- 0.7 psi/ft of depth (true vertical depth, TVD). More
tion determines whether the vertical fracture will generally, as discussed in Chapter 3, the minimum
run north–south, east–west, etc. stress is related to depth and reservoir pressure by
σ c ≅ Ko (σ v − pr ) + pr + T ,
• Stress differences between different geologic lay-
(5-4)
ers are the primary control over the important
parameter of height growth (Fig. 5-9). where Ko is a proportionality constant related to the
• Through its magnitude, the stress has a large bear- rock properties of the formations (possibly to both
ing on material requirements, pumping equipment, the elastic properties and the faulting or failure prop-
etc., required for a treatment. Because the bottom- erties), σv is the vertical stress from the weight of
hole pressure must exceed the in-situ stress for the overburden, pr is the reservoir pore pressure, and
fracture propagation, stress controls the required T accounts for any tectonic effects on the stress (for
pumping pressure that well tubulars must with- a relaxed, normal fault geology, T is typically small).
stand and also controls the hydraulic horsepower Ko is typically about 1⁄3. For fracture design, better
(hhp) required for the treatment. After fracturing, values are required than can be provided by such a
high stresses tend to crush the proppant and reduce simple relation, and methods of measuring or infer-
kf ; thus, the stress magnitude dominates the selec- ring the in-situ stress are discussed in Chapters 3 and
tion of proppant type and largely controls postfrac- 4. For preliminary design and evaluation, using
ture conductivity. Eq. 5-4 with Ko = 1⁄3 is usually sufficient.
Therefore, the detailed design of hydraulic fracture
treatments requires detailed information on in-situ 5-3. Reservoir engineering
stresses. An engineer must know the magnitude of
the minimum in-situ stress for the pay zone and As previously mentioned, because the ultimate goal
over- and underlying zones and in some cases must of fracturing is to alter fluid flow in a reservoir,
know the direction for the three principal stresses. reservoir engineering must provide the goals for a
For a simple, relaxed geology with normal pore pres- design. In addition, reservoir variables may impact
the fluid loss.
(a) (b) (c)
∆σ
σv
σh
0.9
0.8
0.7
0.6
pnet/∆σ
Shale 0.5
pnet
hf hfo 0.4
3 4 σH 0.3
2
0.2
0.1
0
1 1 2 3
h f /h fo
Shale
Figure 5-9. Fracture height growth. (a) Idealized fracture profile of the relation of fracture geometry to in-situ stresses.
σh = minimum horizontal stress, σH = maximum horizontal stress. (b) Typical fracture vertical cross section illustrating the
relation of the total fracture height hf to the “original” fracture height hfo. (c) Theoretical relation among hf /hfo, pnet and the
in-situ stress difference ∆σ (Simonson et al., 1978).

5-3.1. Design goals fracture and also assumes infinite conductivity. Prats
correctly accounted for the pressure distribution
Historically, the emphasis in fracturing low-perme- around a fracture and provided a general relation
ability reservoirs was on the productive fracture between dimensionless conductivity and rw´ for
length xf. For higher permeability reservoirs, the con- steady-state conditions (see Chapter 1). The relation
ductivity kfw is equally or more important, and the shows that for infinite-conductivity fractures, the
two are balanced by the formation permeability k. upper limit on rw´ is slightly less than that from the
This critical balance was first discussed by Prats flow area balance in Eq. 5-6. For infinite kfw, Prats
(1961), more than 10 years after the introduction of found
fracturing, with the important concept of dimension-
less fracture conductivity CfD: rw′ = 0.5 x f . (5-7)
kw
C fD = f . (5-5)
k xf Cinco-Ley et al. (1978) later integrated this into a
This dimensionless conductivity is the ratio of the full description of reservoir response, including tran-
ability of the fracture to carry flow divided by the
ability of the formation to feed the fracture. In gen- 5B. Highway analogy for dimensionless
eral, these two production characteristics should be fracture conductivity
in balance. In fact, for a fixed volume of proppant,
maximum production is achieved for a value of CfD A simplistic analogy for dimensionless fracture conductivity
CfD is a highway system. The numerator of this dimension-
between 1 and 2, as discussed in Chapters 1 and 10, less variable is kfw, which is the capacity of the highway or
with an analogy to highway design in Sidebar 5B. the ability of the highway to carry traffic. The denominator is
kxf ; this is the ability of the feeder roads to supply traffic to
Prats also introduced another critical concept, the the highway.
idea of the effective wellbore radius rw´. As shown in The famous old U.S. highway known as Route 66 ran, for
much of its length, across sparsely populated areas where
Fig. 5-10, a simple balancing of flow areas between feeder roads were few, narrow and far between. The ability
a wellbore and a fracture gives the equivalent value of the feeder road network to supply traffic to the highway
was limited (similar to the conditions existing when a prop-
of rw´ for a propped fracture (qualitative relation ped hydraulic fracture is placed in a formation with very low
only): permeability). In this case, the width, or flow capacity, of the
highway is not an issue (kfw does not have to be large).
2 What is needed (and was eventually built) is a long, narrow
rw′ ≈ xf . (5-6) (low-conductivity) highway.
π As a comparison, consider Loop 610, the “superhighway”
surrounding the city of Houston. The feeder system is locat-
However, this simple flow area equivalence ed in a densely populated area, and the feeder roads are
ignores the altered pore pressure field around a linear numerous and wide. Here, the width, or flow capacity, of the
highway is critical. Making this highway longer has no effect
;;;
on traffic flow, and the only way to increase traffic flow is to
widen (i.e., increase the conductivity of) the road. This is
obviously analogous to placing a fracture in a higher perme-
. . Flow area = 2πrwh
. .. . .. . .. . . ability formation, with the postfracture production limited by
.. .. .... . . . .. .. . . .. . the fracture width (or, more accurately, limited by kfw).
. .. . ..
. .. . .. . .. .. .. .. . If CfD is the ratio of the ability of a highway to carry traffic
. . .. . .. .. .. ... . .. .. . to the ability of the feeder system to supply that traffic to the
. . . . .. . .
highway, clearly a highway should be engineered to approxi-
mately balance these conditions. That is, a CfD value > 50 is
;;;
seldom warranted, because a highway would not be con-
structed to carry 50 times more traffic than the feeder system
Flow area = could supply. In the same way, a value of 0.1 makes little
2πrw´h 2
rw´ = x sense. Why construct a highway that can only carry 10% of
π f
the available traffic? In general, an ideal value for CfD would
be expected to be about 1 to result in a balanced, well-
engineered highway system.
A balance of about 1 is certainly attractive for steady-flow
traffic conditions that may exist through most of the day.
Flow area = 4xfh However, during peak traffic periods the feeder system may
supply more traffic than normal, and if this rush hour or tran-
sient traffic period is a major consideration, then a larger ratio
of CfD may be desirable. Thus, a CfD of 10 may be desirable
for peak flow (transient) periods, as opposed to a CfD value of
approximately 1 for steady-state traffic conditions.
Figure 5-10. Equivalent wellbore radius rw′.

sient flow. For pseudoradial flow, Cinco-Ley et al. ex- 5-3.2. Complicating factors
pressed rw´ as a function of length and CfD (Fig. 5-11).
The chart in Fig. 5-11 (equivalent to Prats) can be These principal concepts give a straightforward
used (when pseudoradial flow is appropriate) as a method for predicting postfracture production; how-
powerful reservoir engineering tool to assess possi- ever, complications can reduce postfracture produc-
ble postfracture productivity benefits from propped tivity below the levels expected or give better pro-
fracturing. For example, the folds of increase (FOI) ductivity than that calculated. The major complica-
for steady-state flow can be defined as the postfrac- tions include non-Darcy (or turbulent) flow, transient
ture increase in well productivity compared with flow regimes, layered reservoirs and horizontal per-
prefracture productivity calculated from meability anisotropy (particularly any natural fissure
permeability).
ln (re / rw ) + s For high-rate wells, non-Darcy or turbulent flow
FOI = , (5-8)
ln (re / rw′ ) can be an important factor that causes an increased
pressure drop along the fracture. This creates an
where re is the well drainage or reservoir radius, rw apparent conductivity that is less than the equivalent
is the normal wellbore radius, and s is any prefrac- laminar flow conductivity. The apparent CfD is also
ture skin effect resulting from wellbore damage, reduced and productivity is less than that expected.
scale buildup, etc. An equivalent skin effect sf result- Another complicating effect that can reduce produc-
ing from a fracture is tivity from expected levels is formation layering,
s f = − ln (rw′ / rw ) (5-9) where a fracture is in multiple layers with signifi-
cantly different values for porosity, permeability or
for use in reservoir models or other productivity cal- both. Unlike radial flow into a wellbore, average val-
culations. Equation 5-8 provides the long-term FOI. ues of permeability and porosity do not apply, and
Many wells, particularly in low-permeability reser- for layered formations, postfracture performance
voirs, may exhibit much higher (but declining) early- falls below simple calculations based on average per-
time, transient FOI. The preceding relations are for meability (Bennett et al., 1986). These and other
transient pseudoradial flow before any reservoir effects are discussed in Section 12-3.
boundary effects; the case for boundary effects is For lower permeability formations and for some
discussed in Section 12-2.6. time period, postfracture performance is dominated
by transient flow (also called flush production) as
discussed by Cinco-Ley et al. (1978). For transient
conditions, reservoir flow has not developed into
0.5 pseudoradial flow patterns, and the simple rw´ rela-
tions are not applicable. In the example in Fig. 5-12,
Effective well radius rw ´
CfD > 30, xf limited

Fracture half-length xf
0.2 rw´ = xf/2

pseudoradial flow did not develop until about 48
,
0.1 months. During the prior transient flow regimes,

0.05
8000 Gas well
CfD < 0.5, kfw limited
0.02 rw´ = 0.28 kfw/k 5000 k = 0.1 md, xf = 1000 ft
Numerical h = 50 ft, kfw = 200 md-ft
0.01 model
0.1 0.2 0.5 1 2 5 10 20 50 2000
Rate (Mcf/D)
Dimensionless fracture conductivity, CfD

1000 Radial flow
× FOI = 3.9
Figure 5-11. Equivalent wellbore radius as a function of 500
dimensionless fracture conductivity and fracture length.
Radial flow
200
100
12 24 36 48 60 72 84 96 108
Time (months)
Figure 5-12. Late development of pseudoradial flow.

productivity was better than that predicted from the where qi is the total injection rate and tp is the pump-
pseudoradial flow rw´. The duration of the transient ing time for a treatment. Equally simple, the fracture
flow period is a function of permeability, CfD and xf 2 volume created during a treatment can be idealized
such that for moderate- to high-permeability wells as
the period is too short to have practical significance
for fracture design. However, it may be important for Vf = h f × w × 2 L = η × Vi , (5-11)
postfracture well test analysis. For low-permeability –
wells with long fractures, transient flow may domi- where hf is an average, gross fracture height, w is the
nate most of the productive well life. average fracture width, L is the fracture half-length
or penetration, and η is the fluid efficiency. Finally,
as discussed by Harrington et al. (1973) and Nolte
5-3.3. Reservoir effects on fluid loss (1979), the volume lost while a hydraulic fracture
treatment is being pumped can be approximated by
Reservoir properties such as permeability to reservoir
fluid, relative permeability to the fracturing fluid fil-
trate, total system compressibility, porosity, reservoir VLp ≅ 6CL hL L t p + 4 LhL Sp , (5-12)
fluid viscosity and reservoir pressure all play a role in
where CL is the fluid-loss coefficient (typically from
fluid loss while pumping (see Section 6-4). Thus, cer-
0.0005 to 0.05 ft/min1/2), hL is the permeable or
tain reservoir information is required for treatment
fluid-loss height, and Sp is the spurt loss (typically
design, as well as for specifying design goals.
from 0 to 50 gal/100 ft2). Because material balance
must be conserved, Vi must equal VLp plus Vf, and
Eqs. 5-10 through 5-12 can be rearranged to yield
5-4. Rock and fluid mechanics
qi t p
Rock and fluid mechanics (along with fluid loss) L≅ , (5-13)
6CL hL t p + 4hL Sp + 2 wh f
considerations control the created fracture dimen-
sions and geometry (i.e., fracture height hf, length showing a general relation between several impor-
L and width w). These considerations all revolve tant fracture variables and design goals.
around the net pressure pnet given by Eq. 5-3. Modeling of hydraulic fracture propagation in
However, pnet, which controls hf and L, is itself a low- to medium-permeability formations typically
function of hf and L, and the various physical behav- shows an average width of about 0.25 in. (±50%)
iors connecting height, net pressure, width, etc., over a fairly wide range of conditions (e.g., Abou-
interact in many ways. This makes simple statements Sayed, 1984). Using this value, the effect of the pri-
about the relative importance of variables difficult or mary variables height hf and fluid-loss coefficient
impossible. However, the basic physical phenomena CL on fracture penetration L are investigated in
controlling fracture growth are understood and are Fig. 5-13. This is for a simple case of a constant
well established. 0.25-in. fracture width. Figure 5-13a shows length
as a strong, nearly linear function of hf; e.g., dou-
bling hf cuts fracture penetration by 50%. For similar
5-4.1. Material balance conditions, Fig. 5-13b shows that the fluid-loss coef-
The major equation for fracturing is material bal- ficient is not as important; e.g., doubling CL reduces
ance. This simply says that during pumping a certain L by only about 20%. However, with fracturing, such
volume is pumped into the earth, some part of that is simple relations are never fixed. As seen in Fig.
lost to the formation during pumping, and the 5-13c, for a higher loss case, doubling CL from 0.005
remainder creates fracture volume (length, width and to 0.01 reveals a nearly linear relation between CL
height). It is the role of fracture models to predict and L, just as for height in Fig. 5-13a. Basically, for
how the volume is divided among these three dimen- Figs. 5-13a and 5-13b, the loss term (first term in the
sions. The volume pumped is simply denominator of Eq. 5-13) is small compared with the
fracture volume term (third term in the denominator).
Vi = qi × t p , (5-10) Therefore, the fluid loss is relatively low and fracture

(a)
fluid efficiency (η, as defined in Eq. 5-11) is high. In
Fig. 5-13c, the loss term is large compared with the
1600 volume term (high loss and low efficiency), and the
CL = 0.001 ft/min1/2
1400 qi = 30 bbl/min loss coefficient becomes the dominant variable, with
–
w = 0.25 in. hf = hL = 100 ft L less sensitive to variations in hf or equivalently w if
1200 Sp = 0
it varies from the fixed value of 0.25 in.
1000
2:1
L (ft)
800
600 5-4.2. Fracture height
400 hf = hL = 200 ft Equation 5-13 demonstrates that fracture height hf
200 and fluid-loss height hL are important parameters for
0 fracture design. Loss height is controlled by in-situ
0 20 40 60 80 100 variations of porosity and permeability. Fracture
Time (min) height is controlled by the in-situ stresses, in particu-
lar by differences in the magnitude or level of stress
(b) between various geologic layers. More formally,
height is controlled by the ratio of net pressure to
1600
hf = hL = 200 ft stress differences ∆σ, as illustrated in Fig. 5-9, where
1400 q i = 30 bbl/min
w = 0.25 in.
∆σ is the difference between stress in the boundary
1200 Sp = 0 shales and stress in the pay zone. Ignoring any pres-
1000 sure drop caused by vertical fluid flow, the relation
CL = 0.0005 among fracture height, initial fracture height, pnet
L (ft)
800
and ∆σ can be calculated as demonstrated by
600
Simonson et al. (1978). This relation is included
400 CL = 0.001 in Fig. 5-9c.
200 For cases when pnet is relatively small compared
0 with the existing stress differences (e.g., less than
0 20 40 60 80 100 50% of ∆σ), there is little vertical fracture growth
Time (min) and the hydraulic fracture is essentially perfectly
confined. This gives a simple fracture geometry
(c)
(Fig. 5-14a) and increasing net pressure (Fig. 5-14b).
For cases when pnet is much larger than the existing
600
hf = hL = 100 ft stress differences, vertical fracture height growth is
qi = 30 bbl/min
500 w = 0.25 in. essentially unrestrained. Again, the geometry is a
Sp = 0
CL = 0.005 fairly simple radial or circular fracture (Fig. 5-14c)
400
and declining net pressure (Fig. 5-14b).
For more complex cases when pnet is about equal
L (ft)
300
1.8:1 to ∆σ, fracture geometry becomes more difficult to
200 predict, and significant increases in height can occur
CL = 0.01 for small changes in net pressure. Also, for this case,
100
the viscous pressure drop from vertical flow retards
0 fracture height growth (see Weng, 1991), and the
0 20 40 60 80 100 equilibrium height calculations in Fig. 5-9 are no
Time (min) longer applicable.
Figure 5-13. Effect of hf and CL on L.

(a) Depth
detail in Chapter 6), the slit opens into an elliptical
(ft) shape, with a maximum width
4900
2 pnet d
wmax = , (5-14)
∆σ = 1500 psi
E′
5000
where E′ is the plane strain modulus (E′ = E/(1 –
ν2), ν is Poisson’s ratio and typically equals about
5100 0.2), and d is the least dimension of the fracture.
pnet < 1/3∆σ
For a confined-height fracture with a tip-to-tip length
5200
greater than hf, d equals hf. This shows a direct rela-
3500 5000 100 200 300
tion between net pressure and width and introduces
Stress (psi) Fracture penetration (ft) an important material property, the plane strain mod-
ulus. However, because typically ν2 < 0.1, the plane
(b) strain modulus seldom differs from Young’s modulus
E by a significant amount.
2000
CL = 0.002
1000 qi = 20 bbl/min
∆σ = 1500 psi hfo = 100 ft
500 5-4.4. Fluid mechanics and fluid flow
Net pressure (psi)
E = 4E+6 psi
µ = 200 cp
200 The major fluid flow parameters are the fluid viscos-
∆σ = 50 psi
100 ity (resistance to flow) µ and injection rate qi. The
50 rate also effects the pump time and hence is impor-
tant to fluid-loss and material-balance considerations,
20
Shut-in as discussed previously. Both parameters are critical
10
1 2 5 10 20 50 100 200
for proppant transport, and both parameters also
Pump time (min) affect net pressure and thus affect fracture height
and width.
As an example, consider a Newtonian fluid flowing
(c)
Depth laterally through a narrow, vertical slit (i.e., fracture)
(ft) (Fig. 5-15). For laminar flow (the general case for
4800 flow inside hydraulic fractures), the pressure drop
∆σ = 50 psi
4900 along some length ∆x of the slit is
∆pnet 12µq
5000
= . (5-15)
5100
∆x hf w3
5200 Assuming a simple case of a long, constant-height

pnet > 4∆σ and -width fracture with two wings and zero fluid
5300
loss (i.e., the flow rate in each wing is q = qi /2) and
3200 3800 100 200 300 400 500 also assuming zero net pressure at the fracture tip,
Stress (psi) Fracture penetration (ft)
Figure 5-14. Relationship of pnet to stress differences.
q = qi ⁄ 2
– q
5-4.3. Fracture width w v =
hfw
Consider a slit in an infinite elastic media (i.e., the

earth). Also consider that the slit is held closed by
a fracture closure stress but is being opened by an
internal pressure equal to the closure stress plus a net
pressure pnet. Under these conditions (discussed in Figure 5-15. Fluid flowing laterally through a narrow verti-
cal fracture.

Eq. 5-15 is integrated from the fracture tip back to tion in Eq. 5-16 assumes no net pressure at the frac-
the wellbore to give ture tip; i.e., fracture tip effects or fracture propaga-
tion effects are ignored. When tip effects are taken
6µqi L
pnet = . (5-16) into account, the fracture width is affected by both
hf w3 fluid viscosity and tip effects (Shlyapobersky et al.,
1988a, 1988b). As shown by Nolte (1991), tip
For this long, confined-height fracture, hf is the
effects can be approximated by considering the net
minimum fracture dimension for Eq. 5-14, and the
pressure within the tip region to equal ptip (as
fracture width and net pressure are related by
opposed to zero) in Eq. 5-16. For a positive tip pres-
E ′w sure, the net pressure equation becomes
pnet = . (5-17)
2h f 1/ 4
 E′3 
Combining the two equations gives the proportionality pnet ≈  4 {κµqi L} + ptip
4
 , (5-20)
 h f 
µq L
1/ 4
w ∝  i  . (5-18) where ptip is the pressure required at the fracture tip

 E′  to open new fracture area and keep the fracture prop-
agating forward. This simple relationship serves to
The exponent of 1⁄4 for this simple fracture geome-
illustrate that there are always two components to net
try and for Newtonian fluids implies that the fracture
pressure: a viscous component and a fracture tip-
width is virtually constant; e.g., doubling the pump
effects component. The relative magnitude of the
rate from 20 to 40 bbl/min increases the width only
two effects varies from case to case, and because of
by about 20%. The same effect is found for all the
the small exponent, the combined effects are much
variables in Eq. 5-18. Generally, for non-Newtonian
less than the direct sum of the individual effects. For
fluids, the exponent is approximately 1⁄3.
example, when the viscous component and the tip
This relationship for fracture width can also be
component are equal, the net pressure is increased by
used with Eq. 5-17 to give net pressure expressed as
only 20% over that predicted when one of the com-
E ′ 3/ 4 ponents is ignored.
pnet = {κµqi L} ,
1/ 4
(5-19)
hf • Fracture toughness and elastic fracture mechanics
where κ is a constant (see Eq. 6-11) to provide an The fracture tip propagation pressure, or fracture
equality for this expression. tip effect, is generally assumed to follow the
Thus, as a result of viscous forces alone, net pres- physics of elastic fracture mechanics. In that case,
sure inside the fracture develops as a function of the the magnitude of the tip extension pressure ptip is
modulus, height and (qµ)1/4. From the nature of this controlled by the critical stress intensity factor KIc
relation, however, it is clear that modulus and height (also called the fracture toughness). Fracture
are much more important in controlling net pressure toughness is a material parameter, and it may be
than are pump rate and viscosity, the effect of which defined as the strength of a material in the pres-
is muted by the small exponent for the relation. ence of a preexisting flaw. For example, glass has
a high tensile strength, but the presence of a tiny
scratch or fracture greatly reduces the strength
5-4.5. Fracture mechanics and fracture (i.e., high tensile strength but low fracture tough-
tip effects ness). On the other hand, modeling clay has low
strength, but the presence of a flaw or fracture
The fluid mechanics relations show pnet related
does not significantly reduce the strength. Lab-
to modulus, height, fluid viscosity and pump rate.
oratory-measured values for the material property
However, in some cases, field observations have
KIc show toughness ranging from about 1000 to
shown net pressure (and presumably fracture width)
about 3500 psi/in.1/2, with a typical value of about
to be greater than predicted by Eq. 5-19 (Palmer and
2000 psi/in.1/2. These tests (after Schmidt and
Veatch, 1987). In such cases the fluid viscosity has a
Huddle, 1977; Thiercelin, 1987) include a range
smaller effect on fracture width than predicted by
of rock types from mudstones and sandstones to
Eq. 5-19. This is probably because the simple rela-

carbonates and consider confining pressures from phenomena. Tip phenomena are discussed in
0 to 5000 psi. detail in Chapters 3 and 6.
From elastic fracture mechanics, for a simple Measured values for tip extension pressure
radial or circular fracture geometry with a pene- that are higher than predicted from laboratory-
tration of L, the fracture tip extension pressure is measured rock toughness KIc can be accounted
for in hydraulic fracture calculations through the
π
ptip = K Ic , (5-21) use of the effective, or apparent, fracture tough-
48 L ness KIc-apparent (Shlyapobersky, 1985). In practice,
because KIc-apparent is not a material constant, the
and it decreases as the fracture extends. For even
tip effects should be defined or calibrated by frac-
a small fracture penetration of 25 ft, this gives a
turing pressure data for a particular situation (see
tip extension pressure of 29 psi, whereas viscous
Sidebar 9B).
pressures (Eq. 5-19) are typically 10 or more
times larger. Thus normal linear elastic fracture
mechanics considerations indicate that fracture 5-4.6. Fluid loss
mechanics, or the tip extension pressure, generally
plays a negligible role for hydraulic fracturing. As seen from the material balance (Eq. 5-13), fluid
• Apparent fracture toughness loss is a major fracture design variable characterized
by a fluid-loss coefficient CL and a spurt-loss coeffi-
Field data typically show fracture extension pres- cient Sp. Spurt loss occurs only for wall-building flu-
sure to be greater than that given by Eq. 5-21, ids and only until the filter cake is developed. For
with 100 to 300 psi as typical values and even most hydraulic fracturing cases, the lateral (and ver-
higher values possible. This difference is due to tical) extent of the fracture is much greater than the
several behaviors not included in elastic fracture invasion depth (perpendicular to the planar fracture)
mechanics calculations. One important (and long- of fluid loss into the formation. In these cases, the
recognized) consideration is that the fracturing behavior of the fluid loss into the formation is linear
fluid never quite reaches the fracture tip; i.e., there (1D) flow, and the rate of fluid flow for linear flow
is a “fluid lag” region at the tip that increases the behavior is represented by Eq. 5-1.
apparent toughness and tip pressure (Fig. 5-16). This assumption of linear flow fluid loss giving the
In other cases, tip pressure may be even greater. CL ⁄√t relation has been successfully used for fractur-
Other tip phenomena include nonelastic rock ing since its introduction by Carter (1957). The rela-
deformation near the fracture tip and tip plugging tion indicates that at any point along the fracture, the
with fines, with these mechanisms acting alone or rate of fluid loss decreases with time, and anything
in conjunction with the fluid flow and/or fluid lag that violates this assumption can cause severe prob-
lems in treatment design. For example, fluid loss to
Closure stress natural fissures can result in deep filtrate invasion
into the fissures, and the linear flow assumption may
no longer be valid. In fact, for the case of natural fis-
sures if net pressure increases with time, the fluid-
loss rate can increase, and treatment pumping behav-
ior may be quite different from that predicted. The
Fluid
lag total fluid loss from the fracture is controlled by the
total fluid-loss coefficient CL, which Howard and
Fast (1957) decomposed into the three separate
p1 p2
mechanisms illustrated in Fig. 5-17 and discussed
in Section 6-4.
The first mechanism is the wall-building character-
p1 = fracture pressure istics of the fracturing fluid, defined by the wall-
p2 ≤ reservoir pressure building coefficient Cw. This is a fluid property that
helps control fluid loss in many cases. For most frac-
Figure 5-16. Unwetted fracture tip (fluid lag). turing fluid systems, in many formations as fluid loss

Invaded Reservoir Closure

For fluid to leak off from the fracture, the reservoir
zone control pressure fluid must be displaced. This sets up some resistance
+ to fluid loss, and this reservoir effect is characterized
pnet
by the compressibility coefficient Cc. As discussed in
Pressure
Chapter 6, the parameter for this calculation is gov-

erned by a pressure difference ∆p between the pres-
Reservoir pressure
sure inside the fracture (i.e., closure pressure + pnet)
and the reservoir pressure, permeability to the mov-
able formation fluid k, total system compressibility
Distance into formation for the reservoir ct, and the viscosity of the reservoir
fluid (gas or oil) µ. This parameter is usually more
important for a liquid-saturated reservoir (low com-
pressibility and relatively high reservoir fluid viscos-
ity) and when a filter cake does not develop.
Each of these three mechanisms provides some
resistance to fluid loss, and all three act as resistors
in series (although the fluid-loss coefficient itself is
defined in terms of conductance, or the inverse of
resistance). The three mechanisms variously com-
Filter cake
bine in different situations to form the total or com-
Figure 5-17. The three regions of fluid loss. bined fluid-loss coefficient CL, which is used for
fracture design (see Chapter 6). This clearly complex
occurs into the formation, some of the additives and situation makes it desirable to measure fluid loss
chemicals in the fluid system remain trapped on or from field tests (just as permeability must be mea-
near the formation face, forming a physical filter- sured from field flow, buildup tests or both) when-
cake barrier that resists fluid loss. ever possible (see Chapter 9).
Outside of the filter cake is the invaded zone,
which is the small portion of the formation that has
been invaded by the fracturing fluid filtrate. This
5-4.7. Variable sensitivities and
mechanism is the filtrate effect, or invaded zone
interactions
effect, and it is characterized by the viscosity or rela- The complexity of hydraulic fracture design comes
tive permeability control coefficient Cv. As discussed from the interactions of the major design variables
in Chapter 6, Cv can be calculated, and this parame- (hf, E, ∆σ, KIc and CL) and that different variables
ter is governed by the relative permeability of the affect different aspects of fracturing in different ways.
formation to the fracturing fluid filtrate kfil, the pres- As discussed in Section 5-4.1 concerning the sensitiv-
sure difference ∆p between the pressure inside the ity of fracture penetration to hf and CL, the impor-
fracture (i.e., closure pressure + pnet) and the reser- tance of various variables can change from case to
voir pressure, and the viscosity of the fracturing fluid case. Several examples of this are discussed here.
filtrate µfil. This mechanism is usually most impor-
• Net pressure
tant in gas wells, where the invading fluid has much
higher viscosity than the reservoir fluid being dis- The magnitude of net pressure for a specific frac-
placed, or where relative permeability effects pro- ture treatment is a major concern, because the
duce a filtrate permeability that is much less (<k/10) ratio of net pressure to stress differences between
than the permeability to the reservoir fluid. Other the pay zone and bounding zones controls fracture
cases are where a clean fluid is used such that no fil- height. Also, net pressure directly controls width.
ter cake develops or for fracturing high-permeability However, what controls net pressure varies signif-
wells where no filter cake develops and high-viscos- icantly from case to case.
ity crosslinked gel may be lost to the formation (i.e., In the case of hard-rock formations (i.e., forma-
µfil is very high). tions with values for Young’s modulus of 2 × 106
psi or greater) with height confinement and for

treatments pumping viscous fluids at normal frac-

(a)
turing rates, the viscous term of the net pressure
equation dominates any fracture tip effects. 2000
Toughness or tip effects become important for µ = 200 cp
1000
cases where fracture height is unconfined (e.g., µ = 50 cp
radial or circular fractures) or for very soft rocks 500
pnet (psi)
(e.g., formations such as unconsolidated sands 200
with E ≤ 0.5 × 106 psi). For treatments using low-
100
viscosity fluid or pumping at very low rates, the Nolte-Smith plot
50 ∆σ = 1000 psi
viscous term of the net pressure equation becomes
qi = 15 bbl/min
small, and fracture toughness becomes a dominant 20 hf = 25 ft
E = 5 × 106 psi
parameter. Although many cases fall into one of
0
these extremes, neither effect should be over- 0 0.2 0.5 1 2 5 10 20 50 100 200 500
looked for the prudent application of fracturing. Pump time (min)
The magnitude of net pressure may also be con-
(b)
trolled by in-situ stress differences between the
Depth
pay and the bounding layers. Consider a case (ft)
where barrier zones (e.g., formations with higher
closure stress) surround the pay zone (Fig. 5-9)
and further assume that because of either viscous
4900
or toughness effects, pnet increases to the level of
the stress differences. Massive height growth then Perforated
begins, and only very small increases in the net interval
pressure are possible. Net treating pressure is now 5000

controlled directly by ∆σ and is essentially inde-
pendent of both fluid viscosity and apparent frac-
ture toughness effects. This case is illustrated in 5100
the next section.
• Fracture height and net pressure
For a fracture with significant stress barriers and
in a formation with a medium to high value for 3500 4000 4500 –0.2 0 0.2
In-situ stress (psi) Width (in.)
the modulus, the viscous term in Eq. 5-20 controls
the net treating pressure. In such a case, pnet Figure 5-18. Height growth example in a hard-rock forma-
becomes a strong function of fracture height. tion.
However, as illustrated in Fig. 5-9, fracture height
hf is controlled by net pressure. To put it in anoth-
er form, fracture height is a function of fracture 1000 psi (i.e., ∆σ is controlling pnet), and exten-
height. As discussed in Chapter 6, this is where sive height growth occurs. Because ∆σ is control-
fracture models become important. ling the allowable net pressure, increasing the
As an example, consider the case of a thin fluid viscosity fourfold has essentially no effect on
(h = 25 ft) sandstone pay zone in a hard-rock for- net pressure after the first few minutes. The verti-
mation (E = 5 × 106 psi). Further assume that this cal fracture width profile plotted in Fig. 5-18b
zone is surrounded by shales with an in-situ stress shows that for pnet about equal to ∆σ, fracture
1000 psi greater than the stress in the sand, mak- width in the bounding layers may be too small for
ing them what would normally be considered proppant admittance. This is discussed in the sub-
good barriers to vertical fracture growth. As seen sequent section on proppant admittance.
in Fig. 5-18a, even for pumping a moderate (50-cp) Now consider the same case but with a 50-ft
viscosity fluid at a moderate rate, net pressure thick sandstone section. As seen in Fig. 5-19, pnet
immediately jumps to a level slightly greater than stays below ∆σ for the 50-cp fluid case and little

(a)
2000
µ = 200 cp
1000 2000 Nolte-Smith plot
Confined height
500 1000 Tip-dominated behavior
µ = 50 cp
pnet (psi)
200 500
µ = 50 cp
pnet (psi)
100 200
∆σ = 1000 psi
50 q = 15 bbl/min 100
i
hf = 50 ft
20 E = 5 × 106 psi 50
∆σ = 1000 psi
0 q = 15 bbl/min
20 hi = 25 ft
0 0.2 0.5 1 2 5 10 20 50 100200 500 f
E = 5 × 106 psi
Pump time (min)
0.2 0.5 1 2 5 10 20 50 100 200 500
Figure 5-19. Height growth example in a thicker hard-rock Pump time (min)
formation. (b)
Depth
(ft)
height growth occurs. For a more viscous (200-cp)
fluid, net pressure again approaches the stress dif-
ference of 1000 psi, and again extensive height 4900
growth occurs. These examples show that fracture Perforated
interval
height is a function of fracture height. 5000
Finally, consider the original (h = 25 ft) case
again, but assume this is a soft-rock (unconsoli-
dated sand with E < 0.5 × 106 psi) zone. Further 5100
assume that because of high permeability, fluid

loss is much greater than for the previous two
3500 4500 100 200 300 400
cases. Figure 5-20 shows pnet is much less than
In-situ stress (psi) Fracture penetration, L (ft)
∆σ, with essentially no height growth. Also, the
flat nature of the net pressure behavior in the
Figure 5-20. Height growth example in a soft-rock
Nolte-Smith log-log plot of pnet versus time indi- formation.
cates that fracture tip effects are dominating net
pressure behavior, as expected from Eq. 5-20. permeability formation it is probable that the frac-
Chapter 9 discusses net pressure behavior and the turing fluid cannot build a filter cake to control
means to determine the controlling conditions. fluid loss, and the only fluid-loss control will
• Fluid viscosity come from the viscosity (or invaded zone) effect
Fluid viscosity provides an example of how vari- Cv (see Section 5-4.6). Viscosity is therefore a
ables affect different parts of the fracturing pro- major factor for fluid selection, despite having no
cess in different ways. Consider a case of radial effect on geometry and not being critical for prop-
fracture growth in a soft rock (E < 1 × 106 psi). pant transport.
Toughness dominates pnet and fracture width, and
viscosity becomes unimportant in controlling frac-
ture geometry. However, viscosity can remain a 5-5. Treatment pump scheduling
critical consideration for proppant transport if a The fracture design process involves reservoir engi-
long fracture is desired and for fluid-loss control. neering to define the xf and kfw goals. It involves
Further assume this case is a very high perme- rock mechanics to consider the possibility of obtain-
ability formation, such that only a short fracture is ing a desired fracture geometry. It includes fluid
required. Thus, high viscosity is not required for mechanics considerations to confirm that the
proppant transport. However, in this very high required proppant transport is possible and rheology
to determine if the required fluid properties are pos-

sible. It also includes material selection and on-site 5-5.2. Pad volume
operational considerations as discussed in Section
5-6. The product of this process is a treatment pump For a treatment using viscous fluid, the fluid carries
schedule. This includes the pad volume necessary the proppant out to the fracture tip. For these cases
to create the desired fracture penetration, along with the pad volume determines how much fracture pene-
acid or proppant scheduling to achieve the desired tration can be achieved before proppant reaches the
postfracture conductivity. For propped fracturing, tip and stops penetration in the pay zone. Once the
pump scheduling includes fluid selection, proppant pad is depleted, a fracture may continue to propagate
selection, pad volume, maximum proppant concen- into impermeable layers until the proppant bridges in
tration to be used and a proppant addition schedule. low-width areas. Thus, pumping sufficient pad to
After the design goals and variables are defined, the create the selected length is critical. For treatments
proppant addition schedule is usually obtained by using very low viscosity fluid (i.e., “banking”-type
using a fracture simulator, although for many cases treatments), proppant settles out of the fluid and
analytical calculations based on fluid efficiency are essentially replenishes the pad. The pad volume must
also easily implemented. Chapter 10 provides addi- only be sufficient to open enough fracture width for
tional detail for treatment design. proppant admittance, and the carrying capacity of the
fluid, as opposed to the pad volume, determines the
final propped length.
5-5.1. Fluid and proppant selection On the other hand, too much pad can in some
instances be even more harmful, particularly for
Fracturing materials are discussed in Chapter 7, and cases requiring high fracture conductivity. The frac-
their performance characterization is discussed in ture tip continues to propagate after pumping stops,
Chapter 8. The major considerations for fluid selec- leaving a large, unpropped region near the fracture
tion are usually viscosity (for width, proppant trans- tip. Significant afterflow can then occur in the frac-
port or fluid-loss control) and cleanliness (after flow- ture, carrying proppant toward the tip and leaving
back) to produce maximum postfracture conduct- a poor final proppant distribution. This afterflow
ivity. Other considerations that may be major for occurs because the widest section of the fracture is
particular cases include near the wellbore at shut-in, and most of the prop-
• compatibility with reservoir fluids and reservoir rock pant pumped is stored there. However, the highest
• compatibility with reservoir pressure (e.g., foams fluid-loss rates are near the fracture tip. Thus, prop-
to aid flowback in low-pressure reservoirs) pant-laden slurry continues to flow toward the tip of
the fracture. Afterflow continues until either the frac-
• surface pump pressure or pipe friction considerations ture closes on the proppant, stopping proppant move-
• cost ment, or until proppant-laden slurry reaches the frac-
• compatibility with other materials (e.g., resin- ture tip. At that point the slurry dehydrates and stops
coated proppant) any additional fracture propagation. Ideally, of
• safety and environmental concerns (see Chapter 11). course, it is better to have the proppant at the frac-
ture tip at shut-in and thus minimize afterflow.
Proppant selection must consider conductivity at An ideal schedule for a normal treatment (as
in-situ stress conditions (i.e., the effect of stress on opposed to subsequently discussed TSO designs) is
proppant permeability kf). Proppant size must also be one where the pad depletes and proppant reaches the
considered. In general, bigger proppant yields better fracture tip just as the desired fracture penetration is
conductivity, but size must be checked against prop- achieved and also just as pumping stops. This is the
pant admittance criteria, both through the perforations sequence in Figs. 5-2, 5-3 and 5-4.
and inside the fracture (see Section 5-5.4). Finally, The critical parameter of the pad volume or pad
the maximum in-situ proppant concentration at shut- fraction fpad is related directly to the fluid efficiency
in must be selected, as it determines how much of the for a treatment (Nolte, 1986b). This relation from
hydraulic width created by the fracture treatment will Sidebar 6L gives the pad volume expressed as
be retained as propped width once the fracture closes. a fraction of the entire treatment volume:

1− η
f pad ≈ . (5-22) 1.0
1+ η % of fracture filled by proppant pack
Pack porosity = 0.35
0.8
That is, a treatment with an expected efficiency η
of 50% would require a pad fraction of about 1⁄3. As
Fill fraction, F
0.6
discussed in Chapter 9, the efficiency for a specific Proppant s.g. = 2.65
(sand)
formation and fluid system can be determined by a 0.4
calibration treatment. Proppant s.g. = 3.2
(intermediate strength)
This discussion of pad volume has so far concen- 0.2
trated on the fluid-loss aspects of the pad volume;
i.e., the pad is pumped first to serve as a sacrificial 0.0
0 4 8 12 16 20 24
stage of the treatment to enable the fracture to pene-
Proppant concentration (lbm/gal)
trate into permeable formations. This important effect
of the pad volume may be the critical aspect govern- Figure 5-21. Fill fraction versus proppant concentration.
ing the size of the pad for most applications. How-
ever, hydraulic fracturing is complicated, in that most
things are done for at least two reasons, which where F is the fill fraction (Fig. 5-21), the con-
applies to pad volume specification. The second pur- stant 8.33 converts the units to lbm/gal, γprop is
pose of the pad volume is to create sufficient fracture the specific gravity (s.g.) of the proppant, C is the
width to allow proppant to enter the fracture (see final in-situ proppant concentration at shut-in
Section 5-5.4 on proppant admittance). Even for a expressed as pounds of proppant per fluid gallon
case of very low fluid loss, some minimum pad vol- (ppg), and φ is the porosity of the proppant pack,
ume is required. Both of these aspects of the pad vol- typically about 0.35.
ume must always be considered for treatment design. Increasing the concentration from 8 (F ≈ 0.4)
• Propped width to 16 ppg (F ≈ 0.6) significantly increases the
propped fracture width (50% increase in the fill
A major design goal is fracture conductivity kf w,
fraction). However, this large increase in propped
which consists of proppant pack permeability and
width is accomplished at the expense of additional
propped fracture width. Proppant permeability kf
risk to the job and to the well, because of either
is a function of the proppant selected, in-situ
surface mechanical problems or an unexpected
stress and residual damage from fluid additives
total screenout somewhere in the fracture or in the
(see Chapter 8). Propped width is controlled by
near-wellbore region between the well and the far-
the treatment design.
field fracture (see the discussion of tortuosity in
The effective propped width wp-eff is a function
Section 6-6). In practice, most treatments use a
of the average fracture width wf at shutdown (i.e.,
maximum concentration of about 8 to 14 ppg,
hydraulic width at the end of pumping a treat-
although concentrations of 20 ppg have been
ment), proppant concentration C in the fracture at
pumped.
that time (i.e., giving the ideal propped width wp)
Another manner of increasing propped width is
and the volume of proppant wlost that is lost on the
to increase fracture width. Theoretical and numer-
faces of the fracture to embedment, gel residue,
ical models generally show that the fracture
etc. (usually expressed as lbm/ft2 “lost”). In terms
width, while the fracture is growing, is relatively
of these parameters, the effective propped width
insensitive to the controllable job variables of
can be expressed as
pump rate and fluid viscosity. For a simple frac-
w p−eff = w p − wlost = w f × F − wlost (5-23) ture geometry, width is proportional to rate and
viscosity raised to a small power. For Eq. 5-18
with the exponent 1⁄4, doubling the pump rate
C
F= increases fracture width by only 18%, at the
( )
,
8.33 × γ prop + C × (1 − φ)
(5-24)
expense of significant pipe friction and surface

pressure increases. Viscosity is easily increased by ture (see Section 6-5). First is what may be termed
an order of magnitude (e.g., 10 times increase in µ simple or single-particle settling. Behavior of this
increases the width by 77%), but only at the type is governed by Stokes law, in which the veloc-
expense of using more fluid additives and with ity of a single particle falling through a liquid medi-
additional conductivity damage potentially negat- um is
ing the extra width.
Thus, the hydraulic fracture width is fairly fixed v fall = 1.15 × 10 3
d 2prop
µ
( )
γ prop − γ fluid , (5-25)
(±50%, at least in terms of the treatment’s control-
lable parameters), and the proppant fill fraction has where vfall is the settling rate in ft/s, dprop is the aver-
a practical limit of about 0.5 (±0.1). Without TSO age proppant particle diameter in in., µ is the fluid
designs (discussed in the following) the final, viscosity in cp, and γprop and γfluid are the specific
effective propped width is almost fixed by nature. gravity of the proppant and the fluid, respectively.
The goal for a normal fracture design is then to The settling rate, and thus the efficiency with which
achieve a required kf w within these limits, with proppant can be transported into the fracture, is
proppant concentration, proppant selection and directly related to the fluid viscosity. This is usually
fluid selection allowing a large range of values. the main consideration for how much viscosity is
• Tip-screenout designs required for a fracture treatment. However, there are
As mentioned previously, as long as a fracture is additional considerations for calculating settling fol-
free to propagate, the hydraulic fracture width is lowing Stokes law. At low proppant concentrations
relatively insensitive to the controllable treatment (e.g., less than 1 or 2 ppg) particles may clump, pro-
parameters of fluid viscosity and pump rate. If ducing an apparent diameter greater than the actual
more conductivity is required than can be particle diameter and accelerating settling. Higher
achieved from a normal design, the only effective particle concentrations act to increase the slurry vis-
manner to increase the propped width is to stop cosity and retard settling (also known as hindered
the fracture from propagating but to continue to settling). The pump rate is also an important param-
pump. This technique has come to be called TSO eter controlling proppant transport for simple settling
fracturing (Smith et al., 1984). by Stokes law.
For a normal treatment, the pad volume is As shown in Fig. 5-22, for a Newtonian fluid the
designed to deplete just as pumping stops. What distance D a proppant particle is transported into a
would happen if pumping simply continued fracture, before that particle can fall from the top of
beyond that time? If the pad is depleted, then
proppant-laden slurry will be located everywhere
around the fracture periphery. If there is fluid loss,
then this slurry will dehydrate and leave packed
proppant around the periphery. Even with no fluid v1
loss, the proppant may bridge in the narrow frac- h
ture width around the periphery, particularly in vfall
places where the width is extremely narrow as a
result of the fracture penetrating a boundary layer.
In either case, any additional propagation is D
restricted and further pumping causes an increase
D ⁄ h = v1 /v 2
of net pressure and thus an increase of fracture
width. TSO designs are discussed in detail in v1 = fluid velocity
Chapter 10. ∝ qi/hw ∝ qi/h(µqi)1/4
∝ qi3/4/hµ1/4
vfall = fall rate ∝ 1/µ
5-5.3. Proppant transport D/h ∝ (qiµ)3/4/h
D is independent of h
Several modes of proppant settling can occur while
proppant is being transported into a hydraulic frac- Figure 5-22. Stokes law.

the fracture to the bottom, is related to (qiµ)3/4. This fluid (which describes most fracturing fluid systems)
distance is independent of the fracture height and, flowing down a channel imparts a normal force to
more significantly, shows that for some given trans- particles entrained in the fluid such that the particles
port distance, less viscosity can be used at higher tend to migrate to and concentrate in the center of
pump rates. This relation can be important for higher the channel. For low average concentrations, this can
temperature applications, where fluid viscosity can result in a center core of high-proppant-concentration
degrade significantly with time. At higher rates (and slurry, with a region of essentially clean fluid on
hence shorter pump times), less viscosity is required either side. This heavier core of concentrated slurry
for proppant transport. Also, the fluid is exposed to tends to fall owing to its greater density, carrying the
the high formation temperature for less time, so the entrained proppant toward the bottom of the fracture
fluid system maintains better viscosity. In general, at a faster rate than for a dispersed slurry (Nolte,
considering how fluid viscosity degrades down a 1988b).
fracture, including the effect of proppant concentra- Finally, any calculations for proppant settling must
tion increasing the effective slurry viscosity, and consider geologic reality. Detailed examinations of
considering the non-Newtonian nature of most hydraulic fractures both at the wellbore using televi-
fracturing fluids, if a fracturing fluid retains 50- to sion cameras (Smith et al., 1982) or away from wells
100-cp viscosity (at reservoir temperature and at a in mineback tests (see Warpinski, 1985) show some-
shear rate of 170 s–1) at the end of the fracture treat- thing other than the smooth fracture walls assumed
ment, it will provide essentially perfect proppant for settling calculations. Small shifts and jogs of the
transport (Nolte, 1982). fracture probably have no significant impact on fluid
The next mode of proppant settling is termed con- flow or on lateral proppant transport into the frac-
vection, and it was probably first included in fracture ture. However, these small irregularities could signif-
modeling in the context of a fully three-dimensional icantly impact settling. Calculations for proppant set-
(3D) planar model by Clifton and Wang (1988). This tling that ignore these effects will be a worst-case
type of settling is controlled by density differences scenario.
(i.e., buoyancy) between two fluids. For example,
a proppant-laden fluid stage with an 8-ppg concen-
tration has a slurry density of 11.9 lbm/gal (s.g. = 5-5.4. Proppant admittance
1.44). If this slurry is placed directly next to a clean Proppant admittance is critical to hydraulic fracturing
fluid stage with a density of 8.5 lbm/gal (s.g. = 1.02), in two forms: entrance to the fracture through perfora-
the heavier slurry will tend to sink and underride the tions and entrance of proppant into the fracture itself.
lighter clean fluid, simply carrying the proppant These effects were recognized early, and the original
toward the bottom of the fracture. However, a treat- fracture width models were used primarily for deter-
ment does not normally follow clean pad fluid with mining a pad volume that would allow admittance by
a heavy 8-ppg slurry. Rather, the treatment increases generating a fracture width greater than 2.5dprop,
proppant concentration slowly to account for fluid- where dprop is the average proppant particle diameter.
loss effects and mitigate convection effects. Only Before these models, operators were reluctant to
near the end of pumping (when the need for trans- pump significant volumes of pad as it was considered
port decreases), when the initial proppant stages expensive and potentially damaging.
have undergone significant dehydration, can a signif- The laboratory data in Fig. 5-23 (Gruesbeck and
icant density difference begin to develop. In general, Collins, 1978) illustrate two important ideas:
rigorous numerical modeling of this phenomena
shows convection is not a major factor during pump- • A minimum perforation diameter is required for
ing (Smith et al., 1997). If excessive pad is used, proppant to flow through the perforations.
such that a large unpropped region of the fracture • Minimum perforation diameter is a function of the
exists after shut-in, convection can occur during the slurry concentration.
shut-in after flow, with potentially significant adverse At low concentrations (e.g., 1 ppg), the perforation
effects on the final proppant placement. hole diameter must be only slightly greater than that
The third effect on proppant transport is termed of the proppant particles. The required hole diameter
migration (see Chapter 6). In brief, a viscoelastic increases with concentration until at about 6 ppg

Maximum particle concentration (vol/vol) Table 5-1. Proppant admittance criteria.

8 0.185 0.312 0.405
Proppant†
7 Concentration
—
average particle diameter
(lbm proppant/gal fluid) w /dprop

Perforation diameter/
6
Experimental Correlation
5 Bridge Formation‡ Bridge
100-cp HEC solution
4 0.5 to 2 1.8 1.15 to 2.0
Tap water
3
2 to 5 2.2 2.0 to 3.0
2
5 to 8 2.6 3.0
1 † Sand as proppant
‡ Data from van der Vlis et al. (1975)
0
0 5 10 15 20 25 30 35
Sand concentration
(lbm sand/gal fluid)
6 to 8 ppg, the required average fracture width
Figure 5-23. Proppant admittance through perforations increases to 3dprop.
(Gruesbeck and Collins, 1978). This critical width is important to the hydraulic
fracturing process. Should proppant enter a part of the
fracture where sufficient width does not exist, the
(solid volume fraction of about 0.20), the perforation proppant will bridge and no longer flow down the
hole diameter must be 6 times the average particle fracture. Additional slurry flowing in this direction
diameter. will cause proppant to pile up, dehydrate and block
This same trend applies for slurry flow down a nar- that part of the fracture. Should this occur near the
row fracture. An approximate proppant bridging or wellbore, possibly as a result of some form of near-
proppant admittance criteria can be derived by calcu- wellbore width restriction (see tortuosity discussion in
lating an equivalent hydraulic radius for a narrow Section 6-8), a total screenout can result with serious
slot, rhyd = w/2, where w is the average width of the consequences for the success of the fracture treatment.
fracture. For a round perforation hole, the hydraulic
radius is d/4, where d is the perforation hole diameter.
Equating the two hydraulic radius values shows that 5-5.5. Fracture models
2w is equivalent to the diameter of a round hole.
Using this along with two lines fitting the data Clearly, developing a final treatment pump schedule
of Gruesbeck and Collins leads to an approximate must consider many options. The interactive roles of
admittance criteria for a hydraulic fracture: the various major variables (hf, E, CL, KIc-apparent, µ
and qi) must be considered along with the various
• For a proppant solid volume fraction fv less than roles of fluid viscosity for net pressure, width, prop-
0.17, the average width must be greater than pant transport and fluid loss. In addition, the design
(1 + 2fv /0.17) × dprop. must consider the various roles of the pad volume
• For fv greater than 0.17, the average width must be concerning fluid loss and creating fracture width.
greater than 3dprop (i.e., a width greater than three Fracture simulators, or fracture placement models,
proppant grain diameters). provide the means to handle this complexity and to
consider the interaction of the multitude of variables.
This approximate correlation also compares well
For this reason, a final schedule is generally devel-
with other experimental data from proppant-laden
oped using a fracture geometry model. However, as
slurry flowed through a narrow slot (van der Vlis et
discussed in Section 5-5.2, Sidebar 6L and Section
al., 1975), although the correlation may be optimistic
10-4, in many instances an acceptable pump schedule
for low proppant concentrations. As shown in Table
can be developed more simply for a treatment on the
5-1, the behavior for bridging in a fracture is similar
basis of the expected fluid efficiency (as determined
to bridging in perforation holes. At low proppant
from a calibration treatment). The use of a properly
concentrations, the average fracture width must be
calibrated fracture geometry model also enables the
only slightly greater than the average particle diame-
consideration of multiple scenarios for designing the
ter. As the proppant concentration increases toward

optimum treatment for a specific application. This with a large hydraulic fracture may drain the reser-
approach is briefly discussed in Section 5-6.1. voir much faster, making the economics much more
attractive despite the additional cost of the treatment.
Carrying this forward, 2, 10 or 100 or more wells
5-6. Economics and operational could be drilled and/or fractured. Between these
considerations extremes is the optimum plan, which is the number
of wells, number of fractured wells or both that max-
The preceding discussion covers most of the techni-
imize the economic value of the production com-
cal aspects of hydraulic fracturing (reservoir engi-
pared with the development capital costs and the
neering, fluid mechanics, rock mechanics, etc.) and
ongoing operating costs.
reviews the complex interactions that exist between
As a simple example, the process (at least for a
the various, often competing design variables.
single well) could proceed as pictured in Fig. 5-24
However, to complicate things further, hydraulic
(Veatch, 1986). First, reservoir engineering calcula-
fracturing and treatment design are generally gov-
tions provide a production forecast for various com-
erned by—or are at least sensitive to—two final
binations of fracture half-length xf and conductivity
considerations: economics and field operations.
kf w (including the case of no fracture at all). Based
on some future price forecast, this allows calculation
5-6.1. Economics of a present value, which is the future revenue from
the production less future operating costs and dis-
At the most basic level, hydraulic fracturing is about counted back to the present. Hydraulic fracturing
time and money: “economics.” Given reasonable calculations based on fluid loss, fracture height, etc.,
geologic continuity, a single well would, given suffi- are used to determine the treatment volumes required
cient time, drain an entire reservoir. However, the to generate various combinations of fracture length
operating costs of maintaining a well over the and propped fracture width, and these calculations
decades required to accomplish this drainage would are easily converted into estimated treatment costs.
probably make the entire operation unattractive from Some form of net revenue economic analysis is then
a commercial viewpoint. Alternatively, a single well used to determine the best type of proppant, desired
Discounted revenue ($)
Cumulative production
x f = x f2
kfw = kfw2 kfw = kfw2
x f = x f1
kfw = kfw1 kfw = kfw1
No fracture
Revenue less cost ($)
Time xf kfw = kfw1
kfw = kfw2
Treatment volume
xf
Cost ($)
kfw = kfw2
kfw = kfw1
xf xf
Figure 5-24. Veatch (1986) economics diagrams.

fracture length and other requirements for the opti- to the hydrostatic head of the fluid in the wellbore
mum treatment. phead and the pipe friction ppipe friction:
There are, of course, many variations of this basic
process. For example, full-cycle economics includes psurf = σ c + pnet + p pipe friction − phead (5-26)
drilling and other completion costs, along with frac-
ture treatment costs, in determining the optimum
fracture design. This type of analysis is usually hhp ∝ qi × psurf . (5-27)
appropriate in any case involving multiple wells
(e.g., should a resource be developed using 10 wells Pipe friction is a major term, and thus the size
with huge fractures or 20 wells with smaller or no of the well tubulars has a strong influence on
fracture treatments?). Point-forward analysis, on the allowable pump rates (because pipe friction is typ-
other hand, considers only the fracture treatment ically related to ve, where v = qi/A is the flow
costs (because drilling and other completion costs velocity down the tubing, and e is typically about
are already expended) and is most appropriate for 1.1 to 1.7). Also, the strength and condition of the
working over existing wells. tubulars (along with the associated wellhead
equipment) set an allowable surface pressure and
thus indirectly set an allowable injection rate. In
5-6.2. Operations addition, the size, type and condition of the well-
As discussed in the preceding section, economics bore tubulars may limit (or prohibit) future work-
provides the final design consideration for hydraulic over and recompletion opportunities.
fracturing, whereas field conditions provide the prac- A critical aspect of wellbore considerations is a
tical limits within which the design must fit. Even good cement job around the casing or liner to pro-
beyond defining these limiting conditions, however, vide zonal isolation. In general, a fracture grows
any design is only as good as its execution; thus the where nature dictates, and the engineer has little
treatment must be pumped as designed. Field opera- control over fracture height growth. The only con-
tions and operational considerations impact trol possible is the ability to specify where the
hydraulic fracturing in two ways: perforations are placed and the fracture initiates.
If that ability is compromised by a poor cement
• prefracture condition of the wellbore, quality of sheath around the casing that allows the perfora-
the cement job, perforations, pressure limits, etc., tions to communicate directly with an undesired
with these considerations defining practical limits interval, then even this minimal level of control is
that the design must meet lost, and the hydraulic fracture treatment may be
• quality assurance and quality control (QA/QC) seriously compromised.
before and during the actual treatment. Another important consideration is the perfora-
These operational considerations are discussed in tions that allow the fluid to leave the wellbore and
Chapters 7 and 11, with some of the major items create the fracture. The number and size of the
highlighted in the following. perforation holes are important for proppant
admittance, as discussed briefly in Section 5-5.4
• Wellbore considerations and in detail in Section 11-3.
Some of the major wellbore considerations for • Quality assurance and quality control
hydraulic fracturing include
Quality issues are critical for hydraulic fracturing.
– size and condition of wellbore tubulars After proppant pumping starts, a treatment cannot
– quality of the cement job for zonal isolation be stopped because of problems without signifi-
– perforations cantly compromising the design goals. For this
time period, everything must work, including
– wellbore deviation.
the wellbore equipment, pumping and blending
During a hydraulic fracture treatment, the pre- equipment and chemicals (i.e., the fluid system).
dicted surface pressure psurf and the hydraulic To cite a simple example, if a treatment uses
horsepower required for a treatment are related 10 tanks of batch-mixed fluid, and one of the

tanks is bad, then the QA score is a relatively high

90%. However, if the bad fluid is pumped just
after proppant addition starts, it may easily cause
total failure of the treatment, and if successful
treatment is critical to economic success of the
well, this causes total economic failure. Typically,
this type of failure cannot be overcome without
completely redrilling the well (refracturing opera-
tions are usually a risky procedure), and thus 90%
can be a failing grade for QA.

◆
Appendix: Evolution of Hydraulic
Fracturing Design and Evaluation
K. G. Nolte, Schlumberger Dowell
Overview reservoir management (i.e., control of both the verti-

cal and horizontal flow profiles within the reser-
This Appendix to Chapter 5 reviews the evolution voir). Similar speculation in a 1985 lecture sug-
of hydraulic fracturing design and evaluation gested that development of the technical foundation
methods. Complementary reviews are the applica- for the TSO generation would quickly bring higher
tion of fracturing by Smith and Hannah (1996) and permeability formations into consideration as typical
fracturing fluids by Jennings (1996). This review of fracturing candidates (i.e., “moderate k (2×)” on
design and evaluation considers three generations of Appendix Fig. 1a, with 2× indicating a target for
fracturing: damage bypass, massive treatments and folds of increase [FOI] in the production rate, in
tip-screenout (TSO) treatments. contrast to 10× for tight gas and massive treat-
The first two generations of fracturing and their ments). However, the advent of this generation was
links to practices are emphasized because these con- considerably delayed because of two factors that
tributions are not likely well known by the current have generally dominated technical considerations
generation of engineers. The review focuses on during the history of fracturing. These dominating
propped fracturing and does not explicitly consider factors are hydrocarbon prices and resistance to try-
acid fracturing. Although the principles governing ing something new until established practices fail to
the mechanics of both are essentially the same, the allow the economic development of a prospect.
fluid chemistry for obtaining fracture conductivity The cycles of fracturing activity in Appendix Fig.
is quite different (see Chapter 7). These principles 1a clearly reflect the timing of the first two fractur-
have their roots in civil and mechanical engineering, ing generations. Appendix Fig. 1b identifies eco-
more specifically in the general area of applied nomic drivers for corresponding cycles in the U.S.
mechanics: solid mechanics for the rock deforma- rig count. The first surge of activity resulted when
tion and fluid mechanics for the flow within the rotary drilling was introduced, which enabled the
fracture and porous media. For the porous media development of deeper reserves. Fracturing activity
aspects, fracturing evaluation has benefited greatly followed this trend soon after its commercialization
from the reservoir engineering practices discussed in 1949 because it was found to be an effective,
in Chapters 2 and 12. low-cost means of mitigating the resulting drilling
This review reflects the author’s perspective and mud damage to reservoir sections (i.e., the damage
bias in interpreting the impact of past contributions, bypass generation). Both drilling and fracturing
and therefore parts of this review should be antici- activities began a long-term decline after 1955
pated to raise objections from others with an exten- because of degrading prices caused by imported oil
sive knowledge of fracturing. In addition to this and regulated gas prices. Similarly, both activities
volume, the Society of Petroleum Engineers (SPE) began a rapid increase at about 1979 as prices
Monograph Recent Advances in Hydraulic Frac- increased because the Organization of Petroleum
turing (Gidley et al., 1989) provides balanced, Exporting Countries (OPEC) reduced its oil supplies
detailed coverage of the diverse areas of fracturing and a natural gas shortage developed in the United
from the perspectives of more than 30 fracturing States. The gas shortage, and its 10-fold-plus
specialists. increase in price, encouraged the development of
This review concludes with speculation concern- tight gas reserves and an associated demand for
ing a future generation, in which fracture design and massive fracturing treatments to develop the tight
reservoir engineering merge into fracturing for reserves. The failure of past fracturing practices for
Reservoir Stimulation A5-1

packing to achieve low-skin completions for a signif-

(a) icant venture in the Gulf of Mexico.
4000
The potential for a future reservoir management
generation was demonstrated in 1994 for the Nor-
(2×)
tivity
wegian Gullfaks field. The potential is to use TSO
rate k
ls,
onduc
treatments and indirect vertical fracturing for
ria
eq ding 10×)
me ate
age
Treatments
Mode
s(
increased reserves recovery, formation solids control
uip , m
High c
dam
nt
ga
de ight
and water management. However, the unique bene-
ove
an
T
?
rst
fits and favorable economics for this different
Rem
Un
approach to reservoir “plumbing” were slow to
materialize because of the industry’s comfort with
0
1950 1955 1971 1981 deviated drilling and more traditional completions.
Year Another observation from this historical perspec-
tive is the 1985 forecast of a flat drilling level
(b)
(Appendix Fig. 1b). However, activity continued to
4500 decrease rapidly, to less than one-half of the forecast,
Annual average rotary rig count
OPEC overextends
4000 U.S. gas prices regulated Prices fall and subsequently declined by another one-half. Stable
3500 Middle East discoveries
activity levels within the petroleum industry are not
3000 U.S. production peaks
seen in the historical cycles and remain the product
ear
OPEC develops
%/y
2500 price authority 1985 forcast flat

r
ea
of wishful thinking.
15.1
Do
/y
wn
9%
2000
Down 25
6.
Up
9.
1%
Up
/ye
1500 ar
.5%/yea
1000
500 Rotary displaces
The beginning
r
cable tool drilling

0
1940 1950 1960 1970 1980 1990 2000 The concept of hydraulic fracturing within the petro-
Year leum industry was developed during the last half of
the 1940s within Stanolind (now BP Amoco; e.g.,
Appendix Figure 1. (a) Trends in fracturing activity treat- Clark, 1949; Farris, 1953; Howard and Fast’s
ments per month (courtesy of K. G. Nolte and M. B. Smith, Hydraulic Fracturing Monograph, 1970) by building
1985–1986 SPE Distinguished Lecture). (b) U.S. drilling rig
activity shows five major trends (updated from Oil & Gas on the industry’s experience with injection tech-
Journal, 1985). niques that had experienced increased injectivity
by fracturing: acidizing (Grebe and Stoesser, 1935),
large treatments spurred a significant research and squeeze cementing and brine injection wells. A re-
development effort that beneficially impacted every issued patent was granted (Farris, 1953, resulting
aspect of fracturing and essentially developed the from an initial filing in 1948) that was comprehen-
fracture design and evaluation framework presented sive in scope and covered many recognized practices
in this volume. The industry’s rapid contraction dur- and products: proppant, gelled oil, breakers, fluid-
ing the early 1980s resulted again from OPEC, but loss additives, continuous mixing, pad-acid fractur-
this time because of OPEC’s failure to maintain arti- ing, emulsified acids and the use of packers for frac-
ficially high prices. The TSO treatment for creating turing multiple zones. Several aspects of the patent
the very wide propped fractures required for high that later became important included the implication
permeability evolved during this time. This tech- that fractures were horizontal and the use of a “low-
nique allowed the development of a troublesome penetrating” fluid or with viscosity > 30 cp.
soft-chalk reservoir in the North Sea by fracturing. The first experimental treatments were performed
However, the significant potential of the TSO gener- in 1947 on four carbonate zones in the Houghton
ation did not materialize until about 10 years later, field in Kansas (Howard and Fast, 1970). The zones
when its application was required on a relatively had been previously acidized and were isolated by a
large scale to achieve viable economics for two high- cup-type straddle packer as each was treated with
permeability applications: bypassing deep damage in 1000 gal of napalm-thickened gasoline followed by
the Prudhoe Bay field and its coupling with gravel 2000 gal of gasoline as a breaker. These unpropped
A5-2 Chapter 5 Appendix: Evolution of Hydraulic Fracturing Design and Evaluation

treatments did not increase production and led to the worthy design and evaluation methods from this gen-
incorrect belief for some time that fracturing had no eration are fracture orientation (horizontal or verti-
benefit over acidizing for carbonate formations. cal), in-situ stress and fracture width models, FOI
A subsequent treatment of the Woodbine sand in prediction and fracture conductivity in production
the East Texas field was highly successful. It con- enhancement.
sisted of 23 bbl of gelled lease crude, 160 lbm of
16-mesh sand at 0.15 ppa and 24 bbl of breaker
(Farris, 1953). Halliburton originally obtained an Fracture orientation and in-situ stress
exclusive license from Stanolind and commercial- The application of mechanics to fracturing was cat-
ized fracturing in 1949. Activity rapidly expanded to alyzed by the horizontal orientation of fractures
about 3000 treatments per month by 1955 (Appendix implied in the Stanolind patent and the desire of sev-
Fig. 1a). Before a universal license was granted to eral operators to avoid paying the nominal patent
other service companies, water or “river” fracturing royalty of $25–$125, based on volume (C. R. Fast,
became popular in lower permeability areas such as pers. comm., 1997). Significant research activity was
the San Juan basin (C. R. Fast, pers. comm., 1997). conducted to show that fractures can be vertical, as is
As implied by the name, the treatments used river now known to be the general case for typical fractur-
water and sand. The water was outside the definition ing conditions. The fracture orientation debate even-
of a nonpenetrating fluid within the patent’s specified tually led to a lawsuit that was settled before the trial
filtrate rate through filter paper or viscosity greater ended. The settlement accepted the patent and nomi-
than 30 cp. nal royalty payments and stipulated that other service
companies receive a license to practice fracturing.
However, the royalty benefits were more than nomi-
The first generation: damage bypass nal to Stanolind because about 500,000 treatments
Applications of first-generation fracturing were pri- were performed during the 17-year period of the
marily small treatments to bypass near-wellbore patent (C. R. Fast, pers. comm., 1997). Key to
drilling fluid damage to formations with permeability the favorable settlement for Stanolind was its well-
in the millidarcy range. An inherent advantage of documented demonstration of a horizontal fracture
propped fracturing, relative to matrix treatment for in the Pine Island field (see fig. 7-1 in Howard and
damage removal, is that a fracture opens the com- Fast, 1970).
plete section and retains a conductive path into the The central issue in the orientation debate was the
zone. The complete opening overcomes the diversion direction of the minimum stress. The pressure
consideration for matrix treatments (see Chapter 19), required to extend a fracture must exceed the stress
but adds the consideration of producing from bot- acting to close the fracture. Therefore, the fracture
tomwater or an upper gas cap. For lower permeability preferentially aligns itself perpendicular to the direc-
formations, large amounts of produced water are tion of minimum stress because this orientation pro-
generally not a problem. For higher permeability for- vides the lowest level of power to propagate the frac-
mations, water production can be significant, which ture. The minimum stress direction is generally hori-
provided the historical preference for matrix treat- zontal; hence, the fracture plane orientation is gener-
ment in higher permeability applications. However, ally vertical (i.e., a vertical fracture). The preference
the precision of fracturing improved significantly, for a horizontal fracture requires a vertical minimum
and TSO treatments have been routinely performed stress direction.
in Prudhoe Bay oil columns only 50 ft thick and In the following review, the orientation considera-
above very mobile water (Martins et al., 1992b). tion is expanded to also cover the state of stress in
The technology for this fracturing generation is more general terms. The stress at any point in the var-
summarized in the Howard and Fast (1970) Mono- ious rock layers intersected by the fracture is defined
graph. The breadth of this volume is shown by its by its magnitude in three principal and perpendicular
comprehensive consideration of candidate selection directions. The stress state defines not only the frac-
(see Chapter 1) and optimal design based on eco- ture orientation, but also the fluid pressure required to
nomic return (see Chapters 5 and 10). Other note- propagate a fracture that has operational importance,
vertical fracture growth into surrounding formation

layers and stress acting to crush proppant or to close Harrison et al. also reported the Sneddon and
etched channels from acid fracturing. The crushing Elliott (1946) width relation for an infinitely extend-
stress is the minimum stress minus the bottomhole ing pressurized slit contained in an infinitely extend-
flowing pressure in the fracture. The orientation ing elastic material. This framework has become the
debate resulted in three papers that will remain signifi- basis for predicting fracture width and fracturing
cant well into the future. pressure response (see Chapters 5, 6 and 9). They
The first paper to be considered is by Harrison et used the fracture length for the characteristic, or
al. (1954). Some of the important points in the paper smaller and finite, dimension in this relation. Sel-
are that the overburden stress (vertical stress σv) is ecting the length for the characteristic dimension
about 1 psi per foot of depth, fracturing pressures are resulted in what is now commonly termed the KGD
generally lower than this value and therefore frac- model. Selecting the height, as is the case for a very
tures are not horizontal, and an inference from elas- long fracture, is termed the PKN model. These mod-
ticity that the minimum horizontal stress is els are discussed in the next section and Chapter 6.
Harrison et al. considered a width relation because
σ h = Ko σ v , (1) of its role in fracture design to determine the fluid
volume required for a desired fracture extent.
where Ko = ν/(1 – ν) = 1⁄3 for ν = 1⁄4 (see Eq. 3-51). The role of volume balance (or equivalently, the
Using Poisson’s ratio ν of 1⁄4, Harrison et al. con- material balance in reservoir terminology) is an
cluded that the horizontal stress is about one-third of essential part of fracture design and fracture simula-
the vertical stress and therefore fractures are vertical. tion code. As shown schematically on the left side of
Appendix Eq. 1 provides the current basis for using Appendix Fig. 2, each unit of fluid injected Vi is
mechanical properties logs to infer horizontal stress, either stored in the fracture to create fracture volume
with Poisson’s ratio obtained from a relation based or lost to the formation as fluid loss. (However,
on the shear and compressional sonic wave speeds Harrison et al.’s 1954 paper does not discuss fluid
(see Chapter 4). Another assumption for Appendix loss.) The stored volume is the product of twice the
Eq. 1 is uniaxial compaction, based on the premise fracture half-length L, height hf and width w. If the
that the circumference of the earth does not change latter two dimensions are not constant along the frac-
as sediments are buried to the depths of petroleum ture length, they can be appropriately averaged over
reservoirs and hence the horizontal components of the length. The half-length is then obtained by sim-
strain are zero during this process. Therefore, Ap- ply dividing the remaining volume, after removing
pendix Eq. 1 provides the horizontal stress response the fluid-loss volume, by twice the product of the
to maintain the horizontal dimensions of a unit cube
constant under the application of vertical stress.
Vi
However, there is one problem with this 1954 L=
conclusion concerning horizontal stress. Appendix 2hf (w + CL √8t )
Eq. 1 is correct for the effective stress σ′ but not for

2hf wL
the total stress σ that governs fracture propagation: η=
Vi
σ′ = σ – p, where p is the pore pressure, which also Fluid loss
has a role in transferring the vertical stress into hori- CL √ t
Proppant
zontal stress as explicitly shown by Appendix Eq. 2.

Harrison et al. (1954) correctly postulated that shales
Geometry
have higher horizontal stresses and limit the vertical Proppant
(% area = η)
fracture height. The general case of higher stress in hf 2L w
shales than in reservoir rocks was a necessary condi-
Pad
tion for the successful application of fracturing
Volume
because fractures follow the path of least stress. If
the converse were the general case, fractures would Appendix Figure 2. Volume balance for fracture place-
prefer to propagate in shales and not in reservoir ment (equation from Harrington et al., 1973) (adapted
zones. courtesy of K. G. Nolte and M. B. Smith, 1984–1985 SPE
Distinguished Lecture).

average height and the average width. The fluid-loss Hubbert and Willis also provided an important set
volume depends on the fluid-loss surface area, or a of postulates: the rock stresses within the earth are
height-length product. Furthermore, as shown on the defined by rock failure from tectonic action and the
right side of Appendix Fig. 2, the ratio of stored to earth is in a continuous state of incipient faulting.
total volume is termed the fluid efficiency η and dir- From this perspective, the stress is not governed
ectly affects the proppant additional schedule (Har- by the behavior of the intact rock matrix, but by an
rington et al., 1973; Nolte, 1986b) (see Sidebar 6L). active state of failure along discrete boundaries (e.g.,
The second paper to be discussed from the orienta- by sand grains within fault boundaries, which
tion era is by Hubbert and Willis (1957). The lessons explains the application of Appendix Eq. 2 to micro-
from this paper extend beyond fracturing and into darcy-permeability sandstones). This insightful con-
the area of structural geology. This work provides clusion about the role of failure is at the other
simple and insightful experiments to define the state extreme of the behavior spectrum from the elastic
of in-situ stress and demonstrate a fracture’s prefer- assumptions that Poisson’s ratio (Appendix Eq. 1)
ence to propagate in the plane with minimum stress governs the horizontal stress and that failure has no
resistance. For the latter experiments, the “forma- effect on the stress. This extreme difference in the
tion” was gelatin within a plastic bottle preferentially assumptions for Appendix Eqs. 1 and 2 is often
stressed to create various planes of minimal stress. overlooked because of the similar value of Ko = ~1⁄3
They also used simple sandbox experiments to obtained in the case of a tectonically relaxed region
demonstrate normal and thrust faulting and to define and Poisson’s ratio near 1⁄4. However, the role of elas-
the state of stress for these conditions (see Sidebar ticity becomes important in thrusting areas (see
3A). They showed that Ko, or equivalently the hori- Section 3-5.2) because of the difference in horizontal
zontal stress, within Appendix Eq. 1 is defined by stress resulting for layers with different values of
the internal friction angle (ϕ = 30° for sand) and is Young’s modulus (stiffness). More of the tectonic
1
⁄3 for the minimum stress during normal faulting and action and higher levels of stress are supported by the
3 for the maximum stress during thrust faulting. For stiffer layers.
the normal faulting case and correctly including pore Additional considerations for horizontal stress out-
pressure in Appendix Eq. 1, the total minimum hori- lined by Prats (1981) include the role of long-term
zontal stress becomes creep. Creep deformation allows relaxation of the
stress difference between the overburden and hori-
σ h = ( σ v + 2 p) 3 , (2) zontal stresses, thereby enabling the horizontal stress
to increase toward the larger vertical stress governed
where Ko = 1⁄3 with ϕ = 30°. For this case the horizon- by the weight of the overburden. This effect is well
tal stress is much less than the vertical stress except known for salt layers that readily creep and can col-
in the extreme geopressure case of pore pressure lapse casing by transferring most of the larger over-
approaching overburden, which causes all stresses burden stress into horizontal stress. The role of stress
and pore pressure to converge to the overburden relaxation is an important mechanism for providing
stress. For the thrust faulting case, the larger horizon- favorable stress differences between relatively clean
tal stress (i.e., for the two horizontal directions) is sands governed by friction (i.e., Appendix Eq. 2) with
greater than the overburden and the smaller horizon- minimal creep and sediments with higher clay con-
tal stress is equal to or greater than the overburden. tent. In the latter case, the clay supports some of the
Both the extreme geopressure case and an active intergranular stresses. The clay structure is prone to
thrust faulting regime can lead to either vertical or creep that relaxes the in-situ stress differences and
horizontal fractures. The author has found Appendix increases the horizontal stress for a clay-rich formation.
Eq. 2 to accurately predict the horizontal stress in tec- Hence, both clay content and Poisson’s ratio pro-
tonically relaxed sandstone formations ranging from duce the same effect on horizontal stress. Because
microdarcy to darcy permeability. The accuracy at clay content also increases Poisson’s ratio, there is
the high range is not surprising, as the formations a positive correlation of clay content (creep-induced
approach the unconsolidated sand in the sandbox stress) to larger Poisson’s ratios (and elastic stress,
experiments. The accuracy obtained for microdarcy- from Appendix Eq. 1) inferred from sonic velocities.
permeability sands is subsequently explained. The implication of the correlation is that clay-rich

formations can also have horizontal stresses greater injection) and decreasing stress within the region
than those predicted by either Appendix Eq. 1 or 2, of decreasing pore pressure (e.g., production). The
which is consistent with the general requirement to long-term impact of Lubinski’s paper is that the
calibrate elastic-based stress profiles to higher levels importance of poroelasticity increases as routine
of stress (e.g., Nolte and Smith, 1981). The correla- fracturing applications continue their evolution to
tion of clay and Poisson’s ratio links the conclusions higher permeability formations. This is apparent
of Hubbert and Willis and Prats that horizontal stress from the thermal analogy—as the area of expansion
is governed primarily by nonelastic effects and the increases the induced stresses also increase. For
general correlation between the actual stress and poroelasticity, the area of significant transient change
elastic/sonic-based stress profiles. in pore pressure increases as the permeability
The third significant paper from this period is by increases (see Section 3-5.8).
Lubinski (1954). He was a Stanolind researcher who Appendix Fig. 3 shows an example of significant
introduced the role that poroelasticity can have in poroelasticity for a frac and pack treatment in a
generating larger stresses during fracturing. (Poro- 1.5-darcy oil formation. The time line for the figure
elasticity could increase horizontal stress and lead begins with two injection sequences for a linear-gel
to horizontal fractures, as in the Stanolind patent.) fluid and shows the pressure increasing to about
Lubinski presented poroelasticity within the context 7500 psi and reaching the pressure limit for the oper-
of its analogy to thermoelasticity. His use of the ther- ation. During the early part of the third injection
mal stress analogy facilitates understanding the poro- period, crosslinked fluid reaches the formation and
elastic concept because thermal stresses are generally the pressure drops quickly to about 5600 psi (the
more readily understood than pore stresses by engi- native fracturing pressure) and remains essentially
neers. The analogy provides that when pore pressure constant during the remainder of the injection.
is increased in an unrestrained volume of rock, the The first two injections, without a crosslinked-fluid
rock will expand in the same manner as if the tem- filtrate (or filter cake) to effectively insulate the for-
perature is increased. Conversely, when the pore mation (as in the thermal analogy) from the increas-
pressure is lowered, the rock will contract as if the ing injection pressure, resulted in pore pressure
temperature is lowered. When the rock is con- increases of significant magnitude and extent within
strained, as in a reservoir, a localized region of pore the formation. The pore pressure increase provides up
pressure change will induce stress changes: increas- to a 1900-psi horizontal and poroelasticity stress
ing stress within the region of increasing pore pres- increase that extends the fracturing pressure beyond
sure (e.g., from fracturing fluid filtrate or water the operational limit, leading to the shut-in for the
10,000 50
Linear gel Crosslinked gel
8000 Injection 40
Bottomhole pressure, BHP (psi)
Step rate
Linear gel
Injection rate (bbl/min)
6000 30
4000 20
Injection
Step rate Minifracture Propped fracture
2000 10
BHP
Injection rate
0 0
0 0.5 1.0 2.0 13.0 13.5 14.0
Time (hr)
Appendix Figure 3. High-permeability frac and pack treatment (Gulrajani et al., 1997b).

second injection. This increase is about one-third of large tip-region stresses to a level that can be accom-
the native stress. However, during the two subsequent modated by the in-situ condition. The presence of
injections the insulating effect of the crosslinked the lag region has been demonstrated by field experi-
fluid’s internal cake and filtrate allows fracture exten- ments at a depth of 1400 ft at the U.S. Department of
sion within essentially the native stress state. The Energy (DOE) Nevada Test Site (Warpinski, 1985).
pressure drop supported by the cake and filtrate is Appendix Fig. 4 compares the Khristianovich and
about 1300 psi, as reflected by the rapid pressure Zheltov analytical results for width and pressure to
decrease after the third injection. This decrease the corresponding parameters from the Warpinski
occurs because of the rapid closure and cessation of field results. For the analytical results, decreasing
fluid loss (that activated the pressure drop), which is values of the complex variable angle ϑ0 toward the
the same reason that surface pressure decreases at the right side of the figure correspond to relatively
cessation of injection and loss of pipe friction. The smaller lag regions and larger differences between
last injection for the proppant treatment is also of the minimum stress and pressure in the lag region
interest because of the absence of a poroelastic effect (i.e., generally deeper formations). The width pro-
during the initial linear-gel injection. This observation files clearly show the clamping action at the tip, and
indicates that the insulating effect remained effective the field data appear to be represented by a ϑ0 valve
from the prior injection of crosslinked fluid. of about π/8 for the analytical cases. Also notewor-
For a normally pressured and tectonically relaxed thy of the experimental results is that tests 4 through
area, the maximum increase in horizontal stress 7 with water and test 9 with gel show similar behav-
before substantial fracture extension is about one- ior when test 4, which had a relatively low injection
third of the native horizontal stress (Nolte, 1997), rate, is ignored. Tests 10 and 11 were with a gelled
as was found for the case shown in Appendix Fig. fluid and clearly show progressively different behav-
3. Also, for any pore pressure condition in a relaxed ior from the preceding tests because of the altered tip
area, the stress increase will not cause the horizontal behavior resulting from prior gel injections and the
stress to exceed the overburden (i.e., cause horizontal residual gel filter cakes that fill the fracture aperture
fracturing). However, as the example shows, without after closure. The cakes have the consistency of sili-
fluid-loss control, poroelasticity can significantly con rubber and functionally provide an analogous
increase the fracturing pressure and extend it beyond sealing affect for subsequent tests.
operational limits for high-permeability reservoirs. The practical importance of the lag region cannot
be overemphasized. The extent of the region, which
is extremely small in comparison with commercial-
Width models scale fractures, adjusts to the degree required to
The first rigorous coupling of fluid flow and the elas- essentially eliminate the role of the rock’s fracture
tic response of the formation was reported by resistance or toughness (e.g., see SCR Geomechanics
Khristianovich and Zheltov (1955). They used a two- Group, 1993) and to isolate the fluid path from all
dimensional (2D) formulation based on a complex but the primary opening within the multitude of
variable analysis. Their formulation was equivalent cracks (process zone) forming ahead of the fracture
to the length becoming the characteristic, or smaller, (see Chapters 3 and 6). The field data show the
dimension and provides the initial “K” for the KGD width at the fluid front is well established (i.e., gen-
width model discussed later and in Chapter 6. In erally greater than 5% of the maximum width at the
addition to being the first paper to provide the cou- wellbore) and that fluid enters only a well-established
pling of fluid flow and rock interaction that is the channel behind the complexity of the process zone.
embodiment of the hydraulic fracturing process, the These aspects of the lag region provide great simpli-
paper also identified the role for a fluid lag region at fication and increased predictablility for applying
the fracture tip. This low-pressure region, beyond the commercial-scale hydraulic fracturing processes.
reach of fracturing fluid and filling with pore fluid, A paper by Howard and Fast (1957), and particu-
has a large, negative net pressure and acts as a clamp larly the accompanying appendix by R. D. Carter,
at the fracture tip. The fluid lag’s clamping effect provides the current framework for fluid loss. The
provides the natural means to lower the potentially paper identifies the three factors controlling fluid
loss: filter-cake accumulation, filtrate resistance into

0.6
Test 1.0
4
5 0.9
0.5 6
7 0.8
9 ϑ0 = 3π
16
0.4 10 0.7
11 ϑ0 = π
8
0.6
ϑ0 = 3π ϑ0 = π
w/wo
w/wo
0.3 8 16
0.5
ϑ0 = π
Width at
fluid arrival 4
0.4
0.2
0.3
0.2
0.1
0.1
0 0
0.25 0.20 0.15 0.10 0.05 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Normalized distance from tip, (L – x)/L Normalized distance from well, x/L
1.0 1.0
Test
4
5
0.8 6 0.8
ϑ0 = π
7
9 8
10 ϑ0 = 3π
0.6 8
0.6
p/po
p/po
ϑ0 = π
16
0.4 0.4
ϑ0 = π
4
0.2 0.2
ϑ0 = 3π
16
0 0
0.5 0.4 0.3 0.2 0.1 0 0 0.2 0.4 0.6 0.8 1.0
Normalized distance from tip, (L – x)/L Normalized distance from well, x/L
Appendix Figure 4. Comparison of Warpinski (1985) field data (left) and Khristianovich and Zheltov (1955) analysis
(right). wo and po are the wellbore values of width and pressure, respectively; x is the distance from the well.
the reservoir and displacement of the reservoir fluid height and tip-to-tip length. This equation, based on
(see Fig. 5-17 and Chapters 6 and 8). All three fac- the assumption of a spatial and temporal constant
tors are governed by the relation 1/√t (where t is fracture width, provided the first rigorous inclusion
time) for porous flow in one dimension. The coeffi- of fluid loss into the fracturing problem (see Chapter
cient for this relation was termed the fluid-loss coef- 6). Equation 6-18, which is solved by Laplace trans-
ficient CL. The authors also provided the means to formation, is in terms of exponential and comple-
determine the coefficient for all three factors using mentary error functions and is not “engineer friendly.”
analytical expressions for the filtrate and reservoir This difficulty was soon overcome by developing a
contributions and to conduct and analyze the filter- table for the more complicated terms in the equation
cake experiment, which is now an American Petro- using a dimensionless variable (see Eq. 6-19) that is
leum Institute (API) Recommended Practice. proportional to the fluid-loss coefficient (loss vol-
Also of significance was presentation of the Carter ume) divided by the width (stored volume) and
area equation, with area defined as the product of the hence also related directly to the fluid efficiency

η illustrated in Appendix Fig. 2. Nomographs for The one shortcoming acknowledged by Perkins
the complete equation were also developed (e.g., and Kern was not rigorously accounting for the flow
figs. 4-17 and 4-18 of the Howard and Fast Mono- rate change in the fracture required by continuity
graph). Eventually a simple and approximate expres- (i.e., material balance). They assumed that the volu-
sion (Harrington et al., 1973) for the Carter equation metric flow rate was constant along the fracture’s
provided the basis for fracture design into the 1980s. length, which does not account for the effects of
The approximate expression is based on the relation fluid loss and local rates of width change (storage
at the top of Appendix Fig. 2. For these applications, change). This assumption was later addressed by
the average width was first determined from either the Nordgren (1972), who provided closed-form equa-
KGD or PKN model, as discussed in the following. tions for the bounding cases of negligible fluid loss
Another 1957 paper was by Godbey and Hodges and negligible fracture storage (i.e., most fluid
(1958) and provided the following prophetic phrases: injected is lost during pumping) for a long-aspect
“By obtaining the actual pressure on the formation fracture and Newtonian fluid (see Section 6-2.2). The
during a fracture treatment, and if the inherent tec- initial letters of the last names of the authors of these
tonic stresses are known, it should be possible to two papers form the name of the PKN model.
determine the type of fracture induced. . . . The The remaining paper of historic importance for
observation of both the wellhead and bottomhole width modeling is by Geertsma and de Klerk (1969).
pressure during fracturing operations is necessary to They used the Carter area equation to include fluid
a complete understanding and possible improvement loss within the short-aspect fracture, as previously
of this process.” These statements anticipated two of considered by Harrison et al. (1954) and Khristian-
the important enablers for the second generation of ovich and Zheltov (1955). Their initials coupled with
fracturing: the use of pressure in an manner analo- those of the authors of the latter paper form the name
gous to well test characterization of a reservoir and of the KGD (or KZGD) width model.
employment of a calibration treatment to improve
the subsequent proppant treatment (see Chapters 5,
9 and 10). Reservoir response to a fracture
In 1961 Perkins and Kern published their paper Until the advent of numerical simulators, production
on fracture width models, including the long aspect models for a fracture did not consider transient flow
ratio fracture (length significantly greater than height) effects and were based on the FOI relative to the
and radial model (tip-to-tip length about equal to reservoir’s radial flow response with no damage (skin
height) as described in Section 6-2.2. They considered, effect = 0). The increase in production, relative to the
for the first time, both turbulent fluid flow and non- case before fracturing, can be significantly greater
Newtonian fluids (power law model) and provided than the FOI measure because fracturing also by-
validating experiments for radial geometry and the passes near-wellbore damage. The enhanced stimula-
role of rock toughness. tion benefit increases as the magnitude of the damage
Perkins and Kern also discussed fracture afterflow increases. For example, removing a skin effect of
that affects the final proppant distribution within the about 25 increases production by about a factor of 4,
fracture. After pumping stops, the stored compres- whereas during the first generation a typical FOI target
sion in the rock acts in the same fashion as com- was about 2, relative to zero skin effect.
pressible fluids in a wellbore after well shut-in. After Papers considering finite-conductive fractures
fracture shut-in, fluid flow continues toward the tip began to appear in 1958 and are summarized in chap-
until either proppant bridges the tip or fluid loss ter 10 of the Howard and Fast (1970) Monograph.
reduces the fracture width and stored compression Craft et al. (1962) considered the combined effects of
to the extent that the fracture length begins to recede fracture stimulation and damage bypass. Also of his-
toward the wellbore (Nolte, 1991). The magnitude torical interest is that most of this work was per-
of the fracture afterflow is large compared with the formed on analog computers with electrical circuits
wellbore storage case, as discussed later for representing the reservoir and fracture components.
Appendix Eq. 4. Recognition of the role of conductivity was important
because the idealized assumption of infinite conduc-

tivity, with no pressure loss in the proppant pack,

1.0
cannot result from an economics-based optimized rw′ = 0.5x f
treatment. The incremental production increase, by
Effective wellbore radius, rw′/xf

0.5
1
achieving the infinite-acting case, would not offset 1 CfD = 30
k fw
the operational cost for the additional proppant. rw′ = 0.28
k
McGuire and Sikora (1960) presented a significant 0. 1
study of the production increase in a bounded reser-
voir for a fracture with a finite conductivity kf w for CfD =
k fw
kx f
the proppant pack, where kf is the fracture permeabil- CfD = 0.2
ity. The boundary and conductivity effects are sum-

marized in the set of pseudosteady-state curves 0.01
shown in Appendix Fig. 5. The curves reflect differ- 0.1 1.0 10 100
ent ratios of the fracture length relative to the Dimensionless fracture conductivity, CfD
drainage radius re, with the vertical axis reflecting
the FOI as J/Jo and the horizontal axis reflecting Appendix Figure 6. Effective wellbore radius versus
dimensionless fracture conductivity (Nolte and
dimensionless conductivity based on the drainage Economides, 1991, adapted from Cinco-Ley and
radius. The McGuire and Sikora curves were the pri- Samaniego-V., 1981b).
mary reservoir tool for fracture design and evalua-
tion until the late 1970s.
Prats (1961) used mathematical analyses to con- The effective wellbore radius, coupled with the radial
duct a comprehensive consideration of finite-conduc- flow equation, provides a powerful tool for efficiently
tivity fractures with the assumption of steady-state calculating the FOI, or negative skin effect, provided
flow (i.e., constant-pressure boundaries). He intro- by the fracture. Prats also considered fracture face
duced a dimensionless conductivity α that is essen- damage (or skin effect) and provided an optimized
tially the inverse of the dimensionless fracture con- treatment based on a fixed amount of proppant.
ductivity commonly used for transient analyses (i.e.,
CfD = kf w/kxf = π/2α). Prats also introduced the con-
Treatment optimization
cept of an effective (or apparent) wellbore radius rw′.
The effective radius allows describing the fracture Optimizing a fracture treatment is an essential part
response in terms of an enlarged wellbore radius of maximizing its benefit (see Chapters 5 and 10).
within the radial flow equation. This concept is illus- For this reason Prats’ (1961) optimization considera-
trated in Appendix Fig. 6 for pseudoradial flow tion is of historic importance, although proppant vol-
(adapted from Cinco-Ley and Samaniego-V., 1981b). ume is generally not a realistic criterion because
proppant cost is only part of the investment for a
fracture treatment (e.g., Veatch, 1986; Meng and
14 Brown, 1987). Prats’ proppant optimization condi-
L/re = 1
0.9
tion at CfD = 1.26 could be a practical target for
)
12 0.8
0.7
ln 0.472 rw
re
10 0.6 high-permeability reservoirs; however, this value is

7.13
8
0.5 about an order of magnitude lower than the optimum
0.4
case for the long transient period of a very low per-
6 0.3
meability reservoir.
(
4 0.2
Additional lessons are also provided by the appar-
(J/Jo)
0.1
2 entwellbore concept. The first is that a fracture is
0 equivalent to enlarging the wellbore and not increas-
102 103 104 105 106 ing the formation’s global permeability. Incorrectly
kfw 40 considering a fracture to be a permeability increase
Relative conductivity,
k √A can lead to incorrect conclusions concerning reser-
voir recovery and waterflood sweep. Another insight
Appendix Figure 5. McGuire and Sikora (1960) curves for
folds of increase (J/Jo) in a bounded reservoir of area A
is the generally favorable economics for an effec-
(acres). tively designed and executed fracture. A fracture

treatment is equivalent to excavating a very large Transition between the first and
diameter borehole (e.g., hundreds of feet in most
cases) and therefore is an extremely cost-effective second generations
way to provide an equivalent excavation. By 1961, the design and evaluation tools for most
The most important optimization lesson is found of the next two decades had been established by the
in Appendix Fig. 6 for the roles of conductivity kf w contributions discussed. Incremental development of
(achieved by proppant cost) and fracture penetration these tools slowed because fracturing was considered
(achieved by fluid and other additive costs; see a mature technology. Also affecting technical devel-
Chapter 7). The figure indicates that as CfD increases opment was the degrading economics for lower qual-
beyond 10, the effective wellbore radius approaches ity reserves as oil import-export increased and frac-
one-half of the fracture length and there are dimin- turing activity decreased (Appendix Fig. 1). This
ishing returns for additional increases in conductivity condition did not change until the mid-1970s
(i.e., incurring proppant costs without an effective brought natural gas shortages and higher gas prices
increase in production rate). For this part of Appen- to the United States. Higher prices produced the
dix Fig. 6, the effective radius is constrained only incentive to develop extensive regions of tight gas
by length and is termed the length-limited case. reserves with fractures targeting the FOI = 10 range
However, increasing both fracture length and con- of the McGuire and Sikora curves (Appendix Fig. 5).
ductivity to maintain a constant CfD achieves the Before this period, typical fracturing targets were oil
most efficient conversion of length into an effective reservoirs with an FOI of about 2, with FOI relative
wellbore radius. This conversion is the basis for to an undamaged wellbore. However the FOI = 10
effectively fracturing low-permeability formations. target required about an order-of-magnitude increase
The practical limits for the length-limited case are in the volume and cost for a typical treatment and
reaching the drainage radius, increasing conductivity was hence termed massive hydraulic fracturing.
within the limits of achievable fracture width and This new target introduced higher temperature
efficiently extending a fracture when the pressure reservoirs, typically of tight gas, that generally
reaches the formation capacity, as discussed later. As exceeded the performance limits for fracturing fluid
permeability increases, and proportionally decreases systems. These conditions stretched the so-called
CfD, the ability to increase conductivity becomes mature technology in almost every conceivable way
the constraint. As CfD progressively decreases, the and resulted in a bumpy journey because of the pro-
conductivity-limited case is reached. The figure indi- portionally large economic penalty when a treatment
cates that as CfD decreases below 1, a log-log unit failed to meet expectations. However, reports of suc-
slope is approached that relates rw′ to kf w/k, with the cessful field development (e.g., Fast et al., 1977)
obvious absence of an effect from length. When the encouraged continued interest in tight gas develop-
unit slope is reached, near a value of 0.2, the well- ment.
bore drawdown completely dissipates within the
fracture before reaching the tip, and the extremities
of the fracture cannot provide a production benefit. Realistic estimate of conductivity
For the conductivity-limited condition, the produc- Cooke (1975) reported realistic experiments for char-
tion rate can be increased economically only by pro- acterizing the conductivity of proppant packs. His
viding more conductivity kf w, with an obvious con- procedure formed proppant packs from a slurry com-
straint from the available fracture width developed posed of polymer-based fluids by using a cell with
during the treatment. This constraint was significantly rock faces that allowed fluid loss and the subsequent
extended by the third fracturing generation of TSO application of closure stress. The Cooke cell is now
treatments, which is discussed toward the end of this a standard apparatus for a fracturing fluid laboratory
Appendix.

(see Chapter 8). The experiments showed that the infinitely above and below the pay section with each
retained pack permeability could be very small. barrier having the same magnitude of stress). The
These results were unexpected because prior testing three-layer case provided insight into how to adapt
procedures did not use fracturing fluids or stress lev- more general relations to any number of layers (e.g.,
els for deeper gas reserves. The primary difference Nolte and Smith, 1981; chapter 3 of Gidley et al.,
resulted because the rock acts as a polymer screen 1989). These relations led to the calculations employed
at moderate and smaller permeability levels, which in pseudo-three-dimensional (P3D) fracture simulators
significantly increases the polymer concentration (see Section 6-3.2).
remaining within the proppant pack porosity after Novotny (1977) outlined a comprehensive basis for
fracture closure. proppant transport calculations and in particular identi-
Cooke also provided a simple mass-balance rela- fied the important roles of channel shear rate and frac-
tion for this important consideration. The concen- ture closure in determining the ultimate placement of
tration factor for the polymer and other additives proppant (see Section 6-5.3. Both effects produce more
remaining in the fracture relative to the original proppant fall. For non-Newtonian fluids, the effective
concentration can be expressed as viscosity for sedimentation is determined from the vec-
toral sum of the shear rate in the channel and that
CF = 44 / ppa , (3) caused by proppant fall (as for stagnant fluid). This
for a typical proppant pack porosity of 0.33 and sum is generally dominated by the channel flow and
proppant specific gravity (s.g.) of 2.65. The relation is much greater than that for a particle in stagnant fluid
depends on the average concentration <ppa> defined (i.e., higher shear rate and lower viscosity). In addition,
as the total pounds of proppant divided by the total the closure period prolongs the time for proppant fall
gallons of polymer-based fluid. This relation indicates and maintains the channel flow to reduce the effective
a polymer concentration increase of 20 or greater for viscosity. Novotny also provided a brief analysis of the
typical treatments at that time (e.g., <ppa> of 1 to volume balance during closure, which is the essential
2 lbm). This unexpected discovery of a significant ingredient for the fracturing pressure decline analysis
reduction in retained permeability, coupled with the (e.g., Nolte, 1979) that is used for calibration treat-
prior discussion on conductivity and effective well- ments (see Section 9-5).
bore radius, partly explains the difficult transition to
massive treatments.
Cooke’s pioneering work had obvious effects on Transient reservoir response
proppant schedules for treatments and laboratory The FOI consideration for fracture production was
testing procedures. Equally important, the work initi- found to be completely inadequate for the substantial
ated substantial product development activities, as period of transient flow that occurs in tight formations
discussed in Chapter 7. These include improved (see Section 12-2). The first tool for finite-conductivity
proppants, beginning with Cooke’s work on bauxite transient flow was type curves provided by Agarwal
for high crushing stress, improved breaker chemistry et al. (1979). Although these curves were developed
and breaker encapsulation, large reductions of poly- from numerical simulators, access to computers was
mer concentration for crosslinked fluids, foams and generally outside the reach of most engineers. These
emulsions, and residue-free viscoelastic surfactant and similar type curves became the standard evalua-
systems. The evolution of fracturing fluid chemistry tion tool to assess production from a fracture treat-
was reviewed by Jennings (1996). ment. Type curves were also used for optimizing treat-
ment design. By the mid-1980s, as general access to
computers increased, the use of type curves began to
Height growth and proppant transport decrease accordingly.
Simonson et al. (1978) presented the mechanics gov- Cinco-Ley and Samaniego-V. (1981b) provided
erning fracture growth into a layer with higher stress, several advancements for understanding and quanti-
complementing the postulate by Harrison et al. (1954) fying the transient behavior of a reservoir fracture
concerning the role of stress for height confinement. system. In addition to advancing the effective well-
The analysis considered a three-layer case for two bore concept (e.g., Appendix Fig. 6) and type curves,
symmetric barriers (i.e., two barriers extending they identified and provided comprehensive descrip-

tions for the distinctive transient regimes resulting These advancements and insight from Bennett
from a finite-conductivity fracture (see Section 12-2). et al. (1986) for layered formations provide a solid
The bilinear flow regime, generally the first to foundation for the reservoir response to fracturing.
occur during production or a well test, was para-
mount for bridging the gap between fracture design
and subsequent evaluation based on production or The second generation:
well tests. For permeability in the range of 10 µd, the
bilinear period can last on the order of a year or more
massive fracturing
for a long fracture (>2500 ft from the well). During As indicated in the preceding section, the bumpy
bilinear flow the stabilized pressure drawdown pro- road to successful massive fracturing also included
gresses along the fracture length. During this period, massive penalties because the cost of a fracture treat-
it is not possible to determine the length of the frac- ment could become equivalent to the well cost. The
ture from a well test or production data because the combined effect of many companies experiencing
total length has not had time to effectively experience $500,000 treatments that did not provide commercial
the wellbore drawdown. Therefore, a meaningful wells resulted in a significant investment for fractur-
evaluation for fracture length cannot be obtained until ing research. One result of this effort is the SPE
the bilinear period ends and the transient response Monograph Recent Advances in Hydraulic Frac-
progresses toward pseudoradial flow (potentially sev- turing (Gidley et al., 1989). The manuscripts for this
eral years). An obvious implication in this case is that comprehensive volume, with more than 30 contribu-
a standard well test cannot be used to determine frac- tors, were completed in 1984, only five years after
ture length; the length can be determined only from the 1979 SPE annual meeting provided the first
long-term production data. They also identified meaningful number of papers from this research
another important aspect of bilinear flow that occurs effort. The papers presented at this meeting were
because of the transient flow condition within the significant also because they presented a key that
proppant pack: the fracture conductivity can be char- enabled the reliable application of massive fracturing
acterized, independent of length and hence most reli- and rapid progression of the treatment size record
ably, by the slope of a plot of pressure versus the from 2 million lbm in 1979 to more than 7 million
quarter-root of time. lbm by 1986.
Recognition of these consequences for bilinear The key was that, for the first time in its 30-year
flow also explains the difficult transition to the suc- history, fracturing was considered in a framework
cessful application of massive treatments. Well test similar to that used for reservoir characterization.
interpretations misinformed instead of informed. The reservoir framework consists of pressure tran-
They indicated relatively short fracture lengths that sient analysis for the flow characteristics, wireline
were assumed to be treatment placement failures and logs for the formation parameters and geophysics
led to the common and contradicting result: how can for the macroview. The 1979 papers include the fol-
1 million lbm of sand be contained in a fracture lowing (a different reference year indicates the publi-
length of only 100 ft? Much longer propped lengths cation date):
were later substantiated by production data after the • Logging: Rosepiler (1979) introduced application
bilinear period had ended (e.g., values of fracture of the long-spaced sonic tool to infer stress in dif-
half-length xf > 5000 ft; Roberts, 1981). Another ferent layers (see prior discussion of stress con-
contribution to incorrect interpretations was ignoring cerning Appendix Eq. 2 and Chapter 4). Dobkins
Cooke’s (1975) report of very low retained-pack per- (1981) presented improved cased hole logging
meability, which led to overly optimistic estimates of procedures for inferring the fracture height that
conductivity and proportionally pessimistic estimates were also used by Rosepiler to qualitatively vali-
of length. The coupling of these two factors pro- date his novel use of mechanical property logs.
duced incorrect and negative assessments for many
early attempts to establish massive fracturing as a • Pressure transient analysis (PTA): Nolte and Smith
viable means of developing tight gas formations. (1981) introduced the role of pumping pressures by

using a log-log plot as a diagnostic tool (similar to width prediction, provides an analog of the reservoir
PTA practice) for fracture growth characteristics, pressure and reflects the height-averaged minimum
the role of pressure simulation for quantifying stress for the pay zone (see Sidebar 9A). The frac-
geometry (including height growth) and the role of ture width is proportional to the net pressure. The
calibrated stress profiles obtained from mechanical data in Appendix Fig. 7, one of the first recordings
property logs. Nolte (1979) introduced the role of of bottomhole pressure during a treatment, are simi-
pressure during the postinjection closing period to lar to the reservoir response for an injection test with
quantify fluid loss and predict fracture width and a pressure increase (pumping) and subsequent falloff
length by using a specialized time function in a (closing). The injection pressure is governed by the
manner analogous to the Horner plot. The combi- evolving fracture geometry, and the closure data are
nation of these two papers provided a foundation governed by the fluid loss. These two conditions,
for the common use of the calibration treatment respectively, enable characterizing the stored and lost
and pressure-history matching for defining design components of the volume-balance equation shown
parameters (see Chapter 9). Appendix Fig. 7 illus- in Appendix Fig. 2. After closure, the pressure is
trates the fracturing pressure for three distinct phas- independent of the fracture parameters and depends
es: pumping, closing and the after-closure period. on the reservoir response to the fluid lost during the
treatment.
• Geophysics: Smith (1979) introduced the role of
The fundamental analogy between reservoir and
mapping fracture trajectories by using surface tilt-
fracturing behavior results because a diffusion-type
meters and borehole passive seismic techniques to
process governs both behaviors. The respective reser-
improve reservoir recovery by the correct place-
voir and fracturing equivalents are kh/µ → w2h/µ
ment of infill wells (see Section 12-1).
(transmissibility), where k is the permeability, h is the
A companion paper in 1980 showed the synergis- reservoir thickness, w is the width, and µ is the
tic benefit when these individual considerations are appropriate fluid viscosity, and φct → h/(wE) ∝ 1/pnet
unified for tight gas exploitation (Veatch and (storage capacity of the reservoir), where φ is the
Crowell, 1982). porosity, ct is the total system compressibility, and E
is the formation’s elastic modulus. The last expres-
sion for storage contains an inverse proportionality to
Pressure from bottomhole bomb
Inferred pressure
the net fracture pressure pnet. This can be written in
terms of the fracture volume Vf, fluid pressure pf and
Bottomhole pressure, pw (psi)
9000 Fracture Pressure decline

treatment Fracture Transient reservoir closure pressure pc.
closing pressure near wellbore
8000
Fracture closes on 1 dVf 1 1
Net fracture
pressure proppant at well = = (4)
7000 pnet = pw – pc Vf dp f pnet p f − pc
Closure pressure
pc = horizontal rock stress
6000
Reservoir pressure 1 dw 1
= for constant h and L. (5)
5000 w dpnet pnet
38 40 42 44 46 48 50 56 58
Clock time (hr) This equation implies that the elastic formation,
compressed to contain the fracture’s volume, pro-
Appendix Figure 7. Bottomhole fracturing pressure
(Nolte, 1982). duces a system compressibility analogous to an equal
volume of perfect gas at a pressure equal to the frac-
ture’s net pressure. The result is a significant storage
capacity considering typical conditions with more
Fracturing pressure: analog of reservoir than 1000 bbl for fracture volume and only hundreds
response of pounds per square inch for net pressure. The last
An important component of fracturing pressure storage relation, for constant lateral dimensions, is
analysis is the closure pressure. The closure pressure important for a TSO, as discussed later.
is the datum for the net pressure that constrains the

Fracture simulators wells (Warpinski et al., 1996; see Section 12-1). The
importance of these measurements for fracture design
Describing a hydraulic fracture produces a signifi- and evaluation cannot be overemphasized. Indepen-
cantly more complex role for the diffusive process dent measurements for each component of the frac-
than the reservoir case because the basic parameter ture volume (Appendix Fig. 2) provide a long-awaited
groups change continuously with time, with a nonlin- benchmark for validating fracture models.
earity for the equivalent permeability, and the far- Like the first generation’s failure to find a consen-
field elastic coupling between width and pressure sus for width models (e.g., Perkins, 1973), pressure-
produces local parameters that have a general depen- history matching could not resolve the second gener-
dence on the pressure everywhere within the frac- ation’s conflicting adaptations of the P3D framework
ture’s unknown boundaries. For these reasons, frac- (see Chapter 6). The convergence of modeling
ture simulators that rigorously and robustly couple assumptions failed for several reasons. The first was
these parameters in a general manner (see Section fundamental to the pressure-matching process and
6-3) have not progressed at the same rate as reservoir results because of the multitude of opportunities for
simulators. nonuniqueness. Another reason was the failure to
The modeling difficulties led to widespread use of achieve a dominant industry opinion on either the
simulators based on P3D assumptions that partially technique or procedures for a specific technique to
circumvent the far-field elastic-coupling condition. define closure pressure (e.g., Plahn et al., 1997). This
The two most common means were relaxing the lat- state of affairs allowed selecting a closure pressure
eral coupling in the long direction of the fracture (as procedure to validate particular modeling assump-
for the PKN model) to allow a cellular representation tions and therefore justify relatively arbitrary and ad
and vertical height growth of the cells (e.g., Nolte, hoc modeling assumptions. Techniques to determine
1982) or prescribing the boundary and width profiles the closure pressure are discussed in Section 3-6 and
by elliptical segments and a lumped dependence on the Appendix to Chapter 9.
the governing parameters (e.g., Settari and Cleary, Because of nonuniqueness in the reservoir response
1986). P3D models, or more precisely P2D models, and the basing of reservoir models on overly idealized
evolved to include automated proppant scheduling modeling assumptions for a fracture, the reservoir
and the temperature-exposure history for polymer and response cannot generally provide an effective
additive scheduling (e.g., Nolte, 1982), acid fractur- constraint on the achieved fracture length (Elbel and
ing (e.g., Mack and Elbel, 1994), economic optimiza- Ayoub, 1991; Nolte and Economides, 1991). Mapping
tion for treatment design (e.g., Veatch 1986; Meng constraints on all three fracture dimensions provide a
and Brown, 1987), automated pressure-history match- unique, objective test of the geometry model assump-
ing (e.g., Gulrajani and Romero, 1996; Gulrajani et tions (e.g., Gulrajani et al., 1997a) and a basis for
al., 1997b) and rigorous 2D slurry flow (e.g., Smith rationally judging and selecting the model complexity
and Klein, 1995). appropriate for the specific application, available data
Originally restricted to in-office use, these models and simulation resources.
merged with on-site fracture monitoring systems to
provide treatment evaluation and simulation in real-
time mode. An equally important advance was the Treatment design and evaluation
parallel evolution of process-controlled mixing and
blending equipment for reliable execution of more The primary fracture evaluation advance from the
demanding treatment schedules and progressively massive treatment generation is the calibration treat-
more complex chemistry that requires precise pro- ment performed before the proppant treatment to
portioning (see Chapters 7 and 11). define placement parameters. Combining the calibra-
tion treatment and the purpose-designed TSO treat-
ment produced the primary treatment innovation of
Fracture mapping and model validation the second generation. The calibrated TSO treat-
ment, developed by Smith et al. (1984), became the
An important achievement was the definition of frac- key to the third fracturing generation (discussed
ture length, height and width by employing passive later) and essentially removed width as a constraint
seismic measurements and tiltmeters in observation for the conductivity required to successfully fracture

very high permeability formations. This capability

Field Data
and timing produced the overly optimistic prediction
in 1985 for the beginning of the TSO generation, as Variable injection rate
Net pressure, pnet (psi)

2000
indicated by Appendix Fig. 1a. Proppant
begins II III-a
I
Transition between the second and

1000 * Proppant
begins
x
II
III-b
[5 MPa] II IV
third generations I
III-a
500
The following paragraphs link several aspects of the 40 60 100 200 400 600 1000
massive and TSO generations by using the informa- Time (min)
tion available from the diagnostic log-log plot for
fracturing in Appendix Fig. 8. Appendix Table 1 lists
Idealized Data
the interpretations for various slopes exhibited in the III-b
figure by the net pressure during fracturing. The data
log pnet
II
III-a
I
are from two massive treatments in tight gas forma- IV
tions. The top curve is a treatment in the Wattenberg Inefficient extension for
pnet ≥ formation capacity pfc
field, the first microdarcy-permeability field develop-
ment (Fast et al., 1977). The behavior shown by the log time or volume
lower treatment curve, which was designed by this
author, provided insight for developing the TSO Appendix Figure 8. Log-log diagnostic plot for fracturing
(Nolte, 1982).
treatment that enables successfully fracturing darcy-
scale oil formations. The treatment related to the
lower curve was not particularly successful. How- capacity. The capacity (Nolte, 1982) defines the pres-
ever, it was one of the first 2 million lbm treatments sure limit for efficient fracture extension and is anal-
and hence functioned better as a “sand-disposal” ogous to the pressure-capacity rating for a pressure
treatment than a gas-stimulation treatment. The sand vessel. The cited reference has an unsurprising
was disposed of with 900,000 gal of crosslinked fluid theme of the negative effects of excesses of pressure,
containing 90 lbm/1000 gal of polymer, or approxi- polymer and viscosity.
mately 80,000 lbm of polymer. Three mechanisms for a formation can define its
The marginal success of the treatment is readily pressure capacity before “rupture” accelerates fluid
understood by considering Appendix Eq. 3. For the loss from the formation’s pay zone. The subsequent
treatment average of 2.1 ppa, the equation predicts fluid loss also leaves proppant behind to further
1900 lbm/1000 gal crosslinked fluid (in reality, a enhance slurry dehydration and proppant bridging.
solid) remaining in the proppant pack porosity after Each mechanism is defined by the in-situ stress state
the treatment. However, the size and viscosity for and results in a constant injection pressure condition,
this treatment provided an ideal test condition of or zero log-log slope, when the net pressure reaches
how a formation responds to fluid pressure and an the mechanism’s initiation pressure. The mecha-
excellent illustration for the concept of formation nisims are
Appendix Table 1. Slopes of fracturing pressures and their interpretation in Appendix Fig. 8.
Type Approximate log-log slope value Interpretation
I ⁄8 to 1⁄4
1
Restricted height and unrestricted expansion
II 0 Height growth through pinch point, fissure opening
or T-shaped fracture
III-a 1 Restricted tip extension (two active wings)
III-b 2 Restricted extension (one active wing)
IV Negative Unrestricted height growth

• opening the natural fissures in the formation, gov- the required penetration could be obtained by not
erned by the difference in the horizontal stresses exceeding the formation capacity. A subsequent treat-
ment designed using 150,000 gal and 900,000 lbm of
• extending the height through a vertical stress bar-
sand (an average of 6 ppa) became the prototype for
rier and into a lower stress (and most likely per-
the remaining development of the field (Nolte, 1982).
meable) zone, governed by the difference in the
The lower curve on Appendix Fig. 8 is for the
horizontal stress for the barrier and pay zone
aforementioned sand-disposal treatment in the Cot-
• initiating a horizontal fracture component when ton Valley formation of East Texas. As previously
the pressure increases to exceed the level of the discussed, the treatment provided an opportunity to
overburden stress. observe a large range of fracturing behavior with five
An important observation for the pressure capacity types of interpretive slopes occurring, including
is that it depends on the in-situ stress state and there- • Type I indicating extension with restricted height
fore does not change for the formation in other well growth
locations unless there are significant local tectonic
effects. As a result, all future treatments for the field • Type II defining this formation’s lowest pressure
can generally be effectively designed on the basis of capacity at 1000 psi for the penetration of a stress
only one bottomhole pressure recording and its barrier
detailed analysis (see Section 9-4). • Type IV, with decreasing pressure, indicating unre-
The upper curve on Appendix Fig. 8, for the stricted vertical growth through a lower stress
Wattenberg treatment, illustrates the fissure-opening zone after the barrier was penetrated.
mechanism with the Type II zero slope occurring at
a net pressure of 1700 psi. This value provides one of The Type IV condition continued until proppant
the largest formation capacities ever reported. The fis- was introduced. Almost immediately after proppant
sure opening is preceded by restricted height growth entered the fracture the pressure increased, most
and unrestricted extension (Type I slope) that provide likely because the proppant bridged vertically in the
the most efficient mode of fracture extension. There- width pinch point formed by the penetrated stress
fore, conditions in this formation are favorable for barrier and restricted additional height growth.
propagating a massive fracture; not by coincidence, During the preceding 6-hr period of significant verti-
this was the first field successfully developed in the cal growth, the horizontal growth was retarded. As a
massive treatment generation (Fast et al., 1997), and it result, the very high polymer concentration formed a
provided incentive to continue the development of thick polymer filter cake at the fracture tip that proba-
massive treatment technology. Returning to Appendix bly restricted further horizontal extension. Thus, the
Fig. 8, after the period of constant pressure and extremities of the fracture were restricted either by
enhanced fluid loss, a Type III-a slope for a fracture proppant or polymer cake, and continued injection
screenout occurs because slurry dehydration forms was stored by increasing width indicated by the Type
frictional proppant bridges that stop additional exten- III-a unit slope. After a significant increase in pres-
sion (i.e., a generally undesired screenout for a tight sure, the pressure became constant for a short period
formation requiring fracture length over conductivity). at 1200 psi with a Type II slope that probably resulted
After the penetration is arrested, the major portion of from opening natural fissures to define a second,
the fluid injected is stored by increasing width (see higher capacity. Subsequently the slope increased to
Appendix Eq. 4) and the net pressure develops the unit an approximately 2:1 slope indicated as Type III-b.
slope characteristic of storage. The amount of width This latter slope for a storage mechanism indicates
increase is proportional to the net pressure increase. that about one-half of the fracture area had become
The Wattenberg treatment consisted of 300,000 gal restricted to flow, which could have resulted from one
of fluid and 600,000 lbm of sand with an average wing of the fracture being blocked to flow near the
concentration of 2 ppa, similar to the previous exam- well because of slurry dehydration from the fissure
ple. However, the treatment was successful because fluid loss. The wellbore region experiences the largest
a polymer-emulsion fluid with low proppant pack pressure and is most prone to adverse fluid-loss
damage was used. After the treatment defined the effects from exceeding a capacity limit.
formation capacity, model simulations indicated that

Subsequent treatments were improved after under- The third generation: tip-screenout
standing the formation’s pressure behavior as in the
Wattenberg case and for this area after understanding treatments
the implications of Appendix Eq. 3 for concentrating A proper historical perspective of this third genera-
polymer. In addition, the observation that proppant tion requires perspective from the next generations;
bridging could restrict height growth was developed however, several of its developments are reviewed
for treatments to mitigate height growth (Nolte, here. A more comprehensive presentation and refer-
1982). An effective and relatively impermeable ence are by Smith and Hannah (1996).
bridge can be formed within the pinch point to retard Demonstration of the ability to routinely place a
height growth by mixing a range of coarse and fine successful TSO treatment opened the door for effec-
sand for the first sand stage after the pad fluid. tive fracture stimulation of higher permeability
Smith et al. (1984) later sought a means to signifi- formations. Another component for the successful
cantly increase fracture width for the development of fracturing of high permeability was the continued
a chalk formation within the Valhall field in the development of synthetic proppants that can produce
Norwegian sector of the North Sea. The additional a cost-effective 10-fold increase in permeability rela-
width was required because laboratory tests indicated tive to sand for higher closure stresses (see Chapter
the likelihood of substantial proppant embedment 7). Coupling this increase in permeability with the
into the soft formation that would lead to the loss of similar increase for propped width achieved by a
effective propped width. Fracturing was considered TSO treatment in a moderate- to low-modulus for-
for this formation because other completion tech- mation provides about a 100-fold increase in con-
niques would not sustain production because of chalk ductivity over a conventional sand fracture. The
flow. The resulting treatment design was based on the conductivity increase also translates into a 100-fold
behavior on the log-log plot in Appendix Fig. 8 for increase of the target permeability for fracturing, as
the sand-disposal treatment: a purpose-designed TSO implied by Appendix Figs. 5 and 6. The increases for
treatment. For the disposal treatment, they observed width and conductivity also mitigate nondarcy (or
that after the initial screenout occurred, 2 million lbm turbulent) flow effects in the fracture for high-rate
of proppant could be placed, and the net pressure wells, particularly gas wells (see Sections 10-7.3 and
increase indicated that this occurred by doubling the 12-3.1).
width after the screenout initiated. However, the anticipated growth rate shown on
Smith et al. designed and successfully placed a Appendix Fig. 1a was slowed not only by the unan-
TSO treatment in which proppant reached the tip and ticipated, extensive contraction of activity in general,
bridged to increase the width by a factor of 2 during but also by two prevailing mind sets: high-perme-
continued slurry injection after the purpose-designed ability formations cannot be successfully fracture
TSO occurred. This design, with successful place- stimulated and why fracture a commercial well?
ment of progressively larger propped width increases, Additional field proof for the benefits of a TSO treat-
became the tool that enabled the development of this ment came from two successful programs: a signifi-
formation. The ability to significantly increase the cant improvement over conventional fracture treat-
width after screenout results from the large storage ments for the Ravenspurn gas field in the southern
capacity of a fracture, as detailed in the discussion North Sea (Martins et al., 1992b) and high-perme-
following Appendix Eqs. 4 and 5. Additional discus- ability applications in the Prudhoe Bay field (Hannah
sion on the fracture completion in Valhall field and and Walker, 1985; Reimers and Clausen, 1991;
the TSO treatment is in the “Reservoir and Water Martins et al., 1992a).
Management by Indirect Fracturing” section.
As a historical note, a similar concept for a TSO
was disclosed in a 1970 patent (Graham et al., Deep damage
1972), with the bridging material consisting of petro- Fracturing in Prudhoe Bay was particularly successful
leum coke particles (approximately neutral density to because deep formation damage induced by prior pro-
ensure transport to the extremities). The patent’s goal duction (i.e., beyond the reach of matrix treatments)
was increased width to enable placing larger size facilitated sidestepping the mind set of not applying
proppant in the fracture.

fracturing to high permeability. The incremental pro- fracture width that mechanically must continue
duction from only one year of the fracturing program around the wellbore; i.e., if the formation is pushed
would have ranked as the 10th largest producing field apart 2 in. over the large surface area of the fracture,
in the United States (e.g., Smith and Hannah, 1996), the rock around the wellbore must be displaced
without including similar results achieved by another accordingly. For a well-designed and executed frac
operator in the other half of the field. Another signifi- and pack, the initiating screenout at the tip is pro-
cant aspect of the Prudhoe Bay application is that the gressively packed back to the well to completely
fractures were routinely placed in a relatively small oil pack the resulting ring.
zone above a rising water zone without entering the The continuing success of the initial frac and
water zone (Martins et al., 1992a), which demon- packs started a rapid conversion to this completion,
strated that fracturing is a viable, potentially superior with the frac and pack becoming the preferred Gulf
alternative to matrix treatments in high-permeability of Mexico sand control completion. In addition to
formations. This precise fracturing was achieved by continued use offshore Indonesia, technology trans-
coupling an initial detailed fracture modeling study fer resulted in a wider geographical distribution for
with a calibration treatment before each proppant this sand control technique (e.g., West Africa,
treatment. Gulrajani et al., 1997b).
As for other applications of TSO treatments, on-site
redesign after a calibration treatment became a stan-
Frac and pack dard frac and pack practice. An important observation
The frac and pack completion consists of a TSO is that the same analysis procedures and design mod-
treatment before a conventional gravel pack. During els introduced for the massive treatments of tight gas
the early 1990s, frac and pack treatments were formations in the late 1970s were transferred directly
applied on a limited basis around the world, notably to frac and pack treatments in soft formations.
offshore Indonesia. Prior to the TSO treatment era,
this technique was tried at various times but without
External gravel pack
sustained success. The large propped width from a Casing connecting all perforations
TSO treatment was a necessary ingredient for suc- with propped fracture
cessful frac and pack applications, as discussed later.
The frac and pack boom was in the Gulf of
Mexico. The first successful application began
because of economic considerations and therefore Packed-back
overcame the mind set of not fracturing a commer- fracture
cial well. A significant field development was not
meeting production expectations because standard Appendix Figure 9. Successfully packed-back TSO treat-
ment.
gravel-packed completions could not consistently
achieve a low skin effect; the skin effect ranged
between 7 and 30. The skin effect was 10 after the
first frac and pack treatment and progressively
decreased to near zero from improvements in the
Reservoir and water management
treatment design and the use of larger size proppant
by indirect fracturing
(Hannah et al., 1994). Another application of TSO treatments is reservoir
The threefold-plus increase in production rate, by management. The prototype example for this applica-
eliminating the skin effect, resulted from more than tion was in the Norwegian Gullfaks field (Bale et al.,
just adding a TSO treatment to the procedure. An 1994a, 1994b). The reservoir section had a multi-
important feature of a frac and pack is reduction of darcy-permeability upper zone that graded downward
the inherent flow restriction around and through the to a permeability of about 100 md. The standard com-
perforations. The ring of proppant around the casing pletion was to perforate and gravel pack the upper
(Appendix Fig. 9) acts as an excellent external gravel zone. However, an edge-water drive would encroach
pack for reducing the pressure drop through the per- through the high-permeability zone and turn a prolific
forated region. The ring results from the large TSO oil well into an even higher water producer.

A solution was found from the pioneering work from the lower sections by fracture stimulation and
of the Valhall TSO treatment discussed for Appendix a significant increase in drawdown. This application
Fig. 8. This application in the early 1980s was for completes the link between the sand-disposal thight
more than mitigating proppant embedment. The pri- gas treatment in Appendix Fig. 8 to reservoir and
mary objective was for controlling chalk production water management with the intermediate develop-
from the primary producing zone above where the ment of the TSO-IVFC for solids control in the
TSO treatment was placed. The upper chalk zone Valhall field.
was very soft with high porosity and composed of
almost as much oil as chalk. When this zone was put
on production, chalk filled the tubing and led to cas- Screenless sand control
ing collapse. The zone was produced by placing the Another apparent role of the IVFC is to eliminate the
TSO treatment in the more competent zone below need for a screen in many sand-control environments
and extending the fracture height into the bottom of by selecting and perforating only competent sections
the very high porosity formation. This completion within or near the unconsolidated sections of the for-
enabled chalk-free production from both the upper mation. The zone selection method can potentially be
and lower zones (Smith et al., 1984). enhanced by a sonic log application. This application
This indirect access to the primary producing zone takes advantage of the generally considered negative
has come to be known as an indirect vertical fracture effect of near-wellbore refracted and relatively slower
completion (IVFC) and is illustrated in Appendix waves caused by the wellbore mechanical damage
Fig. 10. The technique of perforating and fracturing that routinely occurs in weak or highly stressed for-
only from competent sections and producing from mations (Hornby, 1993). However, for screenless
incompetent sections is a robust method for control- completions, the negative effect becomes a positive
ling the production of formation material and effect because the change in the wave speed for the
increasing recovery from the lower permeability refracted wave is a direct indication of the state of
zones by fracture stimulation. From this perspective, rock failure around the well, which is caused by the
a TSO-IVFC becomes a solids control and reservoir wellbore stress concentration within the in-situ stress
management application (see Section 5-1.2). field. Therefore, the layers with a minimal near-well
The Gullfaks adaptation by Bale et al. (1994a) change in wave speed relative to the far-field speed
also placed a TSO-IVFC in a lower competent part are the more competent candidate zones for perforat-
of the formation. In addition to providing sand coning and applying a TSO-IVFC to achieve screenless
trol and managing reservoir depletion, it was a water formation-material-controlled production.
management treatment because it delayed water A second method of achieving a screenless sand-
breakthrough and greatly increased reserves recovery control completion is applied without strategically
placed perforations (e.g., Malone et al., 1997). This
method couples the proppant ring around the casing
from a TSO treatment and proppant with effective
High flowback control (e.g., fibers, curable-resin-coated
permeability
proppant or both). The combination with a success-
ful packed-back TSO achieves an external gravel
Propped Low or pack of stable proppant (i.e., an external formation-
fracture moderate material screen as illustrated by Appendix Fig. 9).
permeability
Perforation and completion considerations are
addressed in Section 11-3.5.
The screenless completion obviously eliminates
the cost of the screen and related tools, but more
importantly it enables economic development of sig-
nificant behind-pipe reserves that do not warrant the
mobilization and operational costs for a rig on an
Appendix Figure 10. Indirect vertical fracture for reservoir offshore production platform, as generally required
management (Bale et al., 1994a).
for a standard gravel-pack completion.

A future generation: fracturing and of using cemented casing for effective treatment
diversion tends to be overlooked because of an appar-
reservoir engineering merger? ent failure to appreciate lifecycle economics or the
The previous discussion of the TSO generation clearly effectiveness of good cementing techniques (see
shows the blurring of what can be controlled on the Chapter 11). Staged fracturing, from correctly placed
inside and outside of the casing and of what have perforated sections, enables highly effective damage
been the traditional roles of a fracture design engineer bypass, as demonstrated by the first fracturing genera-
and a reservoir engineer. This blurring of past distinction’s rate of 100 treatments per day in 1955.
tions provides prospects for additional innovations The general benefit for a horizontal well, particularly
and the advent of a fourth fracturing generation. with vertical variations of permeability, is magnified
by the fracture adding a large vertical permeability
component (see Chapters 11 and 12). Simply stated,
Optimal reservoir plumbing an extended reach well cannot drain what it is not con-
From a broader viewpoint, the IVFC and strategically nected to nor can it efficiently drain what it is isolated
placed perforations provide the means to extend opti- from by wellbore damage. The addition of a vertical
mized plumbing into the reservoir. Optimized plumb- fracture allows efficient drainage of all isolated sec-
ing, through a NODAL analysis, is generally practiced tions that the propped fracture reaches. The location
only for the surface facilities and within the wellbore. of the fracture, or plumbing source, can be specified by
Extended optimization requires additional considera- correctly placed perforations within a cemented casing
tions for designing the plumbing system provided by and an effective fracture design and execution. Cased
the fracture in the reservoir and also within the frac- hole logging and logging while drilling can be used
ture itself. to identify IVFC target locations for connection to by-
The outline for these considerations was defined by passed reserves and management for their exploitation.
Bale et al. (1994b) for the Gullfaks application. They
considered the role of the permeable fracture plane on
Fracturing for well testing
the reservoir’s 3D flow pattern and how tailoring the
distribution of conductivity can advantageously affect The after-closure portion of Appendix Fig. 7, labeled
this flow pattern (e.g., reducing the conductivity as the “transient reservoir pressure near wellbore,” shows the
fracture approaches the high-permeability upper zone return of the fracturing pressure to the reservoir pres-
to delay water production while increasing the con- sure and demonstrates the well testing potential for
ductivity in the lower permeability zone and applying any injection above fracturing pressure. This potential
a large drawdown to accelerate production from this for a fracture is ensured by the well-known result that
zone; see Section 5-1.2). Therefore, the analysis and the long-term reservoir response is pseudoradial flow
design tools have evolved for considering the role of (e.g., Cinco-Ley and Samaniego-V., 1981b) and is the
fractures in NODAL analysis for reservoir, formation same flow regime used for standard well testing. An
material and water management. attractive aspect of the use of fracturing for testing is
that the fracture enhances the likelihood that all the
zones are open and captured by the test. This is an
Achieving full potential for horizontal important consideration for layered formations and
wells and laterals particularly thinly layered zones that can be missed
The preceding discussion of the IVFC is in the con- by open perforations. Another attraction of fracturing
text of single, essentially vertical wells. The potential or injection testing is that the wellbore is generally
for innovative strategies to drain a reservoir increases filled with water that provides minimal wellbore stor-
several fold by adding consideration of horizontal and age and formation volume factor effects.
lateral wells. These highly deviated wellbores are typ- The long-term radial response following fracture
ically placed without cemented casing because of eco- closure was developed and presented in a pair of
nomic considerations and therefore do not generally papers: Gu et al. (1993) from the application perspec-
reach their full potential because they lack an effective tive and Abousleiman et al. (1994) from the theoreti-
technique to remove wellbore damage. The solution cal perspective. They recognized that the radial

response from fracture injection met the assumptions tifying spurt loss is particularly important for high-
for a slug (or equivalently an impulse) test and that permeability formations and is not practically attain-
they could directly apply this developed area of reser- able by any other means than after-closure analysis.
voir technology. The after-closure analyses are presented in Section
Another well-known flow regime for a fracture is 9-6, and a method to quantify reservoir parameters
pseudolinear flow. Incorporating the analysis of this during the closure period is presented in Section 2-8.
after-closure flow regime was the last link of the frac- These applications from the reservoir behavior of
turing–pressure analysis chain between the beginning fracturing complement the 1979 adoption of reservoir
of injection and returning to reservoir pressure. methodologies and achieve a direct merging of frac-
Consideration of this regime by Nolte et al. (1997) turing into the classic realm of reservoir testing and
indicated that the reservoir “memory” of the fractur- characterization (see Chapters 2 and 12). Reservoir
ing event can validate several aspects for analysis of characterization from a calibration testing sequence
a calibration treatment (e.g., closure time and hence to define fracturing parameters provides the ingredi-
the critical closure pressure, fracture length and hence ents essential for on-site, economics-based treatment
the fluid-loss coefficient, and the division of fluid loss optimization.
between normal wall diffusion and tip spurt). Quan-

◆
Mechanics of
Mark G. Mack, Schlumberger Dowell
Norman R. Warpinski, Sandia National Laboratories
6-1. Introduction allel rock faces with roughness comparable to frac-

tured rock, no assumptions need to be made about
The mechanics of hydraulic fracturing is a convenient how the fluid behaves rheologically. Instead, how
description of the processes and mechanisms that are it behaves is simply observed. Unfortunately, phys-
important to fracturing technology. Mechanics gener- ical models are usually expensive to build and use.
ally refers to an engineering discipline that is con- In addition, there are major issues of scale-up if
cerned with the mechanical properties of the material the model is significantly smaller than the actual
under consideration and the response of that material structure. For example, in a model of a bridge, the
to the physical forces of its environment. Hydraulic weight is proportional to the scale factor cubed,
fracturing is complicated because it involves four dif- but the length of any element is proportional only
ferent types of mechanics: fluid, solid, fracture and to the scale factor. Thus, even elements that do
thermal. In fracturing, fluid mechanics describes the not fail in the model may fail in practice. Never-
flow of one, two or three phases within the fracture; theless, scale models are useful provided an appro-
solid mechanics describes the deformation or opening priate dimensional analysis is performed and if the
of the rock because of the fluid pressure; fracture scale factor is not too great (de Pater et al., 1993).
mechanics describes all aspects of the failure and
• Empirical models are developed by observation.
parting that occur near the tip of the hydraulic frac-
Typically, laboratory or field data are gathered and
ture; and thermal mechanics describes the exchange
combined to create design charts or empirical
of heat between the fracturing fluid and the forma-
equations, which can then be used to predict or
tion. Furthermore, the responses are coupled and
design future cases. For example, if 100 wells in
depend on each other. To develop tools for the design
an area have been fractured with different-size
and analysis of a process as complicated as hydraulic
treatments, 6 months of cumulative production
fracturing, it is necessary to build models that
could be plotted against treatment size. Provided
describe each of the responses sufficiently. This chap-
the scatter is not too great, the production response
ter describes the history and technology associated
from a new treatment can be predicted from the
with these models.
historical data.
A model of a process is a representation that cap-
The advantages of empirical models are that no
tures the essential features of the process in a manner
assumptions need to be made about any behavior
that provides an understanding of the process (Star-
and there is no scale effect. The primary disadvan-
field et al., 1990). The construction of the model
tage is low confidence in extrapolation outside the
depends on the type of question it is required to
range of the data. The 100-well data set may be
answer. The three main types of models are physical,
useful in the same field, even for treatments slightly
empirical and mechanistic (or analytic). Each has
larger than any in the data set, but is most likely
advantages and disadvantages, which are outlined
irrelevant in another area. For an empirical model
in the following.
to be useful, the data must be arranged in terms
• Physical models are scale models of actual of suitable dimensionless variables, so that it is
processes. The primary advantage of such models as general as possible. For example, the 100-well
is that, by definition, they incorporate the correct data set may be useful in a different area provided
assumptions of material behavior. For example, if the results are normalized with respect to perme-
a fracturing fluid is pumped between a pair of par- ability and pay thickness. To obtain the right

dimensionless quantities, however, it is generally • perform economic optimization (i.e., determine

necessary to have at least some understanding of what size treatment provides the highest rate of
the mechanics of the physical process being modeled. return on investment)
• Analytical models are mathematical representations • design a pump schedule
of a physical reality in which the governing mech- • simulate the fracture geometry and proppant place-
anics are stated in the form of equations. The equa- ment achieved by a specified pump schedule
tions typically define both physical laws, such as
• evaluate a treatment (by comparing the predictions
conservation of mass, and constitutive laws, such
of a model with actual behavior).
as elasticity. The former are inviolable laws of
nature, whereas the latter are hypotheses of physical In each of these cases, the objective is a quantitative
behavior that require laboratory work to confirm estimate of either the volume of fluid and proppant
that they are applicable and to determine the con- required to create a fracture with a desired conduc-
stants in the model. tivity and geometry or the geometry produced by a
The major advantages of analytical models are specified pump schedule.
that they may be extrapolated outside the range in
which they were developed, provided the various
component laws still hold. Thus, if the elastic con- 6-2. History of early hydraulic
stant of a spring has been measured, the force fracture modeling
required for a given displacement of the spring can
be predicted, even for displacements that have not 6-2.1. Basic fracture modeling
been tested. If the spring is cut in half, the behavior
of each half can be predicted. Perhaps the greatest Sneddon (1946) and Sneddon and Elliot (1946) devel-
limitation of analytical models, however, is the oped the solutions for the stress field and pressure
assumptions that are made in developing the model. associated with static pressurized cracks. They
For example, it is typically assumed that rock is showed that the width of a static penny-shaped (i.e.,
homogeneous, yet there are many cases where it is circular) crack of radius R under constant pressure is
fractured or otherwise variable from point to point, given by the expression
and this may not be accounted for in the model. 8 pnet R(1 − ν 2 )
w( r ) = (1 − (r R) ,
2
A simulator is a computational implementation of (6-1)
πE
a model. Many analytical models are tractable only if
they are solved numerically, unless a large number of which describes an ellipsoid, and the volume of the
approximations or simplifying assumptions are made. crack V by
With the widespread availability of computers, it is 16(1 − ν 2 ) R3
now generally accepted that better answers may be V= pnet , (6-2)
obtained by numerically solving a more general model 3E
rather than by solving a simplified model exactly. where the net pressure pnet is defined as the pressure in
Nevertheless, it must be emphasized that useful rules the crack minus the stress against which it opens, ν is
of thumb and relations between quantities can often Poisson’s ratio, and E is Young’s modulus. Sack
be developed much more easily using analytic solu- (1946) showed that the pressure required to extend a
tions, which provide insight into the relations between crack of radius R under constant pressure is given by
parameters affecting the results for more complex πγ F E
pnet = ,
2(1 − ν 2 ) R
conditions. Some of the simplest rules would probably (6-3)
not be “discovered” from a numerical solution without
a great deal of effort, if at all. An extensive presenta- where γ F is the specific fracture surface energy.
tion of analytic-based solutions and approximations Equations 6-1 and 6-2 are derived using the theory
for the mechanics of hydraulic fracturing was provided of linear elasticity, and Eq. 6-3 is derived using linear
by Valkó and Economides (1996). elastic fracture mechanics. The basis of Eq. 6-3 is that
Four important reasons for developing and using the energy required to create the surface area when a
models of hydraulic fracture treatments are to crack is propagated must equal the work done by the
6-2 Mechanics of Hydraulic Fracturing

pressure in the crack to open the additional width. A ing the importance of different aspects. Carter (1957)
more detailed discussion of fracture mechanics is in neglected both fluid viscosity effects and solid
Chapter 3. Combining Eqs. 6-2 and 6-3, Perkins and mechanics and concentrated on leakoff. Khristianovich
Kern (1961) showed that the pressure for propagation and Zheltov (1955) made some simplifying assump-
of a radial fracture is tions concerning fluid flow and focused on fracture
1/ 5 mechanics. Perkins and Kern (1961) assumed that
 2π3γ 3 E 2  fracture mechanics is relatively unimportant and
pnet = F
 . (6-4)
 3(1 − ν ) V 
2 2 focused on fluid flow. These three basic models are
each described in some detail in following sections.
Thus, if the fracture volume is known, pnet can be The first work on hydraulic fracture modeling was
calculated and Eq. 6-2 used to determine R. For exam- performed by several Russian investigators (summa-
ple, if the injection rate qi is constant, fluid friction in rized by Khristianovich et al., 1959). The first refer-
the fracture is negligible, and there is no leakoff, Eq. ence in English is Khristianovich and Zheltov’s (1955)
6-4 can be substituted into Eq. 6-2 with the volume paper. The other major contribution was the work of
V replaced by qit as Perkins and Kern (1961). These models were devel-
16(1 − ν 2 ) R3  2 π 3 γ 3F E 2 
1/ 5
oped to calculate the fracture geometry, particularly
qi t =   , (6-5) the width, for a specified length and flow rate, but did
 3(1 − ν ) qi t 
2 2
3E not attempt to satisfy the volume balance. Carter
(1957) introduced a model that satisfies volume bal-
where t is the time. Rearranging and solving for R, ance but assumes a constant, uniform fracture width.
 
1/ 5 This model was used into the late 1970s for determin-
9 Eqi2 t 2
R= 2 
. (6-6) ing volume balance, with more realistic width profiles
128πγ F (1 − ν )  from the aforementioned geometry models to ensure
Sneddon and Elliot (1946) also showed that for that the fracture width was sufficient for proppant
fractures of a fixed height hf and infinite extent (i.e., entry. This approach was made obsolete by extensions
plane strain), the maximum width is to the Khristianovich and Zheltov and Perkins and
Kern models developed by Geertsma and de Klerk
2 pnet h f (1 − ν 2 ) (1969) and Nordgren (1972), respectively. These two
w= (6-7)
E basic models, generally known as the KGD and PKN
and the shape of the fracture is elliptical, so that the models after their respective developers, were the first
average width — w = (π/4)w. The term E/(1 – ν2) ap- to include both volume balance and solid mechanics.
pears so commonly in the equations of hydraulic frac- The PKN and KGD models, both of which are
turing that it is convenient to define the plane strain applicable only to fully confined fractures, differ in
modulus E´ as one major assumption: the way in which they convert
E a three-dimensional (3D) solid and fracture mechanics
E′ = , (6-8) problem into a two-dimensional (2D) (i.e., plane strain)
1 − ν2 problem. Khristianovich and Zheltov assumed plane
which is used for this chapter. (A plane strain defor- strain in the horizontal direction; i.e., all horizontal
mation is one in which planes that were parallel cross sections act independently or equivalently, and
before the deformation remain parallel afterward. This all sections are identical (Fig. 6-1), which is equiva-
is generally a good assumption for fractures in which lent to assuming that the fracture width changes much
one dimension [length or height] is much greater than more slowly vertically along the fracture face from
the other.) any point on the face than it does horizontally. In
practice, this is true if the fracture height is much
greater than the length or if complete slip occurs at the
6-2.2. Hydraulic fracture modeling boundaries of the pay zone. Perkins and Kern, on the
Several introductory and key papers published between other hand, assumed that each vertical cross section
the late 1950s and early 1970s that developed the acts independently (Fig. 6-2), which is equivalent to
foundation of hydraulic fracture modeling approach assuming that the pressure at any section is dominated
the problem by making different assumptions concern- by the height of the section rather than the length of

• Derivation of Perkins and Kern model

of a vertical fracture
L(t)
Perkins and Kern (1961) assumed that a fixed-
w(x,t)
height vertical fracture is propagated in a well-
confined pay zone; i.e., the stresses in the layers
ux
above and below the pay zone are sufficiently large
Approximate
ww = w(o,t) shape of fracture to prevent fracture growth out of the pay zone.
rw They further assumed the conditions of Eq. 6-7, as
shown in Fig. 6-2, that the fracture cross section is
elliptical with the maximum width at a cross section
proportional to the net pressure at that point and
independent of the width at any other point (i.e.,
hf
vertical plane strain). Although Perkins and Kern
developed their solution for non-Newtonian fluids
and included turbulent flow, it is assumed here that
the fluid flow rate is governed by the basic equation
Figure 6-1. KGD fracture. for flow of a Newtonian fluid in an elliptical section
(Lamb, 1932):
dp 64 qµ
=− , (6-9)
dx πh f w 3
L(t) where p is the pressure, x is the distance along the
fracture, and µ is the fluid viscosity.
Substituting Eq. 6-7 into Eq. 6-9, replacing the
flow q by one-half of the injection rate (qi/2) and
assuming that the flow rate is constant along the
fracture length (which implies that both leakoff and
w(x,t) storage in the fracture resulting from width increases
are neglected) obtains
w(o,t)
4 µqi E ′ 3
3
pnet dpnet = − dx . (6-10)
hf π h f4
Integrating this expression along the fracture
half-length L obtains, with pnet = 0 at the fracture
tip,
1/ 4
Figure 6-2. PKN fracture. 16µqi E ′ 3 
pnet =  L , (6-11)
 πh f
4

the fracture. This is true if the length is much greater from which Eq. 6-7 implies that
than the height. This difference in one basic assump-
 µq ( L − x ) 
1/ 4
tion of the models leads to two different ways of solv- w( x ) = 3 i  . (6-12)
ing the problem and can also lead to different fracture  E′ 
geometry predictions. In the case of the PKN model,
In oilfield units (with qi in bbl/min and w in in.),
fracture mechanics and the effect of the fracture tip are
the width at the wellbore (x = 0) is
not considered; the concentration is on the effect of
q µL
1/ 4
fluid flow in the fracture and the corresponding pres- ww = 0.38 i  . (6-13)
sure gradients. In the KGD model, however, the tip  E′ 
region plays a much more important role, and the fluid For this model, the average width in the fracture
pressure gradients in the fracture can be approximated. is π/4 (about 80%) of the wellbore width. With a

Newtonian fluid, the model width is independent of where qL is the leakoff rate over the whole fracture
the fracture height. and qf is the volume rate of storage in the fracture.
Perkins and Kern (1961) noted that the average – is assumed to be constant in
If the fracture width w
net pressure in the fracture would greatly exceed both space and time, Eq. 6-15 can be written as
the minimum pressure for propagation, calculated
∂A f
Af ( t )
by an equation similar to Eq. 6-4, unless the fluid
flow rate was extremely small or the fluid had an
qi = 2
0
∫ u dA ∂t
L ,f +w (6-16)
unrealistically low viscosity. Thus, under typical where Af is the fracture face area. Carter showed
hydraulic fracturing conditions, the pressure result- that Eq. 6-16 can be rewritten as
ing from fluid flow is far larger than the minimum t
∂A f ∂A
pressure required to extend a stationary fracture. qi = 2 ∫ uL (t − λ ) dλ + w f . (6-17)
This justifies neglecting fracture mechanics effects 0
∂λ ∂t
in this model. Furthermore, they pointed out that
Substituting Eq. 6-14 into Eq. 6-17 and using
the fracture would continue to extend after pump-
Laplace transformations, he showed that this could
ing stopped, until either leakoff limited further
be solved to obtain
extension or the minimum pressure for fracture
propagation was reached. qi w  s 2 
Af = e erfc( S ) + S − 1 ,
2
2 
(6-18)
Several important observations concern this 4 πCL  π 
solution:
where
– assumption of plane strain behavior in the verti-
cal direction 2CL πt
S= . (6-19)
– demonstration that fracture toughness could be w
neglected, because the energy required to propa-
The fracture wing length L as a function of time
gate the fracture was significantly less than that
is then obtained by dividing the area by twice the
required to allow fluid flow along the fracture
fracture height. Harrington and Hannah (1975)
length
showed (see Sidebar 6A) that Eq. 6-18 could be
– assumption that leakoff and storage or volume simplified with little loss of accuracy to
change in the fracture could be neglected
qi t
– assumption of fixed height Af = , (6-20)
w + 2 CL 2 t
– no direct provision of fracture length as part of which is much easier to work with for simple cal-
the solution. culations.
• Inclusion of leakoff Designs were performed by iterating between the
Although Perkins and Kern (1961) suggested that Carter technique to obtain the fracture length as a
their paper could be used in practical applications, function of time (Eq. 6-19) and the Perkins and
they neglected both leakoff and storage of fluid in Kern model to determine the width (Eq. 6-13) until
the fracture. They assumed that some other method a consistent solution was found, and then Eq. 6-11
would be used to calculate the fracture length, such was used to determine the pressure.
as that proposed by Carter (1957). Nordgren (1972) added leakoff and storage with-
Carter introduced the basic equation for leakoff, in the fracture (resulting from increasing width) to
which is discussed in detail in Section 6-4. The the Perkins and Kern model, deriving what is now
leakoff velocity uL at a point on the fracture wall is known as the PKN model. To add storage and
CL leakoff, the equation of continuity (i.e., conserva-
uL = , (6-14) tion of mass) is added to the set of equations (6-7
t − texp and 6-9) used by Perkins and Kern:
where CL is the leakoff coefficient, t is the current ∂q ∂A
time, and texp is the time at which point uL was + qL + = 0, (6-21)
∂x ∂t
exposed. Carter introduced a simple mass balance:
qi = q L + q f , (6-15)

Dimensionless time tD is a stronger function of

6A. Approximation to the Carter equation for leakoff
the leakoff coefficient (CL10/3) than time t1.
Equation 6-20 was derived by assuming that the exposure Because Nordgren’s solution was ultimately
time texp in Eq. 6-14 is equal to t/2, for which integration gives obtained numerically, it is not possible to express
the volume lost per unit area of the fracture face as
it analytically. However, some useful approxima-
u L = 2C L t 2 = C L 2t . (6A-1) tions to the fracture geometry for the limiting cases
Harrington and Hannah (1975) introduced efficiency as:
of high and low efficiency can be obtained (see
Sidebar 6B). These expressions provide useful
Vf Vf
η= = , (6A-2) physical insight into the behavior of fractures. For
V i Vf + V L
example, the equation for length when leakoff is
where Vf is the fracture volume, Vi is the volume of fluid
injected, and VL is the leaked-off volume, which in terms high (i.e., low efficiency) indicates that the length
of Eq. 6-20 becomes is determined simply by a mass balance between
η=
w
(6A-3)
leakoff and flow into the fracture; i.e., the length
w + 2C L 2t increases just fast enough for the leakoff rate to
or balance the inflow. Analytical extensions to the
η=
1
. (6A-4)
PKN model that include power law fluids and
1+ 2C L 2t w explicit consideration of the efficiency between the
This approximation allows the efficiency η and S in Eq. 6-19 bounding values of 0 and 1 can be obtained.
to be related by
It is important to reemphasize that even for con-
1
η= (6A-5) tained fractures, the PKN solution is valid only
1+ 0.8S
when the fracture length is much greater than the
or
height. Typically, if the height is less than about
1− η
S = 1.25 , (6A-6) one-third of the total (tip to tip) fracture length, the
η
error resulting from the plane strain assumption is
which also shows that S tends to 0 as the efficiency tends
to 1 (negligible fluid loss) and that S tends to infinity for zero
negligible.
efficiency (i.e., negligible fracture volume relative to the fluid-
loss volume). An improved approximation for √2t is in
Chapter 9 (i.e., g0√t, where g0 is within 5% of 1.5 and varies 6B. Approximations to Nordgren’s equations
with efficiency).
Nordgren (1972) derived two limiting approximations, for
storage-dominated, or high-efficiency (tD < 0.01), cases and
for leakoff-dominated, or low-efficiency (tD > 1.0), cases, with
where q is the volume flow rate through a cross tD defined by Eq. 6-24. They are useful for quick estimates of
section, A is the cross-sectional area of the fracture fracture geometry and pressure within the limits of the
approximations. Both limiting solutions overestimate both the
(πwh f /4 for the PKN model), and qL is the volume fracture length and width (one neglects fluid loss and the
rate of leakoff per unit length: other neglects storage in the fracture), although within the
stated limits on tD, the error is less than 10%.
q L = 2 h f uL , (6-22) The storage-dominated (η → 1) approximation is
1/ 5
 E ′q i3  4 / 5
where uL is from Eq. 6-14. The cross-sectional area L(t ) = 0.39 4 
t (6B-1)
A is not Af, the area of the fracture face. Substi-  µhf 
tuting for pressure in terms of width, similar to the  µq 2 
1/ 5
w w = 2.18  i  t 1/ 5 , (6B-2)
method of Perkins and Kern, Eq. 6-21 can be writ-  E ′hf 
ten as
and the high-leakoff (η → 0) approximation is
E ′ ∂2 w 4 8CL ∂w q i t 1/ 2
= + . (6-23) L(t ) = (6B-3)
128µh f ∂x π t − texp ( x ) ∂t 2πC L hf
2
1/ 4
Nordgren solved this equation numerically in a  µq 2  1/ 8
w w = 4 3 i  t . (6B-4)
dimensionless form to obtain the width and length  π E ′C L hf 
as a function of time. The dimensionless time tD Equation 6B-3 could also be obtained from the approxima-
used in the solution is defined by tion in Sidebar 6A, with the fracture width set to zero and
2√2t replaced by π√t, which is more correct. Once the width
2/3
 64C 5 E ′h  is determined from Eq. 6B-2 or 6B-4, the pressure can be
tD =  3 L 2 f  t . (6-24) found from Eq. 6-7.
 π µqi 

• Derivation of the Khristianovich–Geertsma–

de Klerk model
Khristianovich and Zheltov (1955) derived a solu-
tion for the propagation of a hydraulic fracture by
assuming the width of the crack at any distance
from the well is independent of vertical position
(i.e., a rectangular cross section with slip at the
upper and lower boundaries), which is a reasonable
assumption for a fracture with a height much
greater than its length. Their solution includes the
fracture mechanics aspects of the fracture tip. They
recognized that to solve this problem analytically
it was necessary to simplify the solution. They did
Figure 6-3. Barenblatt’s tip condition.
this by assuming that the flow rate in the fracture
is constant and that the pressure in the fracture
pnet ( x )dx
L
could be approximated by a constant pressure in
the majority of the fracture body, except for a small ∫
0 1 − ( x L)
2
= 0. (6-27)
region near the tip with no fluid penetration, and
hence no fluid pressure. This assumption can be The width profile with a small unpressured tip
made because the pressure gradient caused by fluid region is close to that obtained for a constant net
flow is highly sensitive to fracture width and there- pressure over the entire fracture, which is equiva-
fore occurs primarily in the tip region. The concept lent to Eq. 6-7 with hf replaced by 2L:
of fluid lag remains an important element of the
4
mechanics of the fracture tip and has been vali- ww = Lpnet . (6-28)
dated at the field scale (Warpinski, 1985). They E′
showed that provided this dry region is quite small Solving Eqs. 6-26 through 6-28, they found
(a few percent of the total length), the pressure in expressions of the form given by Perkins and Kern
the main body of the fracture is nearly equal to the (1961):
pressure at the well over most of the length, with a 1/ 4
sharp decrease near the tip.  21µqi 
pnet ,w ≈  E′3  , (6-29)
Using Khristianovich and Zheltov’s result that  64 πh f L
2

the tip region is very small, Geertsma and de Klerk
(1969) gave a much simpler solution to the same with the wellbore width given by
1/ 4
problem. Their derivation is outlined in the following.  84 µqi L2 
For a rectangular cross section, the equivalent of ww =   . (6-30)
Eq. 6-9 is  π E ′h f 
∂p 12 qµ For no leakoff, the equations can be solved for

=− , (6-25) length and width, respectively:
∂x hf w3 1/ 6
 E ′q 3 
which can be written in integral form as L(t ) = 0.38 3i  t 2 / 3 (6-31)
 µh f 
6µqi
L
dx
pnet =
hf ∫w
0
3
. (6-26)
 µq 3 
1/ 6
ww = 1.48 i 3  t 1/ 3 . (6-32)
It can be shown that applying Barenblatt’s tip  E ′h f 
condition (which requires that the fracture tip must
close smoothly, as illustrated in Fig. 6-3) implies The high-leakoff solution for the PKN model
that the stress intensity factor (see Chapter 3) is zero: (Eq. 6B-3) also applies to the KGD model, but
Geertsma and de Klerk did not provide an explicit
width relationship for the KGD model in the case
of high leakoff.

Geertsma and de Klerk also extended the model are still used to design treatments and are usually
to include fluid leakoff, following Carter’s (1957) available as options in simulators.
method. Fluid loss is incorporated by assuming that Similar solutions can be derived for radial frac-
it has no effect on fracture shape or pressure distri- tures (see Sidebar 6C).
bution. The volume of a two-wing KGD fracture is
π 6C. Radial fracture geometry models
Vf = h f Lww . (6-33)
2
Both Perkins and Kern (1961) and Geertsma and de Klerk
Performing a volume balance and solution proce- (1969) considered radial fractures, which grow unconfined
from a point source. This model is applicable when there are
dure similar to that of Carter, they obtained no barriers constraining height growth or when a horizontal
fracture is created.
qi ww  s 2 
L=  e erfc( S ) + S − 1 ,
2
(6-34) Geertsma and de Klerk formulated the radial model using
64CL2 h f  π  the same arguments outlined in “Derivation of the
Khristianovich–Geertsma–de Klerk model” (page 6-7). The
fracture width is
where
 µq R 
1/ 4
8C πt w w = 2.56  i  (6C-1)
S= L . (6-35)  E′ 
πww
and the radial length R is
To include the effects of spurt loss Sp, ww should
( )
q i 4w w + 15S p  s 2 2 
be replaced by ww + (8/π)Sp, which is equivalent to R=
30π 2C L2  e erfc(S ) + S − 1 , (6C-2)
– + 2S and  π 
the Carter relation with w replaced by w p
–
w = πw/4.
where
• Assumptions of the PKN and KGD models 15C L πt
S= . (6C-3)
Both the PKN and KGD models contain a number 4w w + 15S p
of assumptions that are revisited in this section.
An explicit relationship for pressure can be derived by con-
They assume that the fracture is planar (i.e., that it sidering the solution for flow from a point source, in which
propagates in a particular direction, perpendicular case the pressure in the fracture is a function of the expres-
sion ln(rw /R), where rw is the radius of the wellbore.
to the minimum stress, as described in Chapter 3). The no-fluid-loss approximations for the radial model are
They also assume that fluid flow is one-dimen- 1/ 9
 µ 2 q i3  1/ 9
sional (1D) along the length of the fracture. In w w = 2.17  2 
t (6C-4)
 E′ 
the case of the models described, they assume
Newtonian fluids (although Perkins and Kern also 1/ 9
 E ′q i3  4 / 9
R = 0.52  t . (6C-5)
provided solutions for power law fluids), and  µ 
leakoff behavior is governed by a simple expres-
The large-fluid-loss approximation for radial length is
sion derived from filtration theory (Eq. 6-14). The
1/ 4
rock in which the fracture propagates is assumed to 1  q i2t 
R=   . (6C-6)
be a continuous, homogeneous, isotropic linear π  C L2 
elastic solid; the fracture is considered to be of
An expression for width in the case of large fluid loss was
fixed height or completely confined in a given not provided but can be found from Eqs. 6C-1 and 6C-6.
layer; and one of two assumptions is made con-
cerning the length to height ratio of the fracture—
i.e., height is large (KGD) or small (PKN) relative
to length. Finally, the KGD model includes the 6-3. Three-dimensional and pseudo-
assumption that tip processes dominate fracture three-dimensional models
propagation, whereas the PKN model neglects
fracture mechanics altogether. The simple models discussed in the previous sections
Since these models were developed, numerous are limited because they require the engineer to spec-
extensions have been made that relax these ify the fracture height or to assume that a radial frac-
assumptions, the most important of which are the ture will develop. This is a significant limitation,
solutions for power law fluids. These two models because it is not always obvious from logs and other

data where or whether the fracture will be contained.

Also, the fracture height usually varies from the well
(where the pressure is highest) to the tip of the frac-
ture. This limitation can be remedied by the use of
planar 3D and pseudo-3D (P3D) models.
The three major types of hydraulic fracture models
that include height growth are categorized according
to their major assumptions.
• General 3D models make no assumptions about the
orientation of the fracture. Factors such as the well-
bore orientation or perforation pattern may cause
the fracture to initiate in a particular direction
before turning into a final preferred orientation (per-
pendicular to the far-field minimum in-situ stress). Figure 6-4. Conceptual representation of the lumped model.
Simulators incorporating such models are computa-
tionally intensive and generally require a specialist
inherent assumptions that fluid flow is along
to obtain and interpret the results. They are most
streamlines from the perforations to the edge of the
applicable in research environments, for which they
ellipse and that the streamlines have a particular
are used for studying details of fracture initiation
shape, derived from simple analytical solutions.
and near-well complexities such as those discussed
Cell-based models treat the fracture as a series of
in Section 6-8, rather than overall fracture growth.
connected cells. They do not prescribe a fracture
One example of such a study was published by
shape, but generally assume plane strain (i.e., each
Brady et al. (1993). These models are not discussed
cell acts independently) and do not fully couple the
further in this volume.
calculation of fluid flow in the vertical direction to
• Planar 3D models are based on the assumption that the fracture geometry calculation.
the fracture is planar and oriented perpendicular to
In the fixed-height models described previously, no
the far-field minimum in-situ stress. No attempt is
consideration is given to the layers surrounding the
made to account for complexities that result in devi-
fractured zone. The planar and P3D models use data
ations from this planar behavior. Simulators based
about the properties of the surrounding zones to pre-
on such models are also computationally demand-
dict the rate of growth into these zones. For the pur-
ing, so they are generally not used for routine
pose of this chapter, planar 3D models are defined as
designs. They should be used where a significant
those in which calculation of the full 2D fluid-flow
portion of the fracture volume is outside the zone
field in the fracture is coupled to the 3D elastic
where the fracture initiates or where there is more
response of the rock, and P3D models are defined as
vertical than horizontal fluid flow. Such cases typi-
those that approximate either the coupling or the 3D
cally arise when the stress in the layers around the
elasticity in some manner.
pay zone is similar to or lower than that within the
Regardless of which type of model is used to
pay. This type of model is described in more detail
calculate the fracture geometry, only limited data are
in Section 6-3.1.
available on typical treatments to validate the model
• P3D models attempt to capture the significant used. For commercial treatments, the pressure history
behavior of planar models without the computa- during treatment is usually the only data available to
tional complexity. The two main types are referred validate the model. Even in these cases, the quality of
to here as “lumped” and cell-based. In the lumped the data is questionable if the bottomhole pressure
(or elliptical) models, the vertical profile of the frac- must be inferred from the surface pressure. The bot-
ture is assumed to consist of two half-ellipses joined tomhole pressure is also not sufficient to uniquely
at the center, as shown in Fig. 6-4. The horizontal determine the fracture geometry in the absence of
length and wellbore vertical tip extensions are cal- other information, such as that derived from tiltmeters
culated at each time step, and the assumed shape is and microseismic data (see Sidebar 6D). If a simulator
matched to these positions. These models make the incorporates the correct model, it should match both

With multiple seismic receivers, triangulation techniques

6D. Field determination of fracture geometry can be employed and greater accuracy obtained. With either
approach, however, the objective is to locate the zone of
Fracture geometry can be determined by using the two tech- microseisms surrounding the hydraulic fracture and deduce
niques of microseismic activity and tiltmeters. Microseisms the size and shape of the fracture from this information.
can be used to locate the fracture, thus providing estimates of
its length and height, whereas tiltmeters can provide informa- Downhole tiltmeters
tion about fracture width.
Width development in a hydraulic fracture results in elastic
deformation of the formation. This deformation can be used
Microseisms for fracture diagnostics to provide significant information about
Although all models of hydraulic fracturing assume that the fracture height and width and also about formation character-
rock is a continuous medium, it is well known that reservoirs istics.
have natural fractures, bedding planes and other weakness As a fracture is opened, the deformation of the rock
features that respond as a noncontinuum. Such features have extends for large distances into the reservoir. Although the
been used to image hydraulic fractures using seismic tech- deformation is small at distances of more than a few tens of
niques. feet, highly sensitive tiltmeter devices can measure these
Hydraulic fractures induce two large changes in the reser- small changes in position. A tiltmeter does not actually mea-
voir as they are created. The stress in the surrounding rocks sure the displacement of the earth, but rather the curvature of
is perturbed because of fracture opening, and the pore pres- the displacement, and it is capable of measuring up to nano-
sure is increased as a result of leakoff of the high-pressure radian resolution (a nanoradian is the angle induced by
fracturing fluid. Both of these features can result in the gener- stretching a line from New York to Los Angeles and raising
ation of large shear stresses on many of the weakness planes the New York side by the diameter of a pencil). Tiltmeters
near the hydraulic fracture, resulting in small shear slippages have long been used for surface diagnostics of earth move-
called microseisms or microearthquakes. ment, but the application of a string of downhole tiltmeters
Microseisms generate seismic waves that can be detected provides highly sensitive fracture data.
by sensitive seismic receivers in nearby wells. As shown in Figure 6D-2 shows a schematic of the tilt response of the
Fig. 6D-1, both compressional waves (P-waves) and shear formation measured in a well offset to the fracture treatment.
waves (S-waves) can be generated by the microseism, and The characteristic S-shaped curve is typical of tilt, as opposed
these two waves travel with different velocities. If a receiver to strain, and can be simply explained. Straight across from
can detect both the P- and S-waves, the time separation can the fracture, the rock is pushed away, but is not tilted on the
be determined and the distance to the source inferred from geometric axis of the fracture, and there is zero tilt. Above the
fracture, the earth experiences curvature that is defined as
[t S − t P ] ,
uPuS negative for this example. The curvature reaches a maximum
d= (6D-1)
uP − uS at a well-defined point and then decreases to zero as the dis-
tance from the fracture increases. The bottom is identical to
where uP and uS are the compressional and shear velocities, the top, except that the curvature has the opposite direction
respectively, and tS and tP are the shear and compressional and opposite sign.
arrival times. Two aspects of this distribution are important for diagnos-
The direction in space can be determined by using a tri- tics. First, the locations of the maximum tilt values are a func-
axial receiver to examine the amplitude of the P-wave. The tion of the height hf of the fracture relative to the distance d
P-wave has the characteristic that its particle motion (how the away. Thus, fracture height can be quickly estimated. Second,
rock mass vibrates) is aligned with the direction of travel of the amplitude of the tilt is a function of the width of the frac-
the wave. By obtaining the orientation of the resultant ampli- ture, so the width during fracturing, and possibly the final
tude vector at any time, the microseism can be traced back to propped width, can be estimated as well.
its source. Branagan et al. (1996) provided an example of the applica-
tion of tiltmeters to the calculation of hydraulic fracture geometry.
Example traces Shear slippage

x
Hydraulic Tiltmeters
fracture
S(t1)
P(t1)
P S hf
y y x
S(t2) d
P(t2)
Receiver
Figure 6D-1. Microseismic traces at the receiver resulting Figure 6D-2. Tiltmeter response to hydraulic fracture
from shear slippage. width.

treating pressure and fracture geometry. These issues section it is assumed that this process is described by
are addressed in Section 6-12 and Chapter 9. linear elastic fracture mechanics (LEFM). If the LEFM
failure criterion is exceeded at any point on the fracture
periphery, the fracture will extend until the criterion
6-3.1. Planar three-dimensional models is again met. For simple shapes and pressure distribu-
A planar fracture is a narrow channel of variable tions, such as ellipses under constant pressure, the cri-
width through which fluid flows. The fracture geome- terion can be specified analytically, similar to Eq. 6-3.
try is defined by its width and the shape of its periph- For more complex shapes and pressure distributions,
ery (i.e., height at any distance from the well and analytical solutions are not available. In these cases,
length). Both the width at any point and the overall it can be shown that a relatively simple criterion can be
shape vary with time. They depend on the pressure written in terms of the width near the tip and the criti-
distribution, which itself is determined by the pressure cal stress intensity factor or fracture toughness KIc,
gradients caused by the fluid flow within the fracture. which is introduced in Chapter 3:
Because the relation between pressure gradient and 4 2 K Ic
flow rate is highly sensitive to the fracture width w( x ) = x, (6-37)
π E′
(Eq. 6-9), the geometry and fluid flow are tightly cou-
pled. Although the mechanics of these processes are where x is the distance measured from the tip. Rela-
described separately in this section, the complexity of tions between fracture mechanics parameters such as
solving any fracture model lies in the close coupling the specific surface energy (used in Eq. 6-3) and the
between the different processes. Three separate prob- fracture toughness are provided in Chapter 3.
lems are considered: The fluid flow is described by equations for conser-
vation of mass (a general form of Eq. 6-21, including
• width profile in a fracture of known shape and pres- the density ρ and expressed in terms of velocity u):
sure distribution
• shape of the fracture (
 ∂(ρwux ) ∂ ρwuy  ∂ )
 +  + (ρw ) + 2ρuL = 0, (6-38)
• flow of fluid in a fracture of known shape and width  ∂x ∂y  ∂t
(i.e., known geometry).
Hirth and Lothe (1968) and Bui (1977) showed which can be written as a vector equation:
how the pressure and width in a fracture may be ∂
∇ ⋅ (ρwu ) + (ρw) + 2ρuL = 0, (6-39)
related. Basically, the width at any point (x,y) is ∂t
determined by an integral of the net pressure over
and the conservation of momentum (a general form of
the entire fracture, expressed as
Eq. 6-9) is
w( x, y) = ∫∫ f ( x − x ′, y − y ′ )( p( x ′, y ′ ) − σ( x ′, y ′ ))dx ′dy ′, Du
S
ρ = −∇p − [∇ ⋅ τ] + ρg , (6-40)
Dt
(6-36) where τ is the shear stress and g– is the acceleration of
where σ is the stress. gravity.
The details of the elastic influence function f in The first two terms in Eq. 6-38 relate to the spatial
Eq. 6-36 are beyond the scope of this volume. Use- change of the mass-flow vector, and the second two
able forms of Eq. 6-36 can be derived generally only terms represent the storage resulting from width
for homogeneous linear elastic materials (see Side- increases and leakoff, respectively. Equation 6-40 is
bar 6E). In fracturing applications, the rock is usually a vector equation. The term on the left-hand side is the
also assumed to be isotropic. rate of change of momentum, and the terms on the
The shape of the fracture evolves with time. In right-hand side are the pressure, viscous and gravita-
essence, the boundary (i.e., the vertical and horizontal tional forces, respectively. It simply states that a small
tips) moves outward as the fluid provides sufficient element of fluid accelerates because of the forces act-
energy to fracture the rock at the boundary. More coming on it. This equation can be expanded and then sim-
plex tip behavior is discussed subsequently, but in this plified for the geometries of interest in hydraulic frac-

6E. Lateral coupling in pseudo-three- yt,i = yt,k

dimensional models
Assume that a fracture has a fixed height and that it consists (xc,i,yc,i) (xc,k,yc,k)
of a number of elements each of constant width over the height
(i.e., a KGD fracture). Let the grid points be represented by
points xi in the center of the elements with corners (xl,i,yb,i),
(xl,i,yt,i), (xr,i,yt,i) and (xr,i,yb,i), as shown in Fig. 6E-1. Crouch and yb,i = yb,k
Starfield (1983) developed a boundary element solution tech-
nique called the displacement discontinuity method. They
xl,i xr,i xl,k xr,k
showed that the pressure at any point is given by
p (x i ) = ∑A w ,
k
ik k
(6E-1) Figure 6E-1. Geometry for displacement continuity
solution.
where Aik is an influence function of the form
G
Aik = I (x c ,i , y c ,i ; x l ,k ; x r ,k y b ,k y t ,k ) , (6E-2) Equation 6E-1 can then be written as
4 π(1− ν)
where the influence function I is defined as p(xi) = p(wi) + pcorr , (6E-6)
[(x c ,i − x r ,k ) 2 + (y c ,i − y t ,k ) 2 ]1/ 2 where

I =
(x c ,i − x r ,k )(y c ,i − y t ,k )
[(x c ,i + x l ,k ) 2 + (y c ,i − y t ,k ) 2 ]1/ 2 p (w i ) = w (x i ) ∑A ik
(6E-7)
−
(x c ,i + x l ,k )(y c ,i − y t ,k )
+
[(x c ,i + x l ,k ) 2 + (y c ,i + y b ,k ) 2 ]1/ 2 p corr = ∑ A ∆w ik ki . (6E-8)
(x c ,i + x l ,k )(y c ,i + y b ,k )
−
[(x c ,i − x r ,k ) 2 + (y c ,i + y b ,k ) 2 ]1/ 2
. (6E-3)
Σ
The term w(xi) Aik thus represents the pressure induced
by a fracture of constant width w(xi). For a fracture of infinite
(x c ,i − x r ,k )(y c ,i + y b ,k ) length, this pressure would be exact if calculated using the
plane strain solution. The term p(wi) can therefore be
To accurately solve Eq. 6E-1 requires a large number of obtained as the sum of the plane strain solution and the effect
elements. Also, it is difficult to extend directly to other shapes of two semi-infinite fractures of w – wi attached at the tip of
such as ellipses or for nonconstant heights. To overcome each fracture wing.
these problems, the equation is modified as follows. The From Eq. 6E-2, the influence functions decrease with
width at any point can be written as distance from an element. The advantages of the form of
Eq. 6E-8 are that the corrections are smallest near the ele-
w (x k ) = w (x i ) + ∆w ki , (6E-4) ment where the widths are almost the same and that the
self-correction is exactly zero by definition. The number of
where ∆wki is defined as elements required to obtain an accurate solution is significantly
reduced, and variable heights and other shapes are easily
∆w ki = w (x k ) − w (x i ) . (6E-5) introduced. Lateral coupling is relatively easy to introduce to
the explicit solution method because the pressure correction
is simply added before the fluid velocities are calculated.
turing (see Sidebar 6F). For a particular component, tions, the equations for the stress in a Newtonian fluid
such as the x component, Eq. 6-40 can be written as reduce to
Dux ∂p  ∂τ ∂τ ∂τ   ∂u 
ρ = − −  xx + yx + zx  + ρgx . (6-41) τ xz = τ zx = −µ x 
Dt ∂x  ∂x ∂y ∂z   ∂z 
A constitutive law relating the stresses τ to the flow  ∂u 
τ yz = τ zy = −µ y  , (6-42)
rate is required to complete the description of fluid  ∂z 
flow. In the case of steady flow in a narrow channel and Eq. 6-41 can be written as
such as a fracture, the full details of the constitutive
law are not required, because the narrow fracture Dux ∂p  ∂2u 
ρ = − + µ 2x  + ρgx . (6-43)
width results in the complete dominance of some Dt ∂x  ∂z 
stress terms. The only terms of interest are the shear
For the special case of a narrow channel (Poiseuille
stresses induced by velocity gradients across the frac-
flow), where velocity gradients parallel to the flow are
ture. In addition, use is made of the lubrication
small and there is no flow perpendicular to the chan-
approximation, so flow perpendicular to the fracture
nel, the time-dependent term simplifies to a partial
wall (the z direction) is neglected. With these assump-

6F. Momentum conservation equation for The first three components of Eq. 6F-5 are the normal
hydraulic fracturing stresses, and the last three are the shear stresses. The last
term of the normal components is zero for incompressible
fluids. In the case of 1D flow between parallel plates, without
Equation 6-40 is a vector equation, for which a component leakoff, two of the velocity components are identically zero.
can be written as In addition, conservation of mass implies that the third com-
ponent cannot vary with position. Hence, all three normal
Du i ∂p  ∂τ xi ∂τ yi ∂τ zi 
ρ =− − + +  + ρg i , (6F-1) components are identically zero. The equations thus reduce
Dt ∂x i  ∂x ∂y ∂z  to those for shear flow. Although these assumptions are not
strictly true in general, they are used for the flow calculations
where u is the velocity, g is the gravitational acceleration, and in hydraulic fracture modeling. It can also be shown that for
i is x, y or z. The term on the left side of Eq. 6F-1 is termed a narrow channel, the velocity gradients perpendicular to the
the substantial derivative, which is the rate of change seen by walls (the z direction) are much greater than those in the par-
an observer moving with the fluid motion. It can be related to allel directions. Finally, therefore, the stress components for
the usual partial derivative (i.e., the rate of change seen by a a Newtonian fluid in a hydraulic fracture can be written as
stationary observer) as
 ∂u 
D ∂ ∂ ∂ ∂ τ zx = −µ x 
= + ux + uy + uz . (6F-2)  ∂z 
Dt ∂t ∂x ∂y ∂z  ∂u y 
τ zy = −µ . (6F-6)
Thus, Eq. 6F-2 can be expanded to  ∂z 
 ∂u ∂u i ∂u i ∂u i  Substituting Eq. 6F-6 into Eq. 6F-4 obtains
ρ i + u x + uy + uz 
 ∂t ∂x ∂y ∂z 
 ∂u x ∂u x ∂p ∂ 2u x
∂p  ∂τ xi ∂τ yi ∂τ zi  ρ u x + uy =− +µ
=− − + +  + ρg i . (6F-3)  ∂x ∂y  ∂x ∂z 2
∂x i  ∂x ∂y ∂z 
 ∂u y ∂u y ∂p ∂ 2u y
ρ u x + uy =− +µ + ρg . (6F-7)
This completely general equation can be simplified for a  ∂x ∂y  ∂y ∂z 2
narrow channel in an impermeable medium. Leakoff does not
occur in this case, so components in the z direction can be For 1D flow along the fracture length, as typically assumed in
neglected. In addition, the flow is assumed to be steady state, P3D models, Eq. 6F-7 can be simplified to
so time derivatives can be ignored. In this case, Eq. 6F-3 sim- ∂ 2u x 1 ∂p
plifies to = . (6F-8)
∂z 2 µ ∂x
 ∂u i ∂u i 
ρ u x + uy 
 ∂x ∂y  Assuming zero slip (i.e., zero velocity at the fracture wall), the
solution to Eq. 6F-8 is
∂p  ∂τ xi ∂τ yi ∂τ zi 
=− − + +  + ρg i ,
( )
(6F-4) 1 ∂p 2
∂x i  ∂x ∂y ∂z  ux = z − (w 2) .
2
(6F-9)
2µ ∂x
for i = 1 or 2. Even for a permeable medium, Eq. 6F-4 is
used. In this case, leakoff is treated as a sink term and Integrating to obtain the average velocity across the channel,
included in the mass balance, but it is assumed not to affect
−w 2 ∂p
the equations relating pressure, stress and fluid velocity. ux = . (6F-10)
12µ ∂x
Newtonian fluids The flow rate per unit height is obtained by multiplying the
To make Eq. 6F-4 useful, the stress components must be average velocity by the width w.
determined, which is done by assuming a model of fluid In the case of 2D flow, the left-hand sides of Eq. 6F-7 are
behavior. For example, a Newtonian fluid is a model with one zero if inertia may be neglected. In this case for the y direc-
parameter, the viscosity µ. The stress components are tion, an equation can be formed similar to Eq. 6F-10, except
that it includes a gravitational term.
∂u x 2
τ xx = µ + µ (∇ ⋅ u )
∂x 3
∂u y 2
τ yy = 2µ + µ (∇ ⋅ u )
∂y 3
∂u z 2
τ zz = 2µ + µ (∇ ⋅ u )
∂z 3
 ∂u x ∂u y 
τ xy = τ yx = −µ + 
 ∂y ∂x 
 ∂u ∂u 
τ yz = τ zy = −µ y + z 
 ∂z ∂y 
 ∂u ∂u 
τ zx = τ xz = −µ z + x  . (6F-5)
 ∂x ∂z 

derivative of velocity with respect to time. It is usually a small fracture, initiated at the perforations, divide it
assumed that the flow is steady state, which finally into a number of equal elements (typically 16 squares)
obtains and then begin solution of the equations. As the
∂p  ∂2u  boundary extends, the elements distort to fit the new
= µ 2x  + ρgx (6-44) shape. One difficulty with such a solution is that the
∂x  ∂z 
elements can develop large aspect ratios and very
and a similar equation for the y component. small angles, as shown in Fig. 6-5. The numerical
Equations 6-36 through 6-44 summarize the planar schemes typically used to solve the equations do not
3D model for Newtonian fluids. Similar results can be usually perform well with such shapes.
obtained for non-Newtonian fluids (see Sidebar 6G). A different formulation was described by Barree
These equations are generally not amenable to ana- (1983), and numerous field applications have been
lytic solutions, but require a numerical simulation. reported (e.g., Barree, 1991). It neatly avoids the prob-
In addition, although it is relatively straightforward lem of grid distortion by dividing the layered reservoir
to write the conceptual equations, efficient and robust into a grid of equal-size rectangular elements, which
numerical solutions are difficult to obtain. The pri- are defined over the entire region that the fracture may
mary reasons for this difficulty are the extremely cover. In this case, the grid does not move. Instead, as
close coupling of the different parts of the solution the failure criterion is exceeded, the elements ahead of
(e.g., fluid flow and solid deformation), the nonlinear the failed tip are opened to flow and become part of
relation between width and pressure, and the com- the fracture, as shown in Fig. 6-6. Two limitations of
plexity of a moving-boundary problem. this approach are that
The first numerical implementation of a planar • the number of elements in the simulation increases
model was reported by Clifton and Abou-Sayed as the simulation proceeds, so that the initial num-
(1979). In essence, their approach was to define ber may be small, resulting in inaccuracy
and for a Bingham plastic
6G. Momentum balance and constitutive equation
τ0
for non-Newtonian fluids µa = µ0 + . (6G-7)
I2
The definition of a Newtonian fluid is the one-parameter rela- 2
tion between stress and velocity (Eq. 6G-5). In tensor notation,
this can be written as The commonly used consistency index K´ is dependent on
the flow geometry and is related to a basic fluid property, the
τ ij = −µ∆ ij , (6G-1) generalized consistency index K (Eq. 6G-6). For parallel
plates (i.e., in a slot), which can represent a fracture, the rela-
where ∆ is the rate of deformation tensor, with components tionship is
 2n + 1
n
∂u i ∂u j K′ =K  . (6G-8)
∆ ij = + . (6G-2)  3n 
∂x j ∂x i
The viscosity may be a function of pressure and tempera- For a pipe it is
 3n + 1
n
ture or other variables, including the history of the fluid, but not
of ∆. For non-Newtonian fluids, an equation similar to Eq. 6G-1 K′ =K  . (6G-9)
 4n 
may be written:
The maximum difference between the two expressions is less
τ ij = −µ a ∆ ij , (6G-3) than 4% for all values of n. For 1D flow of a power law fluid
between parallel plates, the average fluid velocity is given by
where µa is a function of ∆. For flows of the type of interest
1+ n
in fracturing, it can be shown that µa may depend only on ∆  1   ∂p 
1/ n 1/ n
n w  n
through a relation of the form ux =       . (6G-10)

 K   ∂x  1+ 2n  2 
µ a = µ a (I 2 ), (6G-4)
For the special case of the power law exponent n = 1, this
where I2 is the second tensor invariant: reverts to the equation for a Newtonian fluid, with K´ replaced
by the viscosity. Table 6G-1 summarizes useful information for
I2 = ∑∑∆ ∆ ij ji . (6G-5) the laminar flow of both Newtonian and power law fluids
i j under different geometries. However, the expressions for
For example, for a power law fluid, the function µa is pressure drop are not generally applicable for drag-reducing
fluids such as those used in hydraulic fracturing.
n −1
I2
µa = K (6G-6)
2

6G. Momentum balance and constitutive equation for non-Newtonian fluids (continued)
Table 6G-1. Summarized expressions for laminar flow of Newtonian and power law fluids.
Fluid Type Pipe Parallel Plates Ellipse (zero

eccentricity)
Reynold’s number (NRe) Newtonian ρuD 2ρuw πρuw

µ µ 2µ
Power law 81−n ρD nu 2 −n 31−n 22 −n ρu 2 −nw n π 2 −n ρu 2 −nw n

K p′ K pp′ 2n K ell′
Hydraulic diameter (DH) D 2w πw/2

Friction factor 16/NRe 24/NRe 2π2/NRe
  DH  1 
 f =  ∆p  /  ρu 2  
 2 
 4L
Velocity distribution Newtonian   2r  2  3u   2y  2    2 y  2  2z  2 

u r = 2u 1−    ux = 1−    u x = 2u 1−   −  
  D   2   w     w   hf  
Power law  n +1
n
  n +1
n

 3n + 1   2r    2n + 1   2y  
ur =   u 1−    ux =   u 1−   
 n +1    D    n +1    w  
   
Pressure drop (∆p/L or dp/dx) Newtonian 128µq 12µq 64µq

πD 4 hf w 3 πhf w 3
Power law 2 5n + 2 q nK  4n + 2  2q K
n n
See Eq. 6-57
 
π n D 3 n +1  n  hfnw 2n +1
K´  2n + 1 –
n
 3n + 1
n
K p′ = K   ′ =K
K pp 
 4n   3n 
y Elements that could

open to advance
the fracture
x
Open elements within the fracture
Figure 6-5. Planar 3D fracture divided into elements that

were initially square.
Figure 6-6. Fixed-grid solution showing elements open to

• the general size of the fracture must be estimated in advance the fracture.
advance of the simulation to ensure that a “reason-
able” number of elements is used.
strength criterion is used for fracture extension, rather
In addition, this particular implementation has two
than a fracture mechanics effect. The failure criterion
simplifying assumptions, that a simplified method is
is used to compare the stress at the center of all
used for representing modulus contrasts and a tensile
boundary elements with the material tensile strength.

If the strength is exceeded, then the element is

assumed to open. However, the fracture-induced stress
in the material near the tip of a fracture varies with the
square root of the distance from the tip. Hence, the
failure criterion is grid-resolution dependent.
6-3.2. Cell-based pseudo-three-

dimensional models
(6-46)
In cell-based models, the fracture length is divided
into a number of discrete cells. This is directly analo-
gous to the planar models, except that only one direc-
tion is discrete instead of two. Fluid flow is assumed
to be essentially horizontal along the length of the
fracture, and the solid mechanics is typically simpli-
fied by assuming plane strain at any cross section. As
in the PKN model, these assumptions make these
models suitable primarily for reasonably contained
fractures, which are long relative to their height.
These two assumptions allow separating the solid (6-47)
and fracture mechanics solution from the fluid flow
where ρf is the fluid density, hcp is the height at the
as follows. Plane strain implies that each cross section
center of the perforations, and hi is the height from
acts independently of any other. In addition, the
the bottom tip of the fracture to the top of the ith
assumption of 1D fluid flow implies that the pressure
layer, as shown in Fig. 6-7.
in the cross section is always
This set of nonlinear equations can be solved by
p = pcp + ρgy , (6-45) iteration. Assuming that the solution (two vertical
tip positions plus the pressure) at one value of pcp is
where pcp is the pressure along a horizontal line known, a height increment is assumed. The incre-
through the center of the perforations and y is the ver- mental height growth in the two vertical directions
tical distance from the center of the perforations. is then calculated such that Eqs. 6-46 and 6-47 are
Equation 6-45 is valid only if vertical fracture exten- both satisfied, and pcp to obtain these positions is
sion is sufficiently slow that the pressure gradient calculated. Finally, the width profile associated
resulting from vertical flow can be neglected. This with this solution can be obtained as
assumption that the vertical tips of the fracture are
approximately stationary at all times is called the
equilibrium-height assumption.
w( y ) =
4
E′
( ( )
pcp + ρ f g hcp − y − σ n ) y( h − y )
f
4 n −1
∑ (σ i +1 − σ i )
• Solid mechanics solution
+
With the equilibrium-height assumption, the solid πE ′ i = 1
mechanics solution simplifies to the determination   y h f − 2 hi hi 
of the fracture cross-sectional shape as a function (hi − y)cosh  +
–1
of the net pressure, or pcp. Simonson et al. (1978)   | y − hi | h f | y − hi | 

derived this solution for a symmetric three-layer  h − 2 hi  
case. Fung et al. (1987) derived a more general + y(h f − y) cos −1  f  , (6-48)
solution for nonsymmetric multilayer cases.  h f  
Following Fung et al. the stress intensity factors
at the top and bottom tips KIu and KIl, respectively, where y is the elevation measured from the bot-
can be written in terms of the pressure at the center tom tip of the fracture.
of the perforations pcp and the closure stresses in
the layers σi as

Consider, for example, the symmetric three-layer

case shown in Fig. 6-8. If the gravitational compo-
nent is neglected, so that the problem is symmetric,
then the penetrations into the two barriers are
n equal. In this case, Eq. 6-46 can be simplified sig-
nificantly and written as (Simonson et al., 1978)
n–1
πh f  2 ∆σ  h 
h K Iu = K Il =  pcp − σ pay − cos −1  pay   ,
2  π  h f  
y (6-49)
i where ∆σ is the difference in stress between the
central layer (pay zone) and the surrounding layers,
and hpay and σpay are the thickness and stress of the
hi
pay zone, respectively. Figure 6-9 shows fracture
height as a function of net pressure, as calculated
2
by Eq. 6-49.
Layer 1 Although Eq. 6-49 is for a special case, it shows
two interesting practical results. First, penetration
into the barrier layers occurs at a critical net pressure:
2 K Ic2
pnet ,crit = . (6-50)
πh f
Figure 6-7. Definition of variables for the fracture contain-
ment problem. For example, if KIc is 2000 psi/in.1/2 and hf is 20 ft
[240 in.], the critical net pressure for breakthrough
1.5
Symmetric case σ2
hs
σ1 h p
1.0
σ2
hpay hf hs
h
∆σ
0.5
KIc = 0, all σ2 – σ1
KIc = 1000 psi/in.1/2

σ2 – σ1 = 1000 psi
σ2 – σ1 = 500 psi
σ2 – σ1 = 250 psi
0
0 0.2 0.4 0.6 0.8 1.0
σ2 – p
σ2 – σ1
σpay
Figure 6-9. Fracture height versus net pressure for sym-

Figure 6-8. Simple three-layer height growth problem. metric barriers. hs = penetration into the boundary layer.

is only about 100 psi. Second, the net pressure can- Solving Eq. 6-52 with Eq. 6-53 with the no-slip
not reach the stress contrast because this would boundary condition (i.e., zero velocity at the frac-
result in infinite fracture height. ture wall), the average velocity across the channel is
In a typical cell-based simulator, a table of these 1/ n 1+ n
solutions is calculated prior to simulating the frac-  ∂p   ∂p  n  w n

ux = −sgn    K , (6-54)
ture evolution rather than at each time step of the  ∂x   ∂x  1 + 2n  2 
calculation, and the relations among width, pres-
where sgn represents the sign of the quantity.
sure and height are used to greatly speed up the
In the special case of a Newtonian fluid, n = 1
solution of the fluid flow equations (conservation
and µ = K, and Eq. 6-54 becomes
of mass and momentum).
w 2 ∂p
• Fluid mechanics solution ux = − . (6-55)
12µ ∂x
One of the major differences between planar 3D
and P3D models is the fluid flow calculation. The To obtain the total flow rate across the height of
fluid flow model in most P3D models is the same the cross section, and hence an average velocity for
as that introduced by Nordgren (1972) (i.e., a 1D substitution in Eq. 6-51, Eq. 6-54 is integrated from
version of the model described for the planar 3D the bottom to the top tip of the cross section:
q = ∫ w( y)ux ( y)dy .
model). In this model, both vertical flow and the
(6-56)
variation of horizontal velocity as a function of hf
vertical position are neglected. This results in the
inability of typical P3D models to represent several The average velocity is thus determined as
aspects of behavior, namely (Smith and Klein, 1995) 1+ n 1/ n
q  ∂p   ∂p  A  
– effect of variations in width in the vertical direc- u = = −sgn   Φ  2K    ,
A  ∂x   ∂x   hf  
tion on fluid velocity  
– local dehydration, which is approximated as (6-57)
simultaneous dehydration over the entire height
of the fracture where the channel function Φ is
1+ 2 n
– fluid loss after tip screenouts (TSOs), when fluid h  n 1 1+ 2 n
∫ w( y )
n
Φ= f n dy . (6-58)
flow through the proppant pack is ignored  A 2 + 4n h f hf
– proppant settling resulting from convection or

Relations for the PKN model with power law
gravity currents.
fluids can be derived following this approach (see
The average velocity and width are used (width Nolte, 1979, 1991).
is replaced by cross-sectional area divided by – Laminar and turbulent flow
height) to simplify the conservation of mass
When fluid flows between parallel plates at a low
(Eq. 6-38 for an incompressible fluid) to
rate without leakoff, any fluid element remains
∂Au ∂A a fixed distance from the wall of the channel,
+ = −2 ∑ (uL hL )i , (6-51) except in a small entrance region. This is known
∂x ∂t i
as laminar flow. By contrast, in turbulent flow,
where u is the average cross-sectional velocity and eddies occur, and fluid is continually mixed. This
uL and hL are the leakoff rate (Eq. 6-14) and height mixing results in added friction and different flow
in each layer. Similarly, the conservation of behavior. The Reynold’s number NRe (defined in
momentum simplifies to Table 6G-1) enables determining whether laminar
∂p ∂τ or turbulent flow will occur. If NRe exceeds 2100,
= − xz . (6-52) flow will be turbulent. Inside the fracture, NRe is
∂x ∂z
typically well below this value, except for partic-
For a power law fluid with properties n and K, ularly thin fluids, such as acid.
 ∂u 
n
τ xz = K  x  . (6-53)
 ∂z 

– Rheology of fracturing fluids

Fracturing fluids are generally treated as power
law fluids, and because they are shear thinning A4 A3 A2 A1
(i.e., viscosity decreases with increasing shear
rate), n is usually less than 1. The effective para- x5 x4 x3 x2 x1
meters of the power law model K´ and n´ are u5 u4 u3 u2 u1
typically derived from laboratory measurements
(see Chapter 8) over a range of shear rates. For Figure 6-10. Fracture divided into elements with positions
shear-thinning fluids, the apparent viscosity and velocities defined at grid points.
(derived from K´ and n´) decreases as shear rate
increases, and the viscosity would be infinite at
of new elements at the well, as necessary.) Mass-
zero shear rate. In reality, limiting low- and high-
conservation Eq. 6-51 can be rewritten as
shear viscosities occur and must be considered.
∂A ∂Au
Fracturing fluid properties change with time = −2 ∑ (hL uL )i − , (6-59)
and temperature. Typically, exposure to high ∂t i ∂x
temperatures reduces fluid viscosity. However, with the derivatives replaced by central finite-
crosslinkers may cause initial viscosity increases difference approximations to obtain
prior to the degradation. The effects of tempera-
∆t
VL + Ait (ui +1 − ui ) ,
h
ture and time are included in numerical hydraulic Ait + ∆t = Ait − (6-60)
fracture simulators, typically by means of tables ∆x ∆x
of K´ and n´ versus time at a series of tempera- where VL represents the volume leaked off over the
tures, which are similar to those in service com- element of length ∆x in time step ∆t. The velocities
pany handbooks. are calculated at the grid points, and the area is
• Numerical solution of the model assumed constant in each element. The cross-sec-
The three basic solutions described for height- tional area can thus be updated from the values of
growth mechanics (pressure-width-height relation), the velocities and areas at the previous time step.
conservation of mass and conservation of momen- Once this has been done, the pressure at each cross
tum (velocity-pressure relation) are coupled and section can be obtained from the solid mechanics
solved simultaneously. There are several methods solution by looking up the pressure in the precalcu-
by which the coupled equations may be solved, two lated pressure-height-width relation table for the
of which are introduced here. Either a fixed or corresponding area A. Pressure gradients can then
moving mesh may be used for the two methods, be calculated using the approximation
as described previously for planar 3D models. In ∂p pi −1 − pi
≈ (6-61)
∂x i ( xi −1 − xi +1 ) 2
this section, the explicit finite-difference method
is introduced with a grid that moves with the fluid
and an implicit method is described. In each case, and new velocities obtained using Eq. 6-57. Once
prior to starting the simulation of the fracture evo- all the velocities are known at a given time, the
lution, a table of the pressure-height-width relation positions of the grid points are updated using
(from the equilibrium-height solution) is calculated
as described for “Solid mechanics solution” in xi (t + ∆t ) = xi (t ) + ui (t + ∆t )∆t . (6-62)
Section 6.3-2. This method is known as a Lagrangian method
For the explicit finite-difference method, the fluid because the grid coordinates move with the fluid.
in the fracture at any time is divided into n ele- Leakoff causes each element to shrink and possibly
ments, each with a cross-sectional area Ai and even disappear as it penetrates farther into the frac-
bounded by two vertical surfaces at xi and xi + 1, ture, limiting the usefulness of this method for
moving at velocities ui and ui + 1, respectively, as modeling hydraulic fracturing treatments. In addi-
shown in Fig. 6-10. (The grid is numbered such tion, new elements must continually be added at
that i = 1 represents the tip to facilitate the addition the wellbore. This makes it difficult to control how

many elements are present at any time or the sizes preceding text, is that the time step used in the cal-
of the elements. Another approach is to introduce a culation may not exceed a critical value to ensure
fixed grid, as discussed for planar 3D models. This stability. Because only quantities from the previous
has the advantage that the number of elements in step are used in moving forward, numerical errors
the simulation is relatively small near the beginning can grow larger from step to step if the time step is
of the simulation when less accuracy is required too large. In the development of a general hydraulic
and increases as the simulation progresses. Yet fracturing simulator using such differencing
another approach is to introduce a moving mesh schemes, the time step must be chosen carefully to
in which the grid points move at some reasonable avoid stability problems and yet minimize the com-
velocity, for example, such that the fracture is putation time. A simple stability analysis is in
always divided into a fixed number of equal-size Sidebar 6H.
elements (i.e., using a stretching coordinate system; It has been found that in cases of high leakoff
see Sidebar 6H). or large widths (such as TSO designs), the critical
One of the primary limitations of explicit finite- time step for stability may be too small for efficient
difference methods, such as those introduced in the solution of the system, limiting the utility of the
6H. Stretching coordinate system and Substituting Eq. 6H-5 into Eq. 6-59 and applying the chain rule,
stability analysis
∂A D  1+n ∂A 
1/ n −1
 1+n ∂ 2 A  ∂A  
2
= A A A + (1+ n )A n   , (6H-7)
Stretching coordinate system ∂t n  ∂x   ∂x 2
 ∂x  
One way to simplify grid point bookkeeping is to use a where absolute values must be assumed for all quantities,
stretching coordinate system. If because an error analysis is being performed, and D is
defined as
x
X = , (6H-1)
L(t ) ΦC p1/ n
D= . (6H-8)
(2Kh )
1/ n
1+ n
then X will always remain bounded between 0 and 1 while x f
varies between 0 and L(t). Placing a grid on X will fully cover The highest order term in Eq. 6H-7 is
the fracture regardless of the growth characteristics. However,
1/ n −1
although the gridding is simplified, the complexity of the differ- D  1+n ∂A   1+n ∂ 2 A 
ential equation is increased. The derivatives are found as A A
n  ∂x  A ∂x 2  . (6H-9)
 
∂ 1 ∂
= (6H-2) If the derivative is expanded using a central difference
∂x L ∂X
approximation, the term in A i becomes
∂ ∂ X dL ∂ −2D ∂A
1/ n −1
= − . (6H-3) A 3 +1/ n . (6H-10)

∂t ∂t L dt ∂X n ( ∆x )
2
∂x
Equation 6-59 becomes
To investigate the effect of an error introduced into A, A is
∂A X dL ∂A 1 ∂q replaced by A(1 + ε), which can be approximated (for small
− =− − qL , (6H-4) ε) as
∂t L dt ∂X L ∂X
1/ n −1
and the other equations of the system are similarly trans- −2D ∂A   1 
A 3 +1/ n 1+  3 +  ε . (6H-11)
formed. n ( ∆x ) ∂x   n 
2
Stability analysis If a time step is taken (discretizing Eq. 6H-7 similar to Eq. 6-59),
then the error ε grows to
A full stability analysis for a nonlinear system is difficult, but
1/ n −1
an approximate time-step limitation can be found as follows. −2D ∂A   1 
Assume that the pressure gradient can be written as E = ∆t A 3 +1/ n 1+  3 +  ε . (6H-12)
n ( ∆x ) ∂x   n 
2
∂p ∂A
= Cp . (6H-5) For this error to reduce in magnitude, it must be smaller than
∂x ∂x Aε, which can occur only if
(∆x )
2
In the case of the PKN model, where the fracture height hf is
fixed, Cp = βh f , where β is defined by ∆t < , (6H-13)
2AC p C v
p = βw . (6H-6) where the viscosity leakoff control coefficient Cv is
Cv = −
(3n + 1)v . (6H-14)
∂p 2
n
∂x

explicit scheme. An implicit finite-difference (1961) assumption that q is constant (Eq. 6-10),
scheme, with no time-step limitation, may elimi- replacing q by πuhfw/4 allows writing Eq. 6-65 as
nate this limitation. In essence, implicit and explicit
dw 32µh f
methods can be distinguished by the fact that = u (6-66)
explicit methods solve for quantities at one time dx E ′w 2
step on the basis of only values at the previous time or
steps, whereas implicit methods include the use of 32µh f
quantities at the current time. This implies that a set w 2 dw = udx . (6-67)
E′
of equations is set up and solved, because the quan-
tities at the current time step must all be found Integrating over a distance ∆x obtains
simultaneously. For linear problems, a set of linear 1/ 3
 3 x + ∆x
96µh f 
equations results, and these are easily solved by w( x + ∆x ) = w ( x ) + ∫ udx  . (6-68)
standard methods such as gaussian elimination. For  x
E′ 
the 1D flow problem, the implicit finite-difference
If the terms under the integral can be assumed to
formulation yields a tridiagonal system of equa-
be constant, this simplifies further to
tions (i.e., a sparse matrix with only three diagonals
1/ 3
filled with nonzeros). Highly efficient solution  96µh f 
techniques are available to solve such systems w( x + ∆x ) = w 3 ( x ) + u∆x  . (6-69)
 E′ 
(e.g., Carnahan et al., 1969). For nonlinear prob-
lems, however, such methods can be complex and If the height is not constant and the fluid is non-
are not always much more efficient than explicit Newtonian, a similar equation can be written for
methods. Iteration is frequently required, because the cross-sectional area of the fracture by using the
a nonlinear system is linearized. If the linearization power law rheological parameters:
approximation is inaccurate, it must be corrected 1/( n + 2 )
 (n + 2)πKh nf +3 n 
A = ( Ait ) +
n+2
and resolved. t
i +1 u ∆x  , (6-70)
Another method without the time-step limitation,  E ′Φ n 
and which avoids forming a system of equations,
where
is a method using integrated or analytical elements.
1+ 2 n
A similar method to that described in the following n  w( y ) 
∫
n
Φ= dy . (6-71)
was the basis of the commercial time-sharing h f (2 + 4 n) h  w 
method made available by Amoco between 1981 f
and 1983 (Nolte, 1982, 1988a). Consider once For an analytical solution, ∆x would be the entire
again the basic equations of the PKN model with fracture length (Nolte, 1991), and this would be
x = φ at the tip: combined with a tip criterion and a volume-balance
2 pnet h f equation. The numerical solution proceeds simi-
w= (6-63) larly, except that ∆x is chosen sufficiently small to
E′ obtain an accurate solution. Fluid loss is integrated
dp 64 qµ over the time step, which allows obtaining accept-
= . (6-64) able accuracy, even with large time steps. The solu-
dx πh f w 3
tion method at each time step is as follows:
Substituting Eq. 6-64 for p into Eq. 6-63 obtains 1. Estimate a tip velocity.
E ′ dw 64 qµ 2. For each element, working in from the tip to the
= . (6-65) well,
2 h f dx πh f w 3
a. calculate an average fluid velocity based on
Detailed numerical simulations have shown that the velocity at the outer side of the element
the velocity varies much more slowly than the flow and the estimated velocity at the inner side
rate q because the reduction in width toward the tip (At the first iteration, assume the inner fluid
partially compensates for fluid leakoff and storage velocity is equal to the outer fluid velocity.)
in the fracture. Instead of the Perkins and Kern

b. determine the cross-sectional area at the inner tion is quite accurate, provided that the fluid is
side of the element for the estimated velocity moving relatively slowly in the vertical direction so
by using Eq. 6-70 that the pressure drop resulting from vertical fluid
c. determine the velocity at the inner side such flow is negligible. This assumption is violated if
that the leakoff and element volume change high-permeability zones are exposed, because fluid
during the time step result in a mass balance must then move rapidly because of the increased
d. repeat the iteration using the new velocity. leakoff in such layers. Also, if the stress in the sur-
3. Compare the actual flow into the fracture with rounding zones is insufficient to confine the frac-
the wellbore velocity calculated by the iterative ture and the vertical tips extend quickly, then the
scheme in the preceding step. fluid must move quickly to fill the resulting frac-
ture. In either of these cases, the pressure gradient
4. Refine the estimate of the tip velocity using a
resulting from vertical fluid flow may become
Newton-Raphson method until volume balance
large, and the equilibrium-height assumption
is achieved, which typically takes two to four
becomes invalid at these locations in the fracture.
iterations.
To remove this assumption and obtain valid
This method of solving the equations is efficient results from a simulator, some restriction must
because the velocity does not vary significantly be placed on height growth. For nonequilibrium-
along the fracture for typical cases. For typical height growth, the pressure gradient because of
PKN cases with a single fluid, the fracture can fluid flow in the vertical direction must be approxi-
be divided into about 10 elements. For non-PKN mated, based on the rate of height growth. It is
cases, the grid must be chosen sufficiently fine common to base this approximation on the KGD
that the integrand in Eq. 6-68 (which includes model (e.g., Settari and Cleary, 1982). In Section
effects of fluid rheology and fracture height) is 6-7 on tip effects, an analytical near-tip solution
approximately constant in each element (because developed by Lenoach (1994) is discussed that pro-
the solution scheme is derived with the assump- vides an expression for the net pressure of the form
tion that it is constant).
Regardless of whether a moving- or fixed-grid
method is used, usually only a small number of
elements (about 10) is necessary to obtain a
reasonably accurate solution to the equations
described so far. However, other information may
be required at a much finer resolution. To achieve
this, the schedule is typically divided into a large (6-72)
number of substages (about 100). Quantities such
as proppant concentration, fluid temperature and
acid concentration can then be tracked on this where utip is the tip velocity and β is 2/(2 + n). As
finer grid. In addition, particularly in acid frac- previously noted, for a fracture under constant pres-
turing, it is desirable to track leakoff and etching sure, the stress intensity factor is related to the net
on a finer grid. To do this for methods using pressure by
a moving grid, a second grid that does not move K Ic
pnet = . (6-73)
is established. Quantities such as reservoir tem- πh f 2
perature, proppant bank height and leakoff vol-
The Lenoach tip solution can be used to obtain
ume in the reservoir are tracked on this solid-
an apparent fracture toughness caused by the non-
based grid.
zero tip velocity by combining Eqs. 6-72 and 6-73.
• Nonequilibrium-height solution This effect can be added to the actual rock tough-
It was noted in “Solid mechanics solution” in ness, and the sum is used in Eqs. 6-46 and 6-47
Section 6.3-2 that the assumption of slow height instead of the actual rock toughness to determine
growth allows creating a pressure-height-width the fracture height growth. The basic algorithm
table prior to solving the equations of fracture evo- used to move from one pair of vertical tip positions
lution. This so-called equilibrium-height assump- to another during a time step is as follows:

1. Estimate the top and bottom tip velocities for ally coupled PKN models. The pressure predicted
the cell. by the laterally coupled model is always higher
2. Calculate the new fracture tip positions at the than either the KGD or PKN solution would pre-
end of the time step, using these velocities. dict. It can also be shown that the width at the well
is always smaller than that predicted by either of
3. Calculate the stress intensity factors from Eqs.
the simple models. The point in Fig. 6-11 where
6-46 and 6-47.
the pressure from the laterally coupled model is
4. Determine the excess stress intensity factors lowest (and where the pressures from the KGD and
(i.e., the calculated value minus the rock PKN models are equal) corresponds to a square,
toughness). where the fracture wing length is one-half of the
5. Calculate the velocities required to generate height. The pressure calculated by the laterally cou-
these excess stress intensity factors, using Eqs. pled model exceeds that predicted by the KGD or
6-72 and 6-73. PKN model at this time by approximately 40%,
6. Compare the velocities with the guessed values which is comparable to the pressure in a radial
and iterate until the correct velocities are found. fracture of similar dimensions.
One of the advantages of the equilibrium-height
models is the speed gained by precalculating a 1400
table of the fracture height-pressure relation. Not
only is this not possible for the nonequilibrium 1300
Pressure (psi)
model, but the iterative process to determine the tip
positions can be time consuming. The nonequilib- 1200
Lateral coupling
rium-height algorithm should therefore be used KGD
PKN
only when necessary because of the apparent rapid 1100
height growth indicated by the equilibrium-height
calculation. 1000
0 5 10 15 20 25
• Lateral coupling Time (min)
In the description of the solid mechanics solution
provided previously, the basic assumption is that Figure 6-11. Pressure record with and without lateral
coupling.
individual cross sections act independently (i.e.,
plane strain in the horizontal direction, or laterally
decoupled). This is implicit in the assumption that
the pressure and width at any point are uniquely 6-3.3. Lumped pseudo-three-
related. In reality, the pressure at any point is depen- dimensional models
dent not only on the local width, but also on the
Lumped models are an alternative to cell-based mod-
width distribution over the entire fracture, as dis-
els and were first introduced by Cleary (1980b).
cussed in Section 6-3.1 on planar 3D models. This
Although more details are presented in subsequent
lateral coupling is generally not important, unless
paragraphs, it is worthwhile at this point to quote two
the fracture wing length is less than the height. Even
sentences from the conclusions of his paper: “the heart
then, the fracture geometry will not be significantly
of the formulae can be extracted very simply by a
different if lateral coupling is neglected, although
nondimensionalization of the governing equations; the
the pressure response may be underestimated. Lat-
remainder just involves a good physico-mathematical
eral coupling can be included in the solutions
choice of the undetermined coefficients” and “results
described previously (see Sidebar 6E).
could be presented in the usual format of design
The effect of lateral coupling during pumping is
charts, based on dimensionless groups extracted, . . .
to increase the pressure at and near the well and to
[a] more appealing procedure may be to program the
decrease it near the tip. Figure 6-11 shows the evo-
solutions for a suitable pocket calculator, with the sep-
lution of pressure during a treatment for a confined
arately determinable γ or Γ coefficients and job para-
fracture simulated using the KGD, PKN and later-
meters as input.” Although numerous papers have

been presented on the use of this model (e.g., Cleary addition of γ3, and Eq. 6-77 is identical to that of
et al., 1994), which is now on laptop computers rather Geertsma and de Klerk (1969) with γ3 replacing
than pocket calculators, these sentences capture the 4(1 – ν2). Superficially, these equations are extremely
essence of lumped models in that they are extremely simple, but the values of the γ coefficients are not
simple models and the key to their successful use always obvious and may not be constant. As noted
is the selection of appropriate coefficients for the by Crockett et al. (1989), these models are extremely
problem analyzed. general, with the degree of accuracy limited ultimately
Like all the models presented previously, the start- only by the effort invested in determination of the γ
ing point of the lumped model is the set of basic equa- coefficients by detailed simulations, laboratory experi-
tions defining the hydraulic fracturing process, which ments or field studies.
are mass conservation (Eq. 6-39), the relation between For more general fracture shapes (i.e., with height
the distribution of crack opening over the length of growth), it is typically assumed that height growth is
the fracture 2L, net pressure distribution (similar to governed by a KGD-type solution and length growth
Eq. 6-36) expressed as by a PKN-type solution (Cleary et al., 1983), although
L this is not a theoretical limitation of lumped models.
pnet ( x ) = ∫ I ( x, x ′)w( x ′)dx ′,
−L
(6-74) One area in which lumped models have been
exploited extensively is in the development of com-
and conservation of momentum (Eq. 6-40) expressed as puter software systems to apply and use pressure data
during a treatment. Some of the key characteristics
˜ m−1 = − γ 4 w 2 n− m+1∇
qq ˜ p µ, (6-75) and requirements for such a software system are that
(Crockett et al., 1989)
where γ4 is the channel factor (1⁄12 for a Newtonian
• the physics is realistic and general
fluid), and various combinations of the power law fac-
tors m for turbulence and n enable consideration of • execution time is much faster than treatment time
both non-Newtonian fluids and turbulent flow. to allow repetitive execution for pressure history
In the lumped models, these equations are simpli- matching
fied by assuming a fracture shape and adopting a spa- • the software can use improved estimates of parame-
tial averaging approach to reduce them to ordinary ters obtained in real time (i.e., during the treatment).
differential equations in time. This approach implicitly
Although these real-time software systems are gen-
requires the assumption of a self-similar fracture
erally referred to as real-time hydraulic fracture mod-
shape (i.e., one that is the same as time evolves,
els (e.g., Crockett et al., 1989), the model itself is only
except for a length scale). The shape is generally
a small part of the software and should address the
assumed to consist of two half-ellipses of equal lateral
first requirement listed (i.e., realistic and general
extent, but with different vertical extent.
physics). The second and third requirements are com-
It is instructive to consider some of the lumped
puter hardware and software design constraints.
equations for the KGD model (Cleary, 1980b). The
Because lumped models were developed for pocket
mass balance is obtained by averaging over the frac-
calculators in 1980, they impose a minimal impact on
ture length:
computer hardware and systems. As computing power
ρ(qw − Lq L ) = d ( γ 3ρwL) dt , (6-76) continues to improve, it will become possible to run
increasingly sophisticated models during treatment
where
execution either at the wellsite or remotely. There are
w ≈ ( γ 1 pnet E ) L (6-77) other software design issues, such as robust execution
with a wide variety of parameter values, easy import
q m w m ≈ γ 5m ( w )
2 n+2
L2 , (6-78) and superposition of actual data on model output, and
where graphical display, that are required for a useful soft-
ware system for real-time applications. Discussion of
γ 5m ≡ γ 2 γ 4 E ( γ 1µ ) , (6-79) these issues is beyond the scope of this volume.
which is the 1D form of Eq. 6-75. Equation 6-76 is
similar to Eq. 6-15 (based on Carter, 1957) with the

6-4. Leakoff 6-4.1. Filter cake

One of the key issues in designing a fracture treatment In laboratory filtration experiments performed at con-
is accurate knowledge of how rapidly fluid will leak stant pressure, the rate of filtration is proportional to
out of the fracture into the reservoir. Without this the square root of time (see Chapter 8). A model for
information, it would be impossible to design a treat- this process can be derived by assuming that
ment that provides a specified fracture geometry. • the amount of cake deposited is proportional to the
Mini-fracture treatments are performed to estimate the volume of fluid VL passed through a unit surface
leakoff coefficient (see Chapter 9). Equation 6-14 area
introduced in “Inclusion of leakoff” in Section 6-2.2
• cake permeability kcake is independent of its thickness
is the basic equation of filtration and was first used for
fracturing by Carter (1957). He showed that it was • flow through the cake obeys Darcy’s law
applicable for three separate leakoff processes: • pressure drop across the cake ∆pcake is constant
• displacement and compressibility of reservoir fluid to write
• invasion of the formation by filtrate or fracturing fluid dVL kcake α∆pcake
uL = = , (6-80)
• buildup of an external filter cake. dt µ fil VL
Williams (1970) divided the leakoff into three time where µfil is the viscosity of the filtrate and it is assumed
periods, rather than considering the three processes. that the cake thickness Lcake is proportional to the fluid
During the initial period, leakoff is quick, followed volume lost; i.e., VL = αLcake. By integrating for VL, it
by a decreasing leakoff rate and finally a steady-state can be shown that
leakoff rate. In the initial period, filter cake has not
uL = Cw t, (6-81)
formed, so the leakoff rate is controlled by the resis-
tance of the formation to flow of the fracturing fluid. where the fluid-loss coefficient through the wall filter
The external filter cake builds during the second cake is
period. Finally, the cake stops building, because the kcake α∆pcake
high-velocity fluid in the fracture prevents further Cw = . (6-82)
2µ fil
polymer deposition. This last stage is referred to as
dynamic leakoff. Williams lumped all leakoff prior Carter (1957) proposed that the volume leaked off
to dynamic leakoff into a quantity he called spurt can then be determined as
volume, although spurt loss has since been generally
accepted to refer only to the initial high-leakoff period VL = 2Cw t + Sp , (6-83)
before the cake starts building. where Sp is the volume that leaks off without forming
Settari (1985) presented an excellent review of the a filter cake and can be interpreted as an integration
classic leakoff model, as well as an even more general constant. A more appropriate physical model is to
model of leakoff that represents an excellent frame- assume that the initial volume that leaks off, without
work for leakoff modeling. In the following sections, building a cake, is the spurt volume and that Eq. 6-80
each of the three processes (displacement and com- applies after the cake is established. This interpreta-
pressibility of reservoir fluid, invasion of the formation results in
tion by filtrate or fracturing fluid and buildup of filter
cakes, either externally on low-permeability rocks or VL = 2Cw t − tsp + Sp , (6-84)
internally on high-permeability rocks) is considered where tsp is the spurt time.
as if it is the only one acting and then in combination. Another approach to account for leakoff by using
Finally, Settari’s general model is summarized. standard petroleum engineering concepts of reservoir
The fluid-loss derivations in Eqs. 6-82 through 6-91 permeability and treating the filter cake as a pressure-
are for the pressure drop for individual loss mecha- dependent resistance is outlined in Chapter 2 along
nisms. For general application, they are in terms of the with high-permeability conditions (Valkó and Econ-
total pressure drop between the fracture and initial omides, 1997). The following approach to the leakoff
reservoir pressures in Eqs. 6-94 through 6-96 and in coefficient is as presented by Settari (1985).
oilfield units in Eqs. 8-26 and 8-27.

6-4.2. Filtrate zone the front can be treated as the face of an infinite
porous medium, and an analytical solution (Collins,
The first zone inside the reservoir is called the filtrate 1961) is used to obtain
zone or invaded zone. It is assumed that
C
• pressure drop ∆pv across the zone is constant uL = c , (6-90)
t
• filtrate fully displaces the mobile phase(s) within where the compressibility control leakoff coefficient
the formation, resulting in piston-like displacement Cc is
and 100% filtrate saturation
kr ct φ
• the fluid and rock are incompressible Cc = ∆pc , (6-91)
πµ r
to write
where kr is the permeability of the reservoir rock and
dV k ∆pv µr is the reservoir fluid viscosity.
uL = L = fil , (6-85)
dt µ fil Lv
where kfil is the permeability related to the filtrate and
6-4.4. Combined mechanisms
Lv is the length of the invaded zone. Integrating this
equation, with the assumption that In practice, all three processes occur simultaneously.
1 The leakoff velocities in Eqs. 6-80, 6-85 and 6-90
φ∫
Lv = VL φ = uL dt , (6-86) must be equal, and the sum of the pressure drops must
equal the total pressure difference between the reser-
where φ is the porosity, obtains voir pressure and the fracturing pressure:
C
uL = v , (6-87) Ct kcake α∆pcake k fil φ∆pv kr ct φ
t = = = ∆pc
t 2µ fil (t − tsp ) 2µ fil t πµ r t
where the viscosity control leakoff coefficient Cv is
(6-92)
k fil φ∆pv
Cv = (6-88)
2µ fil ∆pcake + ∆pv + ∆pc = ∆ptotal , (6-93)
and the leakoff volume at any time is where Ct is the total leakoff coefficient and ∆ptotal is
the difference between the pressure in the fracture and
VL = 2Cv t . (6-89)
the far-field reservoir pressure pr. If the spurt volume
The permeability to the filtrate kfil reflects the rela- and time can be neglected, these equations can be
tive permeability of the formation to flow of the fil- combined (Williams et al., 1979) to yield the total
trate. This effect may be significant when a water leakoff coefficient:
filtrate enters a hydrocarbon zone at nearly irreducible 2Cc Cv Cw
water saturation. Ct = Cwcv =
Cv Cw + Cw Cv + 4Cc2 (Cv2 + Cw2 )
2 2
(6-94)
6-4.3. Reservoir zone
with the coefficients Cw, Cc and Cv calculated using
Although the uninvaded reservoir does not contain
the overall pressure difference.
fracturing fluid, pressure is required to displace the
Equation 6-94 is valid only if the cake permeability
reservoir fluid away from the fracture face. Assuming
is independent of pressure. If the cake is highly com-
• constant pressure drop ∆pc between the pressible and the cake permeability is approximately
filtrate/reservoir interface and the far-field reservoir proportional to 1/∆p, Nolte (1988a) has shown that the
• compressible flow with constant total compres- fluid loss is limited either by the cake or the reservoir.
sibility ct In that case, the fluid-loss rate is the minimum of
Eq. 6-81, with the pressure drop equal to the total
• relatively slow movement of the front of the
pressure drop, or
invading fluid
C
• an infinite reservoir, uL = cv , (6-95)
t

where the combined leakoff coefficient is product of gel concentration and fluid velocity. In this
2Cc Cv case, Eq. 6-98 is replaced by
Ccv = (6-96)
Cv + Cv2 + 4Cc2 2Cgel Cw2 t
Cw = , (6-99)
with the coefficients Cc and Cv again calculated using Mgel
the total pressure difference. where Cgel is the gel mass concentration in the fluid
and Mgel is its specific density. There is an implicit
assumption that the term CgelCw2 is constant. This
6-4.5. General model of leakoff assumption is consistent with laboratory work reported
A great deal of complexity can be added to the leakoff by Mayerhofer et al. (1991). The viscosity of water
model in an attempt to account for detailed behavior decreases with increasing temperature, and this effect
such as the compressibility of the invading fluid and on the leakoff coefficient should be included in the
the moving boundary of the reservoir fluid. Given fracture model.
the accuracy with which the other parameters in a
hydraulic fracture treatment are known, the inclusion
of such effects is generally unnecessary. This section 6-4.6. Other effects
describes modification of the models to incorporate • Pressure evolution during treatment
the two effects of the variable pressure difference and
changing fluid properties. If an estimate of the leakoff coefficient has been
The model described in previous sections can be obtained from a mini-fracture decline analysis, then
generalized to account for multiple fluids. Settari the most likely way to use the simulator would be
(1985) showed that the fluid loss in the invaded zone to enter the total leakoff coefficient derived from the
can be described by replacing the term Cv with an analysis. It is thus assumed that the leakoff behavior
equivalent term: during the minifracture is the best representation
of what will occur during the main treatment. How-
2Cv2 t ever, if the leakoff behavior is unknown and is to
Cv = , (6-97)
VL be determined from fluid and reservoir properties,
the best approach is to enter a laboratory-determined
where Cv is calculated using the average viscosity and
wall-building coefficient for each fluid and use the
relative permeability of all the filtrate leaked off to
simulator to determine the total leakoff coefficient
the current time, and VL is the fluid volume that previ-
at each position in the fracture as a function of time
ously leaked off into the reservoir. Settari also showed
and on the basis of the continually evolving pressure
that replacing the wall-building coefficient for the
difference between the fracture and the reservoir.
fluid under consideration with an equivalent value
Except in overpressured reservoirs, the assumption
would account for variations in leakoff behavior
of a constant total pressure difference is generally
between fluids in a treatment. The equivalent value
reasonable. The ratio of leakoff coefficients between
in this case is
the lowest pressure difference during the treatment
2Cw2 t (when the net pressure is zero and the pressure
Cw = , (6-98)
VL difference between the closure stress and the reser-
where the previously leaked-off fluid volume VL has voir pressure is just ∆p = σc – pr) and at the end of
also contributed to wall building. pumping (when it is ∆p = pnet + σc – pr) is given for
The critical fluid component affecting wall building reservoir-controlled leakoff by
is the gel and/or fluid-loss additive concentration. An Cc ( pnet = 0) σ c − pr
= (6-100)
Cc ( pnet = pnet ,max )
extension to Settari’s model can be derived by consid-
pnet ,max + σ c − pr
ering Eq. 6-80, in which the thickness of the cake is
assumed to be proportional not to the volume flowed and for the wall-building or viscosity-controlled
through the wall but to the volume of gel deposited. cases by
Thus, the thickness is proportional not to the time-
CL ( pnet = 0) σ c − pr
integrated velocity but to the integral over time of the = .
CL ( pnet = pnet ,max )
(6-101)
pnet ,max + σ c − pr

If the ratio is close to 1, a constant leakoff model 6-5.1. Effect of proppant on fracturing
can be used, in which the leakoff coefficients (Cc, fluid rheology
Cv and Cw) for each fluid in the treatment can be
precalculated on the basis of an assumed typical Generally the viscosity of a proppant-laden slurry µslurry
net pressure (e.g., 250 psi). If the effect of pressure is higher than that of the carrying fluid µbase alone.
changes is large, a variable leakoff model can be Experimental relations are well established for New-
used, in which the pressure changes in the fracture tonian fluids, but much less so for power law fluids.
are accounted for as the simulation proceeds. In this Nolte (1988b) showed that relations for power law
case, the reservoir component of leakoff Cc should fluids could be obtained by using the relations for New-
be determined using a convolution of the pressure tonian fluids and raising them to the power of n. For
history during the treatment. example, the viscosity ratio µr could be obtained as
• Pressure-sensitive leakoff µ slurry 1
µr = = 2.5 n , (6-102)
One of the major assumptions of the analysis in the µ base (1 − fv fvM )
previous sections is that the permeability remains where fv is the proppant volume fraction, and fvM is the
constant. In fact, many reservoirs may have fissures maximum fraction for a mobile slurry.
or other features that may open under the influence
of the fracture treatment. The effect of this opening
is to increase the leakoff rate (Warpinski, 1991). 6-5.2. Convection
Pressure-sensitive leakoff is addressed more fully
in Chapter 9. Density differences between fluids may result in the
denser fluid flowing under the lighter fluid or the
• Poroelasticity and backstress lighter fluid overriding the denser fluid. This phenom-
Chapter 3 discusses the influence of pore pressure enon, known as convection or gravitational flow, is
on rock mass behavior. Poroelastic effects are important in many fields, such as saltwater intrusion
changes in stress that occur as a result of changes under fresh water (Badon Ghyben, 1888; Herzberg,
in pore pressure in the reservoir. As fluid leaks out 1901). In fracturing, it may be relevant if a high-
of the fracture into the reservoir, the affected part density slurry stage flows under a previously pumped
of the reservoir dilates, and a “backstress” develops, stage or pad, as well as for other 2D aspects of fluid
which increases the effective closure pressure. This flow, such as those considered by Clifton and Wang
effect is generally small, but it may be important (1988).
in some cases, as discussed by Nolte et al. (1993). The fluid flow equations for a Newtonian fluid can
Chapter 3 provides the solution for a fracture in an be written as
infinite, homogeneous medium. Although the addi- − w 2 ∂p
tional pressure results in an increased net pressure ux =
12µ ∂x
in the fracture, it generally has little effect on frac-
− w 2  ∂p 
ture geometry. uy =  + ρg (6-103)
12µ  ∂y 
and (Eq. 6-38 for incompressible fluids)
6-5. Proppant placement
( )
 ∂( wux ) ∂ wuy  ∂w
+
The objective of hydraulic fracturing is to place prop-  + + 2uL = 0. (6-104)
pant to provide a conductive path for production. The  ∂x ∂y  ∂t
presence of proppant introduces three important issues
in the behavior of fluids in hydraulic fractures: Substituting Eq. 6-103 into Eq. 6-104 obtains
• effect of proppant on fluid rheology 1  ∂ 2 w 3 p ∂ 2 w 3 p  ∂w 1 ∂w 3ρg
• convection or gravity currents  +  = + 2 u − .
12µ  ∂x 2 ∂y 2  ∂t 12µ ∂y
L
• proppant transport.
(6-105)

The last (gravitational) term on the right-hand side centrated and may be close to the channel walls, which
of Eq. 6-105 is the convective term. This can be treated causes two effects: hindered settling, which implies
as a source term, just as the other two terms are storage that particles get in the way of each other, and clus-
or sink terms, resulting from width change and leakoff. tered settling, in which particles join together, effec-
Baree and Conway (1994), Unwin and Hammond tively increasing the diameter in Eq. 6-106. Novotny
(1995) and Smith and Klein (1995) showed that this (1977) presented a correlation for the particle velocity
is generally not significant for most properly designed usol in hindered settling in terms of the volume fraction
fracturing treatments. Smith and Klein showed that if of solids fv:
usol = u∞ F( fv ) ,
excess pad was pumped, the fluid flow after pumping
(6-107)
stops (i.e., afterflow) could lead to convection until the
pad leaked off. Also, Eq. 6-105 shows the extreme where
F( fv ) = (1 − fv )
sensitivity of convection to fracture width. If the width β
(6-108)
is large (e.g., in a low-modulus rock), convection may
be more critical. Fortunately, such low moduli are usu- and the exponent β ranges from 5.5 at low values of
ally associated with high permeabilities, in which case NRe to 2 at high values of NRe. For power law fluids,
TSO designs and rapid leakoff after shut-in effectively a generalized form of Stokes law (Eq. 6-106) is used:
( )
prevent convection. Cleary and Fonseca (1992) pre- 1/ n
 ρsol − ρ f gdsol
n +1

sented a dimensionless number that reflects the ratio u∞ =   . (6-109)
 3 18 K ′ 
n −1
of buoyant and viscous forces. This ratio can be used
to estimate the effect of different conditions on the
severity of convection. Equation 6-108 can still be used to account for hin-
Finally, Clark and Courington’s (1994) and Clark and dered settling. Other correlations have been developed,
Zhu’s (1994) experiments on convection largely verify but a definitive correlation has not appeared in the lit-
the theoretical and numerical results described here. erature. Many fracturing fluids are designed for almost
perfect transport, so the settling rate is usually not
important unless the fracture remains open for a long
6-5.3. Proppant transport time after pumping stops.
Another effect on proppant placement is fluid
Hydraulic fracturing produces a conductive channel migration (Nolte, 1988b) or encapsulation (Cleary
by placing proppant in a crack created in a pay zone. and Fonseca, 1992). Fracturing fluids are generally
Hence, an essential consideration in fracturing fluid viscoelastic. Although it is beyond the scope of this
design is to accomplish proppant transport. The effect section to discuss this phenomenon in detail, one of
of convection on proppant transport was previously its important effects is to drive proppant to the center
discussed. There are two other factors that may impact of the flow channel. This migration could result in a
proppant placement. The first, and most commonly dense sheet near the center of the channel, surrounded
understood, is settling. If a bottle containing a mixture by clear fluid. This has the effect of accelerating parti-
of sand and water is shaken up and then left on a cle settling, especially for low proppant concentrations.
table, the sand will settle out of the water. It can be Unwin and Hammond (1995) presented simulations
shown theoretically that the terminal velocity of a sin- showing the effect of this migration on proppant
gle particle far from any walls in a stagnant Newton- placement.
ian fluid is given by Stokes law:
u∞ =
( )
g ρsol − ρ f dsol
2
, (6-106)
18µ 6-6. Heat transfer models
where ρsol is the solid particle density, ρf is the fluid The properties of many fracturing fluids show some
density, and dsol is the solid particle diameter. dependence on temperature. In addition, the rates
The assumptions of this equation are of limited of the reactions that occur in acid fracturing are
applicability in hydraulic fracturing because the fluids dependent on temperature. In a typical fracturing
are non-Newtonian and the particles are highly con- treatment, the fluid is pumped at a temperature signifi-

cantly below the reservoir temperature. As the fluid

1.0
penetrates farther into the fracture, heat transfer occurs TD =
T – Ti
0.8 Tr – Ti
between the fracturing fluid and the rock, resulting 90%
in an increase in fluid temperature. 0.6
70%
cy
50%
ien
The temperature gradient in the direction perpendic-
TD
fic
30%
ef
ular to the fracture wall is significantly larger than 0.4
uid
20%
Fl
those in other directions, so the temperature gradients 0.2
10%
in the other directions can be neglected. In addition,
5%
heat conduction in the fluid can be ignored because 0
0 0.2 0.4 0.6 0.8 1.0
it is small relative to both conduction in the rock and
x⁄L
transport of heat with the moving fluid. These assump-
tions reduce the heat transfer problem to a 1D problem Figure 6-12. Temperature profile in a fracture for different
perpendicular to the fracture wall, with conduction fluid efficiencies. TD = dimensionless temperature,
T = absolute temperature, Ti = fluid temperature at the
through the rock to the fracture face and convection fracture mouth, Tr = reservoir temperature.
from the rock face into the fluid.
6-6.1. Historical heat transfer models 6-6.2. Improved heat transfer models
The first significant thermal model for hydraulic frac- Meyer (1987) developed a solution that accounts for
turing was published by Whitsitt and Dysart (1970). a finite-film, or convective, coefficient for heat transfer
To obtain an analytical solution, they assumed a leakoff between the rock and the fluid and also introduced the
rate that varies linearly from zero at the well to a maxi- power law Nusselt number to determine the value of
mum at the fracture tip and accounted for the inhibit- the convective heat transfer coefficient. This showed
ing effect of the leakoff, which occurs in the opposite that the effect of the finite-film coefficient is to reduce
direction to the heat transfer. Unfortunately, the solu- the rate of fluid heat-up.
tion they obtained contains an integral that must be Kamphuis et al. (1993) developed a numerical sim-
evaluated numerically. Two of their more significant ulator that accounted for similar effects. One of the
contributions are demonstration of the effect of tem- advantages of the numerical model is that it allows
perature on acid reaction rates for acid fractures and including more effects, such as variable pump rate
that the temperature in much of the fracture is well during the treatment, and, of more practical impor-
below the reservoir temperature, so that fluids could tance, calculating temperature changes after shut-in.
be designed for lower temperatures than previously This model requires the introduction of a calculation
believed. grid in the rock.
Sinclair (1971) obtained a solution to a similar Another algorithm has been developed to solve the
problem, except that he assumed a uniform leakoff heat transfer problem (see Sidebar 6I). It has many
rate along the fracture. An example of the results is of the advantages of the numerical solution mentioned
shown in Fig. 6-12. The significance of this figure is previously but is extremely computationally efficient.
the relatively small fluid heat-up that occurs when the The equation for the fluid temperature is uncondition-
fluid efficiency is low. For an efficiency of 10%, the ally stable; i.e., there is no upper limit on the time
temperature in the fracture is approximately the inlet step. The results of simulation with this method com-
temperature over about 80% of the fracture length. pare favorably with the full numerical solution of
At higher efficiencies, a more rapid heat-up occurs, Kamphuis et al. (see Sidebar 6J).
so that about 50% or more of the fracture length is
at or close to the reservoir temperature.
6-7. Fracture tip effects
All fracture models include the effects of rock defor-
mation (width), mass transport and fluid loss in similar
ways. However, the failure and opening of the fracture

6I. Efficient heat transfer algorithm Fn = h (T surf

n
− Tfln ) . (6I-8)
Mack and Elbel (1994) presented an efficient algorithm for the Equations 6I-3, 6I-6 and 6I-8 can be solved for Tfl n, Fn and
calculation of temperature changes in hydraulic fractures. Tsurfn to yield
Consider a semi-infinite rock mass with constant surface
Tr 0 − E n + C 2Tfln −1
flux F0 starting at time zero. The temperature change of the Tfln = (6I-9)
rock surface ∆Tsurf as a function of time t is (Carslaw and 1+ C 2
Jaeger, 1959)
D 1 (Tfln − Tfln −1 )
Fn = (6I-10)
∆T surf = −C 0F0 t , (6I-1) ∆t n
Fn
where n
T surf = Tfln + (6I-11)
h
2  κ
1/ 2
C0 =   ,
( ) ∆Dt
(6I-2)
kh  π C 2 = C 0 ∆t n + 1 h 1
. (6I-12)
where kh is the thermal conductivity of a solid and κ is the
n
thermal diffusivity of a solid. Defining D0 as 1/C0,
( )
For a piecewise constant-flux history, Eq. 6I-1 can be gen-
eralized to hD 0 ∆t n (Tr − E n ) + D 1 h ∆t n + D 0 Tfln −1
Tfln =
( )
. (6I-13)
hD 0 ∆t n + D 1 h ∆t n + D 0
= C 0 ∑ (Fi − Fi −1 ) t n − t i −1 ,
n
∆T n
surf
(6I-3)
i =1
It has been shown (Kamphuis et al., 1993) that the effect of
where ti and Fi represent the time and the surface flux, leakoff on the heat flux is equivalent to reducing the conduc-
respectively, at the end of the ith time step. Thus, tivity by the factor
exp( −P 2 )
∆T surf
n
= Tr 0 − E n − C 0Fn ∆t n , (6I-4) , (6I-14)
1+ erf(P )
where Tr0 is the initial reservoir temperature, ∆tn = tn – tn – 1, where P = CL/√kh.
and En represents the effect of all previous time steps, which Meyer (1987) showed that the Nusselt number
can be written as
hw
, (6I-15)
E n = C 0 ∑ (Fi − Fi −1 ) t n − t i −1 − C 0Fn −1 ∆t n .
n −1
(6I-5) k fl
i =1
where kfl is the thermal conductivity of the fluid, for non-
Now consider an element of fluid of height ∂y, length ∂x Newtonian fluids ranges from 6 for n´ = 0 to 4.11 for n´ = 1.
and width w/2 that experiences a change in temperature from A Nusselt number of 4.3 is most representative of typical frac-
Tfl n–1
at the beginning of a time step to Tfl n at the end of the turing fluids.
step. The quantity of heat required to cause this temperature If a fracture treatment is simulated with an explicit finite dif-
change is ρfCpflw∂x∂y (Tfl n – Tfl n – 1)/2. Assuming a constant ference scheme, it is not practical to retain the flux history of
flux over time step ∆tn, this implies that the flux and tempera- each solid grid point because many thousands of time steps
tures over the area ∂x∂y are related by may be required to simulate the entire treatment. It has been
found that using 5 to 10 steps to represent the flux history is
(Tfl
n
– Tfln –1 ) =
Fn
D1
∆t n , (6I-6)
sufficient, provided the time steps are merged in such a way
that the overall heat loss from the formation is conserved.
This results in an accurate representation of the most recent
where temperature changes in the simulation, yet retains computa-
tional and storage efficiency.
ρf C pflw
D1 = . (6I-7)
2
Finally, consider the effect of heat transfer by convection

from a rock surface at temperature Tsurfn to fluid at a tempera-
ture Tfl n. If the heat transfer coefficient is h, the flux is
at its tip boundary are addressed in numerous ways. • failure or opening process (normal LEFM)
Nevertheless, certain general principles can be • disturbed zone in the rock ahead of the fracture tip
described that apply to this region, and different imple- (damage not incorporated in the LEFM model)
mentations can be considered as modifications of the
• unwetted zone (fluid lag region)
general principles. If the fracture tip is envisioned as
the zone between the fracturing fluid and the undis- • disturbed zone along the fracture face (e.g., dila-
turbed rock ahead of the fracture, then there are four tancy or compaction).
possible features of this region that must be addressed:

6J. Verification of efficient thermal calculations

(mm)
The paper by Kamphuis et al. (1993) includes results for a 0 10 20 30 40 50 60
1000
Fracture toughness, KIc (kN•m –3/2 )

KGD fracture in sandstone. The parameters are all held con-
stant except for the leakoff coefficient CL, which has values 1000 900
of 4E–5, 1E–4, 2E–4 and 3E–4 m/s1/2. Figure 6J-1 compares 900 800
the results obtained using the method described in Sidebar
6I and those obtained by Kamphuis et al. The dimensionless 800
700
(psi/in.1/2 )
temperatures along the fracture are shown as a function of 700
the leakoff coefficient, with the lowest curve representing the Geometry 600
largest leakoff coefficient. The agreement is good, consider- 600 Divider
Arrester 500
ing the relatively small number of elements (eight) used in 500 Short transverse
this simulation and the relatively coarse nature of the heat Series 400
transfer algorithm compared to Kamphuis et al.’s detailed 400
A
finite-difference calculations. 300 300
B
200 200
1.0 0 0.4 0.8 1.2 1.6 2.0 2.4
Crack length (in.)
0.8
Figure 6-13. Scale dependence of fracture toughness in
(T – Ti) ⁄ (Tr – Ti)
0.6 the laboratory (Schmidt, 1976).
0.4
• Stresses around a crack tip
0.2 Irwin (1957) identified three different types of sin-
gular stress fields (i.e., stress approaches infinity)
0 around a crack tip and characterized these as
0 0.2 0.4 0.6 0.8 1.0 Mode I (opening), Mode II (in plane sliding) and
Distance ⁄ length Mode III (antiplane sliding). For hydraulic fracture
modeling, Mode I is of primary interest, although
Figure 6J-1. Comparison of temperature calculations
(solid lines) with the results of Kamphuis et al. (1993) the other modes come into play in more compli-
(dashed lines). cated situations such as fracture turning from devi-
ated wells. For a 2D crack opened by a constant
internal pressure, Irwin showed that the stress
These four mechanisms are typically neglected intensity factor KI is simply
or handled in an ad hoc manner because of a lack of
understanding and data, particularly on a field scale, K I = πL pnet , (6-110)
about this complex zone.
where L is the crack length and pnet is the net inter-
nal pressure opening the crack. Similarly, for a
6-7.1. Linear elastic fracture mechanics radial crack
R
Although early studies of fracture in rock used Griffith’s KI = 2 pnet , (6-111)
(1921) crack theory and surface energy (Barenblatt, π
1962; Perkins and Krech, 1968; Friedman et al., 1972), where R is the crack radius (see Sidebar 6K).
most analyses of rock fracture are now formulated in LEFM, as postulated by Irwin, holds that the
terms of LEFM. The advantage of LEFM over earlier crack will advance when the value of KI exceeds
theories is that it incorporates, within a simple frame- some critical value KIc of the material, called the
work, some degree of dissipative energy processes, critical stress intensity factor. More commonly
such as plastic flow and microcracking, when the zone known as fracture toughness, KIc can be related
of dissipation is small compared with the fracture to the surface energy of previous studies through
length (see plateau region for KIc on Fig. 6-13). How-
ever, when this zone is not relatively small, energy- 2 Eγ F
K Ic = , (6-112)
release methods should be used, as discussed in 1 − ν2
Chapter 3.

stant” value as the crack size reaches some thresh-

6K. Crack tip stresses and the Rice equation
old value. Figure 6-13 shows example results for
Irwin (1957) found that the opening mode of a 2D crack has Indiana limestone that led Schmidt to conclude that
a singular stress distribution in the near-crack-tip region: KIc is a material constant. However, the small size
of the laboratory samples is several orders of mag-
σx =
KI
2πr
[
cos (θ 2) 1− sin (θ 2) sin (3θ 2) ] (6K-1)
nitude different than that of field-size hydraulic
fractures, and size effects, which would invalidate
σy =
KI
2πr
[
cos (θ 2) 1+ sin (θ 2) sin (3θ 2) ] (6K-2) the application of LEFM to this process, cannot be
definitely ruled out.
τ xy =
KI
2πr
[
cos (θ 2) sin (θ 2) sin (θ 2) , ] (6K-3) Even without size difficulties, most rock material
“constants” (such as Young’s modulus) are not con-
where θ is the angle measured from the crack axis, r is the stant and vary with confining stress, temperature,
distance from the crack tip, and KI is the stress intensity factor.
By comparing the stress field given by Eqs. 6K-2 and 6K-3 strain rate and the size of the rock mass tested.
with the solution of the stress field around the tip of a 2D Similarly, in Schmidt and Huddle’s (1977) work
crack extending from –L to L, Rice (1968) showed that KI can
be calculated as
with Indiana limestone, a significant increase of the
critical stress intensity factor with confining stress
L+x
L
1
πL ∫
KI = p (x ) dx , (6K-4) was measured. Thiercelin (1987) confirmed this
−L
L−x
behavior but also showed that the amount of the
where p(x) is the pressure distribution in the crack. For con- increase is strongly dependent on the rock fabric
stant pressure in the crack, Eq. 6K-4 reduces to Eq. 6-110.
For a radial crack, the equivalent equation is and other factors.
R
rp (r )
Assuming that KIc is a material constant and scal-
2
πR ∫
KI = dr , (6K-5) ing is not a problem, fracture toughness can be
0 R2 −r 2
incorporated in a 2D or P3D crack model by inte-
which, for constant pressure, reduces to Eq. 6-111.
grating Eq. 6K-4 with the model-derived pressure
distribution (or using Eq. 6-110 with a weighted-
where the specific fracture surface energy γF average pressure) to obtain the stress intensity fac-
includes localized dissipative effects and can tor KI. Even an unwetted region can be included by
be determined in the laboratory, as discussed in modifying the pressure distribution used in
Chapter 3. The attractiveness of this theory is its Eq. 6K-4, as discussed later. For the calculated
ability to include all the complicated failure pro- value of KI:
cesses in one parameter, which, hypothetically, is – If KI exceeds the input value of KIc, the crack is
a material constant similar to modulus or strength. allowed to advance.
However, for the general case of dissipative effects, – If the value of KI is less than or equal to KIc, the
the linear elastic stresses given by Eqs. 6K-1 crack remains in its same position.
through 6K-3 may no longer apply near the crack tip.
– If KI becomes negative, the crack must retreat
• Application of fracture toughness to hydraulic (i.e., the width reduces to zero) until KI becomes
fracturing in rocks ≥0, as negative values indicate that the internal
Although LEFM is attractive in its simplicity, two pressure is insufficient to support the entire crack
questions remain concerning its application to size.
hydraulic fracturing: For field-size hydraulic fractures, L or R and pnet
– Is KIc a material property of rocks and what are are so large that the stress intensity at a hydraulic
its characteristics? fracture crack tip is much larger than typical labo-
– Does KIc require modification for hydraulic frac- ratory KIc values of 500–2000 psi/in.1/2. Hence, it is
turing applications? usually assumed that KIc for this normal crack
The first question is difficult to answer because growth process, as understood from the laboratory,
of scaling problems. In the initial application of KIc is negligible except for small fractures and initial
to rocks, Schmidt (1976) showed a clear size effect growth. The following sections discuss modifica-
at small scales, but KIc appears to approach a “con- tions to this theory.

6-7.2. Extensions to LEFM L

L
ptip
In the practical application of hydraulic fracturing, the
K Iclag = 2 ∫
π L−d L − x 2
2
dx
measured net pressures are usually larger than those L π L − d

= 2 ptip  − arcsin (6-114)
L  
predicted by models (Shlyapobersky, 1985, 1988a, 
π 2
1988b; Cleary et al., 1991). Although these high net
pressures could be due to better than expected confor a 2D (KGD) crack, where ptip is the net pressure
tainment, poor measurement of the closure stress, (pressure near the tip minus the closure pressure) in
near-wellbore effects, complex fracturing, poor under- the nonwetted region and d is the fluid lag distance.
standing of rheology and many other factors, the gen- Typically, ptip values are between –σc + pr and –σc +
eral tendency has been to focus on the tip region as pvapor, where σc is the closure stress, pr is the reser-
the source of these anomalous results. Considerable voir pressure, and pvapor is the vapor pressure of
work on fracture tip effects began after Shlyapobersky the fluid. The fracture propagation criterion now
(1985) suggested that hydraulic fracture data could becomes
be interpreted to show a scale effect on KIc for field-
p(t )
L−d
L
size fractures. Since then, three additional major 2
π ∫ L −x
2 2
dx = K Ic + K Iclag . (6-115)
mechanisms have been proposed to account for the 0
high net pressures at the crack tip: fluid lag effects Similar equations can be written for radial cracks
(Jeffrey, 1989; Gardner, 1992; Advani et al., 1993), (Jeffrey, 1989).
dilatancy (Cleary et al., 1991) and damage (Yew and Detailed study of the crack tip has led to the dis-
Liu, 1993; Valkó and Economides, 1993a). covery of concepts unique to hydraulic fracturing.
• Fluid lag region Modeling of the crack tip region by the SCR Geo-
As applied to hydraulic fracturing, the unwetted mechanics Group (1993; Lenoach, 1995) shows
zone near the crack tip has pressure less than the that even when effects of fracture toughness KIc
closure pressure and hence acts to clamp the frac- are ignored, the consequences of coupled fluid flow
ture tip closed and reduce the stress intensity in the and leakoff still result in a singularity at the crack
rock. This zone was first introduced by Khristian- tip. For an impermeable rock, the power of the
ovich and Zheltov (1955) and successfully used hydraulic fracture singularity is not 1⁄2 as for the
by Geertsma and de Klerk (1969) in modeling 2D rock behavior in LEFM, but rather n/(2 + n), where
fractures. From their initial formulations, it is clear n is the power law index of the fluid. For perme-
that this unwetted region could have an impact on able rocks, the power of the singularity is 3n/(4 + 4n),
fracture parameters if it were sufficiently large. In which is stronger than the impermeable singularity.
a simple 2D geometry, the application is straight- Thus, an important singularity in stress may exist
forward, but the size of the unwetted zone and the at the tip even under conditions in which the frac-
exact pressure in the zone must be assumed. The ture toughness plays no role. They also made
pressure within the unwetted region is most likely numerical simulations that show that the size of
the reservoir pressure for permeable rocks and the fluid lag region adjusts to meet the fracture
could be as low as the vapor pressure of the fluid propagation criterion.
for impermeable rocks, so limits can be placed on Models that determine the size of the unwetted
its value. The size of the unwetted region is a more region (Jeffrey, 1989; Gardner, 1992; Yew and Liu,
difficult problem and has been the subject of con- 1993; SCR Geomechanics Group, 1993) generally
siderable investigation. produce small unwetted lengths, except at small
Fluid lag can be incorporated into the standard confining stresses. However, only a small region is
KIc form by defining an effective fracture toughness required near the tip to overshadow the effect from
(Jeffrey, 1989): the fracture body. One shallow field experiment
(relatively low confining stress) had sizable fluid
K Iceff = K Ic + K Iclag , (6-113) lag zones (Warpinski, 1985), but no careful field
study of fluid lag distances at higher confining
where
stresses has been made. Fracture models where the
lag distance is calculated generally show only a

small influence on global fracture parameters; a ing the rock experiences dilatant-plastic yielding,
good example is calculation of the 3D aspects of (Papanastasiou and Thiercelin, 1993) show no over-
the lag effect by Advani et al. (1993). Earlier in all increase in the width at the fracture tip because
this chapter, it was shown that the dimension that plastic deformation dominates the dilation effect.
controls fracture width is the smaller one. Hence, De Pater et al. (1993) tried to identify dilatancy in
the KGD model is most applicable for short frac- laboratory tests and through detailed modeling of
tures and the PKN model for long fractures. the fracture tip, but no clear evidence of such
Advani et al.’s study shows that a similar effect behavior was obtained.
occurs for the fracture mechanics; i.e., the length • Other behavior: damage- and fluid-induced effects
L in Eq. 6-114 must be replaced by the fracture
Another approach to modeling near-tip rock behav-
height if the fracture length exceeds the height.
ior is to use a cumulative damage approach, in
• Dilatancy which the microcrack damage ahead of the crack
Although the LEFM concept can include small- forces the undamaged material to accept more of
scale damage and plasticity within its framework, the load. Valkó and Economides (1993a) formulated
the possibility exists that the damage zone around a fracture model using this approach. Their model
a field-size hydraulic fracture could be sufficiently scales the damage with the fracture length, which
large that the near-tip stress distribution becomes is consistent with the KGD model and applies only
invalid, or other effects could alter the stress distri- to relatively short fractures. This analysis would
bution. For example, Cleary et al. (1991) suggested be applicable for longer fractures if applied to a tip
that the mechanism responsible for elevating crack element or with a scaling criterion.
tip pressures is dilatancy just behind the fracture Because most of the modeling efforts associated
tip. They postulated that if this dilatancy occurs with hydraulic fracturing deal with either fluid
during the rock failure process, then the fracture mechanics or rock mechanics, the strong chemical
width just behind the tip may be pinched slightly interactions that can affect rock behavior are often
by the expanded fracture. Dilatation of material, forgotten. An extreme example of this effect is the
which is essentially a volumetric expansion caused large reduction in strength that can occur as a result
by failure, cannot be accommodated by the sur- of stress corrosion cracking. Similarly, there has
rounding elastic material, so the rock stress in the been some evidence that the chemistry, rheology
near-tip region must increase. Yew and Liu (1993) or molecular structure may influence tip behavior.
developed a modified fracture toughness to include Holder et al. (1993) conducted laboratory tests in
dilatation of the material ahead of the crack tip for the which the inferred fracture toughness values using
case where a plastic zone is created around the tip. crosslinked gels were substantially greater than
An approximate equation for this behavior, in terms those with linear gels or Newtonian fluids. Dunning
of an effective fracture toughness, was given as (1980) found that surfactants can have a major
effect on the crack propagation stress (or, alterna-
144α 2 1 − ν  E − Et 
K Iceff = K Ic +   KI , tively, the fracture toughness). The effects of pH,
(
3π 2 1 + 3α ) 1 + ν  EEt  total ions, and breakers and other fluid additives
(6-116) can cause additional chemical effects. However,
these effects can influence tip failure only in the
where α is found by solving absence of a fluid lag region or where the fluid pen-
3α etrates a damaged area ahead of the fracture tip.
tan ϕ = , (6-117)
1 − 12α 2
where ϕ is the friction angle and Et is the Young’s 6-7.3. Field calibration
modulus of the plastic material.
Obtaining a definitive description of the fracture tip
Throughout the body of a fracture such dilatancy
behavior is complex and difficult. From a practical
is negligible, but it could play an important role in
standpoint, because the tip pressure must be consistent
the near-tip region, where the width is small. How-
with field observations, its magnitude and impact can
ever, numerical simulations of crack growth, assum-
be estimated using observed pressure data, as dis-

cussed in Chapters 5 and 9. To maintain a credible 6-8.1. Fracture geometry around a wellbore
description of fracture geometry, this should not be an
ad hoc procedure. Shlyapobersky et al. (1988a, 1988b) Several researchers have investigated mechanisms
proposed a method to calibrate a field-scale effective related to fracture initiation in vertical and deviated
fracture toughness that assumes that the fracture stops wells. Behrmann and Elbel (1991) and Daneshy
growing soon after shut-in and fluid flow within the (1973) found that the perforation must be oriented
fracture stops. By determining an initial shut-in pres- within about 10° to 20° of the plane normal to the min-
sure (ISIP) at the time when fracture growth stops (if imum far-field stress for a fracture to initiate at the
possible) and the closure stress in the formation, the perforation and extend. Other experiments show that
method measures a value of pnet associated with an when the perforations are not oriented in the direction
open fracture at the end of pumping. Given pnet, the of far-field fracture propagation and the well is devi-
effective fracture toughness is calculated from ated, the fractures can be nonplanar or S shaped
(Weijers, 1995; El Rabaa, 1989). However, predicting
K Iceff = α g pnet Reff , (6-118) the near-wellbore pressure drop in deviated wells is
difficult because of the uncertainty of the near-well
where Reff is one-half of the fracture height for a long fracture geometry.
2D crack or the crack radius for a penny-shaped A symmetric bi-wing planar fracture is generally
crack, and αg is a geometry coefficient, which depends assumed to develop when a hydraulic fracture treat-
on the geometry for a rectangular crack or is 0.64 for ment is performed. Hydraulic fracture models such
a penny-shaped crack. as the planar and P3D models described previously
This and other procedures to define tip effects by do not account for fracture initiation and near-wellbore
field calibration require an accurate measurement of effects. Apart from multiple fractures, the near-
pnet as well as reliable information on the fracture wellbore effects described here have no effect on
geometry (hf, L and R), fluid flow within the fracture the overall fracture geometry, except if a near-wellbore
after shut-in and the expected nature of the rock’s fail- screenout is caused by near-wellbore effects. This
ure behavior. is in contrast to fracture tip effects (see Section 6-7),
which may affect fracture geometry significantly.
The purpose of modeling near-wellbore effects is
6-8. Tortuosity and other near-well twofold: to understand the source of near-wellbore
screenouts, so that they may be predicted and pre-
effects vented, and to correctly remove the near-wellbore
High near-wellbore friction losses have been observed contribution from the measured “net pressure” so that
in fracture treatments, particularly in deviated wells or the remaining net pressure may be interpreted correctly
when the perforations are inadequate or poorly as a characteristic of the overall fracture geometry.
designed. Some attempts have been made to under-
stand the effect of near-wellbore geometry on the
placement of hydraulic fractures (Aud et al., 1994) and 6-8.2. Perforation and deviation effects
to develop methods to prevent unplanned screenouts The three assumed components of near-wellbore pres-
(Cleary et al., 1993; Stadulis, 1995). Near-wellbore sure loss are friction through the perforation, fracture
friction losses have been attributed to phenomena such turning (i.e., tortuosity) and perforation misalignment
as wellbore communication (perforations), tortuosity friction, which are also assumed to be additive:
(fracture turning and twisting), perforation phasing
misalignment and induced rock pinching, and multiple ∆pnear wellbore = ∆p pf + ∆ptort + ∆pmisalign . (6-119)
fractures (e.g., Stadulis, 1995). These effects have been
It is not possible to predict near-wellbore effects, other
identified as detrimental to the success of a fracturing
than friction through perforations. Rather, models for
treatment because of the increase in net pressure and
these mechanisms of pressure increase are provided,
the increased likelihood of unplanned screenouts
and each mechanism has one or more parameters,
caused by the limited fracture width near the wellbore.
which can be evaluated from field data.

6-8.3. Perforation friction

A discussion of perforating requirements for hydraulic
fracturing is in Chapter 11. Insufficient or poor perfo-
rations can have a significant effect on the execution Initial perforation C = 0.56
and evaluation of a fracturing treatment because they
affect the breakdown and treating pressure. Improper
Casing
perforating can result in near-wellbore screenouts if
the perforations do not provide an adequate pathway
to the main body of the fracture. The equation com-
Final perforation C = 0.89
monly used to calculate perforation friction implicitly
assumes that the perforation is a short cylindrical tun-
nel (McClain, 1963):
q 2ρ
∆p pf = 0.2369 , (6-120) Figure 6-14. Evolution of the coefficient of discharge C
n 2 Dp4 C 2 with the erosion of entrance (Crump and Conway, 1988).
where q is the total flow rate, ρ is the fluid density,
n is the number of perforations, Dp is the perforation 400
diameter, and C is the discharge coefficient. The dis- Net pressure with perforation friction
350 Net pressure without perforation friction
charge coefficient represents the effect of the perfora- Proppant
tion entrance shape on the friction pressure. 300 enters
Net pressure (psi) perforations
The effect of perforation friction on fracture treating 250
pressure is usually negligible if the perforations are 200
correctly sized and phased. If this is not the case, per-
150
foration friction is assumed to be constant during the Start of
100 shut-in
entire treatment. When sand slurries are pumped at
high differential pressure across the perforations, the 50
pressure drop changes, owing to erosion. There are 0
two effects of erosion on the pressure drop through 0 100 200 300
a perforation: smoothing of the entrance of the perfo- Time (min)
ration, with a resulting increase in the discharge coef-
ficient C, and an increase in diameter Dp. Figure 6-14 Figure 6-15. Effect of perforation friction on measured net
pressure.
shows the related evolution of the coefficient of dis-
charge with the perforation geometry (Crump and
Conway, 1988). These effects, and their implementa-
tion in a fracture simulator, are described in more
detail in Romero et al. (1995).
6-8.4. Tortuosity
Figure 6-15 illustrates the difference between the Tortuosity is defined here as a convoluted pathway
resulting pressure responses when perforation friction connecting the wellbore to the main body of the frac-
and erosion are included in the calculation and when ture. Several studies (Aud et al., 1994; Cleary et al.,
they are neglected for a PKN geometry model. The 1993) have identified tortuosity as an important phe-
pressure increases as expected for a confined fracture, nomenon that could affect the execution of a fracture
until proppant reaches the perforations. Then the pres- treatment when the wellbore and stress fields are mis-
sure decreases, mainly because of the increase in the aligned. The simplified schematic of fracture geome-
discharge coefficient. After about 2000 lbm of sand try in Fig. 6-16 shows how a fracture may turn and
is injected, the slope becomes positive again, almost twist to align itself with the preferred fracture plane.
paralleling the slope prior to the sand, which indicates The fracture width is proportional to the difference
a constant discharge coefficient and a slow increase between the pressure in the fracture and the stress
of the perforation diameter. against which the fracture opens. When the fracture

of the curved path is attached to the main body of the

fracture. Equation 6-121 is used to define the shape of
the path, and the outer boundary conditions (i.e., width
and pressure at the end of the tortuous region) are the
conditions at the wellbore, obtained from the simulator
Wellbore
without tortuosity. Although the calculated pressure at
Fracture
plane the well may be higher than that predicted when tortu-
at infinity osity is neglected, the width may be lower because the
stress against which the fracture is opening is higher.
This may result in screenouts caused by near-wellbore
bridging, which can be accounted for in the model by
Fracture preventing proppant from entering if the width is too
reorientation
small.
The effect of tortuosity is largest near the beginning
of the treatment and decreases as the treatment pro-
ceeds. This occurs because an increment in the closure
θ stress, relative to that on a planar fracture, has a fixed
absolute effect (∆w) on the width w. However, the
pressure drop is, roughly speaking, inversely propor-
Figure 6-16. The fracture twists and turns to align itself
with the preferred direction of propagation. tional to the width cubed, so that a change in width
from w to w – ∆w has a much greater effect when w
is small (i.e., when the fracture is first created). The
is opening against a stress higher than the minimum model also shows that the pressure drop caused by
in-situ stress, the fracture width is reduced relative to tortuosity can be reduced by increasing the fluid vis-
that without turning. If the ratio between the stress cosity, which has been reported in practice (Aud et al.,
against which the fracture is opening and the minimum 1994) as an effective means of preventing near-well
in-situ stress is higher than about 1.5, the fracture screenouts. Both added pump time prior to the intro-
mouth acts as a nozzle, allowing fluid to enter, but with duction of proppant (i.e., increased pad) and increased
a large pressure drop associated with the pinching of viscosity may reduce near-wellbore screenouts
the fracture width at the well. This process of fracture because they cause the width to be greater in the tortu-
width reduction along the reorientation path restricts ous region than it would have been. However, these
flow and could cause near-wellbore screenouts. treatment changes can be detrimental to height con-
The radius of curvature R of the reorientation path finement and proppant placement and permeability in
can be determined for a Newtonian fluid as (Romero the resulting fracture, so other approaches to mitigate
et al., 1995) the cause of tortuosity should be considered.
2
E 3µq  1 
R=λ (6-121)
h f  σ h,min (κ − 1) 
 , 6-8.5. Phasing misalignment
Perforating practices (i.e., hole size, spacing and ori-
where λ is an experimental coefficient, q is the flow
entation) vary widely. In general, not all the perfora-
rate, σh,min is the minimum horizontal stress, and κ is
tions in a well are aligned with the preferred fracture
the ratio between the stress against which the fracture
plane. Indeed, it would be quite coincidental for this
is opening and the minimum stress. The coefficient λ
to be the case, unless special efforts are made to
is obtained from experimental data (e.g., Abass et al.,
obtain reliable information on the stress directions at
1994) or field data and can be considered a fitting
a particular well. If 0° phasing is used, the orientation
parameter.
of the perforation to the plane of the hydraulic fracture
Fracture simulators such as the planar or P3D simu-
may be as large as 90°. On the other hand, nearly per-
lators discussed previously represent the behavior of
fect alignment or 0° phasing causes preferential prop-
the main body of the fracture, but an additional com-
agation of one wing of the fracture with limited
ponent is required to represent the tortuosity. A model

penetration of the companion wing because of the

pressure drop resulting from flow around the annulus
to the nonconnected wing.
Nolte (1988a) pointed out that if the fracture does Displacement
(in.)
not initiate at the perforations, the fluid must commu- –0.07
nicate with the fracture through a narrow channel
around the side of the casing. This channel can cause
higher treating pressures because of the width restric-
0
tions (Fig. 6-17). As with the tortuosity effect dis-
Cement ⁄ rock
cussed previously, this can cause both increased interface
pressure and screenouts because of proppant bridging.
Also, proppant may erode the restrictions. The circle
in Fig. 6-17 represents a relatively stiff wellbore (cas-
ing and cement). If the fluid exits the well through the Fracture
perforation, it must traverse the microannulus and
pass the restriction area before entering the main body
Figure 6-18. The displacement field around a fracture
of the fracture. A geometry effect occurs as the rock shows a tendency to pinch closed at the wellbore.
is displaced by a distance w away from the cement,
resulting in a channel around the annulus with a width
of w2/8D at the fracture entrance (point A in the fig- 0.040
ure), where w is the fracture width and D is the well-
bore diameter. In addition, an elastic response (Pois-
0.030
son’s effect) occurs in which the fracture opening
Displacement (in.)
results in movement of the rock toward the wellbore,

reducing the fracture width. 0.020
Figure 6-18 shows the displacement (in the direction
of the fracture) obtained around the wellbore for a typ- 0.010
ical case in which the microannulus and fracture are
subject to a constant fluid pressure. The negative dis-
0
placement of the rock at the intersection between the 0 200 400 600 800 1000
wellbore and the fracture represents the wellbore Net pressure (psi)
pinching from the net pressure in the fracture (Poisson’s
effect). To maintain flow into the fracture through the Figure 6-19. Evolution of wellbore pinching with increasing
pinch point, the microannulus must be pressurized to net pressure.
a higher level than the fracture. Figure 6-19 shows the
pinching displacement when the net pressure varies from 0 to 1000 psi for a typical case. The effect
increases as fracturing pressure increases, in contrast
to the tortuosity effect, which decreases as pressure
Restriction and width increase.
Wellbore If the pinch point is present when proppant attempts
to enter the fracture, bridging may occur, resulting in
premature screenout. The fluid travels through the
pinch point at a high velocity, and either fluid or slurry
A may erode the pinch point, provided this occurs
before bridging. The degree of erosion is affected by
Channel to Perforation
fracture wings the viscosity of the fluid, proppant concentration and
rock strength. The reported effectiveness of proppant
slugs (Cleary et al., 1993; Stadulis, 1995) may be due
Figure 6-17. Nonalignment of perforations and the fracture
plane causes pinch points. to this erosion. Because the slugs are small, they do

not bridge everywhere, so fluid entry at higher veloc- 6-9.1. Historical acid fracturing models
ity continues and erodes some channels. Even prior
to the use of slugs, it was common to inject proppant Williams et al. (1979) provided a detailed discussion
at low concentration to erode restrictions when high of acid fracturing models prior to 1980, and Li et al.
pressures occurred during the fracture initiation stage. (1993) reviewed some of the more recent work. One
In contrast to tortuosity, pinching is increased by large of the main drawbacks of most of the early models
pads and higher net pressures. This may explain why, was that the fracture geometry calculation was sepa-
in some cases, prepad slugs can be injected at low net rated from the acid reaction calculation to develop
pressures when the pinching is smaller, but in the analytically tractable solutions. Since then, computer-
main treatment a near-well screenout occurs. based models have overcome these limitations, and
the preceding models are no longer used. For exam-
ple, Settari (1993) presented a detailed description of
a comprehensive model with the fracture geometry
6-9. Acid fracturing and acid reaction calculations coupled, including a
Hydraulic fracturing with acid (usually hydrochloric comprehensive leakoff model, coupled heat transfer
acid [HCl]) is an alternative to propped fractures in and the capability to include multiple fluids with vary-
acid-soluble formations such as dolomites and lime- ing rheology.
stones. The major difference between acid and propped Much of what follows is based on the work
fractures is that conductivity is obtained by etching the described by Settari. Roodhart et al. (1993) presented
fracture faces instead of by using a proppant to prevent a model in which they developed the heat transfer
the fracture from closing. Acid fracturing may be pre- calculations extensively, using the work of Kamphuis
ferred operationally because the potential for unin- et al. (1993). The other factor they included was the
tended proppant bridging and proppant flowback is effect of the boundary layer thickness for acid reaction
avoided. However, designing and controlling the depth developing as the fluid enters the fracture, resulting
of penetration of the live acid into the formation and in a thinner layer near the well and causing a higher
the etched conductivity are more difficult than control- etching rate. The aforementioned publications contain
ling proppant placement. Acid penetration is governed extensive lists of references.
by the chemical reaction between the rock and the
fracturing fluid (as opposed to a simple mass balance
in propped fractures), and conductivity is determined 6-9.2. Reaction stoichiometry
by the etching patterns formed by the reacting acid (as The main chemical reactions of interest in acid frac-
opposed to being a property of the proppant under a turing are those between HCl and calcium carbonate
given stress). In both cases, acid fracturing introduces (limestone) or calcium-magnesium carbonate (dolo-
a dependence on rock properties that is not present in mite). The chemical reaction for limestone is written as
propped fracturing. In addition, the properties that acid
fracturing design and control depend on are usually more 2HCl + CaCO3 H2O + CO2 + CaCl2
difficult to determine than other formation properties. and for dolomite as
The geometry of acid fractures can be determined
by the same models used for propped fractures, with 4HCl + CaMg(CO3)2 2H2O + 2CO2 + CaCl2
the exception of the impact of etched width on the + MgCl2
width-pressure relation. However, several additional The first reaction equation indicates that two mole-
aspects of acid fracturing must be considered: cules of HCl react with one molecule of calcium car-
• acid transport to and reaction at the rock surface bonate to form one molecule each of water, carbon
dioxide and calcium chloride. The second equation
• heat transfer, because the reaction releases heat, and
shows that four molecules of HCl react with one mole-
the reaction rate is temperature sensitive
cule of calcium-magnesium carbonate to form two
• leakoff, because acid leakoff behavior is signifi- molecules each of water and carbon dioxide and one
cantly different from that of nonreactive fluids. each of calcium chloride and magnesium chloride.

These so-called stoichiometric equations allow com- 6-9.3. Acid fracture conductivity
puting the volume of rock dissolved by a given volume
of acid. These equations can be used to determine the Acid fracture conductivity is much more poorly
dissolving power XC of the acid, which is the volume understood than propped fracture conductivity. The
of rock dissolved per unit volume of acid reacted. The flow rate through an open channel of width w is pro-
mass dissolving power (i.e., the mass of rock dissolved portional to w3. If the etched channel were under no
per unit mass of acid reacted) is first defined as stress, this proportional relation would be used to
determine the conductivity of an acid fracture. How-
β=
(molecular weight of rock ) × (rock stoichiometric coefficient ) . ever, the stress in the reservoir acts to close the channel.
(molecular weight of acid ) × (acid stoichiometric coefficient ) If the etching were completely uniform, this closing
(6-122) could be calculated in a manner similar to that used
to calculate the width of an elliptical fracture, except
For the limestone reaction,
that the net pressure is negative. As an approximation,
100.09 × 1 for a uniform etched width, the shape of the resulting
β= = 1.372 (6-123)
36.47 × 2 closed fracture would be
4σ c
so that each gram of 100% pure HCl dissolves w( z ) = wetch − h 2f − z 2 , (6-126)
1.372 g of rock. To obtain the dissolving power, E′
the masses must be converted to volumes: where wetch is the etched width, and the width is set
ρCβC to zero wherever Eq. 6-126 predicts a negative width.
XC = , (6-124) It is apparent from this equation that the width in most
ρCaCO3
of the channel would be much lower than that of an
where ρC and ρCaCO3 are the densities of the acid solu- open channel, as most of the channel would have
tion and calcium carbonate, respectively, and C is the closed completely. This would clearly reduce fracture
weight-fraction concentration (e.g., 0.28 for 28% conductivity significantly.
acid). For example, the specific gravity of 28% acid Fortunately, acid etches the rock surface in a non-
is 1.14, whereas that for 15% acid is 1.07. A complete uniform manner, because of rock heterogeneity and
table of densities is in Williams et al. (1979). Applying fingering of the acid through the wider previously
this calculation for the limestone-HCl reaction, X15 is etched channels. This results in numerous horizontal
0.082, and X28 is 0.161. Similarly, for dolomite the “pillars” supporting the channels between them, for
values are 0.071 and 0.141, respectively. which Eq. 6-126 could be used with the fixed fracture
The stoichiometric equations for acid reactions pro- height hf replaced by the distance between the pillars.
vide a relation for the coupling between fracture It is not practical to model this in detail, because the
geometry and acid spending. Because there are many pattern is not generally known. Because conductivity
unknowns in acid fracturing, the modeling can be is higher in formations where numerous small chan-
simplified by neglecting the variation in density of the nels occur supported by numerous pillars, uniform
acid and using that of 10% acid, which is a suitable etching is not desirable. If, however, the pillars lack
average for most acid fracture treatments. In this case, the strength to support the additional load required to
X100 can be approximated as 10X10 and XC as CX100. keep the channels open, some of the pillars will col-
Now, consider the volume of a fracture element of lapse, reducing the conductivity. Fracture conductivity
cross-sectional area (width times height) A and length is thus dependent not only on the etching pattern, but
δx in which the acid concentration changes by an also on the rock strength and closure stress. Nierode
amount ∆C. The volume of acid spent is A ⋅ δx ⋅ ∆C, and Kruk (1973) developed an empirical equation for
and the volume ∆A ⋅ δx of rock dissolved is conductivity:
∆Aetch = X100 A∆C , (6-125) (wk )eff = C1 exp( −C2 σ) , (6-127)
–
where Aetch is the etched area and C is the average acid where
concentration in the cross section. C1 = 0.265wk 0fi.822 (6-128)

 (13.9 − 1.3 ln SRE ) × 10 −3 0 psi < SRE < 20, 000 psi rate constants are functions of temperature, following
C2 = 
(3.8 − 0.28 ln SRE ) × 10 20, 000 psi < SRE < 500, 000 psi
−3
the Arrhenius equation:
(6-129)  − ∆E  1 1 
ζ Tf = ζ f exp  −
T
 . (6-133)
ref
3

wk fi = 9.36 × 10  etch  ,
w 13  R  T Tref  
(6-130)
 12 
The acid reaction rate at a surface is thus a complex
where σ is the effective stress in psi, SRE is the rock function of the activities of all species involved in the
embedment strength in psi, and wkfi is the conductivity reaction. Detailed modeling of the reaction in terms of
in md-in. A typographical error in the original paper is these activities is not required for a hydraulic fracture
corrected in Eq. 6-129. simulator, because of the large amount of uncertainty
in the other parameters. Instead, the reaction rate can
be assumed to be governed by the simple equation for
6-9.4. Energy balance during acid fracturing the rate of acid consumption r (Settari, 1993):
The total heat generated (per unit volume) by a ∂Macid
– –
( )
m
change in acid concentration ∆C is ∆C∆H, where ∆H r= = − Kr Cwall − Ceqm , (6-134)
∂t
is the heat of reaction. Coupling of the acid and heat
transfer models is provided by assuming that all the where the temperature-dependent reaction rate
heat initially increases the fluid temperature, resulting constant is
in a fluid temperature change of
 − ∆E  1 1 
∆ C∆H Kr = k0 exp  −  , (6-135)
∆Tfl = , (6-131)  R  T Tref  
ρ f Cpfl
where Macid is the moles of acid per unit rock face
where ρf is the fluid density and Cpfl is the fluid heat
area, t is the time, Cwall is the surface acid concentra-
capacity. Section 6-6 describes how heat is transferred
tion, Ceqm is the equilibrium concentration, m is the
between the fluid and the formation. The magnitude
order of reaction, k0 is the reaction rate constant at the
of the temperature change resulting from acid reaction
reference temperature Tref (298K [25°C]), ∆E is the
may be sufficient to cause the temperature of some
activation energy, R is the universal gas constant, and
fluids to exceed the reservoir temperature. It is thus
T is the absolute temperature.
particularly important to use a numerical temperature
Ceqm is generally zero for the reactions of interest.
calculation when simulating acid fracturing.
6-9.5. Reaction kinetics 6-9.6. Mass transfer

Before the reaction can occur at the fracture wall, the
Surface reactions such as the acid-rock reactions dis-
acid molecules must be transported to the wall. In a
cussed here are complex, even under laboratory condi-
stagnant fluid, diffusion in an ideal case can be
tions. In general, the liquid-phase reaction between
described by Fick’s law:
species A and B to form products C and D is gov-
erned by an expression of the form ∂CA
v A, x = − DA
, (6-136)
∂x
r = ξ f aAn aBn − ξ r aCn aDn ,
A B 1C D
(6-132)
where vA,x is the velocity of species A, DA is the mole-
where ξ f is the forward rate constant, ξ r is the reverse cular diffusion coefficient, CA is the acid concentra-
rate constant, and aX represents the chemical activity tion, and the derivative represents the concentration
of species X. In the reactions of interest in acid frac- gradient. Williams et al. (1979) proposed accounting
turing, reverse reactions are usually much slower than for leakoff by adding a term to the right-hand side of
forward reactions and can be neglected. In very dilute Eq. 6-136, resulting in
systems, the chemical activity is equal to the concen- ∂CA
tration. It is also usually observed that the reaction v A, x = − DA + CA u L . (6-137)
∂x

In flowing fluids, this equation is no longer valid, In this case, the reaction rate is termed mass-trans-
because acid transport is by convection rather than fer limited, because the rate at which it occurs is con-
diffusion. For acid fracture modeling, Eq. 6-136 is trolled by the rate at which live acid can be brought to
simply replaced by the rock surface. Similarly, if Kg + uL is very large
compared with Kr, then Eq. 6-143 is satisfied when
v A, x = K g (CA − Cwall ) + (CA − Cwall )uL , (6-138) – –
Cwall is approximately equal to C. In this case, C can
where the mass-transfer coefficient is replace Cwall on the left-hand side of Eq. 6-143, and
Eq. 6-134 can be written as
K g = Deff N Sh w , (6-139)
∂Macid
( )
m
r= = − Kr Cwall − Ceqm . (6-145)
where Deff is the effective acid diffusion coefficient. ∂t
The Sherwood number NSh is determined from the Equation 6-145 represents the reaction-rate- or
correlation (Lee and Roberts, 1980) kinetics-limited case in which the rate of acid con-
6.26 N Sc1/ 3 N Re < 1800 sumption is limited by the rate at the wall.

N Sh = 0.001104 N Re N Sc 1800 < N Re < 7000 ,
1.153 1/ 3
0.026 N 4 / 5 N 1/ 3 N Re > 7000 6-9.8. Acid fracturing: fracture geometry

 Re Sc
(6-140) model
where the Reynold’s and Schmidt numbers are The movement of acid perpendicular to the fracture
defined respectively by wall is considered in this section. The preceding sec-
tions discuss the fluid flow equations typically solved
2 wvρ f in fracture models. Acid movement within the fracture
N Re = (6-141)
µ can be modeled similarly to the movement of prop-
pant. For a fracture simulator to simulate acid fractur-
µ
N Sc = . (6-142) ing treatments accurately, several specific require-
Deff ρ f ments must be met relating to
• fluid tracking in the fracture and reservoir
6-9.7. Acid reaction model • recession of the active fracture length
• effect of etching on the relation between pressure
If reaction occurs, the acid concentration varies across
and width.
the fracture width, and the surface concentration is less
than the bulk acid concentration. The surface concen- Although typical fluid flow calculation schemes use
tration is such that the amount consumed at the surface a coarse grid (about 10 elements), accurate fluid front
is balanced by transport to the surface by diffusion. tracking can be obtained only by following up to 50
The wall concentration for a given bulk concentra- fluid stages. Typical treatments include only about 10
tion is obtained by equating the right-hand sides of different stages, but stages can be subdivided for bet-
Eqs. 6-134 and 6-138 to obtain ter tracking of the large gradients that may occur in
( ) = (K )(
acid concentration within a single stage. Also, a finer
)
m
Kr Cwall − Ceqm g + uL C − Cwall . (6-143) grid is required to track leakoff volumes into the for-
This equation, which is a general model of acid mation and formation exposure to fluid stages for
reaction, can easily be solved if m = 1 but is solved accurate modeling of the extreme differences in leak-
iteratively otherwise. If Kr is very large compared off characteristics and viscosity between acid and
with Kg + uL, then Eq. 6-143 is satisfied when Cwall nonacid stages.
is approximately equal to Ceqm. In this case, Ceqm can Acid fracturing treatments are typically designed
replace Cwall on the right-hand side of Eq. 6-143, and with sudden changes in flow rate because the different
Eq. 6-138 can be written as fluids in the treatment have significantly different fric-
tional properties. These sudden changes, as well as the
∂Macid
r=
∂t
( )(
= − K g + uL C − Ceqm . ) (6-144) high leakoff that may occur during pumping of the

acid stages, may cause recession of the active fracture may open and propagate in n layers. At any time, the
length. The simulator must model this recession, sum of the flow rates into all layers must equal the
which seldom occurs in proppant treatments. total injection rate. In addition, the sum of the closure
For a confined fracture in a homogeneous isotropic stress in a zone plus the pressure drops through the
elastic material, the relation between the net pressure path from the tip of the fracture in that zone to a refer-
and cross-sectional area A can be written as ence point in the well must be the same for each frac-
2E′ ture. This set of conditions can be expressed as
pnet = A. (6-146) n
πh 2f qi = ∑ qi , j (6-147)
j =1
A modification to this relation is required to account
for dissolution of the rock by acid. Only the elastic
area Aelas contributes to the net pressure in the fluid, pref = σ c , j + ∆pw , j ( qi , j ) + ∆pnear wellbore,i ( qi , j ) − ph, j + pcf , j ( qi , j ),
although the total area A (where A = Aelas + Aetch) is
(6-148)
available as a flow channel and to store the fluid mass.
Mack and Elbel (1994) presented example prob- which is applied for each fracture. The terms on the
lems illustrating the effects of some of the features right-hand side of Eq. 6-148 represent the closure
of acid fracturing models. stress, pressure drop in the fracture, pressure drop in
the near-wellbore region including the perforations,
hydrostatic pressure and casing friction, respectively.
6-10. Multilayer fracturing There are thus n + 1 unknowns (n flow rates qi,j and
reference pressure pref) and n + 1 equations describing
Many fracture treatments are performed in settings that the system. Equation 6-148 is highly nonlinear, but
result in the formation and extension of nearly isolated the system can nevertheless be solved by standard
fractures in different zones. Frequently it is desirable to techniques, as shown by Elbel et al. (1992).
fracture multiple zones simultaneously, because treat- Figure 6-21 shows an example of a multilayer frac-
ment of each zone separately would not be practical or ture treatment modeled as a set of PKN fractures. The
would be significantly more expensive. However, the fluid partitioning was measured using a spinner flow-
design of treatments for multiple zones requires some meter, and the downhole pressure was recorded. The
special considerations. For example, the amount of model accurately captures the behavior of the system.
each fluid stage entering each zone cannot be con- Figure 6-22 shows a more complex case. The effect
trolled by the engineer. Fluid partitioning is important, of a screenout in a layer reduces the flow into that
because it dictates the size of the individual fractures layer while increasing it into others. Another interest-
formed. In addition, if the partitioning is unfavorable, ing effect that the model shows is the effect of cross-
premature screenouts may occur in some zones. flow, in which fluid may flow between fractures after
Some early work on the propagation of multiple pumping ends. If this rate is excessive, proppant may
fractures (Lagrone and Rasmussen, 1963; Ahmed et be drawn out of one or more fractures and that flush
al., 1985; Cramer, 1987; Ben Naceur and Roegiers, fluid may be injected into other fractures, impairing
1990) considered fluid partitioning in a limited way near-wellbore fracture conductivity. The crossflow
(e.g., using a limited representation of the formation or also violates the assumptions of pressure decline
at only a single point in time). In the method described analysis, possibly resulting in an incorrect estimate
in this section, fluid partitioning is calculated through- of fluid loss.
out the treatment. Extension of the model to cases with height growth
To simulate the simultaneous propagation of multi- was reported by Mack et al. (1992). They showed that
ple fractures, a single-fracture model (either analytical significant differences in both fracture geometry and
or numerical) is integrated with a set of constraints flow partitioning can occur if the P3D representation
coupling the individual fractures. For the present, it is used for the individual fractures, because fracture
is assumed that the individual fractures are well sepa- height growth changes the relation between net pres-
rated, with no mechanical interaction or any fluid flow sure in the fracture and the flow rate into the fracture.
between fractures except via the well. In this case, the Figure 6-23 shows an example comparing the pres-
fractures can be represented as in Fig. 6-20. Fractures sure response and the resulting fracture geometry. In

qi
qi
B
σc pref pref
∆pcf,1
A
q1 ∆ph,1
q1
σc,1 ∆pw,1 ∆ppf,1 ∆pcf,2
q2 ∆ph,2
q2
σc,2 ∆pw,2 ∆phw,2
∆pcf,n
qn ∆ph,n
qn
σc,n ∆pw,n ∆phw,n
Figure 6-20. Relationships for multiple fractures propagating simultaneously.
5100 8
Measured Layer 1
Calculated Layer 2
Layer 3
Injection rate (bbl ⁄ min)
Pressure (psi)
4900 4
4700 0
0 2 4 6 8 10 0 2 4 6 8 10
Time (min) Time (min)
Figure 6-21. Multilayer fracture treatment modeled as a set of PKN fractures.

an analytical solution such as that described in Side-

21
bar 6L. However, the analytical solutions are generally
Screenout
applicable only for simple (i.e., radial, PKN or KGD)
18 Bottom models. Another alternative is to develop a pump
schedule generator that uses a numerical simulator.
15 This tool uses the simulator in a so-called inverse
mode to determine the schedule. It requires all the for-
Top
12 mation and fluid data necessary for a forward simula-
Rate (bbl ⁄ min)
tion. Instead of the schedule, however, the desired

9 propped length, minimum and maximum proppant
concentrations, and step in concentration between
Middle Screenout
6 stages are specified. A typical concentration range
could be 2 ppg minimum to 12 ppg maximum, with
3
steps of 2 ppg.
To obtain the desired schedule, the simulator is
started with a schedule derived from an analytic
0
Start proppant
approximation or a schedule with a small pad stage
and an arbitrarily sized slurry stage with the maxi-
–3
0 20 40 60 80 90 100 mum proppant concentration. As the simulation pro-
Time (min) ceeds, the simulation software monitors the leakoff
of individual fluid elements in the fracture. As the
Figure 6-22. Fluid rate into three fractures, showing effects fluid leaks off, the proppant concentration increases.
of screenout and crossflow.
If the user-specified maximum is exceeded, the simu-
lator adjusts the proppant concentration down to the
maximum value and keeps track of how much prop-
4500 pant has to be “converted” conceptually to fluid to
maintain this. In addition, the fracture length is
Depth (ft)
5250 tracked and the schedule continually extended until

the user-specified length is reached. When the desired
length is reached, it is relatively simple to determine
6000
3000 6000 –0.2 0.2 0 2500 how much proppant (if any) is left in each fluid ele-
Stress (psi) Width (in.) Length (ft) ment. This represents the amount of proppant that
should be in that element when pumped, providing a
design proppant schedule. If proppant does not reach
Figure 6-23. Pressure response and geometry of fractures
modeled with the P3D model. the fracture tip (i.e., some of the pad should have been
slurry), this can be accounted for.
There are two issues that make this process more
complex than as described. First, as previously noted,
general, the multiple fractures would not connect into proppant affects fluid rheology, so modification of the
a continuous fracture unless the wellbore were per- amount of proppant during the simulation affects the
fectly aligned (e.g., <2°) with the minimum stress fracture length. This is minimized by repeating the
directions. Except for this rare case, height growth calculation with the schedule generated by the previ-
would be inhibited after the tips overlapped. ous iteration as input. Three iterations are usually suf-
ficient to converge on a suitable schedule. A more
critical issue is that this method cannot easily account
6-11. Pump schedule generation for bridging. Except during the initial small pad stage,
there is always proppant everywhere in the simulated
It is time consuming to design the schedule required
fracture, although some of it may later be converted to
to achieve a fracture of desired length and proppant
fluid. Bridging is therefore ignored and accounted for
concentration. One way to simplify the task is to use

6L. Approximate proppant schedules In addition to the pad, the schedule requires the volume
fraction of proppant fv to be added following the pad
K. G. Nolte (Nolte, 1986b)
Schlumberger Dowell
fv ( τ) = fo τ ε ; ε =
(1− f p ); 0 < τ <1
−η
(6L-5)
The facility to approximate proppant schedules for routine and
η
tip screenout (TSO) designs based on fluid efficiency (Nolte,
1986b) is an important design tool. For a specified amount of to approximate a spatially uniform concentration of fo at the
fluid or proppant, the technique requires an efficiency esti- end of pumping. The dimensionless slurry time τ is 0 when
mate for the proppant treatment that can be determined from proppant addition begins and unity when pumping stops. This
a fracture calibration treatment after an appropriate adjust- definition provides that fp + τ reflects the total time. Equation
ment for differences in the injection times of the treatments. 6L-5 is illustrated as the curve in Fig. 6L-1, which also shows
Efficiency is the ratio of the fracture volume to the injected the division of volume between the pad fraction and the slurry
volume before pumping is stopped. As illustrated on Fig. 6L-1, fraction fs. The definition of ε leads to the shaded area under
it defines the area under the ramp addition curve. This sched- the fv curve, which is equal to the efficiency. The remaining
uling technique is reviewed here and generalized to include area is 1 – η, which reflects the ratio of the loss volume to the
the effect of fluid-loss spurt. injected volume.
Conceptually, spurt loss occurs only during the addition of For scheduling a TSO treatment to achieve the final frac-
new fracture area and before proppant reaches the fracture ture volume Vf relative to that at screenout Vfso, Nolte’s (1990)
tip and halts fracture extension. Therefore, for normal design result can be extended to include spurt:
practices (see Section 5-1.1), spurt loss Sp occurs only for the
pad fluid and must be isolated from the efficiency η to esti- Vf ( ∆t D ) ∆t D 1− ηso  g ( ∆t D ) 
mate the pad volume. The modified efficiency ηC excluding = 1+ +  − 1 (6L-6)
Vfso ηso κηso  g 0 
spurt, which reflects only the CL component of fluid loss, can
( )
be found from Nolte (1989) as
ηso Vf ( ∆t D ) t p − t so
1− η Sp ηp = ; ∆t D = (6L-7)
ηC = 1− ; κ = 1+ (6L-1) 1+ ∆t D Vfso t so
κ g 0CL t S
g ( ∆t D ) = (1+ ∆t D )sin −1(1+ ∆t D )
−1/ 2
+ ∆t D1/ 2 , (6L-8)
(κ − 1)(1− η) =
fLS = (1− η) − (1− η ) , (6L-2)
where ∆tD is the dimensionless time after screenout, ηso and
κ
C
ηp are the respective efficiencies at screenout and end of

where κ is the spurt factor for the case of total fluid loss with pumping, tp is the total pumping time, and tso is the time at
spurt relative to the case with no spurt and κ = 1, g0 ≅ 1.5 screenout. The term g(∆tD) is the low-efficiency dimensionless
(see Chapter 9), CL is the leakoff coefficient, tS is the time of fluid-loss function defined in the Appendix to Chapter 9, which
tip screenout or injection without a screenout, and fLS is the also provides additional TSO relations. Low efficiency is typi-
volume fraction lost to spurt during pumping. Various means cal for TSO treatments. The ratio of fracture volumes defined
for obtaining the value of κ are discussed in Chapter 9. by Eq. 6L-6 can be replaced by the ratio of corresponding
The pad fraction is defined as the ratio of the pad volume average widths, as graphically represented in Fig. 10-15 for
to the total fluid and proppant volume injected during time tS. various efficiency values. For proppant scheduling, the pad
In the absence of spurt loss (κ = 1), the pad fraction can be volume to achieve the TSO is found by using Eq. 6L-4 with ηso
expressed in various forms: and fLS corrected for spurt by Eqs. 6L-1 and 6L-2. The prop-
fp ( η) ≅ (1− η)
2 pant addition is obtained from Eq. 6L-5 in terms of the final
efficiency ηp. From a practical standpoint and to avoid prop-
≅
(1− η) = 1 ≅ 1 , (6L-3) pant screen\out midway in the fracture, the pad can be
(1+ η) (1+ 2ffL ) (1+ G * ) extended by a low-proppant-concentration stage, as
discussed in Section 10-4.2.
where ffL is the ratio of fracture to loss volume during injection
and is equal to η/(1 – η), and G* is the decline analysis vari-
able discussed in Chapter 9. The pad relation (1 – η)2 in Dimensionless slurry time, τ 1
Eq. 6L-3 gives a smaller value than the relation (1 – η)/(1 + η), 1
Dimensionless proppant
which can be alternatively expressed as shown. Numerically 0

simulated pad data fall between the two relations. When the
concentration, fv/fo
spurt loss becomes significant (κ > 1), the pad fraction is

composed of two components: the first is equivalent to the
no-spurt case given by Eq. 6L-3 and uses a value of ηC that 1–η
excludes spurt loss, and the second is the contribution of
spurt using Eq. 6L-2: η
fp (κ > 1) = fp ( ηC ) + fLS (6L-4)
V
Vi
0
0 1
fp fs
Figure 6L-1. Dimensionless proppant concentration Dimensionless injected volume
versus dimensionless injected volume.

only when the final design is simulated. Depending on are rarely well known. Obtaining such data is fre-
the required conductivity, proppant of smaller diame- quently difficult, expensive or both. This section
ter may be used for the treatment. describes a method to obtain data from the postjob
The method described here can be extended to TSO analysis of pressure recorded during a treatment.
designs by specifying both the fluid concentrations The only direct output from the formation during
and the desired areal concentration (e.g., 1 lbm of a fracture treatment is the pressure history measured
proppant per square foot of fracture face area). The during and after pumping the treatment. Chapter 9
simulator is run as previously described, except that discusses the interpretation of these pressure records
once the design length is reached, length extension is in detail. However, these analyses can be only quanti-
artificially prevented and pumping continued until the tatively accurate for relatively simple fracture geome-
fracture width is sufficient to obtain the desired areal tries. This section considers the application of a formal
concentration. The design of TSO treatments is dis- theory of inversion (see Sidebar 6M) to complement
cussed in Chapter 10. qualitative interpretation and to increase the quantita-
tive information available from the pressure record.
Inverse analysis is a method of characterizing a sys-
6-12. Pressure history matching tem from its response to an imposed input. In the case
of hydraulic fracturing, the system is the reservoir,
One of the most difficult and expensive aspects of a surrounding layers, the well and all associated para-
well-engineered fracture design is obtaining the input meters. The input is the pumping of a fluid, and the
required for the design simulators. Formation data, response is the pressure recorded during the treatment.
such as stresses, permeability and elastic properties, The pressure record is analyzed to extract the proper-
6M. Theory and method of pressure inversion Two cases can be distinguished: the measured data defined
by
p meas = Fmeas (x )
The first step in the application of pressure history inversion is
(6M-4)
parameterization of the problem. This involves defining which
properties are to be determined as well as setting bounds on
their values and relations between values of different parame- and the simulated data defined by
ters. For example, it may be assumed that the stress in a
layer is between 5000 and 6000 psi and that the stress in a p sim = Fsim (x ) . (6M-5)
neighboring layer is between 500 and 1000 psi higher. If the
—
parameters are represented by the vector x and the pressure
—
record by p : Equations 6M-4 and 6M-5 imply that if a model is used to
calculate the pressure data for a given set of parameters, it
p = F (x ) ,
will generate a pressure record. Similarly, in the field, a pres-
(6M-1)
sure record is generated by the system with a set of parame-
ters. The function F also has subscripts sim and meas to
where F represents the mechanics of fracture development emphasize that the model is not an exact representation of
and relates the observed pressure to the input parameters. reality, so even if the correct x– is found, the calculated and
The pressure vector is the sequence of discrete pressures measured pressures may not agree. For example, if the PKN
—
measured during the treatment. The vector x may be a list model is selected to match the data but if significant height
of selected parameters, such as growth has occurred, the pressure record generated by the
correct x– will not match the measured pressure.
x = [hf ,E ′, σ] , (6M-2) The objective of pressure history inversion is to minimize
the difference between the measured and calculated pressure
records, defined using an error function:
indicating that the parameters to be found are the fracture
(∑W P )
height, Young’s modulus and stress, and it is assumed that all r
1/ r
other parameters are specified. Symbolically, the inversion ε= i sim ,i − Pmeas ,i , (6M-6)
i
process can be written as
where the weighting factors Wi are typically set to 0 for points
x = F −1 ( p ) , (6M-3) to be ignored and to 1 for all other points. The points can also
be weighted according to the range of interest. For example,
if only the decline period is to be matched, Wi is set to 0 for all
which is analogous to inverting a matrix to solve a set of lin-
points during pumping. The minimization of ε can be per-
ear equations with a known right-hand side. In this case, how-
— formed numerically by a routine in a standard numerical
ever, the known vector p is the sequence of pressure read-
library. Essentially, the algorithm consists of selecting a
ings, the relation is highly nonlinear and cannot be solved
sequence of sets of parameter values until a satisfactory
directly, and there are many more pressure readings than
match is obtained, similar to the 1D Newton-Raphson method
there are unknown parameters.
(Press et al., 1986) for solving a single nonlinear equation.

ties of the formation so that the fracture geometry can Research Institute (Robinson et al., 1991) to evaluate
be determined. a pressure history inversion algorithm. These experi-
This approach is common in well testing, and some ments are ideal for this purpose, because more data
of the same pitfalls and limitations should be noted. were gathered in these wells than in typical commer-
For example, the pressure record should not be cial wells. Figure 6-24 shows the pressure match
assumed to be the only information available. Other obtained by inverting the perforation diameter and
information, such as logs, should be used to narrow the stresses in the layers bounding the pay zone in
the expected ranges of parameters or to specify rela- one well. For comparison, the inverted values of the
tions between them. In addition, the selection of the diameter and stresses are listed in Table 6-1 (Robinson
model types to be used should be made logically on et al., 1991). Gulrajani et al. (1996) also presented
the basis of other data. This is analogous to well test several field applications of pressure history inversion.
interpretation (Gringarten et al., 1974), in which diag- These examples show the wide range of applicability
nostic plots and the knowledge of boundary condi- of the technique, as well as the quality of the results
tions are used to specify model type (e.g., infinite that can be obtained by its application.
reservoir versus rectangular bounded reservoir) before
using an analysis package to determine the best esti-
mates of permeability, height, etc. If this preanalysis
400
is not done, there is a high risk of obtaining a good
match to the pressure history with the incorrect para-
300
meters because of the nonuniqueness of the response;
Pressure (psia)
Measured
i.e., two different sets of inputs may provide the same
200
output pressure. Gulrajani et al. (1996) discussed
nonuniqueness in detail. Other limitations of pressure 100
Calculated
history inversion analysis are the ability of the algo-

rithm to represent the mechanics and the time require- 0
ments for computer processing if a sophisticated 0 5 10 15 20 25 30 35 40 45 50
fracture model is used. Time (min)
Piggott et al. (1992) described a method for per-
Figure 6-24. Pressure match obtained using pressure his-
forming fracturing pressure history inversion to obtain tory inversion (Piggott et al., 1992).
formation properties. These properties can be used in
future designs for wells in the same field and also to
confirm or refute the assumptions of the design of the
pumped treatment. For example, if the postjob appli- Table 6-1. Parameters assumed and from data
inversion (P3D model single-layer simulation)
cation of pressure history inversion analysis indicates (Robinson et al., 1991).
that the stresses in the barriers were smaller than
expected, resulting in the occurrence of significant Assumed parameters
Young’s modulus 8 × 106 psi
height growth, the effect on geometry would be quan-
Poisson’s ratio 0.3
tified (i.e., significant height growth at the expense of
Fluid-loss height 42 ft
reduced length in the pay zone, possibly reducing pro- Closure pressure 6300 psi
duction significantly). This information could then be Number of perforations 35
used to adjust predictions for production from that Leakoff coefficient 0.0037 ft/min1/2
well and to modify input parameters for future well Initial fracture height 120 ft
designs. Pressure history inversion applied on a cali-
bration treatment could be used to redesign the main Parameters from inversion
treatment. Stress contrast below pay zone 337 psi
Stress contrast above pay zone 186 psi
A well-characterized data set is desirable for evalu-
Perforation diameter 0.18 in.
ating any pressure history inversion algorithm. Piggott
et al. used field experiments conducted by the Gas


◆
Fracturing Fluid
Chemistry and Proppants
Janet Gulbis, Schlumberger Dowell
Richard M. Hodge, Conoco
7-1. Introduction 7-2. Water-base fluids

The fracturing fluid is a critical component of the Because of their low cost, high performance and
hydraulic fracturing treatment. Its main functions ease of handling, water-base fluids are the most
are to open the fracture and to transport propping widely used fracturing fluids. Many water-soluble
agent along the length of the fracture. Consequently, polymers can be used to make a viscosified solution
the viscous properties of the fluid are usually con- capable of suspending proppants at ambient temper-
sidered the most important. However, successful ature. However, as the temperature increases, these
hydraulic fracturing treatments require that the flu- solutions thin significantly. The polymer concentra-
ids have other special properties. In addition to tion (polymer loading) can be increased to offset
exhibiting the proper viscosity in the fracture, they thermal effects, but this approach is expensive.
should break and clean up rapidly once the treat- Instead, crosslinking agents are used to significantly
ment is over, provide good fluid-loss control, exhibit increase the effective molecular weight of the poly-
low friction pressure during pumping and be as eco- mer, thereby increasing the viscosity of the solution
nomical as is practical. Characterization of these (Fig. 7-1). The specific chemistry and performance
performance properties is addressed in Chapter 8. of crosslinkers are discussed in more detail in
Because reservoirs to be stimulated vary markedly Section 7-6.
in terms of temperature, permeability, rock compo-
sition and pore pressure, many different types of 800
fluids have been developed to provide the properties Noncrosslinked HPG (40 lbm/1000 gal)
700 Noncrosslinked HPG (60 lbm/1000 gal)
described. The first fracturing fluids were oil-base; Borate-crosslinked HPG (30 lbm/1000 gal)
Viscosity (cp at 170 s–1)
in the late 1950s, water-base fluids thickened with 600

guar became increasingly popular. In 1969, the first 500
crosslinked guar treatment was performed. By this 400
time, only about 10% of fracturing treatments were
300
conducted with gelled oil. Currently, more than 65%
of all fracturing treatments use water-base gels vis- 200
cosified with guar or hydroxypropylguar. Gelled oil 100
treatments and acid fracturing treatments each
0
account for about 5% of the total. About 20%–25% 50 75 100 125 150 175 200 225
of all treatments contain an energizing gas. Addi- Temperature (°F)
tives are also used to enhance viscosity at high tem-
peratures, to break viscosity at low temperatures or Figure 7-1. Effect of temperature and crosslinker on the
to help control leakoff of the fluid to the formation. viscosity of hydroxypropylguar solutions.
This chapter describes the chemistry of commonly
used fracturing fluids and additives. In addition, it One of the first polymers used to viscosify water
discusses how the chemistry is practiced at the well- for fracturing applications was guar gum. Guar is a
site. long-chain, high-molecular-weight polymer com-
posed of mannose and galactose sugars (Whistler,
1959). Polymers composed of sugar units are called
polysaccharides. Guar gum comes from the

endosperm of guar beans, which are grown mainly in

Pakistan and India. The beans are removed from the
bean pod and processed to separate the endosperm
from the bean hull and embryo (splits), and the splits
are ground into a powder (Fig. 7-2). The guar poly-
mer has a high affinity for water. When the powder
is added to water, the guar particles swell and
hydrate, which means the polymer molecules
become associated with many water molecules and
unfold and extend out into the solution. The guar
solution on the molecular level can be pictured as
bloated strands suspended in water. The strands tend
to overlap and hinder motion, which elevates the
viscosity of the solution.
The structure of the guar molecule is usually rep- Figure 7-2. Guar pods, beans, splits and powder.
resented as in Fig. 7-3. For a number of years, it was
thought that guar consisted of a mannose backbone
with galactose side chains on every other mannose reaction changes some of the OH sites to –O–CH2–
unit (one galactose unit to two mannose units). The CHOH–CH3, effectively removing some of the
galactose and mannose sugars differ in the orienta- crosslinking sites. The additional processing and
tion of the OH groups on the ring. Recent studies washing removes much of the plant material from
indicate that the arrangement of galactose units may the polymer, so HPG typically contains only about
be more random, with galactose appearing on two 2% to 4% insoluble residue. HPG was once consid-
or three consecutive mannose units (Guar and ered less damaging to the formation face and prop-
Derivatives, 1986). Also, the ratio of mannose to pant pack than guar, but recent studies (Almond et
galactose may range from 1.6:1 to 1.8:1, instead al., 1984; Brannon and Pulsinelli, 1992) indicate that
of 2:1 as indicated in Fig. 7-3. guar and HPG cause about the same degree of pack
The process used to produce guar powder does not damage.
completely separate the guar from other plant materi- Hydroxypropyl substitution makes HPG more sta-
als, which are not soluble in water. As much as 6% ble at an elevated temperatures than guar; therefore,
to 10% insoluble residue can be present in guar flu- HPG is better suited for use in high-temperature
ids. Guar can be derivatized with propylene oxide to (>300°F [150°C]) wells. The addition of the less
produce hydroxypropylguar (HPG) (Fig. 7-4). The hydrophilic hydroxypropyl substituents also makes
Galactose
CH2 OH CH2 OH substituents
O H
HO H
Mannose
H OH H
backbone
OH H OH
O O
CH2 CH2OH CH2 CH2

O O O O
H H H H H H H H
O O O O
O OH HO H OH HO H OH HO H OH HO H
H H H H H H H H
n
Figure 7-3. Structure of guar.
7-2 Fracturing Fluid Chemistry and Proppants

CH2 OR CH2 OR CH2 OH
O O O
HO H H H H H H
O
O
H H H
OH OH OH
H OH H OH H OR
O
Figure 7-5. Repeating-unit structure of hydroxyethylcellu-
CH2 OR CH2 lose, R–CH2CH2OH.
O O
composed of glucose sugar units. Although similar
H
to the mannose backbone of guar, there is a signifi-
O O
cant difference. Guar contains hydroxyl pairs that are
H
OH HO OH RO positioned on the same side of the sugar molecule
(cis orientation). In HEC, the OH groups are on adja-
H H H H cent carbons, but they are on opposite sides of the
ring (trans orientation). Because of their close prox-
imity, the cis arrangement for guar is easily cross-
Figure 7-4. Repeating-unit structure of hydroxypropylguar, linked, whereas the increased separation of the trans
R–CH2–CHOH–CH3.
arrangement makes HEC more difficult to crosslink.
However, HEC can be crosslinked at a pH of 10 to
the HPG more soluble in alcohol. A common quality 12 with Zr(IV) (Underdown et al., 1984) or with lan-
assurance check is to add an equal volume of meth- thanides (Dovan and Hutchins, 1993). To crosslink
anol to the polymer solution. Guar precipitates, while HEC under milder conditions, the carboxymethyl
HPG with the standard level of hydroxypropyl sub- group can be added to make carboxymethylhydroxy-
stitution does not (Ely, 1985). HPG containing less ethylcellulose (CMHEC), which makes crosslinking
hydroxypropyl substitution than the standard gener- with metal ions such as Al(III), Ti(IV) and Zr(IV)
ally fails the test. possible at a pH of approximately 4 to 6.
Another guar derivative used in recent years is Still another type of polymer is xanthan gum (Fig.
carboxymethylhydroxypropylguar (CMHPG). This 7-6). Xanthan is a biopolymer, produced metaboli-
“double-derivatized” guar contains the hydroypropyl cally by the microorganism Xanthomonas campestris
functionality of HPG as well as a carboxylic acid (Lipton and Burnett, 1976). Xanthan solutions
substituent. CMHPG was first used for low-tempera- behave as power law fluids even at low shear rates
ture wells (Almond and Garvin, 1984). For these (Kirkby and Rockefeller, 1985), whereas HPG solu-
applications, it is usually crosslinked with Al(III) tions become Newtonian. Clark et al. (1985) showed
through the carboxy groups. This provides a less that at shear rates less than 10 s–1 the low-shear
expensive fluid than HPG crosslinked with Ti and properties enable xanthan solutions to suspend sand
Zr complexes. More recently, CMHPG has been better than HPG. These properties may increase the
crosslinked with Zr crosslinker to produce fluids future use of xanthan for fracturing, but currently
with higher viscosity at high temperatures than those xanthan is more expensive than guar or cellulose
made with comparable amounts of HPG (Hunter and derivatives, and it is used less frequently. Davies
Walker, 1991). et al. (1991) reported using a different biopolymer,
Cellulose derivatives have occasionally been used scleroglucan, because of its near-perfect proppant
in fracturing fluids (Carico and Bagshaw, 1978). suspension and because it does not require a breaker.
Hydroxyethylcellulose (HEC) (Fig. 7-5) or hydroxy- Partially hydrolyzed acrylamide polymers are used
propylcellulose (HPC) is used when a very clean as friction-reducing agents. These polymers can be
fluid is desired. These polymers have a backbone used at low loading (less than 10 lbm/1000 gal) to

CH2 OH CH2 OH
O O
H H H H
O O
OH H
H OH H OH n
O
O
=
CH2 OCCH3
O
H H H
– +
HO OH
COO M
O O H
H H
– +
COO M OH H
O CH2
C O
O H OH +
H M = Na, K, 1⁄2Ca
CH3
O OH HO
H H
Figure 7-6. Repeating-unit structure of xanthan gum.
reduce the horsepower required to pump water at insulated from the aqueous environment. When the
high rates. aqueous environment contains an optimum concen-
Acrylamide copolymers are used to viscosify acid- tration of salts (usually potassium or ammonium
based fracturing fluids. Acrylamide homopolymers chloride [KCl or NH4Cl] solutions), the micelles
hydrolyze in aqueous solution to produce acrylate assume a rodlike shape. If the surfactant is present in
groups, which cause the polymer to precipitate in the a sufficient concentration (usually >1% by volume)
presence of Ca2+ ions. Because high levels of Ca2+ the micelles associate with one another. The result-
are found in spent acid after fracturing a limestone ing hindered movement causes the fluid to become
formation, Ca2+-sensitive acrylamide homopolymer both viscous and elastic. These associations are elec-
should not be used. To improve performance at ele- trostatic in character; therefore, VES fluids are not as
vated temperatures, an acrylamide copolymer is syn- sensitive to shear history as polymer-base fluids. If
thesized using a monomer with functional groups the micelles are disrupted owing to shear, they will
that protect the acrylamide group from hydrolysis. quickly reaggregate and recover when shear ceases.
Polymer-free, water-base fracturing fluids can be Like for polymer-base fluids, the performance of
prepared using viscoelastic surfactants (VES) VES fluids is sensitive to temperature; therefore, the
(Stewart et al., 1995). These surfactants (typically a surfactant concentration (and in some cases, the salt
quaternary ammonium salt of a long-chain fatty acid; concentration) must be adjusted accordingly.
Fig. 7-7) consist of two regions: the head group is The micellar structure of VES fluids is permanent-
the quaternary ammonium portion of the molecule ly disrupted by two mechanisms: contact with hydro-
and the tail group is the long-chain hydrocarbon por- carbons and dilution by aqueous fluids such as for-
tion of the molecule. The head group is hydrophilic, mation water. In both cases, the viscosity of the VES
meaning that it prefers to be in contact with water. fluid falls greatly (Brown et al., 1996; Samuel et al.,
The tail group is hydrophobic, meaning that it pre- 1997). Because one or both scenarios normally occur
fers to be in contact with oil. When the surfactant is during postfracture production, no additional breaker
added to water, the molecules associate into struc- chemicals are required. The principal advantage of
tures called micelles (Fig. 7-8). VES fluids is that, unlike polymer-viscosified fluids,
In a micelle, the hydrophilic head groups are on little residue is left after cleanup. As a result, less
the outside, in direct contact with the water phase. damage to the proppant pack and fracture face is
The hydrophobic tail groups form an inner core, observed. The typical retained permeability of prop-

H H Cl–
CH2–CH2–OH
C C
CH 3 –( C H 2 ) 7 (CH2)11–CH2–N+ –CH3
CH2–CH2–OH
Figure 7-7. Molecular and structural formulas for a viscoelastic surfactant thickener.
Figure 7-8. Micellar association.
pant packs treated with VES fluid systems is >95%.

VES systems can also be foamed with nitrogen. No
additional foaming agents are required.

7-3. Oil-base fluids of different organic compounds and may contain

paraffins and asphaltenes. Some high-molecular-
Heavy oils were used originally as fracturing fluids, weight compounds, especially paraffins and asphalt-
primarily because these fluids were perceived as less enes, are not compatible with the aluminum phos-
damaging to a hydrocarbon-bearing formation than phate ester gelling system. Many crude oils may be
water-base fluids. Their inherent viscosity also makes gelled, but it is good practice to test them prior to
them more attractive than water (Howard and Fast, attempting to gel on location.
1970). Oil-base fluids are expensive to use and oper- The R groups can be pictured as forming an oil-
ationally difficult to handle. Therefore, they are now compatible shield around the polar core of aluminum
used only in formations that are known to be ions (McKenzie, 1980). Polar species (such as water,
extremely water-sensitive. acids, bases or salts) are incorporated into the polar
In the 1960s, the industry used aluminum salts of core and affect the association of the aluminum ions
carboxylic acids (e.g., aluminum octoate) to raise the and phosphate ester groups. These materials can
viscosity of hydrocarbon fracturing fluids (Burnham make the gel structure more rigid, or they can
et al., 1980). This improved the temperature stability destroy the gel structure.
and proppant-carrying capability of the fluids. In the The viscosity of the standard aluminum phosphate
1970s, the aluminum carboxylate salts were replaced ester gel is controlled by varying the quantities of
by aluminum phosphate ester salts. Again, the tem- aluminum compound and phosphate ester. To improve
perature range of the fluids was extended and prop- high-temperature performance, the viscosity of the
pant transport was enhanced. Today, aluminum phos- gel can be increased by increasing the amount of
phate ester chemistry remains the preferred method polymer; however, this results in very high viscosities
of gelling hydrocarbons for fracturing purposes. on the surface, which make it difficult to draw the
Both methods of thickening oil rely on an “associa- fluid out of the tanks to the pumps. One approach
tive” mechanism (Baker et al., 1970). As suggested used is to add part of the gelling materials “on the
in Fig. 7-9, interactions between the aluminum com- fly” so that high viscosity is not achieved until the
plexes and phosphate ester molecules produce a long fluid reaches the fracture (Harris et al., 1986; Cramer
polymer chain (Burnham et al., 1980). et al., 1991). On-the-fly addition means that the mate-
The R groups shown in Fig. 7-9 are hydrocarbon rials are added to the fluid as the fluid is pumped
chains that must be soluble in the oil to be gelled. downhole. Another approach is to maximize thermal
The soluble R groups keep the aluminum phosphate stability by carefully controlling the composition of
ester polymer in solution. Generally, the R groups the solution to provide optimum conditions for asso-
are hydrocarbon chains containing 1 to 18 carbon ciation of the aluminum and phosphate ester species
atoms (Crawford et al., 1973). The R groups have a (Gross, 1993).
high affinity for oils such as kerosene and diesel that Typically, these gels take several hours to form
comprise 12- to 18-carbon (and somewhat higher) once the chemicals are mixed together. Recent devel-
chains. Crude oils are composed of a larger number opments in gelled oil chemistry make a true continu-
ous-mix (all materials added on the fly) gelled oil
R R R R possible. By changing the aluminum source, the alu-
O O O O
minum/phosphate ester ratio in the gel and/or the
phosphate ester mix (Daccord et al., 1985; McCabe
P P
et al., 1990; Huddleston, 1992), a rapidly thickening
O H O O H O gel composition can be achieved. With this chemistry,
Al O Al O Al the aluminum source and phosphate ester can be
O O O O added to the hydrocarbon as it is pumped downhole.
P P The gel is formed on the way to the perforations. The
O O O O expense of premixing the gel is eliminated, as well as
the disposal problem of unused gel.
H R H R
Figure 7-9. Proposed structure of the aluminum phos-

phate ester polymer chain (Burnham et al., 1980).

7-4. Acid-based fluids initiate the fracture and to deposit a filter cake that
acts as a barrier to acid leakoff. The ability of a sin-
Acid fracturing is a well stimulation process in gle viscous pad fluid to control fluid loss is question-
which acid, usually hydrochloric acid (HCl), is able. Studies by Nierode and Kruk (1973), Coulter et
injected into a carbonate formation at a pressure suf- al. (1976) and Crowe et al. (1989) show that the fil-
ficient to fracture the formation or to open existing ter cake deposited by the pad is quickly penetrated
natural fractures. As the acid flows along the frac- by wormholes resulting from acid leakoff. Once this
ture, portions of the fracture face are dissolved. occurs, the acid fluid loss is identical to that occur-
Because flowing acid tends to etch in a nonuniform ring if no pad were used. In recent years, multiple
manner, conductive channels are created that usually stages of viscous pad have been used to control acid
remain when the fracture closes. The effective length fluid loss (Coulter et al., 1976). In this widely used
of the fracture is determined by the etched length, technique, the fracture is initially created by a gelled
which depends on the volume of acid used, its reac- pad, after which alternating stages of acid and addi-
tion rate and the acid fluid loss from the fracture into tional polymer pad are pumped. These additional pad
the formation. The effectiveness of the acid fractur- stages are designed to enter and seal wormholes cre-
ing treatment is determined largely by the length of ated by the preceding acid. Each alternating pad
the etched fracture. stage in this treatment is usually equal to or larger
In some cases, especially in carbonates, a choice than the acid stage that preceded it.
exists between acid and propped fracturing treat- In addition to fluid-loss additives, two-phase fluids
ments. Operationally, acid fracturing is less compli- (foams and emulsions) have been shown to effect-
cated because no propping agent is employed. Also, ively control fluid loss during acid fracturing treat-
the danger of proppant screenout and the problems of ments. Nierode and Kruk (1973) presented data
proppant flowback and cleanout from the wellbore showing that an acid external emulsion, consisting
after the treatment are eliminated. However, acid is of an oil inner phase with gelled acid as the outer
more expensive than most nonreactive treating fluids. phase, provides good fluid-loss control. The use of
The major barrier to effective fracture penetration these acid external emulsions in well stimulation has
by acid appears to be excessive fluid loss (Nierode been rather limited. The use of foamed acid is one of
and Kruk, 1973). Fluid loss is a greater problem the most effective methods for controlling acid fluid
when using acid than when using a nonreactive fluid. loss. Scherubel and Crowe (1978) and Ford (1981)
The constant erosion of fracture faces during treat- showed that foamed acids provide excellent fluid-loss
ment makes it difficult to deposit an effective filter- control. Fluid-loss control is further enhanced by the
cake barrier. In addition, acid leakoff is extremely use of a viscous pad preceding the foamed acid.
nonuniform and results in wormholes and the However, foaming the acid reduces the effective
enlargement of natural fractures. This greatly amount of acid available for etching because there
increases the effective area from which leakoff is less acid present per unit volume injected. As a
occurs and makes fluid-loss control difficult. result, a high acid concentration (e.g., 28% HCl)
should be used in preparing foamed acid to maximize
the amount of acid available for fracture etching.
7-4.1. Materials and techniques for acid Acid fluid loss can also be reduced by gelling
fluid-loss control the acid. This method of control has become widely
Various additives and treating techniques have been used since the development of more acid-stable
developed to control acid fluid loss. Among these are thickening agents. Commonly, thickeners include
particulates (oil-soluble resins and 100-mesh sand) xanthan biopolymers, various acrylamide copoly-
and gelling agents. In general, acid fluid-loss addi- mers and certain surfactants that thicken acid by
tives have not been used extensively because of per- micellar association.
formance and cost limitations. As a result, alternate A gelling agent must be sufficiently stable to allow
methods of fluid-loss control usually are employed. the gelled acid to retain its viscosity at the treating
The most common technique involves the use of a temperature. However, slow cleanup or actual plug-
viscous pad preceding the acid. The pad is used to ging of the well may result if the viscosity of the

enhanced by incorporating a second phase into the dorff and Ainley, 1981). The thicker the continuous
fluid. Foams are created by adding gas to the fluid. phase, the more difficult it is for the gas bubbles to
Emulsions are created by mixing oil and water move together and coalesce. Guar, HPG and xanthan
together. The different systems are described in this gum have been used as stabilizers. Still, a relatively
section. high quality, although not as high as 52%, is required
to maintain dispersion of the gas phase.
A further improvement in foam stability can be
7-5.1. Foams achieved by crosslinking the polymer in the aqueous
A foam is a stable mixture of liquid and gas. To phase (Watkins et al., 1983). The liquid phase then
make the mixture stable, a surface-active agent (sur- becomes viscous enough to maintain dispersion of
factant) is used. The surfactant concentrates at the the gas bubbles, even at foam quality less than 40%.
gas/liquid interface and lowers the interfacial ten- Thickening the liquid phase also improves foam rhe-
sion. The surfactant stabilizes thin liquid films and ology and fluid-loss control. Proppant concentrations
prevents the cells from coalescing. in the foamed fluid are generally lower than the con-
Pressurized gas (nitrogen or carbon dioxide) in centration achieved with single-phase, liquid treat-
a foam expands when the well is flowed back and ments. Therefore, a larger volume of foam may be
forces liquid out of the fracture. Foams accelerate required to place the desired amount of proppant.
the recovery of liquid from a propped fracture and Nitrogen and carbon dioxide are used as energizing
thus are excellent fluids to use in low-pressure reser- gases. N2 is less dense than CO2. CO2 creates a
voirs. Also, the liquid phase is minimal because denser foam and, consequently, lower surface treating
foams contain up to 95% by volume gas. In the case pressures because of the increased hydrostatic head in
of a water-base fluid, foaming the fluid significantly the wellbore. Lower treating pressures reduce pump-
decreases the amount of liquid in contact with the ing costs. On the other hand, because CO2 is much
formation. Therefore, foams perform well in water- more soluble in oil and water than N2, it takes more
sensitive formations (Ward, 1984; Ainley, 1983). CO2 to saturate the liquid and to create the foam.
Foams yield pseudoplastic fluids with good transport Reductions in pumping costs may be offset by
properties (King, 1982; Reidenbach et al., 1986). increases in material costs.
They provide good fluid-loss control in low-perme-
ability formations where the gas bubbles are approxi-
mately the size of the rock pore openings (Harris,
7-5.2. Emulsions
1985). An emulsion is a dispersion of two immiscible
Foams are described by their quality: phases such as oil in water or water in oil stabilized
gas volume with a surfactant. Emulsion-based fracturing fluids
foam quality = ×100 . (7-1) are highly viscous solutions with good transport
foam volume
properties. The higher the percentage of the internal
Originally, foam quality was considered to range phase, the more resistance there is to droplet move-
from 52% to 95%. Above 95%, the foam usually ment, resulting in a higher viscosity.
changes to a mist, with gas as the continuous phase. Emulsion-based fracturing fluids have been used
Below 52%, a stable foam does not exist because for a number of years (Kiel, 1971). The most com-
there are no bubble/bubble interactions to provide mon fluid, termed polyemulsion, is composed of
resistance to flow or to gravity separation (Mitchell, 67% hydrocarbon internal phase, 33% viscosified
1969). Above 52% gas, the gas concentration is high brine external phase and an emulsifying surfactant.
enough that the bubble surfaces touch. Viscosifying the aqueous phase improves the emul-
Stable dispersions of gas in liquid can be prepared sion stability and significantly reduces friction pres-
with qualities less than 52% (Watkins et al., 1983). It sure during pumping because the polymer acts as a
may not be appropriate to call them foams, but they friction reducer. The polymer concentration used is
can be used effectively as energized fluids. Viscos- generally 20 to 40 lbm/1000 gal, so the fluid con-
ifying the liquid phase with a polymer is an effective tains only one-sixth to one-third as much polymer as
method for increasing the stability of foams (Wen- a standard water-base fracturing fluid. The emulsion

usually breaks because of adsorption of the emulsi- molecules overlap, the complex in Fig. 7-10a can
fier onto the formation rock; because so little poly- react with other polymer strands to form a crosslinked
mer is used, this type of fluid is known for causing network (Menjivar, 1984) illustrated in Fig. 7-10b. A
less formation damage and cleaning up rapidly species is created with 2 times the molecular weight
(Roodhart et al., 1986). of the polymer alone. Because each polymer chain
Disadvantages of polyemulsions are high friction contains many cis-hydroxyls, the polymer can be
pressure and high fluid cost (unless the hydrocarbon crosslinked at more than one site. Networks with
is recovered). Polyemulsions also thin significantly a very high molecular weight develop, especially under
as the temperature increases, which limits their use static conditions, resulting in highly viscous solutions.
in hot wells. One of the simplest crosslinkers, the borate ion, is
used to produce very viscous gels with guar and HPG
that can be stable above 300°F. At a pH above 8,
7-6. Additives borate ions and guar form an extremely viscous gel
A fracturing fluid is generally not simply a liquid in a matter of seconds. To maximize the thermal
and viscosifying material, such as water and HPG stability of the crosslinked gel, the pH and borate
polymer or diesel oil and aluminum phosphate ester concentration must be increased, with an optimum
polymer. Various additives are used to break the fluid pH of 10 to 12 depending on the borate compound
once the job is over, control fluid loss, minimize for- and borate ion concentration (Harris, 1993). The
mation damage, adjust pH, control bacteria or borate ion B(OH)4– is believed to be the crosslinking
improve high-temperature stability. Care must be species. Regardless of the borate source (boric acid,
taken when using multiple additives to determine borate salt or borate complex), a high pH is required
that one additive does not interfere with the function to shift the equilibrium and maintain an adequate
of another additive. concentration of borate ions (Prud’homme, 1991):
H3BO3 + OH– → B(OH)4–
7-6.1. Crosslinkers The fraction of boric acid present at ambient tem-
A number of metal ions can be used to crosslink perature as the effective crosslinking compound,
water-soluble polymers (Conway et al., 1980). B(OH)4–, is shown in Fig. 7-11 as a function of pH.
Borate, Ti(IV), Zr(IV) and Al(III) compounds are As illustrated, increasing the pH results in a higher
frequently used crosslinkers. The borate compounds concentration of B(OH)4–. Increasing the tempera-
(Deuel and Neukorn, 1949) and transition metal com- ture reduces the pH, resulting in a lower crosslinker
plexes (Chrisp, 1967) react with guar and HPG concentration and lower viscosity. Attempting to
through cis–OH pairs on the galactose side chains compensate for the detrimental effects of tempera-
to form a complex, as illustrated in Fig. 7-10a. As the ture by increasing the H3BO3 concentration can
cause syneresis (overcrosslinking) of the gel.
H H
O O O O
H H
O O O O
H H
HO OH O O HO OH O O
n n
B B
HO OH O O
(a) (b)
H
OH
O
Figure 7-10. Proposed crosslinking mechanism (Menjivar, 1984).

Regardless of the gel composition or viscosity, all

Effective crosslinker concentration 100 fracturing gels thin with shear and heat. However,
some gels return to their original state once shear
80 or heat is removed. Typically, borate crosslinking is
reversible; crosslinks form and then break, only to
60
(%B(OH)4)
form again (Deuel and Neukorn, 1949). If the poly-

mer is not thermally degraded, this reversible behav-
40
ior continues to accommodate changes in shear rate
20 or temperature. The transition metal–polymer bond
is sensitive to shear. High shear irreversibly degrades
0 transition metal–crosslinked fluids (Craigie, 1983).
6 7 8 9 10 11 12 Unlike borate crosslinker, once the bond between the
pH at ambient temperature transition metal crosslinker and polymer is broken, it
does not reform. Therefore, if the crosslinking rate is
Figure 7-11. Borate in concentration as a function of pH. very rapid in the high shear region of the tubing, an
irreversible loss of viscosity occurs. The effect that
Transition metal crosslinkers have been developed tubing shear has on fluid viscosity is illustrated in
for high-temperature applications and/or low-pH Fig. 7-12. A fluid that is crosslinked rapidly under
environments (i.e., CO2-energized fluids). Titanium unrealistic conditions of low shear is very viscous at
and zirconium complexes have been used most fre- high temperatures (curve A). The same fluid cross-
quently because of their affinity for reacting with linked at high-shear-simulating conditions in the tub-
oxygen functionalities (cis-OH and carboxyl groups), ing loses much of its viscosity because of shear de-
stable +4 oxidation states (Cotton and Wilkinson, gradation and behaves like curve C. Other character-
1972) and low toxicity. The bond formed between istics of commonly used crosslinkers are compared
the titanium or zirconium complex and the polymer in Table 7-1.
is thermally stable. The upper temperature limit for
these gels is 350° to 400°F [150° to 200°C]. It 500
Fast crosslinker, low-shear mixing
appears that the stability of the polymer backbone, Slow crosslinker, high-shear mixing
Fast crosslinker, high-shear mixing
rather than the polymer–metal ion bond, is the limit- 400
ing factor. Very hot wells (>400°F) can be fractured A

with these fluids if the treatments are designed to pro- 300
vide adequate cooldown by injecting sacrificial fluid
immediately before the fracturing treatment. How- 200
ever, it must be emphasized that rudimentary heat B
transfer calculations suggest that the cooldown of 100 C
a formation is only moderate and confined near the
well unless fluid leakoff is substantial. The vast 0
0 2 4 6 8
majority of the fracturing fluid is likely to be exposed
Time (hr) at 250°F
to the static reservoir temperature.
Figure 7-12. Effect of shear and crosslinking rate on
viscosity.
Table 7-1. Characteristics of commonly used crosslinkers
Crosslinker Borate Titanate Zirconate Aluminum

Crosslinkable polymers Guar, HPG, Guar, HPG, Guar,‡ HPG,‡ CMHPG, CMHEC
CMHPG CMHPG, CMHEC† CMHPG, CMHEC†
pH range 8–12 3–11 3–11 3–5
Upper temperature limit (°F) 325 325 400 150
Shear degraded No Yes Yes Yes
† Low-pH (3–5) crosslinking only
‡ High-pH (7–10) crosslinking only

To avoid the detrimental effects of high shear for the metal ion between the polymer and other
occurring in the tubing, the crosslinking rate is usu- ligands. A hypothetical titanium complex with two
ally reduced to limit viscosity development until the ligands (L) capable of binding at two sites (biden-
fluid is near the target zone. In addition to minimiz- tate) and two ligands (A) capable of binding at one
ing shear degradation, delaying crosslinking also site (monodentate) is illustrated in Fig. 7-13a. On
reduces frictional pressure losses and, therefore, addition to water, complexes of titanium and
hydraulic horsepower requirements. A number of zirconium form colloidal particles (Fig. 7-13b)
factors can be manipulated to control the rate of (Prud’homme et al., 1989). For crosslinking to
crosslinking. These include fluid temperature and occur, polymer molecules must displace the organic
pH, shear conditions, crosslinker type and the pres- compounds at the coordination sites on the surface
ence of other organic compounds that react with the of the colloidal particles. If the ligands are easy to
crosslinker. For example, increasing the temperature displace, crosslinking occurs rapidly. If the organic
or pH usually accelerates the crosslinking reaction. compounds are difficult to displace or are present in
Fortunately, some of these parameters can be con- a high concentration, crosslinking occurs slowly
trolled to slow down the crosslinking reaction so that (Payne and Harms, 1984; Rummo, 1982; Hodge,
it does not occur in the high-shear region (generally 1988a, 1988b). Forming slow-reacting complexes to
500 to 1500 s–1) of the tubing, while allowing the control the crosslinking rate can also be used with
bulk of the crosslink reaction to occur in the low- borate ions to minimize friction pressure and greatly
shear region (generally 10 to 200 s–1) of the fracture. improve thermal stability (Dawson, 1991).
By manipulating the chemistry, shear degradation and Slowly dissolving crosslinkers and activators can
frictional pressure loss can be minimized. The effect be used to delay crosslinking. As the crosslinking
of reducing the crosslinking rate on viscosity is illus- ion or activator concentration increases, the crosslink
trated by comparing curves B (delayed crosslinking) density increases, producing a rise in viscosity. Some
and C (rapid crosslinking) in Fig. 7-12. The effects borate compounds, such as colemanite and ulexite,
of high shear can be reduced, but not eliminated, by dissolve slowly in water, producing a controllable
slowing the crosslinking rate (i.e., curve B does not crosslinking rate and delaying viscosity development
reach the viscosity values of curve A). (Mondshine, 1986; Tan et al., 1992). Also, slowly
A number of techniques can be used to control dissolving bases or acids can be used to delay the
the reaction rate of the metal ion and polymer. For crosslinking rate of pH-dependent crosslinkers. The
example, many different organic molecules (ligands) pH of the fluid containing the desired concentration
are capable of reacting with the metal ion (cross- of crosslinker begins at a value that does not initiate
linker), which can strongly affect the properties of significant crosslinking. The base or acid dissolves at
the ion. Crosslinkers can be delayed by competition a controlled rate, producing the desired pH change
L Ti
O
O–
A Ti
H+
O
Ti Ti OH
A O
R = Sugar
HOR
Ti R = Ti
L OR
(a) (b) O
Ti
L, A = Complexing agents
Figure 7-13. Hypothetical titanium complex (a) hydrolyzed to a colloidal titanium dioxide particle and (b) providing polymer
crosslinking on the particle surface.

and initiating crosslinking. This method can be used generally not recommended. Factors to consider in
to control the crosslinking rate of borate ions by optimizing crosslinker performance are reviewed in
using a rapidly dissolving borate compound, such as Sidebar 7A.
boric acid, and a slow-dissolving base, such as CaO
or MgO.
When using delayed crosslinkers, viscosity should 7A. Ensuring optimum crosslinker performance
be building before the fluid reaches the producing
The composition of crosslinked fluids is carefully optimized
interval, although complete crosslinking is not neces- to obtain the desired performance (rheology and proppant
sary at that time. If complete crosslinking occurs too transport, thermal stability, crosslinking rate, cleanup, etc.).
Many factors influencing the performance must be consid-
soon, high friction pressure and shear degradation ered during the selection of a fracturing fluid candidate and
may result. If crosslinking occurs too slowly, prop- monitored during on-site preparation. To obtain the best pos-
sible performance from a crosslinked fluid, the following
pant may settle in the wellbore or in the fracture near issues must be addressed.
the wellbore, resulting in poor proppant transport
and potential screenout (proppant blocking fluid pas- Crosslinker concentration
sage in the fracture). Therefore, considerable effort Each fluid composition has an optimum range for crosslinker
concentration dictated by the type of polymer, polymer con-
is spent on location to produce a fluid composition centration and fluid pH. If the crosslinker concentration is too
with the desired crosslink time. Unfortunately, as low, the crosslinking rate will be slower and the viscosity
development will be lower than anticipated. If the crosslinker
pointed out by Baranet and Ainley (1985) and concentration exceeds the optimum range, the crosslinking
Hodge and Baranet (1987), commonly used field rate will be faster than anticipated and the final viscosity may
methods for determining crosslink time may not be be much lower because of syneresis. Syneresis is the pre-
cipitation of the polymer from solution caused by the collapse
reliable. Typically, these methods do not simulate of the polymer network. In the most severe cases, “free
the shear conditions in the tubing during a fracturing water” may be observed at ambient sampling conditions.
However, detection of syneresis in delayed crosslinking fluids
treatment and produce a fluid composition that usually requires heating the fluid to the anticipated downhole
crosslinks too slowly. To avoid the problems associ- temperature to fully react all the crosslinker.
ated with overdelayed crosslinked fluids, crosslink pH control
times of one-half to three-fourths of the tubing resi- All crosslinked fluids have a specific pH range for optimum
dence time may be recommended (Cawiezel and performance. If this pH is not maintained, the desired
Elbel, 1990; Aud et al., 1994) or dual-crosslinker crosslinking rate and thermal stability cannot be obtained. To
minimize pH variation in crosslinked fluid, buffers are includ-
systems may be used (Baranet and Ainley, 1985; ed with the crosslinked fluid. However, gross contamination
Royce et al., 1984; Hodge and Baranet, 1987). Dual- of the fracturing fluid can overwhelm these buffers and com-
promise crosslinked-fluid performance. Proper pH control is
crosslinker systems combine a fast crosslinker to critical to crosslinked-fluid performance and must be moni-
ensure adequate viscosity at the perforations and a tored diligently.
slow crosslinker accelerated by heating in the frac-
Chemical contamination
ture to produce a viscous, temperature-stable fluid.
A variety of common compounds and oilfield products can
There are many benefits from using a delayed- interfere with the performance of crosslinker compounds.
crosslinking fluid. Delayed crosslinkers produce flu- Typically, these contaminants reduce or eliminate crosslink-
ing and produce a fluid with a slow rate of viscosity develop-
ids with better long-term stability at elevated temper- ment (extremely long crosslink time) and significantly lower
atures. In some areas, this has allowed reducing poly- viscosity. In the most obvious cases, no viscosity increase
may be produced by the crosslinker. Among the “naturally”
mer loadings. Also, reducing the friction pressure occurring compounds that can be present in mix water are
allows higher injection rates and lower horsepower bicarbonate, phosphate and silicates. The source of many
requirements. However, delayed crosslinking intro- of these contaminants is the mix water used to prepare the
fracturing fluid. In addition to naturally occurring contami-
duces some risk of near-wellbore screenout associated nants, many surfactants, clay stabilizers and foaming agents
with overdelaying the crosslinking rate. Little or no can interfere with crosslinked-fluid performance. To avoid
contamination with incompatible additives, fracture tanks and
benefit from avoiding shear degradation is realized mixing equipment should be empty and clean before the mix
for treatments with low wellbore shear rates (<300 s–1) water is loaded and fracturing fluid prepared. Furthermore,
substitution of additives should not be approved without com-
or short wellbore residence times (time to perfora- patibility testing (typically, rheology testing at the anticipated
tions < 60 s). In these low-shear and/or shallow jobs, fluid temperature).
delayed crosslinking to avoid shear degradation is

7-6.2. Breakers tion occurs more rapidly. With increased tempera-

ture, the breaker becomes too reactive. As little as
Relatively high viscosity fluids are used to transport 0.1 lbm/1000 gal of dissolved persulfate causes rapid
proppant into the fracture. Leaving a high-viscosity viscosity loss above 180°F [80°C]. At high tempera-
fluid in the fracture would reduce the permeability of tures, the reaction of dissolved persulfate with the
the proppant pack to oil and gas, limiting the effec- polymer generally occurs during pumping rather
tiveness of the fracturing treatment (Penny, 1987; than after placement of the proppant and closure of
Brannon and Pulsinelli, 1992). Gel breakers are used the fracture. This undesirable high reactivity of dis-
to reduce the viscosity of the fluid intermingled with solved persulfate at elevated temperatures was a
the proppant. Breakers reduce viscosity by cleaving major limitation of these breakers. The introduction
the polymer into small-molecular-weight fragments. of encapsulated breakers, which is discussed later,
It has been estimated that fluid loss during the treat- greatly improved persulfate performance at elevated
ment and during closure increases the polymer con- temperatures.
centration in the fracture after closure 5–7 times Certain enzyme breakers of the class hemicellulase
(Penny, 1987) to as much as 20 times (Hawkins, are also used to reduce the viscosity of water-base
1988) higher than the surface concentration. The fluids. Enzymes have been in use for some time, but
increased polymer concentration causes a major prior to 1994 their use was thought to be limited to a
increase in viscosity. For example, the viscosity relatively mild environment: pH range of about 3.5
of an unbroken guar fluid containing polymer at to 8 and temperatures less than about 150°F. Because
400 lbm/100 gal (40 lbm/1000 gal gel concentrated they are active at ambient temperature, enzymes
10 times because of fluid loss on fracture closure) begin to degrade the polymer immediately upon
has been estimated to be in excess of 1000 poise mixing and under some conditions can be too reac-
(Pope et al., 1994 ). Significant effort has gone into tive, like persulfates.
designing breakers to address this problem. Ideally, Recent work in the area of conductivity improve-
a gel breaker put into the fluid at the surface should ment and breakers has led to a broadening of appli-
have minimal effect on the gel until pumping ceases cations for both oxidizers and enzymes. Before the
(and the fracture closes) and then should react rapid- introduction of encapsulated breakers in 1989, only
ly with the gel. The viscosity of the gel and the mol- low levels of peroxydisulfate breaker were used for
ecular weight of the polymer should be significantly wells above 150°F. Such low levels of breaker were
reduced to allow rapid cleanup of the sand pack not effective at breaking the highly concentrated gels
(Almond et al., 1984; Gall and Raible, 1985). existing in the pack after closure (Brannon and
The most widely used fracturing fluid breakers are Pulsinelli, 1992). To significantly improve pack per-
oxidizers and enzymes. The most common oxidative meability, breaker concentrations 5–10 times higher
breakers are the ammonium, potassium and sodium are required. Encapsulated breakers were developed
salts of peroxydisulfate (S2O82–). Thermal decomposi- to allow high concentrations of breaker to be used
tion of peroxydisulfate (persulfate) produces highly without compromising fluid viscosity during pump-
reactive sulfate radicals that attack the polymer, reducing (Gulbis et al., 1992). In an encapsulated breaker,
ing its molecular weight and its viscosifying ability: the active breaker is coated with a film that acts as a
O3S – O:O – SO32– → ·SO4– + ·SO4– barrier between the breaker and the fracturing fluid.
The breaker may be released as a result of crushing
Free-radical breakers have the potential to create
(Nolte, 1985), osmotic rupture (Walles et al., 1988)
free radicals on the polymer and produce a chain
or diffusion of the breaker chemical (Gupta and
reaction that increases breaker efficiency. Decom-
Cooney, 1992) through the barrier polymer. Any type
position of persulfate, and therefore reactivity, is
of breaker, including enzymes and acids, can be
very temperature dependent, as is typical of chemical
encapsulated. Encapsulation is usually expensive, so
reactions. Thermal decomposition is too slow below
a mixture of dissolved and encapsulated breakers can
about 125°F for persulfates to be used alone, but
be used to achieve the desired level of breaker at the
free-radical generation can be accelerated by the
lowest cost. Because temperature has a significant
addition of amines (Hinkel, 1981). When the fluid
effect on breaker activity and coating permeability, a
temperature exceeds 125°F, sulfate radical genera-
single encapsulated breaker cannot cover the temper-

ature range of 70° to 300°F [20° to 150°C]. A higher fluid, before it heats up in the reservoir, and prema-
temperature usually requires that the breaker have a ture viscosity loss may result.
less permeable coating, which is achieved by Since 1994, enzymes have been used in conditions
increasing the coating thickness or by using a differ- previously thought impossible. Oxidizers and
ent barrier polymer. The breaker chemical itself may enzymes are also being used together, which may
have to be changed for high-temperature applica- provide the best performance. Breaker selection cri-
tions. For example, the half-life of persulfate at teria are listed in Sidebar 7B. It remains to be seen
160°F [70°C] is about 6.8 hr, but at 200°F it is only whether they will evolve to displace the workhorse
about 15 min. Thus, even encapsulated persulfate oxidizers.
breakers have limited utility above 200°–225°F Other materials have been considered for use as
[90°–110°C]. breakers. Oxidizers that are not as active as persul-
Consideration of the theoretical advantages of fates may be used (Misak, 1975). Under controlled
enzyme breakers has led to a renewed interest in conditions, many organic peroxides do not begin to
these materials. Enzymes are biocatalysts, meaning produce free radicals unless the temperature
they are not used up when they react with guar. approaches 200°F (Norman, 1968). Unfortunately,
Theoretically, a single enzyme molecule can react these materials are usually hazardous and difficult to
with many different guar molecules (i.e., turnover handle. Inorganic peroxides, such as CaO2, can also
rate) so polymer degradation might go on for a be used (Mondshine, 1993). Inorganic peroxides are
longer time and be more complete than with an oxi- controlled-release breakers because of the limited
dizer. These enzymes are globular proteins with solubility of the inorganic peroxides at high pH. The
three-dimensional (3D) structures. They promote release of the peroxide ion, and therefore the rate of
reactions with molecules that are able to become reaction with the polymer, is pH dependent. Con-
properly oriented at the enzyme’s 3D active site. taminants such as metal ions can rapidly accelerate
Therefore, enzymes are fairly specific in terms of the decomposition of both organic and inorganic per-
their reactions. In the case of oilfield polymers, an oxides (Sheppard and Kamath, 1978), which can
enzyme that catalyzes the hydrolysis of the inter-ring make their performance unpredictable.
linkages is required. It has been suggested that the Another category of breaker acts by eliminating
most effective enzyme breaker for guar would con- crosslinking. A solution of fluoride ions at low pH
tain a beta-mannanase for attacking the polymer has been used as an overflush to dissolve filter cake
backbone and an alpha-galactosidase for removing formed from a titanate- or zirconate-crosslinked fluid
the side chains (Brannon and Tjon-Joe-Pin, 1994). (Norman et al., 1989).
Commercial oilfield enzyme mixtures used in frac- Acids degrade guar polymers. The release must be
turing fluids contain these enzymes. The ratio of controlled, because fracturing fluids for high-temper-
mannanase to galactosidase varies. ature formations generally are high pH to achieve the
Recently, enzymes have been identified that are desired stability. A combination breaker/fluid-loss
effective up to a pH of about 10, making it possible additive was developed using an acid condensation
to use them in high-pH fluids such as borates. Also, product (Cantu and Boyd, 1988). The material has a
it has been found that commercial enzymes retain particle-size distribution appropriate for a fluid-loss
some level of activity, for at least a short period of additive and hydrolyzes to slowly release an organic
time, above 150°F. They can be used at elevated acid. High concentrations of acid are typically
temperatures if the concentration is increased signifi- required to reduce the pH of the fracturing fluid suf-
cantly. This is particularly true above 200°F, where ficiently for conductivity improvement, so this type
higher concentrations are required because of dena- of breaker has not been widely used.
turing (permanent loss of the 3D structure) of the Guar polymers decompose at elevated tempera-
enzyme. A mitigating factor is pressure. Enzymes tures because of hydrolysis or reaction with dis-
can have increased lifetimes at high pressure solved oxygen, so it would seem that at some tem-
(Michels and Clark, 1992), which can reduce the perature breakers would not be necessary. However,
concentration necessary at temperatures of the benefits of a high-temperature breaker have been
150°–200°F. As with oxidizers, enzymes can be observed even in 300°F wells (McConnell, 1994).
extremely reactive when first introduced into the The reason breakers are beneficial even at very high

7B. Breaker selection
Which type of breaker is the best is currently a topic for much debate. Both of the common types of breakers have strengths and
weaknesses.
Selection Criterion Oxidizers Enzymes Comments

Performance at high temperature + Oxidizers have been identified for high-temperature applications.
Current enzymes have some activity up to 225°F [105°C].
Higher temperature versions will most likely be identified in the
future.
Completeness of break + Theoretically, enzymes should have the advantage because of

their catalytic nature. However, enzyme sensitivity to tempera-
ture, pH and other chemicals can significantly shorten the life-
time of the enzyme. Under ideal conditions (less than approxi-
mately 180°F [80°C], pH 5–8), enzymes break the polymer into
smaller frag|ments than oxidizers do, but there is no information
in the literature to document that significant production of simple
sugars results.
Duration of breaking + Enzymes, unless exposed to extremes of pH or temperature,

react with the polymer over a more extended time period (days)
than oxidizers (hours).
Fast break + A fast break to allow quick turnaround of the well is accomplished
much better with oxidizers.
Chemical sensitivity + + Enzymes are highly sensitive to pH, so control of pH within

a limited range is necessary for predictable performance.
Oxidizers are affected by the presence of curable-resin-coated
proppants, whereas enzymes are not.
temperatures may be that thermal decomposition is permeability rocks. Fluids containing these polymers
reduced when the polymer is in highly concentrated are called wall-building fluids because of the layer of
form as it is in the closed fracture, or it may be that polymer and particulates that builds up on the rock.
the breaker causes a faster break and enhances early This layer, called a filter cake, is generally much less
production. permeable than the formation. If the fluid contains
The breaker situation for oil-base fluids is some- particulates of the proper size, these particulates tend
what different (McKenzie, 1980). Acids and bases to plug the pore spaces and enhance the formation of
are known to rapidly break the aluminum phosphate filter cake. The fluid volume lost before an effective
ester gel. Usually, a slowly soluble acid or base is cake forms is called spurt loss. Pore-size distribution
added to the gel on the fly. Gels break fairly readily for the rock matrix varies from formation to forma-
at elevated temperatures but can be difficult to break tion. Generally, lower permeability formations have
below 100°F [35°C]. smaller pore openings. A 0.1-mD rock may have an
average pore diameter of less than 1.0 µm whereas a
500-mD rock may have an average pore diameter of
7-6.3. Fluid-loss additives 20 µm. The range of pore size may be quite large,
Good fluid-loss control is essential for an efficient which makes it beneficial for fluid-loss additives to
fracturing treatment. Several types of materials are have a wide range of particle sizes so that all pore
used to provide fluid-loss control, but the effective- spaces can be bridged.
ness of the various types depends on the type of In high-permeability formations, polymer and
fluid-loss problem: loss to low- or high-permeability additives may be able to penetrate most pore throats
matrix or loss to microfractures. and form an internal filter cake. In this case, most of
During leakoff into the rock matrix, fluid enters the resistance to leakoff, and therefore pressure drop,
the pore spaces of the rock. Some polymers, such as occurs inside the rock, leaving only a small fraction
guar and HPG, are filtered out on the surface of low- of the total pressure drop in the external cake
(Navarrete and Mitchell, 1995). This makes any

spent acid is too high or if the polymer degrades the fluid loss of acid gelled with surfactant-type
to form insoluble reaction products. Procedures for thickeners, which break during spending, is not sig-
evaluating acid gelling agents and the advantages nificantly different from that of ungelled acid. Only
and limitations of various materials were described those gelled acids that retain their viscosity during
by Crowe et al. (1981). Guar and cellulose base leakoff are capable of providing effective fluid-loss
thickeners lack sufficient stability for use in acid at control. Another disadvantage of this gelling agent
temperatures above 125°F [50°C]. Xanthan biopoly- is the limited temperature range over which it can
mers can be used at temperatures up to 200°F be used. Although the surfactant itself is stable in
[90°C]; at higher temperatures acrylamide copoly- acid, it provides little increase in viscosity above
mers are generally employed. Certain acrylamide 150°F [65°C]. High temperatures disrupt the micellar
copolymers are stable at high temperatures and do association responsible for acid viscosity.
not produce insoluble reaction products. However,
acrylamide homopolymers should not be used as
acid thickeners. They are rapidly hydrolyzed and 7-4.2. Materials and techniques for acid
combine with calcium to form insoluble precipitates reaction-rate control
upon spending of the acid. Reducing the reaction rate of acid to achieve
A number of crosslinked gelled acids have been increased etched fracture penetration is an important
developed. These systems employ various acid consideration. In low- to moderate-temperature wells,
gelling agents and use either polyvalent metal ions or retardation is not critical. However, in applications
aldehydes as crosslinkers. In general, the crosslinked greater than about 250°F [120°C] retardation can be
acids tend to be shear sensitive and are usually critical to the success of the treatment (Nierode and
unstable at elevated temperatures. Breaking of the Kruk, 1973). One of the most common methods of
crosslinked gel in spent acid also is a problem. It extending live acid penetration involves the injection
may be claimed that the gel will be broken by acid of a viscous nonreactive pad preceding the acid. The
degradation, and data are usually presented showing pad reduces the acid reaction rate by increasing the
a loss of viscosity with time in live acid. However, in fracture width and by cooling the fracture surfaces.
actual practice, the acid spends rapidly with no live Weak organic acids (acetic or formic acid) are used
acid available to degrade the polymer. As a result of as retarded acids. These weakly ionized acids react at
their high viscosity, the crosslinked systems present a much slower rate than HCl, even at very high tem-
a greater risk of formation damage and have not peratures. Additives, such as retarders, reduce the
been widely used. reaction rate by forming a protective hydrophobic
Certain surfactants can also be used as thickeners film on carbonate surfaces that acts as a barrier to
for gelling acid (Norman, 1978). These gelling slow acid attack or by blanketing carbonate surfaces
agents thicken acid by forming micelles that associ- with a thin layer of carbon dioxide foam. Oil-outside
ate in chainlike structures and therefore behave much emulsions are the most effective retarders because
like polymers. Gelled acids of this type are quite the external oil phase physically separates the acid
shear-stable, because micellar chains reform quickly from the reactive carbonate surface.
following shearing. Another advantage of the surfac- Gelled acids are usually considered retarded.
tant thickeners is the low viscosity of the spent acid. However, the amount of retardation provided by the
These acid thickeners are usually designed to pro- increased acid viscosity is generally small and can
vide considerable viscosity in live acid but to thin actually accelerate the acid reaction rate under flow-
during spending. This occurs as the result of the ing conditions, as shown by examining the data pre-
disruption of micellar association brought about sented by Gdanski and Norman (1986) and Crowe
by changes in the ionic environment caused by et al. (1990).
increased concentrations of reaction products formed
during spending of the acid. The lowered spent-acid
viscosity aids the recovery of treating fluids follow- 7-5. Multiphase fluids
ing the treatment. However, this can also be a disad- There are situations in which the properties of stan-
vantage. Acid fluid loss is dependent largely upon dard water-base, oil-base or acid-based fluids can be
the leakoff viscosity of the spent acid. As a result,

external filter cake vulnerable to shear degradation Deformable particles such as starches are also
by the fluid. During a fracturing treatment, fluid loss good fluid-loss additives. Starches are polysaccha-
occurs under dynamic conditions (i.e., fluid flows rides of long chains of glucose molecules, compris-
along the face of the formation). Prud’homme and ing 20%–30% linear amylose molecules and
Wang (1993) proposed that the extent to which the 80%–70% branched amylopectin molecules. They
thickness of the filter cake grows is controlled by the are extracted in granular form from the cells of cer-
shear stress τ exerted by the fluid at the wall of the tain plants, such as potato, maize, wheat and tapioca,
cake and the yield stress of the cake. The cake stops which determine their shape and size. Starches can
growing when the fluid stress becomes equal to the be viewed as soft particles that can deform under
yield stress of the cake, and it starts to erode when pressure or stress. This makes them well suited to
the fluid stress is larger than the yield stress of the produce low-permeability filter cakes, because they
cake. The yield stress of a polymer cake depends on can deform and block pore throats or fill in empty
the polymer concentration and pressure gradient in spaces, thus reducing the porosity of the cake. On
the cake, whereas the shear stress of the fluid is the other hand, the deformability of a starch can play
determined by the rheological properties of the fluid an adverse role when pore throats are too big. In that
and the shear rate γ at the formation face. case, starch penetrates into the formation without
Silica flour has been shown to be an effective reducing spurt loss. Combining starch with particu-
fluid-loss additive for helping establish a filter cake. lates can provide an overall fluid-loss additive that
Penny et al. (1985) reported a 10-fold reduction in reduces spurt and forms a low-permeability filter
spurt loss for 5- to 100-mD rock when silica flour cake. Starches are considered less damaging than
was used. Navarrete et al. (1994) determined that the inorganic particles because they degrade naturally
effectiveness of a particulate fluid-loss additive, such or by the addition of amylase enzymes and oxidizers
as silica flour, depends on whether it can reach the (Williamson and Allenson, 1989).
wall of the rock and avoid being sheared off the sur- Oil-soluble resins also can be used to control fluid
face. This depends on the relative ratio of the force loss. These materials, when sized properly with an
driving the particle toward the wall Fy and the shear adequate thermal-softening point, can bridge and
force driving the particle tangentially to the wall Fx. seal the pore spaces to reduce fluid loss. An advan-
The chance that a particle will reach the wall and tage that these materials have over silica flour or
stay there increases as Fy /Fx increases: starches is that they are oil-soluble and dissolve in
produced liquid hydrocarbons. On dissolving, forma-
Fy
~ (q L /dτ w ) . (7-2) tion or proppant-pack damage is minimized. How-
Fx ever, oil-soluble resins are considerably more expen-
sive than other types of fluid-loss additives.
The ratio Fy /Fx increases with
Another method for controlling fluid loss is to use
• larger leakoff flux toward the wall (qL ) dispersed fluids. These fluids are oil-in-water disper-
• smaller particle size (d) sions that contain a fairly small oil concentration and
exhibit good fluid-loss control. For low permeability,
• smaller shear stress (τw) of the fluid along the wall.
the dispersed droplets are effective as deformable
The shape of a particle plays an important role in particles to block pore throats. The dispersion also
its effectiveness to approaching and staying at the produces two-phase flow through the filter cake,
wall. Gauthier et al. (1971) showed experimentally which significantly reduces its permeability to water.
that axisymmetric particles (rods, platelets, etc.) in The effectiveness of 5% diesel as a fluid-loss addi-
a shear flow tend to align their major axis with the tive has been demonstrated in dynamic tests (Gulbis,
flow. Under dynamic conditions, an axisymmetric 1983; Penny et al., 1985). However, Gulbis pointed
particle approaches the fracture surface with its out that diesel is not nearly as effective in a dynamic
major axis parallel to the surface. Once the particle test as in a static test. In a static test, 5% diesel
is on the surface, there is little torque exerted on the reduced the fluid-loss rate by a factor of 5. In a
particle for removal. Therefore, larger aspect ratio dynamic test, the reduction was only by a factor of
axisymmetric particles are better fluid-loss additives 1.5. A possible explanation for the difference is that
than low-aspect-ratio particles. dynamic filter cake containing oil is very thin com-

pared with one without oil, indicating that oil has a saccharides (sugar polymers) used to thicken water
detrimental effect on cake durability and thickness. are an excellent food source for bacteria. Bacteria
Penny et al. (1985) and Nolte (1982) also empha- not only ruin gel by reducing the molecular weight
sized the fragile nature of filter cakes that contain oil. of the polymer, but some can turn the reservoir fluids
Controlling fluid loss to natural fractures that sour. Once introduced into the reservoir, some bacte-
intersect the main fracture is more difficult than con- ria can survive and reduce sulfate ions to hydrogen
trolling fluid loss to the matrix because the openings sulfide (H2S), an extremely dangerous gas with a
to be blocked can be larger and can expand from the characteristic rotten-egg odor (i.e., sour).
pressurized filtrate. Solid materials (such as silica Materials such as glutaraldehyde, chlorophenates,
flour) can bridge a fracture and plug it off (Hall and quaternary amines and isothiazoline are used to con-
Houk, 1983; Wood and Cramer, 1984), but their trol bacteria (Ruseska et al., 1982). Usually the
effectiveness depends on the size of the intersecting materials kill the bacteria, but they do not always
fracture. Silica flour smaller than 200 mesh is very inactivate the enzymes they have produced that are
useful against microfractures (<50 µm wide), but responsible for breaking down the polysaccharides.
larger particles (such as 100-mesh sand) are neces- For this reason, it is common practice to add bacteri-
sary after the macrofracture aperture expands cide to fracture tanks before the water is added to
(>50 µm wide). These large particulates also are ensure that the bacterial enzyme level is kept low.
available as oil-soluble resins. Improvements and changes in bactericides are made so
As stated earlier, the primary purpose of a fractur- that if a strain develops that is resistant to one type of
ing treatment is to generate a highly conductive flow bactericide, another can be used. Bactericides are not
channel through the producing formation to the well- necessary in acid-based or oil-base fracturing fluids.
bore. Ideally, fluid-loss additives should not damage
the permeability of the formation, fracture face or
proppant pack. In reality, many of the fluids and 7-6.5. Stabilizers
additives form long-lasting filter cakes that are not Stabilizers are used to prevent degradation of poly-
readily removed from the fracture. As demonstrated saccharide gels at temperatures above 200°F
by Holditch (1979b), a 90% reduction in permea- (Thomas and Elbel, 1979). The common stabilizers
bility of the fracture face can be tolerated, because are methanol and sodium thiosulfate (Na2S2O3).
of the linear flow regime caused by the fracture. This Methanol is more hazardous to handle and is used as
is true for relatively low-permeability reservoirs 5% to 10% of the fluid volume. Sodium thiosulfate
(e.g., k < 5 mD) where the fracture conductivity is generally used at 10 to 20 lbm/1000 gal. Sodium
is moderate to large and the fracture length is large thiosulfate is the more effective of the two, increas-
(invariably longer than 200 ft). Fracture face dam- ing the viscosity at elevated temperatures by a factor
age, caused by fluid leakoff, is important in high- of 2 to 10, depending on the temperature and time of
permeability fracturing where the fracture length exposure to temperature (Thomas and Elbel, 1979).
may be as short as 15 ft. This damage is one of the The mechanism for these stabilizers is not fully
dominant reasons for post-treatment positive skin understood. It is believed that they act as oxygen
effects (Mathur et al., 1995). For moderate-perme- scavengers and prevent the rapid gel degradation
ability reservoirs, the conductivity of the propped caused by dissolved oxygen. Walker et al. (1995)
fracture, rather than formation permeability, tends studied several oxygen scavengers and found sodium
to be the limiting factor in long-term productivity. thiosulfate to be the most effective at maintaining gel
Therefore, formation face damage caused by fluid- stability. Sodium thiosulfate must be used in great
loss additives is usually a secondary concern. excess because its reaction with oxygen is not favor-
able. Other materials that are better oxygen scav-
engers, such as sodium sulfite and sodium erythor-
7-6.4. Bactericides bate, are not good gel stabilizers, apparently because
Bactericides are added to polymer-containing aque- the reaction products also cause gel degradation.
ous fracturing fluids to prevent viscosity loss caused Fluid pH also should be considered in maximizing
by bacterial degradation of the polymer. The poly- fluid stability. Guar and its derivatives are hydrolyzed

at low pH, especially at elevated temperatures such as zirconium oxychloride (Veley, 1969) and
(>200°F). Therefore, if long-term fluid stability is hydroxyaluminum (Haskin, 1976), are used primarily
desired, a high-pH (e.g., 9–11) fluid should be used. in matrix-acidizing treatments to neutralize the sur-
face charge on clays (Thomas et al., 1976). Unfor-
tunately, they have limited compatibility with higher
7-6.6. Surfactants pH fracturing fluids.
A surface-active agent, or surfactant, is a material Quaternary amines possess a positively charged
that at low concentration adsorbs at the interface group that is attracted to the negatively charged clay
between two immiscible substances. The immiscible particle. Once the quaternary amine is attached to the
substances may be two liquids, such as oil and water, clay particle, the hydrocarbon chain portion extends
a liquid and a gas, or a liquid and a solid. The sur- from the particle, forming an organic barrier and
factant becomes involved in the interface and lowers minimizing the cationic cloud. This type of clay
the amount of energy required to expand the inter- stabilizer is used in water-base fracturing treatments.
face (Rosen, 1972). More detailed information about
surfactant structure and function is in Chapter 15.
Some applications for surfactants in fracturing 7-7. Proppants
fluids have already been discussed. They are neces-
sary ingredients in foams to promote the formation Proppants are used to hold the walls of the fracture
of stable bubbles. They are used in polyemulsion flu- apart to create a conductive path to the wellbore after
ids to stabilize the oil-in-water emulsion. In addition, pumping has stopped and the fracturing fluid has
they are used as surface-tension-reducing agents and leaked off. Placing the appropriate concentration and
formation-conditioning agents (Penny et al., 1983) to type of proppant in the fracture is critical to the suc-
promote cleanup of the fracturing fluid from the frac- cess of a hydraulic fracturing treatment. Factors
ture. Some bactericides and clay-control agents are affecting the fracture conductivity (a measurement
surfactants. of how a propped fracture is able to convey the pro-
duced fluids over the producing life of the well) are
• proppant composition
7-6.7. Clay stabilizers
• physical properties of the proppant
Clays are layered particles of silicon and aluminum • proppant-pack permeability
oxide averaging 2 µm in size (Moore, 1960). Nega-
tively charged particles result when the charge bal- • effects of postclosure polymer concentration in the
ance between positive (aluminum) and negative fracture
(oxygen) is disrupted through displacement of • movement of formation fines in the fracture
cations or breaking of the particles. Cations, from • long-term degradation of the proppant.
solution, surround the clay particle and create a posi-
tively charged cloud. Such particles repel each other
and are prone to migration (Crowe, 1979). Once clay 7-7.1. Physical properties of proppants
particles are dispersed, the particles can block pore
spaces in the rock and reduce permeability. The physical properties of proppants that have an
Solutions containing 1% to 3% KCl are commonly impact on fracture conductivity are
used as the base liquid in fracturing fluids to stabilize • proppant strength
clays and prevent swelling. In addition to KCl, the • grain size and grain-size distribution
organic cation tetramethyl ammonium chloride is an
• quantities of fines and impurities
effective stabilizer (Himes and Vinson, 1991). All
these salts help maintain the chemical environment of • roundness and sphericity
the clay particles, but they do not provide permanent • proppant density.
protection. More permanent methods for controlling
To open and propagate a hydraulic fracture, the in-
clay migration involve the use of quaternary amines
situ stresses must be overcome. After the well is put
or inorganic polynuclear cations. The latter materials,
on production, stress acts to close the fracture and

confine the proppant. If the proppant strength is

inadequate, the closure stress crushes the proppant, 1000
creating fines that reduce the permeability and con-

ductivity of the proppant pack. Proppants can be pro-
duced from a variety of materials and in a variety of High-strength
proppant
size ranges to meet the conductivity requirements of
Intermediate-
the fracture design. strength proppant
The difference between the bottomhole fracturing
Permeability (darcy)
pressure and bottomhole producing pressure pro-
vides an estimate of the maximum effective stress
100
(or closure stress) on the proppant. During flowback
and testing operations, the bottomhole producing
pressure is usually held constant and at a low value Resin-coated
sand
to maximize the production rate. The potential for
maximum crushing can occur during flowback and
testing operations, when the flowing pressure at the Sand
perforations may be low, or initially in the produc-

tion of a well because the fracture gradient is at its
maximum. However, if the well is initially completed 10
and produced at a higher bottomhole pressure and 2000 6000 10,000 14,000
with a nearly constant production rate, the maximum Closure stress (psi)
effective stress on the proppant is less. By producing
a well in this manner, the stress on the proppant can Figure 7-14. Strength comparison of various types of
increase with time, but it never exceeds the bottom- proppants.
hole fracturing pressure. Because the producing pres-
sure is lowest at the well, the effective closure stress or those subject to significant fines migration, are
is highest at the well, and higher strength proppant poor candidates for large proppants. The fines tend
can be used as a tail-in segment after the fracture has to invade the proppant pack, causing partial plugging
been packed with a lower strength proppant. and a rapid reduction in permeability. In these cases,
Strength comparisons are shown in Fig. 7-14. The smaller proppants, which resist the invasion of fines,
following general guidelines may be used to select are more suitable. Although smaller proppants offer
proppants based on strength and cost: less initial conductivity, the average conductivity
over the life of the well is higher and more than off-
• sand—closure stresses less than 6000 psi
sets the initial high productivity provided by larger
• resin-coated proppant (RCP)—closure stresses less proppants (which is usually followed by a rapid pro-
than 8000 psi duction decline).
• intermediate-strength proppant (ISP)—closure Larger grain sizes can be less effective in deeper
stresses greater than 5,000 psi but less than wells because of greater susceptibility to crushing
10,000 psi resulting from higher closure stresses (as grain size
• high-strength proppant—closure stresses at or increases, strength decreases). Larger proppants have
greater than 10,000 psi. more placement problems. Placement problems are
twofold—a wider fracture is required for the larger
Proppant type and size should be determined by grains, and the particle settling rate increases with
comparing economic benefits versus cost. increasing size. If the grain-size distribution is such
Proppants with larger grain sizes provide a more that the mesh range contains a high percentage of
permeable pack because permeability increases as smaller grains, the proppant-pack permeability and
the square of the grain diameter; however, their use therefore conductivity are reduced to about that for
must be evaluated in relation to the formation that is a pack of the smaller grains.
propped and the increased difficulties that occur in The roundness and sphericity of a proppant grain
proppant transport and placement. Dirty formations, can have a significant effect on fracture conductivity.

Proppant grain roundness is a measure of the relative coated proppant may be used as an alternative to ISP,
sharpness of the grain corners, or grain curvature. which is discussed next. Resin-coated sands have a
Particle sphericity is a measure of how close the specific gravity of about 2.55.
proppant particle or grain approaches the shape of The resin coating on some RCPs is cured (at least
a sphere. If the grains are round and about the same partially) during the manufacturing process to form
size, stresses on the proppant are more evenly dis- a nonmelting, inert film. Proppants processed in this
tributed, resulting in higher loads before grain failure fashion are called precured-resin-coated proppants.
occurs. Angular grains fail at lower closure stresses, The major application for precured-resin-coated
producing fines that reduce fracture conductivity. proppants is to enhance the performance of sand
Proppant density has an influence on proppant at high stress levels.
transport because the settling rate increases linearly A curable resin coating may also be applied to
with density. Therefore, high-density proppants are sand or other types of proppants. The major applica-
more difficult to suspend in the fracturing fluid and tion of curable-resin-coated proppants is to prevent
to transport to the top of the fracture. Placement can the flowback of proppants near the wellbore. The
be improved in two ways: using high-viscosity fluids curable-resin-coated proppants are mixed and
to reduce settling or increasing the injection rate to pumped in the later stages of the treatment, and the
reduce treatment time and the required suspension well is shut in for a period of time to allow the resin
time. Also, high-density proppants require more to bind the proppant particles together. Theoretically,
mass of material to fill a given fracture volume. the RCP cures into a consolidated, but permeable,
filter near the wellbore.
Although they provide versatile and reliable per-
7-7.2. Classes of proppants formance, RCPs contain components that can inter-
Sand is the most commonly used proppant. It is the fere with common fracturing fluid additives, such as
most economical, is readily available and generally organometallic crosslinkers, buffers and oxidative
provides sufficient fracture conductivity for closure breakers. These undesirable interactions have been
stresses less than 6000 psi. Its specific gravity is about reported to interfere with the crosslinking of organ-
2.65. Depending on the overall balance of physical ometallic crosslinkers, suppress fracturing fluid
properties, sand can be subdivided into groups: cleanup by consuming oxidative breakers and com-
promise proppant-pack bonding, leading to reduced
• northern white sand permeability, proppant flowback and increased prop-
• Texas brown sand pant crushing (Dewprashad et al., 1993; Nimerick
• Colorado silica sand et al., 1990; Stiles, 1991; Smith et al., 1994). Fiber
technology, as discussed in Section 11-6.4, is an
• Arizona silica sand.
alternative technique for proppant flowback prob-
American Petroleum Institute (API) standards can lems that introduces no chemical compatibility
be used to similarly qualify and group any sand source. issues or special curing requirements for time and
Resin coatings are applied to sand (usually north- temperature. Guidelines for minimizing the undesir-
ern white sand) to improve proppant strength and able effects of RCPs are listed in Sidebar 7C.
to reduce flowback during production. Resin-coated ISP is fused-ceramic (low-density) proppant or
sand is stronger than conventional sand and may be sintered-bauxite (medium-density) proppant. The
used at closure stresses less than 8000 psi, depending sintered-bauxite ISP is processed from bauxite ore
on the type of resin-coated sand. At closure stresses containing large amounts of mullite. This is in con-
greater than 4000 psi and without adverse fluid trast to a high-strength proppant, which is processed
effects on the resin, resin-coated sand has a higher from bauxite ore high in corundum. ISP is generally
conductivity than conventional sand. The resin helps used at closure stresses greater than 5,000 psi, but
spread the stress over a larger area of the sand grain less than 10,000 psi. The specific gravity of ISP
and reduces the point loading. When grains crush, ranges from 2.7 to 3.3.
the resin coating helps encapsulate the crushed por- High-strength proppant is sintered bauxite contain-
tions of the grains and prevents them from migrating ing large amounts of corundum, and it is used at clo-
and plugging the flow channel. In some cases, resin- sure stresses greater than 10,000 psi. High-strength

it means that all ingredients (except fluid-loss addi-

7C. Minimizing the effects of resin-coated
proppants tive, breaker and proppant) are blended together in
the fracture tanks (typically, 500-bbl capacity) before
To minimize the effects of curable-resin-coated proppant and pumping begins. The tanks are usually mixed the
fluid interactions, the following guidelines are recommended.
day before pumping because the gel takes several
• Use fiber reinforcement. hours to form. A fluid-loss additive and a breaker are
added on the fly as the gel is pumped. These materi-
• Minimize the amount of RCP. If proppant flowback control als are added on the fly to prevent the fluid-loss addi-
is required, consider alternate materials for addressing
the problem. Card et al. (1994) described the incorpora- tive from settling out in the fracture tanks or the
tion of fibers in the proppant pack as a means to prevent breaker from prematurely reducing the gel viscosity
flowback.
prior to pumping.
• Use precured-resin-coated proppants. These materials For batch-mixed, water-base fluids, the bacteri-
are cured (at least partially) and are typically less reactive cide, polymer, salt, clay stabilizer, etc., are mixed
with fracturing fluid additives than fully curable-resin-
coated proppants.
together before pumping. The polymer is given suffi-
cient time to hydrate in the tanks before the job
• Avoid using curable-resin-coated proppants in conjunction begins. The pH of the gel is adjusted for optimum
with high concentrations of oxidative breakers. RCPs
decrease breaker effectiveness. Conversely, oxidative
crosslinking. Crosslinker is added on the fly in the
breakers can compromise the strength development of case of transition metal (Ti and Zr) crosslinkers.
an RCP. The reactivity of RCPs with oxidative breakers Because borate crosslinking occurs only at a high
varies, and some RCPs have little effect on breaker activ-
ity. Additional breaker may be required when using RCPs pH, boric acid can be added to the polymer in the
to ensure adequate cleanup. tanks, and a base such as NaOH can then be added
on the fly to raise the pH and to initiate crosslinking.
• Always determine the compatibility of the fracturing fluid
and the RCP before the treatment. Typically, 30%–60% As discussed later, batch mixing affords the best
of the curable-resin coating can be lost to the fluid. Resin opportunity for quality assurance. Unfortunately, it
loss does not behave in a straightforward manner with
changes in temperature or time at temperature. Resin also results in wasted materials. There are always
loss increases as exposure to shear increases. tank bottoms, the fluid that cannot be drawn out of
the fracture tanks. Typically, tank bottoms represent
• Never batch mix RCPs in fracturing fluids. Minimize han-
dling of RCPs to keep dust levels low. Solid resin is a at least 7% of the total volume of fluid in the tanks,
good emulsion stabilizer and can create an emulsion with resulting in the waste of 7% of the batch-mixed
the fracturing fluid that is not miscible in water.
chemicals and requiring costly disposal. Also, this
fluid must be broken and the fracture tanks should be
cleaned. If the job is postponed and the gel degrades
proppant is the most costly proppant. Its specific
because of bacterial action, the entire batch of gel
gravity is 3.4 or greater.
may have to be discarded. From a cost standpoint,
continuously mixed fluid is more desirable. In this
7-8. Execution mode, all materials are added on the fly, so there is
During the fracturing treatment, fluid chemistry no wasted fluid and no unnecessary expense. Poly-
comes together with proppant handling, mixing and mer slurries (concentrated suspensions of guar or
pumping equipment to create the desired propped HPG in diesel) were developed so that polymer
fracture. The field environment is often quite differ- could be accurately metered and so that it would dis-
ent from the ideal laboratory conditions in which the perse and hydrate rapidly enough for continuous
fracturing fluid or additive was developed. The fol- mixing (Constien et al., 1988; Yeager and Bailey,
lowing sections address the field environment. 1988). This type of operation requires accurate
metering of all materials and makes quality assur-
ance more difficult. Techniques for on-site rheology
7-8.1. Mixing measurement have been developed so that the linear
(precrosslinked) gel viscosity can be closely moni-
Fluids may be batch mixed or continuously mixed. tored. Because of environmental considerations and
Batch mixing has slightly different meanings, disposal costs, most aqueous-based fluids are now
depending on the fluid prepared. For oil-base fluids, continuously mixed.

7-8.2. Quality assurance fluid back and forth between two containers. The
time at which the fluid appears very thick (stringy or
Testing the components of the fracturing fluid prior pours as one mass from container to container) is the
to the treatment is the only way to ensure that the crosslink time. The falling-ball viscometer measures
fluid used will have the properties expected of it. the time required for the crosslinked fluid to develop
In the case of a water-base fluid, the mix water must suspending properties. Hodge and Baranet (1987)
meet certain specifications. It should be relatively point out that the vortex closure and static stiffness
free of iron (<25-ppm Fe2+) and other reducing tests are misleading because the procedures do not
agents that promote gel degradation or catalyze the simulate the shear the fluids experience in the tubing
reaction of peroxydisulfate (persulfate) breakers or the heat-up that occurs in the wellbore. Both fac-
(Ely, 1985). The pH of the water affects the polymer tors greatly impact the crosslinking rate and viscosity
hydration rate. If the water pH is low, hydration may development. As discussed in Section 7-6.1, the use
be rapid and lumping of the gel results. If the water of dual-crosslinker systems can eliminate the neces-
pH is too high, the gel hydrates too slowly. Water sity of accurate control of the crosslink time.
temperature also affects hydration rate. Increasing Not all potential mix water sources or fracturing
the water temperature causes faster polymer hydra- fluid additives are compatible. When preparing the
tion (viscosity development). A sample of polymer test sample with materials from location, all compo-
from location should be used to prepare a linear nents should be included. Any severe compatibility
(precrosslink) gel with the water from the fracture problems should show up in this test, preferably well
tanks. Testing should ensure that proper dispersion before the treatment and in a laboratory setting. To
(no lumps) and hydration (viscosity development) determine how well the fluid will perform at elevated
will be obtained on location. temperatures, a Fann 50 rheology test (see Chapter
If the proper viscosity can be achieved, crosslinker 8) is usually performed. This preliminary test should
performance can be evaluated. A fracturing fluid reveal problems with the chemicals or problems with
sample containing the appropriate concentration of the mix water or oil source.
crosslinker can be heated to accelerate crosslinking Once the job is in progress, the viscosity of any
and produce a thick gel. In addition to gross cross- continuously mixed linear gel should be monitored
linker performance, a measure of how long it will to quickly identify problems that could develop.
take for the gel to crosslink should be performed Also, a sample of the crosslinked fluid should be
(crosslink time). Common on-site tests for measuring inspected periodically. A detailed discussion of qual-
crosslink time are the vortex closure test, static stiff- ity assurance procedures is provided by Ely (1985).
ness test (Hodge and Baranet, 1987) and falling-ball
viscometer method (Cawiezel and Elbel, 1990). In
the vortex closure test, polymer solution is stirred
in a blender to create a vortex. The crosslinker is Acknowledgments
added, and the time required for the vortex to close,
Curtis Crowe contributed the acid fracturing section.
forming a smooth surface, is termed the vortex clo-
Reinaldo Navarrete updated the fluid-loss section.
sure time. The static stiffness test involves mixing
The proppant section was written originally by Brian
the fluid components together and then pouring the
Ainley and edited by Bob Thrasher.


◆
Performance of
Fracturing Materials
Vernon G. Constien, Constien and Associates
George W. Hawkins, Schlumberger Dowell
R. K. Prud’homme, Princeton University
Reinaldo Navarrete, Kelco Oilfield Group
8-1. Introduction • quality control before or during the treatment.
Hydraulic fracturing employs special fluids that are Characterization during the development process
intended primarily to create an appropriate fracture for an additive or fluid system is typically used to
geometry while transporting proppants. Key to the determine if a new composition is an improvement
entire exercise is the fluid rheology, which affects the over an existing system or if it can provide similar
fluid viscosity, its proppant-carrying capability and its performance at a lower cost.
propensity for leaking off into the porous medium. Characterization of fluid systems that obtains rep-
This chapter begins with fracturing fluid character- resentative performance data in critical areas such as
ization, followed by the translation of laboratory- rheology, pipe friction pressure, fluid-loss rates, frac-
obtained information to field conditions. There is a ture conductivity and formation damage is conducted
clear link between the chemistry of fracturing fluids, to obtain data that can be used in fracture design and
most of which are polymer solutions with their prop- production simulators.
erties augmented by several additives, and the physi- Characterization of fluid systems at the point of use
cal properties expected from these fluids. Rheology for quality assurance purposes usually involves meth-
and its modeling and control take center stage in the ods that can be applied at less than ideal conditions to
chapter. This includes not only traditional water-base indicate how the systems are performing. Examples
polymer solutions but also complex fluids such as of these tests are provided in Davidson et al. (1994),
foams. Proppant-transporting slurries are also exam- Ely (1985) and Section 7-8 of this volume.
ined in relative detail with regard to both their rheol- The American Petroleum Institute (API) has
ogy and potentially problematic proppant settling. developed recommended practices for several of the
The general discussion of fracturing fluid loss con- laboratory characterization methods described in this
tains some of the classic thinking, especially with chapter. An excellent review of laboratory methods
regard to laboratory-derived data. This discussion is also in the Society of Petroleum Engineers (SPE)
should be read in conjunction with Chapters 9 and Monograph 12, Recent Advances in Hydraulic
2, which address leakoff modeling and the interpre- Fracturing (Gidley et al., 1989).
tation of fracturing pressure behavior, which are inti- This chapter addresses the characterization meth-
mately related phenomena. Finally, damage from ods used for additive and system development and
fracturing fluids to the formation face and the result- for obtaining input data for use in fracture design
ing proppant pack, the first from fracturing fluid simulators. The focus is on the most commonly used
leakoff and the second because of unbroken polymer fluids: water-base systems utilizing guar or deriva-
following the treatment, are outlined. These prob- tized guar as the polymer viscosifier (see Chapter 7).
lems are tackled in detail in Chapter 12. In general, methods used for water-base fluids can be
applied—with appropriate modification—to other
8-2. Fracturing fluid characterization fluids such as gelled oils, emulsions and foams.
Whatever the fluid system, a set of data describing
Fracturing fluid additives and fluid systems are char- the fluid rheology, fluid loss, pipe friction, fracture
acterized for the following purposes: conductivity and possible formation damage should
• additive and system development be determined before the fluid system is used in field
operations.
• obtaining input data for use in fracture design
simulators

8-3. Characterization basics 8-5. Molecular characterization

Before any significant laboratory work to determine of gelling agents
fluid system properties is begun, representative mater- An important set of techniques for the understanding
ial samples of all components must be obtained. and prediction of fracturing fluid performance
These include samples of the water, hydrocarbon and involves characterization at the molecular level. As
acid. In some instances sampling may require obtain- discussed in Chapter 7, water-base fracturing fluids
ing multiple lots of materials to prepare a composite are formulated primarily from guar or derivatized guar
blend. At a minimum, all materials must be identified polymers. Methods such as intrinsic viscosity mea-
by the manufacturer and lot number so that they can surements, dynamic light scattering, nuclear magnetic
be traced back to the point of origin. In the case of resonance (NMR), size-exclusion chromatography
proppants, special care must be taken because it is (SEC) and freeze-fracture electron microscopy pro-
generally difficult to obtain representative samples vide information on guar molecular size, crosslinking,
from any size lot (Recommended Practices for Testing degradation mechanisms and gel structure.
Sand Used in Hydraulic Fracturing Operations Before extensive investigations are made of the
[RP 56], 1983). This difficulty arises because of size properties of fully formulated guar-based fracturing
segregation during transport and transfer. fluid systems, it is worthwhile to consider the proper-
In addition to obtaining representative material ties of the guar itself. Guar is a natural product and,
samples, calibrating and maintaining the laboratory as such, is subject to variation of its properties.
instruments before and during testing is required. Suppliers of oilfield-grade guar blend batches to
Adherence to the prescribed testing methods is cru- obtain consistent performance from the materials.
cial for comparing different systems and avoiding fre- However, before an investment is made in the devel-
quent rerunning of the data to obtain an exact match opment of a new fluid system, the basic properties of
to the testing conditions. Complete records containing blended representative batches of the base polymer
material sample and calibration data information are should be determined.
essential for product or system comparisons because
the usable lifetime of the data can span several years.
8-5.1. Correlations of molecular weight
8-4. Translation of field conditions and viscosity
to a laboratory environment Guar and its derivatives are the polymers most com-
The range of conditions that fracturing fluids experi- monly used in fracturing fluids. Like many polysac-
ence during their use is so diverse and the physical charides, guar is conformationally mobile and adopts
scale so large that the characterization process is usu- a disordered or “random-coil” geometry in dilute
ally divided into a series of small-scale tests that sim- solutions.
ulate key portions of the large, complex physical Viscosity–molecular weight relations for several
environment. The scope of the characterization can guar samples were determined by Robinson et al.
potentially range from the molecular level to testing (1982) using dilute solution viscometry and light-
fluids in large-scale equipment, which may approach scattering measurements. They concluded that guar
actual mixing, wellbore and fracture dimensions. acts as a random-coil polymer and that the methods
The exact details of the characterization method commonly used to study synthetic polymers could be
may change depending on the composition of the used to determine information such as intrinsic chain
fluid. For example, oil-base fluids, emulsions and flexibility.
foams may have compressibility or bubble- or drop- Intrinsic viscosity µi is a convenient index of the
size effects that require the determination of proper- size or hydrodynamic volume of isolated polymer
ties over a range of pressures. coils in solution. For random-coil polymer solutions,
µi can be expressed as a function of the end-to-end
8-2 Performance of Fracturing Materials

distance of a random-coil polymer and the molecular Table 8-2. Mark-Houwink coefficients for fracturing
weight. An experimental determination of the intrinsic µi = K(Mv)a).
fluid gelling agents (µ
viscosity is obtained with
Polymer K a Reference
ln µ r  ln µ µ 0 
µi =  = , (8-1) HEC 1.03 × 10–3 0.70 Sharman et al. (1978)
 C  C=0  C  C=0
Galactomannan 6.7 × 10–4 0.80 Sharman et al. (1978)
where C is the concentration and the viscosity para- Guar 3.8 × 10–4 0.723 Robinson et al. (1982)
meters are defined in Table 8-1. Guar 3.67 × 10–5 0.884 Pope et al. (1994)
Table 8-1. Viscosity nomenclature

(Kaufman and Falcetta, 1977).
8-5.2. Concentration and chain overlap
Solvent viscosity µ The interaction of concentration and the molecular
Shear-rate-independent µ0 size or weight of the gelling agent is important for
(zero-shear) viscosity understanding the critical concentrations necessary
 ln(µ µ 0 )  for crosslinking and the shear stability of dilute
Intrinsic viscosity µi =   crosslinked fracturing fluids. As the concentration C
 C  C =0
of the gelling agent in solution is increased, the poly-
Relative viscosity µr = µ/µ0 mer coils begin to interact at a concentration called
Specific viscosity µsp = µr – 1 the critical overlap concentration C*. The value of C*
is determined by measuring the solution viscosity as
ln(µ r ) a function of concentration. The results are plotted as
Inherent viscosity µ inh =
C the log specific viscosity µsp versus log polymer con-
centration (i.e., the Kraemer method). The specific
To determine the intrinsic viscosity of a polymer, viscosity is used to remove the contribution of the
viscosity measurements are made on a series of dilute solvent to the bulk viscosity:
samples at very low shear rates. The extrapolation to
µ sp = (µ bulk − µ solvent ) µ solvent . (8-3)
zero concentration at zero shear is the value of µi.
The Mark-Houwink equation provides a close As discussed previously, extrapolation to zero poly-
approximation of the variation of µi with molecular mer concentration provides the intrinsic viscosity µi.
weight: As the polymer concentration is increased from zero,
µ i = K ( Mv ) , C* is indicated by a significant increase in slope. C*
a
(8-2)
is also affected by the solvent, other soluble species
where the Mark-Houwink coefficient K in dL/g and such as salts and temperature.
the exponent a are experimentally determined con- C* is the theoretical minimum concentration at
stants for a polymer-solvent system at a specific tem- which intermolecular crosslinking is possible. Below
perature. Mv is the viscosity average molecular C*, it is not possible to form chemical crosslinks
weight. The value of a provides an indication of the because the polymer molecules are too far apart.
degree of coiling or extension of the polymer chains. Reaction of the crosslinker below C* results predomi-
For spheres, a = 0; for rigid rods, a = 2; and for ran- nantly in intrapolymer crosslinks but not interpolymer
dom chains, a ≅ 1. For guar, values of a typically have crosslinks (Menjivar, 1984; Pezron, 1988a, 1988b).
been reported between 0.72 (Robinson et al., 1982) For solutions of polysaccharides, C* is about 4/µi
and 0.80 (Sharman et al., 1978). (Robinson et al., 1982). Pope et al. (1994) determined
Data for the Mark-Houwink coefficients for guar the relation (Fig. 8-1) between specific viscosity and
and hydroxyethylcellulose (HEC) are listed in the concentration and intrinsic viscosity for the high-
Table 8-2. molecular-weight guar commonly used in hydraulic
fracturing. From Fig. 8-1 and the µi value reported for

to water-soluble polymers is more difficult than for

10,000
50°F (high molecular weight) + polymers soluble in organic solvents (e.g., polyethyl-
+
65°F
80°F
(high molecular weight)
ene, polystyrene), useful data can be obtained if care
1000
+
95°F
110°F
is taken. SEC is a highly effective method for study-
+
125°F
140°F
ing the degradation of water-soluble polymers. Its
100 50°F (low molecular weight) + advantage over simple measurements of intrinsic vis-
+
65°F (low molecular weight)
+
µsp
+ 80°F (low molecular weight) cosity is that it provides information on the molecular
+
95°F (low molecular weight)
10 + weight distribution in the sample and relative differ-
+
+
+
ence between samples for evaluating guar degrada-
1 + tion processes.
+
+ An example from an SEC analysis of HPG

0.1 degraded by an ammonium persulfate breaker is
0.1 1 10
shown in Fig. 8-2. This technique is not only useful
Cµi
in laboratory studies, but important for analyzing
Figure 8-1. Relationship between specific viscosity and
flowback samples of the fracturing fluid to determine
the concentration and intrinsic viscosity for two molecular how breakers work in actual field conditions. When
weights of guar over a temperature range of 50° to 140°F
[10° to 60°C] (Pope et al., 1994).
12
0
Pope et al.’s guar sample of 15.5 dL/g, C* is about
10
0.21 dL/g, or in terms of common oilfield units, about Weight difference (%)
0.812 × 10–4
17.5 lbm/1000 gal. This value is in good agreement
8
with the lower concentrations of guar used in cross- 1.62 × 10–4
APS concentration (N)
linked fracturing fluids, which are in the range of 3.25 × 10–4
6
15–20 lbm/1000 gal (Harris and Heath, 1998). 6.50 × 10–4
Values for µi and C* for various molecular weights 4

of hydroxypropylguar (HPG) reported by Menjivar
(1984) are listed in Table 8-3. 2
0
8-5.3. Molecular weight distribution 1E+4 1E+5 1E+6
Molecular weight
Size-exclusion chromatography (Gall and Raible,
1984, 1985; Sattler et al., 1985; Hawkins, 1986) is a Figure 8-2. Molecular weight reduction and broadened
technique for measuring the molecular weight distrib- distribution of HPG with increasing ammonium persulfate
ution of polymers. Although the application of SEC (APS) breaker concentration at 199°F [93°C] for 5 hr.
Table 8-3. Critical overlap concentration C* for several HPG samples (Menjivar, 1984).
HPG Sample Intrinsic Viscosity Critical Overlap Concentration Estimated Average

(dL/g) (g/dL) Molecular Weight
A 18.1 0.19 3.0 × 106
B 16.4 0.21 2.6 × 106
C 15.9 0.22 2.5 × 106
D 15.4 0.22 2.4 × 106
E 13.8 0.25 2.0 × 106
F 13.3 0.26 1.9 × 106
G 12.5 0.27 1.8 × 106

the molecular weight (or intrinsic viscosity) and con- Relatively few studies have identified the actual
centration of guar are known, the viscosity can be chemical species involved in the reactions with metal
predicted from a master curve. This information ion crosslinkers and guar or guar derivatives (see
enables estimating the tendency for viscous fingering Chapter 7). Some borate-crosslinked fracturing fluid
to occur during flowback of the fracturing fluid (Pope systems have been studied using NMR. Using this
et al., 1994). technique, Jasinski et al. (1996) were able to deter-
mine the stoichiometries of the borate anion and
high-molecular-weight guar to derive the equilibrium
8-5.4. Characterization of constants. The number of borate crosslinks for a gel
insoluble components can be calculated from measurements of the borate-
Although molecular size studies of the water-soluble to-monomeric-sugar equilibrium-binding constants.
portion of polymers provide much understanding of These constants have been measured for model sug-
performance properties such as viscosity, some por- ars and the borate-to-sugar equilibrium-binding con-
tions of the guar molecule are not soluble. The insol- stants used to model gel rheology (Kesavan et al.,
uble components can be present in the polymers from 1993; Pezron et al., 1988b) and correlate computed
the beginning, or they can be created during fluid crosslink concentrations with the elastic storage mod-
degradation. The insoluble components are difficult ulus G´ (see Section 8-6.7) (Jasinski et al., 1996).
to characterize for determining their influence on The overall borate ion equilibrium is
properties such as fluid loss and fracture conductivity B(OH)3 + H2O B(OH)4– + H+
damage. Analysis of the insoluble components of
guar is also difficult. For guar systems, insight can Ka
be gained through basic elemental analysis (e.g., total with the computed fraction of borate anion versus pH
carbon, hydrogen and nitrogen) and more complex shown in Fig. 8-3. The acid equilibrium constant Ka
techniques such as fourier transfer infrared spec- of the borate ion equilibrium equation is 9.2.
troscopy (FTIR). These analyses can also provide The multiple equilibria associated with borate ion
information on the change in the galactose:mannose crosslinking of guar can be calculated from the known
ratio (see Chapter 7 for the guar structure) in the sol- equilibrium constants shown in Fig. 8-4. The figure
uble and insoluble guar residues following exposure illustrates that as a function of pH value, the concen-
to breakers or elevated temperatures. tration of active crosslink sites (denoted as A2B–)
changes by over 1 order of magnitude from a neutral
pH of 7 to a pH of 9.2. The number of active cross-
8-5.5. Reaction sites and kinetics link sites decreases at very high pH values. The equi-
of crosslinking libria are also a function of temperature. Increasing
Crosslinking is used to enhance the performance of temperature produces an exponential decrease in
fracturing fluids by effectively increasing the molecu- crosslink concentration, as shown in Fig. 8-3. The
lar size of the gelling agent. Figures 8-1 and 8-2 equilibria enable understanding the mechanism for
show that viscosity can be increased by increasing
the polymer concentration, molecular size or both. 1.0
20°C
For economic and performance reasons (especially 0.8 40°C
for fracture conductivity), it is desirable to use as low
Fraction B(OH)4–
60°C
80°C
a concentration of gelling agent as possible. If the 0.6
concentration is fixed at a value slightly higher than 0.4

C*, then to achieve high viscosity, µi (or the volume
0.2
occupied by the polymer chain) must be increased.
Crosslinking the polymer chains together increases 0
the effective size of the species up to a size limited 7 8 9 10 11 12
pH
by the stoichiometry of the reaction and the shear to
which the fluid is exposed. Shear effects on fluid Figure 8-3. Borate anion fraction in a 0.01M boric acid
solution (= 0.7M NaCl, no guar) versus pH and tem-
properties are discussed in Section 8-6.4. perature.

10
A2B – AB – B(OH)3
+ ABOH B(OH)–4 A0
1 Torsion spring
Transducer
Concentration (M)
0.1
++ ++ To digital display
++ ++
0.01 + + and/or recorder
+
+
0.001
+
+
0.0001 Rotor
5 6 7 8 9 10 11
pH
Figure 8-4. Species concentrations calculated from the

equilibrium model versus pH. A2B – = crosslink site,
ABOH = borate in the nonionic form attached to one
cis-diol, AB – = a borate ion attached to one cis-diol, Bob
B(OH)4– = borate ion, B(OH)3 = neutral boric acid,
A0 = uncomplexed cis-diol.
controlling borate gel rheology. By selecting suitable

Rotation
chemistry to control the concentration of borate ions
or the base (pH) in solution (Cawiezel and Elbel,
1990; Nelson et al., 1997; Dawson, 1992), it is possi- Figure 8-5. Rotational concentric cylinder viscometer.
ble to produce systems with high viscosity at high
temperatures that do not overcrosslink and synerese
(i.e., phase separate) while the fluid is heating up to 2
the temperature within the fracture.
∆p
8-6. Rheology
The most common laboratory evaluations conducted
on fracturing fluids are steady-shear rheological mea- 5
6
surements. The property that is determined is the
apparent viscosity of the fluid as a function of shear 1
3
rate, temperature, fluid composition and time. These 4 7
relations are commonly determined in viscometers,
such as the rotational concentric cylinder (Fig. 8-5),
capillary (Fig. 8-6) and large pipe or slot-flow devices.
The data are usually related to a mathematical 1 Measurement capillary
model for predicting the fluid viscosity in the various 2 Differential pressure ∆p transducer
environments that occur in the fracturing process. 3 Floating piston accumulator containing test fluid
on top of piston and hydraulic fluid below piston
Because of the difficulties in evaluating the rheol- 4 Programmable-temperature oil bath
ogy of fluids containing proppant in small-scale labo- 5 Hydraulic fluid directional control valve
6 Variable rate positive-displacement pump
ratory instruments, most of the slurry rheology data 7 Hydraulic fluid reservoir
in the industry have been derived from large pipe or
slot-flow devices, as discussed in Section 8-6.9. Figure 8-6. Reciprocating capillary viscometer.

8-6.1. Basic flow relations Newtonian behavior implies that fluids have a con-
stant viscosity at all shear rates. Water, low-viscosity
Simple viscosity characterizations involve measure- oils and gas are examples of fluids that exhibit this
ments of the stress that results from applying a behavior. Fracturing fluids have predominantly non-
known shear rate on the fluid. Shear in laminar flow Newtonian behavior. This means that the apparent
can be thought of as a process in which infinitely viscosity of the fluid is dependent on the shear that
thin, parallel planes slide over each other (Fig. 8-7). the fluid is experiencing at a specific point. Figure 8-8
The shear rate γ is defined as the velocity difference illustrates the apparent viscosity of a simple fracturing
between the planes divided by the distance between fluid over a wide range of shear.
the planes:
du u1 − u2 100
γ= = . (8-4) 60°F
dx x 80°F
100°F
+
The usual rate of shear reported in viscometric exper- 125°F
150°F
iments is the value at the wall of the instrument and
Viscosity (Pa·s)
+
+
+
+
+
+
+
+
+
is referred to as the apparent Newtonian or nominal
+
++
+
10–1
shear rate.
+
+
The shear stress τ is the shearing force per unit Newtonian
+
region
+
area of surface:
+
+
F Power law
τ= . (8-5) region
+
A
+
10–2
10–2 10–1 100 101 102 103 104
In most measurements, the shear stress is determined
by measuring the torque exerted on a measurement Shear rate (s–1)
bob or by the pressure drop across a tube.

Figure 8-8. Shear viscosity of hydroxypropylguar (0.48%
The apparent viscosity µa is the ratio of the shear in water) (Guillot and Dunand, 1985).
stress to the shear rate:
τ
µa = . (8-6) A fracturing fluid will have considerably different
γ
values of µa, depending on the shear that is exerted
on the fluid. The non-Newtonian behavior of most
u1
fracturing fluids plays a significant role in the friction
pressure developed in the tubing and the fracture and
also in the ability of the fluid to transport proppant.
F
Rheological characterization of a non-Newtonian
fluid requires that the response of the fluid to changes
A in shear rate must be determined and related to a
model so that the apparent viscosity can be calculated
under the different shear conditions in the fracture.
x
8-6.2. Power law model

The power law model is the most widely used model
to represent fracturing fluid behavior in fracture
A
design simulators:
τ = Kγ n (8-7)
u2
Figure 8-7. Shear rate depicted as the difference in veloc- µ a = K γ (1−n ) , (8-8)
ity between two plates divided by the distance x between
the plates.

where K is the consistency index in lbf-sn/ft2 or kPa·sn three flow devices are listed in Table 8-4. The device-
and n is the flow behavior index (dimensionless). independent values are used for the general flow
These relations hold for most fracturing fluids over equations in Chapter 6. Tables 9.1 and 9.2 in Gidley
the range of shear rates in which the fluid displays et al. (1989) provide a comprehensive summary of
non-Newtonian behavior. A log-log plot of τ versus γ these flow considerations.
(Fig. 8-9) usually yields a straight line over a portion
of the shear range. The slope of the straight-line por- Table 8-4. Viscometer geometry corrections.
tion is equal to the behavior index n, and the value of
τ at γ = 1.0 s–1 is equal to the consistency index K.  B 2 n (B 2 − 1) 
−n
A log-log plot of µa versus γ has a straight-line slope K = Kv  2 

 n (B − 1)B 
2n
of n – 1 when the power law model is applicable per

Eq. 8-8 (Fig. 8-8). The slope is zero for Newtonian  3n + 1
n
behavior. K pipe = K  
 4n 
 2n + 1
n
10 K slot = K  
 3n 
47, 880K
µa =
Shear stress (lbf/ft2)
γ( )
1− n
1 K = geometry-independent consistency index (lbf-sn/ft2)

Kv = consistency index for a concentric cylinder viscometer
B = rcup /rbob
Slope = n
rcup = cup inner radius
rbob = bob radius
Kv
0.1 Kpipe = consistency index for pipe flow
1 10 100 1000 Kslot = consistency index for slot flow
Shear rate (s–1) µa = apparent viscosity (cp) at a specific shear rate
γ = shear rate (s–1)
Figure 8-9. Determination of power law coefficients from
capillary viscometer data.
8-6.3. Models that more fully describe

Fracturing fluid rheology data are usually deter- fluid behavior
mined under laminar flow conditions in a rotational A limitation of a model that uses only n and K is its
concentric cylinder viscometer (Fig. 8-5) and inability to accurately describe the Newtonian region
reported in terms of the power law parameters n and of fluid behavior that generally occurs at both very
K. However, K is dependent on the flow geometry for low and high shear rates. The prediction of viscosity
concentric cylinder devices and is referred to as the using only high-shear values of n and K can vastly
viscometer consistency index Kv. For a power law overestimate the viscosity at low shear rates.
fluid, the shear rate depends on the value of n in addi- Because shear rates vary across the fracture width
tion to the flow rate and conduit dimension. As a from high at the wall to zero in the center, the viscos-
result, determination of the shear rate is coupled to ity of the fracturing fluid also varies. The viscosity at
the determination of n from the flow data. Savins the wall is much lower than the viscosity in the center
(1958) overcame the coupling problem by assuming of flow. Figure 8-10 is a laser doppler velocimetry
that the shear rate was the same as for a Newtonian (LDV) plot of a fracturing fluid profile in a slot-flow
fluid. This assumption enables determining the value geometry (Lear, 1996). Equation 8-4 implies that the
of n and a device-independent value of K (i.e., Kpipe shear rate is proportional to the slope of the flow pro-
or Kslot; Table 8-4) for each flow rate. The corre- file; therefore, Fig. 8-10 indicates a relatively small
sponding device-independent values depend on the shear rate over the central portion of the slot. The
ratio of the actual and apparent Newtonian shear variations in shear rate and viscosity are important for
rates, raised to the nth power. Savin’s relations for the accurately predicting the fracture width and proppant

Similar behavior was observed for fluids viscosified

0.45
0.40
with hydroxyethylcellulose (HEC) (Torrest, 1982;
0.35 Roodhart, 1985a).
Velocity (ft/s)
0.30
0.25
0.20 100
0.15 = Concentric cylinder viscometer
0.10 = Capillary viscometer
0.05
0
0 0.10 0.20 0.30 0.40 0.50
Viscosity (Pa·s)
Distance across 0.5-in. gap (in.) Temperature = 70°F
10–1
Figure 8-10. Laser doppler velocimetry plot of a 35 lbm
guar/1000 gal fluid at a flow rate of 40 gpm in a slot-flow
geometry (Lear, 1996).
transport. Fracture width is related to the viscosity at 10–2

the wall whereas proppant transport is related to the 10–2 10–1 100 101 102 103 104
viscosity gradient across the slot (Guillot and Shear rate (s–1)
Dunand, 1985; Kirkby and Rockefeller, 1985;
Figure 8-11. Effect of shear rate on the viscosity of an
Roodhart, 1985a; Acharya and Deysarkar, 1987). HPG aqueous solution (0.48%) (Guillot and Dunand,
Low-shear viscosity is also important for understand- 1985).
ing the cleanup of fracturing fluids within proppant
packs (Penny and Jin, 1995; Pope et al., 1994). These
effects are discussed in Section 8-7.
To better predict the full range of fracturing fluid 2.00
viscosity, a rheology model must utilize not only n
and K but also a zero-shear viscosity term. The Ellis 1.75
model adds zero-shear viscosity µ0 to the power law
+
model to improve viscosity prediction: +
+
+
+
1.50
+
+
1 1 1
= + , (8-9)
+
+
µ a µ 0 Kγ n−1 1.25
+
Reduced velocity
+
+
where n and K are defined from the high-shear data

1.00 +
+
(i.e., power law region; Fig. 8-8). +

+ +
It is difficult experimentally to obtain both µ0 and

+
0.75
the high-shear viscosity behavior because of the wide
+
Shear
+
range of torque or pressure drop values that an instru- ID (mm) rate (s–1)
+
ment must be capable of accurately measuring. 0.50 = 10 2.1

+
+
= 10 13.3
Guillot and Dunand (1985) used LDV to determine = 10 80.6
+
+
= 10 353
+
the velocity profile of HPG gels flowing in a tube, 0.25

+
which in turn was used to calculate the fluid viscosity.

+
To verify the method, their first step was to compare 0

viscosities measured in a concentric cylinder geometry –1.00 –0.50 0 0.50 1.00
viscometer and their capillary viscometer (Fig. 8-11). Reduced diameter
The agreement was good, and the results showed a
deviation from power law behavior at shear rates Figure 8-12. Measured and predicted velocity profiles of
an HPG solution (0.48%) (Guillot and Dunand, 1985).
below 10 s–1. Guillot and Dunand found that the pre-
dicted data from the Ellis model fit the experimen-
tally determined velocity profiles very well (Fig. 8-12).

8-6.4. Determination of fracturing

680
fluid rheology 640 Shear in casing
600 KGD model
• Shear history and temperature conditioning
During hydraulic fracturing treatments, the fluid 400
360
Shear rate (s–1)

experiences wide variations in shear and tempera- 25% of fluid pumped
320
ture. The fluid properties in the fracture depend on 280
the chemistry in the formulation and also the his- 240
tory of shear and temperature to which the fluid 200
50%
has been exposed, from initial mixing through final 160
75%
120
disposition in the fracture. High shear is experi- 80
100%
enced by the fluid during pumping through the 40
tubulars and perforation tunnels. Once in the frac- 0
ture, the shear on the fluid is significantly less, but 0 60 120 180 240 300 360 420 480 540 600
Distance along fracture (ft)
the fluid temperature increases until it eventually
reaches formation temperature. 4.2 37 74 111 148
Examples of a shear and temperature history Time (min)
experienced by a fluid for a specific set of condi-

Figure 8-13. Shear rate profile during a fracture treatment
tions and the KGD fracturing model (Table 8-5) (Table 8-5) (Worlow, 1987).
are illustrated in Figs. 8-13 and 8-14, respectively.
The shear rate experienced by a specific volume
of fluid increases with time. The formation temper- 260
Table 8-5. Input data for shear and temperature 220

Temperature (°F)
simulation (KGD model).

180
Depth to perforations 8000 ft
% of
140
Fracture gradient 0.70 psi/ft fluid pumped
25%
Porosity 10% 100 50%
75%
Permeability 0.10 md 100%
60
Well spacing 320 acres 0 120 240 360 480 600
Reservoir pressure 4000 psi Distance along fracture (ft)
Fracture height 300 ft
Figure 8-14. Temperature profile during a fracture treat-
Shear modulus 2.4 × 106 psi
ment (Table 8-5) (Worlow, 1987).
Spurt-loss coefficient 0 gal/ft2
Total fluid-loss coefficient 0.002 ft/min1⁄2
Ambient surface temperature 80°F [26°C] ature gradient experienced by the fluid as it enters
Reservoir temperature 200°F [95°C] the fracture is highest at the beginning of the treat-
ment and decreases thereafter. As a result, the dis-
Fluid flow behavior index n 0.50
tance the fluid travels down the fracture before it
Fluid consistency index K 0.02 lbm-sn/ft2
reaches maximum temperature increases with the
Fracture treatment volume 250,000 gal total fluid volume pumped (see Section 10-2.9).
Pump rate 40 bbl/min These are typical trends in the shear and tempera-
Casing ture conditions experienced by a fluid. Changing
OD 5.5 in. the treatment conditions or the fracturing model
ID 4.67 in. can significantly change the predicted values. For
Shear rate in casing 647 s–1 example, if the height is reduced by 75% in the
Time at shear in casing 4.24 min example in Table 8-5 (from 300 to 75 ft), the wall
shear rate would increase by a factor of 4.

Measurement of the rheological and leakoff

160
properties of a fracturing fluid under all conditions 675 s–1, 0 min
that approach realistic shear and temperature con- 140 675 s–1, 3.3 min
675 s–1, 6.6 min
ditions can be an expensive and time-consuming 1350 s–1, 6.6 min
Viscosity (mPa·s at 170 s–1)

120
operation. The extent to which field conditions
must be simulated depends on the composition 100
of the fluid and the intended use of the data. For
80
example, fluids that do not show time or shear
effects on their properties require less elaborate 60
equipment and testing procedures and are easier to 40
scale up to field conditions than fluids that have a
strong dependence on shear history. Also, if the 20
intent is to study the effect of shear and tempera- 0
ture on the fluid and to optimize fluid composition 0 10 20 30 40 50 60 70 80
for maximum stability, different testing methods Time (min)
may be required than what is appropriate for gen-
Figure 8-15. Effect of shear history during crosslinking of
erating rheological data strictly for fracture design fast-reacting titanate crosslinker (Prud’homme et al., 1989).
purposes.
Because of the sensitivity of the rheology of
crosslinked fluids to their shear and temperature diameter capillary viscometers also was found to
history, considerable efforts have been made to enhance the reproducibility (Prud’homme, 1986).
develop industry standard testing methods that will Several other references (Fan, 1992; Prud’homme
enable comparison of fluid performance data from et al., 1988, 1989; Prud’homme, 1990) discuss
different laboratories. The API Work Group on these procedures in detail.
Fracturing Fluid Rheology recommended a Cameron et al. (1990) and Cameron (1993)
method that evaluates fluids under fixed tubing and studied the nonhomogeneous flow character of
fracture shear history conditions. This method does delayed crosslinked gels and proposed methods for
not intend to represent actual fracturing conditions improving the evaluation of fluids with these char-
nor necessarily provide data for fracture design acteristics.
purposes, but rather to establish baseline data for • Shear effects on fluid rheology
comparing fluid systems. Crosslinked fracturing fluids fall into two distinct
An example of the effect of shear history on typi- categories: gels with reversible crosslinks and gels
cal crosslinked fluid systems is shown in Fig. 8-15. with irreversible crosslinks. Both types of gels
The data were measured using the device illustrated may look equally “solid” or “liquid” in bulk, but
in Fig. 8-16. The technique maintains the fluid the reversibility of the crosslink sites makes pro-
under flowing conditions from the time the cross- found differences between the flow properties and
linker is added until the test is complete. The test microstructures of the gels.
procedure subjects the fluid to a high-shear envi- Shear affects the rheology of reversible and irre-
ronment to simulate flow down the tubing or casing versible gels differently. For guar crosslinked with
followed by a reduced shear rate and increased borate at a guar concentration of 40 lbm/1000 gal,
temperature to simulate fracture conditions. Cross- the polymer strand density is almost uniform in
linker is continuously metered into the fluid at the space because the concentration is 6 times higher
start of the high-shear conditioning. The equip- than the critical overlap concentration C* (see
ment for conducting the experiment varies from Section 8-5.2). For this state, shear history effects
small laboratory devices to large coiled-tubing are minor. For borate gels in the range of 20 lbm/
devices connected to slot-flow viscometers. These 1000 gal, the polymer chains are nearing the value
instruments generally provide reproducible results of C*. In this regime shear can destabilize the gel
as long as the crosslinker is carefully metered into and cause shear-induced phase separation. The
the fluid under controlled shear conditions. The fluid begins to separate into regions of high and
use of narrow-gap concentric cylinder or small- low concentration at a microscopic scale. With

Concentric
cylinder
viscometer
Shear History Simulator
Static mixer
∆p ∆p ∆p
fracturing
Crosslinker
Water
fluid
Additive
Tubular
Simulation
Oil Fracture
Simulation
High-pressure
syringe pumps
Triplex
fracturing
fracturing
Water
Water
fluid
fluid
pump Triplex Reciprocating
pump capillary
Oil viscometer Oil
Oil Oil
Figure 8-16. Tubing shear history simulator plus rotational or reciprocating capillary viscometers (Constien et al., 1986).
only minor segregation, the low-concentration cently formed bundles of approximately 25 to 50

regions fall below C*, homogeneous crosslinking molecular chains that aggregated into an open net-
is no longer possible, and the gel viscosity drops. work with pore sizes of 100–250 Å and the appear-
Because the borate crosslinks may reform over ance of felt. After the fluid was sheared, the gel
time, the gel will regain its initial viscosity if the was broken into fragments. Each fragment acted
shear is removed. However, the time scale for as a gel particle, and the rheology became that for
reequilibration may be on the order of hours. To a suspension of elastic spheres.
minimize shear-induced phase separation at poly- To minimize shear damage to the crosslinked
mer concentrations near C*, the fluid chemistry structure, delayed crosslinking systems are used to
can be adjusted to control the crosslinking rate prevent crosslinking during the high-shear-rate flow
during the time the fluid is exposed to high shear down the wellbore. However, even in the lower
in the tubing or casing. This approach can produce shear environment of the fracture, there are shear
uniform and highly viscous fracturing fluids at low rate effects on the ultimate viscosity (Fig. 8-17).
gelling agent concentrations. Fluids crosslinked in the example in Fig. 8-17 at
Shear also affects the rheology of irreversible shear rates between 60 and 87 s–1 show differences
titanate and zirconate gels differently. Shear during in viscosity of 40%.
the crosslinking reaction degrades the gel domains.
Under shear, irreversible gel networks break apart
into small domains or particles of crosslinked gel. 8-6.5. Rheology of foam and emulsion fluids
This has been demonstrated by freeze-fracture The viscosity of foams or emulsions is dependent
microscopy on sheared and unsheared titanate- on the volume fraction of the internal phase (known
crosslinked HPG gels (Zasadzinski et al., 1987). as the quality Γ), bubble- or drop-size distribution
The texture of a fluid that was crosslinked quies- and shape (known together as the texture), interfacial

10,000 40
Metal A 400 s–1
Metal B 800 s–1
Apparent viscosity (cp)

Metal C 30
Percent increase in apparent viscosity (at 170 s–1)
20
1000
10
0
0 10 20 30 40 50 60 70 80 90 100
Foam quality (%)
100 Figure 8-18. Apparent viscosity data for N2 foam with a

quality of 50% to 80% at 400- and 800-s–1 shear rates
(Cawiezel and Niles, 1987).
fluid. The change in bubble or drop size depends on

pressure, temperature and shear effects, making the
1 characterization of behavior for these multiphase sys-
10 100 1000 tems more difficult than for single-phase fluids.
Shear rate (s–1) In addition to the standard shear rate and shear
stress measurements for viscosity determinations, the
Figure 8-17. Effect of shear during crosslinking on appar- volume fraction of the internal phase, bubble- or
ent viscosity (constant shear on a Fann 39 viscometer for
20 min at 80°F [26°C]) (Conway et al., 1980). drop-size distribution and interfacial tension should
be characterized and reported for foams and emul-
sions. These data are required to adequately compare
tension and rheological properties of the continuous different fluid systems and aid in understanding fluid
phase (see Sections 7-5.1 and 7-5.2). Increasing the behavior at various temperature and shear conditions.
quality of the foam or emulsion generally increases the Unfortunately, in most of the literature on foam or
viscosity until the concentration of the internal phase emulsion fluids, data for the bubble or drop size were
reaches a point at which the system is no longer sta- not obtained. A good example of test results that
ble. Figure 8-18 illustrates the effect of increasing include drop-size information in addition to pressure
quality on a simple nitrogen (N2) foam (Cawiezel and drop and flow rate data is by Harris (1985).
Niles, 1987). Without data relating to the bubble- or drop-size
The foam texture is dependent on several variables, distribution, accurately correcting the pressure drop
including the surfactant chemistry, mixing and flow data during flow for wall slip is difficult. During wall
conditions, and pressure (Cameron and Prud’homme, slip, the shear rate of the flow is concentrated and
1989). In general, the finer the bubble or drop size increased at the slip boundary, which can change the
and the narrower the size distribution, the higher the bubble or drop and result in a viscosity change for the
viscosity of the system becomes. fluid (Fig. 8-19) (Prud’homme and Khan, 1996).
Foams and emulsions have been characterized The equipment used to characterize a foam or
using several different rheological models, including emulsion should be selected so that the dispersed
Herschel-Bulkley, Bingham and power law (Reiden- drop or bubble size is less than 1⁄10 of the distance
bach et al., 1986; Harris and Reidenbach, 1987; between the measurement boundaries. This geometry
Mitchell, 1971; Wendorff and Ainley, 1981), and a limitation restricts the use of standard narrow-gap
wide variety of experimental methods. One of the dif- concentric cylinder viscometers for all but the smallest
ficulties that occurs during testing is that the bubble bubble- or drop-size systems. The majority of rheo-
size and distribution may change with the different logical characterizations have been conducted with
shear and temperature conditions experienced by the either capillary or pipe viscometers with the condi-

Table 8-6. Parameters for N2 foams

(Reidenbach et al., 1986).
r
Foam Quality (%) True Yield Point (lbf/ft2)
Γ ≤ 0.6 0.07Γ
Γ > 0.6 0.0002e 9Γ
True Foam Consistency Index (lbf-sn/ft2)

External Phase K C1 n
τo
Water 0.00002 3.6 1.0
10 lbm HPG/1000 gal 0.00053 2.1 0.75
Figure 8-19. Flow of foam in a tube. The foam slips at the
wall and is in plug flow at the center of the tube, where 20 lbm HPG/1000 gal 0.00256 1.7 0.607
the shear stress is below the yield stress of the foam τo. 40 lbm HPG/1000 gal 0.0152 1.2 0.45
r = tube radius.
Table 8-7. Parameters for CO2 foams

tions made to match the actual fracturing treatment (Reidenbach et al., 1986).
conditions of shear, temperature and pressure as
closely as possible. The viscometers are high-pres- Foam Quality (%) True Yield Point (lbf/ft2)
sure circulating systems (Reidenbach et al., 1986; Γ ≤ 0.6 0.042Γ
Harris and Heath, 1996) or single-pass pipe viscome- Γ > 0.6 0.00012e 8.9Γ
ters (Harris, 1985; Cawiezel and Niles, 1987; Harris
True Foam Consistency Index (lbf-sn/ft2)
et al., 1995).
One of the more extensive studies of N2 and External Phase K C1 n
carbon dioxide (CO2) foams was conducted by Water 0.00002 4.0 1.0
Reidenbach et al. (1986). A model for calculating 10 lbm HPG/1000 gal 0.00053 2.6 0.75
rheological properties based on the foam quality Γ, 20 lbm HPG/1000 gal 0.00256 2.2 0.607
yield point τyp, and n and K for the liquid phase was
40 lbm HPG/1000 gal 0.0152 1.0 0.45
proposed. The basic equation is
n
∆pDp  8v  geometry-independent approach for the prediction of
= τ yp + K foam   , (8-10)
4L  Dp  foam properties in pipes at different conditions:
where τyp is related to the foam quality and gas com- vˆ ρl
εS = = , (8-12)
position, Dp and L are the pipe diameter and length, vˆl ρ
respectively, and ∆p is the pressure drop. The consis-
tency coefficient Kfoam was found to be dependent on where the specific volume expansion ratio εS is
the liquid-phase consistency coefficient K and quality: defined as the ratio of the specific volume of the
foam v̂ to the specific volume of the base liquid v̂l and
K foam = Ke
( C Γ +0.75 Γ )
1
2
, (8-11) is also equal to the ratio of the liquid density ρl to the

foam density ρ. In terms of the VEPL model, the
where the constant C1 varies with the external phase. effective viscosity is (Enzendorfer et al., 1995)
Typical values for τyp, C1 and K are listed for N2 and
1 6n + 2 
n
CO2 foams in Tables 8-6 and 8-7, respectively. µ eff = Kε1S−n u n−1 D1−n  . (8-13)
One of the difficulties that typically occurs in relat- 8 n 
ing laboratory-generated data on foams to field-scale Foams with qualities from 30% to 75% were gen-
tubing is the effect of the pipe diameter on the result- erated in a 40-ppg HPG solution and the pressure was
ing stress, which may result from texture differences measured in a series of pipe diameters. Figure 8-20
near the pipe wall. Winkler et al. (1994) proposed illustrates the shear stress versus volume-equalized
using a volume-equalized power law (VEPL) model shear rate for these foams pumped through four dif-
as a constitutive flow behavior equation to provide a ferent pipe diameters.

100 400
Pipe
diameter (m)
Peak viscosity (cp at 170 s–1)

0.0495 75% quality
0.0407 300 50% quality
0.0338
0.0213 25% quality
τ/εS (Pa)
200
10
100
0
1 50 100 150 200 250 300
100 1000 10,000 Temperature (°F)
γ/εS (s–1)
Figure 8-21. Curve fit of borate-crosslinked foam with
Figure 8-20. Volume-equalized wall shear stress versus 30 lbm/1000 gal guar.
volume-equalized wall shear rate of 40-ppg HPG foams of
quantities from 30% up to 75% at pipe inlet conditions of
different pipe sections and a temperature of 68°F [20°C]
(Winkler et al., 1994). 800
700
Peak viscosity (cp at 170 s–1)
25% quality
600 50% quality
Measuring the rheology of crosslinked foams
500 75% quality
involves many of the same considerations discussed
in Section 8-6.4 for single-phase crosslinked fluids. 400
Because crosslinked foams are generally used to 300

transport higher proppant concentrations at higher 200
formation temperatures than foams prepared from 100
linear gels, crosslinked foams should be measured
0
at the temperatures and pressures that represent these 50 100 150 200 250 300
conditions (Watkins et al., 1983). The temperature Temperature (°F)
and pressure conditions are important for correctly
simulating the quality and texture. Figure 8-22. Curve fit of borate-crosslinked foam with
Harris and Heath (1996) used a recirculating flow- 40 lbm/1000 gal guar.
loop viscometer to measure borate-crosslinked N2
foams. The viscosity of the borate-crosslinked foams
was predicted by linear regression analysis with a 8-6.6. Effect of viscometer geometry
temperature adjustment factor: on fluid viscosity
borate foam viscosity = mo To (1 + To c) + µ100 , (8-14) In addition to testing fluids with shear and tempera-
ture history simulation, understanding the effect of
where mo = –0.5, To is the offset temperature (i.e., the the viscometer geometry on the inferred viscosity of
actual temperature minus 100°F), the factor c = 80, the fracturing fluid is important. For rheological mea-
and µ100 is the viscosity at 100°F [40°C]. Figures surements in concentric cylinder viscometers, the
8-21 and 8-22 are curve fits for foamed borate-cross- more narrow the gap between the rotor and the bob,
linked fluids by this technique. the more uniform the shear field for non-Newtonian
The viscosity for borate-crosslinked foamed fluids fluids. Harrington et al. (1979) illustrated the effect
was found to increase with the gas quality and poly- on the shear rate across the gap of a rotational vis-
mer concentration, with a 3- to 10-fold increase over cometer as a function of the behavior index n. Fan
linear foam of the same polymer concentration. and Holditch (1993) reported the effects of gap width
on the inferred viscosity of several fracturing fluids.

As the gap width increased, the shear stress variation πω

γb = , (8-16)
in the gap increased and the average or effective   r  2 n′ 
shear rate on the fluid decreased as a function of the 15n ′ 1 −  bob  
  rcup  
width and the behavior index n´ of the fluid. Figure  
8-23 illustrates the effect on crosslinked fracturing
fluid viscosity of two different bob sizes in a Fann where ω is the viscometer angular velocity. Equations
50 viscometer. 8-15 and 8-16 can be applied to Newtonian fluids
Comparison of data gathered using the same when n´ = 1.
Newtonian wall shear rate in wide- and narrow-gap Figure 8-24 illustrates use of the volumetric aver-
viscometers finds that the inferred viscosity of the age shear rate method to test a titanate-crosslinked
fluid in the wide-gap viscometer is higher because the fracturing fluid (Fan and Holditch, 1993).
effective shear rate is lower. Data can be determined
independent of the gap width if the experiments are
200
performed using an equivalent volume-averaged
shear rate. The volume-averaged shear rate for a
power law fluid is determined as follows: 180
 rcup  2−2 n′ 
n′ 160
γv = γb   − 1 , (8-15)
 rcup  2   rbob  
(n′ − 1)  − 1
 rbob  
140
where rcup is the cup radius, rbob is the bob radius, 120
and γb is the shear rate at the bob:
100
EX-B2: VASR = 59.3 s–1, rpm = 265
300
EX-B2: VASR = 71.6 s–1, rpm = 323
80 EX-B2: VASR = 85.1 s–1, rpm = 380
EX-B5: VASR = 85.6 s–1, rpm = 118
250 60
0 20 40 60 80 100 120 140

200 Time (min)
Figure 8-24. Comparison of viscosity measurements for

delayed titanate HPG gel using the volumetric average
150 shear rate (VASR) method at 150°F [65°C] (Fan and
Holditch, 1993).
100 8-6.7. Characterization of fluid micro-

EX-B2 bob structure using dynamic oscillatory
EX-B5 bob
measurements
50 Shear rheology measurements provide the parameters
0 20 40 60 80 100 120 140 n and K. This information is necessary for optimizing
Time (min)
fluid stability and viscosity and for predicting laminar
Figure 8-23. Viscosity measurements of delayed titanate
flow behavior in fractures. Additional information
HPG gel using a nominal shear rate of 100 s–1 at 200°F concerning the microstructure of the fluid can be
[95°C] and two bob sizes (Fan and Holditch, 1993). obtained using dynamic oscillatory rheological mea-
surements. Dynamic oscillatory measurements may
be used to evaluate the viscous, linear elastic and

nonlinear elastic (i.e., normal force) properties of constant of the fluid can be obtained. The importance
polymer fluids. The linear elastic properties, which of the time constant is that it can be related to the
are known as the elastic storage modulus G´ and vis- shear rate for the fluid by the dimensionless Deborah
cous loss modulus G´´, are sensitive to changes in number:
polymer structure that may not be observed with
λ
steady shear measurements. For example, Prud’- De = , (8-17)
homme (1986) used dynamic oscillatory measure- t
ments to determine differences in the extent of where λ is the characteristic relaxation time and t is
hydration of guar mixed by two different methods, the characteristic flow time, which is proportional to
although the steady shear viscosity measurements 1/γ. If De > 1, then the relaxation time of the polymer
were the same. Similarly, Knoll (1985) studied the gel is longer than the flow time. In this case, the fluid
differences in crosslink density and wall-slip tenden- can not relax and behaves more like an elastic solid
cies of fluids that had been dynamically mixed. than a fluid. As a solid it slips at interfaces rather than
Several other studies have used dynamic mechanical flowing as a continuous fluid. The inferred viscosity
testing to obtain information on such properties as of a “slipping” fluid depends on the measurement
crosslinking kinetics and the effects of elasticity on geometry and wall roughness, and hence does not
proppant suspension capabilities and flow properties reflect the actual fluid composition and process con-
(Clark, 1979; Acharya, 1986, 1988; Acharya and ditions such as fluid temperature but is an artifact of
Deysarkar, 1987; Menjivar, 1984; Prud’homme, 1985; the testing device.
Clark and Barkat, 1989; Clark et al., 1993; Power et Measurements from multiple geometries are neces-
al., 1994; de Kruijf et al., 1993). sary to separate the effects of slip from the true fluid
viscosity (Kesavan et al., 1993). For shear rates such
that De < 1, the fluid will display normal shear thin-
8-6.8. Relaxation time and slip ning viscosity.
The relaxation time of fracturing fluids defines the
critical shear rate for reversible behavior. This time 8-6.9. Slurry rheology
can be determined from dynamic oscillatory measure-
ments. Dynamic oscillatory moduli as a function of Fluids containing proppant generally account for 20%
temperature are shown in Fig. 8-25. If the moduli are to 80% of the total volume of a fracturing treatment,
fitted to a Maxwell model (i.e., a rheological model yet little rheological data exist for these slurries.
for a linear viscoelastic fluid with one time constant Determining the rheology of fracturing slurries is a
and one elasticity constant), the relaxation time considerable problem because of the dependence on
fluid composition, flow geometry, temperature, time,
and proppant size, density and concentration. The
101
+++
majority of instruments used to determine the rheo-
+++++++ +++++ logical properties of “clean” fluids are not suited for
+ + + ++
+ these studies because their geometry does not accom-
++
100 modate the distance between the flow boundaries (i.e.,
G′ (Pa)
the gap or slot should be >10 times the particle diam-

eter) or high concentrations (i.e., up to 20 ppa). Also,
the proppant must be kept in uniform suspension.
10–1 The common means for characterizing fracturing
65°C + 25°C
55°C 20°C fluid slurries are large slot-flow devices, “wiped-disc”
45°C 15°C
35°C 10°C
concentric cylinder viscometers and wide-gap con-
10–2 centric cylinder viscometers. Examples of slurry
10–1 100 101 102 viscosity data are presented in Figs. 8-26 and 8-27
ω (rad/s) (Ely, 1987).
Gardner and Eikerts (1982) used a large closed-
Figure 8-25. Storage modulus G′ of 48% by weight HPG
crosslinked by 0.12% by weight borax gel at a pH of 9.15 loop pipe viscometer to study crosslinked aqueous
(Kesavan et al., 1993). fracturing fluids containing proppant in laminar flow.

1.0 10
0.8
Shear stress (lbf/ft2)

0.6
n′
0.4 1
0.2
Base gel (no sand)
0 1.25 lbm/gal 20/40
0 1 2 3 4 5 6 7 8 2.7 lbm/gal 20/40
4.82 lbm/gal 20/40
Sand concentration (lbm/gal) 6.11 lbm/gal 20/40
0.1
500 100 1000 10,000
400 Nominal shear rate (s–1)
300
K′ (dynes-sn′/cm2)
Figure 8-28. Effect of proppant on fluid viscosity (Gardner

200 and Eikerts, 1982).
that the apparent viscosity of the fluid at 511 s–1 with

100
proppant was 2.7 times that predicted by Ford (1960)
for a Newtonian fluid with an equivalent proppant
concentration. These large effects of proppant are
50
0 1 2 3 4 5 6 7 8 contrary to those expected for shear-thinning fluids
Sand concentration (lbm/gal) (Nolte, 1988b).
In a more extensive work that considered both lam-
Figure 8-26. Flow behavior index n′ and fluid consistency inar and turbulent flow, Shah and Lee (1986) studied
index K′ as functions of sand concentration (Ely, 1987).
the relation of friction pressure to proppant concen-
tration and size in four different HPG base fluids in
different pipe sizes and correlated the laboratory pre-
104
dictions with field measurements. Shah and Lee’s
studies illustrate the complexity of characterizing
Apparent viscosity (cp)
slurry rheology. The friction pressure of fluids con-

taining proppant was found to increase with increas-
103 ing proppant concentration (Figs. 8-29 and 8-30). The
predicted amount of friction pressure increase dimin-
ished with turbulent flow rates (e.g., >12 bbl/min).
A relation for the increase in friction pressure in
Brookfield turbulent flow resulting from the presence of prop-
Pipe viscometer
102 pant is
0 1 2 3 4 5 6 7 8
∆p friction = (µ r ) (ρr )
m (1– m )
Sand concentration (lbm/gal) , (8-18)
Figure 8-27. Apparent viscosity measured with two types
of viscometers as a function of sand concentration (Ely, where ∆pfriction is the friction pressure ratio with and
1987). without solids, µr is the ratio of the apparent slurry
viscosity to the apparent fluid viscosity, ρr is the ratio
They found that the apparent viscosity of a cross- of the slurry density to the fluid density, and m is
linked fracturing fluid increased up to 230% with the the log-log slope of the friction plotted versus the
addition of 6 ppa (Fig. 8-28). Their data also indicate Reynold’s number. Hannah et al. (1983) used m =
0.2 for the equation.

50 8-7. Proppant effects

Predicted
In addition to the change in apparent viscosity with
8 lbm/gal the addition of proppant to the fluid, an equally
40 important factor to determine is the possibility of
6 lbm/gal interaction of the fluid chemistry with the proppant
to change the stability or maximum viscosity that the
Pressure drop (psi/100 ft)
system can generate. This is particularly the case with

30 resin-coated proppants. Methods for evaluating fluid
and proppant interactions were described by Norman
et al. (1990) and Nimerick et al. (1992).
4 lbm/gal
20
8-7.1. Characterization of proppant

2 lbm/gal
sand concentration transport properties
10
Measuring the ability of a fluid to transport proppant
is one of the more difficult tasks in fracturing fluid
40-lbm HPG base gel
characterization. Many factors, such as fluid rheol-
ogy, fluid velocity and proppant concentration and
0
0 2 4 6 8 10 12 14 16 18 20 density, affect proppant settling rates.
Flow rate (bbl/min) Most of the experimental studies to characterize
proppant transport properties use one or more of the
Figure 8-29. Comparison of actual friction pressures at following three approaches:
various flow rates for 40 lbm HPG/1000 gal and sand-
laden 40 lbm HPG/1000 gal fluids in 2 7⁄8-in. tubing (Shah • measuring the rheological properties of clean fluid
and Lee, 1986).
and using these as the basis for predicting the
transport properties
70
Predicted • measuring proppant settling velocities in stagnant
fluids
60
• observing proppant transport in slot-flow devices,
Percent friction pressure increase over base gel
flow loops or concentric cylinder devices.

50 Particle settling velocities have been measured in
Sand concentration a variety of experimental devices. Kern et al. (1959)
and Babcock et al. (1967) studied the flow and depo-
40 8 lbm/gal
sition of sand slurries in a vertical slot-flow model.
Schols and Visser (1974) and Sievert et al. (1981) also
6 lbm/gal
30 used a vertical slot-flow model to develop equations
for both the height and length of deposition beds.
4 lbm/gal Clark et al. (1977, 1981) and Sievert et al. (1981)
20 used a large vertical slot-flow model to study the set-
tling of particle clusters in non-Newtonian fluids.
2 lbm/gal Novotny (1977), Hannah et al. (1978), Harrington
10
et al. (1979) and Clark and Guler (1983) used a con-
centric cylinder device to study proppant settling
0 velocities. Novotny’s study included wall effects,
0 4 8 12 16 20 24 28 concentration effects and shear effects on a series
Flow rate (bbl/min) of Newtonian and non-Newtonian fluids. Hannah et
Figure 8-30. Effect of sand concentration on friction pres- al. and Harrington et al. used two different concentric
sures of 40 lbm HPG/1000 gal fluid in 2 7⁄8-in. tubing at vari- cylinder devices to study noncrosslinked and cross-
ous flow rates (Shah and Lee, 1986).

(1+ n ) n
linked fluids. The results of both studies indicate that ut ∝ d prop . (8-20)
the settling of particles in the fluids deviated from
Stokes law settling velocities. Clark et al. (1981) and For Newtonian fluids, the relation is
Quadir (1981) reported results from a study using
ut ∝ d prop
2
. (8-21)
both a parallel-plate device and a concentric cylinder
device. They also reported significant deviations from Equation 8-19 can be used to determine only single-
Stokes law settling velocities. particle fall rates over shear ranges in which the fluid
Flow loops with particles suspended in a vertical follows power law behavior. As shear rates approach
section were used by Shah (1986). very low or very high values, limiting values of the
More recently, proppant transport has been evalu- apparent viscosity are reached. In actual treatments,
ated in large slot-flow devices at commercial testing values of high-shear-limiting viscosity µ∞ are not
laboratories (Barree and Conway, 1994; Zhao et al., approached in the fracture. However, at the center of
1995) and universities (Clark and Zhu, 1995a; Shah the fracture channel the shear rate is zero and the
and Subramanian, 1997; Goel et al., 1997; Hejjo et fluid viscosity approaches the value for zero-shear
al., 1997; Shah et al., 1998; Shah and Asadi, 1998). viscosity µ0 (Fig. 8-10). Roodhart (1985b) and other
The transport efficiency in these tests is usually deter- investigators determined that low-shear viscosity
mined by visual observation of proppant fall, and plays a considerable role in proppant transport during
comparisons are made to the performance of other flow conditions (Kirkby and Rockefeller, 1985). To
fracturing fluids under similar test conditions. correct the limitation of the power law model to
When the results of the experimental measure- describe flow fields where the shear rates approach
ments are translated to predictive models, Stokes set- zero, Slattery and Bird (1961), Turian (1967),
tling law for a Newtonian fluid (see Eq. 6-106) is Dunand et al. (1984) and Roodhart (1985b) studied
most frequently used as an initial frame of reference. the Ellis fluid model, which incorporates a µ0 term
Most fracturing fluids are non-Newtonian, with (Eq. 8-9).
fluid viscosity decreasing as shear increases. Novotny Combining Eq. 8-9 with Stokes law leads to a set-
(1977) determined that the most important variables tling velocity equation of the form
affecting proppant settling are the non-Newtonian
1− n′
characteristics of the fluid, wall effects and proppant g∆ρd prop
2
g∆ρd prop
2
 ut 
ut = + , (8-22)
concentration. For a non-Newtonian transporting 18µ 0 18 K ′  d prop 
fluid, Novotny found that the shear rate within a fluid
had a significant effect caused by its reduction in the where ∆ρ = ρp – ρf and the shear rate induced by par-
apparent viscosity affecting particle settling. ticle settling is assumed to be (Novotny, 1977)
A modified version of Stokes law for the terminal ut
settling velocity ut of a single particle in a quiescent γ= (8-23)
d prop
power law fluid is
(versus 3 times this value in Eq. 8-19).
( )
1n
( n +1) 
 ρ p − ρ f gd prop Roodhart (1985b) tested the validity of Eqs. 8-9 and
ut = K   , (8-19)
 18 K (3)
( n −1) 8-22 under static and flowing conditions. He found

excellent correlation with experimental data for the
where ρp and ρf are the densities of the particle and apparent viscosity versus shear rate (Fig. 8-31) and
the fluid, respectively, g is the gravitational accelera- static settling velocities for glass and steel particles
tion, and dprop is the particle diameter. This equation (Fig. 8-32). Roodhart also found that the shear from
and other forms of Stokes equations in this section the horizontal fluid flow could be ignored for fracture
assume that the particle Reynold’s numbers are less shear conditions from 0 to approximately 25 s–1 and
than 2. that an extended power law model that includes a
With Eq. 8-19, the particle settling velocity becomes term for zero-shear viscosity was applicable for deter-
a function of the fluid parameters n and K. The set- mining proppant settling rates within the fracture for
tling velocity in power law fluids is proportional to pump rates for these wall shear conditions.

Another approach for predicting proppant settling

100 1000
Fit with Eq. 8-23 rates from rheological measurements was reported by
Experimental curves Acharya (1986), Shah (1986) and Meyer (1986b).
This approach correlates a generalized drag coeffi-
Apparent viscosity (poise)

Apparent viscosity (Pa·s)
10 100
cient defined as
IV V
III
4 gd prop  ρ p − ρ f 
Cdrag = (8-24)
3 ut2  ρ 
1 10
II
and a generalized particle Reynold’s number NRe:
0.1 I 1
d proput ρ f
N Re = , (8-25)
µ
0.1 1 10 100 1000 where µ is equal to K(γ)n – 1 for power law fluids in
Shear rate (s–1) laminar and turbulent flow. Using the definitions for
Cdrag and NRe, a correlation of particle settling veloci-
Figure 8-31. Comparison of experimental data (Table 8-8) ties and fluid properties was made by a generalized
with values predicted from the extended power law
(Eq. 8-9) (Roodhart, 1985b).
plot of √Cdrag
2 – nN 2 versus N . This relation results in
Re Re
a family of curves that are functions of the behavior
index n (Fig. 8-33). The significance of representing
x
data in this fashion is that they can be used to predict
10–1 x Power law (Eq. 8-8)
Extended power law (Eq. 8-9) x xx x particle settling velocities in other fluid systems if the
x
x particle density and size and the fluid properties n and
x x
Measured settling velocity (m/s)
xx
x K are known (Shah, 1986). This method has been
10–2 x x
xx
x reported to predict experimentally determined settling
x x x
x x velocities to within ±20% for a quiescent fluid
x x
x (Meyer, 1986b).
10–3 x
x
x
8-7.2. Particle migration and concentration

10–4
One of the factors that can influence proppant trans-
100% agreement port is the tendency of particles to migrate and con-
10–5 centrate in preferred areas of the flow field. Under
shear gradient conditions in a pipe or fracture slot,
10–5 10–4 10–3 10–2 10–1
proppant particles can move to the center of the fluid
Calculated settling velocity (m/s)
for viscoelastic fluids or toward the wall for non-
Newtonian fluids that are not viscoelastic. Central
Figure 8-32. Experimental versus calculated settling veloc-
ities in viscous fluids (NRe < 2) (Roodhart, 1985b). migration of the particles results in proppant settling
rates that are greater than expected (Nolte, 1988b).
Table 8-8. Data plotted in Fig. 8-31 (Roodhart, 1985b).
Fluid Type Gelling Agent Power Law Indicies Zero-Shear

Concentration n K (Pa⋅⋅sn) Viscosity
(kg/m3) (Pa⋅⋅s)
I Guar gum 3.6 0.52 0.33 0.1
II HEC 4.8 0.45 1.40 0.54
III HEC 7.2 0.37 4.0 2.0
IV Guar gum 9.6 0.29 8.5 4.2
V HEC 12.0 0.22 40.0 32.0

are a function of only the fluid viscosity, and they act

2 × 102
in the narrow spaces between particles. The net effect
102
is that the resistance encountered by a particle to
n = 0.281 movement in the suspension increases with the parti-
NRe
101 cle volume fraction.

2–n 2
0.427
On the average, particles migrate away from high-
√ Cdrag
0.47
0.553 concentration zones, where the frequency of the inter-
0.762
100 1.0 actions is higher, to low-concentration zones, where
the frequency of the interactions is lower. This type
2 × 10–2 of particle migration mechanism is referred to as
10–2 10–1 100 101 102 5 × 102
shear-induced self-diffusion, as observed by Gadala-
NRe
Maria and Acrivos (1980) and explained by Leighton
Experimental data and Acrivos (1987).
Newtonian 35 lbm HPC/1000 gal
20 lbm HEC/1000 gal 50 lbm HPG/1000 gal
If there is a concentration gradient in the suspen-
30 lbm HPG/1000 gal 80 lbm HPG/1000 gal sion caused by the migration of particles resulting
Open symbols: static data
Solid symbols: dynamic data from non-Newtonian or inertial effects, a net diffu-
sional flux will oppose the migration. Unwin and
Hammond (1995) used a phenomenological model
Figure 8-33. Relationship of √C2– n 2
dragNRe to n (Shah, 1986).
that considers all these effects simultaneously to
solve for the particle concentration profiles in con-
centric cylinder and slot-flow geometries.
There are different techniques for measuring parti-
cle migration in fracturing fluids of varying rheologi- 8-8. Fluid loss
cal properties under different flows. The effect of rhe-
ological properties of fracturing fluids on proppant Fluid loss to the formation during a fracturing treat-
migration has been studied in large slot-flow models ment is a filtration process that is controlled by a
by videotaping the particle positions in the gap of the number of parameters, including fluid composition,
slot. The slot width was divided into thin slices, and flow rate and pressure, and reservoir properties such
the number of particles traveling in each slice during as permeability, pressure, fluid saturation, pore size
a certain interval of time was counted from individual and the presence of microfractures.
videotape frames. Several controlling mechanisms can be involved
Tehrani (1996) reported particle migration experi- in limiting fluid loss, as discussed in Section 6-4.
ments in pipe flow. The slurry consisted of nearly Filtrate viscosity and relative permeability can control
spherical, transparent acrylic particles with a density fluid loss when their ratio is greater than that for the
of 1.180 g/cm3 dispersed in borate-crosslinked HPG reservoir fluid. The filtrate- (or viscosity-) controlled
fluids. A video camera with a variable shutter speed fluid-loss coefficient in ft/min1⁄ 2 is described by
was used to record images of the flow field in the k fil ∆pT φ
pipe. A vertical sheet of laser light illuminated the Cv = 0.0469 , (8-26)
µ fil
flow field. The rheological properties of the fluids
were measured, including shear viscosity and normal where kfil is the filtrate permeability in millidarcies
stress as functions of the shear rate and G´ and G´´ as into the saturated reservoir, ∆pT is the total differen-
functions of the frequency. Particle migration was tial pressure between the fluid in the fracture and the
found to be controlled by the elastic properties of the initial reservoir pressure in psi, φ is the formation
suspending fluid and the shear rate gradient. porosity (fraction), and µfil is the apparent viscosity
Particle concentration has the effect of increasing in cp of the filtrate flowing into the formation.
the frequency of interparticle interactions. The bulk The filtrate control mechanism is most likely in
viscous stresses that drive particles together are a effect when a gas reservoir is fractured with a non-
strong function of the suspension viscosity, which wall-building, high-viscosity fluid or for a formation
is a function of the particle volume fraction. The at irreducible water saturation. Also for non-wall-
lubrication forces that resist interparticle interactions building polymers, the apparent viscosity of the fil-

trate is lower than that in the fracture because of Zitha, 1995), resulting in lower leakoff rates. This
higher shear rates for the filtrate. The viscosity of the mechanism is called internal filter-cake formation.
filtrate from a wall-building fluid is generally much The rate of filtrate leakoff is specific to the particu-
lower than the viscosity of the fracturing fluid. lar fluid composition and formation properties and
The viscosity and compressibility of the reservoir must be determined experimentally, either in the lab-
fluid also help to control leakoff into the formation. oratory or by in-situ calibration treatments, as dis-
This control mechanism is most effective when the cussed in Section 9-5. The following provides an
reservoir fluids have high viscosities and are not introduction to the laboratory tests.
greatly compressible (e.g., heavy oil). The reservoir-
(or compressibility-) controlled fluid-loss coefficient
is calculated as follows: 8-8.1. Fluid loss under static conditions
kr φct Static fluid-loss tests for fracturing fluids have been
Cc = 0.00118∆pT , (8-27) used for many years (Howard and Fast, 1970; Rec-
µr
ommended Practice for Standard Procedures for
where kr is the reservoir permeability, ct is the com- Evaluation of Hydraulic Fracturing Fluids [RP 39],
pressibility of the reservoir in psi–1, and µr is the vis- 1983). A simple schematic of a static fluid-loss test
cosity of the reservoir fluid in cp. These equations for is shown in Fig. 8-34.
Cv and Cc differ from Eqs. 6-88 and 6-99, respec- In static fluid-loss tests, fluid is heated to the test
tively, because they are expressed in common refer- temperature, a differential pressure is applied across
ence to the total differential pressure.
Because it is possible for both filtrate and reservoir N2
effects to be factors in controlling fluid loss, it is
common practice to combine the two coefficients (see
Eq. 6-96):
2Cv Cc
Ccv = . (8-28)
Cv + (Cv2 + 4Cc2 )
1/ 2
A third fluid-loss control mechanism is in opera-

tion with wall-building fluids. When a wall-building
fluid is forced onto a rock, a thin film of material Fuid
(i.e., polymer, fluid-loss additives or both) is filtered
out and deposited on, or within the surface region of,
Core holder
the formation face surface. This thin film of material,
known as the filter cake, is less permeable than the Core
Seal
rock. Fluid loss can then become controlled by the
filter cake rather than by the rock (i.e., Cv or Cc).
Depending on the pore size of the formation, poly-
mers and fluid-loss additives in the fracturing fluid
may be filtered out on the formation surface or may
invade into the matrix and deposit by bridging and
adsorption mechanisms in the pore space. The depth
of invasion of the polymers and fluid-loss additives is
a function of their size, degree of entanglement and
ionic character. High deformation gradients resulting
from pressure differentials can cause polymer mole-
cules to transform from coiled conformations to a
stretched state, which enables them to flow into pore
sizes smaller than their largest dimension. If adsorp-
Filtrate
tion or plugging occurs in the pore throats, the poly-
mer chains form bridges in the pore throats (Pacelli Figure 8-34. Schematic of a static fluid-loss cell.

the core (usually 1000 psi), and the rate of filtrate vol-
70
ume forced through the core is measured versus time.
Cumulative fluid-loss volume (L × 10–3)

For wall-building fluids, the filter cake continues to 60
grow with time and the fluid-loss rate decreases. For
50
an ideal wall-building fluid, a plot of the filtrate vol-
Crosslinked fluid 80°F
ume versus the square root of time results in a straight 40
line (see Sidebar 5A). The slope of the straight line is 30
used to calculate the wall-building coefficient Cw and
the intercept is used to calculate the spurt loss Sp. The 20
value of Cw is directly proportional to the leakoff 10

velocity through the established filter cake. The spurt
0
value represents the fluid that leaks off during the for- 0 2 4 6 8 10 12 14 16 18 20
mation of an effective filter cake. By using the inter- Time (min)
cept to calculate spurt, the assumption is made that the
filter cake is instantaneously established. However, Shear rate Core
(s–1) permeability (md)
because a finite time or volume is required for an
123 0.19
effective filter cake to form, calculated values of spurt 82 0.21
only approximate the fluid-loss behavior during filter- 41 0.15
41 0.21
cake formation.
Cw is calculated in ft-min1/2 as
Figure 8-35. Cumulative fluid-loss data for different shear
0.0164 m rates (McDaniel et al., 1985). ∆p = 300 psi, core length =
Cw = , (8-29) 0.02 m.
A
where m is the slope of the leakoff plot in mL/min1/2 similar results for noncrosslinked HPG and for
and A is the area in cm2 of core exposed to fluid. borate- and transition-metal-crosslinked HPG (Gul-
From the y-axis intercept b, the spurt value is bis, 1983), if the test times were less than approx-
determined in gal/100 ft2: imately 30 min. Gulbis (1983) also observed that
shear rates greater than 80 s–1 in dynamic tests caused
24.4b transition-metal-crosslinked fluids to have higher
Sp = . (8-30)
A leakoff rates than noncrosslinked or borate-cross-
Most fluid-loss data are generated in static condi- linked fluids (Fig. 8-36). The shear rate at the fracture
tions, and these data may be misleading because the surface changes significantly (i.e., decreases) with
filter cake is allowed to grow without being subjected
to erosion of the flowing fluid along the fracture sur- 100
TM-XL, 1.02 Pa·s
face.
Fluid loss per 40.5 cm2 (mL)
TM-XL (delayed), 1.47 Pa·s

80 B-XL, 0.91 Pa·s
Non-XL, 0.06 Pa·s
60
8-8.2. Fluid loss under dynamic conditions
Filter-cake erosion and fluid degradation under condi- 40
tions of shear and temperature have been the subject
20
of considerable study. Studies through the mid-1980s
include Hall and Dollarhide (1964, 1968), Sinha 0
(1976), McDaniel et al. (1985), Gulbis (1982, 1983), 0 10 20 30 40 50 60 70 80
Penny et al. (1985) and Roodhart (1985a). The Time (min)
results from these studies show that dynamic filtra-
Figure 8-36. Effect of fluid composition on fluid loss (Gul-
tion tends to increase as the shear rate and tempera- bis, 1983). The borate-crosslinked (B-XL) system is in
ture increase (Fig. 8-35). Penny et al. found that equilibrium. These systems approach the noncrosslinked
dynamic fluid-loss tests conducted at 40 s–l produced (Non-XL) system when subjected to shear. TM-XL = transi-
tion metal crosslinked.
data similar to static test results. Gulbis (1982) found

time for a specific element of the fracture surface. shear rate that a rock segment sees at a fixed position
These considerations are addressed in the next sec- from the wellbore decreases with time as a result of
tion. the widening of the fracture width (Fig. 8-38).
The effect that a decreasing shear rate history has
on the rate of leakoff and filter-cake formation is
8-8.3. Shear rate in the fracture and illustrated in Fig. 8-39. The leakoff volume is plotted
its influence on fluid loss versus time for three different cores with permeabili-
During a fracturing treatment, fluid loss occurs under ties of 1, 10 and 62 md. The initial spurt-loss vol-
dynamic conditions. Prud’homme and Wang (1993) umes increase with core permeability. The after-spurt
proposed that the extent to which the filter cake data show that once the shear rate falls below about
grows in thickness is controlled by the shear stress 140 s–1, the slope of the fluid-loss curves (i.e., the
exerted by the fluid at the wall of the cake and the leakoff rate) drops, denoting a change in the leakoff
yield stress of the cake. behavior.
The cake stops growing when the fluid stress
becomes equal to the yield stress of the cake. It starts 350
to erode when the fluid stress is larger than the yield
300
stress of the cake. The yield stress of the cake Calculated
depends on the concentration and pressure gradient in 250
Shear rate (s–1)

Approximation
the cake, whereas the shear stress of the fluid is deter- 200
mined by its rheological properties and the shear rate 150
to which it is subjected.
100
Navarrete et al. (1996) compared dynamic and sta-
tic fluid loss using crosslinked guar gels. The data 50
show that the effect of shear increased leakoff by 0
increasing spurt and after-spurt leakoff. The extent 0 20 40 60 80 100 120 140
to which the gel was able to invade the core was con- Time (min)
trolled by the degree to which the crosslinked struc-
Figure 8-38. Shear rate history over a rock segment from
ture was broken by the shear. Also, the magnitude of Fig. 8-37 at 50 ft from the wellbore (Navarrete et al., 1996).
the shear rate determined the rate of leakoff after
spurt by limiting the extent of filter-cake growth.
The shear rate that the fluid experiences during a 16 350
fracturing treatment varies with distance from the tip 14 300
Cell 3 – 62 md
of the fracture and time, as shown in Fig. 8-37 for the 12
250
Shear rate (s–1)

test case in Table 8-5 at a pump time of 145 min. The
Volume (mL)
10
Cell 1 – 1 md 200
8
Cell 2 – 10 md 150
6
450 Shear rate
10% job 4 100
400
350 25% job 2 50
Shear rate (s–1)
50% job
300 0 0
75% job 100% job 0 20 40 60 80 100 120 140
250
200 Time (min)
150
100 Figure 8-39. Comparison of fluid loss for a 40 lbm/1000 gal
50 borate-crosslinked guar gel under static and dynamic condi-
0 tions (permeability = 0.5 md, pressure drop = 1000 psi,
0 100 200 300 400 500 temperature = 150°F) (Navarrete et al., 1996).
Distance along fracture (ft)
Figure 8-37. Shear rate profiles during a fracturing job

(fracture height = 300 ft, pump rate = 40 bbl/min, pump
time = 145 min) (Navarrete et al., 1996).

8-8.4. Influence of permeability and 8-8.5. Differential pressure effects

core length The effect of the pressure drop is significant on spurt
Permeability and core length are important variables in high-permeability cores. Parlar et al. (1995)
in fluid-loss tests on high-permeability cores, where reported a linear dependency of spurt with the pres-
the polymer and fluid-loss additives can invade the sure drop under static conditions in 1000-md, 5-in.
rock and plug pore throats deeper into the formation. long cores at pressure drops of 500 to 1500 psi. The
The length of high-permeability cores used in fluid- effect of the pressure drop on Cw in high-permeability
loss tests must be longer than the invasion zone cores was found to scale with ∆p0.5 or follow the
length of the polymer. Otherwise, polymer may pene- pressure dependency of an incompressible filter cake
trate the entire core length, and the measured spurt (see Eq. 6-82). This implies that the internal filter
values will be unrealistically high. Parlar et al. (1995) cake, which dominates the resistance to fluid loss
used 1000-md cores under static conditions to deter- in high-permeability cores, behaves incompressibly.
mine that spurt loss became insensitive at core The effect of the pressure drop on Cw under static
lengths of 5 in. and larger. For 500-md cores under conditions in low-permeability cores (i.e., where the
dynamic conditions, Navarrete and Mitchell (1995) filter cake is external) was found to scale with ∆p0.6
determined that most of the polymer accumulation for borate-crosslinked fluids and ∆p0.56 for zirconium-
occurred in less than about 1 in. crosslinked fluids at low pressure drops (100 psi < ∆p
< 500 psi). At higher pressure drops (600 psi < ∆p <
1400 psi), Cw ~ ∆p0.1 (Mayerhofer et al., 1991). If the
classic models for filter cake are considered in the
interpretation of these results, the resulting conclu-
sions are that as the pressure drop increases, the filter
cake becomes less permeable and Cw becomes insen-
sitive to the pressure drop. This indicates that the fil-
ter cake is compressible at those differential pres-
sures. Similar results were found by Nolte (1982,
1998a) and Penny et al. (1985).

◆
Fracture Evaluation Using
Pressure Diagnostics
Sunil N. Gulrajani and K. G. Nolte, Schlumberger Dowell
9-1. Introduction response after fracturing characterizes its hydrocar-

bon production potential. Finally, the versatility and
The hydraulic fracture design process requires sub- ease of use of these techniques lend their application
stantial information about the reservoir and the frac- to most field situations, enabling an assessment of
turing fluid. The effectiveness of a design is also the fracturing process, either on site and in real time
dependent on the quality of the required data. The or after completion of the treatment for the improve-
fracture design parameters can be inferred from wire- ment of future designs.
line logs and through laboratory fluid and core testing Figure 9-1 shows a recording of the bottomhole
procedures. The reliability of the data inferred from pressure measured during a fracture stimulation.
these methods is reduced by factors such as the scale Fluid and proppant were injected for approximately
of measurement, variability in the geologic environ- 3 hr, followed by an extended period of shut-in that
ment, assumption of an overly simplified fracture lasted for more than 18 hr. The pressure response
response and significant deviations of the test envi- was recorded for all facets of the fracture and reser-
ronment from in-situ reservoir conditions. Fracture voir response—the pumping period, pressure decline
parameter uncertainties can result in a suboptimal as the fracture closes, time at which the fracture
fracture, at best, or in a complete failure of the stimu- closes on the proppant and finally the asymptotic
lation procedure under the worst-case scenario. approach to the reservoir pressure. Of particular
An on-site procedure for predicting fracture interest to the fracture design process is the analysis
dimensions is desirable but extremely difficult to of pressure measured during a calibration treatment.
obtain. Numerous fracture-mapping techniques, such A calibration treatment, or a mini-fracture treatment,
as radioactive tracers, surface and bottomhole tilt- precedes the main fracture treatment and is per-
meters and various electromagnetic measurements, formed without the addition of proppant under con-
have been applied to infer fracture dimensions (see ditions that mimic the main treatment. The pressure
Section 12-1.1). The techniques, however, provide response for this test follows the sequence in Fig. 9-1.
only a limited amount of information (e.g., fracture Fracturing pressures during each stage of fracture
azimuth or wellbore height) that is available gener- evolution (i.e., growth, closing phase and after-
ally only after completion of the fracture treatment. closure period) provide complementary information
Although sophisticated microseismic measurements pertinent to the fracture design process. The frame-
have been developed to infer the created fracture
dimensions, their restricted spatial range of observa-
Bottomhole pressure, pw (psi)
Injection Shut-in
tion and expensive instrumentation currently limit 9000
Fracture Transient reservoir
widespread application. closing pressure near the wellbore
In contrast, pressure analysis during and after a 8000

Net fracture Fracture closes on
fracture treatment is recognized as a powerful tech- pressure proppant at well
7000 pw – pc
nique for developing a comprehensive understanding
Closure pressure
of the fracturing process. A recording of the well- 6000 pc = horizontal rock stress
bore pressures provides an inexpensive measurement Reservoir pressure
for fracture diagnostics. The specialized analysis of 5000

38 40 42 44 46 48 50 56 58
pressure provides a qualitative indicator of the frac-
Time (hr)
ture growth, as well as estimates of the primary frac-
ture parameters. An analysis of the reservoir Figure 9-1. Bottomhole fracture pressure history (Nolte,
1982, 1988c).

work for analyzing fracturing pressure and its variation pressure response during this period loses its depen-
during a calibration test, along with example applica- dency on the mechanical response of an open fracture
tions, is the focus of this chapter. and is governed by the transient pressure response
Pressure analysis is based on the simultaneous con- within the reservoir. This transient results from fluid
sideration of three principles that are central to the loss during fracturing and can exhibit either linear
fracturing process: material balance, fracturing fluid flow or a long-term radial response. Each of these
flow and solid mechanics or the resulting rock defor- flow patterns can be addressed in a manner analogous
mation. Section 9-3 describes these principles. to conventional well test analysis for a fixed-length
Material balance, or the conservation of mass, is conductive fracture, as discussed in Chapter 12. The
important for analyzing the pumping and closing after-closure period characterizes the reservoir’s pro-
phases of fracturing. It also provides a framework for duction potential. Its analysis enhances the objectivity
determining the fluid volumes and proppant schedule of otherwise uncertain preclosure pressure interpreta-
for a basic fracture design. Fluid flow and solid tion. The theoretical, operational and application
mechanics define the interaction between the fluid aspects for describing the pressure following fracture
pressure and the formation. The fracture width and closure are outlined in Section 9-6.
pressure relations are derived by combining these two Each fracturing phase, from fracture creation
principles. In addition, solid mechanics also intro- through the after-closure period, provides a sequence
duces the concept of closure pressure, which is the of complementary information pertinent to the frac-
reference pressure for fracture behavior. Closure pres- turing process. A comprehensive assessment thus
sure is the most important parameter for fracturing requires an integrated procedure that combines the
pressure evaluation. information provided by each phase. Sections 9-7 and
The pressure measured during pumping provides 9-8 present a generalized pressure analysis method-
an indication of the fracture growth process. The pri- ology that unifies the various interpretative, analytical
mary diagnostic tool for this period is the slope of and numerical analysis techniques discussed in this
the log-log plot of net pressure (i.e., the fracturing chapter.
pressure above the reference closure pressure) versus
pumping time. The slope of the log-log plot is used
to characterize the fracture geometry by combining 9-2. Background
the fundamental relations, as outlined in Section 9-4.
The log-log plot is complemented by the pressure Injection pressure has been measured for safety con-
derivative to identify complex fracture growth pat- siderations since the inception of hydraulic fracturing.
terns and the effects of proppant injection. The pump- Its importance for characterizing a fracture was rec-
ing phase additionally characterizes the formation ognized as early as 1954 by Harrison et al., who
pressure capacity, which is the fracture pressure above developed a relation between pressure and the frac-
which only limited fracture propagation occurs. ture volume. The potential importance of understand-
The pressure response during fracture closure is ing the fracturing pressure response was noted by
governed largely by the rate of fluid loss. The analysis Godbey and Hodges (1958). They concluded, “by
of pressure during this period estimates the fluid effi- obtaining the actual pressure on the formation during
ciency and the leakoff coefficient. These parameters a fracture treatment, and if the inherent tectonic
are determined from a plot of the pressure decline ver- stresses are known, it should be possible to determine
sus a specialized function of time, commonly referred the type of fracture induced.” Furthermore, they
to as the G-plot. This specialized plot provides the stated, “the observation of both wellhead and bottom-
fracturing analog to the Horner plot (see Section 2-1) hole pressures is necessary to a complete understand-
for well testing. Theoretical principles and example ing and possible improvement of this process.”
applications of G-function analysis are outlined in Development of two-dimensional (2D) fracture
Section 9-5. This section also describes simple analyt- models by Khristianovich and Zheltov (1955),
ical corrections that extend the basic pressure decline Perkins and Kern (1961) and Geertsma and de Klerk
analysis to nonideal fracture behavior. (1969) provided a theoretical means for estimating
The final fracturing pressure analysis pertains to the the fracture width and its dependence on the net pres-
evaluation of pressure following fracture closure. The sure. The early 1970s were characterized by regulated
9-2 Fracture Evaluation Using Pressure Diagnostics

gas prices in the United States and significant oil dis- Schlottman et al., 1981; Elbel et al., 1984; Morris and
coveries in the Middle East, which caused a lull in Sinclair, 1984; Cleary et al., 1993) and fracturing for
fracturing activity and hence a reduced interest in specialized applications (e.g., Smith, 1985; Bale et al.,
fracturing research. Renewed activity was fostered 1992). It is most effective when used to characterize
during the mid-1970s when massive hydraulic frac- fracture growth behavior during the early stages of
turing led to an emphasis on improved fracturing eco- field development. Variations of the traditional cali-
nomics. The fluid and proppant volumes utilized for bration test have also been proposed for specialized
these fracturing operations significantly increased, purposes, such as the step rate/flowback test to deter-
which increased execution costs. Successful fracture mine closure pressure (e.g., Felsenthal, 1974; Nolte,
execution was recognized as critical for economic 1982; Singh et al., 1985; Plahn et al., 1997), step-
field development. These developments spurred a down test to identify near-wellbore effects (e.g.,
renewed interest in understanding fracture pressure Cleary et al., 1993) and short impulse injection test
behavior for the effective design and analysis of frac- to obtain reservoir permeability (e.g., Gu et al., 1993;
ture treatments. The Appendix to Chapter 5 provides Abousleiman et al., 1994). Additional studies
a detailed discussion of these activities. (Mayerhofer et al., 1993; Nolte et al., 1997) extend
A numerical simulation describing the pressure fracture pressure analysis to the realm of well testing,
decline response during fracture closure was pre- wherein reservoir information typically obtained from
sented by Novotny (1977). Nolte (1979) presented a conventional well tests can be inferred from calibra-
fundamental analysis to estimate the fluid-loss para- tion treatments.
meters and the fracture length for the PKN fracture
geometry. This analysis was subsequently extended
to the other 2D fracture geometry models (Nolte, 9-3. Fundamental principles
1986a). Simonson et al. (1978) characterized height
growth into stress barriers. Nolte and Smith (1981)
of hydraulic fracturing
related trends on the log-log net pressure plot to the Three basic relations govern the hydraulic fracturing
evolution of the fracture geometry, and Clifton and process: fluid flow in the fracture, material balance or
Abou-Sayed (1981) generalized fracture modeling conservation of mass, and rock elastic deformation.
and the related pressure response to three-dimensional These relations are reviewed in Chapter 5, presented
(3D) hydraulic fracture models. in the context of fracture modeling in Chapter 6 and
Later developments attempted to address the non- reformulated in this section to facilitate the develop-
ideal conditions that commonly occur during field ment of pressure analysis techniques.
practice but had been excluded in these early studies.
Soliman (1986a) developed an analysis to consider
the effects caused by fluid compressibility and tem- 9-3.1. Fluid flow in the fracture
perature change during shut-in. A correction to incor- The fracture essentially is a channel of varying width
porate pressure-dependent fluid-loss behavior was over its length and height. The local pressure gradient
proposed by Castillo (1987). Nolte (1991) addressed within the fracture is determined by the fracturing
several nonideal conditions during injection as well fluid rheology, fluid velocity and fracture width.
as shut-in and provided techniques to diagnose the Equations governing fluid flow within the fracture
related pressure responses. A semianalytical approach can be derived using the principle of conservation of
to account for variation of the fluid-leakoff coefficient momentum and lubrication theory applied to a fluid
following the end of treatment as well as the effects traveling in a narrow conduit. The rheology of frac-
resulting from fluid flowback was developed by turing fluids is generally represented by a power law
Meyer (1986a). A decline analysis methodology that model (see Chapter 8) that incorporates the parame-
addresses a comprehensive list of nonideal factors ters K and n. In recognition that fluid flow within a
during fracture closure was proposed by Nolte et al. fracture is laminar for most fracturing applications
(1993). (Perkins and Kern, 1961), the global pressure gradi-
Pressure analysis has been applied since the early ent along the length of a fracture can be expressed as
1980s to improve fracture performance in a variety
dp Kvxn
of applications, including routine fracture design (e.g., ∝ , (9-1)
dx w 1+n

where vx is the average fluid velocity along the length conservation, or volume balance, to replace those
of the fracture and is defined in terms of the volumet- of mass conservation.
ric injection rate qi, fracture height hf and height-aver- Pressure analysis, irrespective of the propagation
aged fracture width w – . Material balance or conserva- model, is based on three expressions of material or
tion of mass suggests that vx is proportional to qi/w– h. volume balance. The first defines the treatment effi-
f
Equation 9-1 then becomes ciency η as the ratio of the volume of the fracture
n created at the end of pumping Vfp and the cumulative
dp K  qi 
∝ 1+2 n h  . (9-2) injected volume Vi:
dx w  f Vfp
η= . (9-4)
In the special case of a Newtonian fluid (n = 1 and Vi
K = µ, where µ is the fracturing fluid viscosity), The second expression states that at the end of
Eq. 9-2 reduces to pumping, Vi is equal to Vfp plus the cumulative vol-
dp µ  q  ume of fluid lost to the formation during pumping
∝ 2  i , (9-3) VLp:
dx w  wh f 
– h . is readily recognized as the aver- Vi = Vfp + VLp . (9-5)
where the term w f
age fracture cross-sectional area. Equation 9-3 is It follows from Eqs. 9-4 and 9-5 that
essentially Darcy’s law with the permeability propor-
– 2. VLp = (1 − η)Vi . (9-6)
tional to w
Equations 9-1 and 9-2 are formulated in terms of Finally, during any shut-in period ∆t, the volume of
the average velocity and implicitly ignore change in the fracture is
the fracture width over its height. The varying width Vf ( ∆t ) = Vfp − VLs ( ∆t ) , (9-7)
profile has an effect on the flow resistance relative to
the case of a constant-width channel, as discussed in where VLs(∆t) is the volume of fluid lost to the forma-
Chapter 6. The increase in the flow resistance is tion between the shut-in time and any time ∆t there-
accentuated during periods of fracture height growth after.
into barriers at higher stress. The varying width pro- At closure (i.e., ∆t = ∆tc) the volume of the fracture
file affects other physical phenomena that are highly is equal to the bulk volume of proppant Vprop that was
sensitive to the velocity (e.g., temperature profile and injected during pumping. At closure, Eq. 9-7
proppant distribution). becomes
Vfp = VLs ( ∆tc ) + Vprop . (9-8)
9-3.2. Material balance or Eliminating Vfp from Eqs. 9-5 and 9-8 gives
conservation of mass Vi − Vprop = VLp + VLs ( ∆tc ) . (9-9)
Fluid compressibility is neglected in this chapter for These material-balance relations are illustrated in
the purposes of simplicity and clarity. For water- and Fig. 9-2. Equation 9-9 simply states that for a cali-
oil-base fracturing fluids, fluid volume changes are of bration treatment in which no proppant is added
secondary importance to the elastic deformation of (i.e., Vprop = 0), all injected volume is lost at closure.
the fracture. The physical effects of pressure and tem- “Mathematical relations for fluid loss” in the
perature changes on the fracturing fluid in the well- Appendix to this chapter provides expressions for
bore could be significant for foamed fracturing fluids. VLp and VLs(∆t) that are based on the derivations pre-
In this case, using a direct measurement of the bot- sented by Nolte (1979, 1986a).
tomhole pressure and incorporating changes in the
wellbore volume during shut-in significantly reduce
errors that may be introduced by the assumption of 9-3.3. Rock elastic deformation
an incompressible fluid. The generally applicable
assumption of an incompressible fracturing fluid The principles of fluid flow and material balance are
therefore enables using simple expressions of volume coupled using the relation between the fracture width
and fluid pressure. The relation defines the fracture

The first relation predicts the width for a planar

Pumping, tp Closure, ∆t 2D crack (Sneddon and Elliot, 1946), with one
dimension infinite and the other dimension with a
Lost finite extent d. The second relation provides a similar
VLp
expression for a radial, or circular (also called penny-
Stored VLs(∆t) VLs(∆tc)
shaped), crack in an infinite elastic body (Sneddon,
Vfp
Vf(∆t) Vprop
1946). In both cases, the fracture width has an ellipti-
cal shape. The maximum width is proportional to the
product of the characteristic dimension (d for 2D
End of Pumping Vi = Vfp + VLp
cracks and R for radial cracks) and the net pressure
( p–f – σmin) and inversely proportional to the plane
Lost
VLp → CL, κ strain modulus E′ = E/(1 – ν2). The formation
Volume in
VLp
Poisson’s ratio ν and Young’s modulus E are typi-
Vi = qitp cally estimated from laboratory experiments using
Stored
Vfp → w, hf, L cored samples of the reservoir rock (see Chapter 3)
Vfp
or sonic logs (see Chapter 4), and E′ is preferably cal-
ibrated from pressure data (see Section 9-7.2). The
Figure 9-2. Fracture volume-balance relations. average width w – and maximum width w for the 2D
max
crack (i.e., fracture) are, respectively,
compliance. Linear elastic deformation of the reser-
voir rock is assumed for the fracturing process. The w=
(
π p f − σ min d ) (9-10)
linear elasticity assumption is justified because field- 2E′
scale fractures produce relatively small additional 4
stresses superimposed on the much larger in-situ wmax = w (9-11)
π
stresses, excluding possibly the more complex defor-
mation occurring in the fracture tip region (see and for a radial fracture:
Section 6-7.2). The rock deformation, or fracture
width, can be predicted using two classic relations for w=
(
16 p f − σ min R ) (9-12)
cracks in an elastic material of infinite extent that are 3πE ′
subjected to a constant internal pressure p–f, with an
external far-field confining stress σmin applied perpen-
dicular to the plane of the crack, as shown in Fig. 9-3.
2D Crack Plane View of

Radial Crack
wmax d
σmin wmax
σmin
pf
2(pf – σmin)d 8(pf – σmin)R

wmax = wmax =
E´ π E´
π 2
w= w w= w
4 max 3 max
Figure 9-3. Sneddon cracks for 2D and radial fractures.

3
wmax = w. (9-13) 9A. What is closure pressure?
2
The fracture closure pressure pc is defined as the fluid pres-
These relations indicate that the fracture has a sure at which an existing fracture globally closes. Mathe-
width greater than zero only if p–f > σmin. The fluid matically, for a linear relation between the fracture width and
pressure at which an idealized unpropped fracture pressure (i.e., Eq. 9-21), pc equals σmin, the minimum princi-
pal in-situ stress in the reservoir. Ideally, the value of σmin is
effectively closes is globally invariant in homogeneous formations. Reservoirs,
however, are commonly characterized by lithology variations
pc = σ min (9-14) and natural fissures. These cause σmin to become a local,
directional quantity. In this case, the choice of pc depends
and is termed the fracture closure pressure. For com- on the scale and orientation of the representative fracture
geometry.
mercial fracturing applications, pc is distinguished Closure pressure thus is a fracture-geometry-dependent
from σmin, which is a local, directional quantity. The quantity. For example, a micro-hydraulic fracture treatment
closure pressure approximates the average stress over creates a fracture with a limited height (≈5 ft) and hence pro-
vides an estimate of σmin only at that scale. Fracture propa-
the scale and orientation of the initial fracture height. gation over this limited dimension is described by a value of
It represents the stress that governs the propagation of pc that equals the measured σmin. A large-scale, “commercial”
fracture that initiates from the same small interval, however,
a fracture over this scale of interest, as discussed in will grow well beyond the limited height dimension. During
Sidebar 9A. Field practices for estimating pc are this process, it certainly will cross various heterogeneities
prior to establishing coverage over the primary or gross inter-
described in “Estimating closure pressure” in the val, which is defined by meaningful stress barrier layers. Its
Appendix to this chapter. closure pressure is not represented by the measured σmin but
by a value averaged over the gross interval. For large-scale
• Basic fracture geometry models fracturing, pc is equal to σmin only in the unlikely event that
the gross reservoir height is devoid of any variations in the
The elastic relation for the 2D fracture (Eq. 9-10) magnitude or direction of the minimum stress.
is used in two fundamentally different ways to Direct measurement of the fracture width cannot be
achieved during routine field fracturing operations. The infer-
model a fracture. Using a more general form of ence of pc based on the pressure-width relation outlined in
this relation, the characteristic dimension d was Fig. 9A-1 is limited in field practice. An indirect approach that
estimates pc is based on formation tests that create fractures
assumed to be the total fracture length 2L by over the scale of interest. Such large-scale fractures are com-
Khristianovich and Zheltov (1955) and Geertsma monly achieved by the high injection rates used during step
rate and calibration tests. However, conventional shut-in diag-
and de Klerk (1969). The latter study also included nostic plots based on these tests (see “Estimating closure
the effect of fluid loss. This model of a fracture, pressure” in the Appendix to this chapter) may contain multi-
ple inflection points caused by fracture length change and
denoted the KGD model, implicitly assumes that height recession across the reservoir layers. Pressure inflec-
the fracture height is relatively large compared tions are also introduced by the after-closure reservoir
with its length. The other way to use this relation response, as discussed in Section 9-6. These additional
physical phenomena introduce uncertainty in identifying pc
is to assume that the characteristic dimension d is
the fracture height hf, as assumed by Perkins and
Kern (1961) and Nordgren (1972). Their approach,
denoted the PKN model, implicitly assumes that
the fracture length is the infinite dimension.
Each of these two basic, idealized models
assumes that one of the fracture dimensions, the
fracture length or its height, is relatively large in
Width
comparison with the other. Elastic coupling is gen-

erally ignored along the direction of the larger
dimension. Which 2D model is pertinent depends
on which physical dimension of the fracture more
closely replicates the assumptions of the corre-
sponding 2D fracture, as shown on Fig. 9-3. As σmin pc
indicated by Perkins (1973) and Geertsma and
Haafkens (1979), the KGD model is more appro- Pressure
priate when the fracture length is smaller than the
height, whereas the PKN model is more appropri- Figure 9A-1. Mathematical definition of closure
ate when the fracture length is much larger than pressure.

9A. What is closure pressure? (continued)
from a calibration treatment. In contrast, the step rate and

flowback tests discussed in “Estimating closure pressure” in 3800
the Appendix avoid the interpretation uncertainty that is intro-
duced by such phenomena. They provide a more objective
Bottomhole pressure (psi)

diagnostic procedure and should be the preferred field tech- 3700
nique for estimating pc.
The relationship between pc and σmin is best illustrated with
a field example adapted from the M-Site fracture experiments 3600
(Branagan et al., 1996). The reservoir is characterized by a
thin, clean layer (≈6 ft) near the top of a larger sandstone inter-
val (≈30 ft), as shown on the well log plotted in Fig. 9A-2. From 3500
a micro-fracture test conducted within this small interval, σmin
3440 psi
was potentially obtained for the thin layer as 3030 psi (Bran-
agan et al., 1996). A step rate test created a fracture that, 3400
because of the higher injection rates, quickly grew within the 0 1 2 3 4
entire 30-ft interval and indicated that pc was 3440 psi Rate (bbl/min)
(Fig. 9A-3). The calibration injection used downhole inclino-
meters (Branagan et al., 1996) to quantify the fracture width,
and its normalized measurement plotted as a function of the Figure 9A-3. Step rate test analysis.
bottomhole pressure on Fig. 9A-4 shows a trend similar to
that in Fig. 9A-1. Fracture opening is seen to occur at about
3030 psi, which equals approximately the measured σmin.
The linear portion of the curve, however, has an intercept of 1.0
3470 psi, which is near the value suggested by the step rate
test. A pc value of 3030 psi characterizes the micro-fracture
test, where the fracture is limited to the thin sandstone interval.
A pc value of 3470 psi is more appropriate for analyzing the
calibration and proppant injection treatments. 0.8
Normalized tiltmeter response
4520
0.6
4530
4540 Sandstone stringer

Gross
+
interval
4550 Micro-fracture 0.4
test
Depth (ft)
4560
Perforations
4570
0.2
4580
4590 3470 psi
4600 0
0 50 100 150 200 250 3000 3200 3400 3600
Gamma ray (API units) Bottomhole pressure (psi)
Figure 9A-2. Gamma ray log showing the micro- Figure 9A-4. Normalized inclinometer response versus
fracture test location and perforated interval. bottomhole pressure.
the height. Consequently, the 2D model is valid in as is the case for a horizontal fracture in a vertical
cases where the fracture length is either relatively wellbore, or as an intermediate condition between
small or large in comparison with the height. In the two limiting cases of the 2D models.
practice, these models are applicable when the • Correction for fluid pressure gradient
dimensions differ by a factor of about 3 or more.
The fundamental elastic relations (Eqs. 9-10
The radial model is most appropriate when the
through 9-13) assume that the pressure in the frac-
total length 2L (2R in Fig. 9-3) is approximately
ture is constant. The fluid flow relation (Eq. 9-2),
equal to the height. This condition occurs for frac-
however, indicates that a pressure gradient exists
ture propagation from a point source of injection,

within a fracture. The fluid pressure varies from its horizontal well, large pressure gradients may occur
maximum value pw at the wellbore to the forma- near the wellbore. In such cases, the value of β is
tion closure pressure pc at a short distance behind relatively small. Following shut-in, the fluid pres-
the fracture tip. As discussed in Section 5.4-5, the sure is relatively constant near the wellbore, and a
fracturing fluid never quite penetrates the near-tip value of β nearer to unity can be expected.
region, referred to as the fluid lag zone (see also It is clear from the preceding discussion that the
Appendix Fig. 4 of the Appendix to Chapter 5). pressure gradients and hence the values of β dur-
The fluid lag zone is characterized by a pressure ing injection and shut-in differ. During injection,
that rapidly decreases from pc at the fluid front to β is selected as the value at the end of pumping βp.
a value at the fracture tip that is approximately Its value for a PKN fracture can be obtained from
equal to the reservoir pore pressure for permeable Nolte (1979, 1991) as
rock or the vapor pressure for relatively imperme-
β p = (n + 2) (n + 3 + a) . (9-19)
able rock.
The effect of the pressure gradient can be incor- The parameter a defines the degree of reduction
porated into the pressure-width relations of Eqs. in viscosity from the well to the fracture tip result-
9-10 through 9-13 by introducing the factor β ing from thermal and shear degradation. For a con-
(Nolte, 1979). β is defined as the ratio of the aver- stant-viscosity profile, a = 0; for a linearly varying
age net pressure in the fracture ∆p–f to the net pres- profile, a = 1 (i.e., effectively zero viscosity at the
sure at the wellbore pnet: tip).
∆p f No expressions similar to Eq. 9-19 have been
β= , (9-15) reported in literature for the KGD and radial mod-
pnet els. An estimate of βp = 0.85 can be inferred from
where Daneshy (1973) for the KGD model. For the radial
model where fluid enters from within a limited set
pnet = pw − pc . (9-16)
of perforations, βp can be much smaller than unity
The term ∆p–f is defined as the net pressure cor- because of the high entrance flow rate and the
responding to a constant internal pressure p–f that consequently enhanced pressure gradient.
would produce the same average width as where During the fracture closing phase, β is selected
a pressure gradient exists along the fracture length. as its value after shut-in βs, and Nolte (1979, 1986a)
Thus, showed that
∆p f = p f − pc . (9-17) (2 n + 2) (2 n + 3 + a) PKN

β s ≈ 0.9 KGD (9-20)
Equation 9-15 can then be expressed as
(3π 2 32) Radial .

p f − pc = βpnet = β( pw − pc ) . (9-18)
The expressions for the KGD and radial frac-
Equation 9-18 includes the pressure gradient tures are approximate and based on the observa-
effect from flow and fluid rheology along the fraction that the pressure gradient is concentrated near
ture. In combination with Eqs. 9-10 through 9-14, the fracture tip for these models. In particular, the
it also relates the average fracture width to the value of 3π2/32 = 0.925 was selected for the radial
bottomhole wellbore pressure pw, which is a com- model to enable subsequent cancellation with its
monly available field measurement. inverse in Eq. 9-22.
The factor β incorporates the effects of fluid Figure 9-4 is an example of pressure and flow
pressure gradients in a fracture. Tip-dominated profiles during pumping and after shut-in based
fracturing behavior (see Sidebar 9B) is character- on a numerical simulation using the PKN model.
ized by a relatively constant pressure profile, and In addition to the associated values of β, the figure
the fractures exhibit a value of β that approaches shows that flow in the fracture continues until the
unity. For injection from a limited number of per- fracture closes because of the redistribution of the
forations, such as a horizontal fracture from a ver- stored volume from the larger width near the well
tical well or a transverse vertical fracture from a to the higher rate of fluid loss near the tip. This

9B. Pressure response of toughness- response and thus dominates the fracture width and fracturing
dominated fractures process. For example, for assumed values of n = 0.7 and
ptip = 1⁄2∆pµ, the final wellbore pressure is increased by only 3%
M. B. Smith, NSI Technologies, Inc. from the case where ptip is negligible. Alternatively, if ptip is 50%
greater than ∆pµ, then the final wellbore net pressure is just
In its broadest terms, the fracture net pressure that is mea- 6% greater than for the case where ∆pµ is ignored.
sured at the wellbore represents the contribution from fluid vis- The higher injection rates used with very viscous fracturing
cosity and the rock resistance to fracture propagation fluids during large-scale, high-volume fracture treatments
(Shlyapobersky, 1985). The net pressure associated with fluid result in both a relatively large and increasing magnitude of
viscosity results from the flow of the fracturing fluid within the ∆pµ, which changes the pressure response from its early-time
narrow fracture. The dependency of the viscous fracturing tip-dominated behavior to one that is viscous dominated. This
pressure on the rock mechanical and fracture geometry para- change in the relative contribution of ∆pµ compared with that of
meters is provided by Eq. 9-24, which is derived by assuming ptip leads to a corresponding change in the net pressure
that the net pressure is negligible at the fracture tip. behavior on a log-log plot.
Various mechanisms at the fracture tip have been postu- An example is exhibited in Fig. 9B-1 by the pressure
lated to explain the contribution from the rock resistance to the response monitored during a calibration injection into a reser-
fracturing pressure, and these are discussed in Section 6-7. voir with a relatively low Young’s modulus E. The value of pnet
From the perspective of fracture pressure analysis, their contri- is initially constant, signifying toughness-dominated behavior.
bution can be cumulatively represented as the tip-extension At approximately 2.5 min, the pressure response gradually
pressure ptip. The tip-extension pressure can be expressed in increases and eventually approaches a slope of 0.18 on the
terms of the apparent fracture toughness KIc-apparent, which is log-log plot, as expected for a PKN-type fracture with a negligi-
better rationalized from the basis of fracture mechanics. The ble tip pressure (Eq. 9-24). The transition from tip-dominated
relationship between these two quantities generally depends to viscous-dominated behavior is identified from this positive
on the fracture geometry, and for the assumption of a semi- slope on the net pressure plot. In contrast, tip-dominated frac-
circular fracture tip is (Shlyapobersky, 1985) ture behavior would continue to exhibit a relatively constant
net pressure response throughout the treatment.
2
K Ic −apparent = p tip h 2 , (9B-1) The nearly constant pressure response during the initial
π stage of the treatment provides an estimate of ptip, which in
conjunction with Eq. 9B-1 enables the inference of KIc-apparent.
where h is the fracture height in the tip region.
For the example in Fig. 9B-1, ptip is obtained as 260 psi.
Using the PKN fracture geometry model for the viscous
Substituting this value in Eq. 9B-1 with a fracture height
pressure contribution ∆pµ beyond the tip region, the wellbore
of 38 ft suggested by well logs obtains
net pressure can be given as (Nolte, 1991)
  p  2n + 2 
1 ( 2n + 2 )
K Ic −apparent =
2
260 psi
(38 ft × 12 ft / in.) = 4430 psi / in.1 2.
= ∆p µ    + 1
tip
p net , (9B-2) π 2
  ∆p µ  
   
This estimate for KIc-apparent is higher than the critical stress
where the fracture half-length L > h/2 and ∆pµ is the PKN pres- intensity factor KIc commonly measured with laboratory tests
sure contribution from the tip to the wellbore but without the (see Section 3-4.6). The larger value can be attributed to any
rock resistance to propagation. An analytical relation for ∆pµ of the tip mechanisms described in Section 6-7 and results in
for the PKN fracture model beyond the tip region is derived as a correspondingly higher resistance to fracture propagation.
This discussion applies exclusively to an elongated fracture
1 ( 2n + 2 )
E ′  Kh  q i   geometry, as approximated by the PKN fracture model. In con-
n
∆p µ ≅ p net ,PKN ≅ 1.5    (L − Lt ) , (9B-3) trast, both ptip and ∆pµ decrease with continued injection for the
h  E ′  h   radial fracture model. As a result, it is less clear which mecha-
nism dominates the pressure response for radial fractures.
where E ′ is the plane strain modulus, K is the fluid consistency
coefficient, qi is the fluid injection rate, and Lt is the length of
the tip region. Equation 5-19 provides a more specific relation- 1000
ship for pnet with a Newtonian fluid.
The relative contributions of ∆pµ and ptip determine whether Toughness dominated Viscous dominated
the fracture growth is viscosity dominated or tip dominated,
respectively. Usually one phenomena dominates; a nearly
equal contribution from both mechanisms is only rarely
observed. Although ptip generally shows little variation during
pnet
a treatment, Eq. 9B-3 shows that ∆pµ gradually increases with

continued extension. This indicates that a tip-toughness-domi- ptip 0.18
nated fracture response is most likely manifested during the 1
early stage of the fracture treatment. The pressure response
could continue to be toughness dominated if the pressure con-
tribution from ∆pµ remains modest, as is the case for low injec- 100
tion rates, low fluid viscosity (i.e., low K), short fracture lengths 0.1 1.0 10.0
or soft rocks (i.e., low E′). Most of these conditions are likely
Time (min)
during micro-fracture stress tests (see Chapter 3) and tip-
screenout treatments in unconsolidated formations.
The exponents for Eq. 9B-2 suggest that usually one mech- Figure 9B-1. Net pressure response for the transition
anism effectively dominates the wellbore net pressure from toughness-dominated to viscous-dominated frac-
ture growth.

h f
= ∆pf
PKN
1.5
πβ 
1.3 cf = 2 L KGD (9-22)
Average net pressure
2E′ 
(32 3π ) R
1.0 Before shut-in
Net pressure
2
After shut-in
Radial,
0.5 β = ∆p f/pnet where L is the fracture half-length.
βp = 1/1.5 = 0.67 (before)
0 βs = 1/1.3 = 0.77 (after) Martins and Harper (1985) derived the compli-
ance for a fracture that grows as a series of con-
focal ellipses. In this case, the fracture width
Flow rate in fracture
1.0
depends on the elliptic integral of the fracture
Injection rate
Before shut-in
After shut-in aspect ratio. This analysis is applicable during
0.5
Decreasing during closure
fracture growth in an unbounded fashion following
0
initiation from a perforated interval that is shorter
Well 1/2 Tip than the fracture height.
Distance into fracture
Figure 9-4. Pressure and flow rate in a fracture before and

after shut-in (Nolte, 1986a). 9-4. Pressure during pumping
Equations for interpreting pressure during pumping
afterflow causes additional extension during shut-
are developed by combining the basic relations of
in (Perkins and Kern, 1961), which can be signifi-
material balance, fluid flow and rock elastic deforma-
cant for high fluid efficiencies and in reservoirs
tion. The relation between the fracture geometry and
that show a moderate to low resistance to fracture
pressure during pumping was initially proposed by
propagation. Furthermore, shortly after shut-in, the
Nolte and Smith (1981), with application to the PKN-
pressure gradient equilibrates in the near-wellbore
type fracture geometry. This analysis was subse-
region. The redistribution of pressure is accompa-
quently generalized for application to each of the
nied by a change in the wellbore width. The
basic fracture geometry models (Nolte, 1986b).
change in the value of β from injection to the shut-
Extensions were also proposed by Nolte (1991) to
in period reflects the fact that the average width
consider deviations in the fracture geometry from the
should remain relatively constant during this
idealized 2D fracture geometry conditions.
period.
The fundamental relation that defines fracture
• Fracture compliance behavior during pumping at conditions of nearly con-
The net pressure within the fracture compresses stant injection rate and rheology can be obtained by
the formation and results in the fracture width. The combining the fluid flow relation (Eq. 9-2) with that
relation between the net pressure and the fracture of the fracture width and compliance (Eq. 9-21):
width averaged over its length and height 〈w –〉 can
n
be expressed by combining Eqs. 9-10 through 9-13 dp K  qi 
∝ 1+ 2 n   .
( )
with Eq. 9-18: (9-23)
dx c f pnet  hf 
w = c f pnet . (9-21)
Integration over the length with the assumption that
Equation 9-21 indicates that the average fracture pc is a constant and that pnet is negligible at the frac-
width is linearly proportional to the wellbore net ture tip gives
pressure. The constant of proportionality cf is 1/( 2 n + 2 )
referred to as the fracture compliance. The use of  K  q n 
pnet ∝  1+2 n  i  L  . (9-24)
compliance to describe the deformation of solid  c f  hf  
materials under externally applied loads is analo-  
gous to the compressibility of fluid systems during The integration implicitly assumes that the flow
reservoir analysis. The fracture compliance profile along the fracture length has a constant shape.
depends on the formation plane strain modulus This is essentially the case for the three fracture mod-
E′, β coefficient and pertinent 2D fracture geome- els; e.g., the velocity is essentially constant for the
try model: PKN model (Nolte, 1991).

Equation 9-24 and the following equations assume qi ηt qi t *

Af = = , (9-28)
that pnet is dominated by the frictional effects from w w
flow of a viscous fluid within the fracture. This
assumption may not be valid under specific condi- where t* is referred to as the reduced time (Nolte,
tions, as discussed in Section 5-4.5. Sidebar 9B out- 1991). The fracture surface area Af for the basic frac-
lines a diagnostic procedure for comparing the net ture geometry models is
pressure contribution from the fracture tip to the net 2 Lh PKN
pressure from fluid flow (or the viscous net pressure) 
predicted by Eq. 9-24. A f = 2 Lh KGD (9-29)
πR 2
Introducing the appropriate compliance relation in  Radial.
Eq. 9-24 for the three models and using L = hf /2 = R Combining Eqs. 9-27 through 9-29, an expression
for the radial model: for the fracture width as a function of reduced time
e
 L  for the three basic models can be obtained:
PKN pnet ∝ ( E ′ Kq )  3n+1 
2 n +1 n e
i
1 ( 2 n +3 )
 h f   Kq n+1 
 1 
e
PKN w ∝ i n  (t *)1 ( 2 n+3)
KGD pnet ∝ ( E ′ Kqi )  n 2 n 
2 n +1 n e
(9-25)  E ′h f 
1 ( 2 n+4 )
 h f L   Kqin+2 
e
KGD w ∝ (t *)1 ( n+2 )
pnet ∝ ( E ′ 2 n+1 Kqin )  3n  ,
n+2 
1 (9-30)
 E ′h f 
e
Radial
R  1 ( 3 n+6 )
 K 2 qin+2 
where the exponent e represents Radial w ∝   (t *)( 2−n ) ( 3n+6 ) .
 E′2 
e = 1 (2 n + 2 ) . (9-26)
Substituting the relation between width and net
On the basis of Eq. 9-21, the fracture width w pressure in Eq. 9-21 into Eq. 9-30 gives the following
at the wellbore is proportional to cfpnet for the three expressions for pnet in terms of the reduced time:
fracture geometry models. Multiplying each side
1/ ( 2 n +3 )
of Eq. 9-25 by the appropriate definition of cf from  KE ′ 3n+1qin+1 
Eq. 9-22 results in PKN pnet ∝  3 n +3  (t *)1/( 2 n+3)
 hf 
pnet ∝ ( KE ′ n+1 )
e 1/ ( n + 2 )
 Kq n  (t *)− n/( n+2 )
[ ] KGD (9-31)
e
PKN w ∝  i  h1f −n L
 E′  Radial pnet ∝ ( KE ′ n+1 )
1/ ( n + 2 )
e
(t *)− n/( n+2 ) .
 Kq n   L2 
e
KGD w ∝  i   n  (9-27) Equation 9-31 indicates that for typical fracturing flu-
 E ′   h f 
e
ids (i.e., n ≈ 0.4–1.0), the fracture pressure during
 Kqin  2−n e
 [R ] .
Radial w ∝  injection is only nominally sensitive to the reduced
 E′  time t* or fluid efficiency η. Consequently, the effi-
These fracturing pressure and width relations indicate ciency, or alternatively the fluid-leakoff coefficient,
that their dependence on the fluid rheology parame- cannot be determined by analyzing pressure during
ters K and n, wellbore injection rate qi and plane fluid injection exclusively.
strain modulus E′ is the same for all the models. Equations 9-30 and 9-31 also show that the net
Their dependence on the fracture extension L or R pressure and fracture width for any efficiency η can
and height hf differs. In addition, Eqs. 9-25 and 9-27 be approximated by their values for the case of no
also show that pnet and w, respectively, have a weak fluid loss, if the time is scaled by ηt. This time scal-
dependence on qi and that for increasing penetration ing is illustrated for the PKN fracture model in
L or R, pnet increases for the PKN model but Fig. 9-5, which shows the net pressure corresponding
decreases for the KGD and radial models. to no fluid loss (i.e., η = 1) and to an efficiency η =
The time dependence of the fracture width and 0.2 at a time of 50 min. The latter case corresponds
pressure is developed using the definition of η from to a reduced time of t* = ηt = 50 × 0.2 = 10 min.
Eq. 9-4 at a constant injection rate qi (i.e., Vi = qit): Figure 9-5 illustrates that the net pressure at a time of
50 min for the fluid-loss case is equal to the net pres-

R ∝ t 1/ 4 η→ 0
1000 Radial (9-34)
η=1 R ∝ t ( 2 n+2 ) ( 3 n+6 ) η → 1.
in
at 50 m
600 η = 0.2
Limiting expressions for the fracture net pressure
Efficiency, η
400 0.4 are similarly outlined in the Appendix:
pnet (psi)
pnet ∝ t 1/ 4 ( n+1) η→ 0
PKN (9-35)
200 0.2 pnet ∝ t 1 ( 2 n+3) η→1
pnet ∝ t − n /2 ( n+1)
Net pressure
100
Efficiency
0.1
η→ 0
KGD (9-36)
2 5 10 20 50 pnet ∝ t − n ( n+2 ) η→1
Injection time (min)
pnet ∝ t −3n /8( n+1) η→ 0
Figure 9-5. Reduced time illustrated for PKN fracture Radial (9-37)
geometry (Nolte, 1991). pnet ∝ t − n ( n+2 ) η → 1.
Each of these bounding expressions for the net
sure derived for the particular case of no fluid loss at pressure is a power law relation. Consequently, the
a time of 10 min. log-log graph of net pressure versus time should yield
This observation is significant because simple ana- a straight line with a slope equal to the respective
lytical expressions for the three basic models are exponent: positive for PKN behavior and negative for
readily available when η → 1. The various fracture KGD and radial behavior. In particular for PKN
parameters for any generalized value of η can then behavior, the log-log slope for commonly used frac-
be obtained from this limiting conditions merely by turing fluids (i.e., n ≅ 0.5) is typically less than 1⁄4 and
scaling the time by a factor of 1/η. decreases as the efficiency decreases. The log-log
plot of the net pressure versus time during injection,
commonly known as the Nolte-Smith plot, forms the
9-4.1. Time variation for limiting fundamental basis for the interpretation of pressure
fluid efficiencies profiles during fracturing and is analogous to the log-
log diagnostic plot for reservoir flow, as discussed in
Approximations for the time dependency of the frac- Chapter 2.
ture penetration and pressure can be derived from the
equations presented in the previous section for the
two extreme values of the fluid efficiency η. These 9-4.2. Inference of fracture geometry
limiting cases are for very high and low fluid effi- from pressure
ciencies, approaching 1 and 0, respectively. This sim-
plification provides bounding expressions for the The primary reservoir interval is bounded on both
fracture penetration and related pressure. A similar sides by shale formations in the majority of fracturing
approach is used in Section 9-5 to derive relations for applications. Shale zones are generally at higher
analyzing pressure decline during the shut-in period. stress and provide the primary barrier to fracture
Following the mathematical derivations outlined in height growth, particularly during the initial stage
“Mathematical relations for fluid loss” in the Appen- of fracture propagation. The restriction of fracture
dix to this chapter, it can be shown that the fracture height growth is important in low- to moderate-per-
penetration is bounded in the following fashion: meability formations, where relatively long fractures
are required for effective stimulation. Figure 9-6
shows the evolution of the fracture geometry and the
PKN L ∝ t 1/ 2 η→ 0 (9-32)
corresponding wellbore pressure for fracture propaga-
L ∝ t ( 2 n + 2 ) ( 2 n +3 ) η→1 tion under these conditions.
The initial character of fracture propagation, labeled
KGD L ∝ t 1/ 2 η→ 0 (9-33)
as stage 1 on the figure, depends on the length of the
L ∝ t ( n+1) ( n+2 ) η→1 perforation interval providing fluid entry into the frac-
ture relative to the reservoir thickness. Two limiting

source, and the fracture area evolves in an elliptical

Point Source Line Source
Stage 1 Stage 1 shape. The KGD geometry model best describes the
Well early phase of this fracture growth (Martins and
Harper, 1985).
+ + + ++ +
Radial
For either the radial or elliptical propagation mode
model KGD model during stage 1, the net pressure decreases with con-
+
tinued injection. It also exhibits a log-log slope

Barrier between –1⁄8 and –1⁄4 depending on the fluid rheology
∆σ exponent n and the efficiency η, as in Eqs. 9-36 and
9-37. The decreasing pressure reflects the fracture’s
PKN model pc = σmin
preference to grow with decreasing resistance and in
1 Stage 3
Stage 2 an unrestrained fashion as it gets larger. Stage 1 may
occur for only a short time for fracture initiation
within a relatively small interval or for the entire
treatment in a massive zone (Smith et al., 1987).
When barriers at higher stress exist above and
Linear Plot of Pressure
below the reservoir pay zone, fracture height growth
Barriers could be confined following stage 1. Under these
conditions, the fracture is prevented from expanding
pressure, pw
Bottomhole
1 3 in its preferred circular shape and fracture length

2
pc = σ
extension is promoted. This mode of propagation is
denoted as stage 2 in Fig. 9-6 and results in increas-
Time ing pressure as the fracture becomes long relative to
its vertical height. This type of fracture propagates
Log Plot of Net Pressure in a manner similar to the PKN model. For this stage,
the log-log slope of the fracturing pressure is between
log(pnet = pw – pc)
3
1
⁄4 and 1⁄8, once again depending on n and η (Eq. 9-35).
2
Confined fracture height with its characteristic
1
positive log-log slope can be expected until the frac-
turing net pressure approaches a value that is approxi-
log (time) mately one-half of the stress difference ∆σ to which-
ever stress barrier bounding the fracture has the lower
Figure 9-6. Evolution of fracture geometry and pressure stress value. At this magnitude of the net pressure,
during pumping.
the fracture begins to penetrate in a restricted, or con-
trolled, fashion into the adjacent barrier layer with the
cases are described: a limited fluid entry interval and lower stress value. The fracturing pressure continues
one where fluid entry occurs over the complete reser- to increase with penetration, although at a rate that is
voir thickness. Short fluid entry intervals (i.e., limited progressively less than for the PKN model. This con-
perforation intervals) may be desired in vertical well- dition of fracture propagation is indicated as stage 3
bores to mitigate the occurrence of near-wellbore in Fig. 9-6.
problems (see Section 11-3.2). They also occur in hor- If one of the formation barriers is absent (i.e., ∆σ
izontally oriented fractures, during the placement of = 0), height growth into the higher stress barrier is
transverse hydraulic fractures in a horizontal well or arrested. The fracture height, however, continues to
in wellbores that are inclined with respect to the plane grow essentially in a radial-like fashion along the
of σmin. The limited fluid entry into the fracture is direction where the barrier is absent and exhibits a
approximated by a point source. As shown in Fig. 9-6, continuously decreasing pressure (stage 1). This frac-
the fracture area increases in a circular shape for a ture height growth pattern could also occur when
point-source fluid entry and hence is best described fractures are deliberately initiated from zones at
by the radial geometry model. Fluid entry over the higher stress and propagated into bounding layers
complete reservoir thickness is approximated by a line at lower stress, as during an indirect vertical fracture
completion (IVFC; see Section 5-1.2).

The magnitude of the net pressure during stage 2 Higher stress barriers normally have only a limited
(i.e., PKN-type fracture growth) can be used to infer extent. Growth through barriers could eventually be
the magnitude of the fracture compliance based on followed by uncontrolled height growth, resulting in
Eq. 9-31 and therefore the average width using adverse effects during fracturing (see Section 9-4.5).
Eq. 9-30. The net pressure exhibits a relatively small It thus is necessary to estimate the primary parame-
and decreasing value during stage 1, and it cannot be ters that govern fracture height growth: the magni-
effectively used to estimate fracture dimensions. In tude of the stress difference ∆σ between the reservoir
such cases, significant errors could be introduced and barrier and the bounding zone thickness. Figure
owing to uncertainties in the closure pressure and its 9-7, for idealized conditions such as Eq. 6-49,
change resulting from poroelastic effects (see Section addresses these requirements. The figure assumes that
3-5.4). the upper and lower barriers have the same stress
Decreasing net pressure during the initial growth value and an infinite extent. It also approximates
period indicates a radially evolving fracture in either height growth into bounding zones with unequal
the horizontal or vertical plane. Alternatively, an stress magnitudes. As shown, controlled height
increasing net pressure with a small log-log slope (i.e., growth depends on the ratio of pnet and ∆σ. For a
between 1⁄8 and 1⁄4) after the initial growth period is ratio of about 0.4, negligible height growth occurs;
indicative of a vertical fracture extending primarily for a ratio of about 0.65, the total fracture height is
in length with restricted height growth. Following this twice the initial fracture height hi and each barrier
period, if a reduction in the rate of pressure increase thickness must be at least one-half of the height of
is observed, fracture height growth into a barrier zone the reservoir to ensure continued controlled height
should be expected. The net pressure during the growth. Figure 9-7 implies that this condition
period of height growth is governed primarily by the requires a barrier thickness that is at least equal to
difference in stress between the primary reservoir and the height of the reservoir for a ratio of about 0.8.
penetrated zones. Consequently, the pressure response It follows from the previous discussion that the
during the height growth period can be used to esti- amount of height growth into the bounding zones
mate the stress difference, as discussed in the next depends on the thickness of the reservoir and the ratio
section. The stress of the bounding formation is an pnet /∆σ. The reservoir thickness is defined using stan-
important parameter for fracturing design, and it can dard well logs whereas the magnitude of the net pres-
also be used to calibrate log-inferred values of stress sure is estimated from a fracture simulator. The
(see Chapter 4). appropriate fracture model for assessing net pressure
In conclusion, this discussion indicates how the and growth into barriers is based on the PKN model
pressure response during pumping can provide infor-
mation on the state of stress, type of fracture created
and fracture geometry, or more generally the fracture
∆σ
volume term for the material-balance relation in
Eq. 9-5.
4 4
hi
9-4.3. Diagnosis of periods of controlled

Width compliance, cf/cf(hi)
fracture height growth

Fracture height, hf/hi
Fracture height growth into bounding barrier zones at

2 2
higher stress requires an increasing pressure response
prior to the height growth period into the barriers.
Height growth into a higher stress barrier thus cannot
occur during stage 1, where the decreasing pressure 1.1
characteristic of the KGD and radial models occurs. 1 1
0.1 0.2 0.4 0.6 0.8 1.0
Height growth into higher stress barriers (stage 3 in Net pressure, pnet/∆σ
Fig. 9-6), however, is a commonly occurring devia-
tion from the constant-height assumption of a PKN- Figure 9-7. Net pressure and compliance for idealized
type fracture. fracture height growth (Nolte, 1986a).

(Fig. 9-6). Equation 9-25 provides this relation and Table 9-1. Treatment parameters and rock
indicates the importance of the initial fracture height mechanical properties for controlled
(i.e., height of the reservoir) on the magnitude of the fracture height growth example.
net pressure and hence the tendency for growth into
E 5.3 × 105 psi n 0.44
barriers. The net pressure is approximately inversely
proportional to the height. Therefore, the net pressure ν 0.22 K 0.248 lbf-sn/ft2
approximately doubles when the initially fractured hL 24 ft a 0 (constant viscosity)
zone height is halved. The general conclusion is that
Calibration test
a smaller zone is more likely to experience height
growth and will require both higher stress differences Vi 95 bbl tp 12 min
and thicker barrier zones for controlled fracture

height growth. height hL of 24 ft and is bounded by higher stress
Following fracture height growth into a barrier, the shale barriers on both sides (Fig. 9-8a). As shown
efficiency is relatively constant because no fluid loss in Fig. 9-8b, the calibration injection lasted for
is expected to occur in the barrier zone. The pressure 12 min and was followed by an extended shut-in
response during this period deviates from its other- period of approximately 40 min. The shut-in
wise straight-line response on the log-log plot and is pressure approached the far-field reservoir pressure
characterized predominantly by height growth behav- (≈8100 psi) within a relatively short shut-in period
ior. It thus can be used to identify the onset of barrier because of the high formation permeability (≈250
penetration and to estimate the magnitude of the stress md). The closure pressure pc for the formation was
difference (Nolte, 1991; Ayoub et al., 1992a). It relies
on the use of a characteristic signature of the pressure (a)
derivative of the fracturing pressure during controlled
height growth. Because of its increased sensitivity,
the pressure derivative magnifies the deviation in the
fracturing pressure from its expected response and
therefore enhances the identification of fracture
height growth, as discussed in the following section.
The specialized pressure derivative diagnostic has
several applications in addition to quantifying frac-
ture height growth, as discussed in Sidebar 9C: to
validate the fracture geometry inferred from the log-
log plot and to objectively confirm fracture closure
pressure, as well as its capacity to identify the onset
of a screenout. (b)
9400 20
9-4.4. Examples of injection 9200
Calibration test 18 Step rate
pressure analysis 16
9000 Bottomhole pressure 14

Two field examples are presented here. The first 8800 12
example describes log-log analysis for a reservoir 10
8600 8
bounded by shale barriers. The second example dis-
8400 6
cusses the pressure response for a radial fracture
4
geometry in a reservoir where higher stress barriers 8200
Injection rate 2
are absent. 8000 0
0 10 20 30 40 50 60 70 80 90 100
• Example of controlled fracture height growth Time (min)
Table 9-1 lists the parameters relevant to the analy-
sis of a calibration test in a gas-bearing sandstone Figure 9-8. Calibration test analysis for controlled height
growth. (a) Well logs. (b) Bottomhole pressure and rate
reservoir. The producing interval has a permeable record.

9C. Pressure derivative analysis for the net pressure response for the simulated treatment in
diagnosing pumping pressure Fig. 9C-2. The pressure derivative registers a rapid increase at
50 min, identifying the occurrence of a TSO, whereas no signifi-
Joseph Ayoub, Schlumberger Dowell cant change in the net pressure response is visible until later.
Another observation from Fig. 9C-2 pertains to the long-
The fracture net pressure exhibits a power law variation with term pressure derivative response after a TSO. The log-log
respect to time, as demonstrated in Section 9-4. This can be slope becomes greater than 1 with continued injection. The
generalized as ratio of the pressure derivative and the net pressure eventu-
ally becomes larger than 1. Consequently, after a TSO the
p net = p w − p c = At b , (9C-1) pressure derivative eventually becomes larger than the net
presure, as shown on the figure.
where A is a constant and the exponent b is the slope of the Figure 9C-2 also shows an increase in the pressure deriva-
log-log plot of the net pressure versus time t. The slope b tive at 25 min. This response is attributed to increased viscos-
depends on the fracture geometry and the fluid rheology and ity caused by the introduction of proppant. The pressure deriv-
efficiency. ative can be used to assess the importance of the apparent
For pressure data measured during a fracturing treatment, rock toughness relative to the viscous pressure (Nolte, 1991).
the slope depends on the choice of the fracture initiation time
and closure pressure. The fracture initiation time is selected
by examining the pressure record during pumping, and it 10,000
Net pressure or pressure derivative (psi)

often coincides with the time when the fracturing fluid first Underestimated pc
reaches the perforations. The closure pressure is indepen-
dently estimated using one of the techniques discussed in
“Estimating closure pressure” in the Appendix to this chapter.
ct pc
Significant uncertainty is often associated with its determina- Corre
tion, which could result in an incorrect interpretation; e.g.,
treatments for which the pressure data exhibit a small 1000 pc
increase in pressure tend to exhibit a relatively constant net ated
estim
Over
pressure response if a lower closure pressure estimate is
e
used. rivativ
ure de
This sidebar introduces the pressure derivative to enhance Press
the fracturing injection pressure diagnosis and analysis
(Ayoub et al., 1992a). The pressure derivative was initially
introduced in well testing, for which it quickly became stan-
dard practice because it significantly enhances the identifica- 100
tion of various flow regimes during the analysis of transient 1 10 100
well test data (Bourdet et al., 1989). Similarly, when applied Time (min)
to the fracturing injection pressure, the derivative “magnifies”
and detects fracturing events earlier in time. It also assists in Figure 9C-1. Closure estimation using pressure deriv-
the determination of closure pressure. Differentiating Eq. 9C-1
ative analysis.
with respect to time gives
d (pw − p c ) dp w
= = Abt b −1, (9C-2)
dt dt
and multiplying both sides of Eq. 9C-2 by t gives 2000
Net pressure or pressure derivative (psi)
Tip screenout
dp w
t = Abt b . (9C-3) Begin proppant
dt
1000
pnet
Defining the left side of Eq. 9C-3 as the pressure deriva-
tive, it follows that
• Pressure derivative versus time exhibits the same log-log 500
slope as the net pressure.
• Net pressure and pressure derivative are separated by a
factor of 1/b on a log-log plot.
The pressure derivative is independent of the particular 200
choice of closure pressure and is thus unaffected by errors Pressure derivative
in its determination. For typical PKN, KGD and radial fracture
behavior, the closure pressure can thus be inferred from the
injection pressure by selecting a value that makes the net 100
pressure response parallel to the pressure derivative on a log- 5 10 20 50 100
log plot (Fig. 9C-1). This feature of pressure derivative analy- Injection time (min)
sis was applied to the calibration test in Fig. 9-8 to confirm the
closure pressure magnitude (Fig. 9-9).
The pressure derivative magnifies fracturing events Figure 9C-2. Pressure derivative analysis for a TSO
because of its enhanced sensitivity. This characteristic of the response (Nolte, 1991).
pressure derivative is used to quantify fracture height growth
into higher stress bounding zones, as for the examples in
Section 9-4.4. The occurrence of a tip screenout (TSO) is also
magnified and can be detected earlier in time. This is noted on

inferred to be 8910 psi from the step rate test (see Thus, the stress difference between the lower
“Estimating closure pressure” in the Appendix to stressed bounding zones and the perforated interval
this chapter) that followed the calibration treatment. is approximately 360 psi.
The log-log plot of the injection pressure in • Example of radial fracture growth
Fig. 9-9 is similar to the idealized example shown
The example of a radially propagating fracture was
in Fig. 9-6 and can be interpreted in the same
inferred during a calibration treatment performed
manner. The initially decreasing pressure response
in a high-permeability, heavy-oil-bearing sand-
is representative of either a KGD or radial mode
stone reservoir (Fig. 9-10 and Table 9-2). The cali-
of fracture propagation. The initial log-log slope
bration treatment was preceded by a short injection
of –0.18 is between the bounds suggested by
using completion fluids (also called a mini-falloff
Eqs. 9-36 and 9-37 for the KGD and radial models,
injection) and a step rate test. The mini-falloff test
respectively, for the value of n in Table 9-1. The
is used to characterize the reservoir producing
subsequent log-log slope of 0.16 indicates a period
parameters and is discussed in Section 9-6. The
of fracture extension in the PKN mode and a low
formation closure pressure was estimated to be
fluid efficiency.
4375 psi from the step rate test (see “Estimating
1000 (a)
Height
Radial PKN growth
pnet or pressure derivative (psi)
100
10
pnet
1 Pressure derivative
0 1 10
Time (min)
Figure 9-9. Log-log net pressure and pressure derivative

analysis for the calibration treatment in Fig. 9-8.
(b)
The reduced rate of pressure increase during the 5500 100
last 3.5 min of injection is attributed to fracture Mini- Step Cali-
5000 falloff rate bration

growth into the higher stress bounding shales. This Bottomhole 75

pressure
diagnostic is supported by the constant 36-psi 4500
value of the pressure derivative during this period. 4000 3726 psi
50
The magnitude of the stress difference is approxi-
mately 10 times the constant pressure derivative 3500
25
value during this period (Nolte, 1991): 3000
Injection rate
1  d ( pnet )  2500 0
≈ 0.1
∆σ  dt 
t (9-38) 25 75 125 175
Time (min)
 d ( pnet )  Figure 9-10. Calibration test analysis for radial fracture

∆σ = t   0.1 = 36 0.1 = 360 psi. (9-39) growth. (a) Well logs. (b) Bottomhole pressure and rate
 dt  record.

Table 9-2. Treatment parameters and rock 9-4.5. Diagnostics for nonideal
mechanical properties for radial fracture fracture propagation
growth example.
The log-log diagnostic for injection pressure, pre-
E 4.5 × 105 psi n 0.40 sented in the previous sections, is based on idealized
ν 0.25 K 0.084 lbf-sn/ft2 behaviors for fracture height growth and fluid leakoff.
This section discusses common conditions that cause
a 0 (constant viscosity)
deviation from the idealized behaviors and could
Mini-falloff test result in treatment failure from a premature near-
Vi 14.75 bbl tp 3 min wellbore screenout during the proppant treatment.
Calibration test • Rapid growth through a barrier—uncontrolled
Vi 3
107 bbl [600 ft ] tp 4.6 min fracture height growth
Moderate or controlled fracture height growth into
higher stress zones following a period of confined
closure pressure” in the Appendix to this chapter)
fracture extension is described in Section 9-4.3.
and an after-closure linear flow analysis (see
Uncontrolled or runaway fracture height growth
Section 9-6).
occurs when the higher stress zone is traversed and
The log-log net pressure plot for the calibration
the fracture extends into a lower stress zone. This
injection (Fig. 9-11) indicates a slope of –0.11.
is shown on Fig. 9-12, where fracture growth dur-
This slope is within the bounds of a radial fracture
ing stages a and b is the same as that in stages 2
suggested by Eq. 9-37 for the value of the fluid
and 3, respectively, on Fig. 9-6 (The initial stage 1
rheology exponent n in Table 2. A radially grow-
period on Fig. 9-6 of radial- or KGD-type fracture
ing fracture should also be expected in this case
growth is not shown on Fig. 9-12.) Stage b ends
because the reservoir lacks significant shale zones
when the fracture enters a lower stress zone. When
(Fig. 9-10a) that potentially could have constrained
this occurs, the fluid pressure is greater than the
the fracture height.
stress of the zone, initiating an accelerated rate of
growth that leads to stage c. This uncontrolled
growth begins at the well, where the pressure is
1000
b
log pnet
a c
log t
pnet
100
0 1 10
Time (min) hi a b c a b c
1 hf/hi
Figure 9-11. Log-log net pressure analysis for the calibra-
tion treatment in Fig. 9-10.
Figure 9-12. Pressure and width for height growth through

a pinch point.

greatest, and progresses farther along the fracture of the fracture geometry was inferred following
as pumping continues. In a similar manner, uncon- a consistent evaluation of the propped treatment
trolled height growth ceases progressively from using a pseudo-three-dimensional (P3D) fracture
the fracture tip to the wellbore when another simulator (see Section 6-3) and the methodology
higher stress barrier is reached. The fracturing outlined in Section 9-8.
pressure then increases again, indicating either As in the case of a radial fracture, decreasing net
height confinement or a controlled rate of barrier pressure generally indicates uncontrolled height
penetration. growth. The pinch point resulting from height
Figure 9-12 also indicates that the pressure growth to a lower stress zone can cause proppant
between stage b and stage c is relatively constant to bridge at its location during the proppant treat-
and is regulated by a pinch point. During the initial ment but will allow fluid to pass through freely.
penetration into the lower stress zone, the pinch Consequently, excessive dehydration of the slurry
caused by the higher stress barrier layer results in a and decreasing width in the primary reservoir zone
fracture width of nearly zero that closes if the pres- can result in a rapid screenout, even at low prop-
sure decreases. The reduced fracture width also pant concentrations.
causes additional pressure loss in the vertical • Horizontal fracture components—pressure greater
direction, which limits fluid flow vertically into the than overburden
lower stress zone. This pinching mechanism regu-
For the normal state of rock stress in moderate to
lates the pressure to a nearly constant value until
deep reservoir depths, the horizontal stress is less
the penetration becomes sufficiently large to main-
than the overburden or the vertical stress, as dis-
tain a reasonable open width within the barrier
cussed in Section 3-5. When the bottomhole treat-
layer. After the pinch point is overcome for stage c,
ing pressure is less than the overburden stress, a
the rate and extent of vertical growth increase sig-
fracture can propagate only in the vertical plane.
nificantly, accompanied by decreasing pressure
A vertical fracture can also contain a horizontal
and width in the primary reservoir.
component when the pressure exceeds the over-
Uncontrolled fracture height growth is charac-
burden or vertical stress component. This condi-
terized by declining pressure, as indicated on
tion may occur
Fig. 9-12. Figure 9-12 also shows that during the
phase of uncontrolled fracture growth (stage c), – at shallow depths where erosion has removed
a rapid increase in the fracture height hf is accom- some of the overburden to reduce the vertical
panied by a corresponding decrease in the frac- stress
turing net pressure pnet. In contrast, the onset of – in reservoirs in tectonically active thrusting
controlled fracture height growth results in a nomi- environments or in geopressured reservoirs.
nal change in the rate of the net pressure increase, Both conditions increase the horizontal stress.
as discussed in Section 9-4.3. Controlled height – in formations with low in-situ shear strength that
growth, therefore, is relatively difficult to identify undergo stress relaxation resulting in an increase
on a net pressure log-log plot, particularly during of the horizontal stress.
its early stage. The fracture geometry under these conditions
The net pressure response for uncontrolled could have both a vertical component and a hori-
height growth is shown by the field example zontal component. This geometry is called a T-
in Fig. 9-13. A pinch point above the perforated shaped fracture. Examples where T-shaped frac-
interval was identified on the stress log developed tures resulted were reported for the fracturing of
using sonic measurements (see Chapter 4). Frac- a shallow coal bed (Mahoney et al., 1981), shallow
ture height growth into the lower stress sandstone limestone formation (Wood et al., 1983) and lami-
above the pinch point commenced after approxi- nated sandstones (Fragachan et al., 1993). The cor-
mately 150 min of injection and was accompanied responding pressure response and vertical cross
by a steady decrease in the fracture pressure. sections of the width profile are illustrated in
Uncontrolled fracture height growth was also Fig. 9-14. The figure indicates that stage c has a
confirmed using radioactive isotopes that were nearly constant pressure response. Growth into the
injected during the treatment. The time evolution higher stress horizontal plane is similar to uncon-

(a)
30
Calculated pressure
Simulated pressure
5500 60
Slurry rate
Proppant concentration
Proppant concentration (ppg)

50
20
Slurry rate (bbl/min)

4500 40
30
10
3500 20
10
2500 0
50 100 150
Treatment time (min)
(b)
5400
5600
Depth (ft)
5800
6000
3500 5000 –0.6 0.6 0 500 1000 1500
Stress (psi) Width (ft) Fracture half-length (ft)
Figure 9-13. Pressure response for accelerated height growth. (a) Bottomhole pressure match plot. Calculated pressure is
from surface pressure. (b) Fracture profile at the end of injection.
trolled vertical growth beyond a pinch point, as in al., 1988). The horizontal fracture component
Fig. 9-12. Although uncontrolled vertical growth increases the area available for fluid loss and
commences at a pressure less than that of the stress decreases the treatment efficiency. In addition, the
barrier, T-shaped fracture growth occurs at a pres- horizontal component readily accepts fluid but pre-
sure slightly larger than the vertical stress. vents proppant from entering because of its limited
The width of the horizontal fracture component width. Both effects can excessively dehydrate the
is narrow and has twin pinch points at the juncture slurry in the vertical component, which could lead
with the vertical component because of the elastic to premature screenout.
interaction of the two components (Vandamme et

The pressure magnitude provides a diagnostic basis

for determining whether the fracture plane is entirely 10,000
vertical or has a horizontal component as well. The
horizontal component occurs when the fracture pres-
sure is nearly constant and approximately equal to the
overburden stress of the formation, as illustrated in

Fig. 9-14. The magnitude of the overburden stress pw = 10,145 psi
can generally be estimated (see Section 3-5.1) and = overburden stress
should always be compared with the magnitude of 1000

the bottomhole fracturing pressure as part of fracture
pressure analysis.
PKN-type fracture T-shaped
fracture
log pnet
b c pw ≈ overburden
100
a
10 100
Time (min)
log t
Figure 9-15. Calibration treatment pressure response for a
Overburden T-shaped fracture.
of the rock matrix is negligible. Natural fissure

systems are highly directional and show a prefer-
σ
ential orientation along a single axis (Warpinski,
1991). The fissures can produce a complicated
fracture behavior because of fracture offsets,
σ < pw < overburden pw ≥ overburden pw ≈ overburden
enhanced fluid friction along the fracture length
(a) (b) (c)
and the creation of secondary fracture strands.
Figure 9-14. Pressure response for T-shaped fractures.
An important effect of natural fissures is enhanced
leakoff, which can lead to a premature screenout
during proppant injection.
This pressure diagnostic was used to infer the
Natural fissures have a negligible effect on the
occurrence of a T-shaped fracture during a calibra-
fluid leakoff process if the reservoir matrix perme-
tion treatment performed in a tectonically active
ability is high. In low-matrix-permeability condi-
reservoir at a depth of 9750 ft. The log-log net
tions, however, the transmissibility of natural fis-
pressure plot (Fig. 9-15) shows a positive slope
sures can be significantly higher than that of the
lasting approximately 75 min, which indicates a
reservoir matrix. The fracturing fluid can readily
confined mode of fracture extension. The pressure
penetrate into natural fissures during the fracturing
subsequently stabilized at approximately 10,145 psi.
process and maintain a pressure nearly equal to
This constant pressure value is the magnitude of
the pressure in the primary fracture. This process
the vertical stress component for an overburden
is illustrated in Fig. 9-16 in terms of the pressure
gradient of 1.04 psi/ft. During the remaining part
response and a horizontal cross section of the
of the treatment, the penetration of the vertical frac-
width profile. Stage a represents the normal frac-
ture component becomes less efficient because of
turing response in which the fluid pressure is
the propagating horizontal fracture component.
lower than the normal stress on the fissure σf, and
• Natural fissure opening—enhanced fluid loss hence there is a relatively small increase in the fis-
Natural fissures can be important for hydrocarbon sure transmissibility. A continued increase in the
production in the majority of low-permeability fluid pressure, however, reduces the effective
reservoirs, particularly where the permeability stress acting to close the fissure. This effect is

Nolte and Smith showed that the wellbore net

pressure required for fissure opening is
c
a b σ H ,max − σ h,min
log pnet pnet , fo = , (9-40)
(1 − 2 ν)
log t
where σH,max and σh,min are the maximum and mini-
σ
mum far-field principal horizontal stresses, respec-
σf σf
tively. Equation 9-40 applies to PKN-type frac-
p < σf p ≈ σf p > σf
tures, and the fissure can have any orientation with
respect to the main fracture. An important observa-
(a) (b) (c)
tion is that a significant difference between the two
principal stresses in a horizontal plane is required
Figure 9-16. Pressure and width for opening natural to ensure effective fracturing in fissured reservoirs.
fissures. The second model, presented by Warpinski
(1991), is applicable to reservoirs where natural
fissures are the primary source of permeability.
partially compensated for by an increase in the
This model provides a more detailed description
normal stress across the fissure σf resulting from
of the effects resulting from natural fissures. It pre-
the effect of Poisson’s ratio from the main
dicts an enhanced rate of fluid loss throughout the
hydraulic fracture. The overall effect, however,
treatment, with an accelerating effect as the frac-
is a decrease in the effective stress on the fissure.
turing pressure increases (stage a of Fig. 9-16).
When the fluid pressure eventually exceeds σf,
The increase in fracture pressure reduces the effec-
the effective stress on the fissure becomes nega-
tive normal stress acting to close the fissures and
tive. The fissure mechanically opens at this stage,
hence increases their permeability, as described in
labeled stage b in Fig. 9-16. A significant portion
Sidebar 9D. For hydraulic fracturing purposes, the
of the injected fluid can be lost during this process
effect of the magnified permeability is reflected as
because of the large number of fissures that can
an increase in the fluid-leakoff coefficient. As dis-
open at this critical pressure. The conductivity of
cussed in Sidebar 9D, the fluid-leakoff coefficient
fissures is increased by orders of magnitude when
in the presence of natural fissures could be as high
the threshold value is exceeded. The corresponding
as 2 to 3 times that for normally occurring pres-
increase in fluid loss at an essentially constant net
sure-dependent leakoff behavior, even under the
pressure greatly reduces treatment efficiency and
net pressure conditions for stage a.
can lead to excessive slurry dehydration in the
Injection pressure during fracturing for these
main body of the fracture and premature screenout
conditions exhibits a continuously decreasing
(stage c).
slope on a log-log plot, indicating a progressively
Optimized fracturing of a fissured formation
increasing rate of fluid loss. Under continued fluid
requires a model to predict the effect of the fis-
injection, the negative effective stress condition
sures on fluid loss. Two fissure models have been
described by the Nolte and Smith model can occur
reported in literature. The first was proposed by
to open fissures and regulate the pressure to a con-
Nolte and Smith (1981). Their assessment pertains
stant threshold value (stage b of Fig. 9-16). Frac-
to fissured formations with only a slight fluid-loss
turing pressure during pumping in the presence
enhancement unless the threshold fluid pressure
of fissures, therefore, behaves in a manner similar
required for fissure opening is exceeded. The fis-
to the pressure response during periods of con-
sures enlarge when the fluid pressure within the
trolled fracture height growth. Because of this sim-
fracture—and the pressure inside the fissures—
ilarity, the primary diagnostic for distinguishing
becomes greater than the normal stress acting to
between height growth and natural fissures is pres-
close them. When this occurs, fluid loss becomes
sure decline data, as discussed subsequently in
significant and represents the worst-case situation
Sidebar 9F.
for fissure-related leakoff behavior.

9D. Fluid leakoff in natural fissures σ H ,max − σ h ,min (9D-1)

p net ,fo =
,
1− 2ν
Norman R. Warpinski, Sandia National Laboratories where pnet,fo is the wellbore net pressure for fissure opening,
σH,max and σh,min are the maximum and minimum horizontal
Control of fluid leakoff is one of the critical elements for opti-
stresses, respectively, and ν is Poisson’s ratio. Equation 9D-1
mizing hydraulic fracture treatments. The primary reasons for
applies to vertical fissures at any orientation and shows that
leakoff control are to ensure that the fracture acquires the
fissure opening is a likely possibility in formations where the
desired penetration, to keep the sand slurry sufficiently
difference in the horizontal stresses is low. When the fissure
hydrated to flow readily and to limit the invasion of potentially
opens, it behaves much like a hydraulic fracture, accepting
damaging fluids into the formation. In most fracture treat-
large amounts of fluid and resulting in leakoff coefficients that
ments, it is assumed that the leakoff coefficient is constant.
increase by orders of magnitude (thus the term accelerated
However, in naturally fissured formations the assumption of a
leakoff).
constant leakoff coefficient may cause considerable problems
For slightly elevated pressures, the conceptual model is
during the execution of the treatment and seriously hinder
more complicated. The fissure porosity begins to open as the
productivity.
pore pressure increases because the elevated pressure
The concept that natural fissures, or fractures, could alter
relieves some of the net stress on the asperity contacts.
leakoff has been examined or accounted for in some studies
Several models of this process have been developed, the
(Nolte and Smith, 1981; Castillo, 1987; Nolte, 1991; Warpinski,
most well known of which is by Walsh (1981). In his model,
1991; Barree and Mukherjee, 1996) through the use of fis-
the change in permeability of the fissure resulting from
sure-opening conditions or pressure-sensitive leakoff equa-
changes in stress and pressure is
tions. Figure 9D-1 shows conceptually the ways that elevated
pressure could affect natural fissures. Fissures with rough 3
  σ  
surfaces and minimal mineralization are most likely highly k = k o C ln ref   , (9D-2)
sensitive to the net stress pushing on them. Under virgin   σ − p  
reservoir conditions (i.e., the pressure p within the fissure
equals the initial reservoir pressure pi), the effective stress is where C and σref are constants determined from empirical
fairly high and the open slot pores are most likely deformed data, ko is the initial fissure permeability, σ is the stress on the
and closed. As the pressure in the fissure increases because fissure, and p is the pore pressure in the fissure.
of leakoff of the high-pressure fracturing fluid (p > pi), the net The Walsh model has been shown to reasonably represent
closure stress is reduced and the fissure porosity opens. In data from a fissured reservoir in the Mesaverde formation in
this regime, the leakoff coefficient is highly pressure depen- the Piceance basin of Colorado, USA (Warpinski, 1991).
dent. As the pressure exceeds the closure stress on the fis- Figure 9D-2 shows a fit of the Walsh model to data measured
sure (p > pfo), the entire fissure opens, yielding an accelerated during injection and drawdown conditions. The reservoir is
leakoff condition. Vuggy porosity, on the other hand, is gener- overpressured, giving a relatively high initial pressure (5400 psi)
ally insensitive to stress and remain unchanged until the pres- relative to the initial in-situ stress (~7000 psi). The initial reser-
sure exceeds the closure stress and opens the entire fissure voir permeability was measured using drawdown and buildup
(i.e., accelerated leakoff). tests when the interval was first completed. The permeability
The case of fissure opening when the pressure exceeds during injection conditions was measured using both injection
that closing the fissure is the simplest case to consider. Nolte and falloff behavior during nitrogen tracer testing to assess the
and Smith (1981) derived a relation for the critical pressure in interconnectability of the fissure system among nearby wells.
the hydraulic fracture for fissure opening to occur: The permeability collapse that occurred during full drawdown
was a common phenomenon in all intervals at the site. Inev-
itably, a hard drawdown would produce a reasonable rate of
No mineralization, Vuggy gas flow until the pressure reached a critical level at which the
rough surface porosity production would drop to an immeasurable level (but produc-
tion would be restored when the pressure built back up). On
the basis of the flow rates at the initial conditions, the perme-
ability was estimated to have dropped 2–3 orders of magni-
p = pi tude to achieve the observed decrease.
100
Injection
Permeabiltiy (normalized)
Measured reservoir permeability

during nitrogen injection tracer tests
10
pi < p < pfo Initial permeability
1
Estimated permeability
0.1 when production shuts
off from full drawdown
0.001
Initial In-situ
pressure stress
p > pfo 0.001
0 1000 2000 3000 4000 5000 6000 7000
Pore pressure (psi)
Figure 9D-2. Measured permeability variations of the

Figure 9D-1. The effects of pressure on fissure open- Mesaverde natural fissure system at the M-Site
ing and porosity. (Warpinski, 1991).

9D. Fluid leakoff in natural fissures (continued)
The importance of Fig. 9D-2 is that it provides a mechanis-

tic model of understanding the behavior of fissured reservoirs 3.0
during and after stimulation. When the fracturing pressure is
low, the fissure permeability is near the initial value and the
leakoff coefficient is relatively low. As the fracturing pressure 2.5
increases, the leakoff coefficient increases rapidly, causing Pressure-sensitive fissures
problems with fracture size, slurry dehydration and screenout.
2.0
However, the situation during cleanup reverses because the
CL,fissure/CL
well is drawn down to extract stimulation fluids and the natural
fissure permeability decreases. Thus, stimulation fluids are 1.5
injected under wide-open fissure conditions but produced
under clamped fissure conditions, making it difficult to clean √∆p
up the reservoir. 1.0
The Walsh and other models can be incorporated into a
fluid-loss equation to represent changing leakoff conditions
0.5
(Warpinski, 1991). Figure 9D-3 shows the calculated pres-
sure-sensitive leakoff of a Mesaverde fissure system com-
pared with the normal pressure sensitivity of a conventional 0
pore space (recall that the filtrate leakoff coefficient is propor- 0 200 400 600 800 1000 1200 1400
tional to the square root of the fracturing pressure minus the
Net treatment pressure (psi)
reservoir pressure). Generally, the pressure sensitivity of con-
ventional reservoirs is ignored because the changes are
small; in this case, it varied from 1.0 to 1.4. However, the Figure 9D-3. Pressure-sensitive leakoff of a Mesa-
pressure sensitivity of the fissures greatly exceeded this
change, and the leakoff of the fissures reached about 3 times
verde fissure system.
the conventional leakoff. Even at low fracturing pressures, the
leakoff of the fissures was greater than conventional leakoff
decreasing slope with decreasing pressure. Fissure-sensitive
because the fracturing pressure was large relative to the
fluid loss can be recognized in the injection pressure behav-
reservoir pressure and the fissure pores had much less stress
ior, but it is difficult to identify because it looks much like
closing them. This example is based on a tight Mesaverde
height growth during the injection (i.e., a nearly constant pres-
coastal zone reservoir with a base leakoff coefficient CL of
sure derivative). Fissure dilation is usually followed by flatten-
0.0004–0.0006 ft/min1⁄2 (measured during pressure declines)
ing of the fracturing pressure, and screenout most likely
and a fracture-calibrated leakoff coefficient CL,fissure of
occurs relatively fast, depending on the injected proppant
0.0015–0.0019 ft/min1⁄2 during injections.
concentration (Fig. 9-16).
Pressure-sensitive fissure behavior is best recognized in
the pressure decline where the G-plot shows a continuously
9-4.6. Formation pressure capacity These three complicating mechanisms produce

excessive fluid loss from the main body of the frac-
As discussed previously, for a PKN-type fracture, a ture and potential treatment problems. They depend
constant pressure response can occur after a period on the stress differences along the three principal
of normal pressure increase resulting from controlled directions in the formation. The formation acts like
fracture height because the pressure a pressure vessel with a pressure capacity governed
• approaches the stress of a barrier and causes signif- by the pressure-limiting mechanism and the corre-
icant height growth into a lower stress zone (deter- sponding in-situ stress difference.
mined by the horizontal stress difference between When the net pressure reaches the formation pres-
the reservoir and adjacent barrier; Fig. 9-12) sure capacity, the fracturing process may exhibit one
• exceeds the overburden, and the initiation of a T- of the three growth patterns discussed in Section 9-4.5.
shaped fracture begins at a pressure slightly greater Each growth pattern makes additional fracture propa-
than the overburden (determined by the horizontal gation relatively inefficient. Under these conditions,
and vertical stress differences in the reservoir; an inaccurate diagnostic could lead to a design change
Fig. 9-14) that either compromises the effectiveness of the frac-
ture treatment or further aggravates the problem. For
• exceeds the normal stress acting on natural fissures
example, the pad volume or the fracturing fluid vis-
and causes them to open (determined by the stress
cosity may be increased to alleviate recurring screen-
difference between the two horizontal stresses in
outs. These approaches, however, could be ineffective
the reservoir; Fig. 9-16).
if the screenouts are not due to excessive fluid dehy-

dration or insufficient fracture width, respectively. of proppant convection and settling is not a limi-
An important function of calibration tests is to iden- tation with the commonly used polymer fluids
tify the formation pressure capacity and the associ- (Nolte, 1988a; Warpinski et al., 1998a; Shah and
ated complicating mechanism. The propped treatment Asadi, 1998). In addition, a reduction in the fluid
can then be planned to attain the desired fracture viscosity only nominally affects the treatment effi-
characteristics in a cost-effective manner. ciency. Therefore, a moderate decrease in the fluid
Numerous alternatives have been proposed to mod- viscosity should not significantly alter the propped
ify fracture treatments that are constrained by the for- fracture characteristics. Fracturing equipment is
mation pressure capacity. They range from the appli- capable of delivering a continuous decrease in fluid
cation of unconventional designs to reduce the frac- viscosity throughout the treatment (see Chapter 11).
turing pressure (e.g., Veatch and Crowell, 1982; Britt All fracture treatments benefit from a reduction in
et al., 1994) and the use of specialized additives (e.g., the polymer mass used during the treatment. This
Nolte, 1982; Nguyen and Larson, 1983; Warpinski, approach limits the gel residue in the proppant pack
1990) to even altering the reservoir stress state to prevent an excessive loss of fracture conduc-
through changes of either the reservoir pressure or tivity (see Chapter 8).
temperature (Cleary, 1980a). The particular choice of Figure 9-17a shows a pressure record during a
an alternative is driven primarily by the nature of the fracture treatment in a gas reservoir with micro-
problem and how quickly the formation pressure darcy permeability. The formation net pressure
capacity is reached during the propped treatment, capacity identified as 1180 psi is attributed to
as discussed in the following. uncontrolled height growth that occurred after
approximately 75% of the propped treatment had
• Pressure capacity approached during the
been injected. Fracture treatments on offset wells
late stage of the treatment
were successfully designed to maintain the pres-
The recommended approach when any of the three sure below the formation pressure capacity by con-
complicating mechanisms occurs is to engineer the tinuously reducing the fluid viscosity during the
propped treatment so that the fracturing pressure proppant treatment (Fig. 9-17b). The initially
remains below the formation pressure capacity. increasing fluid viscosity was required to over-
Limiting the fracturing pressure may require come enhanced degradation of the early treatment
reducing the treatment volume, injection rate or stage from longer exposure to the reservoir tem-
fluid viscosity. The pressure relation for the PKN perature (Nolte, 1982, 1988a).
fracture geometry in Eq. 9-25 provides an assess-
ment for determining which of these changes is
most effective at reducing the fracturing pressure. (a)
For example, using the typical fluid rheology 2000
Gel (lbm/1000 gal) 60
exponent n = 0.5 in Eq. 9-25 shows that a twofold
pnet capacity = 1180 psi
reduction in the injection rate reduces the net pres- 1000
sure by only 11%. A twofold reduction in the fluid

viscosity (or the fluid rheology coefficient K), 500
10 100 1000
however, provides a 22% reduction in the net pres-
sure. A reduction in the fluid viscosity thus will be (b)
2000
more effective in controlling the fracturing pres- Gel (lbm/1000 gal) 60 50 40 30
sure. Reducing the fluid viscosity, however, pnet capacity = 1180 psi
1000
requires additional considerations of the potentially
increased fluid loss, modifications to the breaker 500
and proppant scheduling, fluid degradation from 10 100 1000
exposure to the reservoir temperature and near- Time (min)
wellbore effects (Nolte, 1982, 1988a).
Viscosity requirements for efficient proppant Figure 9-17. Control of fracture height growth through vis-
transport are often overestimated because the effect cosity reduction (Nolte, 1982, 1988a).


1500
intermediate stage of the treatment

Fissure opening
The mitigating design for reaching the pressure pnet capacity = 1000 psi
1000
capacity during the intermediate stage of the treat-
ment depends on the specific mechanism associated Accelerated leakoff
with the formation pressure capacity. Relatively
fine solids such as silica flour or platelet-based
additives (Vinod et al., 1997) can control acceler- 10
ated fluid loss in formations where fissures provide 10 20 30 40 50 60
a meaningful increase in the overall permeability, Time (min)
even before they are open. The fine-size additive 1500

(e.g., 300-mesh particles) should be used through-
pnet capacity = 1000 psi
out the pad so that fluid leakoff into fissures, and 1000
hence the fluid pressure within them, is reduced
100-mesh sand
from the onset of the treatment.
Controlling accelerated leakoff can significantly
delay or even prevent the mechanical opening of
fissures during the propped treatment. The additive 10
10 20 30 40 50 60
will not bridge at the fracture tip because the far- Time (min)
thest it can travel is the boundary of the fracturing
fluid and the fluid lag region (see Fig. 5-16). The Figure 9-18. Control of accelerated leakoff into natural fis-
fracture width in this region is generally much sures using particulate-based additives (Warpinski, 1990).
larger than the extremely small diameter of the
fine-size additive.
growth (Mukherjee et al., 1994; Nolte, 1988a) or
A larger size solid additive (e.g., 100-mesh
the occurrence of T-shaped fractures. The additives
sand) should be used to control accelerated leakoff
bridge in the narrow pinch points associated with
resulting from mechanically opened fissures. This
both of these mechanisms and arrest the further
additive should be scheduled immediately prior to
growth of secondary features. The benefit of the
proppant addition. It will either bridge in the open
particle-size mixture described by Nolte (1988a)
fissure or block its opening, thereby impeding fluid
is exhibited in the field example in Fig. 9-19. The
entry and the associated increase in the rate of
calibration treatment shows that the critical net
fluid loss. The additive alone may be sufficient if
pressure of 800 psi was reached after approxi-
the fissures exhibit negligible permeability before
mately 20 min of fluid injection (Fig. 9-19a). The
they are mechanically opened. It can be used in
pressure subsequently declined, exhibiting a signa-
conjunction with the previously discussed finer
ture for uncontrolled height growth similar to that
300-mesh particles within the pad when fissures
in Fig. 9-12. Rapid height growth led to a prema-
are responsible for increased fluid loss before they
ture screenout soon after the introduction of prop-
are mechanically opened.
pant. A subsequent successfully placed propped
The effectiveness of 100-mesh sand in control-
treatment exhibited a higher net pressure once
ling fluid loss is demonstrated by the calibration
height growth activity was controlled after the
test shown in the field example in Fig. 9-18. Fis-
addition of a proppant mixture (Fig. 9-19b).
sure opening occurred at a net pressure of 1000 psi,
The use of solid-based additives requires caution
defining the pressure capacity of the formation.
and proper design procedures. Elbel et al. (1984)
The accelerated fluid loss was mitigated by the
showed that incorrect scheduling of the larger size
introduction of 100-mesh sand on a subsequent
additives (e.g., 100-mesh sand) within the pad
calibration test, which exhibited a net pressure well
could result in a premature screenout. Particulate-
in excess of the previous net pressure capacity of
based additives should not be scheduled in con-
1000 psi.
junction with proppant stages because their combi-
A 100-mesh sand or specialized proppant sys-
nation reduces proppant pack permeability. In
tems should be used to control excessive height

(a) horizontal stresses in the reservoir. Under the

Conventional Design altered state of stress, undesirable fracture growth
2000
10 MPa
patterns will occur at much higher net pressure
values. A subsequent, larger size fracture treatment
1000 can then be designed to obtain the desired fracture
Proppant
geometry with a fluid pressure that remains below
Net pressure (psi)
600
the magnitude triggering complex fracture growth.
400
5 10 20 50 100 200
(b) 9-4.7. Pressure response after a screenout

Height Control Design
1000 The pressure response in Fig. 9-1 is for a fracture that
Proppant mixture
600
5 MPa is free to extend until shut-in. A common occurrence
during proppant injection is a screenout. Screenouts
400
5 10 20 50 100 200
can occur in the vicinity of the wellbore when the
Time (min)
formation pressure capacity is reached (see Section
9-4.6) or at the fracture tip and around the complete
Figure 9-19. Control of height growth using a mixture of circumference of the fracture area (i.e., a tip screen-
particle sizes (Nolte, 1982, 1988a). out, or TSO) when the proppant slurry bridges
because of insufficient width or dehydrates as a result
general, only small concentrations of the additive of fluid loss. In addition to resulting from proppant
are sufficient to mitigate any of the complicating bridging, screenouts can also be caused by the exces-
mechanisms. Furthermore, even a moderate loss sive buildup of polymer filter cake at the tip or by
of fracture conductivity has a nominal effect on incorrect scheduling of particulate-based fluid-loss
poststimulation production in low-permeability control additives (Elbel et al., 1984). Either of these
reservoirs, where additive use is more likely. circumferential flow restrictions effectively stops
Fissured formations may benefit from the use of fracture propagation, and subsequent injection is
particulate additives because they effectively prop stored primarily by increasing the width. A screenout
the natural fissures and increase their permeability may be undesirable in a low-permeability reservoir
(Miller and Warembourg, 1975). where increasing the fracture length is the primary
objective, or it can be deliberately designed to occur
to increase production in high-permeability environ-
early stage of the treatment
ments (see Chapter 5).
The approach of pressure capacity during the early As the fracture pressure and width increase during
stage of treatment presents the worst-case scenario the restricted growth period, the pressure gradient
for effective fracture stimulation. None of the mea- from fluid flow decreases. The value of ∆p–f approaches
sures discussed previously may be effective in that of pnet. The value of β approaches 1, as discussed
controlling the complicating mechanisms and their in Section 9-3.3 and illustrated on the log-log plot in
damaging consequences. Mitigating strategies Fig. 9-20. Figure 9-20a shows the variation of pnet
under these conditions should focus on altering the and ∆p–f, which defines the average width. Post-
stress state of the reservoir. Section 3-5.4 indicates screenout behavior is depicted in Fig. 9-20b. The
that any change in the reservoir pore pressure time when the wellbore pressure apparently responds
could change the horizontal stresses by 46% to to the restriction is tso /βp, where tso is the time when
80% of the pore pressure change. The stress state, the screenout occurs. From the values of βp deter-
therefore, is most practically reduced through an mined with Eq. 9-19, tso /βp is on the order of 1.5tso or
extended period of hydrocarbon production prior about 50% greater than at the time of injection before
to the fracture treatment. the restriction occurred. In contrast, Sidebar 9C
The best approach is to initially produce the shows that the pressure derivative is much more
reservoir for an extended period, possibly with a sensitive, with its change perceptible shortly after
much smaller fracture. The corresponding reduc- a screenout.
tion in the reservoir pore pressure will reduce the

where ∆tDso is the dimensionless time after a screen-

(a) Unrestricted Extension out, defined as the ratio of the incremental time after
pnet the screenout and the screenout time, and ηso is the
efficiency when the screenout occurs.
βp < 1
Figure 9-20 shows that the log-log slope of the net
∆pf
Log (“averaged” net pressure, ∆pf)

pressure generally changes after a screenout. Its par-
ticular value at ∆tDso = 2 is shown in Fig. 9-21. An
Log (net pressure)
(b) Restricted Extension at Fracture Tip

important conclusion based on this figure is that the
postscreenout log-log slope approaches unity only
tso /βp
for a high treatment efficiency before the TSO. An
tso example showing the net pressure response in a low-
β≈1 permeability oil-bearing reservoir characterized by a
pnet >1 high fluid efficiency is presented in Fig. 9-22a. A unit
βp < 1
1 slope follows the TSO. In lower efficiency situations,
∆pf however, the fracture pressure exhibits a log-log
behavior with a slope that can be larger than 1.0. This
case is shown in Fig. 9-22b, where the TSO is char-
acterized by a log-log slope of 1.7. The pressure data
Log (time) are from a frac and pack treatment in a 1.5-darcy, oil-
bearing, unconsolidated reservoir that exhibited a low
Figure 9-20. Log-log relations of net pressure versus time treatment fluid efficiency.
for restricted and unrestricted extension. Proppant bridging between the wellbore and the tip
increases the rate of the pressure increase over that
The pressure response and fracture width relation described by Eq. 9-41. Successive packing of the
after the screenout can be derived using the material- fracture toward the wellbore also results in a progres-
balance relations from Section 9-3. For the assump- sively increasing log-log slope with continued slurry
tion of an ideal screenout occurring around the frac- injection (Martins et al., 1992c).
ture circumference, the fracture area remains constant
and equal to the area immediately before the screen-
out. The material-balance relations following a 2.5
screenout therefore are the same as those applied to

describe the pressure behavior following the end of
Log-log slope
2.0
pumping, as discussed in Section 9-5. The only dif- ∆tso
∆tDso = =2
tso
ference, however, is that although no additional fluid 1.5
volume is introduced during the decline, slurry injec-
tion continues after the onset of the screenout and
1.0
should be included in the fundamental relations pre- 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
sented in Section 9-3. Efficiency at screenout, ηso
Applying these modifications to the material-bal-
ance relations, it can be shown (see Appendix Eq. 51) Figure 9-21. Log-log slope after a TSO for ∆tDso = 2.
for the commonly used polymer fracturing fluids that
the log-log slope of the net pressure plot after a screen-
out and in the absence of spurt as β approaches 1 is
9-4.8. Fracture diagnostics from log-log
 π
− sin −1 (1 + ∆t Dso )
−1/ 2

plot slopes
2
π
t dpnet  2 − sin (1 + ∆t Dso ) − ∆t Dso (1 + ∆t Dso )
−1 −1/ 2 The previous sections provide conceptual analyses
= of the different types of log-log slopes of net pressure
pnet dt  ηso → 0 versus time. The slopes are characteristic of various

1.0 ηso → 1 , types of fracture geometries and modes of propaga-
 tion. Therefore, the log-log plot, its associated slopes
(9-41)

After the fracture is confined by barriers, the pres-

(a)
sure increases as predicted by the PKN model, with
1000 the range of the slope as given by Eq. 9-34. This
range is from about l⁄8 for low efficiency to about
1
⁄4 for high efficiency.
As the fracture pressure increases, it can reach the
1
pressure capacity of the formation. This leads to a
1 regulator effect, resulting in nearly constant pressure
because of accelerated fluid loss primarily near the
wellbore (see Section 9-4.5). A nearly constant pres-
100
sure measurement that is equal to the overburden
1 10 100 stress indicates a T-shaped fracture. Controlled frac-
Time (min) ture height growth into a barrier is characterized by
a gradually decreasing log-log net pressure slope and
(b) a constant pressure derivative. Net pressures steadily
Field Data
1000 decrease if uncontrolled fracture height growth
beyond a pinch point occurs. Fissure-dominated
fluid-loss behavior regulates the pressure to a con-
stant value when the fissures are mechanically

opened. If the fissure permeability increases before
100
the mechanical opening occurs, this response will be
1.7
preceded by a gradually decreasing log-log slope.
1 A significant pressure increase (i.e., log-log slope
between those suggested by Eq. 9-41) indicates
10 restricted extension or a screenout near the fracture
1 10 100
tip, whereas a significantly higher slope should be
Time (min)
expected for a restriction nearer the wellbore than the
Figure 9-22. Log-log plot of the postscreenout pressure
fracture tip (Nolte and Smith, 1981).
response for (a) low-permeability long fracture and (b) The types of slopes and associated interpretations
high-permeability short fracture. for vertical fractures are listed in Table 9-3. This
table, in conjunction with the interpretation plot
and the pressure derivative provide a diagnostic tool
Table 9-3. Interpretation of log-log plot
for interpreting the fracturing process. The analyses fracture pressure slopes.
presume that the pressure measurement represents the
actual fracturing behavior, corrected for near-well- Propagation Log-Log Slope Interpretation
Type
bore effects (see Section 9-4.9). Also, the pump rate
and fluid properties are assumed to be relatively con- Ia –1⁄6 to –1⁄5 KGD (Eq. 9-36)
stant during the treatment. Therefore, for a correct Ib –1⁄8 to –1⁄5 Radial (Eq. 9-37)
interpretation of the fracturing pressure, it is impor-
II ⁄6 to 1⁄4
1
PKN (Eq. 9-35)
tant to note variations in the pressure response that
occur when proppant is first injected into the forma- III Reduced from II Controlled height growth
Stress-sensitive fissure
tion or with significant changes in the pump rate or
fluid rheology (e.g., at about 17 min on Fig. 9-18). IV 0 Height growth through
pinch point
The basic net pressure interpretation includes the Fissure dilation
initially decreasing pressure before the fracture is T-shaped fracture
influenced by barriers (Fig. 9-6). This time is gener- V ≥1 Restricted extension
ally short, particularly for zones of relatively small
VI Negative following IV Uncontrolled height
height. For this initial behavior, the theoretical range growth
of the slope is given by the exponents in Eqs. 9-36 Note: n = 0.5
and 9-37.

in Fig. 9-23, shows that the log-log plot with its char- contribution and the fracturing-related pressure
acteristic slopes provides a diagnostic tool analogous response are additive:
to the log-log plot for identifying flow regimes within
pmeas = pc + pnet + ∆pnear wellbore , (9-42)
a reservoir.
where pmeas is the measured bottomhole pressure and
∆pnear wellbore is the cumulative near-wellbore pressure
loss. The result of the near-wellbore effects is an ele-
vated pressure measurement during fluid injection
Log net pressure
and potential misinterpretation when the techniques

V described in the previous sections are used.
III IV
Insufficient perforations could result from improper
VI
perforating practices, poor perforation cleanup or
II ineffective formation breakdown procedures. It has
Ia, Ib
been experimentally shown (Crump and Conway,
1988) that the perforation diameter changes only
Log time
when proppant-laden slurry enters the perforation.
Figure 9-23. Log-log interpretation plot for various fracture For a constant injection rate, perforation friction is
propagation modes. reflected as a constant increase in the treatment pres-
sure during a calibration test and the pad stage of the
main fracture treatment. The introduction of proppant
erodes the perforations, increasing their diameter.
9-4.9. Near-wellbore effects
This results in a progressive decrease in the measured
The previous discussion of log-log plot diagnostics treatment pressure during proppant injection until the
assumes that the bottomhole pressure measurement pressure drop across the perforations is negligible, as
reflects the actual fracture response. The injected shown from 25 to 35 min on Fig. 9-24.
fluid may experience pressure loss prior to entering An apparently higher net pressure with a reduced
the main body of the fracture. Near-wellbore pressure log-log slope is indicated on net pressure plots that are
loss could occur as a result of restrictions arising not corrected for the presence of perforation friction.
from the well completion (e.g., perforations; see The bottomhole pressure measurement, however, will
Section 11-3) or the near-wellbore fracture geometry exhibit an instantaneous change during shut-in if sig-
(e.g., a convoluted fracture pathway; Cleary et al., nificant perforation friction is encountered. The
1993). The latter cause can be particularly damaging, instantaneous pressure change provides a measure of
because it reduces the fracture width close to the the excess pressure associated with perforation friction
wellbore and increases the likelihood of a premature and, in conjunction with Eq. 6-120, is used to esti-
screenout. Near-wellbore pressure loss is rate depen- mate the number of perforations accepting fluid.
dent and affects the pressure response exclusively Fracture tortuosity, or the existence of a convoluted
during the fluid injection. The occurrence of near- pathway from the perforations to the main fracture
wellbore problems thus masks the actual fracture body, causes the fracture to open against a normal
response and complicates the interpretation of injec- stress that is higher than the formation closure pres-
tion pressure. sure. Consequently, the fracture width is narrower in
Chapters 6 and 11 identify three mechanisms to the near-wellbore region, which restricts fluid flow
explain near-wellbore pressure loss: insufficient per- and increases the bottomhole pressure. The associated
forations, a convoluted fracture pathway in the vicin- pressure drop in the tortuous region can be approxi-
ity of the wellbore (or fracture tortuosity) and a mated using the methodology outlined in Section
hydraulic annulus between the cement and the rock 6-8.4. Because the fracture width is smallest at the
that connects the perforations with the main body of beginning of a treatment, the fracture reorientation
the fracture. The evolving fracture geometry, away pressure loss is largest during the early stage of the
from the wellbore, is completely independent of the treatment. As fluid injection proceeds, the pressure
near-wellbore effects. It is assumed that their pressure drop associated with fracture reorientation progres-

sively decreases. Figure 9-25, developed using than 10°; Behrmann and Elbel, 1991), the fracture
numerical simulation, illustrates this response for var- may not initiate at the perforations. Rather, the fluid
ious degrees of fracture turning. communicates with the fracture through a narrow
If the orientation of the perforation with the plane annulus around the casing (Fig. 9-26) (see Section
of the hydraulic fracture is large (typically greater 11-3 and Sidebar 11C). As discussed in Section 6-8,
4100
End of injection
Fluid injection 4000 Proppant
at perforations

3900
Casing 3800
3700
3600
Proppant injection
3500
3400
3300
0 50 100 150 200
Figure 9-24. Perforation friction pressure response.
4100
4000
End of injection
3900
15° reorientation
Fracture 3800
Wellbore plane
at infinity 3700
3600 5° reorientation
3500
Fracture
reorientation 3400 Aligned fracture
3300
0 10 20 30 40 50
Figure 9-25. Fracture reorientation pressure response.
Restriction area
Wellbore
Fracture Fracture
Channel to
fracture wings
Perforation
Figure 9-26. Pinch point caused by the rock–casing annulus.

the formation elastic response (i.e., Poisson’s ratio Fracture pinch effects resulting from fracture reori-
effect) results in a pinch point at the intersection of entation or a hydraulic annulus can be surmounted
the channel with the fracture body. This annulus can during fluid injection. Consequently, an instantaneous
cause higher treating pressure because of the width pressure drop at shut-in, as seen for perforation fric-
restriction, as in the case of fracture reorientation. tion, may not be observed when either of these
However, unlike for fracture reorientation, the effect effects is present. An additional contrast to perfora-
of the annulus can increase with increasing fracture tion friction is that the associated pressure loss
width because of the Poisson’s ratio effect (see Fig. depends on the fracture width and hence exhibits
6-19). The annulus effect decreases with erosion of continuous variation during injection, rather than
the pinch point by fluid in soft rocks or following the being a discrete, time-related event. The step-down
injection of proppant in hard rocks. Section 6-8.5 dis- test, as discussed in Sidebar 9E, can be applied to
cusses the near-wellbore pressure loss associated with develop a more objective prediction of the magnitude
perforation phasing misalignment. of and the mechanism governing near-wellbore pres-
sure loss.
9E. Rate step-down test analysis—a diagnostic ∆p pf ≅ k pf q i2 , (9E-1)

for fracture entry
where the proportionality constant kpf is determined in oilfield
Chris Wright, Pinnacle Technologies units by the fluid density ρ, perforation diameter Dp, number of
perforations N and discharge coefficient C:
The step-down test identifies and quantifies near-wellbore tor-
tuosity. It also quantifies perforation effectiveness and can pro- ρ
vide a rough estimate of the number of perforations accepting k pf = 0.237 . (9E-2)
N 2D p4C 2
fluid. With the step-down test, on-site evaluation can be made
of potential problems resulting from near-wellbore fracture tor- In contrast, ∆pnear wellbore is roughly proportional to the injection
tuosity and perforating. Careful analysis of near-wellbore pres- rate raised to an exponent that is less than unity because it is
sure losses, or the fracture entry friction, enables identifying due to laminar flow through a narrow channel in the pressure-
the problem for resolution and subsequently evaluating the sensitive near-wellbore region:
effectiveness of the remedial measures employed.
Rate step-down test concept ∆p near wellbore ≅ k near wellbore q iβ , (9E-3)
Fracture entry friction (Fig. 9E-1) typically is the combination
of the perforation friction ∆ppf and the tortuosity or near-wellbore where knear wellbore is a proportionality constant and the power
friction ∆pnear wellbore. Perforation friction is simply dissipation of law exponent β is between 0.25 and 1.0, with a value of 0.5
the kinetic energy imparted on the fluid as it flows through a appropriate for most engineering applications. This marked
small orifice at high velocity. Perforation friction, therefore, is difference in the injection rate dependence of the two compo-
proportional to the injection rate qi squared times a proportion- nents of fracture entry friction clearly distinguishes them in
ality constant: a rate step-down test.
Test procedure and analysis

1. Measure the surface pressure and slurry rate at a sampling
Near-Wellbore Friction Perforation Friction rate of every 1–3 s. In general, the measurement of the bot-
tomhole pressure assists the analysis by avoiding uncer-
tainties regarding the hydrostatic head (e.g., in foam treat-
Narrow pressure-
dependent opening
ments) or large values of wellbore friction (e.g., high-rate
tubing treatments).
2. After a breakdown injection or calibration treatment, reduce
Wide far-field fracture the injection rate in steps of 1⁄5 to 1⁄3 of the full rate until the
rate becomes zero. Hold the rate constant at each step to
High-velocity kinetic obtain a stabilized pressure measurement for about 15 to
energy dissipation 20 s. The exact rate at each step is not important; rather,
the goal is rapid change between the constant rates. The
easiest way to achieve this is to take one or two pumps off
Cased line at a time. Record the stabilized surface pressure at
borehole each rate step for input to the step-down test analysis.
with
perforations 3. Determine the change in the bottomhole pressure for each
change in the injection rate. Then plot the fracture entry fric-
tion ∆pentry versus the injection rate. Fit the points obtained
from the rate step-down test with two functions using
Figure 9E-1. Near-wellbore friction and perforation fric-

( ) ( )
∆p entry (q i ) = k pf × q i2 + k near wellbore × q i1/ 2 . (9E-4)
tion exhibit different dependencies on the injection rate.

9E. Rate step-down test analysis—a diagnostic for fracture entry (continued)
A dual-curve-fit algorithm can be used to determine the best- result, several proppant slugs were planned for as early
fit values of kpf and knear wellbore, thus defining the respective as possible during the pad period of the propped treatment.
values of ∆ppf and ∆pnear wellbore for any given rate (Fig. 9E-2). Before the slugs reached the perforations, the tortuosity
continued to increase, which limited the injection rate so
Limitations using surface pressure measurements that the surface pressure could be maintained below the
Rate step-down analysis using the surface treating pressure acceptable value of 6000 psi. However, when the proppant
can be difficult if the tubular friction is large compared with slugs arrived at the perforations, the tortuosity was signifi-
∆pentry because the wellbore friction can vary unpredictably cantly reduced, which enabled increasing the injection rate.
from published or expected values. The friction of water can The propped fracture treatment was successfully placed
be significantly reduced by small amounts of gel contamination with 6-ppg maximum proppant loading.
in the wellbore fluids or from a gel hydration unit. Fluid friction • Poor perforation effectiveness and excessive pad volume
with crosslinked gel may vary with small variations in the fluid
composition. Foam friction behavior is extremely unpredictable The rate step-down test in Fig. 9E-4 was performed after
and variable, and the analysis is further complicated by the first KCl injection. The test clearly shows the dominance
changes in hydrostatic pressure. With turbulent flow (i.e., of ∆ppf (i.e., near-wellbore pressure losses that relate to the
essentially all water injections), the friction is functionally closer injection rate to the 1.94 exponent). The estimated value of
to ∆ppf; with laminar flow (i.e., low injection rates with viscosi- ∆ppf was about 4500 psi at 18 bbl/min, the equivalent of
fied fluids), the friction is functionally closer to ∆pnear wellbore. only 4 of 60 holes open. This condition would not allow fluid
injection at the planned rate of 30 bbl/min. The near-well-
Field example
bore fracture tortuosity was low, at about 50 psi at 18 bbl/min.
The following example illustrates the usefulness of rate step- Additional KCl breakdown injections and surging did not
down test analysis resulting from the ability to understand and improve the low injectivity.
apply the information contained in fracture pressure behavior. The well was reperforated with larger holes, which
reduced ∆ppf to 1500 psi at 18 bbl/min. Following a
• Severe near-wellbore fracture tortuosity crosslinked gel calibration treatment (including a 20-bbl,
The rate step-down test following a second injection of 4-ppg proppant slug), the equivalent of 20 holes was open,
potassium chloride (KCl) water showed that the near-well- which provided an acceptable value of ∆ppf of 500 psi at
bore fracture tortuosity in a naturally fractured dolomite for- 30 bbl/min. The value of ∆pnear wellbore remained at less than
mation was extremely high, at 1900 psi (Fig. 9E-3). As a 50 psi, with C equal to approximately 0.6 for fluid injection.
High Near-Wellbore Friction:

High Premature Screenout Potential
Rate and bottomhole pressure
∆p1
∆p1 Total entry friction
Fracture entry friction
∆p2
∆p2
∆p3 ∆p3
∆pnear wellbore ≅ qi1⁄2
∆qi1
∆qi2 ∆p4
ISIP ∆p4
∆qi3
∆qi4 ∆ppf ≅ qi2
Time ∆qi4 ∆qi3 ∆qi2 ∆qi1

Rate
High Perforation Friction:
Only a Small Number of Perforations
Effectively Taking Fluid
Rate and bottomhole pressure
∆p1
Fracture entry friction
Total entry friction

∆p1
∆p2
∆p2
∆qi1 ∆p3
∆qi2 ∆p3 ∆ppf ≅ qi2
∆p4
∆qi3 ISIP
∆p4
∆qi4 ∆pnear wellbore ≅ qi1⁄2
Time ∆qi4 ∆qi3 ∆qi2 ∆qi1

Rate
Figure 9E-2. The step-down test is conducted to measure ∆ppf and ∆pnear wellbore. ISIP = instantaneous shut-in pressure.

Conclusions
Rate step-down tests are simple to implement and can provide
key insights into the nature of the near-wellbore connection
between the wellbore and the far-field hydraulic fracture.
Although rate step-down tests have limitations, they can pro-
vide the rare combination of critical information at minimal
additional cost.
6000 20
1300-psi reduction Surface pressure limitation = 6000 psi
(first slug)
Injection rate (bbl/min) and bottomhole proppant (ppg)

Surface pressure (psi)
Bottomhole proppant (ppg)
4800 16
Step-down 1: 1900-psi
tortuosity and small
perforation friction
3600 12
Step-down 2: Zero tortuosity at
300-psi tortuosity end of pumping
2400 8
Increased maximum
proppant loading
from 4 to 6 ppg
1200 4
0 0
0 30 60 90 120 150
Time (min)
Figure 9E-3. Propped treatment example with severe near-wellbore friction (i.e., fracture tortuosity) that was mitigated
by pumping two proppant slugs.
9-5. Analysis during fracture closure ties as those observed with the interpretation of con-
ventional well test data.
Fracture behavior during shut-in and prior to closure Basic decline analysis, outlined initially in this sec-
is governed by the fluid-loss characteristics and the tion, follows derivations presented by Nolte (1979,
material-balance relation (Fig. 9-2). A mathematical 1986b). A generalization of the technique using ana-
description of the pressure during the fracture closing lytical extensions to address nonideal conditions
period can be developed by also incorporating the (Nolte et al., 1993) is also presented.
fracture compliance relation (Eq. 9-21). These two
relations and that describing fluid loss are combined
to develop the specialized G-plot, which describes the 9-5.1. Fluid efficiency
pressure response during shut-in. Application of the
“Mathematical relations for fluid loss” in the
G-plot is analogous to the Horner analysis used for
Appendix to this chapter derives the fundamental
conventional well tests. The selection of an applica-
relations for fluid loss at the end of pumping VLp
ble slope for the G-plot also has the same uncertain-
(Appendix Eq. 23) and during the subsequent shut-in

(a)
12,000 50 6000 600
Injection rate (bbl/min) Perforation friction (psi)
Near-wellbore friction (psi)

Perforation friction (psi)

Bottomhole pressure (psi) Near-wellbore friction (psi)

10,000 40 4800 480
8000 30 3600 360

Entry friction
6000 20 2400 240
Points used for step-down
analysis (~end of each rate step)
4000 10 1200 120
2000 0 0 0
5.0 5.5 6.0 6.5 7.0 7.5 0 4 8 12 16 20
Time (min) Pumping rate (bbl/min)
(b)
50 10,000 100
Surface pressure limitation = 10,000 psi
Step-downs 1 and 2 show severe Proppant loading (ppg)
perforation friction (only 4 out of 60 Surface pressure (psi)
perforations open) and low tortuosity Slurry rate (bbl/min)
40 8000 80
Step-downs 3 and 4 show
lower perforation friction and
small tortuosity; No reaction
Proppant loading (ppg)
on proppant slug
Slurry rate (bbl/min)

30 6000 60
20 4000 40
Cut pad percentage
from 50% to 25%
Reperforate
10 2000 20
0 0 0
0 60 120 180 240 300
Time (min)
Figure 9E-4. The step-down test (a) diagnosed high ∆ppf that was remedied by reperforating (b).
period VLs (Appendix Eq. 27). These equations are are not routinely used during field practice. Simple
in terms of the dimensionless volume-loss function analytical approximations, however, can be derived
g(∆tD) and its value at shut-in g0. In this context, ∆tD for certain values of α. These values of α are valid for
is referred to as the dimensionless time and is defined the commonly used crosslinked fluids that develop
in Appendix Eq. 15 as the ratio of the shut-in time ∆t a polymer filter cake along the fracture walls. These
to the injection (or pumping) time tp. fluids are the focus of the remainder of this section.
The general expressions for g(∆tD) and g0 in Corresponding relations for the non-wall-building
Appendix Eqs. 17 and 20, respectively, are based on fluids, such as linear gels or viscoelastic surfactant–
the assumption of a monotonically increasing fracture based fluids, are also discussed in “Mathematical rela-
area that is defined by a power law expression with tions for fluid loss” in the Appendix to this chapter.
an exponent α (Appendix Eq. 2). The generalized The value of the area exponent α can be explicitly
expressions are relatively complicated and therefore determined for two limiting cases of fracture growth.

The lower bound α0 is for the case in which fracture

5
behavior is dominated by fluid leakoff (η → 0) and Bounds
the fracture area grows as the square root of time. Upper (α1 = 1)
Lower (α0 = 1/2)
Treatments in which fluid leakoff is negligible (η → 4
1) represent the upper bound of the area exponent α1,
4/3[(1 + ∆tD)3/2 – ∆tD3/2]
g(∆tD)
and in this case the area increases approximately lin- 3
early with time. (1 + ∆tD)sin (1 + ∆tD)–1/2 + ∆tD1/2
–1
“Mathematical relations for fluid loss” in the 2

Appendix to this chapter provides expressions for α0 π/2
and α1. The upper bound α1 is less than 1. It depends 4/3
1
on the fluid rheology exponent n and is given by 0 0.05 0.1 0.2 0.5 1.0 2.0 5.0
Appendix Eq. 40 for the three fracture geometry Dimensionless time, ∆tD = ∆t/tp
models. The lower bound α0 is independent of the
fracture geometry and is defined by the fluid filtrate Figure 9-27. Bounding values of the dimensionless vol-
rheology (Appendix Eq. 41). For wall-building fluids, ume function g(∆tD) for fracture closure (Nolte, 1986a).
the following bounding values are used to provide
simple analytical expressions: implies that an approximation of g(∆tD) based on a
1 suitable value for α, such as that suggested by Eq. 9-44,
< α < 1. (9-43) should be sufficient for field applications. The
2
approximation is best developed by interpolation
Generally the fluid efficiency varies throughout the using its bounding values and a particular value of α:
treatment. The efficiency is at its maximum value
g( ∆t D , α ) = g ∆t D , α =  + η(2α1 − 1)
during the early stage of fracture propagation, and it 1
gradually reduces with fracture propagation because  2
of the increasing surface area available for fluid
 1 
leakoff. As a result, the fracture area extends with a × g( ∆t D , α = 1) − g ∆t D , α =  , (9-45)
decreasing exponent. The time variation of the expo-   2 
nent affects the fluid loss because the rate of fluid loss where the values of g(∆tD) at α = 1⁄2 and α = 1 are
at any time depends on the prior history of the area from Appendix Eq. 29.
evolution. However, the effect is small, and to main- The efficiency of a calibration treatment can be
tain tractable fluid-loss expressions, it is assumed that derived from the relation that the fracture volume at
the exponent remains essentially constant and is the end of injection equals the total volume of fluid
defined by the value of efficiency at the end of injec- lost during shut-in (Eq. 9-8):
tion. This value of the exponent can be found by inter-
polation between its lower bound α0 of 1⁄2 and upper Vfp VLs ( ∆tcD )
bound α1: η= = . (9-46)
Vi VLp + VLs ( ∆tcD )
α = α 0 + η(α1 − α 0 ) = + η α1 −  ,
1 1 Substituting Appendix Eqs. 22 and 26 into Eq. 9-46
(9-44)
2  2 produces the following expression for the efficiency:
where α1 is from Appendix Eq. 40. For typical effi- g( ∆tcD ) − g0
ciency values of 0.3 < η < 0.6, α is equal to approxi- η= , (9-47)
mately 0.6. g( ∆tcD ) + (κ − 1)g0
A much simplified expression for the fluid-loss where the spurt factor κ is defined in Appendix Eq. 24.
volume function g(∆tD) and its initial value g0 can κ denotes the ratio of fluid loss for a case with spurt
then be developed by using the bounding values of to that without spurt. For the propagation period, κ =
α (Eq. 9-43). These are given by Appendix Eqs. 29 1 for no spurt; more generally,
and 30, respectively, and graphically presented in
κ −1 spurt volume loss
Fig. 9-27. Throughout this development, the differ- = . (9-48)
ence between the upper and lower bounds of g(∆tD) is κ total leakoff volume loss
nominal, as illustrated on the figure. This observation

Figure 9-28 shows the relation between the treat-

ment efficiency η and the dimensionless shut-in time 1.0
κ =1
when the fracture closes ∆tcD for no spurt (i.e., κ = 1). Bounds
0.8 Upper (α1 = 1)
The plot was generated in terms of the two bounding Lower (α0 = 1⁄ 2)
values of α from Eq. 9-43. The equation and figure
Fluid efficiency, η
are independent of the geometry model and enable 0.6
defining the efficiency for proppant scheduling (see
Sidebar 6L). 0.4
The fracture penetration is determined using the
following equation, which is obtained by combining 0.2
Eq. 9-6 and Appendix Eq. 22:
Af =
(1 − η)Vi , (9-49)
0
0 0.05 0.1 0.2 0.5 1.0 2.0 5.0 10.0
2 rp κCL t p g0 Dimensionless closure time, ∆tcD
where the fracture surface area Af for the three basic Figure 9-28. Relationship between efficiency and closure
models is from Eq. 9-29, rp is the ratio of permeable time (Nolte, 1986a).
(or fluid-loss) area to fracture area, and CL is the
fluid-loss coefficient. The average fracture width
is then obtained by using Eqs. 9-4 and 9-49: 9-5.2. Basic pressure decline analysis
2 ηrp κCL t p g0 The previous section presents relations for the fracture
V
w = fp = . (9-50) geometry parameters from simple considerations of
Af 1− η material balance and the assumption of power-law-
Finally, the maximum fracture width immediately based fracture area growth. The basic decline analysis
after shut-in for the three basic models is obtained as assumes that the end of injection marks the termina-
tion of additional fracture extension and that the
4 ( πβ s ) PKN change in the fracture volume during shut-in is attrib-

wmax,si = w 4 π KGD (9-51) uted entirely to the change in the average fracture
3 2 Radial, width during this period. The latter assumption also
 implies that the fracture area is invariant throughout
where 〈w– 〉 is from Eq. 9-50 and β is from Eq. 9-20.
s the shut-in period. The fracture geometry models
The maximum width at the end of pumping is relate the fracture width to the net pressure through
obtained similarly: their compliance. Consequently, the combination of
the compliance and fracture geometry equations
β s β p PKN
 enables determination of the fluid-leakoff coefficient
wmax,p = wmax,si ≈ 1 KGD (9-52) on the basis of the rate of pressure decline during
≈ 1
 Radial. shut-in.
Using the material-balance relation during shut-in
The values for the PKN fracture geometry model from Eq. 9-7, it follows for a constant area that
are from Nolte (1979), whereas those for the KGD
and radial models assume an elliptical width profile dVf ( ∆t ) d w
− = − Af = qL . (9-53)
(Eqs. 9-11 and 9-13, respectively) and no change in d∆t d∆t
the fracture volume immediately before and after For an assumed constant fracture compliance that
shut-in. The value of wmax,p for the radial model from is ensured by a constant area, differentiation of
Eq. 9-52 is valid only for a line-source fluid entry Eq. 9-21 and substitution in Eq. 9-53, with the
condition (Fig. 9-6). The high entry velocity and expression for the fluid leakoff rate qL from Appendix
pressure gradient during pumping for the point- Eq. 11, results in
source case produce a nonelliptical width profile
dpnet 2 rp CL A f
(Geertsma and de Klerk, 1969). − Af c f = f ( ∆t D ) , (9-54)
d∆t tp

where the fluid-loss rate function f(∆tD) is from

L
(
( )
1 2 h 2f PKN
Appendix Eq. 16 and equals the derivative of g(∆tD). 2 − η) ′ 
Equation 9-54 can be integrated between ∆tD = 0 and L =
1 V i E
( )
2 g0 κβ s p * 
1 4h f KGD (9-59)
3
∆tD (assuming a constant value of pc): R  32 3π 3
Radial.

πrp CL t p
pws − pw ( ∆t D ) = G( ∆t D ) ,
This equation provides the appropriate value of pene-
(9-55)
2c f tration for determining CL from Eq. 9-58.
The fluid efficiency can be obtained by substituting
where pws is the bottomhole pressure at shut-in. The the definition of the G-function from Eq. 9-56 into
function G(∆tD) was introduced by Nolte (1979): Eq. 9-47 with the approximation g0 = π/2 for α = 1⁄2:
G( ∆t D ) =
4
π
[
g( ∆t D ) − g0 . (9-56) ] η≈
Gc
, (9-60)
2κ + Gc
This function shares the same bounds for α as g
(Eq. 9-45). where Gc is the value of G(∆tD) at closure (i.e., ∆tD
Castillo (1987) recognized that under ideal condi- = ∆tcD). The approximation provided by Eq. 9-60 is
tions, Eq. 9-55 linearly relates the pressure versus the exact at η → 0.
G-function defined in Eq. 9-56 (Fig. 9-29) with a Martins and Harper (1985) extended the applica-
negative slope p*: tion of pressure decline analysis to the case of
expanding confocal ellipses. This propagation model
πrp CL t p is relevant to the early stage of propagation, before
p* = . (9-57)
the radial model is applicable, and includes the KGD
2c f
and radial models as limiting cases. Its consideration
Combining Eq. 9-57 with expressions for the frac- requires including the appropriate definition of the
ture compliance cf from Eq. 9-22 yields the following fracture compliance cf in Eq. 9-57.
equation for determining the fluid-leakoff coefficient:
h f PKN
p * βs  9-5.3. Decline analysis during
CL = 2 L KGD (9-58) nonideal conditions
rp t p E ′ 
(32 3π ) R
2
Radial.
The basic pressure decline analysis (see Section 9-5.2)
This relation provides a direct solution for the PKN implicitly assumes that
model (assuming that hf is known). For the other
• Fluid loss is based on Carter’s (1957) formulation
models, the penetration is required and must be deter-
of the square root of exposure time (see “Mathe-
mined first. This is achieved by substituting CL from
matical relations for fluid loss” in the Appendix to
Eq. 9-58 into Eq. 9-49:
this chapter) and is characterized by a constant
leakoff coefficient that is independent of pressure.
• Fracture area evolution with time is described by
Pressure
Derivative
a power law area relation during injection.
Derivative, dp/dG (psi)
• Fracture area and compliance are constant during

the closing phase.
pw (psi)
Slope, p*
• Fracturing fluid is incompressible.
pnet,si
• Formation closure pressure is constant.
pc All these assumptions are seldom met in routine
G(∆tcD) field practice. A departure from any of them produces
a G-plot with a continuous curve (i.e., not a straight
G(∆tD) line with a constant slope). In such cases, the applica-
tion of basic pressure decline analysis generally pre-
Figure 9-29. G-plot of the G-function response approxi- dicts optimistic estimates of the fluid-leakoff coeffi-
mating idealized fracture propagation conditions. cient and treatment efficiency. However, a rigorous

characterization of the deviations introduced to the – Change in fracture penetration

analysis when its assumptions are violated results in Following shut-in, the initial extension of the
an overly complex analysis. Fortunately, most of the fracture penetration increases the area exposed
deviations from nonideal conditions are amenable to to fluid loss. The correspondingly increased
relatively straightforward analytical modifications. fluid-loss rate relative to the case of constant
Corrections to the basic pressure decline analysis area results in a correspondingly steeper initial
are presented in this section. Change in the fracture slope of the G-plot. The subsequently reduced
area during shut-in and a pressure-dependent fluid- rate of pressure decline during recession results
loss coefficient are discussed because of their com- from a decrease in the fracture area and elimina-
mon occurrence. Considerations for fluid compress- tion of the region of relatively higher fluid loss
ibility, thermal effects and varying formation closure near the fracture tip. The consequences of frac-
stress (i.e., poroelasticity) were presented by Nolte ture penetration change during shut-in are a
et al. (1993) and are not repeated here. Their effects concave-upward profile on the G-plot (Fig. 9-31a)
were found to be minimal when the procedure out- and a continuous change in its slope mG.
lined in Section 9-5.4 was applied for the decline The corrected slope p* that defines the fluid-
analysis. loss coefficient based on Eqs. 9-54 and 9-58
• Fracture geometry change during shut-in represents an instantaneously unchanged prod-
uct of compliance (e.g., dependent on βs) and
Both the fracture length and height can change
area (i.e., d(cfAf)/dt = 0). For the three fracture
during the shut-in period. The fracture penetration
geometry models, this condition occurs near the
initially increases before eventually receding back
transition between extension and recession
toward the wellbore. Continued fracture extension
(Fig. 9-30). Extensive numerical simulation
occurs because of the redistribution of stored vol-
shows that this condition occurs when the well-
ume from the larger width region near the wellbore
bore net pressure reaches about three-quarters
to the fracture tip region. Simultaneously, the
of its value at the shut-in net pressure pnet,si:
height recedes from any higher stress barriers
because of the reducing fluid pressure (Fig. 9-30). pnet 3
= . (9-61)
This variation in fracture geometry changes the pnet ,si 4
character of the otherwise straight-line G-plot and
calls for additional considerations in using basic Thus, the slope of the G-plot can be desig-
pressure decline analysis to estimate the fluid-loss nated as m3⁄4 and is evaluated at the 3⁄4 point.
parameters. This slope is used in conjunction with Eq. 9-58
to eliminate the effect of fracture penetration
changes during shut-in. The validity of selecting
the slope at the 3⁄4 point has been experimentally
Net pressure
established by de Pater et al. (1996).
fracture penetration and height
Dimensionless net pressure,
– Fracture height recession

Fracture Fracture height growth into higher stress bound-
penetration, L ing zones results in an increase in the fracture
compliance, as shown in Fig. 9-7. Equation 9-57
also indicates that the rate of pressure decline is
Height inversely proportional to the average fracture
growth, h
compliance. Therefore, substantial height
tp
growth during injection leads to a decreased
compliance during the initial shut-in period and
Dimensionless time, tD
reduces the rate of pressure decline until the
height recedes from the bounding zones. Figure
Figure 9-30. Time variation of the net pressure and frac-
ture geometry for the PKN fracture geometry.
9-31b depicts this initial period of reduced slope
on the G-plot.

(a) (b)
Length extension
Length recession
pnet
pnet
pnet,si pnet,si
Height recession
Length recession
pnet ≈ 3/4pnet,si
pnet ≈ 0.4∆σ
Ideal response Ideal response
G(∆tD) G(∆tD)
Figure 9-31. Conceptual G-plot response for nonideal fracture behavior during shut-in for (a) fracture extension and (b)
height growth.
Complete height recession from the bounding Height growth into a higher stress barrier
zones occurs when the wellbore net pressure requires a previously increasing net pressure
equals about 0.4 times the stress difference ∆σ. response. It therefore is not consistent with the
Also from Fig. 9-7, the net pressure at shut-in is basic requirements of the radial and KGD mod-
about 0.8∆σ when significant height growth els, both of which exhibit decreasing net pres-
occurs. Therefore, complete height recession sures. Height growth should be anticipated only
from the bounding zones occurs when the well- for PKN-type behavior, with an increasing and
bore net pressure reduces to a value that is relatively large net pressure during injection. As
approximately one-half of that at shut-in. a result, for the PKN model with significant
Following fracture withdrawal into the primary height growth, the correction mGc is required to
reservoir zone, the pressure subsequently the slope of the G-plot following the termination
declines faster than the initially reduced rate and of height recession and prior to fracture closure.
as for a fracture geometry where no height The corrected slope mG′ that accounts for length
growth occurs. Thus, fracture height growth into recession during this latter phase can be inferred
the bounding zones changes the otherwise from numerical simulations and the material-
straight-line G-plot into a convex-upward curve balance relation during shut-in (Nolte, 1991):
(Fig. 9-31b).
(1 + ∆t ) f (∆t ) ,
1
From this discussion, it is clear that when mG′ ≈ mGc (9-62)
βs
cD D cD
height growth occurs, the fracture is still reced-
ing from its bounding zones when without where fD(∆tD) is the dimensionless fluid-loss rate
height recession the fracture area is momentarily function and is given in Appendix Eq. 28. The
stationary (i.e., pnet/pnet,si = 3⁄4). Therefore, the correction in Eq. 9-62 for fracture height reces-
fluid-leakoff coefficient is underestimated by sion complements Eq. 9-61 for length change to
using the slope at the 3⁄4 point in Eq. 9-58. The account for PKN-type behavior during the shut-
equation also assumes constant compliance and in process.
hence is valid only after complete height reces-
sion into the primary reservoir zone. Conse- • Variable fluid-loss coefficient
quently, the decline analysis requires that the Basic pressure decline analysis assumes that the
G-function slope be evaluated after the period fluid-loss volume is defined by a constant leakoff
of height recession. coefficient. This assumption has been shown to

be generally true (Mayerhofer et al., 1991; see

1.1
Chapter 8) for fluid leakoff governed by a polymer
wall cake (i.e., CL ≈ Cw). Fluid loss, however, is
pressure dependent when it is controlled by either 1.0
the invasion of fracturing fluid filtrate into the
reservoir (i.e., filtrate-controlled fluid-loss coeffi-
Cc correction factor, Kc
0.9
cient Cv) or reservoir properties (i.e., compressibil- η = 0.2
ity control leakoff coefficient Cc), as discussed in

Section 6-4. 0.8
η = 0.3
– Reservoir-controlled leakoff
0.7
Reservoir-controlled leakoff commonly occurs η = 0.4
when reservoirs with highly viscous oil are frac- mǴ = γ Kcm3/4
0.6
tured. It can also occur in low-permeability
η = 0.6
reservoirs that exhibit a high water saturation.
In the case of reservoir-controlled leakoff, the 0.5
slope of the G-plot at the 3⁄4 point can be modi- 0 1 2 3
Dimensionless fluid-loss pressure, pDLs
fied to include the pressure dependency on fluid
leakoff (Nolte et al., 1993):
Figure 9-32. G-plot slope correction factor Kc for reservoir-
mG′ = γKc m3/ 4 , (9-63) controlled leakoff (Nolte et al., 1993).
where the product γKc is applied in the same

manner as the correction for mitigating height be substituted into Eqs. 9-49 and 9-50 for calcu-
recession (i.e., Eq. 9-62) and lating the fracture area and width, respectively.
γ = 1 + β s pDLs , (9-64) – Fracturing-fluid-filtrate-controlled leakoff

Filtrate-dominated leakoff is more characteristic
where the dimensionless pressure difference pDLs
of high-permeability reservoirs where the fractur-
for pressure-dependent leakoff is defined as
ing fluid itself (e.g., viscoelastic surfactant
pws − pc or crosslinked polymer) invades the reservoir.
pDLs = , (9-65)
pc − pi It can also dominate in low-permeability forma-
tions at irreducible water saturation, with small
where pi is the reservoir pressure. values of relative permeability to the fracturing
The correction factor Kc in Eq. 9-63 is defined fluid filtrate. This fluid-loss mechanism depends
in Fig. 9-32 for various values of η and pDLs. on the square root of the pressure difference
This correction factor enables evaluation of the between the fracture and the reservoir (see
slope of the G-plot at a common reference point Section 6-4.2). The nonlinear dependence pre-
(i.e., when pnet/pnet,si = 3⁄4) for the decline analysis. cludes a derivation similar to that for reservoir-
For typical values of efficiency η < 0.5 and pDLs controlled leakoff (Nolte et al., 1993). However,
< 0.5, the figure indicates that Kc is equal essen- Eqs. 9-63 and 9-66 provide an acceptable
tially to 1. approximation if they are modified to reflect the
The corrected G-function slope from Eq. 9-63 square-root pressure behavior. The corrected 3⁄4
can be used with Eq. 9-58 for estimating the slope can be obtained in a manner analogous to
effective fluid-leakoff coefficient Cce during Eq. 9-63:
injection. Cce is related to the reservoir-controlled
fluid-loss coefficient Cc (see Section 6-4.3) eval- mG′ = ( γKc ) m3/ 4 .
1/ 2
(9-67)
uated at the pressure difference ∆p = pc – pi:
The equivalent filtrate-controlled leakoff coef-
Cce = γCc . (9-66) ficient Cve can be defined similarly to Eq. 9-66:
Cce is equivalent to the conventional pressure-
Cve = γ 1/ 2 Cv , (9-68)
independent fluid-leakoff coefficient CL and can

where γ is defined by Eq. 9-64 and Cv is the vis- 4. Gc is corrected to include effects resulting from
cosity control leakoff coefficient (see Section nonideal conditions on the G-plot by defining a
6-4.3) evaluated at ∆p = pc – pi. Like for corrected value of G at closure:
Eq. 9-63, the correction factor Kc is obtained pnet ,si
from Fig. 9-32. G* = , (9-70)
In summary, for fluid loss governed by a poly- p*
mer filter cake, the fluid-loss behavior is inde- where pnet,si is the net pressure at shut-in.
pendent of pressure, and the analysis presented 5. The fluid efficiency is determined by modifying
for Eq. 9-61 or 9-62 is used. Closed-form cor- Eq. 9-60:
rections of the type presented in Eq. 9-63 extend
the basic decline analysis for the reservoir-con- G*
η= . (9-71)
trolled fluid-loss mechanism, such as for heavy- 2κ + G *
oil-bearing reservoirs. Filtrate-controlled behav- 6. Finally, CL is estimated by applying Eq. 9-58 with
ior, which is more common in the stimulation of p* determined using Eq. 9-69.
high-permeability reservoirs, shows a nonlinear
dependence on pressure. The corrections devel- In the absence of height growth and pressure-
oped for the reservoir-controlled leakoff case dependent fluid loss, m3⁄4 and mG′ are expected to be
can be adapted to quantify these effects. equal. With height growth but no pressure-dependent
fluid loss, mG′ is expected to be greater than m3⁄4 and
provide a better estimate for p*. Alternatively, for no
9-5.4. Generalized pressure decline analysis significant height growth but with pressure-dependent
fluid loss, m3⁄4 is expected to be greater than mG′
As discussed in the preceding section, commonly (Eq. 9-62) and provide a better estimate for p*.
occurring nonideal conditions modify the otherwise The relative magnitudes for m3⁄4 and mG′ provide
straight-line behavior of the G-plot. The 3⁄4 point (i.e., a diagnostic to infer the occurrence of pressure-
pnet /pnet,si = 3⁄4) provides a reference to extend basic dependent leakoff. An alternative diagnostic is the G-
pressure decline analysis for effects resulting from function semilog derivative, discussed in Sidebar 9F.
change in the fracture length. In addition, the correc- If pressure-dependent leakoff conditions occur,
tion for recession following fracture height closure or Eqs. 9-63 through 9-68 should be used to determine
pressure-dependent leakoff complements the 3⁄4-point the effective fluid-leakoff coefficient. The primary
slope with the following unified methodology for the deficiency of this method is the selection of an appro-
analysis of shut-in pressure decline before fracture priate slope when both significant height growth and
closure: pressure-dependent fluid loss occur.
1. For all geometry models, find the slope at pnet /pnet,si The 3⁄4 point provides a common reference point,
= 3⁄4 (i.e., m3⁄4). where the slope of the G-plot can be used to account
2. If the PKN model is applicable (determined by the for nonideal effects. Accordingly, this generalized
analysis of pumping pressure) and pressure-depen- evaluation methodology is commonly referred to as
dent leakoff is not expected, determine the slope at the 3⁄4 rule for pressure decline analysis.
closure mGc. Obtain the corrected slope at closure • Example field application of PKN fracture
mG′ using Eq. 9-62. decline analysis
3. The value of p* is selected from the following Shut-in pressures monitored during the calibration
equation on the basis of the applicable fracture treatment described in Section 9-4.4 are analyzed
geometry model: in this example to illustrate the application of
decline analysis to a PKN-type fracture (Fig. 9-8).
max( m3/ 4 , mG′ ) PKN The relevant formation mechanical parameters are

p* = m3/ 4 KGD (9-69) listed in Table 9-1.
m Radial. The plane strain modulus E′ is derived from E
 3/ 4 and ν:

9F. G-function derivative analysis Fissure-dominated leakoff

Fissure-dominated leakoff is illustrated by the analysis of a
R. D. Barree, Marathon Oil Company prefracture injection test (Fig. 9F-2) conducted in a low-per-
Ideal fracture closing behavior is characterized by a straight- meability sandstone formation. The data were acquired using
line pressure response on the G-plot, as described in Section a surface pressure gauge and show pressure-dependent
9-5.2. Extensions of this analysis use the derivative of pres- leakoff. The end of pressure-dependent leakoff is clearly indi-
sure with respect to the G-function (dp/dG; Castillo, 1987) and cated by the derivative becoming constant. This point corre-
the semilog or superposition derivative (Gdp/dG; Barree and sponds to the end of a hump on the superposition Gdp/dG
Mukherjee, 1996). The characteristics of these added diag- curve, following which the curve becomes linear. This early-
nostic curves provide a qualitative indication of the change in time hump above the extrapolated straight line on the super-
fracture geometry during shut-in. They also describe the dom- position curve, along with the sharply curving derivative, is a
inant leakoff mechanisms. This information can be reconciled signature of pressure-dependent leakoff resulting from fis-
with the observed pressure behavior during pumping, knowl- sures. Fissure-dominated leakoff presists up to a G-function
edge of in-situ stresses and rock properties, and postfractur- value of about 0.75, as shown by the onset of the straight-line
ing measurements (e.g., tracer and temperature logs) for an section of the superposition derivative. After fissure closure
improved understanding of the fracturing process. the derivative is constant and the superposition curve is linear
The diagnostics and analysis can be performed on site in a (i.e., constant slope), both indicating a constant leakoff coeffi-
simple and straightforward manner using the measured pres- cient. The pressure data alone provide a less clear indication
sure during shut-in. Values of the pressure p and the deriva- of the end of pressure-dependent leakoff. Closure of the pri-
tives dp/dG and Gdp/dG are plotted on Cartesian axes mary hydraulic fracture is apparent on the derivative and
against the G-function. The plot is interpreted primarily superposition curves at a G-function value of about 2.3, corre-
by inspection. As with a type-curve analysis, characteristic sponding to a pressure of approximately 5100 psi. The instan-
shapes in the curves are identified. Once the general nature taneous shut-in pressure (ISIP) is nearly 1500 psi higher than
of the falloff behavior is developed, specific numerical values the closure pressure.
(e.g., closure pressure and fissure dilation pressure) can be
determined.
The following field examples illustrate the application of
these diagnostics to characterize nonideal fracture and fluid- 2000 7000
loss behavior and the associated characteristic curve shapes. 6500
1600
Derivatives (psi)
Pressure (psi)
Fracture height recession Gdp/dG 6000
Pressure
1200
Height recession during shut-in from high-stress bounding lay- 5500
ers results in changes in the fracture compliance and total Fracture
800 closure
fracture surface area relative to the leakoff (i.e., permeable) 5000
area (Figs. 9-7 and 9-36). This behavior causes several obvi-
ous signatures on the pressure and derivative plots (Fig. 9F-1). 400 dp/dG 4500
The pressure versus G-function curve shows a distinct down- Fissure closure
ward bend as height recession progresses, as discussed in 0 4000
Section 9-5.3. This behavior results in an increasing magni- 0 0.5 1.0 1.5 2.0 2.5 3.0
tude of the dp/dG curve and the superposition Gdp/dG curve. G-function
Figure 9F-1 elaborates this diagnostic using shut-in pres-
sure measurements acquired during a water injection test in
a carbonate formation. Injection was confined to a relatively Figure 9F-2. Pressure derivative analysis for fissure-
small perforated interval surrounded by several hundred feet dominated leakoff.
of similar lithology. The continuously increasing values of
dp/dG and Gdp/dG indicate continuous height recession dur-
ing closure. The figure also suggests that complete fracture Pressure-dependent leakoff
closure has not occurred by the end of the shut-in period.
As discussed in Section 9-5.3, the value of dp/dG can contin-
uously decrease during the preclosure shut-in period. This
behavior results because of change in the fracture penetration
500 1500
during shut-in, and it may occur in spite of a constant value of
1450 the fluid-loss coefficient CL. The diminishing rate of change on

400 the G-plot is magnified if the fluid leakoff is predominantly
1400
Derivatives (psi)
pressure dependent (i.e., filtrate or reservoir controlled). In

Pressure Gdp/dG
300 1350 this case, the superposition analysis provides a distinct signa-
ture that identifies the occurrence of pressure-dependent fluid
1300
loss, despite its similarity to fracture length change during the
200 1250 shut-in period.
1200 Figure 9F-3 shows that pressure-dependent fluid loss is
100 dp/dG characterized by a superposition derivative that approaches
1150 a straight line after a short shut-in period. The straight-line
0 1100 period is unaffected by change in the fracture penetration dur-
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 ing shut-in, and its slope is proportional to the total fluid leak-
off. The early-time Gdp/dG curve has a much steeper slope,
G-function and it falls below the straight-line extrapolation (i.e., the hump
is below the extrapolated Gdp/dG curve). This pattern differ-
Figure 9F-1. Pressure derivative analysis for fracture entiates fracture extension or recession from true pressure-
height growth.

The steps outlined previously in this section are

9F. G-function derivative analysis (continued)
then applied to determine the efficiency and fluid-
leakoff coefficient.
The 3⁄4-point pressure p3⁄4 is derived from an
2000 10,500
instantaneous shut-in pressure pISI of 9012 psi and
1600 10,100 the estimated 8910-psi pc:
Derivatives (psi)
Pressure (psi)
( pISI − pc ) = pc + pnet ,si
3 3
1200 9700 p3/ 4 = pc +
Pressure
4 4
9300 3
= 8910 + (9012 − 8910) = 8987 psi. (9-74)
800 Gdp/dG
400
dp/dG
8900
4
8500
The G-function slope m3⁄4 at this 3⁄4 point is esti-
0
0 0.5 1.0 1.5 2.0 2.5 mated to be 136 psi in Fig. 9-33. For the PKN
G-function fracture geometry model, the slope prior to closure
mGc is inferred to be 123 psi. This slope is cor-
Figure 9F-3. Pressure derivative analysis for filtrate- or rected for the length recession following the termi-
reservoir-controlled leakoff.
nation of height growth using Eq. 9-62:
dependent leakoff. One ambiguity exists, however. The same 1
pattern can be generated by extension of a preexisting nat- mG′ = 123 1 + 0.58 × 0.93 = 194 .
ural fissure set oriented nearly parallel to the hydraulic frac- 0.74
ture or leakoff into a swarm of parallel fractures created by
the fracturing process.
9040
pISI = 9012 psi

E
E′ = 9000
1 − ν2
5.3 × 10 5
3/4 point = 8987 psi
pw (psi)
= = 5.6 × 10 5 psi. (9-72) 8960 m3/4 = 136

1 − 0.22 2 3/4p
net, si
Similarly, βs is determined from Eq. 9-20 using 8920

pc = 8910 psi
the fluid rheology exponent n = 0.44 and viscosity
mGc = 123
profile parameter a = 0: 8880
0 0.2 0.4 0.6 0.8 1.0 1.2
(2)(0.44) + 2
βs = = 0.74 . G (∆tD)
(2)(0.44) + 3 + 0
The fracture height hf selected as the zone of pri- Figure 9-33. Pressure decline analysis for a PKN fracture.
mary initial confinement is determined from well
logs (Fig. 9-8a) to be 24 ft. Because separation The value of mG′ greater than that of m3⁄4 signi-
occurs between the resistivity curves throughout fies fracture height growth, which agrees with the
this pay interval, fluid leakoff is inferred to occur pressure derivative analysis for the injection pres-
over the entire initial fracture height. Thus, the sure in Section 9-4.4.
ratio rp of the leakoff height to the initial fracture The value of p* is determined using Eq. 9-69 to
height is unity. be 194 psi, and G* is determined from Eq. 9-70:
The fracture closure pressure is determined from
9012 − 8910
a step rate test (see “Estimating closure pressure” G* = = 0.53 .
in the Appendix to this chapter) and pressure 194
derivative analysis (Fig. 9-9) to be 8910 psi. The Assuming negligible spurt (i.e., κ = 1), the effi-
dimensionless time at closure is ciency is determined from Eq. 9-71:
∆tc 7.0
∆tcD = = = 0.58 . (9-73) η=
0.53
= 0.21 .
t p 12.0 2 + 0.53

Equation 9-58 is then applied to obtain the fluid- 4555 − 4375

G* = = 0.4 .
loss coefficient: 450
194 × 0.74 × 24
CL = = 1.77 × 10 −3 ft min1/2 . As subsequently discussed in Section 9-6.8, a
1 × 12 × 5.6 × 10 5 nominal spurt with a spurt factor κ =1.02 is esti-
mated for this example. From Eq. 9-71 the effi-
• Example field application of radial fracture
ciency is predicted:
decline analysis
0.4
Application of the generalized pressure decline η= = 0.16.
analysis to a radial fracture is illustrated by analyz- (2 × 1.02) + 0.4
ing the shut-in period during the calibration treat-
The fracture radius R is determined from Eq. 9-59:
ment depicted in Fig. 9-10b. The value of rp is
equal to approximately unity on the basis of open-
(1 − 0.16) × 600 × 4.8 × 10 5 32
hole well logs. The plane strain modulus is deter- R=3 = 37 ft.
mined using Eq. 9-72 with the mechanical parame- 2 × ( π 2) × 1.02 × (32 3π ) × 450 3π 3
2
ters listed in Table 9-2:

Finally, CL is determined by applying Eq. 9-58:
4.5 × 10 5
E′ = = 4.8 × 10 5 psi . 450(3π 2 32) 32
1 − 0.252 CL = × 37
1 × 4.6 × 4.8 × 10 3π 2
5
The value of βs for the radial model is 0.925

from Eq. 9-20. = 1.62 × 10 −2 ft / min1/2 .
As in the previous example for PKN fracture
decline analysis, the steps outlined in this section
are applied to determine the fluid-loss parameters.
The 3⁄4 point is calculated as 9-6. Pressure interpretation after
3
fracture closure
p3/ 4 = 4375 + (4555 − 4375) = 4510 psi , Another application of pressure evaluation pertains
4
to the pressure response following fracture closure.
and m3⁄4 is obtained as 450 psi, as shown on
The pressure during this period reflects the transient
Fig. 9-34.
reservoir response to fracturing and is independent
The fracture is inferred to propagate in a radial
of the mechanisms governing fracture propagation.
fashion, so the correction for height growth sug-
Its character is determined entirely by the response
gested by Eq. 9-62 is not required. For the radial
of a reservoir disturbed by the fluid-leakoff process.
model, p* = m3⁄4 = 450 psi.
During this period, the reservoir may initially exhibit
The corrected value of G at closure is derived
formation linear flow followed by transitional behav-
from Eq. 9-70 as
ior and finally long-term pseudoradial flow (see
Section 12-2). Formation linear flow and pseudoradial
flow are hereafter referred to simply as linear and
pISI = 4555 psi
radial flow, respectively.
4500
3⁄4 point = 4510 psi The after-closure response is similar to the behavior
3⁄4 pnet,si
observed during a conventional well test of a propped
pw (psi)
fracture. It therefore supports an evaluation method-

m3⁄4 = 450
4400 ology analogous to the established principles of pres-
sure transient evaluation. However, one important
pc = 4375 psi
aspect differentiates after-closure evaluation. A propa-
gating or receding fracture exposes the reservoir to an
4300
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 unequal distribution of fluid-loss flux over its length.
G(∆tD) For example, fluid loss dominated by the leakoff coef-
ficient exhibits an elliptical profile of the specific vol-
Figure 9-34. Pressure decline analysis for a radial fracture. ume (i.e., volume lost per unit area) (Fig. 9-35a). The

(a) Leakoff rate (b)

Specific volume
Spurt
Physical length, L
Apparent length, xfa
Figure 9-35. Fluid-loss volume and rate distribution (a) in the absence of spurt and (b) for a spurt-dominated treatment.
after-closure response can then be characterized in 9-6.1. Why linear and radial flow
terms of an equivalent and spatially uniform fluid-loss after fracture closure?
flux over an apparent half-length xfa that is generally
less than L, the physical half-length of a propagating A reservoir is disturbed by fluid-loss invasion during
fracture. Consequently, the value of dimensionless fracture propagation and closing. The fluid-loss rate
time T that is expressed in terms of L for after-closure changes with time as well as over the fracture length
analysis can be different from that of tD, the standard (i.e., has temporal and spatial components). This con-
dimensionless time (see Chapter 12) based on the cept is further explained on Fig. 9-36, which illus-
apparent stationary fracture half-length xfa inferred trates the change in fracture length during a calibra-
from the reservoir response. This distinction between tion test. At a time t = 0, the fracture length is zero.
T and tD is consistent with the notation for the fracture The fracture length increases over the injection time
lengths L and xf employed in this volume. tp, and it finally reaches a maximum value shortly
The after-closure period provides information that after the end of injection. Fluid leakoff at a hypotheti-
is traditionally determined by a standard well test cal location in the reservoir (labeled point a) is initi-
(i.e., transmissibility and reservoir pressure). It com- ated at the time ta when the fracture arrives at this
pletes a chain of fracture pressure analysis that pro-
vides a continuum of increasing data for developing
a unique characterization of the fracturing process.
At optimum conditions, it objectively determines L
parameters that either cannot be otherwise obtained
(e.g., spurt) or exhibit considerable uncertainty when
estimated with conventional pressure decline analyses
Distance
(e.g., closure pressure). a τ

Special attention is also required to address the
interpretation aspects for application of the after-
closure reservoir response. These considerations,
along with an overview of related theoretical frame-
work, are discussed in this section. A physical
description of the reservoir response to fluid loss is
presented initially to enhance understanding of the
ta tp tr tc
after-closure interpretation methodology. The syn-
ergy attained by combining the after-closure analysis Time, t
with information derived from the fracture injection

and preclosure periods is also discussed. Figure 9-36. Rock exposure to fluid loss during fracture
propagation.

location. The fluid-loss behavior can be interpreted fracture propagation, which is equivalent to a diffu-
as an injection source at point a with a strength (or sional rate of L2/t. The relative magnitude or ratio of
intensity) equal to the rate of fluid loss that perturbs the diffusional rates determines the shape of the pres-
the reservoir from its initially undisturbed state. Fluid sure patterns and defines the dimensionless time T:
leakoff continues while the fracture remains open for
pressure diffusional rate
the time interval τ and ends when the fracture recedes T=
to close at the reference point time tr. fracture growth diffusional rate
The calibration test creates a fracture that has negli-
=
(k φµc ) =
t kt
. (9-75)
gible conductivity upon its closure at time t = tc,
following which additional fluid loss is terminated.
(L t)
2
φµct L2
Furthermore, following closure, the reservoir loses all Several pressure propagation patterns are possible
recollection of the mechanical aspects (e.g., elastic depending on the value of T. Figure 9-37a, based on
deformation) of fracturing. The pressure disturbances a reservoir simulation for a low value of T, illustrates
created by leakoff, however, are preserved and dissi- the pressure profile shortly after fracture closure in a
pate with time. The wellbore pressure measurements reservoir characterized by low diffusivity and rapid
following fracture closure reflect the dissipation fracture extension, or a low-permeability reservoir.
process. The after-closure reservoir response is the The pressure disturbances near the wellbore have trav-
cumulative effect (i.e., superposition) from injection eled a small distance into the reservoir. They are gen-
sources distributed along the mechanical length of the erally transmitted perpendicular to the previously frac-
fracture and over a time interval representing their tured surface, indicating a one-dimensional (1D)
exposure to fluid loss. pattern or linear flow. The pressure profile also reflects
The pressure disturbances induced by fluid loss dis- the fracture length attained during fluid injection.
perse or diffuse into the reservoir at a rate governed A similar, and potentially unanticipated, pressure
by the ratio of the mobility k/µ and storage φct (i.e., profile is depicted in Fig. 9-37b for the same fracture
the reservoir diffusivity k/φµct) expressed in terms of propagation rate as in Fig. 9-37a but for a reservoir
the permeability k, porosity φ, viscosity µ and total with a much larger diffusional speed and therefore a
compressibility ct. The resulting pressure patterns cre- relatively large dimensionless time. These conditions
ated by the disturbances depend also on the rate of can be anticipated in a high-permeability reservoir
pi 0.25(pw – pi) 0.5(pw – pi) 0.75(pw – pi) pw
(a) (b)
Fracture Fracture
1 0.5re re 1 0.5re re
0 0
0.5re 0.5re
re re
Figure 9-37. Pressure distribution in the reservoir shortly after closure for (a) small and (b) large values of dimensionless
time. re = reservoir radius.

where the pressure disturbance caused by fluid 9-6.2. Linear, transitional and radial flow
leakoff travels much faster than the rate of fracture pressure responses
growth. Figure 9-37b clearly shows that the pressure
distribution is completely devoid of any linear flow The characteristic pressure response at the wellbore
character and is represented largely as a radial flow and beyond the filtrate region following the closure
pattern. The pressure distribution also does not reflect of an unpropped fracture is illustrated in Fig. 9-38.
the fracture length. Its magnitude at the wellbore is The pressure difference ∆p = p(t) – pi, normalized
independent of the nature of fluid injection into the with respect to its value at closure p(tc) – pi, was sim-
reservoir but is characterized by the reservoir mobil- ulated for a dimensionless time T = 0.0001 at fracture
ity and injected volume. closure, or for a fracture that propagated much faster
A final aspect of the volume distribution pertinent than the pressure disturbance in the reservoir. Figure
to the after-closure response must be described. In 9-38 also shows the pressure derivative and variation
contrast to the elliptical profile in the absence of spurt of the characteristic log-log slope defined as d(ln ∆p)/
(Fig. 9-35a), spurt loss acts as a moving injection d(ln t), where ∆p is the pressure above the initial
source and tends to distribute the fluid-loss volume reservoir pressure pi. This slope is used to identify the
equally over the fracture length (Fig. 9-35b). As indi- time intervals over which the various flow regimes
cated previously, radial flow behavior depends pri- prevail. The pressure response is illustrated for time
marily on the volume of fluid injection and is unaf- intervals that extend beyond 10 times the closure
fected by the leakoff profile. However, during linear period. For this limiting situation, known as the
flow, the reservoir remembers this difference in the impulse condition (see Sidebar 9G), the pressure
fluid-loss behavior because it affects the reservoir variation is described by the ratio of the closure time
pressure distribution shortly after closure. Spurt- to the current time (Gu et al., 1993; Abousleiman et
dominated fluid-loss behavior thus is manifested as al., 1994).
a change in the attributes of after-closure linear flow. (a)
Several conclusions can be derived from the pre- 1.0
ceding qualitative descriptions of the reservoir Pressure
Normalized pressure difference
difference
response:
and pressure derivative
• Fluid injection tests that achieve efficient fracture 0.1

Transitional knee time
extension (as in most calibration treatments) are T = 0.2
likely to achieve a low dimensionless time at the Pressure
derivative
end of pumping. Therefore, they could display 0.01
well-defined periods only of linear flow and long-
term radial flow after fracture closure.
• The pressure distribution during linear flow reflects 0.001
0.001 0.01 0.1 1.0 10.0
the fracture geometry attained during injection. The
(b)
related analysis can provide an estimate of the frac-
1.0
ture length.
• Linear flow is also affected by the relative contri- 0.9
bution of spurt, which can be distinguished by Extended Radial
radial
Log-log slope
evaluating the pressure data from this period. 0.8
• Fluid injection tests characterized by a low effi- T = 0.005 T = 0.05 T = 0.5 T = 5.0
0.7
ciency can be devoid of linear flow in spite of the Linear Extended
creation of a fracture. 0.6
linear
• Radial flow is independent of the nature of fluid

leakoff. This period is, however, governed by the 0.5
0.001 0.01 0.1 1.0 10.0
reservoir mobility and can be used to estimate the
Dimensionless time, T
reservoir transmissibility.
Figure 9-38. After-closure pressure difference and deriva-
tive (a) and log-log slope (b) plots for an impulse injection.

9G. Impulse testing where pi is the reservoir pressure, µ is the reservoir fluid vis-
cosity, k is the reservoir permeability, and h is the reservoir
~
Joseph Ayoub, Schlumberger Dowell height. The term q defines the average injection rate into the
reservoir. Equation 9G-3 can be reformulated on the basis
The impulse test is a specialized well testing procedure that that the injected fluid volume is the product of the average
enables analysis of the reservoir response following a rela- ~
injection rate and closure time (i.e., Vi = q tc):
tively short duration of fluid injection or production (Ayoub et
al., 1988). The pressure response to an ideal-rate impulse Vi µ 1  1 tc 
pw ( ∆t ) − p i = 1− 
(i.e., to an instantaneous source) is given by the derivative of 4 πkh ∆t  2 ∆t 
the pressure response to a step rate change. The impulse pro- Vµ 1  1 
cedure can also be applied to a short injection that creates a = i 1− ε , (9G-4)
hydraulic fracture within the reservoir (Abousleiman et al., 1994). 4 πkh ∆t  2 
The dimensionless wellbore pressure response pwD follow-
ing fracture closure resulting from a constant injection rate where ε = tc /∆t. It follows that for ∆t ≥ 10tc, pw(∆t) – pi is pro-
can be expressed by the principle of superposition as portional to 1/∆t within an error of less than 5%.
The pressure derivative ∆tdpw/d∆t can be obtained from
Eq. 9G-4:
p wD (t D > t cD ) = p wD (t cD + ∆t D ) − p wD ( ∆t D ) , (9G-1)
dp w ( ∆t ) Vi µ 1
where tD is the conventional dimensionless time from Eq. 12-1 ∆t
d∆t
=−
4 πkh ∆t
[1− ε] . (9G-5)
and tcD and ∆tD represent the dimensionless times for fracture
closure and any subsequent interval, respectively (Fig. 9G-1). The term ε can be neglected when the shut-in period is rel-
Equation 9G-1 can be reformulated using the standard atively long compared with the closure time. For ∆t ≥ 10tc,
Taylor series expansion: Eqs. 9G-2 and 9G-5 become equal and proportional to 1/∆t
2
within a 10% error. It follows that
p wD (t D > t cD ) = t cD
dp wD 2 1 d p wD
+ t cD + higher order terms .
d∆t D 2 d∆t D2 dp w ( ∆t ) Vi µ 1
p w ( ∆t ) − p i = ∆t = , (9G-6)
(9G-2) d∆t 4 πkh ∆t
Neglecting the higher order terms in Eq. 9G-2 and substi- where ∆t >> tc.
tuting for pwD with its expression during pseudoradial flow from
Eq. 12-15 gives Practical considerations
q˜µ t c  1 tc  In theory, impulse analysis is applicable only for an instanta-
p w ( ∆t ) − p i = 1− , (9G-3) neous fluid injection. In practice, however, a finite injection
2πkh 2∆t  2 ∆t 
time is required to create a pressure disturbance that is mea-
surable for a long shut-in duration. Consequently, the analysis
is valid only if the shut-in time is long in comparison with the
End of pumping finite injection time, as previously discussed.
Equation 9G-6 shows that during reservoir radial flow, a
pc log-log plot of the pressure difference and pressure derivative
versus 1/∆t exhibits a unit slope. Multiplying the pressure dif-
Pressure (psi)
ference and pressure derivative by ∆t provides a horizontal

line (i.e., zero slope) similar to the classic pressure response
for transient radial flow. The two lines also overlie each other.
tc ∆t This behavior provides a diagnostic method not only for iden-
tifying reservoir pseudoradial flow but also for determining the
initial pressure of the reservoir. In addition, the pressure
response during this period can be used in conjunction with
Eq. 9G-6 to estimate the reservoir transmissibility kh/µ. This
Time, t provides the basic parameters required for evaluating the
reservoir production performance.
Figure 9G-1. Reference times for the impulse injec-

tion. pc = closure pressure.
As shown on Fig. 9-38b, linear flow appears ini- also emphasizes the “knee” formed by the intersec-
tially and lasts until T exceeds 0.005 and exhibits the tion of the slopes for the linear and radial flow peri-
anticipated half-slope for the pressure difference and ods. This intersection defines a unique dimensionless
derivative. Reservoir radial flow, characterized by a time that can be used to infer the fracture length
unit log-log slope, is the terminal flow regime that achieved during propagation, as subsequently dis-
occurs only for T > 5. The intermediate transitional cussed in Section 9-6.7.
period lasts over a time period spanning a factor of Figure 9-38 also illustrates a significant shortcom-
approximately 1000 times beyond the end of linear ing for extracting the reservoir transmissibility from
flow. The plot of the pressure difference in Fig. 9-38a the after-closure response in low-permeability reser-

voirs that have extensive, efficient fracture propaga- period of after-closure pressure data (Nolte et al.,
tion. Such conditions are characterized by a small 1997). As discussed in Chapter 8, polymer fluid inva-
value of Tp and require a long after-closure period to sion into a moderate- to high-permeability reservoir
attain radial flow. For the example in Fig. 9-38, a pro- can result in an internal cake of a highly concentrated
hibitively long shut-in period that equals 50,000 polymer gel. This internal cake experiences a sus-
times the injection time would be required prior to tained period of creeping afterflow following fracture
obtaining radial flow. However, from an engineering closure. It supports a fraction of the total pressure
perspective (i.e., an allowable error of 10%), the tran- drop between the fracture and the reservoir, masking
sitional period can be shortened by extending the lin- the reservoir response. After a time that equals
ear and radial flow periods, as shown in Fig. 9-38b. approximately the time before closure, the internal
This approximation considerably reduces the transi- cake begins to flow more freely. The pressure then
tional period from a factor of 1000 to 10, which quickly drops to a value that reflects the pressure dif-
greatly increases the likelihood of obtaining a linear ference in the reservoir beyond the polymer material.
or radial flow analysis with an acceptable accuracy Finally, the increasing effective stress during the
and under the normal constraints of field operations. after-closure period continues to consolidate the filter
cake and fracture faces and further reduce the near-
wellbore permeability. This can eventually result in
9-6.3. Mini-falloff test a loss of communication between the wellbore and
A robust strategy for estimating the reservoir parame- the reservoir that limits the availability of valid long-
ters from radial flow as well as for characterizing term pressure data.
fracture behavior and obtaining linear flow is to apply “Guidelines for the field application of after-closure
two separate tests, with the first test a short injection analysis” in the Appendix to this chapter discusses
test that may or may not create a fracture. This spe- operational considerations for the mini-falloff test.
cialized calibration test is referred to as the mini- Appendix Eq. 1 provides a guideline for determining
falloff test. The mini-falloff test should be performed the injection rate during the test. It follows from this
in an undisturbed reservoir. Except in very tight for- equation that early radial flow can be achieved in a
mations, the test can be engineered using inefficient low-permeability gas reservoir (i.e., in tenths of a mil-
fluids and a low injection rate such that radial flow lidarcy) with an operationally permissible injection
occurs during injection or shortly after closure (i.e., rate (≈ 1 bbl/min). Lower injection rates are required
early radial flow). The subsequent fracture calibration to attain radial flow within the shut-in period in a
test is performed with much higher injection rates reservoir with microdarcy permeability. Under these
and a more efficient fracturing fluid to characterize extreme reservoir conditions, only linear or transi-
the fracture behavior as well as attain after-closure tional flow may occur because of field operational
linear flow. The long-term radial flow behavior that constraints. However, transitional flow can also be
normally occurs only after a prohibitively long shut- analyzed to estimate the reservoir properties, as dis-
in period can be anticipated from the reservoir infor- cussed in “Comparison of fixed-length and propagat-
mation derived with the mini-falloff test. The mini- ing fractures” in the Appendix to this chapter.
falloff test, therefore, also facilitates the integration
of pre- and after-closure analyses for a calibration
test, as discussed in Section 9-6.4, under the time
9-6.4. Integration of after-closure
constraints of normal field fracturing operations.
and preclosure analyses
A nonpolymer fluid (e.g., well completion fluid) The after-closure analysis complements the preclosure
must be injected during the mini-falloff test for two analysis by providing parameter estimates that cannot
reasons. First, this type of fluid generally exhibits a otherwise be derived from pretreatment calibration
larger fluid-loss rate than that of conventional poly- tests. After-closure reservoir radial flow determines
mer fracturing fluids. Consequently, it promotes a the formation transmissibility and initial pore pres-
larger dimensionless time and the earlier emergence sure. These parameters are of paramount importance
of radial flow during the shut-in period. Second, for optimizing a fracture stimulation (see Chapters 5
physical effects resulting from the viscoelastic nature and 10). In addition, spurt can be a significant consid-
of polymer fluids can potentially corrupt an extended eration for wall-building fluids in high-permeability

fracturing. However, it cannot be estimated from the

standard pressure decline analysis described in
Pumping Shut-in After-closure
Section 9-5. As previously mentioned, spurt influ-
• Fracture model • cf • G* • p* •κ • xfa
ences the linear flow response. Consequently, after-
closure linear flow analysis can be used to obtain an
improved on-site assessment of spurt over conven-
tional core-based laboratory tests.
A significant enhancement provided by after- Fluid loss
closure analysis is its potential to validate preclosure •η • CL
analysis. The reservoir linear flow regime retains two
critical pieces of information concerning the fracturing
process: the time when fluid loss terminated (i.e., frac-
ture closure time) and the fracture length. Cessation of
Fracture geometry Analyses Reservoir
fluid loss provides the linear flow period with a more agree
•L •L
distinct marker for fracture closure than that obtained
from conventional shut-in procedures (see “Estimating
closure pressure” in the Appendix to this chapter) that Figure 9-39. Interrelation of pre- and after-closure pres-
may exhibit multiple inflection points or a continuous, sure analyses.
gradual pressure variation during closing. Fluid-loss
discontinuity provides the fundamental basis for using duction parameters. The synergistic, complementary
after-closure analysis to determine the fracture closure character of the pre- and after-closure pressure analy-
pressure (see Section 9-6.8). ses provides a comprehensive suite of crossvalidated
The transitional period exhibits a unique dimen- fracture parameters for the optimized design of frac-
sionless knee time (see Fig. 9-38a) that provides ture treatments based on the economic guidelines
information related to the fracture length through provided in Chapter 10.
Eq. 9-75. The fracture length can also be estimated
from the preclosure pressure decline analysis (see
Section 9-5.2). The after-closure linear flow charac- 9-6.5. Physical and mathematical
teristics are derived from the principles of transient descriptions
reservoir flow, which are fundamentally different
The total pressure difference between the fracture and
from the principles of linear elasticity and material
the reservoir ∆pT can be divided into three compo-
balance for a fracture that govern the preclosure pres-
nents, as shown in Fig. 5-17. Of relevance to after-
sure analyses. A consistent calibration evaluation is
closure analysis is the pressure difference ∆pR in the
indicated by agreement between the fracture lengths
reservoir beyond the filter cake and filtrate regions.
estimated by these divergent analyses.
This pressure difference represents the added contri-
The preceding discussion introduces the interrela-
butions of the two sources of fluid leakoff: a contri-
tion between the pre- and after-closure flow periods,
bution from the Carter-based leakoff loss (i.e., the CL
which is summarized in Fig. 9-39. The four inputs of
component of fluid loss) ∆pRC and that resulting from
cf, G*, p* and κ are required to determine the fluid
spurt ∆pRS:
leakoff coefficient (Eq. 9-57) and the fluid efficiency
(Eq. 9-71). The value of cf is provided by the pressure ∆pR = ∆pRC + ∆pRS . (9-76)
response during injection (see Section 9-7). The pre-
closure shut-in period (see Section 9-5) estimates of As discussed in Section 9-6.3, the pressure changes
G* and p* and the previously discussed after-closure across the filter cake and filtrate disappear within a
analysis can be used to predict the spurt factor κ. The short time after fracture closure. The subsequent bot-
fracture length is independently obtained by these tomhole pressure measurements p(t) then reflect the
two approaches, and a consistent fracture analysis is reservoir response to the pressure changes and fluid-
indicated by agreement of the two length estimates. loss distribution induced during the propagation and
Finally, either the after-closure transitional or radial closing periods:
flow period is also used to assess the reservoir pro- ∆p(t ) = ∆pR (t ) = p(t ) − pi t > tc . (9-77)

A physical and mathematical description for the It follows from Eqs. 9-75, 9-78 and 12-1 that the
after-closure period is in this section. The focus is on dimensionless time tD, which corresponds to the
∆pRC, the reservoir response to Carter-based leakoff. apparent length xfa, is related to T as
The effects resulting from spurt loss are addressed
separately in Section 9-6.6. T = faL2 t D . (9-79)
The two primary changes that the reservoir experi- A general expression for faL in terms of the fluid-
ences following fracture closure are change in its loss parameters was given by Nolte et al. (1997):
boundary condition and change in its representation
in reference to the response during production of a g0  κ 
fixed-length fracture. First, as discussed in “Pressure faL =  . (9-80)
characterization for a propagating fracture” in the
(1 − η)  g0 (κ − 1) + 2 tc t p 
Appendix to this chapter, the pressure difference ∆pRC For the case of inefficient fracture propagation
in the reservoir is constant during fracture propaga- (η → 0) and negligible spurt (κ → 1), g0 = π/2
tion. In contrast, the after-closure period is character- and tc → tp. Equation 9-80 then indicates that
ized by cessation of the fluid-loss flux. Consequently, faL = π/4 (Fig. 9-35a). For spurt-dominated treat-
the boundary condition along the closed fracture ments (κ >> 1), the specific volume-loss distri-
changes from one of constant, uniform pressure to bution over the fracture length is approximately
zero or constant flux. The change in the boundary constant, as illustrated in Fig. 9-35b. As expected
condition is referred to as a mixed-boundary condition. for this case, Eq. 9-80 indicates that faL approaches
The second difference between the preclosure and unity.
after-closure periods pertains to the representation of
• Linear flow
the after-closure period, which is analogous to the
problem of a fixed-length fracture used in pressure The similarity between reservoir transient flow and
transient analysis. The leakoff distribution following heat transfer arises because of the underlying “dif-
fracture closure is shown in Fig. 9-35. The illustrated fusional” process that governs these two physical
volume profile closely reflects the distribution phenomena. The physical concept of heat conduc-
expected for a fixed-length, uniform-flux fracture and tivity is similar to that of reservoir mobility,
can be used to represent the after-closure pressure whereas the heat capacity of a solid body is equiv-
response. This observation contrasts with the pre- alent to the reservoir storage. An expression for
closure conditions of a propagating fracture, for ∆pRC during the after-closure linear flow period for
which the fluid-loss flux distribution is better repre- the mixed-boundary condition can be adapted from
sented by a fixed-length, uniform-pressure fracture. a similar condition presented by Carslaw and
This distinction is elaborated in “Comparison of Jaegar (1959) for heat transfer (Nolte et al., 1997):
fixed-length and propagating fractures” in the πµ
Appendix to this chapter. ∆pRC (t ) = CL t = tc
kφct
• Apparent fracture length πµ  2 −1  tc  
= CL  sin    t > tc
The previous discussion indicates that the after- kφct  π  t  
closure reservoir response can be analyzed by con-
πµ  2 tc 
sidering the leakoff flux to be uniformly distrib- ≈ CL   t > 3tc ,
uted over an equivalent fixed-length fracture. The kφct  π t 
apparent length of the equivalent fracture xfa is (9-81)
generally smaller than the physical length L
attained by the fracture during propagation. Its where t is the time since fracture initiation.
ratio with respect to the physical length is defined Equation 9-81 is based on 1D heat transfer
as the apparent length fraction: through the surface for an infinitely long fracture.
It also applies to the wellbore pressure of a finite-
x fa length fracture in linear flow because the reservoir
faL = . (9-78)
L surrounding the wellbore is experiencing 1D reser-
voir flow (Fig. 9-37a).

• Radial flow alent to the standard Horner function used during

No relation analogous to Eq. 9-81 for a finite- pressure transient testing.
length fracture with a mixed-boundary condition
applicable to radial flow has been presented in the
literature. This reservoir response can alternatively 9-6.6. Influence of spurt loss
be described by a superposition relation with the The preceding discussion of the reservoir response
logarithmic time function that is commonly used does not consider the effect of spurt. Its considera-
to describe the radial flow response for a fixed- tion, however, is important because spurt can domi-
length fracture (Eq. 12-15): nate the wall-building fluid-loss behavior of moder-
ate- to high-permeability formations.
µ Vi  1  λtc + (t − tc )   The role of spurt during either the pre- or after-
∆p(t ) =  ln 
2 πkh tc  2 λ  t − tc   closure period is difficult to describe with simple ana-
µ Vi lytical concepts. However, characterizing its contribu-
≈ t – tc >> λtc . (9-82) tion to the reservoir pressure difference ∆pRS was
4 πkh t − tc
found to be relatively simple, based on the results
The terms preceding the brackets on the right- of numerical simulations (Nolte et al., 1997). These
hand side of Eq. 9-82 are incorporated to ensure indicate that during the linear flow period spurt
that its long-term asymptotic behavior provides the causes a time-dependent increase in the pressure. Its
well-known response for an impulse test (see pressure response ∆pRS is added to the pressure con-
Sidebar 9G). tribution from Carter-based leakoff behavior
Equation 9-82 also incorporates the apparent (Eq. 9-81). The total reservoir pressure difference
closure time λtc. The apparent production time ∆p during the after-closure linear flow period that
used for the mixed-boundary condition in standard includes the spurt contribution can be expressed in
pressure transient analysis (Horner, 1951) is terms of ∆pRC and the spurt factor κ as
defined as the ratio of the cumulative volume and
2  t  κ − 1 tp 
the final production rate. This definition of appar- ∆p(t ) = ∆pRC (tc ) sin −1  c  + t ≥ tc .
ent time is not applicable for a propagating frac- π  t 2 t 
ture. Rather, the multiplier λ was selected from
(9-84)
numerical simulations as the value that provides
the shortest after-closure time for the application In particular, the pressure difference during an
of Eq. 9-82. For the mini-falloff test and its related extended period of linear flow can be approximated
conditions of negligible efficiency and spurt, the from Eqs. 9-81 and 9-84 as
optimized value of λ can be characterized as
πµ  2 tc 
∆p(t ) ≈ CL (1 + fκ )   t > 3tc , (9-85)
λ = 1+
0.14
η → 0, κ → 1. (9-83) kφct  π t 
Tp
where the spurt fraction fκ is defined as
Equations 9-82 and 9-83 provide the reservoir
π tp
transmissibility with acceptable accuracy signifi- fκ = (κ − 1) . (9-86)
cantly before the actual occurrence of radial flow. 4 tc
For example, for a dimensionless pumping time Tp The distinctive time behavior exhibited by spurt
= 0.1, the optimized multiplier expression enables during the after-closure period is illustrated in
the application of Eq. 9-82 at a dimensionless time Fig. 9-40, which is based on the linear flow time
T = 0.5 as opposed to T = 2 required for a standard function FL(t/tc) that is subsequently defined in
radial flow analysis. Furthermore, Eq. 9-83 shows Eq. 9-88. As discussed in Section 9-6.8, this presenta-
that λ is equal essentially to unity for well-devel- tion is consistent with the diagnostic log-log plot for
oped radial flow (i.e., Tp > 5). For these conditions, flow regime identification. The reservoir pressure dif-
the effect of the mixed-boundary condition dimin- ference is normalized with respect to ∆pRC(tc) and is
ishes. The time function in Eq. 9-82 then is equiv- compared for the conditions of negligible spurt (κ =

• increase in the apparent length that delays the knee

time
η = 0.20
Tp = 0.01 • deviation from the expected pressure response dur-
1 ing linear flow that lasts over a shut-in time period
difference and derivative
Knee time
equal to approximately 3tc.
Normalized pressure
κ=5 9-6.7. Consistent after-closure

0.1 diagnostic framework
κ=1 The theoretical relations outlined in the previous sec-
Difference tions can be readily distilled into an elementary set of
Derivative
equations. These simplified relations provide a con-
1 10 100
sistent framework for analyzing after-closure behav-
1/{FL(t/t c )} 2
ior and are summarized in this section.
• Linear flow
Figure 9-40. Simulated pressure for spurt and nonspurt
cases (Nolte et al., 1997). The relation in Eq. 9-84 is used to describe the lin-
ear flow behavior because it provides a general
1) and dominant spurt (κ = 5). An efficiency of 0.2 mathematical representation for this flow regime.
is prescribed for both cases. Under these conditions, The equation can be reformulated as
Eq. 9-48 implies that 5 times more fluid is lost during ∆pR (t ) = mlf FL (t tc ) , (9-87)
propagation for this particular value of κ than for the
condition of negligible spurt. where FL(t/tc) is the linear flow time function and
Figure 9-40 shows that spurt results in a pressure mlf is the corresponding slope on a Cartesian plot:
difference ∆p and derivative that fall below the char-
2 −1  tc 
acteristic half-slope for linear flow over a time period FL (t tc ) = sin   (9-88)
of about t < 3tc (Eq. 9-85). The figure also indicates π  t
that spurt apparently decreases the knee time, which
is the time defined by the intersection of the asymp- πµ
totic pressure responses during linear and radial flow. mlf = CL (1 + fκ ) κ > 1, t > 3tc
kφct
The apparently earlier time of the knee in the pres-
πµ
ence of spurt occurs because the time function used = CL κ = 1, t > tc , (9-89)
in Fig. 9-40 is referenced to the closure time. Spurt, kφct
however, has a significant impact on the closure time
itself for a given value of efficiency (Eq. 9-60). For where the spurt fraction fκ is as defined in Eq. 9-86.
the spurt parameters used to develop Fig. 9-40, the The dependency of fκ on the spurt factor κ, as well
closure time for spurt is 7.5 times larger than its value as the closure time tc, implies that both parameters
in the absence of spurt. In reality, the dimensional cannot be simultaneously determined from the
knee time for the spurt case occurs later than for the reservoir linear flow response. An accurate value
nonspurt case (i.e., by a factor of 1.4 relative to the of tc is therefore required to properly characterize
beginning of pumping for these spurt parameters). κ for spurt-dominated conditions.
The later knee time for the spurt case results because In the absence of spurt, κ = 1, fκ = 0, and the last
of its larger apparent fracture length, as indicated by relation in Eq. 9-89 provides an expression analo-
Eq. 9-80. The relation between the apparent length gous to Eq. 2-50 for estimating the reservoir per-
and knee time is further discussed in Section 9-6.7. meability from the linear flow response, as sug-
In summary, spurt results in gested by Mayerhofer et al. (1993). The square-
root dependence of Eq. 9-89, however, makes this
• an added component to the reservoir pressure dif- procedure sensitive to the general uncertainty in
ference during the pre- and after-closure periods the parameters that define the fluid-loss coefficient

and reservoir storage. Consequently, the technique Equations 9-87 and 9-92 suggest that a log-log
could introduce significant uncertainty in the diagnostic plot of the pressure difference and pres-
inferred permeability (Ispas et al., 1998). sure derivative based on {FL(t/tc)}2 would provide
• Radial flow a consistent basis for diagnosing both linear and
radial flow. The plot also would provide the half-
Equation 9-82 provides the general expression
slope and unit slope behavior during linear flow
of the logarithmic time function that describes the
and radial flow, respectively, as is expected for
reservoir pressure difference during radial flow.
after-closure pressure analysis.
This function can be approximated by the square
of its linear flow counterpart (Eq. 9-88) by intro- • Spurt
ducing a coefficient of π2/8: Spurt is estimated from the linear flow slope mlf
2 using Eqs. 9-86 and 9-89:
1  λt  π 2  2 −1  tc  
ln1 + c  ≈  sin    (1 + ε ) , 4  mlf kφct  t
2 λ  t − tc  8  π  t   κ = 1+  − 1 c . (9-94)
π  CL πµ  tp
(9-90)
The spurt factor κ can be related to the spurt
where ε is the error introduced by the approxi- coefficient Sp with Appendix Eq. 24.
mation. The relative error can be based on the
• Fracture length
pressure derivative because it is the primary mea-
sure for after-closure analysis. Thus, The fracture length is obtained from the dimension-
less time corresponding to the knee formed by the
ε = λ −  c 
4 t intersection of the asymptotic log-log slopes during
∆t > 3tc . (9-91)
 3   ∆t  linear and radial flow, in a plot similar to Fig. 9-38a.
A general relation for the fracture length can be
It follows from Eq. 9-91 that the error introduced
derived from the pressure response for a fixed-
by the approximation in Eq. 9-90 is at a minimum
length fracture and relating its behavior to that of
when the apparent time multiplier λ = 4⁄3. As indi-
the propagating fracture using the apparent length
cated by Eq. 9-83, this value of λ corresponds to a
fraction faL, as discussed in Section 9-6.5.
dimensionless injection time Tp = 0.4, which meets
The fracture length is required to verify the pre-
the desired condition of a mini-falloff test for early
closure analysis for a calibration test, where the
radial flow analysis. The approximation, therefore,
use of a more efficient fracturing fluid generally
is well suited for application to the mini-falloff test.
leads to a well-defined linear flow response. The
More generally, the relative error ε can be shown to
knee subsequently occurs only after a significantly
be less than 5% for the after-closure period recom-
longer after-closure shut-in period (Fig. 9-38).
mended in “Guidelines for the field application of
Under this extended shut-in condition, the after-
after-closure analysis” in the Appendix to this
closure response can be approximated by the
chapter and using the expression for λ provided
impulse assumption (see Sidebar 9G). The asymp-
by Eq. 9-83.
totic expressions for a fixed-length fracture for lin-
The radial flow period can then be represented
ear and radial flow, in terms of the classic dimen-
by substituting the approximation provided by
sionless time tD, follow from Eqs. 12-12 and 12-15,
Eq. 9-90 into Eq. 9-82:
respectively. The derivatives during these periods
{
∆p(t ) = mrf FL (t tc ) , }
2
(9-92) are
 π
dpD  2 t D
where the function FL(t/tc) is defined in Eq. 9-88 Linear flow
and mrf is the corresponding slope on a Cartesian = (9-95)
plot: dt D  1
Radial flow,
 2t D
π Vi µ
mrf = . (9-93) where pD is the standard dimensionless pressure,
16 khtc defined in Eq. 12-2. The knee time tD,knee, or the
dimensionless time for the crossing of the two

derivatives, is determined by equating the two reservoir’s perspective of the fracture length, as
expressions: derived from the knee time concept of Eq. 9-100:
1 κCL t p
t D,knee =
. (9-96) 4 faL mlf tc k φct
π = . (9-101)
(1 − η) π µ π
Equation 9-96 can be expressed in terms of the
dimensionless time T for a propagating fracture by The derivation of Eq. 9-101 also uses the defini-
substituting Eq. 9-79: tion of mrf from Eq. 9-93. The application of
Eq. 9-101 to a calibration test serves as a valida-
faL2
Tknee = faL2 t D,knee = . (9-97) tion check for its evaluation and provides the fun-
π damental relation to verify the parameters pre-
As discussed in Section 9-6.5, for negligible dicted by the pre- and after-closure analyses.
efficiency and spurt, faL = π/4. It follows from
Eq. 9-97 that Tknee = 0.2, as shown for these con-
ditions assumed by Fig. 9-38. An estimate of the 9-6.8. Application of after-closure analysis
actual fracture length L can be obtained in terms A systematic procedure for the application of after-
of the dimensional knee time tknee by combining closure analysis is presented in this section. “Guide-
Eqs. 9-75 and 9-97: lines for the field application of after-closure analy-
1 πktknee sis” in the Appendix to this chapter discusses the
L= . (9-98) operational considerations required to engineer an
faL φµct
injection testing sequence that provides the pressure
The value of tknee can be defined by mlf and mrf, data required for an objective, comprehensive appli-
which are more readily estimated during field cation of the after-closure analysis.
practice. This definition of tknee also uses the
• Background information
extended time approximation for FL (Eq. 9-81):
2
A reliable estimate of the reservoir pressure should
4t  m  be established prior to the evaluation process. This
tknee = 2c  rf  . (9-99)
π  mlf  can be achieved from previous analyses (e.g., well
tests or production evaluation) or data measured
Combining this expression with Eq. 9-98 gives prior to fluid injection. The estimate suitably con-
the fracture length as strains the pressure difference within the flow
regime identification process, which ensures objec-
1 4 ktc mrf
L= , (9-100) tivity in the after-closure analysis. The fluid leak-
faL π φµct mlf off can be determined using the shut-in analysis
where the apparent length fraction faL is from described in Section 9-5. The formation storage
Eq. 9-80. φct is established using standard well logs and
It is unlikely that both linear and radial flow will pressure-volume-temperature (PVT) data or
occur during the same shut-in period. Consequently, correlations.
applying the length relation in Eq. 9-100 for a cali- • Flow regime identification
bration test requires a combination of its linear The occurrence of after-closure linear or radial
flow response and the radial flow analysis from a flow is identified using the diagnostic log-log plot
preceding mini-falloff test. Equation 9-93 can be of the pressure difference and pressure derivative
used to anticipate mrf for the calibration test from based on {FL(t/tc)}2, as discussed in Section 9-6.7.
the injected volume and fracture closure time of The two quantities are plotted as a function of
the calibration test and transmissibility estimate 1/{FL(t/tc)}2 so that the shut-in time conventionally
obtained from a preceding mini-falloff test. increases from left to right along the x-axis
• Validation of pre- and after-closure analyses (Fig. 9-41).
As a final step, the fracture length estimated by During linear flow, the curves for the two quan-
decline analysis in Eq. 9-59 is compared with the tities show straight-line behavior separated by a

• Linear flow analysis

1.0
Pressure difference Generally, linear flow analysis is applicable only
Pressure derivative
to a calibration test (i.e., because a mini-falloff test
is designed to achieve early radial flow). During
Normalized pressure differernce
Linear flow Radial

flow the calibration test, the linear flow period is identi-
and pressure derivative
0.1
fied from the diagnostic log-log plot (Fig. 9-41)
2×
for estimating either the closure time or spurt.
1 Obtaining representative values of these param-
2
eters generally requires a good prior estimate of
0.01 the initial, undisturbed reservoir pressure.
In the absence of spurt, the closure time can
be estimated as the smallest value of tc that, when
used with the independently derived reservoir
1
1
pressure estimate, provides the linear flow diag-
0.001
1 10 100 nostic discussed in the preceding “Flow regime
1/ {FL(t/tc)}2
identification” step. If spurt is expected, the value
of tc should be independently determined using the
Figure 9-41. Identification of linear and radial flow using procedures outlined in “Estimating closure pres-
the diagnostic log-log plot. sure” in the Appendix to this chapter. The after-
closure linear flow analysis can then be used to
estimate the spurt, with the linear flow period ini-
factor of 2 and each has a half-slope. Radial flow
tially identified from the diagnostic log-log plot.
is characterized by a log-log slope that approaches
If linear flow is identified, a Cartesian plot of the
unity as well as an approximate overlying of both
pressure versus FL(t/tc) is constructed to determine
the pressure difference and pressure derivative
the slope mlf. Equation 9-94 is used to obtain the
curves.
spurt factor κ. The reservoir mobility k/µ required
• Reservoir parameters determination by this equation is derived from transitional or
The testing sequence should preferentially include radial flow analysis and CL from the decline analy-
a mini-falloff test to characterize the reservoir sis. The spurt coefficient Sp can then be obtained
parameters. Shut-in pressures monitored during the from Appendix Eq. 24.
mini-falloff test should initially be investigated for • Crossvalidation
radial flow using the log-log diagnostic described
Equation 9-101 is applied to compare the fracture
in the preceding step. If radial flow can be identi-
length predicted by the reservoir response to that
fied, a Cartesian plot of the pressure response dur-
derived from decline analysis (see Section 9-5) for
ing radial flow versus {FL(t/tc)}2 is constructed.
the calibration test. The radial flow slope mrf is
The slope of the straight-line portion on this graph,
anticipated from Eq. 9-93, with the reservoir trans-
or mrf, can be used with Eq. 9-93 to estimate the
missibility as previously determined from the
reservoir transmissibility. The pressure intercept of
mini-falloff test and Vi and tc defined as the vol-
the straight-line portion is an estimate of the initial
ume injected and closure time for the calibration
reservoir pressure. This procedure is analogous to
test, respectively. Agreement between the two
the conventional Horner analysis used for well
independently inferred fracture lengths indicates
testing.
correct evaluation of the calibration treatment.
In the absence of radial flow, the transitional
flow-based analysis described in “Comparison
of fixed-length and propagating fractures” in the 9-6.9. Field example
Appendix to this chapter should be applied to the
mini-falloff test. This type-curve matching proce- Application of the after-closure analysis methodology
dure estimates the initial reservoir pressure and its is illustrated by analyzing the pressure monitored
transmissibility based on the transitional flow during the field calibration tests in Fig. 9-10. Analysis
period. of the injection pressure for the calibration test

described in “Example of radial fracture growth”

1000
in Section 9-4.4 determined that the fracture is best
Pressure difference and pressure derivative (psi)

Pressure difference
described by the radial fracture geometry model. The Pressure derivative
fracture is approximated as having a rectangular shape
of equal area to facilitate the after-closure analysis.
The shut-in pressure is analyzed in “Example field
application of radial fracture decline analysis” in
Section 9-5.4, from which the leakoff coefficient CL is
calculated to be 1.62 × 10–2 ft/min1/2. A closure pres-
sure of 4375 psi was inferred from the step rate test
(see “Estimating closure pressure” in the Appendix
to this chapter).
1
• Background information
1
The stabilized pressure on the bottomhole gauge 100
is 3726 psi in Fig. 9-10. The stabilized pressure 1 10
measurement provides an independent, objective 1/{FL(t/tc)}2
assessment of the reservoir pressure. A log-deter-
mined porosity of 19% and a saturation-weighted Figure 9-42. Identification of after-closure radial flow.
formation compressibility of 8.0 × 10–5 psi–1 were
also obtained. PVT analysis indicated a relatively
4400
high oil viscosity of 4 cp.
4300
• Radial flow identification 4200 660
Pressure (psi)
After-closure pseudoradial flow is observed in the 4100

reservoir for the initial mini-falloff test (Fig. 9-10), 4000
as shown by the diagnostic log-log plot in Fig. 9-42. 3900
A Cartesian plot of the pressure versus {FL(t/tc)}2 3800
during the after-closure period is in Fig. 9-43. An 3700 3736 psi
initial reservoir pressure estimate of 3736 psi is 3600

obtained as the y-axis intercept of the straight-line 0 0.2 0.4 0.6 0.8 1.0
period on the plot in Fig. 9-43. This value indi- {FL(t/tc)}2
cates consistency with the previously inferred esti-
mate of 3726 psi from the bottomhole gauge. The Figure 9-43. After-closure radial flow analysis.
slope mrf of the straight line is 660 psi and the clo-
sure time tc is 3.75 min, based on a closure pres- • After-closure linear flow determination
sure of 4375 psi (see “Estimating closure pressure”
in the Appendix to this chapter). A total volume of The shut-in pressure measured following the cali-
14.75 bbl was injected during the test. The trans- bration test is investigated for after-closure linear
missibility is obtained by using this information flow behavior. The observed potential period of
with Eq. 9-93 reformulated in dimensional form: linear flow is confirmed by pressure derivative
analysis (Fig. 9-44). The initial pressure pi of
kh Vi 1 3724 psi required for the analysis is in excellent
= 2.5 × 10 5
µ tc mrf agreement with the 3736-psi reservoir pressure
14 .75 1 derived from the radial flow analysis and its value
= 2.5 × 10 5 (9-102) of 3726 psi inferred as the stabilized pressure mea-
3.75 (660)
= 1490 md - ft / cp. surement on the bottomhole gauge prior to injec-
tion (Fig. 9-10). The linear flow slope mlf is
The transmissibility estimate implies a reservoir deduced to be 815 psi from the corresponding
permeability of 80 md based on the reservoir para- Cartesian plot.
meters.

the conditions of high viscosity and low compress-

Pressure difference and pressure derivative (psi) 1000 ibility, as in this example.
Pressure difference
Pressure derivative • Reconciliation
Equation 9-80 is used to obtain the apparent frac-
ture length faL:
π2  1.02 
faL =   = 0.86.
100
2×
(1 − 0.16)  π 2 × (1.02 − 1) + 2 5.6 4.6 
1
2 The slope mrf for the calibration injection is pre-

dicted using Eq. 9-102, the dimensional form of
Eq. 9-93:
Vi  µ 
mrf = 2.5 × 10 5 ×
10 tc  kh 
1 10 100
= 2.5 × 10 5 ×
107  1 
1/{FL(t/tc)}2
5.6  1490 
= 3205 psi.
Figure 9-44. Identification of after-closure linear flow.
The fracture length is estimated using the
dimensional form of Eq. 9-100:
• Spurt estimation
The previously determined linear and radial flow 1 ktc m rf
L = 2.37 × 10 −3
slopes are used with the estimates of formation faL φµct mlf
porosity and compressibility to estimate spurt with 1 80 × 5.6 3205
Eq. 9-94, which is reformulated in oilfield units as = 2.37 × 10 −3
0.86 0.19 × 4 × 8.0 × 10 −5 815
4 mlf kφct  t = 33 ft.
κ = 1+ 1.18 × 10
−3
− 1 c
π CL µ  tp The length estimate is in good agreement with
the 37-ft fracture radius determined from the shut-
1.18 × 10 −3 815  in analysis, as described in Section 9-5.4.
4  1.62 × 10 −2  5.6
= 1+  
π  80 × 0.19 × 8 × 10 −5  4.6
× − 1 9-7. Numerical simulation of
 4 
pressure: combined analysis
= 1.02 ,
of pumping and closing
where the closure time of 5.6 min was based on
The pressure analysis techniques presented in this
the closure pressure of 4375 psi determined with
chapter are based on analytical derivations. They
the step rate test (see “Estimating closure pressure”
assume primarily a simplified fracture behavior to
in the Appendix to this chapter).
provide an engineering approximation of the fracture
The spurt coefficient Sp is determined with
parameters. A more comprehensive analysis is
Appendix Eq. 24:
required to determine rational changes in the propped
Sp = 7.48 × 10 2 g0 CL t p (κ − 1) treatment design (e.g., fluid viscosity, leakoff coeffi-
cient changes because of additives) without the nega-
= 7.48 × 10 2 × ( π 2) × 1.62 × 10 −2 4.6 × (1.02 − 1)
tive impact of trial and error adjustments during com-
= 0.8 gal / 100 ft 2 . mercial field development. This understanding is
particularly important for reservoirs where formation
The nominal value of Sp, in spite of the high for- pressure capacity considerations (see Section 9-4.6)
mation permeability, reflects that the fluid leakoff require unconventional designs (Nolte, 1982, 1988a;
is governed by the reservoir. This can occur for

Warpinski and Branagan, 1988) to produce the with any inverse problem is the occurrence of
desired fracture geometry. nonunique solutions. With respect to hydraulic frac-
These objectives are best achieved with an appro- turing, nonuniqueness arises when more than one
priately calibrated numerical simulator that adequately combination of the unknown parameters can be used
replicates the important processes that occur during a to approximate the observed pressure.
fracturing application. The simulator that best meets A multitude of parameters govern fracture behavior
these objectives is most likely not one of the compre- in a nonlinear and coupled fashion. Uncertainty in
hensive, computationally intensive simulators that are some of the parameters leads to erroneous prediction
better suited to research environments (e.g., Plahn et of the unknown parameters inferred by the pressure-
al., 1997). The description of the overall fracture matching process. A nonunique parameter set that
behavior required for routine engineering applications produces an incorrect prediction of the fracture geom-
can generally be satisfied by P3D fracture geometry etry results in an inappropriate redesign of the
models (see Section 6-3). propped treatment. This effect is readily apparent
in Eq. 9-24, in which the pressure response during
pumping exhibits an approximately first-order inverse
9-7.1. Pressure matching dependency on the fracture compliance cf. For the
A typical approach for calibrating a numerical frac- commonly occurring PKN-type fracture, cf ≈ hf /E′
ture simulator is through iterative pressure matching. (Eq. 9-22). Consequently, an inaccuracy in the value
The fracture simulator is executed with varying para- of the plane strain modulus E′ produces a correspond-
meter inputs to visually match the simulated pressure ing error in the predicted fracture height hf and there-
with the field measurements. The quality of the pres- fore the fracture penetration, without affecting the
sure match is assessed by balancing objectivity, phys- quality of the pressure match.
ical insight and constraints from other data sources. Nonuniqueness in pressure matching may also
However, given the limited number of formation occur during periods of controlled and uncontrolled
parameters that can be accurately characterized using height growth. During both these conditions, the net
wireline logs or laboratory tests, the procedure can pressure response is determined largely by the stress
require significant effort when several parameters are difference between the bounding zones and the pri-
not known. A large number of unknown parameters mary interval. The pressure response exhibits a much
may also compromise the objectivity of the matching reduced sensitivity to other fracture parameters, such
procedure. as E′ or CL. Consequently, a reasonable pressure
An alternative approach for pressure matching match during these periods may be attained in spite
relies on an algorithm that repeatedly executes the of errors in the other fracture parameters.
fracture simulator to improve the estimates of the Nonuniqueness is avoided by preventing arbitrari-
fracture parameters within specified bounds until the ness during the pressure-matching process. This is
best match to the field-measured response is obtained. best achieved by initially applying the diagnostic pro-
This semiautomated approach, discussed in Section cedures outlined in this chapter to obtain a first-order
6-12, is based on conventional principles of nonlinear estimate of the unknown parameters and an approxi-
regression analysis for system identification. The mate understanding of the nature of the fracture
technique, initially introduced by Gringarten et al. growth. Only moderate variation of the unknown
(1979) for well test analysis, has also been applied parameters from these initial estimates should be
to log interpretation and reservoir characterization. allowed during the pressure-matching process. In
addition, the numerical simulator must be adequately
constrained so that the simulated fracture geometry
9-7.2. Nonuniqueness conforms to the growth patterns inferred by the pres-
sure diagnostics; e.g., the simulator must predict con-
Pressure matching is an inverse problem for which trolled height growth if this behavior was concluded
the input (i.e., fracture treatment schedule) and output from pressure derivative analysis.
(i.e., fracture pressure) are known and are used to The pressure-matching process should also include
determine the unknown model parameters (e.g., zone all independently derived information. For example,
stresses, leakoff coefficient). An inherent limitation nonuniqueness resulting from cf could be overcome

by a rational selection of the primary zone height

using well logs. Young’s modulus E can then be cali-
brated by matching pressures during the stage of the
Pressure/injection rate
Step rate/
treatment that exhibits PKN behavior, before penetra- Mini-falloff flowback/rebound Calibration
tion of the bounding layers occurs. Alternatively, the
value of E could conform to core tests or sonic logs
of the reservoir (see Chapter 4), and the primary zone
thickness could be calibrated with the pressure match
and verified against well logs. Similarly, variation in
the zone stresses must be based on their trends as pre-
dicted by stress logs. In the absence of stress logs, Time
lithology logs can be used to develop an approxima-
tion for particular geologic environments (Smith et
Figure 9-45. Formation calibration testing sequence.
al., 1989). Furthermore, the simulated fracture
dimensions could be compared against independent
measurements, such as radioactive tracers or micro- process based on these tests.
seismic measurements, to validate the outcome of the 1. Review existing information pertinent to fracturing
pressure-matching exercise (Gulrajani et al., 1998). A properly calibrated sonic log (see Chapter 4)
The inversion solution is confirmed by applying the provides an excellent basis for assessing the for-
numerical simulator based on these calibrated fracture mation stress profile. It can also be used to identify
parameters to predict the fracture behavior during the lithologic aggregates for developing an under-
propped treatment or for offset wells. An objective standing of the zone layers. In its absence, a stan-
pressure evaluation is indicated by agreement between dard gamma ray or spontaneous potential log
the simulated pressure response and actual field mea- should be applied to define the depth and thickness
surements. Conformity also establishes the ability of of the layers. The permeable zone thickness (i.e.,
the calibrated numerical simulator to develop the thickness over which fluid loss occurs) can be
improved designs for future fracture treatments. identified using specialized magnetic resonance
logs that directly measure permeability. This infor-
mation can also be obtained from the separation
9-8. Comprehensive calibration of various resistivity measurements, which indi-
test sequence cates drilling fluid invasion and therefore a finite
permeability.
An optimized hydraulic fracturing program integrates
An estimate of the formation stresses can be
the analyses and diagnostic methods discussed in this
derived by history matching prior fracture treat-
chapter with the proper design process (see Chapters
ments in offset wells. This information can be used
5 and 10) and execution aspects (see Chapter 11) of
to characterize the fluid leakoff behavior as well
the stimulation procedure. These considerations
as calibrate either sonic or gamma ray logs (Smith
include identifying candidate wells that will benefit
et al., 1989) to obtain the stress variation over the
from the stimulation treatment, determining the
depth of interest. Where available, laboratory core
appropriate injection rate and proppant addition
tests should be used to obtain related information
sequence as well as assessing the risks involved with
(e.g., mechanical properties, existence of natural
treatment execution. The techniques described in the
fissures, fluid-loss behavior).
previous sections provide a basis for estimating the
fracture and reservoir parameters required during the 2. Determine reservoir production parameters
optimization process. To ensure completeness and The reservoir pressure and permeability are best
consistency in the evaluation process, fracturing pres- determined from a well test. In its absence, the
sure data should be obtained through a planned information can be obtained using the mini-falloff
sequence of calibration tests, as shown in Fig. 9-45. test. The mini-falloff test attempts to propagate a
The following steps describe the fracture evaluation short, inefficient fracture to attain the required

reservoir transitional or pseudoradial flow (see 6. Characterize fluid leakoff

Section 9-6) within an acceptable data-recording A pressure decline analysis performed during the
period. This objective is best achieved by applying calibration test closing period and the previously
the operational guidelines provided in “Guidelines calibrated fracture compliance produce the most
for the field application of after-closure analysis” reliable estimate of the fluid-leakoff parameters
in the Appendix to this chapter. Using an a priori (see Section 9-5). They also determine the treat-
estimate of the reservoir pressure reduces uncer- ment efficiency, using an estimate of spurt from
tainty in the after-closure analysis. the after-closure linear flow analysis. The fluid-
Pore pressure and permeability are the bases of loss parameters are used to provide a measure of
the reservoir’s producing potential and are required the fracture length (see Section 9-5.2).
for an economics-based design (see Chapters 5 and The shut-in pressure during the closing of a cali-
10). bration test can also be used to infer the closure
3. Define closure pressure pressure (see “Estimating closure pressure” in the
The preferred method for objectively determining Appendix to this chapter). This analysis, however,
the fracture closure pressure is a step rate/flowback is highly subjective, and caution should be exer-
test (see “Estimating closure pressure” in the cised during its application.
Appendix to this chapter). For low to moderate 7. Assess after-closure response
permeabilities, the step rate test typically follows Calibration test shut-in pressure should be moni-
the mini-falloff test (Fig. 9-45). In higher perme- tored for a duration that is at least twice that of the
ability formations, it follows the calibration test closure time tc. This enhances the prospect that a
because the subsequent loss of completion fluid majority of the recorded after-closure response
with which the step rate test is performed is signif- period is not masked by complications arising
icantly reduced relative to fluid loss before the cal- from polymer injection and that a well-established
ibration test. linear flow response is obtained. The pressure dur-
4. Characterize fracture geometry ing the linear flow period is used to estimate spurt
The character of fracture growth is identified dur- (see Section 9-6.8), which in conjunction with the
ing a calibration test by the analysis of pressure shut-in analysis provides the treatment efficiency.
during pumping (see Section 9-4). The analysis Linear flow may be completely absent in a high-
also defines the formation pressure capacity (see mobility reservoir.
Section 9-4.6) and can be used to calibrate the The operational guidelines outlined in “Guide-
fracture compliance through pressure matching. lines for the field application of after-closure
The injection pressure is used to determine the analysis” in the Appendix to this chapter should be
bounding stresses (Eq. 9-39). This additional stress applied to prevent additional contaminating effects
measure, in conjunction with an estimate of the on the after-closure pressure data, which could
closure pressure, provides the basis for a linear lead to an inconclusive linear flow analysis.
calibration for correcting a sonic-based stress log 8. Crossvalidate evaluation results
for the particular lithologic setting (Nolte and Information derived from the after-closure linear
Smith, 1981; Nolte, 1988a). flow response, in combination with the previous
5. Identify near-wellbore problems estimate of reservoir transmissibility, provides an
A short step-down test (see Sidebar 9E) at the end independent perspective for the fracture length
of the injection period can be used to identify the (see Section 9-6.8). Agreement between this length
magnitude and severity of near-wellbore problems. estimate and that determined from the preclosure
Net pressure analyses (e.g., log-log plots) should analysis indicates consistency in the pressure
be based on the bottomhole pressure corrected to analysis.
reflect measurements of near-wellbore pressure
that is not associated with the fracture behavior.

9. Conduct pressure history matching 10. Verify against propped treatment pressure
The previously obtained fracture parameter esti- response
mates (i.e., zone mechanical parameters and fluid The evaluation can be further refined for future
properties) can be refined through history matching treatments by matching pressures during the
the simulated pressure with the pressure measured propped treatment and comparing the required
during the calibration test. Care should be taken, parameter set for this treatment with that obtained
however, to prevent the occurrence of nonunique- from the calibration treatment. Conformity
ness, an inherent limitation in pressure matching between these two parameter sets provides confi-
(see Section 9-7.2). This shortcoming can be pre- dence in the fracture parameters. In the case of
vented by performing a comprehensive evaluation disagreement, a reanalysis should be conducted
and testing procedure to limit the number of with emphasis on deviations from ideal behavior.
unknowns as well as eliminate unrealistic solutions. A more authoritative measure of the correctness
of the predicted parameters is agreement among
parameter sets obtained in offset wells, with suit-
able corrections made to account for differences
in the formation thicknesses, well completions,
fracturing schedules or fracturing fluid systems.


◆
Appendix: Background for
Pressure Analysis Techniques
Sunil N. Gulrajani and K. G. Nolte, Schlumberger Dowell
Estimating closure pressure is created, and the shut-in pressure is commonly

used as a first-order approximation of the stress.
The distinction between the formation closure pres- Significantly higher net pressure, however, occurs
sure and the local stress is discussed in Sidebar 9A. with the procedures employed to determine pc. Thus,
The closure pressure pc is distinguished from the pc can be considerably different from the instanta-
minimum stress σmin because σmin generally varies in neous shut-in pressure (ISIP), and it must be esti-
magnitude and direction over the gross pay interval mated with alternative procedures.
(i.e., the zone between meaningful stress barriers) Techniques commonly used to determine pc are
whereas pc provides a global average for the interval. the step rate, shut-in decline and flowback tests
Field procedures for estimating these two stress (Appendix Fig. 1) (see Section 3.6-2). Step rate and
measures differ in two primary ways. First, estimat- flowback tests are conducted exclusively to deter-
ing the magnitude of the local stress requires the cre- mine pc. The shut-in decline test is used in conjunc-
ation of a small fracture by using a relatively small tion with a calibration test, which is usually con-
fluid injection rate and volume. Determining pc ducted to quantify fluid-loss behavior (see Section
requires the creation of a fracture over the entire 9-5). The value of pc can also be estimated from
thickness of the gross pay interval, which requires injection pressure derivative analysis (see Sidebar
a much larger injection rate and volume. Second, a 9C) or after-closure linear flow analysis (see Section
smaller net pressure occurs when a smaller fracture 9-6).
(a) (b)
Bottomhole pressure
Injection rate
at step end
pc
Extension
pressure
Time Injection rate

(c) (d)
Pump-in
Bottomhole pressure
Bottomhole pressure
Flowback
a pc
pc
b
Possibilities
c
Time √∆t or √∆t + tp
Appendix Figure 1. Tests to determine closure pressure (after Nolte, 1982): (a) step rate test, (b) bottomhole pressure plot-
ted versus injection rate to infer the values of pc and the fracture extension pressure, (c) combined step rate and flowback
tests (a = rate too low, b = correct rate for pc at curvature reversal and c = rate too high) and (d) shut-in decline test dis-
played on a square-root plot (∆t = shut-in time and tp = injection time into the fracture).

Step rate test

9200
The step rate test is conducted solely to determine pc. pc = 8910 psi

In low-permeability reservoirs, the test generally is 8900
performed with completion fluids (e.g., treated
Extension pressure = 8950 psi
water). The use of polymer fluids may be required in
8600
high-permeability reservoirs (Smith, 1985) to control
potential fluid loss and ensure fracture creation at
lower injection rates. 8300
The step rate test can be conducted as the pumping
phase of either the flowback or shut-in decline test. 8000
The time duration of the individual injection steps 0 3 6 9
should be equal (Appendix Fig. 1a) and can be rela-
tively small (i.e., the time required for the pumps to
Appendix Figure 2. Pressure versus rate analysis for the
change and maintain a constant rate and the pressure step rate test in Fig. 9-8.
to be recorded, typically 1 or 2 min). The injection
rate increments over the successive steps should also
be approximately the same. When the step rate test is should not be used in preference to also obtaining a
conducted as part of a flowback or decline test, the well-defined straight line for matrix injection, which
last step is maintained for a longer duration (i.e., 5 is the only means to ensure that fracture extension
to 10 min) to ensure the creation of a sufficient-size occurred during the test.
fracture. Injection rates typically vary from 1 to The y-axis intercept interpretation was used to ana-
10 bbl/min in moderate-permeability formations, and lyze the step rate test illustrated in Fig. 9-10b. The
the rates in low-permeability formations are about pressure versus rate plot is entirely characterized by
one-half of these values. Ideally, the injection rates the fracture extension line (Appendix Fig. 3), which
for three of the steps should fall below the extension has a y-axis intercept of 4375 psi for the estimated
pressure to define matrix injection prior to fracturing, value of pc.
and a similar set of values should be obtained above An indication of a valid step rate test is that the
the extension pressure. extrapolated pressure for the zero rate of the line rep-
The fracture closure and extension pressure are resenting matrix injection should equal approximately
inferred from a crossplot of the bottomhole pressure the bottomhole pressure preceding the test. The zero-
at the end of each injection step and the injection rate rate pressure is the reservoir pressure, if significant
(Appendix Fig. 1b). The plot is characterized by two amounts of fluids were not previously injected. This
straight lines, with matrix injection represented by the quality control check can also be used while conduct-
line with the steeper slope and fracture extension ing the test when no apparent slope change is
characterized by the shallower sloping line. The inter- observed for the pressure versus injection rate plot.
section of the two lines provides an estimate of the
fracture extension pressure, which is the upper bound 4800
for pc. Typically, the extension pressure is 50 to

4700
200 psi greater than pc, as shown on the pressure
versus rate plot in Appendix Fig. 2 for the step rate 4600
test in Fig. 9-8b. This larger value represents effects
from fluid friction pressure within the fracture and 4500
resistance to fracture extension (i.e., toughness).
Laboratory tests (Rutqvist and Stephansson, 1996) 4400
pc = 4375 psi
also indicate that the y-axis intercept of the shallower
4300
sloped line that represents fracture extension on the 0 5 10 15 20 25
crossplot provides a first-order approximation for pc, Injection rate (bbl/min)
even when the steeper line that represents matrix
injection is absent. This interpretation, however, Appendix Figure 3. Pressure versus rate analysis for the
step rate test in Fig. 9-10.
A9-2 Chapter 9 Appendix: Background for Hydraulic Fracturing Pressure Analysis Techniques

Shut-in decline test ful fraction of the in-situ leakoff rate). With the
assumption that a fracture has been created, the pres-
The shut-in decline test can be used with either a step sure response during flowback has two distinctly dif-
rate or calibration test. The decline data are displayed ferent profiles while the fracture is closing and after
on a square-root plot (Appendix Fig. 1d) or a G-plot the fracture closes. Comprehensive simulations (Plahn
(Fig. 9-29) that assumes square-root exposure time et al., 1997) indicate that the fracture pc is identified
for the fluid-loss behavior. The closure pressure is by the intersection of the two straight lines that define
inferred where the slope changes on either plot. The these two periods. The increasing rate of pressure
derivative should be used to magnify the change of decline for the postclosure period results from fluid
slope and enhance its identification. flow through the pinched fracture width (i.e., induced
Either of the specialized plots, however, may be fluid choking) in the near-wellbore region induced by
completely devoid of a significant slope change or fluid flowback. The characteristic “lazy-S” signature
may exhibit multiple slope changes. In general, up exhibited by the pressure during the flowback period
to six events could be associated with a slope change: is in contrast to the multiple inflections commonly
• height recession from the bounding layers observed with the shut-in decline test. Therefore, the
• transition between fracture extension and recession flowback test provides a more objective indication of
closure relative to the decline test.
• fracture closure
Maintaining a constant flowback rate as the pres-
• postclosure consolidation of the polymer filter cake sure decreases is critical for a flowback test. This
and fracture face irregularities objective requires a field-rugged, debris-resistant flow
• reservoir linear flow regulator that both measures and controls the flow-
• reservoir radial flow. back rate. The flow regulator should be preset for the
desired rate at the pressure expected following the
Consequently, the shut-in test commonly fails to end of injection, and it should be isolated by a closed
provide an objective indication of pc and should not valve during pumping. Presetting the flow regulator is
be used as the primary procedure for determining it. best achieved by opening it during the last period of
Experience indicates that the square-root plot may the step rate test to establish the desired flowback rate
provide a better indication of closure for fluids that prior to the actual test. The effect of the additional
do not have effective fluid-loss control from wall- fluid loss can be compensated for by increasing the
building behavior, whereas the G-plot may provide a injection rate. Fluid injection is terminated once the
better indication for fluids with wall-building behav- desired constant flowback rate has been attained, and
ior. The analysis of decline data typically uses both this rate is then maintained throughout the flowback
plots to determine the value of pc. period.
Other specialized plots have been used, although Experience shows that an adjustable choke often
less frequently, to identify pc. These include the log- plugs because of pipe dope and other debris loosened
log shut-in plot (Elbel et al., 1984) and multidimen- into the wellbore during the injection period. A gate
sional derivative analysis. In contrast to the square- valve is preferable for controlling the flowback rate.
root and G-function plots, the interpretation In addition, a pressure sensor and fixed choke at the
philosophy of these plots is based on identifying end of the flowback line can be substituted for a flow
reservoir flow regime changes to obtain bounding meter to reliably measure the flow rate, particularly
values of pc. when the rate is low (i.e., 3 bbl/min or less). Tabu-
lated values for the pressure drop versus the flow rate
through standard choke sizes are used to select the
Flowback test choke size that will provide the pressure change nec-
The preferred method for determining pc is a combi- essary for a reliable pressure measurement at the
nation of the step rate test (with an extended last step) anticipated flowback rate.
and flowback test (Appendix Fig. 1c). The essential The flowback test is frequently repeated for verifi-
element of the flowback test involves a flowback cation and selection of a more optimum flowback
period at a constant rate that is between 1⁄6 and 1⁄4 of rate. The first flowback period should be of sufficient
the last injection rate (i.e., at a rate that is a meaning- duration to ensure that fracture closure in the primary

zone is recorded. Pressure during this test, however, For moderate to high fluid-loss conditions, the cali-
should be controlled to prevent the production of bration test generally exhibits a closing period that is
reservoir fluids into the wellbore. The second flow- less than the injection time, and prior estimate of pc is
back test may be preceded by either a step rate test or not required for a shut-in period that approximates
a constant injection rate and should end after the indi- the injection period. Conducting the step rate/flow-
cation of a curvature reversal that clearly confirms the back test after the fracture calibration test reaps the
pc estimate from the first test. following advantages from the prior fluid injection:
The pressure decline for the second test is limited
• Breakdown of the complete gross pay interval is
to obtain an optimum rebound pressure. The rebound
ensured.
pressure is the nearly constant pressure that occurs
following a short period of increasing pressure after • Residual fluid-loss control from unbroken polymer
shut-in of the flowback test (see Fig. 3-26b). This sta- or additives in the fracturing fluid enables fractur-
bilized pressure is generally near to or smaller than pc ing using moderate injection rates with the com-
and provides its lower bound estimate. The rebound pletion fluid.
pressure is also an effective tool for estimating pc • Because the poroelastic effects have stabilized, the
when near-wellbore flow restrictions, as discussed in step rate/flowback test better replicates the stress
Section 9-4.9, are large. During the flowback period, conditions that occurred during the preceding cali-
these near-wellbore effects cause prediction of a bration treatment.
lower estimate of pc because of the additional flow
restriction in the near-wellbore region. The wellbore
rebound pressure, however, is unaffected by the
restriction and should provide an accurate lower
Guidelines for the field application
bound estimate of pc (Plahn et al., 1997). of after-closure analysis
In conclusion, the combination of the upper bound The reservoir response during linear flow and either
estimate of pc from the intersection of the matrix and radial or transitional flow is required to conduct a
fracture extension lines on a step rate test, the lower comprehensive analysis of the after-closure period.
bound of pc determined from the rebound pressure Several factors, however, can prevent obtaining infor-
and the estimate of pc from the y-axis intercept of the mation for one or more of the flow regimes. For
fracture extension line as well as the intersection of example, radial flow may be not be attained within a
the two lines during flowback provides multiple, reasonable time period for the high injection rates and
independent values that establish a firm basis for leakoff-resistant fracture fluids used during conven-
defining pc. tional fracturing operations. In addition, the after-
closure pressure data may be compromised because
of several undesirable communication effects
Test sequence between the wellbore and formation. Therefore, the
Because the mini-falloff test (see Section 9-6.2) that calibration testing sequence must be engineered to
characterizes the production potential should be con- provide the requisite pressure data (Talley et al.,
ducted in an initially undisturbed reservoir, it should 1999). The following guidelines increase the likeli-
always be the first injection within the calibration hood of obtaining a comprehensive, objective after-
testing sequence (Fig. 9-45). The subsequent applica- closure analysis.
tion sequence of the step rate/flowback and calibra-
• Unless a bottomhole shut-off valve is employed,
tion/shut-in tests generally depends on the formation
the reservoir pressure should ideally be equal to
fluid-loss characteristics.
or greater than the hydrostatic pressure of the well-
For low fluid-loss conditions, the step rate/flow-
bore fluid. Vacuum-induced fluid injection violates
back test to establish pc should precede the fracture
the no-flow assumption of the analysis for the
calibration test where the closure time could be long
after-closure period.
(e.g., up to 3 times the injection time). This prior esti-
mate of pc can be used to ensure that the shut-in pres- • The wellbore must be free of gas to ensure that
sure during the subsequent calibration test is moni- correct values of the hydrostatic pressure and
tored during the complete time of fracture closure. injected volume are used in the analysis and to

minimize wellbore expansion effects during pres- • It is unlikely that both linear and radial flow will
sure falloff. For gas reservoirs, this can be achieved occur during a decline period (see Section 9-6.2).
by conducting the injection test before a production The testing sequence illustrated in Fig. 9-45 is sug-
period. Alternatively, the gas should be circulated gested to increase the likelihood of obtaining infor-
from the wellbore or bullheaded into the formation. mation pertinent to both the linear and radial flow
An extended shut-in period can be required after periods. The mini-falloff test should be applied to
bullheading to allow the pressure transient to dissi- determine the radial flow parameters. The subse-
pate before liquid injection resumes. A relatively quent calibration test is more likely to attain linear
small volume of gas injection ahead of the fluid is flow. The radial flow response for the calibration
acceptable—e.g., gas is circulated from the inter- test can be anticipated from the reservoir informa-
mediate casing but remains in the shorter liner. tion derived from the mini-falloff test (see “Frac-
• Circulating or bullheading (potentially with a long ture length” in Section 9-6.7).
shut-in period) may be similarly required to spot • To attain radial flow within a reasonable time
the fracturing fluid at the perforations for the frac- frame, the mini-falloff test should adhere to the fol-
ture calibration test. The residual reservoir response lowing injection rate criterion, presented in con-
from injection of a significant volume of low- ventional oilfield units:
efficiency wellbore fluid causes the linear flow
qi ( bbl / min ) k ( md ) CL
analysis to indicate an unrealistically high spurt
h f (ft )
≤ 4 × 10 −6
µ(cp) CR
( pc − pi (psi)).
loss (Talley et al., 1999).
• Like preclosure analysis, after-closure analysis is (1)
an inverse problem that is inherently nonunique
If the fluid loss is controlled by the reservoir, as
(see Section 9-7.2). The objectivity of after-closure
desired for the test, the ratio of the fluid-loss coeffi-
analysis can be improved by obtaining an a priori
cients CL and CR becomes unity and a higher injec-
estimate of the reservoir pressure, particularly if
tion rate is possible. The equation provides an
the after-closure period is abbreviated (e.g., before
equality for a dimensionless time of 1.0 (i.e., the
wellbore vacuum) and either the closure time
beginning of radial flow during the injection period
(Nolte et al., 1997) or spurt loss is inferred. The
on Fig. 9-38). The guideline requires using esti-
reservoir pressure estimate can be obtained
mates of the reservoir parameters and fluid-loss
– as the stabilized bottomhole pressure measured characteristics to design the mini-falloff test. In
prior to fluid injection into the reservoir general, Appendix Eq. 1 provides an operationally
– as the stabilized surface pressure measured prior reasonable rate for radial flow with a short moni-
to fluid injection into an overpressured reservoir toring period in a reservoir with a mobility greater
– from the surface pressure and hydrostatic col- than about 5 md/cp. For reservoirs with lower val-
umn estimated through an accurate measurement ues of mobility, transitional flow resulting from
of the fluid injected to completely fill the well- injection rates greater than guideline can be used
bore for an underpressured reservoir to determine the reservoir parameters with a type-
curve-based analysis.
– from an accurate reservoir pressure gradient
established for the field. • Volume has a minimal effect on dimensionless time
and hence the time for development of radial flow
• In deep or hot reservoirs, bottomhole gauges are
because of the quasistationary value of dimension-
necessary because wellbore fluid expansion from
less time for a constant injection rate. However, a
the decreasing pressure and increasing temperature
minimum volume must be pumped to ensure an
during shut-in decrease the hydrostatic pressure.
accurate measure of the volume injected through
Excessive expansion of the fluid may eventually
the perforations because the transmissibility is pro-
violate the no-flow condition to the degree that the
portional to the injected volume (Eq. 9-93).
longer term data are corrupted, particularly for
residual gas in the wellbore. Mitigation of these • If polymer fluids are used (see Section 9-6.3) the
effects, like for a wellbore vacuum, requires a pressure data obtained after fracture closure can be
downhole shut-in device. corrupted by continued consolidation (i.e., squeez-

ing) of the fracture faces and filter cake. The con- treatment efficiencies are derived subsequently in this
solidation period lasts for a time approximating the Appendix. The exponent generally decreases through-
combined injection and closing times. The calibra- out the injection time, but the change is relatively
tion test shut-in pressure should therefore be moni- small and can be ignored. For the commonly used
tored for a time interval that is at least 2 to 3 times crosslinked polymer fluids, the exponent is typically 0.6.
the total closure time tc. Similarly, the shut-in Three mechanisms govern the fluid filtration process
period for a mini-falloff test should be at least 4 to during fracturing, as discussed in Section 6-4. The
5 times the total closure time. These guidelines for square root of exposure time relation, introduced by
the shut-in time increase the likelihood that ade- Carter (1957), is applied to predict the cumulative
quate, representative pressure data are obtained for effect of these mechanisms on the rate of fluid loss.
a valid after-closure analysis. A more general expression used in this Appendix
includes fluid-loss behavior dominated by a non-
Newtonian filtrate. Consequently, in contrast to Eq. 5-1,
the flux uL, defined as the rate of fluid loss per unit
Mathematical relations for fluid loss leakoff area, is expressed as
Derived in this section of the Appendix to Chapter 9
2 CL
are the relations used for pressure evaluation prior to uL = , (5)
fracture closure. (t − τ)1−θ
where t is the time since the beginning of pumping,
CL is the fluid-loss coefficient, and θ is the fluid-loss
Fluid-loss relations exponent. The coefficient 2 in Appendix Eq. 5
Two fundamental relations are required to derive the accounts for leakoff over the two walls of the fracture.
expressions for fluid loss. The first relation describes The specific fluid-loss volume vL, defined as the
the nature of fracture growth with respect to time. fluid-loss volume VL per unit leakoff area, is obtained
The second relation is based on the expression for by the time integration of Appendix Eq. 5:
fluid loss introduced by Carter (1957). t
(t − τ(a)) .
2 CL
vL = ∫ uL dt =
θ
The evolution of the fracture area is assumed to (6)
follow a power law relation with time in which the 0
θ
area monotonically increases with time. The proper- Parlar et al. (1995) showed that θ is related to the
ties of the injected fluid and the pump rate are power law exponent nf of the filtrate that invades the
assumed to be relatively constant. The power law reservoir during the leakoff process:
expression relates any intermediate fracture area a
created at a time τ to the total fracture area A at the nf
θ= . (7)
current time t: 1 + nf
α
a  τ As discussed in Chapter 8, a filter cake is deposited
= (2)
A t by a wall-building fluid along the fracture in low-per-
meability reservoirs and within the formation in high-
1/ α
τ  a permeability reservoirs. A Newtonian filtrate (i.e.,
= , (3)
t  A water) is created during the process. Under these con-
ditions, nf = 1 and Appendix Eq. 5 reduces to Eq. 5-1,
where α is referred to as the area exponent. The which is Carter’s square root of exposure time rela-
exponent α is also the log-log slope of A versus t, as tion for the fluid-loss rate (i.e., θ = 1⁄2). Non-wall-
shown by differentiating Appendix Eq. 3 with respect building fracturing fluids invade high-permeability
to time: reservoirs. The resulting filtrate fluid is typically non-
t dA Newtonian, with nf < 1 and θ < 1⁄2. In addition, the
α= . (4) extensional viscosity behavior of viscoelastic filtrates
A dt
above a threshold filtration rate can exhibit relatively
The value of α depends on the fluid efficiency. large values of nf (Chauveteau et al., 1986). The
Bounding values for α corresponding to low and high value of θ in this case is greater than 1⁄2 and can

approach unity. A value of θ that is different from ∆tD, which is defined as the ratio of the shut-in time
1
⁄2 has also been proposed to model the effect of ∆t to the pumping time tp:
natural fissures (Soliman et al., 1990).
t − t p ∆t
∆t D = = = tαD − 1. (15)
tp tp
Fluid-loss volume with the Carter-based
leakoff model The time functions f(tαD,α,θ) and g(tαD,α,θ) can be
expressed in terms of ∆tD:
The rate of fluid loss associated with Carter-based
dξ
1
f ( ∆t D , α, θ) = ∫
leakoff behavior (i.e., the contribution of CL over an
∆t D ≥ 0 (16)
(1 + ∆t − ξ1/ α )
1−θ
elemental leakoff area da) can be obtained by substi- 0 D
tuting Appendix Eq. 3 into Appendix Eq. 5:
1
g( ∆t D , α, θ) =
1
(1 + ∆t D − ξ1/ α ) dξ ∆t D ≥ 0 . (17)
θ
∫
2r C da
q L ( da, t ) = 1p−θ L 1−θ , (8)
t p (tαD − ξ1/ α ) θ0
where rp is the ratio of the fracture surface area avail- Valkó and Economides (1993b) showed that the
able for fluid loss to the gross fracture area and tp is functions f(tαD,α,θ) and g(tαD,α,θ) are part of the
the injection or pumping time. The dimensionless hypergeometric family of functions or their subset
parameters tαD and ξ are defined as of incomplete beta functions (Meyer and Hagel,
1989). Either of these function families is relatively
t complicated, but simple analytical expressions can
tαD = (9)
tp be obtained for a limited set of values, as discussed
subsequently.
a
ξ= , (10)
Af
where Af is the fracture surface area at the end of
Cumulative fluid-loss volume
pumping. The total fluid-loss volume at the end of pumping
The total rate of fluid loss is obtained by the inte- VLp comprises the cumulative contributions of its
gration of Appendix Eq. 8 over the fracture area: CL fluid-loss component VLp,C and spurt VL,S:
VLp = VLp,C + VL ,S .
( )
2 rp CL A f (18)
qL Af , t = f (tαD , α, θ), (11)
t 1p−θ
An expression for VLp,C can be obtained by substi-
where the function f(tD,α,θ) is defined as tuting ∆tD = 0 into Appendix Eq. 13:
1
dξ VLp,C = VL ,C ( ∆t D = 0) = 2 rp CL t pθ A f g0 (α, θ), (19)
f (tαD , α, θ) = ∫ tαD ≥ 1. (12)
0 (t αD −ξ )
1/ α 1−θ
where the function g0(α,θ) represents the value of the
g-function in Appendix Eq. 17 when ∆tD = 0:
An expression for the CL component of the fluid-
1
(1 − ξ1/ α ) dξ .
loss volume VL,C is similarly obtained by substituting
g0 (α, θ) = g( ∆t D = 0, α, θ) =
1 θ
Appendix Eqs. 9 and 10 into Appendix Eq. 6 and θ0∫ (20)

integrating over the area:
Spurt occurs only for wall-building fluids. It ac-
VL ,C = 2 rp C t A f g(tαD , α, θ),
θ
L p
(13) counts for the fluid-loss volume prior to the creation
of a filter cake and is applicable only during the frac-
where the function g(tαD,α,θ) is defined as ture propagation period. Following the spurt period,
1 wall-building fluids exhibit a Newtonian filtrate during
g(tαD , α, θ) =
1
( tαD − ξ1/ α ) dξ tαD ≥ 1.
θ
θ0∫ (14) the fluid-loss process (i.e., nf = 1 or θ = 1⁄2). The total

volume of fluid lost to spurt follows from definition
The functions f(tαD,α,θ) and g(tαD,α,θ) are usually of the spurt coefficient Sp:
presented in terms of the dimensionless shut-in time

VL ,S = 2 rp Sp A f . (21) θ = 1⁄2 in Appendix Eq. 41, whereas Appendix Eq. 40

provides the upper bound. An upper bound of unity is
The cumulative fluid-loss volume during injection assumed because it provides simple, closed-form
is obtained by substituting Appendix Eqs. 19 and 21 expressions. The bounding values of α for a wall-
into Appendix Eq. 18: building fluid are therefore as in Eq. 9-43:
1
VLp = 2 rp κCL t p A f g0  α, θ = 
1 1 < α < 1.
θ= (22)
 2 2 2
Analytical equations for f(∆tD,α,θ), g(∆tD,α,θ) and
1
VLp = 2 rp C t A f g0 (α, θ)
θ
L p θ≠ . (23) g0(α, θ) are obtained by substituting the bounding
2 values of α into Appendix Eqs. 16, 17 and 20,
Appendix Eq. 22 is expressed in terms of the spurt respectively, and integrating the resulting expressions
factor κ, which provides for the increase in fluid loss (Nolte, 1979; Smith, 1985):
over the no-spurt condition:
sin −1 1 + ∆t −1/ 2
( D)
1
 α=
f ( ∆t D ) θ= 1
Sp
κ = 1+ = 2
( )
. (24)
 1 2 (1 + ∆t D ) − ∆t 1D/ 2 α =1
1/ 2
g0 α, θ = CL t p
2
 
2
The rate of increase in the fracture area decreases (28)
significantly at the end of pumping. Spurt-dependent
 1 + ∆t sin −1 1 + ∆t −1/ 2 + ∆t 1/ 2
( D) ( D)
1
fluid loss therefore also reduces relatively quickly fol- α=
g( ∆t D ) θ= 1
D
lowing the cessation of fluid injection and is assumed = 4 2
to terminate at the time t = tp. The total fluid-loss vol-
ume during a shut-in period is represented entirely by
2
3
(
 (1 + ∆t D ) − ∆t D3/ 2
3/ 2
) α =1
its CL component VLs,C (∆t): (29)
VLs ( ∆t ) = VLs ,C ( ∆t ). (25) π 2 α=

1
g0 θ= 1 = 2 (30)
An expression for VLs(∆t) results from subtracting 2 4 3 α = 1.
Appendix Eq. 19 from Appendix Eq. 13:
All these functions are well behaved. Their values at
VLs ( ∆t ) = 2 rp CL t A f any other value of α can be obtained through simple
interpolation between bounding values.
 1 
× g ∆t D , α, θ =  − g0  α, θ =   θ =
1 1
  2   
2  2
Non-Newtonian filtrate control
(26)
A non-Newtonian filtrate occurs with fracturing fluids
(
VLs ( ∆t ) = 2 rp CL t θ A f g( ∆t D , α, θ) − g0 (α, θ)) θ ≠ .
1 that exhibit a power-law-based rheology and do not
2 develop an effective filter cake. In this case, nf ≠ 1
and if the filtrate controls fluid loss, θ deviates from
(27)
its commonly assumed value of 1⁄2. These conditions
limit the general analytical expressions to g0 only.
Explicit integration of Appendix Eq. 20 gives
Newtonian filtrate control
Fracturing fluids produce a Newtonian filtrate follow- Γ (1 + θ)Γ (1 + α )
g0 (α, θ) = , (31)
ing the deposition of a filter cake. For this case, nf = 1, θΓ (1 + α + θ)
and Appendix Eq. 7 indicates that θ = 1⁄2. Analytical
expressions for fluid loss can be derived for bounding where Γ(x) is the gamma function.
values of α. The lower bound value corresponds to Analytical expressions for the fluid-loss rate and
negligible efficiency and is obtained by substituting volume functions (i.e., Appendix Eqs. 16 and 17,

respectively) can be obtained for only select values The evolution of the fracture area for high fluid
of their arguments that are applicable to specific field efficiencies is obtained from this expression as
applications. The upper bound of α = 1 applies also qi t
for θ ≠ 1⁄2, and the corresponding functions are Af = η → 1, (39)
w
f ( ∆t D , α = 1, θ) =
1
θ
[
(1 + ∆t D )θ − ∆t Dθ ] (32) where 〈w〉 is the fracture width averaged over the
—
fracture area. The bounding expressions for fracture

penetration in Eqs. 9-32 through 9-34 are derived by
g( ∆t D , α = 1, θ) =
1
θ(1 + θ)
( [ 1+θ
]
1 + ∆t D ) − ∆t 1D+θ . (33) combining Appendix Eqs. 37 and 39 with the fracture
width relations in Eq. 9-27, for which the fracture
Simple analytical expressions for these functions area Af is proportional to the half-length L for the
also result for the value of θ approaching unity, PKN and KGD models and to R2 for the radial model.
which is applicable for time-independent fluid-loss For the common occurrence of Newtonian fracturing
conditions. This fluid-loss behavior can be exhibited fluid filtrate (i.e., θ = 1⁄2):
by viscoelastic fluids with a rapidly increasing exten- L ∝ t 1/ 2 η→ 0
sional viscosity above a threshold filtration rate: PKN
L ∝ t ( 2 n + 2 ) ( 2 n +3 ) η→1
f ( ∆t D , α, θ = 1) = 1 (34)
L ∝ t 1/ 2 η→ 0
KGD
g( ∆t D , α, θ = 1) =
1
+ ∆t D . (35) L ∝ t ( n+1) ( n+2 ) η→1
1+ α
The integral forms for both functions are well R ∝ t 1/ 4 η→ 0
behaved and therefore can be evaluated by numerical Radial
R ∝ t ( 2 n+2 ) ( 3 n+6 ) η → 1.
integration for other values of their arguments.
The time dependency of the net pressure pnet for the
bounding values of η is obtained by combining its
Bounding values for fluid efficiencies dependency on the fracture penetration for the three
The fracture penetration and net pressure can be rep- fracture geometry models in Eq. 9-25 with the three
resented for the bounding values of high and low preceding equations to produce Eqs. 9-35 through 9-37:
fluid efficiencies. When the efficiency η → 0, all the
pnet ∝ t 1/ 4 ( n+1) η→ 0
injected volume can be assumed to have leaked off. It PKN
follows from Eq. 9-5 that pnet ∝ t 1 ( 2 n+3) η→1
Vi ≈ VLp . (36) pnet ∝ t − n /2 ( n+1) η→ 0

KGD
From the relation Vi = qit and combining Appendix pnet ∝ t − n ( n+2 ) η→1
Eq. 36 with the generalized expression for fluid loss
during pumping in Appendix Eq. 23: pnet ∝ t −3n /8( n+1) η→ 0
Radial
pnet ∝ t − n ( n+2 ) η → 1.
qi t
Af = η→ 0 (37)
2 rp CL t θ g0
∝ t 1−θ . Bounding values for the area exponent
For the alternative situation of extremely high fluid The bounding values for the area exponent α corre-
efficiencies (η → 1), the fluid-loss volume can be spond to negligible (η → 1) and dominant (η → 0)
ignored. In this case, Eq. 9-5 reduces to fluid loss to the formation. The upper bound α1 in
Eq. 9-43 assumes not only negligible fluid loss but
Vi ≈ Vfp . (38) also a constant fracture width. Because the width
generally increases during injection, the upper bound

for the area growth is less than unity and independent (

VLp ( ∆tso ) = 2 rp A f CL t p g( ∆t Dso ) − g0 , ) (46)
of the loss exponent θ. It is derived from the condi-
tion of η → 1 in Eqs. 9-32 through 9-34: where the dimensionless time ∆tDso is defined as the
ratio of the time after screenout ∆tso to that at the
(2 n + 2) (2 n + 3) PKN onset of screenout tso.

α1 = (n + 1) (n + 2) KGD η → 1. (40) The change in the area-averaged fracture width fol-
( 4n + 4) (3n + 6) Radial lowing a screenout is obtained by dividing the terms

in Appendix Eq. 42 with those in Appendix Eq. 44
The lower bound α0 follows from Appendix Eq. 37: and substituting the fluid-loss expressions from
Appendix Eqs. 45 and 46:
α0 = 1 − θ η → 0. (41)
∆w 1   g( ∆t Dso ) − g0  
For the typical value of θ = 1⁄2, Appendix Eq. 41 = ∆t Dso − [1 − ηso ]  . (47)
shows that α0 = 1⁄2, which is the value used in Eq. 9-43. wso ηso   κ so g0  
An expression for the treatment efficiency at any

Postscreenout pressure relations time after a screenout η(∆tDso) can be obtained by
combining the definition of η from Eq. 9-4 with the
The material-balance relations following a peripheral width multiplier expression in Appendix Eq. 47:
screenout (i.e., constant fracture area following a
screenout) can be written in a manner similar to
η( ∆t Dso ) = 1 −
(1 − η )
so
 g( ∆t Dso ) − g0 
1+
(1 + ∆t Dso )  .
Eq. 9-5 (Nolte, 1990). At any additional time follow- (48)
κ so g0 
ing a screenout ∆tso, the change in fracture volume is
equal to the additional injected slurry volume minus Appendix Eqs. 47 and 48 are equivalent to
the volume of fluid lost to the formation through Eqs. 6L-6 and 6L-7.
leakoff: Based on the proportionality between net pressure
∆ w A fso = qi ∆tso − VLp ( ∆tso ) ,
and width, differentiating Appendix Eq. 47 obtains
(42)
the following expression:
1 dpnet ( ∆tso )
where the subscript so identifies the corresponding
1 d ∆w
parameter value at screenout. Similarly, an expression =
for the fracture volume at screenout Vfso can be wso dt pnet ,so dt
written as
=
1  f ( ∆t Dso ) 
Vfso = ηsoViso , (43) tso ηso 1 − [1 − ηso ] , (49)
 κ so g0 
where ηso is the fluid efficiency at screenout and Viso
is the volume injected prior to the screenout. where pnet,so is the net pressure at the screenout and β
For an assumed constant injection rate Viso = qitso, is assumed to be unity (Fig. 9-20). From the defini-
Eq. 9-6 and Appendix Eq. 43 can be combined: tion of efficiency, it can be readily shown that
(w A) f = ηso qi tso = ηsoVLp (tso ) (1 − ηso ) , (44) tso ηso

=
tη( ∆tso )
. (50)
pnet ,so pnet ( ∆tso )
so
where VLp(tso) represents the cumulative volume of

fluid lost prior to screenout and can be expressed in a Combining Appendix Eqs. 49 and 50 produces the
manner similar to Appendix Eq. 22 for a wall-building following relation for the log-log slope of the net
fluid: pressure after a screenout (Nolte, 1990):
VLp (tso ) = 2 rp κ so CL tso A f g0 (α ) , (45) t dpnet ,so
=
pnet ,so dt
where κso is the spurt factor evaluated at tso.
Because the same set of conditions as that follow- (
1 − [1 − ηso ] f ( ∆t Dso ) (κ so g0 ))
( ( ) (κ g ))
.
1 − [1 − ηso ] [1 + ∆t Dso ] 1 + g( ∆t Dso ) − g0
ing the end of injection is assumed to exist following
so 0
a screenout, the expression for VLp(∆tso) is similar to
that in Appendix Eq. 26: (51)

The bounding values of the log-log slope are for describing the pressure in the reservoir are the
obtained by substituting the appropriate values for η pressure drop caused by the near-face leakoff effects
and the corresponding relations for f(∆tDso), g(∆tDso) (i.e., filter cake and filtrate) ∆pnf and the pressure dif-
and g0 that can be obtained from Appendix Eqs. 28, ference in the reservoir ∆pR:
29 and 30, respectively.
∆pT = p f − pi = ∆pnf + ∆pR t < tc (52)
Comparison of fixed-length ∆pT = ∆pR ; ∆pnf = 0 t > tc . (53)

and propagating fractures† In general, each of the pressure differences depends
Fracture propagation and closure lead to characteristic on time and position as the fracture propagates and
pressure and fluid-loss distributions along the inter- closes.
face between the fracture and the reservoir. The time As shown in Eq. 9-76, ∆pR is represented by the
history of these distributions establishes the boundary added contributions from the two sources of fluid
conditions for the reservoir response before and after loss: the pressure difference associated with the CL
fracture closure. These conditions are generally dif- component of fluid loss ∆pRC and the component
ferent from those experienced by fixed-length frac- associated with spurt loss ∆pRS. Nolte et al. (1997)
tures; however, there are also similar conditions that provided the linear flow expression for ∆pR that
are shared by these two types of fractures. The simi- includes the time dependence for ∆pRS in Eq. 9-84,
lar conditions enable applying established fixed- which is applicable both before and after closure.
length relations to a propagating fracture. The back- The remainder of this section of the Appendix
ground for identifying these differences and assumes that near-face effects and spurt loss are neg-
similarities is reviewed in this section of the ligible and focuses on the CL component of the pres-
Appendix to Chapter 9 along with the fixed-length sure difference:
relations that are applicable to the propagating case. ∆pRC = ∆pT ; ∆pRS = ∆pnf = 0. (54)
Specifically, this section of the Appendix provides
Abousleiman et al. (1994) expressed the general
• conditions that enable adapting fixed-length relations
relation for ∆pRC in terms of an integral equation
to a propagating fracture
using a Green’s function approach and requiring
• definition of a generalized reservoir leakoff coeffi- specification of the fluid-loss history. Another
cient that is applicable to fracture propagation approach for describing ∆pRC is to identify several
within all flow regimes simplifying features for a propagating fracture. These
• framework for type-curve analysis of the after- are the conditions of approximately constant pressure
closure period. and dimensionless time during fracture propagation.
Equations 9-23 and 9-24 characterize the pressure
Throughout this section, the propagating fracture
profile within the fracture and its magnitude at the
is assumed to have a rectangular shape (i.e., Af = 2hf L)
wellbore in terms of net pressure. These equations
and a square root of time dependency of the fluid loss
indicate that the pressure is only weakly dependent
(i.e., θ = 1⁄2 in Appendix Eq. 5).
on fracture length and the position in the fracture,
with the potential exception of the region near its tip.
Pressure characterization for It can, therefore, be concluded that the overall spatial
a propagating fracture and temporal pressure variations within a fracture are
relatively small in comparison with the total pressure
The total pressure difference ∆pT between the fluid drop between the fracturing fluid and the undisturbed
pressure in the propagating fracture pf and the initial reservoir pressure. An approximately constant and
reservoir pressure pi can be divided into three compo- uniform pressure is a primary assumption for deriv-
nents, as proposed by Howard and Fast (1957) and ing the individual leakoff coefficients (see Section 6-4)
illustrated in Fig. 5-17. The components of interest and the primary reason for their successful applica-
tion as invariant parameters. This pressure condition
†
This section by K. G. Nolte, Schlumberger Dowell.
is also reflected in the approximately constant injec-

tion pressure during propagation at a constant injec- whereas the fixed-length case is generally associated
tion rate. The simultaneous condition of constant with a conductive fracture.
pressure and constant rate provides one of the pri-
mary distinctions between a propagating and a fixed-
length fracture. The fixed-length fracture requires an Fluid-loss characterization for
increasing pressure difference to maintain a constant a propagating fracture
production rate or a decreasing rate for a constant- Figure 9-35 illustrates the spatial distributions of the
pressure condition. From the reservoir perspective, fluid-loss rate during propagation and the specific
this difference occurs because a propagating fracture loss volume after propagation. The discussion of this
has an approximately constant or stationary value of figure in Section 9-6 observes that the specific vol-
dimensionless time with its associated condition of ume distribution is similar to that expected from a
time invariance for pressure and flow rate. uniform-flux fixed-length fracture. This similarity
provides the basis for the apparent fracture length
relation given by Eq. 9-78. A second type of fixed-
Propagation with a stationary length fracture assumes uniform pressure, or equiva-
dimensionless time lently infinite conductivity. Both of these fixed-length
The dimensionless time T (Eq. 9-75) can be com- fractures were discussed and characterized by
bined with the power law relation for fracture area Gringarten et al. (1974).
versus time (Appendix Eq. 2) and an efficiency-based The two types of fixed-length fractures can be
approximation for the area exponent α (Eq. 9-44) to compared to the after-closure behavior of the impulse
obtain injection shown in Fig. 9-38 for a propagating frac-
0.5 η ture. Sidebar 9G summarizes the impulse description
T  tp 
=  , (55) of after-closure radial flow behavior by the time
Tp  t  derivative of dimensionless pressure. The impulse
where Tp is the dimensionless time at the end of description can also be applied to linear and transi-
pumping and the approximation used is α ≈ 0.5 + tional flow. For a propagating fracture, this applica-
0.25η, which covers the values given by Appendix tion requires the apparent length relation to transform
Eqs. 40 and 41 for typical fracturing conditions. the dimensionless time (Eq. 9-79) for defining the
Appendix Eq. 55 shows that T is stationary and equal dimensionless pressure of a fixed-length fracture. The
to its value at the end of pumping for a fracture with derivative of dimensionless pressure is the same for
vanishing efficiency (i.e., η → 0). The equation also the two fixed-length fractures in linear and radial flow
shows that T retains only a weak dependence on time (Eq. 9-95) and accurately represents the impulse
for moderate values of efficiency. For example, for behavior in Fig. 9-38 for these flow regimes.
η = 0.5, T decreases only 19% during the second half However, the fixed-length fractures have different
of a treatment. Furthermore, Fig. 9-38 shows that a transitional flow behaviors (e.g., Gringarten et al.,
10-fold change in the dimensionless time is required 1974) and represent the transitional behavior in
before any meaningful change occurs in the reservoir Fig. 9-38 with differing degrees of accuracy. The
flow regime. The reservoir flow regime, therefore, is uniform-flux fracture has a deviation of less than
even more weakly dependent on dimensional time 5% from the propagating case for the complete time
than the case for dimensionless time. It can be con- range shown in the figure (i.e., that applicable to the
cluded that typical conditions for fracture propagation impulse representation). The transitional behavior for
result in essentially a stationary dimensionless time the uniform-pressure case deviates by almost 25%
and reservoir flow regime. from that for the propagating fracture. The maximum
The approximately stationary value of T provides deviation in both cases occurs slightly before the
another primary difference between a propagating knee time. These quantitative results confirm the
fracture and a fixed-length fracture for which dimen- cited qualitative inferences from Fig. 9-35, that the
sionless time increases with increasing dimensional after-closure behavior of a propagating fracture can
time. An additional difference for a calibration treat- be represented by a uniform-flux fixed-length fracture.
ment is that it does not retain fracture conductivity,

In contrast to this after-closure comparison, the where reservoir linear flow is assumed (Howard and
uniform-pressure fracture provides essentially the Fast, 1957) and the definition is in terms of the total
same flux distribution in radial flow as that for the pressure difference between the fracture and the ini-
fluid loss from an inefficient propagating fracture. tial reservoir pressure. A definition in Eq. 6-91 is in
This comparison is shown on Appendix Fig. 4. The terms of the pressure difference ∆pRC between the fil-
fixed-length distribution from Gringarten et al. (1974) trate/reservoir interface and the initial reservoir pres-
was described as the stabilized flux distribution. The sure. This pressure difference is defined in Appendix
spatial variation of the fluid loss during propagation Eq. 54 and denoted as ∆pc in Eq. 6-91. Combining
was obtained by combining the fracture growth Appendix Eq. 56 and Eq. 6-92 gives the following
power law relation (Appendix Eq. 2) for vanishing ratio between the pressure differences and fluid-loss
efficiency (i.e., area exponent α = 1⁄2) and the square coefficients:
root of time leakoff behavior (Appendix Eq. 8). ∆pRC CL
= . (57)
∆pT Cc
Uniform-pressure fixed-length
fracture (Gringarten et al., 1974) When the near-face effects and spurt are negligible,
10
Inefficient, propagating fracture the two pressure differences are equal (i.e., Appendix
8 Eq. 54) and Appendix Eq. 57 then indicates the
Normalized flux
expected result that the combined and reservoir coeffi-

6 cients are also equal. This result reflects the expected
Stabilized distribution
4
radial flow, TD ≥ 1.0 reservoir behavior based on a linear relation between
the pressure difference and flux. This behavior is also
2 required for the radial flow reservoir coefficient:
∆pRC CL
Linear flow = 0.1 0.01 0.001
0
0.5 0.6 0.7 0.8 0.9 1.0 = . (58)
Dimensionless position, ξ
∆pT CR
The previously cited application of pD,up for defin-
Appendix Figure 4. Comparison of flux distribution for ing CR requires expression of the dimensionless pres-
a uniform-pressure fixed-length fracture and a fluid-loss-
dominated propagating fracture. sure (Eq. 12-2) in terms of the quantities for the CL
component of fluid loss:
Gringarten et al. also noted that the stabilized 2π khL ∆pRC t p ( ),
distribution is independent of its prior history. ( )
pD,up Tp =
µq L ,C
(59)
Furthermore, it is spatially the same as that for a
propagating fracture with vanishing efficiency and where the fluid-loss height hL = rphf and qL,C is the
spurt. It follows that the dimensionless pressure leakoff rate associated with Carter-based fluid loss.
response pD,up for the uniform-pressure fixed-length Substituting for the expression for qL,C with vanishing
fracture applies to a propagating fracture when the efficiency and spurt from Appendix Eq. 11 and the
pressure is evaluated at its stationary value of the dimensionless rate function f = π/2 for the specified
dimensionless time Tp (i.e., pD,up(Tp)). This dimen- conditions (Nolte, 1986a):
sionless pressure value applies for radial flow with
vanishing efficiency and spurt and therefore provides
the basis for subsequently defining the reservoir ( )
pD,up Tp =
k tp
µ CL L
∆pRC t p ( ) η → 0, κ → 1. (60)
leakoff coefficient CR for these conditions.
Before defining this coefficient, the conventional An expression for the radial coefficient CR can then
definition of the reservoir leakoff coefficient must be be obtained by combining Appendix Eq. 59 with the
considered: definitions of dimensionless time (Eq. 9-75), Cc
(Appendix Eq. 56) and the pressure ratio (Appendix
k φ ct Eq. 58):
Cc = ∆pT , (56)
πµ Cc
CR = πTp η → 0, κ → 1.
( )
(61)
pD,up Tp

Limiting values for pD,up(Tp) were given by The second relation in Appendix Eq. 64 indicates
Gringarten et al. (1974): an expanded range of application relative to that
( )
assumed for the derivation of Appendix Eq. 61. The
Linear flow pD,up Tp = πTp Tp < 0.01 expanded range results from numerical simulations
Radial flow pD,up (T )
p
1
2
[
= ln Tp + 2.2 ] Tp > 3. (Abousleiman et al., 1994; Nolte, 1998) that indicate
that Appendix Eq. 64 approximates (i.e., within a 5%
(62) error) the reservoir coefficient for transitional flow.
More generally, the simulations indicate that
They also provided a general relation for pD,up(Tp)
Appendix Eq. 64 is approximately valid (i.e., within
in terms of special functions. This relation can be
a 10% error) for moderate values of efficiency
approximated with an error of less than 2% by
(η < 0.5) and with any reservoir flow regime.
 πTp − 0.58Tp Tp < 0.16 Therefore, CR, as defined by Appendix Eq. 61,

( )
pD,up Tp ≅  1 represents the “general reservoir” leakoff coefficient
( [
 2 ln Tp + 0.22 + 2.2 ] ) Tp > 0.16. within the accuracy required for fracture design and
evaluation purposes.
(63)
Appendix Eq. 63 can be used to show that during
Comparison of Appendix Eqs. 62 and 63 shows transitional and radial flow, pD,up(Tp) ≤ √πTp.
that the approximations provided by the latter equa- Appendix Eq. 64 therefore implies that the general
tion are defined by adding a term to each of the limit- coefficient CR is larger than the linear flow coefficient
ing cases. The second approximation in Appendix Cc for fracture propagation under these flow condi-
Eq. 63 can be obtained by applying the apparent time tions. This result has been reported by Hagoort
multiplier (1 + 0.22/Tp) to the dimensionless time. Its (1980) and Valkó and Economides (1997). For exam-
inclusion extends the applicability of the logarithmic- ple, for a dimensionless time Tp = 1, the dimension-
based radial flow relation to a dimensionless time that less pressure pD,up = 1.21 and √πTp = 1.78. For these
is about 1⁄20 of the value normally required for radial values, Appendix Eq. 64 indicates that the general
flow (i.e., Tp = 0.16 in Appendix Eq. 63 versus Tp = coefficient CR is about 1.5 times larger than Cc under
3 in Appendix Eq. 62). The apparent time multiplier these late transitional flow conditions. For a larger
for application with Appendix Eq. 63 has the same dimensionless time Tp = 10, CR becomes larger by
form as that for a similar development introduced in a factor of 2.5.
Eq. 9-83 for the after-closure behavior of a propa- This observation of a larger leakoff coefficient gen-
gating fracture. These two relations are seemingly erally applies to the mini-falloff test because the test
different because the relation for a propagating frac- design should be based on Tp > 1 and the achieve-
ture contains a different constant (i.e., 0.14 instead ment of reservoir-controlled fluid-loss conditions
of 0.22). However, this difference occurs because the (Appendix Eq. 1). Larger values of the reservoir coef-
dimensionless times corresponding to the two cases ficient do not affect the fluid loss for most proppant
differ by the square of the apparent length fraction treatments where near-face effects are designed into
faL from Eq. 9-79. For the assumed conditions of van- the fluid system to control the fluid-loss behavior.
ishing efficiency and spurt, Eq. 9-80 indicates that Combining Appendix Eqs. 58 and 61 provides the
faL = π/4. Applying this value to the dimensionless general relation for the CL component of the pressure
time for the fixed-length case indicates that the con- difference ∆pRC for all flow regimes and in terms of
stants for the two apparent time relations are actually the total pressure difference and combined fluid-loss
equivalent. coefficient.
For reservoir linear flow, Appendix Eq. 62 shows
that pD,up(Tp) ≈ √πTp. It follows from Appendix
Eq. 61 that Type-curve-based analysis
Cc Linear flow The dimensionless pressure pD,up can also be used to
 develop type-curve analyses for general after-closure
CR =  πTp conditions. The normalized pressure difference and
( )
C Transitionaland radial flow.
 pD,up Tp c
 pressure derivative variables and the log-log slope in
(64) Fig. 9-38 illustrate several characteristics of the after-

closure pressure response that motivate its analysis lished by the matching procedure, the dimensionless
within a type-curve framework. For example, the pressure pD,up(Tp) can be obtained using Appendix
character, or shape, of the curves depends on the Eq. 63.
dimensionless time Tp. Also, a suitable match pres- This type-curve analysis has the following relation
sure for the analysis can be defined as the ratio of the between the after-closure pressure and the pressure-
dimensional pressure variables and the normalized difference curve Rp:
[ ]
pressure variables shown on the figure.
The development of type curves applicable to ∆pR (t ) = p(t ) − pi = m p Rp Tp , (t − tc ) tc , η, κ . (69)
generalized fluid-loss conditions requires a relation
For fracture propagation in flow regimes other than
among the average value of the fluid-loss rate q—L, CL
well-established linear and radial flow, Rp must be
component of the fluid-loss rate qL,C, injection rate qi,
defined by numerical simulation.
fluid efficiency η and spurt factor κ. This relation can
The matching procedure uses a pair of type curves:
be obtained by combining the rate versions of Eq. 9-6
one for the pressure difference and one for the pres-
and Appendix Eq. 22:
sure derivative. Each quantity can be defined using an
q L = κ q L ,C = (1 − η) qi . (65) appropriate time function. The square of the linear
flow time function FL(t/tc) given by Eq. 9-88 is pre-
For general values of efficiency and spurt, the ferred because it provides a consistent representation
reservoir pressure difference at the end of pumping of the after-closure period for the reservoir response
∆pR(tp) provides a convenient quantity to use as the in any flow regime, as discussed in Section 9-6.7.
match pressure (i.e., the multiplying factor for the Furthermore, the pressure difference and the pressure
type curves). The relation between ∆pR(tp) and ∆pRC derivative are presented in terms of the inverse of FL2
at the end of injection can be obtained from Eqs. 9-76 because this presentation provides the conventional
and 9-84 and Appendix Eq. 58: representation of increasing time from left to right
along the x-axis. Appendix Fig. 5 illustrates these
κ + 1 κ + 1  CL
( )
∆pR t p =  ∆p t =  ( )
 2  RC p  2  CR T p
∆p t . ( ) curves for the case of vanishing efficiency and spurt
(e.g., applicable to a mini-falloff test).
(66) The type-curve analysis begins by matching the
Rearranging Appendix Eqs. 65 and 66 and substi- character (i.e., shape) of the pressure derivative for
tuting them into Appendix Eq. 59 provides a more the data to the character of one of a collection of type
general form of the dimensionless pressure: curves based on different values of Tp. This character
matching defines Tp. The selected curve for Tp is then
 2κ  2 π kh
( ) ( )
vertically translated to match the pressure derivative
pD,up Tp =   mp ∆pR t p → m p ,
 (1 − η)(1 + κ )  µqi
of the data, and the resulting form of the pressure
derivative defines the match pressure mp:
(67)
where the fluid-loss height hL = h and the role of 1.0
∆pR(tp) is introduced as the match pressure mp. The Pressure difference
Pressure derivative
transmissibility can be determined from this dimen- η→0 κ→0
sionless relation:
kh  (1 − η)(1 + κ )  qi pD,up Tp ( )
Rp(t)
=
µ   2 πm . (68) Tp = 0.001
2κ  p
In addition to determination of the match pressure for

Tp = 2.00 Tp = 0.05
the type-curve analysis, the dimensionless time at the 0.1
end of pumping Tp is required to define pD,up for 0 10 100
1/{FL(t/tc)}2
Appendix Eq. 68. The value of Tp can be obtained by
matching the character, or shape, of the data to that
of the type curves. Once a value of Tp has been estab- Appendix Figure 5. Example type curves for negligible
efficiency and spurt.

  (i.e., ∆pR(tp) inferred from the after-closure data).

dp(t )
{F (t t )} { }
dRp
= m p  FL (t tc )  Relative to Appendix Eq. 66, ∆pT (tp) = p(tp) – pi, with
2 2
2 .
{
d FL (t tc ) }  {
d FL (t tc )} 
L c 2
  p(tp) defined by the ISIP, and when ∆pR(tp) ≈ ∆pT (tp)

in the absence of spurt, the reservoir and total leakoff
(70) coefficients are approximately equal. Under these
conditions, essentially the entire pressure difference
The transmissibility is then estimated from Appendix
∆pT is within the reservoir and the reservoir controls
Eq. 68 using the two type-curve parameters Tp and mp.
the fluid-loss behavior. Conversely, when ∆pR(tp) <<
The initial reservoir pressure is extracted by applying
∆pT(tp), a negligible fraction of the total pressure drop
Appendix Eq. 69 over the time range of the match.
occurs in the reservoir, and the reservoir does not
Additional information concerning the nature of
have a meaningful role in controlling fluid leakoff.
the fluid loss can be obtained from the value of mp

◆
Fracture Treatment Design
Jack Elbel, Schlumberger Dowell

Larry Britt, NSI Technologies, Inc.
10-1. Introduction • Discounted return on investment (DROI) is the ratio

of the project’s NPV to the present value of the total
Fracture stimulation is used to overcome the adverse investment discounted at a stated rate. DROI is a
effects of formation damage and low permeability, measure of capital efficiency, and in the simplest
accelerate production, increase reserves and control terms it is the expected dollars of profit per dollar
the production of water and formation solids. These invested from a discounted viewpoint.
fracturing applications require different fracture stim-
• Return on investment (ROI) is the ratio of the pro-
ulation designs to achieve their objectives; therefore
ject’s undiscounted cash flow to the undiscounted
a means of evaluating treatment designs is important
total investment. This economic parameter does not
to determine which is the optimum approach. Frac-
consider the time value of money and, therefore,
ture economic optimization techniques can be used
artificially favors long-term projects.
to conduct this evaluation (Veatch, 1983; Meng and
Brown, 1987). These references also provide the cal- • Payout is the time for the cumulative undiscounted
culation procedures for economic analysis. The frac- cash flow of a project to reach zero. Payout does
ture optimization process requires the ability to not consider the time value of money or the cash
predict fracture geometry and cost, reservoir perfor- flow recovered after the project reaches payout.
mance and revenue and then to couple these cost and However, it does provide an indication of how long
revenue streams with an economic evaluation. The investment capital is at risk and reflects the time
following sections discuss some considerations for period for a project to reach the break-even point.
fracture optimization. • Rate of return, or profitability index, is the com-
Economic analysis is used routinely to evaluate pounded interest rate that has a discount effect that
investment decisions in the petroleum industry. makes the present value of the net cash flows equal
Because both the magnitude and timing of project to zero. This rate is analogous to the interest rate
cash flow are important yardsticks by which project received if the investment were considered a loan
performance is evaluated, it is important to under- to be paid back at this interest rate.
stand the general economic criteria used by the indus- Each of these criteria should be reviewed within the
try to evaluate any investment decision, including current economic environment to ensure that business
hydraulic fracturing. objectives are achieved. If, for example, the business
• Discount rate is analogous to an interest rate that objective is to maximize profitability, it is necessary
reflects the “time value” of money and is selected to recognize that in an economic environment where
to balance the targeted investment reward relative access to capital is constrained and fracturing candi-
to the inherent risks for the investment. dates are unlimited, optimizing the DROI of each
• Present value is the sum of all future cash flows treatment maximizes overall profitability. If, on the
(income and expenses) discounted to the present other hand, the economic environment is such that
time at a stated discount rate. Another way to think capital and stimulation candidates are unlimited, the
about the present value concept is that the present NPV of each treatment should be maximized. The
value is the current dollars required to be indiffer- design consequences for the two criteria can be sub-
ent to receiving that amount of the future worth. stantial, with maximizing the DROI of the treatment
resulting in a smaller treatment and investment.
• Net present value (NPV) is the difference between
Sidebar 10A provides an example.
the project’s present value of future cash flows and
the present value of the investment.

10A. NPV for fixed costs or designated proppant mass
If it is assumed that the fracturing fluid and injection rate have been selected, the other major design considerations are the treatment
size, type of proppant and proppant scheduling. The size of the treatment should ideally be based on the optimum fracture penetration
determined by economic considerations (see Section 10-1.1) and made after ideal model assumptions are considered (see Section
10-3.1). However, it is not uncommon for a fixed expenditure to be used in determining the size of a treatment. There may be other
constraints, such as proppant mass and fluid volume, resulting from the availability of materials or logistics, which may also limit the
size of a treatment. Even then, the treatment design providing the best NPV or other economic criteria within these constraints can be
made by conducting a number of economic evaluations with different proppants and various maximum concentrations.
NPVs for various treatment volumes and proppant concentrations for a range of fixed costs are shown in Fig. 10A-1. In Fig. 10A-1a
the NPV after 1 year is plotted versus cost of treatment with sand and a higher cost, higher strength, premium intermediate-strength
proppant (ISP). The final proppant concentrations in the simulations are 10, 14 and 16 ppg. If the treatment cost is limited to $200,000,
the maximum NPV after 1 year is $2,080,000 using sand at a final concentration of 16 ppg. This NPV is $200,000 higher than that for
a treatment at the same cost with a premium proppant at 10-ppg final concentration and $310,000 greater than premium proppant at
16 ppg. If the treatment cost is increased to $300,000, the maximum NPV is increased only $20,000 and can also be achieved with
premium proppant at 16- and 14-ppg final concentrations.
A maximum NPV can also be determined for cases where the proppant volume or mass is limited because of availability, location,
size or an arbitrary decision. For this criterion, the curves in Fig. 10A-1b show that the premium proppant is always optimal. The plots
also show that for less than 130,000-lbm proppant, 10-ppg maximum concentration is optimal. For more than 130,000 lbm, a greater
economic benefit is obtained at a concentration of 14 ppg, increasing to 16 ppg with additional increases in proppant mass or treat-
ment cost that achieve more fracture length and benefit from increased concentration for improved conductivity.
Figure 10A-1. Net present value

(a)
versus (a) treatment cost and
2,300,000 (b) proppant type and mass.
The curves are scaled at incre-
ments of 100-ft proppant pene-
2,100,000
tration, beginning at 100 ft.
1,900,000
One-year NPV ($)
1,700,000
1,500,000
Sand 14 ppg
ISP 14 ppg
Sand 10 ppg
1,300,000
ISP 10 ppg
Sand 16 ppg
ISP 16 ppg
1,100,000
100,000 200,000 300,000 400,000 500,000
Treatment cost ($)
(b)
2,300,000
2,100,000
1,900,000
One-year NPV ($)
1,700,000
1,500,000
Sand 10 ppg
ISP 10 ppg
1,300,000 Sand 14 ppg

ISP 14 ppg
Sand 16 ppg
ISP 16 ppg
1,100,000
0 50,000 100,000 150,000 200,000 250,000 300,000 350,000
Proppant mass (lbm)
10-2 Fracture Treatment Design

10-2. Design considerations • variable proppant cost = $/unit × units of proppant

The unit cost includes
10-2.1. Economic optimization – proppant
Economic optimization of hydraulic fracture treat- – proppant transportation to location and storage
ments allows the production engineer to design a frac- – proppant pumping charges.
ture treatment that optimizes the production rate and
• variable hydraulic horsepower (hhp) cost = $/hhp
reserve recovery from a well to maximize well prof-
× injection rate × surface treating pressure/40.8
itability. In addition, a good understanding of the key
× standby hhp factor
parameters for the fracture treatment can be developed
from the optimization study. For example, Fig. 10-1 • other fixed costs
is a plot of 1-year NPV versus the productive fracture – mobilization
half-length xf in a 0.01-md formation. This figure – personnel
shows the relationship among length, conductivity and – well preparation (workover rig, etc.)
profitability. For penetrations from 200 to 600 ft, about
the same NPV is produced by proppant concentrations – cleanup costs (coiled tubing, disposal if not
of 6 to 14 ppg. However, the net production—and included as a fluid unit cost, etc.).
therefore cash flow—will be higher with the higher
concentrations. At 1000-ft penetration, concentrations
from 10 to 14 ppg yield about the same NPV and are 10-2.2. Treatment optimization
significantly greater than when using 6 ppg. Increasing design procedure
the fracture length improves the profitability of a well Optimization procedures require methods to determine
in this reservoir but also requires increasing fracture fracture geometry and production from the propped
conductivity for most penetrations. fracture. They may be in the form of a nomograph,
To estimate the cost of a fracturing treatment, the analytical solutions, two- or three-dimensional (2D or
variable costs can be added to some fixed cost not 3D) models for geometry and productivity index (PI)
directly associated with treatment size: calculations, type curves, or analytical or numerical
• variable fluid cost = $/unit × units of fluid reservoir models for production simulation. The accu-
racy of the optimization should increase with increas-
The unit cost includes
ing sophistication of the models and the accuracy of
– fracturing fluid plus additives the input parameters. Sensitivity studies of input para-
– mixing and blending charges meters with uncertain values are warranted.
– transportation, storage and disposal charges A basic procedure for economic optimization is as
(commonly included in other fixed costs). follows:
1. Select the fluid systems applicable to the formation
to be fractured.
Production time: 1 year
400,000 2. Select the proppant on the basis of stress and con-
ductivity requirements.
Net present value ($)
300,000
3. Determine the maximum allowable pump rate on the
200,000
basis of the pressure limitations of the wellhead and
tubulars. The optimum injection rate is a balance of
100,000 Maximum proppant
decreased fluid loss and increased horsepower as the
concentration 10.0 ppg rate is increased. Shear degradation, for some frac-
6.0 ppg 14.0 ppg
0 turing fluid systems, should also be considered (see
0 500 1000 1500 2000 2500
Chapters 7 and 8).
Productive fracture half-length (ft)
4. Select an appropriate fracture propagation model
Figure 10-1. Net present value versus productive fracture (e.g., pseudo-3D, or P3D) for the formation char-
half-length for a 0.01-md formation. acteristics and pressure behavior on the basis of
in-situ stress and laboratory tests, calibration treat-

ments and log analysis (e.g., stress profile, gamma those with the greatest potential for error are forma-
ray) (see Chapters 3, 4 and 9). tion permeability and fracture conductivity. Fortun-
ately, a number of economic models combine
5. Develop the input data required for the selected
geometry and reservoir models to allow making
geometry model(s).
detailed studies in a reasonable amount of time.
6. Determine fracture penetration and fracture conduc- Economic optimization among different fracturing
tivity for a selected treatment size and proppant fluids such as oil-base, water-base and foam fluids
concentration by forward simulation or determine is difficult. These fluids are usually chosen for com-
fluid and proppant volumes required and fracture patibility with formation fluids or cleanup properties,
conductivity obtained for a selected penetration and their economic benefit cannot be quantified
using inverse simulation (see Chapter 6). Determine unless features such as formation damage, polymer
the optimum pad fraction (see Chapters 5 and 6). If damage to proppant pack and cleanup time can be
the model does not account for polymer conductiv- determined by the production models or included
ity damage, an estimated damaged conductivity by some other means in the analysis. As an example,
should be used (see Chapters 7 and 8). if it is known that a foam fracturing fluid will reduce
7. Determine the production rate and cumulative the cleanup time, the cost of cleanup can be reduced
recovery over a selected time period for a specific for the foam fluid relative to that for a nonfoam
propped penetration and its corresponding conduc- fluid. Some risk factor and uncertainty can also be
tivity. associated with various input parameters and behav-
8. Calculate the present value of the net revenue of ioral assumptions (Nolte and Economides, 1991).
the production based on a discount rate (i.e., the
sum of the present values for each year of the
selected period). 10-2.3. Fracture conductivity
9. Calculate the total treatment cost including the Placing the appropriate amount and type of proppant
costs associated with fluids, proppants and in the fracture is critical to the success of a hydraulic
hydraulic horsepower. fracturing treatment. Independent of fluid residue
10. Calculate the NPV for the fracture by subtracting damage, proppant concentration in the fracture and
the treatment cost from the well’s discounted net resistance to crushing determine the fracture conduc-
revenue (step 9 minus step 8). tivity over the producing life of the well. Proppant
selection is optimized by balancing the potential to
11. Repeat the preceding computational cycle for
create fracture conductivity against the additional cost
incremental increases in length until the NPV
or risk of placement. Factors such as proppant proper-
decreases or a maximum length is reached.
ties (strength, particle size, roundness and fines con-
12. Construct curves showing the fracture NPV or tent), closure stress, polymer damage, drawdown rate,
other appropriate economic criteria versus frac- embedment and the resultant propped fracture width
ture penetration. When the NPV starts to decline affect the fracture conductivity. Several of these fac-
with increasing fracture lengths, the cumulative tors are discussed in more detail in Chapters 7 and 8.
production for the specific lengths will still be For simplicity, the fracture conductivity used for
increasing. production simulation is usually considered to be
The cycle can be repeated for other materials or homogeneous. Advances in geometry and production
conditions such as other fluids and additive concen- modeling (see Chapter 12) enable simulation with
trations, injection rates, proppant types and maxi- spatial variations in horizontal (Bennett et al., 1983)
mum proppant concentrations or even with other as well as vertical conductivity (Poe et al., 1992).
geometry models. The process can easily become Two regions over the vertical profile are affected
time consuming, and the number of iterations should by different criteria for conductivity. The first is a
be governed by the accuracy required as well as the bank buildup region created at the bottom during
accuracy of the input parameters. The cycles can be placement where the pack width is equal to the
repeated for sensitivity to parameters to determine hydraulic width for the region. The second is the
bounds. The most important input parameters and overlying slurried height region at closure where the

propped width and height are affected by the proppant considered for fracturing formations where stresses
slurry concentration and closure time. Ignoring the on the proppant are less than 6000 psi. The perme-
effect of proppant redistribution during closure, ability of precured resin-coated proppants is
Fig. 5-21 gives the percent of “fill” of the propped between that of sand and ISP.
width relative to the width when pumping stops. If a well is produced at a constant bottomhole
producing pressure, the minimum stress, and hence
• Stress on proppant
the closure stress on the proppant, decreases during
When a hydraulic fracture is created, the minimum production as the effective reservoir pressure
in-situ stress must be overcome to open and propa- decreases. If the well is produced at a constant rate
gate the fracture. When the fracture closes on the such that the bottomhole producing pressure must
proppant after the treatment, the effective closure be decreased faster than the effective reservoir
stress (hereafter termed simply closure stress) on pressure decreases, the stress on the proppant may
the proppant is equal to the minimum in-situ stress increase. When the minimum surface producing
plus the additional stress induced by the pack width pressure is reached and the well is produced at
minus the pore pressure in the proppant pack. The a constant pressure, the stress will decrease.
additional stress caused by the width can be esti- Formations with a high stress (usually associated
mated from an appropriate width model (see with abnormally high pore pressures) may not
Chapter 6) using the propped width and the result- always require high-strength proppants as would
ing net pressure as the increased stress. To bring be expected. This is the case where the desired pro-
the well on production, the pressure in the pack duction rate can be maintained at a low drawdown
at the wellbore must be reduced below that of the pressure during early production time, minimizing
reservoir, increasing the stress on the proppant at the effective stress on the proppant. Later, because
this time. the pore pressure has decreased as a result of pro-
The stress on the proppant changes during the duction, the in-situ stress may have decreased to
life of the well. If initiating production requires that a point where subsequent increased drawdown will
the well be swabbed, high stresses on the proppant not cause the stress on the sand to become exces-
near the wellbore may occur. If the well is swabbed sive. From a practical standpoint (i.e., to enable
off the bottom, the bottomhole producing pressure economic production during the period of low
is essentially zero, creating the maximum stress drawdown), this condition requires relatively high
on the proppant. Cyclic loading tests (Kim and values of the permeability-thickness product kh for
Willingham, 1987; Holditch and Blakely, 1992) the formation and effective fracture characteristics.
show that the conductivity is irreversibly reduced. The incongruity of this scenario for relatively shal-
Therefore, as the reservoir pressure is reduced low reservoirs is that if the formation permeability
because of depletion and the closure stress is is low, high-strength proppants may be required,
decreased proportionally (by a ratio of about 1⁄2; and if it is high, sand may be sufficient.
see Chapter 3), the permeability of the proppant
• Propped width
is not restored to greater than what it was when
the maximum stress was applied. Figure 10-2 illustrates the typical relationship of
Figure 7-13 shows the effect of closure stress on fracture conductivity to proppant concentration.
the permeability of various propping agents. When Increasing the proppant concentration results in
the stress on the proppant exceeds 10,000 psi, high- multiple layers of proppant and an increase in the
strength proppants (such as sintered bauxite) are fracture conductivity. Although maximum conduc-
required. In the range of 5,000 to 10,000 psi, the tivity can be obtained by placing proppant in a
use of higher strength, manufactured ceramic prop- partial monolayer (a technique developed for hori-
pant—commonly called intermediate-strength zontal fractures), placement of a partial monolayer
proppant (ISP)—should be considered. These pre- in a vertical fracture is virtually impossible to
mium proppants provide greater conductivity at achieve. Therefore, fracture treatments are designed
higher closure stresses, but their cost makes them for multiple proppant layers. With multiple layers
economically unattractive at lower closure stresses. of proppant, the outer layer may embed in softer
Sand is the most common proppant, and its use is formations, allowing only the inner layers to

monolayer
10
Partial
Full
monolayer Multilayer
105
(lbm ⁄ ft2 of propped area)

Number
of
layers
1 2 3 45
1.0
104
kfw (md-ft)
103
0.1
0.01 0.10 1.0
Propped fracture width (in.)
102 Figure 10-3. Fracture proppant (sand) concentration ver-

0.01 0.10 1.0 10 sus propped fracture width.
(lbm ⁄ ft2 of propped area)
Therefore, treatments are generally designed with
Figure 10-2. Relationship of fracture conductivity to frac- higher proppant concentrations to compensate for
ture proppant concentration for 20/40-mesh sand, no fluid these unknown, negative factors.
damage and relatively small stress.
remain open to flow (Strickland, 1985; Smith et al., 10-2.4. Dimensionless fracture conductivity
1987; Martins et al., 1992c).
Comparison of the effect of fracture conductivity on
Figure 10-3 is a graph of proppant concentration
production can be made easily if conductivity is cast
in the fracture versus propped fracture width for
in dimensionless terms and plotted against dimension-
20/40-mesh sand. The graph illustrates that once
less time and rate, as shown by the constant-pressure,
multilayer packing is achieved, the fracture width
finite-conductivity type curves in Fig. 10-4. These
increases proportionally to the increase in proppant
type curves give the wellbore production response
concentration. In Figs. 10-2 and 10-3, 2 lbm/ft2
as the reciprocal dimensionless rate 1/qD for a range
corresponds to about 0.25-in. propped width and
of dimensionless fracture conductivity CfD.
has about 10 layers for 20/40-mesh proppant. The
In Fig. 10-4, a design with a calculated CfD of 1.0
significance of more than five layers is discussed
and a selected dimensionless time tD of 0.01 is plotted
concerning proppant flowback in Section 11-6.
as point A. This corresponds to production repre-
Proppant packs with more than five layers of prop-
sented by 1/qD equal to 0.8. Doubling the fracture
pant were found to become unstable and produce
length with the same conductivity reduces CfD by
proppant when subjected to the forces from fluid
50% and decreases tD by 25% (point B). The value
flow (Asgian et al., 1995). Thus, proppant consoli-
of 1/qD for point B has essentially not changed, indi-
dation techniques are indicated.
cating that the production at this time is not different
Other factors influencing the final conductivity,
from that at point A. However, if the penetration is
particularly the gel residue in the fracture, are dis-
doubled and the proppant volume is increased by a
cussed in Chapter 8. Influences that are more diffi-
factor of 4 over that of the design for point A, the con-
cult to define are the creation of formation fines
ductivity doubles but CfD remains equal to 1. This is
because of stress cycling (Morita et al., 1988;
shown as point C, with a corresponding 1/qD of 0.55,
Ramakrishnan et al., 1991), movement of forma-
which indicates a 45% increase in the production rate.
tion fines into the fracture, long-term proppant
However, a more significant increase in production
degradation from dissolution and stress corrosion,
over that for the point A design can be achieved by
and permeability loss from precipitate buildup.

10
Reciprocal dimensionless rate (1/qD)
1 CfD
B
A
0.1
C D
0.5
A'
1
10–1 5
B'
kh∆p
10 1/qD =
141.2qBµ
25
tD = 0.0063kt (days)
50 φµctxf2
100 kfw
CfD =
kxf
500
10–2
10–5 10–4 10–3 10–2 10–1 1
Dimensionless time (tD)
Figure 10-4. Finite-conductivity type curve comparisons of production for various fracture designs located as points.
increasing the proppant volume or conductivity by a • Constant proppant volume, optimum CfD
factor of 4 for the same length, as shown by point D. Prats (1961) showed that for a given fracture vol-
Increasing the conductivity for the point A design ume (proppant volume) there is a fracture width to
by a factor of 50 results in a CfD of 50, as shown as fracture length relation for achieving maximum
point A′. This design corresponds to a 1/qD of 0.25. productivity. This relation can be expressed as
Doubling the length with the same conductivity now CfD equal to 1.26, and it is valid when the well is
results in a 1/qD of approximately 0.15, or a 66% at pseudosteady state for the majority of its produc-
increase in production rate at that particular time tive life. However, Morse and Von Gonten (1972)
(point B′). A design with twice the length, but with showed that because of transient production in low-
the same proppant volume as point A′, yields a CfD permeability formations, the higher rates obtained
of 12.5, and it still results in a higher production rate prior to reaching pseudosteady-state conditions can
than that of point A′. This requires the same proppant significantly affect the economics of hydraulic frac-
volume as doubling the length and maintaining the turing treatments. A study by Elbel (1988) using
same CfD. a reservoir simulator showed that for constant prop-
Exercises such as this show the insight gained by pant volume in formations with permeabilities
using type curves in fracture design, and they have led greater than l md, a CfD of 1.26 is optimum. For
to guidelines regarding the use of CfD for design eval- maximum production in formations with less than
uation. Designs with CfD equal to 3 or less cannot be 0.1-md permeability, a CfD of 3 is optimum. How-
improved significantly by increasing the fracture ever, these evaluations, for proppant alone, ignore
length with the same conductivity. A redesign for sig- the cost of fluids and other associated treatment
nificantly more production requires an increase in CfD. costs required to place a fracture.
If CfD is 30 or greater, increasing the length is more
• Constant length, optimum CfD
beneficial than increasing conductivity. The optimum
conductivity should be defined by an economic analy- For a constant fracture length, a CfD between
sis; however, insight into the role of conductivity is 10 and 30 has generally been accepted as an opti-
provided by considering the special cases addressed mum range (Holditch, 1979a). Because a CfD of
in the following. 30 for a given fracture length requires 3 times the

specific proppant volume required for a CfD of 10, during crushing and with fluid residue (Cooke, 1973).
further refinements should be made by inspection The effects of saturation and multiphase flow can
of the constant-pressure type curve (Fig. 10-4). increase the non-Darcy effects (Maloney et al., 1987;
The effect of various values of CfD can be seen Martins et al., 1990). Methods to correct the dimen-
by calculating dimensionless time for 36 days to sionless conductivity term used in fracture design and
observe differences in the corresponding values well test analysis have been developed by Holditch and
of 1/qD on the constant-pressure type curve. If the Morse (1976), Guppy et al. (1982b) and Gidley (1991).
calculated tD at 36 days for the fracture length is However, numerical simulation is usually required for
0.1, theoretically there is little benefit from a CfD accurate prediction and evaluation. Non-Darcy flow
greater than 10. However, if tD is 10–4, a CfD of effects are further discussed in Section 12-3.1.
100 would be of considerably greater benefit. Concern for possible non-Darcy effects is another
• Constant length, varying conductivity reason for the overdesign of conductivity. Because the
velocity is affected by the width, doubling the propped
To increase productivity, the later part of a treatment
width reduces the non-Darcy effects by a factor of 4,
may contain a higher permeability proppant, such
which may be more economical than using a higher
as bauxite or an ISP. Initially these types of treat-
priced proppant that has higher permeability and
ments were made in areas where high closure stress
requires less width. This depends on the portion of
would severely crush the sand, and the cost of using
the total pressure that results from the non-Darcy
an expensive, stronger proppant throughout the
effects. However, it may not be good practice to use
treatment was considered prohibitive. Bennett et al.
a lower permeability proppant at higher concentrations
(1983), Britt and Bennett (1985) and Elbel (1988)
to achieve more width to try to overcome non-Darcy
investigated the effect of varying fracture conduc-
effects. The production velocity for fractures in low-
tivity and high-conductivity tail-in treatments and
permeability formations is usually low enough to not
showed that this design strategy may be beneficial.
have significant non-Darcy effects. Again, numerical
Relating real time to tD (as in the previous section)
simulation is generally required to assess non-Darcy
helps determine if a fracture of varying conductivity
effects and achieve an optimum design.
could be of significant benefit over a fracture of uni-
form conductivity. Figure 10-4 shows that for CfD
of about 10 and a 36-day tD greater than 0.1, varying 10-2.6. Proppant selection
the conductivity or tailing in with a high-strength
proppant would have minimum effect except poten- A major consideration in proppant selection is opti-
tially for some improvement in the fracture fluid mizing permeability or conductivity versus the associ-
cleanup, cases with catastrophic near-wellbore prop- ated cost and benefit. The permeability of various
pant failure and cases with non-Darcy flow effects. proppants versus stress is shown in Fig. 7-14. The
proppant with the highest permeability is not always
the optimum choice. The volume and cost required to
10-2.5. Non-Darcy effects obtain an optimum or desired conductivity should be
considered. Figure 10-5 is a plot of relative proppant
The previous discussion of conductivity is based on
volume versus closure stress for different proppant
Darcy’s law, for which the pressure drop for fluid flow
types (Elbel and Sookprasong, 1987). The relative
is directly proportional to velocity. Cooke (1973) and
proppant volume Vrp in lbm/md-ft3 reflects the amount
Holditch and Morse (1976) demonstrated the role of
of proppant required to achieve a specific conductivity:
conductivity with a non-Darcy flow effect that adds
( )
an additional pressure drop that is proportional to the
product of a turbulence factor and the velocity squared. Vrp = ρ p 1 − φ p k f , (10-1)
They showed that the pressure drop for this component
where ρp is the proppant density in lbm/ft3, φp is the
can exceed that of Darcy flow with high-velocity pro-
porosity of the propped fracture, and kf is the fracture
duction. Problems in design optimization of proppant
permeability (i.e., the permeability of the proppant in
conductivity are that the velocity varies down the frac-
the fracture). As stress increases, the relative proppant
ture and the turbulence factors vary with the proppant
volume (RPV) increases but more significantly for
type as well as the proppant’s change in permeability

0.006 Brown sand 20/40 mesh

Brown sand North white sand
0.0025
Relative proppant cost ($/md-ft3)

Relative proppant volume (lbm/md-ft3)
North white sand Resin-coated sand
0.005 Resin-coated sand Bauxite
Bauxite ISP
ISP 0.0020
0.004
0.0015
0.003
0.0010
0.002
0.0005
0.001
0
1000 3000 5000 7000 9000 11,000 13,000 15,000
0
1000 3000 5000 7000 9000 11,000 13,000 15,000 Closure stress (psi)
Closure stress (psi)

Figure 10-6. Relative proppant cost versus closure stress.
Figure 10-5. Relative proppant volume versus closure
stress for different proppant types (Elbel and Sookprasong,
1987). is less for sand at 10 ppg and the 1-year optimum is
achieved at 500- to 600-ft penetration. The more per-
low-strength proppants because of their loss of both meable premium proppant at 16 ppg with a penetration
permeability and porosity. The product of the RPV of 900 ft increases the NPV by 35%. Although the
and cost of each proppant plotted versus closure stress maximum NPV is achieved for a specific penetration,
in Fig. 10-6 reflects the cost effectiveness to achieve additional penetration results in more production—but
conductivity. These plots should be used within the at a higher cost.
bounds of practical proppant widths in lbm/ft2 of The role of fracturing fluid viscosity and leakoff
propped area (i.e., 1 to 3 lbm/ft2) and maximum slurry characteristics is generally well known for fracture
concentrations to achieve these widths, usually 16 ppg propagation and the placement of the propping agents;
for low-permeability reservoirs. however, other properties must also be considered. The
selected fracturing fluid should correctly balance the
following, usually conflicting, properties and features:
10-2.7. Treatment size • adequate fluid-loss control
If it is assumed that the fracturing fluid and injection • viscosity stability during placement for adequate
rate were selected by considering proppant transport, proppant transport
fluid loss, and horsepower and pressure limits, the • compatibility with the formation rock and reservoir
other major design considerations are treatment size, fluids
type of proppant and proppant scheduling. A general
• low friction loss in the pipe
statement can be made that the greater the propped
fracture length and the greater the proppant volume, • minimal damaging effects on proppant permeability
the greater the production. Limiting effects are • controlled breaking and cleanup properties
imposed by factors such as the size of the production • ease in mixing
string, limit of achievable fracture conductivity and
• minimum disposal problems
fracture height growth, in addition to well spacing.
Within these constraints the size of the treatment • operational safety
should ideally be based on the optimum fracture pene- • environmental safety
tration determined by the economic considerations • economical price.
discussed earlier. A plot of NPV versus propped pene-
tration is shown in Fig. 10-7 for a premium ISP and The last three considerations may eliminate systems
sand at concentrations of 10, 14 and 16 ppg. The NPV that may otherwise be applicable. The first seven con-

Sand 10 ppg
2,300,000 ISP 10 ppg
Sand 14 ppg
ISP 14 ppg
Sand 16 ppg
2,100,000
ISP 16 ppg
1,900,000
One-year NVP ($ )
1,700,000
1,500,000
1,300,000
1,100,000
100 300 500 700 900
Proppant penetration (ft)
Figure 10-7. Net present value versus penetration for various proppant concentrations and types.
siderations are controlled to various degrees by addi- improved by using additives such as solids, surfac-
tives, as discussed in Chapter 7. Experience in an area tants, liquid hydrocarbons and gases (see Chapter 8).
also influences the selection of a fluid. Experience can The time for a fracture to close after a large fractur-
be either positive from “fluid proof testing” or negative, ing treatment can be hours. During this time, a signifi-
impeding the consideration of potentially more effec- cant amount of the proppant can migrate to the bottom
tive fluid systems. Figure 10-8 provides a general of the fracture without connecting to the perforated
guideline for fluid selection with a distinction between interval (Cleary and Fonseca, 1992). Closure times
oil and gas wells. Experience has shown that both increase in cases with fracture height growth, particu-
water- and oil-base fluids have been used successfully larly into nonleakoff barriers and when low proppant
in oil and gas wells. The greatest concern is the use concentrations are used. Schlottman et al. (1981)
of oil-base fluids in dry gas wells; however, they have attributed poor apparent fracture conductivity in the
been used in gas condensate wells. early stages of the development of fracturing the
After the fluid considerations have been balanced for Cotton Valley formation in East Texas to a treatment
the important properties of fluid loss and viscosity, the proppant concentration of only 4 lbm/gal. At this con-
related additive concentrations remain for consideration. centration, only 25% of the fracture height was filled
if the proppant settled completely before closure.
Furthermore, during closure, the fill fraction increases
10-2.8. Fluid loss because of fluid loss (see Fig. 5-21). Increasing the
Fluid loss affects penetration and closure time. The proppant concentration in the fluid improved results
mechanisms that control fluid loss are discussed in on subsequent treatments. For this case, at least 50%
Chapters 6 and 8. There is some degree of depen- of the fracture was filled if the proppant concentration
dence on formation permeability, but the fluid-loss in the fluid was 10 lbm/gal.
control for almost any fracturing fluid system can be Fluid-loss control is also desired to minimize the
extent of damage to the matrix. The damage may be

Gas Well Oil Well
Less 175°F More No Very water sensitive Yes
Low pressure Low pressure Water Low

No or Yes No or Yes No Yes No Yes
water sensitive water sensitive sensitive pressure
Foamed
Less 150°F More fluids Less 300°F More Less 250°F More See Gelled oil Gelled oil + N2
70%–75% gas well
quality guide
Surface-delayed or low pH
crosslinked Low pH Titanate or
borate crosslinked zirconate No Low pressure Yes
crosslinked + +
25% CO2 crosslinked
guar or 25% CO2 +
HPG Super foam 25% N2
borate + N2 Less 200°F More Less 200°F More
Delayed borate,
titanate or Oil: polymer Foamed Less 250°F More
zirconate Zirconate
crosslinked emulsion fluids
crosslinked
guar HPG 70%–75%
Low pH Titanate or
quality
crosslinked zirconate
No kfw > 1000 md-ft + xf > 300 ft Yes + crosslinked
25% CO2 +
25% N2
Linear fluids
Figure 10-8. General guideline for fracturing fluid selection in oil and gas wells.
due to physical or chemical alterations to the matrix 10-2.9. Viscosity effects

that result in a decrease in matrix permeability. It can
also be associated with gel residue causing a reduction The ability of fracturing fluids to transport proppants
in fluid mobility, particularly in formations with high over long distances can be the limiting factor in frac-
permeability. Holditch (1979b) and Pope et al. (1996) ture length optimization. Because of the problem of
showed that for typical invasion depths of a few inches viscosity degradation with time and temperature, treat-
and degrees of mobility reduction as high as 90%, the ments usually start with a higher viscosity than that
damage has a negligible effect on production, but required in the later stages. This has resulted in the
Montgomery and Berthelot (1990) and Mathur et al. development of highly viscous, crosslinked fracturing
(1995) showed that for high-permeability reservoirs fluids. Technology in fracture fluid chemistry contin-
fracture face damange can have a significant impact ues to evolve to minimize the role of temperature on
on well performance and can affect the postfracturing viscosity degradation.
evaluation if not taken into account. Proppant transport concerns typically result in
In fracturing highly permeable formations, the depth designing the treatment with a fluid viscosity higher
of fluid-loss penetration of a highly viscous fluid may than necessary. Nolte (1982) showed that without
be tens of inches. In the absence of an effective breaker complete consideration of the effects of the fluid’s
for the fluid, a pressure drop of several hundred psi can behavior on the proppant settling rate, the design vis-
be required to regain sufficient permeability to not cosity may be up to 50 times greater than required.
affect production significantly. Damage to the proppant The polymer concentration in water-base fracturing
pack is usually of more significance than damage to fluids should be minimized because of the adverse
the matrix, particularly in high-permeability formations effects of residues on proppant conductivity and
that require highly conductive fractures; however, both higher pressures on fracture geometry (Nolte, 1982,
types of damage should be considered in fracturing 1988c).
high-permeability formations. Schlottman et al. (1981), White and Daniels (1981)
and Nolte (1982, 1988a) showed the benefits of tapered

polymer loading during a treatment. The polymer indicates that a viscosity as low as 50 cp at 170 s–1 is
concentration is based on maximum exposure to time sufficient for proppant transport in a crosslinked fluid.
and temperature for different segments of the fractur- Concerns about polymer damage to the proppant pack
ing fluid during injection. Fluid exposure time is dis- and improvements in fluid systems and mixing have
cussed in Sidebar 10B. An example of the time that resulted in treatments using low-guar fluids and poly-
the fracturing fluid is at the maximum temperature in mer-free, water-base fluids (see Chapter 7).
the fracture and a resulting tapered polymer schedule Higher polymer concentrations can result in higher
for a treatment are shown in Fig. 10-9. Not only is fluid efficiency because of the effect of the lower
there a savings in polymer cost, but there is also less fluid-loss rate through the wall filter cake Cw and the
potential polymer damage to the proppant pack per- viscosity control leakoff effect Cv (see Chapter 6),
meability and less potential of exceeding a critical net greater fracture width from the higher net pressure
pressure for efficient fracture extension. Experience or both cases.
10B. Fluid exposure time
Polymer concentration can be excessive because of conservative laboratory test procedures. The tests are usually run at the bottom-
hole temperature for a time equal to the job time, but in fact the maximum exposure of any increment of fluid at the bottomhole tem-
perature will be less. The selected polymer concentration will maintain a greater viscosity than that required for the actual exposure
conditions. A fracture design model should determine the exposure times of designated fluid stages that can be matched to laboratory
data for determining more realistic conditions, thereby preventing the use of excessive polymer. A preferred case is for the laboratory
data to reflect the actual heat-up condition experienced by the fluid.
Figure 10B-1 shows the maximum time in the fracture for different increments of fluid before shut-in and exposure time in the frac-
ture before reaching various temperatures. The time that an increment of fluid is in the fracture can be approximated by a triangle with
the base (x-axis) scaled to the pumping time or treatment volume (pumping time × constant injection rate). The y-axis is the time in the
fracture. Although the total pumping time in this example is 280 min, the maximum time that any increment of fluid is in the fracture is
120 min before it is either depleted early because of fluid loss or pumped later during the treatment. On the figure this condition occurs
at the apex of the triangle.
For continuous polymer mixing, relatively small increments of various polymer concentrations can be considered. For batch mixing in
tanks, the tank size is the obvious increment. For example, fluid from tank 1 is in the fracture for a maximum of about 37 min. The incre-
ment of fluid corresponding to the apex of the triangle (tanks 7 and 8) has the maximum time in the fracture. This corresponds to the
last of the pad fluid required to get the proppant to the tip of the fracture. The fluid reaches the maximum temperature of 360°F [180°C]
about 60 min after its injection, and tanks 7 and 8 span an exposure time of about another 60 min at that temperature. The additives for
tanks 7 and 8 should be based on this time of exposure and the laboratory test selected and performed accordingly. Fluid pumped prior
to this time is lost into the formation before the end of pumping (pad depletion), and fluid injected later has some portion of it left in the
fracture for a time equal to the job time minus the time of injection. For these fluids, the test data should also be selected and performed
accordingly. The closure time, which depends on the local proppant slurry concentration, can be added to the test time for the later flu-
ids to reduce proppant fall during closing. An exception to reducing the duration requirement for the pad fluids would be cases where
fluid leakoff into the matrix is viscosity controlled.
If the proppant slurry should not be exposed to temperatures higher than 360°F during injection, proppant addition should be delayed
until tank 9 for this 12-tank treatment. With this planned delay, proppant will not reach the hydraulic tip of the fracture because of the
excessive pad, referred to as the cool-
120
down pad. Higher proppant concentra-
Total job time 280 min Time in fracture
tions should be considered to account
for the redistribution of proppant during
closure and to reduce the time for clo-
Time in fracture or to reach temperature (min)
100
sure on the proppant.
From this example, and by consid-
ering the other temperatures plotted on
80 the figure, an effective strategy can be
developed for adding fluid breaker con-
360°F centrations and potentially different
60 types that are more effective for various
temperature ranges
40
250°F
20
200°F
Adding proppant
0 Tank 1 Tank 7 Tank 8 Tank 12 Figure 10B-1. Time and tempera-
0 1 2 3 4 5 6 7 8 9 10 11 12 ture exposure in a fracture per tank
Tank sequence number of treating fluid.

corresponding friction pressure typically limit the

Gel schedule (lbm polymer)
injection rates as a result of tubing or wellhead pres-
x50 x50s x50 x40 x30 sure ratings. The increase in surface pressure increases
150 10
the horsepower requirement and cost. These factors
120 8 are discussed in more detail in Chapter 11.
x50s
Time at temperature (min)
Time at Height growth (see Chapter 5) is also affected by
Proppant (lbm/gal)
temperature
90 before 6
the viscosity and injection rate in some cases. Table
depleted 10-1 shows results of P3D simulations at rates of 10,
x50s
Design
schedule
20 and 40 bbl/min. The fracture height growth is con-
60 4
tained by a barrier below and is restricted by an upper
barrier with stress 500 psi greater than that of the
x40
30 2
100-ft pay zone. The upper barrier has no leakoff.
The fluid volumes required to obtain 600-ft penetra-
x30
Pad
0 0 tion decrease with higher rates. However, these sav-
0 80,000 160,000 240,000
ings are offset by the increased horsepower, and the
Fluid (gal)
proppant volume increases with increasing rate
Figure 10-9. Fluid selection for exposure time and prop- because of more width and height growth. The associ-
pant and fluid scheduling (Nolte, 1982). An s with the poly- ated cost of the three variables is shown in the last
mer loading value indicates that it is stabilized.
column. For an upper barrier stress differential of
2000 psi with the height well contained, the compar-
isons are similar. These relationships are generally
applicable for growth into an impermeable stress bar-
10-2.10. Injection rate
rier only, whereas growth through a barrier and into
Selection of the fracture treatment injection rate a lower stress zone can be extremely detrimental
depends on a number of factors. Generally, high injec- (Nolte, 1982, 1988c). Nolte also showed that viscosity
tion rates should be considered because of increased reduction is about twice as effective for height mitiga-
treatment efficiency resulting from decreased fluid- tion as rate reduction, on a comparable percentage
loss time and increased fracture width. Higher rates reduction.
also directly improve proppant transport capabilities Britt et al. (1994) reported a 40% savings in treat-
because of an increase in slurry velocity relative to ment costs by reducing the polymer concentration and
proppant fall rates and a reduced pumping period, injection rate to limit the height growth and improve
leading to less time for proppant fall and less viscosity production. These examples demonstrate the result of
degradation. The size of the treating tubulars and the good engineering practices for specific applications.
Table 10-1. Sensitivity of height to rates of 10, 20 and 40 bbl/min.
Injection Wellbore Slurry Surface Hydraulic Proppant Cost

Rate Height Volume Pressure Horsepower Mass ($)
(bbl/min) (ft) (gal) (psi) (lbm)
Upper barrier stress + 500 psi

10 142 42,500 3264 800 97,000 35,000
20 164 35,000 4040 1980 116,000 39,000
40 203 33,000 5537 5428 142,000 58,000
Upper barrier stress + 2000 psi

10 104 40,700 3300 809 86,000 33,000
20 106 31,000 4075 1998 95,000 35,000
40 108 29,000 5630 5520 121,000 54,000

10-3. Geometry modeling The various fracture geometry models are discussed
in Chapters 5 and 6. The three basic types of geome-
An important step in fracture design is modeling the try models described in Sidebar 10C can be further
geometry and proppant placement expected for the subdivided as follows:
specific treatment conditions. The simulation allows
the design engineer to • 2D
– PKN
• ensure that the proppant addition schedule does not
cause an undesired screenout – KGD
• determine the treatment fluid and proppant volumes – radial
required to achieve a desired fracture penetration • 3D
• ensure that the proppant concentration per unit area – lumped P3D
of the fracture face provides adequate fracture con- – discrete cells P3D
ductivity over the complete extent of the pay zone. – planar 3D
10C. Geometry models
Schematics for fracture width are shown in Fig. 10C-1 for the 2D PKN, P3D and multilayer fracture (MLF) models in a multilayer setting.
• For the 2D PKN model, the fracture height estimated by the engineer remains constant for the simulation. The fracture length
grows from a line source of perforations, and all layers have the same penetration. The simulation can be approximated by the
average modulus of all the layers, with the reduced width from a higher stress layer between the sands accounted for by a multi-
plying correction (e.g., about 2; Nolte, 1982).
• For the P3D model, the fracture initiates in the zone with the lower in-situ stress. The height growth is determined by the bounding
layers’ stress and other mechanical properties. Growth into the other sand layer depends on the stress and thickness of the
interbedded shale layer and the distance between the two; it is independent of the wellbore and perforations in the layer. With a
relatively low stress contrast, the two fractures join rapidly and behave as a single fracture. The height growth beyond the three
layers depends on the stress and modulus profile of the adjoining layers. The simulated penetration is generally greater in the
lower stress zone. The P3D model is a common geometry model for fracture design.
;; ;;
; ;;
; ;;
;
This discussion is for the discrete cell implementation (see Section 6-3). Depending on the implementation, the lumped P3D
model may not provide a variable width profile or differentiate the lengths for two zones.
• The MLF model allows simulating simultaneous fractures. The fractures (PKN) in the layers are initiated when the wellbore pres-
sure is above the layer’s stress. This model is the most applicable when separate fractures initiate and they do not coalesce,
which is the expected case. After the MLF model is used to define the relative injection rates for the zones, the P3D model can be
employed for a more detailed consideration of each zone (see Section 10-5.4). The fractures can have different lengths, and each
fracture’s geometry depends on its height, net pressure, modulus and efficiency. This model can also address the application of
;; ;;
; ;;
; ;;
;
limited entry and determination of the stages required for adequate stimulation of a number of layers.
Gamma ray Layered formation 2D P3D MLF

Perforations
;;;;
;;;;;;;
Perforations
Shale Sand
Figure 10C-1. Fracture geometry of 2D, P3D and MLF models.

• multilayered schedule depends primarily on the efficiency; there-

– PKN fractures fore, models that provide very different widths can
predict the same schedules by calibrating a different
– P3D fractures.
fluid-loss coefficient for the specific width assump-
They can be further classified as to how they model tions. It is still good practice and not unusual to use
fluid loss, postpumping behavior, tip effects, poro- several models during the design phase; however, the
elastcity, various forms of fluid flow (1D and 2D), PKN model is a more appropriate approximation for
proppant transport, etc. (see Chapter 6). fractures where the length is considerably longer than
the height, and the KGD model should be used for
fractures where the fracture length is of the same
10-3.1. Model selection order or shorter than the fracture height. An example
The wide range of models and features available can is presented in Section 10-6.3.
make selecting a model an overwhelming task. Gen- Fracture design for a wildcat or high-risk well calls
erally, the model should be selected to match the level for the collection of a comprehensive set of design
of complexity required for the specific application, data and the use of more sophisticated models. This
quantity and quality of data, allocated time to perform is also the case when treating problems or lower than
a design and desired level of output. expected production rates occur.
Modeling hydraulic fracture propagation with a pla- Geometry models are typically validated with well
nar 3D model can be time consuming. Not as much test analysis and production history matching. The
time is required for 2D modeling, but the results can evaluation of fracture treatments, however, requires
be simplistic. P3D models provide a compromise and sophistication in pressure transient analysis and the
are most often used in the industry for the evaluation determination of rock properties for use in the geome-
of hydraulic fracturing treatments. try models (Elbel and Ayoub, 1992). The determina-
A diagnostic method based on the treating pressure tion of an average permeability may not be sufficient
history and analysis by history matching (Nolte, 1982, for production history matching. Layered reservoirs
1988c) can be used to identify various modes of prop- with large permeability contrasts (Bennett et al., 1986;
agation (see Chapter 9). Using this technique, the Camacho-V. et al., 1987) or reservoirs with horizontal
engineer selects the appropriate model and uses it to permeability anisotropy (Ben Naceur and Economides,
evaluate important fracture parameters such as treat- 1989) appear to have shorter fractures if the well test
ment efficiency and net treating pressure. The concept models used assume a single isotropic layer.
of history matching a calibration injection test and In addition to considering the effects for idealized
using the parameters required for the match in the reservoir assumptions, Nolte and Economides (1991)
design of the propped fracture treatment forms the discussed and attempted to quantify the effects for
foundation for effective treatment design. By applying idealized assumptions related to the design and place-
this method, the engineer becomes familiar with the ment of the fracture. The common assumptions about
advantages and limitations of the various geometry fractures and reservoirs were found to generally result
models and good engineering practices, which with in a productive length that is less than the design
experience leads to successful fracturing programs length. The coupling of the shorter productive length
(Martinez et al., 1992; Johnson et al., 1993; Stadulis, and the shorter apparent length (to compensate for the
1995; Gulrajani and Romero, 1996). reservoir model assumptions) compounds the prospect
When the controversy between 2D KGD and PKN that the actual production will be less than anticipated.
models was high (e.g., Perkins, 1973), it was not An example indicated that an effective reduction of
uncommon to simulate the treatment design with both the fracture to 33% of the design length would not be
models, looking at each for potential problems with an unreasonable condition.
the final design. This practice did not ensure an opti- They cast the reservoir and placement effects in
mum design but was used mainly to minimize prema- terms of a design factor that has the role of a safety
ture screenouts by calibrating the fluid-loss coefficient factor used in other engineering fields. General frac-
on the basis of the results of prior treatments. Also, as turing practice incorporates these effects in an ad hoc
subsequently discussed for Fig. 10-11, the proppant manner and through experience with results that fail to
meet the expected ROI (e.g., increased fracture height,

fluid-loss coefficients, propped width). The design

10D. In-situ stress correlation with lithology
factor, equal to 33% for the cited example and analo-
gous to the inverse of a safety factor, can be incorpo- Correlations based on lithology can be developed from long-
rated directly into the economic optimization process spaced sonic logs, core analyses and in-situ stress tests
from a limited number of wellbores and subsequently used
by effectively reducing the length axis by this factor to estimate properties for similar formations in the region.
for the expected revenue. A formal, comprehensive A common means of correlating stress is using Poisson’s
ratio ν obtained from sonic logs. Table 10D-1 provides
identification process of the various model effects is ranges of ν for various lithologies determined from cores or
not warranted for every treatment; however, consider- inverted from field stress tests (Holditch and Rahim, 1994).
ation of this process for each reservoir unit results in Poisson’s ratio versus the minimum stress gradient for three
reservoir pore pressures is shown in Fig. 10D-1.
more effective treatments for the unit, particularly for
the initial treatment that is early on the learning curve Table 10D-1. Poisson’s ratio for various
(Gatens et al., 1991; Nolte and Economides, 1991). lithologies (Holditch and Rahim, 1994).
Rock Type Poisson’s Ratio

10-3.2. Sources of formation parameters A. Gas-bearing sandstone 0.10 to 0.25
B. Wet sandstone 0.25 to 0.30
The 3D models require more data, primarily in the
C. Wet siltstone 0.20 to 0.30
form of profiles of stress and moduli. The data are
D. Limestone 0.30 to 0.32
obtained from log analysis, measurements on cores
E. Shale 0.28 to 0.43
and interpretation of pressure from injection tests.
Inverse modeling and injection pressure history
matching (Bhalla and Brady, 1993; Gulrajani et al., 50
1996; see Chapters 6 and 9) can also be used to deter- 40
E
mine unknown parameters.

Poisson's ratio
D
30 B C
Usually, log data must be averaged over intervals A
within the bounds of layers of higher contrasts. The 20
height of the layers averaged should be consistent
10
with lithology changes (Holditch and Rahim, 1994).
Default values are commonly used for the lithology 0
0.10 0.30 0.50 0.70 0.90 1.10
and porosity. If possible, data should not be accepted Stress gradient (psi ⁄ ft)
on a stand-alone basis but checked for consistency
with other values either measured or implied. There Figure 10D-1. Poisson’s ratio versus stress gradient
for three reservoir pore pressures (modified from
is usually correlation among lithology, porosity and Holditch and Rahim, 1994).
Young’s modulus (Morales and Marcinew, 1993);
e.g., a porous section of an otherwise dense limestone Figure 10D-1 is a modification of a Holditch and Rahim
(1994) figure of the linear relation of stress versus pore pres-
should have a lower modulus. There is also a correla- sure. The effect of tectonic stress (i.e., the ratio of maximum
tion among reservoir pressure, Poisson’s ratio and to minimum horizontal stress σH,max /σh,min) was not taken
into account when inverting for ν from stress tests, as it is
minimum stress as introduced by Rosepiler (1979); embedded in the Poisson’s ratio correlation. The stress
therefore, pressure depletion should increase the stress ratio can increase the minimum stress. The effect that a
σH,max /σh,min of 1.3 has on the correlations is shown in Fig.
contrasts between the reservoir and nonpermeable 10D-1 by the dashed curves. If the ratio is known (from the
barriers. In general, log-inferred stress profiles require differential strain curve, breakdown pressure or wellbore
calibration from stress tests (Nolte, 1982, 1988c) or breakout analyses) and accounted for, the inverted value of
ν will be lower. For example, for a stress gradient of 1.0 psi/ft
directly from injection pressure data (Nolte and Smith, in a formation with a pore pressure gradient of 0.7, the
1981; Gulrajani et al., 1997a). Sidebar 10D provides a inverted ν is 0.43 without tectonic effect and only 0.30 for
a stress ratio of 1.30.
discussion of in-situ stress correlation with lithology. As the pore pressure gradient increases, the effect of ν
becomes less. It can also be seen that decreases in stress
resulting from pore pressure depletion are greater in cleaner
sands. Consequently, depletion can increase the stress con-
trasts among sandstones, siltstones and shales. The pore
pressure in impermeable shales does not decrease because
of production, which further increases stress contrasts with
the pay zone. The magnitude of the stresses obviously

10-4. Treatment schedule a catastrophic event such as an undesired screenout,

which can be caused by insufficient width, pad deple-
Most fracturing treatments are performed to bring eco- tion or slurry dehydration near the wellbore resulting
nomic viability to formations with low permeability; from a high proppant concentration. Historically,
however, as discussed in Chapter 5, the rapid growth proppant scheduling has consisted of gradual incre-
of fracturing after its introduction was primarily for mental increases in proppant concentration during the
applications to bypass wellbore damage. The sensi- course of the treatment and was based on experience.
tivity of fracturing economics to formation perme- The schedules were conservative to avoid screenouts.
ability and skin effect is discussed in Sidebar 10E. During the treatment, the slurry concentration was
The goal of a treatment design is to provide a typically increased if the treating pressure decreased.
schedule for injecting the treating fluid and proppant. If the pressure started to rise, the concentration was
The schedule reflects the volume of fluid based on the decreased under the belief that treatment pressure was
desired penetration and viscosity profile and the mass greatly affected by proppant concentration at the well-
and type of proppant based on the desired conduc- bore. Nolte and Smith (1981) introduced the monitor-
tivity. Scheduling the proppant addition rate during ing of net fracturing pressure and showed that there is
the treatment is important. A major goal is to prevent a characteristic signature of pressure increase caused
10E. Fracturing economics sensitivity to formation permeability and skin effect
To illustrate some economic aspects of fracturing, consider the optimization of fracture treatments in formations with 50 ft of net pay
and permeabilities ranging from 0.001 to 10 md. For this example, all the reservoir and formation properties are the same except for
the permeability and wellbore skin effect s.
Three-year NPVs are used to determine
the optimum penetration of the fracturing Table 10E-1. Optimum penetration for different permeabilities
treatments for the different values of per- based on three-year net present value.
meability and skin effect. Table 10E-1 lists
the optimum penetration values, which for Formation permeability (md)
this example do not change with the skin
effect. 10 1 0.1 0.01 0.001
The production revenues and NPVs are
shown in Fig. 10E-1. The total production xf (ft) 400 1100 1300 1300 700
revenue is the greatest for the 10-md for- CfD 0.5 2.1 19 189.2 3000
mation and decreases, as expected, with
each decrease in permeability. However, Fluid (gal) 60,000 423,000 588,000 588,000 174,000
the net (postfracture minus prefracture) Proppant (lbm) 37,000 136,000 170,000 170,000 76,000
revenue is low for the 10-md case
Cost ($) 47,000 205,000 275,000 275,000 120,000
because of the high prefracturing produc-
tion. Net revenue increases for the 1.0- Note: 12-ppg maximum concentration
and 0.1-md cases and decreases for the
0.01- and 0.001-md cases. The NPVs
follow the net revenue pattern.
The results of repeating the optimization with s = 20 and 100 are also plotted on Fig. 10E-1. A positive skin effect has the effect
of decreasing the prefracture revenues; the postfracture revenues remain the same. For s = 100, the net revenue and NPV for the
10-md case increase considerably; however, the 1-md case now has the highest net revenue and NPV. Success in fracturing high-
permeability formations is attributed to overcoming the
14
large wellbore skin effect usually associated with these
Total revenue (all)
formations.
12 Net revenue with s = 0
Increasing the tubing size and utilizing tip screenouts
NPV with s = 0 for high-permeability cases and optimizing proppant
10 and proppant concentration and increasing penetration
Revenue (million $)
in low-permeability cases are good engineering prac-

8 s = 100
tices; however, they will not significantly alter the
points made by this example.
6
4 s = 20
0
10 1 0.1 0.01 0.001 Figure 10E-1. Economic sensitivity to formation
Formation permeability (md) permeability and wellbore skin effect.

by tip screenout (TSO) and that the continued injec- halfway down the length. Comparison of the simulation
tion of high, or even higher, concentrations of prop- of the 1-ppg 40,000-gal TSO case with the same
pant for a considerable time is possible. Although the propped volume injected at 2 ppg shows about the
fracture extension is arrested, it may be possible to get same concentration in lbm/ft2 for the tip half of the
the designed final concentration of proppant into the fracture but less concentration over the wellbore half.
fracture to achieve the desired conductivity. This is because the concentration at the wellbore is only
1 ppg. Increasing the proppant concentration (ppg) dur-
ing the treatment is required to obtain a fairly uniform
10-4.1. Normal proppant scheduling propped width down the fracture.
The optimum design for a conventional fracture treat- The efficiency of the treatment determines the prop-
ment is one in which the pad volume has leaked off pant addition schedule that will achieve a specific
into the formation and the proppant has reached the slurry concentration in the fracture at the end of
tip at the end of pumping, leaving the fracture filled pumping (Nolte, 1986b). For this approximation
with the proppant-laden slurry to provide a fairly uni- method, the volume and efficiency of the treatment
form propped width and sufficient conductivity to must be known, and the efficiency can be estimated
minimize the pressure drop during production. from a calibration treatment. The following discussion
The proppant concentration (pounds of proppant uses this approach; however, for general application a
added to 1 gal of fluid, or ppg) in any segment of slurry numerical placement simulation should also be used
increases because of fluid loss as the slurry moves to define the volume, schedule and placement (see
down the fracture. The propped concentration in lbm/ft2 Section 6-11).
of fracture area depends on the rate of fluid loss from Figure 10-11 shows plots of addition schedules for
the slurry and the fracture width profile. This is illus- a uniform concentration at shut-in and in terms of
trated in Fig. 10-10 by simulations of treatments with various efficiencies based on the equations in Sidebar
only one proppant concentration. All the treatments 6L. If the efficiency is low, the ramp is almost linear;
have the same pad volume: two of the simulations are if it is high, the initial concentrations must be ramped
for 20,000 gal of slurry at 1 and at 2 ppg, and the third up rapidly. However, this figure does not represent
is for 40,000 gal at 1 ppg with the last half of the slurry equal-penetration fractures. Small errors in the initial
injected after a TSO. The top set of curves shows the ramping or efficiency estimate can cause dehydration
proppant concentrations in the slurry increasing over near the tip, so in practice, the early stages are usually
the fracture length. The bottom set of curves shows the not increased as rapidly.
corresponding propped width profile decreasing slightly Figure 10-11 or the associated equations can be
away from the wellbore and then starting to rise about used to provide a schedule to place a specific, uniform
5
0.7
4
3 2 ppg in 20,000 gal 0.6
Concentration (lbm/ft2)
2
Concentration (ppg)
1 ppg in 20,000 gal 0.5

1
1 ppg in 40,000 gal, TSO after 20,000 gal
0.4
0.3
2 ppg in 20,000 gal 0.2

1 ppg in 40,000 gal, TSO after 20,000 gal
1 ppg in 20,000 gal 0.1
0
0 100 200 300 400 500 600 700
Fracture half-length (ft)
Figure 10-10. Proppant concentration profiles for slurries injected at 1 and 2 ppg.

profile of proppant concentration and also to provide Figure 10-12 shows these proppant addition schedules
a quick calculation routine. An inconvenience for this for a treatment with an efficiency of 0.5; the stair-
approach is that the figure is for proppant volume stepped schedule shows slightly less concentration
fractions and must be converted to ppg units for prac- initially compared with the ramp and reflects a com-
tical applications. The figure is presented in a dimen- mon cautionary practice to avoid a screenout prior to
sionless framework and the area under any particular the reaching the tip. This stairstepped schedule is used
curve is equal to the fluid efficiency for that treatment. in Fig. 10-13 for a simulation illustrating proppant
The fraction of area under any segment of a curve— movement from the end of pumping to closure.
i.e., from the end toward the beginning pad portion— The schedule in Fig. 10-13 results in a uniform slurry
is the same fraction of the fracture volume from the concentration at the end of pumping. If there were no
wellbore toward the tip and therefore defines the further movement of slurry during closure, the prop-
placement location of that portion of the proppant. pant concentration would range from 1.6 lbm/ft2 at
The concentration for a segment can be changed rela- the wellbore to 1.1 lbm/ft2 near the fracture tip. How-
tive to the reference value by a proportional change in ever, the slurry continues to move after shut-in, as
the value of the ordinate for the segment on the figure. shown by Nolte (1986b), and is seen here to result in
Similar concepts for placement and proportional a fairly uniform concentration (lbm/ft2) profile over
changes in concentration can be incorporated into a most of the fracture with an additional penetration
more accurate, automated numerical scheduling proof about 5%. This result is the basis for a design with
cedure presented by Nolte (1982, 1988c) for a fracture constant ppg at shut-in as a first target for a final
placement model to determine a specific ramping schedule (e.g., Fig. 10-11).
schedule and pad volume (see Chapter 6). The efficiency can also be used in Eq. 5-22 to esti-
The proppant concentration in the slurry determines mate the pad fraction required for the total treatment
the ratio of the final propped width to hydraulic width. volume. The pad fraction, in addition to creating
This relation can be inferred from the plot of fracture adequate fracture width and preventing premature
fill versus proppant concentration in ppg shown in screenout, can also influence slurry movement during
Fig. 5-21. If the last injected stage is at a concentra- the fracture closure stage. If excess pad remains ahead
tion of 8 ppg of sand, the average propped width at of the slurry, it can continue to extend the fracture,
the wellbore cannot be greater than 40% of the and the position of the slurry front will continue to
hydraulic width. move forward until it dehydrates. Figure 10-14 com-
The continuous curve, or ramp, schedule is typically pares the proppant concentration profile for the previ-
represented as a stairstepped, incremental schedule ous simulation with a profile that has twice the pad
to facilitate simulation or operational requirements. volume. Additional penetration from the excess pad
Efficiency 10
0.7 0.5 0.3 0.1 9
0.6 0.4 0.2
1
8
Slurry concentration (ppg)
0.9
7
Fraction of final concentration
0.8
6
0.7
0.6 5
0.5 4
0.4 3
0.3
2
0.2
1
0.1
0 0
0 0.2 0.4 0.6 0.8 1 0 5 10 15 20 25 30 35 40 45 50
Fraction of job time Job time (min)
Figure 10-11. Proppant addition schedules for different Figure 10-12. An optimally designed proppant addition
treatment efficiencies. schedule, both ramped and stairstepped.

25 2
Proppant added (EOJ)
Total proppant (ACL)
20
End of pumping concentration (lbm/ft2 of fracture area)
Concentration (ppg)
15
Redistribution during 56-min closure (lbm/ft2)
1
10
End of pumping (ppg)
0
0 100 200 300 400 500 600 700
Figure 10-13. Change in fracture proppant concentrations from the end of pumping to closure. EOJ = end of job,
ACL = after closure.
12 2
10
8
Concentration (ppg)
Optimum pad
2× optimum pad
6 1
2
Proppant added (EOJ)
Total proppant (ACL)
0 0
0 100 200 300 400 500 600 700
Figure 10-14. Proppant concentration profiles with different pad volumes.
volume is limited to about 10% because of slurry to decrease CfD. Fractures with an equal or less than
dehydration near the tip. optimum CfD (i.e., conductivity limited) have no addi-
Excessive extension after shut-in can cause the final tional benefit from the increased length and may have
concentration of the proppant in lbm/ft2 of fracture lower production (see Section 10-2.2).
area to decrease as the width of the fracture decreases, Insufficient pad results in less fracture penetration
creating a longer fracture with less conductivity. Both and possibly premature termination of the treatment.
increased length and decreased conductivity work This is the case particularly with stiffer fractures that

are either short or in high-modulus formations. In the cient way to obtain increased conductivity than by
stiffer cases the pressure increase is usually relatively relying on only the increase in width.
rapid, generally preventing placement of the design
• Tip-screenout design
volume of proppant or in extreme cases not allowing
enough time for flushing of the wellbore with only Smith et al. (1987) introduced a method of TSO
a portion of the proppant placed. design utilizing an analytical radial simulator
through pad depletion and calculating the extra
leakoff over a constant area after bridging. Nolte
10-4.2. Tip screenout (1986b, 1990) published analytical relations based
on the efficiency at screenout for the ramp schedule
Hydraulic fracturing in high-permeability reservoirs and width increase for the additional injected vol-
differs from conventional fracturing in that the objec- ume after a TSO (see Sidebar 6L). Martins et al.
tive is more to generate fracture conductivity than (1992c) modified Nolte’s concepts by extending
length. Techniques for increasing fracture conductivity the initial low-concentration proppant stage to min-
include increasing proppant size for cases that do not imize subsequent screenout at an intermediate dis-
produce formation fines into the proppant pack, tance that could lead to detrimental rapid backward
increasing proppant concentration (all applications), packing of proppant and a pressure increase, partic-
using clean fluid (all applications) and using TSO ularly for stiff fractures.
design techniques. Each of these techniques has posi- TSO treatments can be designed using either an
tive results on increasing fracture conductivity and the appropriate placement simulator or analytical meth-
success of high-permeability fracturing; however, the ods based on efficiency (see Section 6-11). The
application of TSO fracturing, introduced by Smith et analytic technique uses the control of the TSO
al. (1984), has revolutionized the completion and fracturing process by pad depletion and material-
stimulation of wells in higher permeability reservoirs balance considerations. Because the TSO design
(Smith et al., 1987; Monus et al., 1992; Martins et al., is predicated on fracture storage and fluid loss
1992c; Hannah et al., 1994). for a fixed penetration, width increases are readily
TSO treatments can achieve proppant concentrations determined as a function of fluid efficiency and
in excess of 20 lbm/ft2 of fracture area, which can fracture volume increases resulting from the
represent a 10- to 20-fold increase in fracture conduc- injected volume after screenout, as illustrated in
tivity. Similar conductivity increases can be achieved Fig. 10-15 (K. G. Nolte, pers. comm., 1984). For
by using synthetic proppant (relative to sand) at higher example, at 50% efficiency at the time of screenout
stress. The compounding effect can result in an and to achieve a twofold increase in the hydraulic
approximately 100-fold increase in conductivity. fracture width ratio of channel width w to the width
Well productivity may be increased 4- to 7-fold if after screenout wso requires injecting a total volume
production from the well is conductivity limited (see that is 1.8 times that injected before the screenout.
Fig. 10-4). An analytic-based proppant addition schedule
A TSO is designed to deliberately cause proppant for a TSO fracture stimulation can be designed in
to pack at a specific location because of width restric- a manner similar to that for conventional treat-
tion, pad depletion or slurry dehydration. Once pack- ments. First, the pad fraction is determined from
ing occurs, further fracture propagation ceases at this the efficiency before screenout and preferably
point, usually at the tip, and generally along the obtained from a calibration treatment. Second, the
restricted width over the entire perimeter. Continued fluid efficiency and desired fracture width are used
injection increases the hydraulic fracture width and as in Fig. 10-15 to determine the additional slurry
final conductivity. The ratio of propped to hydraulic volume relative to the volume at screenout. Finally,
widths at the wellbore is controlled by the final prop- with these parameters and the final efficiency, the
pant concentration in the injected slurry (see Section proppant addition schedule can be defined as in
10-4.1), and proppant conductivity down the fracture Sidebar 6L.
is governed by the proppant concentration schedule, As discussed previously and illustrated in Fig. 10-16,
fluid-loss rate and backward packing of proppant from Martins et al. (1992c) modified the proppant sched-
the fracture tip to the wellbore. Increasing the slurry uling method by increasing the initial low-concentra-
concentration after the onset of a TSO is a more effi-

Efficiency at screenout, ηso

1.0 0.8 0.7 0.6
0.5
2.4
0.4
2.2
0.3
2.0
Injection time/time to screenout
0.2
1.8
0.15
1.6
0.1
1.4
0.05
1.2
1.0 0
1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0
Screenout width ratio, w/wso
Figure 10-15. Width ratio increases after screenout as a function of fluid efficiency and injection time.
tion stage, which still reaches the tip and bridges at final 10-ppg stage continues for the additional time.
the same penetration, and designing a new schedule Another TSO schedule extrapolating the slurry
based on the end of job (EOJ) efficiency and a ramping to a higher concentration of 14 ppg for the
longer job time. The new schedule starting regular additional time is shown by curve C. These tech-
proppant addition at a later time is initially a more niques are prone to creating intermediate dehydra-
aggressive schedule because of the higher efficiency tion locations and rapid pressure rises. Nolte’s
after screenout. The following discussion of Figs. method (see Sidebar 6L) of modifying the ramping
10-16 and 10-17 provides insight to these various schedule on the basis of the estimated EOJ effi-
scheduling assumptions. The figures show a normal ciency and using the efficiency at screenout for the
schedule and four different TSO design assumptions. pad staring time is shown as curve E. For these
A normal treatment design with 0.5 efficiency latter two methods, the lower initial concentrations
and final proppant concentration of 10 ppg is minimize intermediate screenouts or backward
labeled A in Fig. 10-16. The average propped frac- packing. Martins et al.’s method is shown as curves
ture width for this schedule is 1.2 lbm/ft2. From D and D – i. The main ramping curve D is based
Fig. 10-15, increasing the hydraulic width by a fac- on the EOJ efficiency, but proppant must be initiated
tor of 2.5 with a TSO requires about 2.15 times earlier for the screenout, which is shown as curve
more slurry (35,000 to 75,000 gal). Curve B shows D – i. For this example, with a long fracture and
the same schedule up to the TSO, after which the high efficiency, the lower concentrations in early

14
A. Normal 10-ppg treatment
B. 10 ppg continues after TSO
12 C. Linear extrapolation to 14 ppg
D. 14-ppg final stage after TSO
D – i low-concentration stage C
Slurry concentration (ppg )

10 E. Nolte method (see Sidebar 6L)
B
8
A E
4
D
2
D–i
0
0 20 40 60 80 100
Time (min)
Figure 10-16. Various proppant scheduling practices after tip screenout.
100 5
E. Nolte method
(395,000-lbm
proppant) 4
D. TSO with linear extrapolation to 14 ppg
(420,000-lbm proppant)
Concentration (ppg)
50
B. TSO with low-concentration stage
C. TSO with 10-ppg final stage (375,000-lbm proppant) 2
1
A. Normal treatment to 10 ppg
0 0
0 100 200 300 400 500 600 700 800
Figure 10-17. Proppant concentration profile comparisons for the schedules in Fig. 10-16.
ramping in curves D and D – i allow about 10% they incorporate three basic methods to control
greater penetration. The corresponding proppant sand production in the operations.
concentration profiles for the normal and four TSO The first method is a four-step operation. The
schedules are in Fig. 10-17. TSO is performed, the wellbore is cleaned up, the
• TSO with sand control application screen assembly is run, and the gravel-pack opera-
tion is performed (Monus et al., 1992).
A large number of TSO treatments are in marginally
The second method is a one-step operation that
consolidated or unconsolidated formations, and
overcomes possible damage created by the cleanup

stage of the first method. The designed fracture

250
treatment is pumped with the screen in place, the
crossover tool in the circulating position and the
Productivity ratio (%)

200
Frac and pack range
choke on the annulus closed. The annular choke is 150
opened when the gravel-pack volume is pumped at Gravel pack range
a slower rate to ensure annular packing. After 100
screenout of the casing/screen annulus, the tool is 50
placed in reverse position and the excess slurry is
reversed out of the tubing (Hannah et al., 1994; 0
–4 0 4 8 12 16 20 24 28
Meese et al., 1994).
Skin effect
The third method is a TSO with a proppant that
is stabilized with a back-production control addi- Figure 10-18. Typical productivity ratio and skin effect for
tive (e.g., curable resin, fibers or both). This conventional gravel-packed wells and wells completed with
a frac and pack treatment in the U.S. Gulf Coast.
method uses a more aggressive schedule designed
to promote backward packing to an extent such that
a continuous external pack is formed as a ring
around the annulus on the outside of the casing cre- 10-5.1. Limited entry
ated by the fracture width, which eliminates the The technique of limited-entry perforating is used to
need for an internal screen. achieve a large frictional pressure drop across certain
Executing a planned backpacking of the annulus perforations (e.g., in a given layer) in an attempt to
outside the casing creates an external pack, pro- ensure fluid injection through each perforation in the
vides maximum perforation efficiency with mini- completed interval. By limiting the number of perfora-
mal completion skin effect and should be con- tions for multiple intervals, successful application of
sidered for all three completion methods. the limited-entry technique may increase the number
The improved productivity of frac and pack of intervals that can be penetrated by the fracturing
completions over those with only gravel packs has fluid. During injection, the frictional backpressure
been shown by Mullen et al. (1994), Monus et al. offsets the stress differences between the zones to
(1992), Fletcher et al. (1984), Hannah et al. (1994), enhance injection into all the perforated zones (La-
Papinczak and Miller (1993) and Stewart et al. grone and Rasmussen, 1963). The technique does not
(1995). As for hard-rock applications, optimization consider the distribution of the injected fluid resulting
studies of proppants, fluids and breakers result in from different layer properties.
more effective frac and pack designs. The plot of During the 1960s, the injection of perforation ball
productivity ratio versus skin effect in Fig. 10-18 sealers was commonly used in conjunction with the
compares gravel-packed wells and wells completed limited-entry technique. The ball sealers were injected
with a frac and pack treatment in the U.S. Gulf at various time intervals during one treatment to
Coast area. increase the wellbore pressure even more. One limited-
entry method (Webster et al., 1965) used the differ-
ences in stress between the individual zones. The
10-5. Multilayer fracturing lowest stressed zone was fractured first at an injection
pressure below the stresses of the other intervals. A
Vertical wells usually penetrate a number of potential
specific number of ball sealers was then injected to
producing zones. In low-permeability formations, the
seal off the perforations for this zone to cause the next
zones may span hundreds of feet with no individual
to the lowest stress zone to be fractured. This method
zone capable of economic production. For maximum
found limited application because of the magnitude of
reserve recovery, it is desirable to fracture each zone
the stress differences required to confine each stage to
individually to ensure adequate penetration. To mini-
a single interval. Limited entry is discussed further in
mize fracturing costs and completion time, it is desir-
Chapter 11 and Section 10-6.4.
able to fracture all zones with one treatment or at least
to minimize the number of individual fracturing treat-
ments by efficiently grouping the layers.

10-5.2. Interval grouping modulus or higher stress. For meaningful differences

in any of the ratios, application of the limited-entry
The initial task in designing treatments for multilayer technique should be investigated, but it may not
formations is to minimize the number of fracturing achieve the desired results.
treatments by grouping the maximum number of lay-
ers that can be treated with a single fracture treatment.
The simplest case is one fracture covering all layers. 10-5.3. Single fracture across multilayers
The one-stage treatment is effective where all produc-
tive layers have a small stress contrast relative to the The simplest multilayer case is a zone with one frac-
adjoining nonproductive layers. Many high-rate, mas- ture covering all the layers, which have relatively
sive hydraulic fracturing treatments are designed for small heights and low stress contrasts with the adjoin-
this situation. ing nonproducing layers. If these intervals are in turn
If the layers are separated by higher stress, nonpro- bounded by massive shales of higher stress, the height
ductive intervals, the problem is more complex. For of a hydraulic fracture will be limited to the pay, and
example, for an application with four separate produc- a 2D geometry model may be sufficient for the
tive zones over a large interval, eight grouping options design. Nolte (1982, 1988c) showed that for height
are available. growth equal to the pay zone height into either barrier
Grouping can be made by intuitive reasoning based zone, the fracture penetration and net pressure are
on the zones’ thicknesses and their proximity to each reduced only by about 10%, and a constant-height
other and experience gained in similar conditions. PKN model should be adequate for the design.
Equation 10-2 can be used to estimate injection parti- Design and optimization of a single fracture in
tioning between zones A and B (Elbel, 1993): a homogeneous formation were previously discussed.
Production for a layered reservoir requires some
( 2 n +3 ) / ( n +1) 3 1/ ( n +1) adjustments relative to the case for a homogeneous
qiA  p f − σ h,minA   h fA   EB′   ηB 
2
=  h   E ′   η  , formation.
qiB  p f − σ h,minB   fB  A A If the layers penetrated by a single fracture have
different permeabilities or porosities, the initial pro-
(10-2)
duction response during transient time will be less
where qi is the injection rate in bbl/min, pf is the frac- than predicted using the average permeability over
ture pressure in psi, σh,min is the minimum stress in psi, the fractured interval. This behavior has been shown
hf is the fracture height in ft, E′ is the plane strain by Bennett et al. (1986), who introduced the dimen-
modulus in psi, and η is the efficiency (ratio of the sionless reservoir conductivity term CRD (see Section
fracture volume to the injected volume). 12-3.5 and Eq. 12-44) and showed that the Agarwal
The equation shows that the injection rate partition- et al. (1979) type curves can be used to evaluate the
ing between the two zones A and B is governed by transient response in a layered reservoir if the tD term
four zone parameter ratios of the gross height, net is replaced by tD/CRD2. Without this substitution, the
pressure, modulus and efficiency. The difference in effective fracture length is in error by a factor equal
the injection rates occurs primarily because of differ- to CRD, and the value of CRD is always less than 1 for
ent fracture widths caused by a height- and modulus- a layered reservoir. As a consequence, the response
dependent stiffness effect. The net pressure and from the fracture treatment, based on the physical
efficiency are also dependent on the height and modu- fracture length and average reservoir properties, will
lus; therefore, calculating the actual ratio for design be less than predicted. Economic optimizations for
requires an iterative procedure. It is evident that a fracture length should take this layer effect into
zone with a small height should generally not be account; e.g., if CRD is 0.5, the optimization should be
grouped with a much larger height zone. An exception based on production during transient flow for a frac-
would be where the formation is sufficiently charac- ture penetration one-half of that predicted by the
terized to ensure the existence of the unlikely condi- geometry simulation.
tion that the thicker zone has a correspondingly higher

10-5.4. Two fractures in a 657 ft for the top zone and only 325 ft for the bottom
multilayer reservoir zone. The injection rate in the top zone is 9.6 bbl/min
and 2.4 bbl/min in the bottom zone. The bottom zone
Two perforated intervals can have different fracture screens out at about 80% of the job time and the top
geometries as illustrated by Eq. 10-2. Elbel (1993) layer at 95%. The simulation continues as a TSO in
discussed the considerations required for a design in both zones.
this setting and pointed out the difficulty in achieving The model assumes constant height of the fractures
equal proppant penetration in each layer because of and, as previously discussed, is valid for cases with
different pad volume and perforation frictional pressure high stress contrasts that limit growth into a barrier to
drop requirements. The coalescence of two individual a distance of about the pay section. P3D simulations
fractures, originating at different perforated intervals, for the individual zones at their partitioned rates can
would generally not occur unless at the unlikely condi- be made to determine height growth. Figure 10-19
tion of near-perfect alignment for the wellbore axis and shows the results of this exercise and indicates that for
preferred fracture planes. Also, for the case of the two a mid-barrier stress contrast of 1100 psi, fractures
fracture heights extending and overlapping each other, begin to overlap near the wellbore after about 80%
the effect of the width of each fracture would increase of the treatment time. However, even if the fractures
the surrounding stress and impede additional height have coalesced, the mid-barrier stress should result in
growth into the overlap region. a pinch point for fracture width and proppant bridging
Consider a case with two perforated intervals with would prevent vertical fluid flow communication.
the same properties except that the heights are 80 and An additional P3D simulation in Fig. 10-19 for a
40 ft for the top and bottom layers, respectively; they mid-zone stress contrast of only 600 psi shows fracture
are separated by a 40-ft interval at higher stress. overlap for the two zones occurring earlier, after about
Because both intervals are perforated and have the 30% of the treatment time and before the first proppant
same stress, fractures will be initiated in both layers. stage. The similar penetrations from the idealized mul-
A common practice for this case is to design a tilayer model and for the multiple simulations with a
pumping schedule for a single fracture over the three P3D model for each of the stress cases demonstrate
intervals. The consequences of ignoring the different the utility of the multilayer model. This model pro-
stresses for the intervals, for the single-fracture vides the generalization of Eq. 10-2 for multilayer
assumption, are addressed in the following for a treatment design from two to an essentially unlimited
design using an injection rate of 12 bbl/min. Other number of layers. The model enables consideration
design details are not critical for this purpose and are of the different penetrations that result because of dif-
not provided. The design is to achieve a penetration ferent layer properties. These effects have also been
of 560 ft. First, consider the single-fracture design demonstrated by Rahim and Holditch (1992).
within the constant-height multilayer fracturing model The significance of different penetrations is illus-
described in Chapter 6. This results in penetrations of trated by production forecasts made at a constant well-
9400 2000 9400

Multizone fracture penetration
9500 9500
9.6 bbl/min, 657-ft penetration 1500
Net pressure (psi)
Fracture height (ft)
9600 9600
Well depth (ft)
2.4 bbl/min, 325-ft penetration

9700 1000 9700
Bottom
9800 9800
Top 500
9900 9900
0
0 20 40 60 80 100 0 200 400 600 800 1000
Treatment time (min) Fracture half-length (ft)
Figure 10-19. P3D simulations for individual zones with 2D partitioned rates.

head pressure of 300 psi for a single fracture and multi- mum recoverable. Comparison of the 10-year cumula-
fracture cases (Table 10-2). The four combinations tive production from the design with the maximum
of permeability listed all have a permeability-thickness recoverable in the column labeled percent of maxi-
kh of 3.6 md-ft. Cases A and B have 0.03-md permea- mum provides another perspective on design and indi-
bility in both zones and cases C and D have different cates if a different completion or spacing should be
permeability combinations to obtain the same average considered. Smaller spacing with more wells or rec-
permeability-thickness. tangular spacing with the long axis parallel to the frac-
The 10-year cumulative production is listed in ture azimuth, if known, would improve the percent
Table 10-2 for the individual zones and their com- recovery (Holditch et al., 1978; Elbel, 1986; Meehan
bined production. Case A has the highest production, et al., 1988).
~12,000 MMscf, and case C has the lowest at Application of the limited-entry technique can be
~9,800 MMscf, or 82% of that for case A. Cases B considered for the two-layer cases B, C and D in
and D each have about 95% percent of the production Table 10-2. If the density of the open perforations is
of case A. Table 10-2 also lists the percent of gas in 1 shot per foot (spf) in each layer, the top 80-ft zone
place recovered from the individual and combined will take most of the fluid, and the number of perfora-
zones. The 40-ft zone in case D, which has the lower tions should be limited in that zone. Reducing the
permeability of 0.01 md and a shorter fracture length, number of perforations from 80 to 14 would reduce
has only 7.9% recovery of the gas in place. In case C, the initial injection rate from 9.5 to 7.5 bbl/min with
the same zone simulated with a higher permeability of an initial perforation friction pressure of 100 psi in
0.07 md recovered 31% of the gas in place. These this layer. The perforation friction for the bottom
results show the role of multilayer modeling in deter- interval remains negligible, and the number of perfo-
mining the benefits for two separate treatments that rations does not require changing. The penetration is
can individually optimize the recovery of each zone. 550 ft into each layer. The net pressure would be
The maximum recovery possible over a specific about 100 psi higher in the bottom interval, which
time period can be determined by the simulation of could increase its fracture height growth depending
production from a fracture of infinite-acting conduc- on the stress in its barriers.
tivity penetrating to the drainage boundary. Production The previous discussions assume that the perforated
from a fracture penetrating 1500 ft, or 80% of the zone fractures when the wellbore pressure is above its
assumed 320-acre spacing, was simulated for all minimum stress. This generally does not occur with-
zones in cases A through D and listed as the maxi- out pretreatment injection procedures using acid, per-
Table 10-2. Ten-year production forecasts for a single fracture (A) and multifracture cases (B–D).
Remaining Gas in Place

Case xf Net Perme- 10-yr Gas-in- Maximum Percent of Difference
(ft) Height ability Cumulative Place Recoverable Maximum Design – Design Maximum
(ft) (md) (MMscf) Recovery (MMscf) Maximum (MMscf) (MMscf)
(%) (MMscf)
A 560 120 0.03 11,975 22.4 22,134 54.1 –10,159 41,485 31,326
B 660 80 0.03 8,596 24.1 14,819 58.0 –6,223 27,072 20,849
330 40 0.03 3,118 17.5 7,440 41.9 –4,322 14,720 10,398
Commingled 11,713 21.9 22,259 52.6 –10,546 41,795 31,249
C 660 80 0.01 4,273 12.0 8,239 51.9 –3,966 31,395 27,429
330 40 0.07 5,531 31.0 9,980 55.4 –4,449 12,311 7,862
Commingled 9,804 18.3 18,219 53.8 –8,415 43,770 35,355
D 660 80 0.04 10,409 29.2 16,718 62.3 –6,309 25,238 18,929
330 40 0.01 1,411 7.9 4,132 34.1 –2,721 16,450 13,729
Commingled 11,820 22.1 20,850 56.7 –9,030 41,664 32,634
Note: All cases have a constant wellhead pressure of 300 psi and a combined kh of 3.6 md-ft.

foration ball sealers or both to mitigate the stress con- 10-5.5. Field example
centration around the wellbore. The effect of the stress
concentration is enhanced when the two horizontal This section reviews a study by Morales et al. (1995)
stresses are similar. Plugged perforations and wellbore of multiple fractures in reservoirs with shales inter-
damage can prohibit the breakdown fluid from pene- bedded in soft to unconsolidated sands. The review
trating into the formation and beyond the stress con- includes some of the features of multilayer fracturing.
centration. These restrictions increase the breakdown Fracture evaluation in multilayer zones is discussed
pressure in low-permeability zones. However, unre- in Sidebar 10F. In one example, Morales et al. showed
stricted fluid penetration into high-permeability zones that stress contrasts as small as 200 psi can result in
that allow large-scale pore pressure increases raises the simultaneous propagation of essentially indepen-
the breakdown pressure as a result of poroelastic- dent fractures, particularly where the shales have
induced stress increases. Therefore, if these conditions higher moduli than the softer sands. They also showed
affecting breakdown are different from zone to zone, how calibration treatment pressure decline and deriva-
the magnitude of the initiation pressures over that of tive plots can indicate more than one closure pressure
the minimum stress will be different. Once all the ini- (approximately 5600 and 5200 psi; Fig. 10-20); vali-
tiation pressures are reached, the zonal rate partition- dation was conducted by matching the pressure
ing discussed in this section should apply. A more decline with a simulation from a multilayer model
detailed presentation of the stress concentration and (see Chapter 6).
breakdown pressure is provided in Section 3-5.7.
10F. Fracture evaluation in multilayer zones
Bennett et al.’s (1986) work concerning the transient behavior of a fracture in a layered reservoir introduced the dimensionless reser-
voir conductivity term CRD (see Eq. 12-44) to enable modeling the correct multilayer transient response with an equivalent single-
layer model. Camacho-V. et al. (1987) expanded this work to include layers with unequal fracture lengths and conductivities. They
showed that the equivalent single-layer fracture penetration Lapp is the sum of the products of each layer’s CRD times the layer’s frac-
ture penetration L:
n
Lapp = ∑C
j =1
RD , j Lj . (10F-1)
An example for this relation was provided by Elbel and Ayoub (1992) for transient well tests in a well with two zones. The zones
were treated separately and designed to achieve 400- and 689-ft penetrations in zones 1 and 2, respectively. After the treatment
cleanup period, a well test was performed and considered by a single-zone type curve. The analysis indicated an average perme-
ability of 0.76 md and a fracture penetration of 230 ft, considerably less than those of the design lengths.
Cores from these zones in other wells in the area have a thin high-permeability layer in most zones and a wide range of very low
permeability. The interpretation was therefore modified by characterizing the two zones as having three layers with different heights,
permeabilities and porosities based on the logs and cores. The zones and layer groupings are shown in Table 10F-1. The CRD value
for each layer was calculated using Eq. 12-44. Using the designed fracture penetrations of 400 ft in zone 1 and 689 ft in zone 2 in
Eq. 10F-1 showed that the transient behavior would be similar to an apparent single-layer fracture penetration of 223 ft in a forma-
tion with an average permeability of 0.76 md. That the single-layer transient analysis gave a result of 230 ft indicates a probability
that the designs were achieved, although the apparent value is considerably less than the penetration in either zone.
This example makes several important points:
• The postfracturing reservoir
response is inherently Table 10F-1. Effect of different fracture penetrations in
nonunique in the absence of a two multilayer zones (Elbel and Ayoub, 1992).
comprehensive reservoir char-
acterization. Zone 1 Zone 2 Average
• Validating fracture placement
designs from well test or pro- Height (ft) 49.2 8.9 1.0 39.3 38.4 1.0
duction data is difficult. Permeability (md) 0.01 0.20 80.00 0.02 0.10 20.00 0.76
• The reservoir inference of frac-
ture length xf and the physical Porosity 0.05 0.08 0.14 0.05 0.08 0.17 0.06
fracture length L should be iden- CRD 0.04 0.04 0.11 0.04 0.11 0.06
tified by different notations, as
used in this volume. xf (ft) 400 400 400 689 689 689

smaller zone and its lower leakoff rate. The rates of the
7000 0.5
First closure Second closure
other two layers changed during injection because of
6500
0.4 their relative fluid loss and stiffness. During closure,
Pressure (psi)
crossflow occurred because of the different zone prop-
Derivative
0.3
6000 erties; for the zones with dominating fractures, layer 1
Derivative 0.2 gave up fluid to layer 3 until layer 3 closed at its higher
5500
0.1
stress.
The high net pressure of 1300 psi in the injection
5000 0 matched the multilayer model, whereas simulation
0 5 10
of a single fracture over the entire interval estimated
Closure time (min)
a net pressure of only 200 psi. The high net pressure
Figure 10-20. Treatment pressure decline and derivative could be matched by increasing either the modulus
showing the simulated pressure decline match (Morales of the fracture by sixfold, which is an unusually large
et al., 1995).
value, or the toughness by an even larger factor.
(Toughness in calibration treatments is discussed in
For this field example, the logs in Fig. 10-21a show Chapter 5.) The properties used in the simulation are
three distinct gas-bearing zones (layers 1, 2 and 3 are listed in Table 10-3.
10, 6 and 14 ft thick, respectively) separated by 22 and The multilayer simulation of the rate distribution for
20 ft of shale. Simulations of the calibration treatment the actual fracture treatment is shown in Fig. 10-23.
using the multilayer model provide the layer rate distri- For these rate distributions, the fracture lengths and
butions shown in Fig. 10-22. The pressure and closure heights obtained using a P3D simulator are shown in
time matches were made with fluid-loss coefficients Fig. 10-21b. The barriers were assumed to have 0.1-
of 0.015 ft/min1/2 for layer 3 and 0.0006 ft/min1/2 for psi/ft greater gradients than the pay zones. Because of
layers 1 and 2. Very little fluid entered the 6-ft thick height growth, the fracture lengths are less than those
layer 2 because of the relatively higher stiffness of the predicted by the multilayer model.
(a) (b)
7600 7600
1
Depth (ft)
7650 7650 2
7700 7700
–45 30 0.1 1 10 0 50 100 150 200
SP (mV) Focused resistivity xf (ft)
(ohm-m)
Figure 10-21. Logs of 10-, 6- and 14-ft gas zones and P3D simulation of the fractures (Morales et al., 1995).

15 12
Screenout
10
1
3
10 3
8
Shut-in
Rate (bbl/min)
Rate (bbl/min)
2 6
5
1 3 4
1
0
2
2
0 2
–5
1
–2
–10 –4
0 5 10 15 0 10 20 30
Figure 10-22. Multilayer fracture (MLF) PKN simulation Figure 10-23. Simulated rate distribution and layer screen-
showing the layer rate distributions during calibration injec- out for the actual fracture treatment (Morales et al., 1995).
tion (Morales et al., 1995).
Table 10-3. Properties for multilayer 10-6. Acid fracturing

simulation example.
Hydrochloric acid (HCl) inhibited to prevent the cor-
Property Value rosion of wellbore tubulars was first used in 1932 by
Reservoir pressure 4700 psi the Pure Oil Company to stimulate a limestone forma-
tion. The previously dead well responded by produc-
Height
ing 16 BOPD. Acidizing rapidly became the preferred
Layer 1 10 ft
stimulation method for carbonate reservoirs and began
Layer 2 6 ft to replace explosive stimulation of openhole comple-
Layer 3 14 ft tions. Grebe and Stoesser (1935) observed that during
Fluid-loss coefficient acid injection, the formation “lifting pressure” was
sometimes obtained, indicating that the formation was
Layer 1 0.0006 ft2/min
also being fractured. This was the first description of
Layer 2 0.0006 ft2/min
hydraulic fracturing applied to petroleum reservoirs.
Layer 3 0.015 ft2/min During the late 1940s, the process of creating a frac-
Young’s modulus 2,200,000 psi ture by the injection of oil and propping the fracture
Poisson’s ratio 0.25
with sand was developed for stimulating sandstone
formations. The primary difference between acid and
Closure pressure propped fracturing is the means of achieving fracture
Layer 1 5200 psi conductivity after the fracture closes: an etched pat-
Layer 2 5200 psi tern of voids on the fracture faces and propping the
Layer 3 5500 psi faces apart, respectively.
Advances in hydraulic propped fracturing in the
Fracture toughness 2000 psi-in.1/2
1970s, particularly the ability to model the fracture
geometry and proppant placement, resulted in the suc-
cessful use of propped fracturing for stimulating car-
bonate formations. The use of acid fracturing also
decreased because modeling of the etched acid geom-
etry was more complex and less developed. However,

advances in acid fracturing modeling (Settari, 1991; measuring and modeling acid-etched fracture conduc-
Mack and Elbel, 1993) and determining reaction para- tivity.
meters (Li et al., 1993; de Rozières, 1994) began dur- Most geometry models incorporate Nierode and
ing the 1990s. These developments support more Kruk’s correlation with various corrections for fluid
reliable treatment designs and form a basis for choos- loss and field-scale inhomogeneity. Beg et al. (1996)
ing between acid fracturing and propped fracturing reported the same effect of rock embedment strength
stimulation of carbonates. Section 6-9 addresses vari- and closure stress on conductivity. Large volumes of
ous aspects of modeling acid fracturing. acid and the dissolution of large volumes from the
Operationally, acid fracturing is less complicated fracture faces result in negative net fracturing pres-
because no propping agents are used, which elimi- sures in modeling. Laboratory tests would be more
nates the risk of a screenout and subsequent problems meaningful if they could be run at appropriate pre-
of proppant flowback and cleanout from the wellbore. stressed conditions.
However, the effective length of an acidized fracture Increased etched conductivity resulting from frac-
is limited by the distance the acid travels along the ture-face heterogeneity was reported by Van Domelen
fracture before spending. At high temperatures this (1992). Beg et al. (1996) reported that tests with fluid
limit is a greater problem. Another barrier to effective loss typically result in higher fracture conductivity
acid penetration is excessive fluid loss (Nierode and than tests on similar cores without fluid loss. Fluid
Kruk, 1973). Continuous acid corrosion and erosion loss enhances surface reactions by removing the by-
of the fracture faces during treatment make it difficult products of the acid reaction. They also observed, in
to deposit an effective filter-cake barrier. In addition, repeated experiments, significantly higher conductiv-
acid leakoff is highly nonuniform and typically results ity when acid forms a deep channel.
in wormholes and the enlargement of natural fractures The difficulty in accurately modeling acid fracture
and matrix permeability. These natural consequences conductivity is apparent. Conductivity can be cali-
of acid treatments increase the effective area and the brated with well test data after a treatment for use
volume of leakoff. Acid fluid loss and its control are in subsequent treatment simulations. After analyzing
discussed in greater detail later in this chapter. field production data, Aud et al. (1992) and Elbel
(1993) suggested increasing Nierode and Kruk’s cor-
relation by 1 order of magnitude.
10-6.1. Acid-etched fracture conductivity Fortunately, except for soft formations or extremely
The factors controlling the effectiveness of acid frac- large closure stresses, the conductivity obtained from
turing are the etched fracture penetration and conduc- fracture acidizing is usually excessive because of the
tivity. Laboratory measurements of acid-etched high reactivity of limestone and the relatively large
fracture conductivity are not always reproducible and volumes of acid required to achieve the desired pene-
may not represent actual conditions because of the tration. Dimensionless fracture conductivity values
small size of the laboratory samples. The normal vari- greater than 30 are common. This makes errors in con-
ation of sedimentary layers, over the larger scale of ductivity of little consequence for production forecasts.
the fracture face, provides local preferred paths for However, conductivity predictions are important for
acid reaction and subsequent channeling of the fol- optimizing acid volume for dolomite formations and
lowing acid stages. This situation makes it difficult to particularly for temperatures less than 120°F [50°C],
properly model conductivity (see Chapter 6 for more for which acid reaction rates are relatively small.
discussion). In an attempt to make conductivity pre- A technique, described as equilibrium acidizing,
dictions, Nierode and Kruk (1973) used empirical lab- was reported by Tinker (1991) to be successful for
oratory data and developed a conservative method of enhancing conductivity in low-temperature dolomites.
calculating fracture conductivity (Williams et al., With this technique, a fracture of the desired penetra-
1979). Their predictions are based on correcting the tion is created by a volume of acid and the injection
theoretically ideal conductivity, based on the etched rate is reduced to maintain equilibrium with the acid
width cubed (Eq. 6-130), for the effect of fracture clo- leakoff rate or, more practically, to a rate that achieves
sure stress and rock embedment strength (similar to a fracturing pressure between the extension and for-
Brinnell hardness). Sidebar 10G further discusses mation closure pressures. The lowered rate also mini-
mizes vertical growth into water-producing zones

10G. Acid-etched conductivity
Measurements of acid-etched conductivity in the laboratory are usually not reproducible or representative of large-scale in-situ
behavior because of heterogeneities in the rock and the small size of laboratory samples. This makes it difficult to calculate and vali-
date acid-etched fracture conductivity; however, Nierode and Kruk (1973) developed an empirical equation based on closure stress
and the reservoir’s embedment strength from laboratory tests on formation cores (see Section 6-9.3). The amount of formation dis-
solved is used to calculate a uniform (ideal) etched width. The ideal conductivity for a uniform open fracture with a width of 0.1 in. is
about 4.5E+6 md-ft. The empirical correlation for zero closure stress reduces this by almost 3 orders of magnitude to about 6.5E+3
md-ft. Further reductions are made as closure stress and embedment strength effects are considered. This correlation is a conserva-
tive (i.e., lower bound) estimate, and calibration with other tests was recommended by Nierode and Kruk.
Figure 10G-1a is a plot of Nierode and Kruk’s conductivity calculations for average etched widths of 0.1 and 0.025 in. The undam-
aged conductivities of 20/40-mesh ISP and sand proppants at a concentration of 1.5 lbm/ft2 are shown for comparison. These com-
parisons have supported the preference of propped fracturing over acid fracturing in some areas.
The difficulty in accurately modeling acid fracture conductivity is apparent. Calculated etched conductivity has been calibrated with
well test data after treatments for use in subsequent treatment simulations. After analyzing field production data, Aud et al. (1992)
and Elbel (1993) suggested increasing the Nierode and Kruk correlation by 1 order of magnitude. Figure 10G-1b shows conductivi-
ties using this correction and a comparison with the conductivity of proppant corrected for 50% polymer damage (see Chapter 8).
The comparison indicates that the correct choice between acid and propped fracturing should not be based on the assumption of low
and etched conductivity; in fact, acid fracturing of competent carbonate formations should always be considered as the most
prospective treatment (see Section 10-6.7).
(a) Nierode and Kruk's correlation (b) 10× Nierode and Kruk's correlation
and undamaged proppant pack and 50% retained propped permeability
10,000,000
Ideal kfw (0.1 in.)

Acid (0.1 in., 200,000-psi embedment)
1,000,000 Ideal kfw (0.025 in.)
20/40 ISP (1.5 lbm/ft2)
20/40 sand (1.5 lbm/ft2)
100,000
Conductivity (md-ft)
10,000
1000
1
0 1500 3000 4500 6000 8000 10,000 0 1500 3000 4500 6000 8000 10,000
Effective stress (psi) Effective stress (psi)
Figure 10G-1. Comparison of etched acid and propped fracture conductivity (Nierode and Kruk, 1973).

while creating more conductivity from additional vol-

800
umes of acid. 30 bbl/min, 0.5 md
A similar technique allows the initial etched frac- 15 bbl/min, 0.5 md

30 bbl/min, 2.0 md
600
ture to close while injection continues at a pressure 15 bbl/min, 2.0 md
below the closure pressure. The continued injection

400
promotes etched-channel development for increased
conductivity (Sizer et al., 1991). This technique most
likely has limits for etched penetration (Tinker, 1991). 200
Using a viscous pad and limiting fluid entry points

by selected perforating was reported by Davies et al. 0
0 600 1200 1800 2400 3000
(1987). This technique should promote channeling of Reservoir pressure (psia)
the acid through the viscous pad, localizing the acid
reaction to enhance etched conductivity. Figure 10-24. Effects of injection rate for permeability and
reservoir pressure spanning the bubblepoint on etched
fracture penetration (Aud et al., 1992).
10-6.2. Acid fluid loss
Excessive fluid loss is generally considered to be the explained the selective enlargement of certain large
factor that limits fracture extension and etch penetra- pores present in limestone. Schechter and Gidley
tion for acid fracturing carbonate formations. The (1969) later addressed this same problem in greater
parameters that control fluid loss are formation perme- detail. Nierode and Kruk (1973) evaluated the effect
ability and porosity, reservoir fluid compressibility, of various acid fluid-loss additives during wormhole
leakoff filtrate viscosity and the differential pressure development. Crowe et al. (1989) showed that worm-
between the fracture and reservoir (see Chapter 6). Oil hole depth is limited by spending of the acid along the
reservoirs, above the bubblepoint pressure and free of wormhole channel (Fig. 10-25). The missing, critical
gas, can have a sufficiently low compressibility to parameter for effectively modeling wormholes is the
control the leakoff of fluids that would otherwise have areal density of the wormholes that penetrate the frac-
a high fluid loss. ture face.
The role of free gas was reported by Aud et al. Various additives and treating techniques have been
(1992) in a study of acid refracturing. The study used developed to control acid fluid loss. Except in the pad
reservoir modeling to indicate relatively long fractures stages, acid fluid-loss additives with particulate mater-
from treatments performed prior to reservoir deple- ial have not been used extensively because of poor
tion. However, they showed that the same treatments performance and cost limitations. Various strategies
performed after the reservoir was below bubblepoint for fluid-loss control are outlined as follows.
resulted in significantly less etched penetration. They
concluded that the low compressibility of the reservoir 30
fluid above the bubblepoint pressure effectively con-
trolled fluid loss by compressibility-controlled fluid-
loss behavior. Subsequent steps to control the fluid
20
Penetration (cm)
loss with viscous pads and emulsified acid, which

were not required when the reservoir was above bub-
blepoint pressure, resulted in greater etched penetra-
tions and increased production. The simulations from 10
Aud et al. in Fig. 10-24 show the change in penetra-
tion above and below bubblepoint pressure for various
injection rates and formation permeabilities.
Fluid loss occurs in a highly selective manner, 0
0 10 20 30 40 50
which can create wormholes and enlarge natural frac- 15% HCl (mL)
tures during acid treatment. This phenomenon was
first described by Rowan (1957), who mathematically Figure 10-25. Wormhole growth rate during leakoff of 15%
HCl at 0.065 ft/min and 150°F [65°C] (Crowe et al., 1989).

• Pad stages for fluid-loss control

10H. Fluid-loss control in wormholes
The use of a water-base or gelled-water viscous
pad preceding the acid stage is commonly used to The creation of wormholes by acid fluid loss during acid frac-
turing can cause excessive leakoff and limit the etched pene-
initiate the fracture and deposit a filter cake that can tration. Alternating the acid and pad stages is a technique
act as a barrier to acid leakoff. The actual ability of used to minimize fluid loss in this scenario (Coulter et al., 1976).
the pad fluid to control fluid loss in this manner is The effectiveness of this technique has been demonstrated
in laboratory tests (Fig. 10H-1). The core on the left illus-
questionable. Studies by Nierode and Kruk (1973), trates the results of uncontrolled acid fluid loss in which
Coulter et al. (1976) and Crowe et al. (1989) show channeling occurred completely through the core in about
15 s. The center core was first treated with a pad composed
that the filter cake deposited by the pad is quickly of water gelled with guar polymer and then treated with acid.
penetrated by wormholes resulting from acid leak- Channeling occurred after only 20 s. In the core on the right,
wormholes were initially created by injecting a limited amount
off. Once this occurs, acid fluid loss can be the of acid, followed by the pad fluid and then by more acid. In
same as if no pad were used. this test, channeling by the second acid stage was delayed
for 15 to 25 min, demonstrating the effectiveness of the tech-
Multiple stages of gelled pads have been used to nique for controlling acid fluid loss. This acidizing procedure
control acid fluid loss (Coulter et al., 1976). In this of multiple alternating stages of acid and pad is widely used
technique, the fracture is initially created by a pad, in acid fracturing treatments.
after which alternating stages of acid and pad are

pumped. These additional pad stages enter and seal Filter
cake
wormholes created by the preceding acid. The
sequential pad and acid stages are effective for con-
trolling the leakoff of acid into wormholes and
enlarged natural fractures. Sidebar 10H provides
additional discussion of fluid-loss control in worm-
holes.
The pad has other useful functions. Cooling
effects decrease acid reaction within the tubulars and
on the formation face, whereas the relatively large
Acidized core Gel followed Acid followed by
viscosity increases width, which can improve etched (channels in by acid gel followed by acid
penetration. A wider fracture reduces the areal reac- 15 s) (channels in (channels
20 s) in 15–25 min)
tion rate by reducing the surface area relative to the
transported acid volume. Viscous pads also promote
Figure 10H-1. Laboratory core tests show how a gel-
acid fingering for improved conductivity. filled channel (right) resists leakoff whereas simple
For a treatment design, the optimum volumes for filter cake has little effect (Coulter et al., 1976).
each stage of the pad and acid sequence should be
determined by iterative simulations using a com-
prehensive acid fracturing model (see Section 6-9). nated. Fine salt can be used for acid fracturing
• Particulate materials for fluid-loss control water injection wells (Schriefer and Shaw, 1978).
During such treatments, sufficient salt is added to
Fine particulate material is typically added to the
saturate the acid solution and to provide an excess
pad stages to aid in fluid-loss control. The particu-
of particulate material at the bottomhole treating
late material fills and bridges wormholes and natural
temperature. Following the treatment, any remain-
fractures and improves fluid efficiency. The most
ing salt is dissolved by water injection.
common material used is 100-mesh sand (Coulter
et al., 1976), which is usually added at a concentra- • Gelled acid for fluid-loss control
tion of 1 to 3 lbm/gal. Oil-soluble resins for oil Acid fluid loss also can be reduced by gelling the
reservoirs and salt of a similar particle size are also acid. The advantages, limitations and disadvantages
used. Although much more expensive, oil-soluble of this method are discussed in Section 7-4.1.
resins eliminate the possibility of conductivity • Acid emulsion systems for fluid-loss control
impairment resulting from the 100-mesh sand in
Acid emulsions, both acid-external-phase and acid-
the fracture. Potential problems resulting from the
internal-phase systems, have been used in acid
return flow of sand into the wellbore also are elimi-
fracturing. The leakoff is controlled by the high

viscosity of the emulsion. Excessive fluid friction

350
pressure for emulsions limits their application. The qi = 10 bbl/min
oil-external-phase system is used because it isolates h =
–
50 ft
vL = 0.0005 ft/min
the acid and produces excellent retardation proper- 300 w= 0.1 in.
ties (see Section 10-6.3) compared with regular
acid. This system is also more efficient than the
250 28
acid-external phase because it increases the relative
Acid penetration distance (ft)

%
HC
l
amount of acid in the fluid system.
• Foamed acid systems for fluid-loss control 200
15
%
HC
The use of foamed acid is an effective method for l
controlling acid fluid loss (Scherubel and Crowe, 150 Dolomite

28% HCl
1978; Ford, 1981). Fluid-loss control is further
enhanced by the use of a viscous pad preceding the 100 15% HCl
Limestone
foamed acid. However, foaming the acid reduces
the effective amount of acid available for etching
because less acid is present per unit volume injected. 50
As a result, 28% HCl should be used in preparing
the foamed acid to maximize the amount of acid 0
available for fracture etching. The relatively unique 100 120 140 160 180 200 220
non-Newtonian flow properties obtained by foam- Temperature (°F)
ing or emulsifying acid can promote fingering of
the fluid during fracturing to enhance channels on Figure 10-26. Penetration of 15% and 28% HCl in
dolomite and limestone. qi = injection rate, h = height,
the fracture face. —
v L = average fluid-loss velocity, w = fracture width.
10-6.3. Acid reaction rate Differences in mineralogy can affect the reaction
Acid reaction rates are defined by laboratory tests and properties of HCl for different formations. Also, dif-
are usually performed on pure limestone or dolomite. ferent acid systems, such as gelled and emulsified sys-
The effect of temperature in accelerating the reaction tems, may produce different effects even on the same
rate affects the depth of live acid penetration. The con- formation. These effects further emphasize the role of
centration has a similar affect. Figure 10-26 shows that representative and formation-specific laboratory tests.
15% and 28% HCl at 100°F [40°C] produce more than • Effect of acid emulsion systems on reaction rate
twice the penetration in a dolomite than in a limestone.
Oil-external-phase emulsions are the most common
Increasing the temperature to 220°F [105°C] reduces
because the external oil phase physically separates
the penetration distance in a limestone about 10% but
the acid from the reactive carbonate surface. The
reduces penetration 50% in dolomite. This is because
acid reaction rate can be further slowed by surfac-
the acid reaction rate on limestone is mass-transfer lim-
tant retarders that cause carbonate surfaces to
ited whereas the acid reaction rate on dolomite is lim-
become oil-wet. The combination of emulsification
ited by reaction kinetics that approaches the rate on
and surfactant-induced oil-wetting reduces the acid
limestone only at high temperatures. These compar-
reaction rate to an extremely low level. These
isons clearly indicate the need for representative labora-
emulsion systems produce good retardation under
tory tests and model capabilities to effectively design
static and flowing conditions (de Rozières, 1994).
acid treatments.
The effective diffusion rate of the acid spending at
Retarded acids should always be evaluated under
150°F [65°C] is 1.3E–5 ft2/min for HCl and only
pressure, temperature and flow conditions simulating
3E–9 ft2/min for emulsified acid.
those existing in the formation and including the effect
of fluid loss to remove by-products and transfer unre- • Effect of gelled acid systems on reaction rate
acted acid to the surface. Static reaction rates measured Gelled acids are used in acid fracturing treatments
under atmospheric pressure are practically meaningless. and are usually considered to be retarded because

of the role of the viscosity in reducing the mass tration distances on the basis of considerations of sur-
transfer of acid at the fracture wall. However, face kinetics, flow and temperature conditions in the
extensive laboratory results by Crowe et al. (1989) fracture and fluid loss into the fracture faces. The mod-
indicate that this role for viscosity is most likely els are useful in performing sensitivity studies on such
not significant. As discussed previously, the addi- parameters as rate and pad volume for various acid sys-
tional fracture width resulting from higher viscosity tems to improve the design for treatment optimization.
also increases the penetration of live acid.
The effect of retardation (i.e., diffusivity effect)
on the etched conductivity, penetration and produc- 10-6.5. Parameter sensitivity
tion was compared by de Rozières (1994) for neat, A number of options are available for achieving
gelled and emulsified acids (Fig. 10-27). The same deeper penetration of an etched fracture: increasing
treatment was simulated with only the laboratory- the acid volume, modifying the injection rate, mod-
determined diffusion coefficients differing for the ifying the acid concentration, including a pad fluid
neat, gelled and emulsified acids. The penetration, and applying retardation methods. The sensitivity of
with an etched fracture greater than 1000 md-ft, is penetration to these parameters individually and in
1100 ft for the emulsified acid, 800 ft for the gelled combination is shown in Figs. 10-28 through 10-31
acid and 600 ft for the neat acid system. The results of etched conductivity profiles for various treatment
also confirm the observation made in Section 10.6-1 schedules. Most of the simulations are for an oil well
that conductivity may be sacrificed to create addi- with the reservoir pressure above the bubblepoint. As
tional penetration. The simulated cumulative pro- discussed in Section 10-6.2, this condition has a sig-
duction listed in Fig. 10-27 shows that the highest nificant effect on controlling fluid loss. Some simula-
production corresponds with the greatest penetration. tions are made with the reservoir pressure below the
bubblepoint for fluid-loss comparison. The other para-
meters are listed in Table 10-4.
10-6.4. Acid fracturing models
• Acid volume and concentration
There are various mathematical models for predicting
the results of acid fracturing treatments (see Section 6-9). Increasing the volume of acid increases the etched
The models are designed to predict effective acid pene- penetration, depending on the fluid loss and reac-
100,000
Diffusion (cm2/s) Cumulative oil production (10 yr)
Neat acid 2.13E–5 76,672 STB
Gelled acid 7.92E–6 83,951 STB
Emulsified acid 4.29E–8 91,953 STB
Conductivity (md-ft)
10,000
1000
0 100 200 300 400 500 600 700 800 900 1000 1100
Length (ft)
Figure 10-27. Effect of retardation on the etched conductivity and penetration of neat, gelled and emulsified acids
(de Rozières, 1994).

Table 10-4. Reservoir and fluid properties for acid

100,000
and propped fracture simulations. 10 bbl/min
20,000-gal 15% HCl
20,000-gal 28% HCl
Conductivity, kfw (md-ft)

Property Value 10,000
10,000-gal 15% HCl
Well type Oil
10,000-gal 28% HCl
Bottomhole static pressure 5000 psi 1000
5000-gal
15% HCl
Bottomhole static temperature 150°F [65°C]
100
Fluid surface temperature 70°F [20°C]
5000-gal 28% HCl
Bubblepoint pressure 3000 psi

0
Reservoir fluid compressibility 6.8E–6 psi–1 0 100 200 300
Reservoir fluid viscosity 1.2 cp
Zone height 50 ft Figure 10-28. Etched conductivity profiles for various vol-
umes of 15% and 28% HCl.
Embedment strength 60,000 psi
Young’s modulus 5,784,000 psi

from fracture acidizing. Increasing the acid concen-
Porosity 20% tration from 15% to 28% while maintaining the
Permeability 1 md
same volume of acid increases the etched penetra-
tion another 20% to 25%. These simulations
Fluid rheology assume that the fluid-loss rate is not increased
Linear gel 20 cp by the increased acid volumes.
Crosslinked gel 200 cp The dissolving power of 28% HCl is about twice
15% HCl 2 cp that of 15% HCl, and therefore 5,000 gal of 28%
28% HCl 2 cp HCl should dissolve about the same amount of car-
bonate as 10,000 gal of 15% HCl. The comparison
HCl reactivity
of 10,000 and 20,000 gal of 15% HCl with 5,000
Effective diffusivity 6.46E–6 ft2/min and 10,000 gal of 28% HCl in Fig. 10-28 shows
Rate constant 5.4E–6 gmol/cm2 that the larger volumes of 15% acid achieve about
Rate order 0.5 the same etched penetration as the lower volumes
Activation energy 22,600 J/gmol of 28% HCl. However, the average conductivity
obtained with the larger volumes of 15% HCl is
Heat of reaction 40,000 J/gmol
about twice that of the compared 28% HCl. The
Acid dissolving power 0.53 cm3/cm3
difference is attributed mostly to the increased acid
contact time in the fracture prior to leakoff. This
tion rates, and results in an increase in etched con- will not be true for all cases, as it depends on syn-
ductivity. This effect is shown in Fig. 10-28 for ergy with other parameters.
various volumes and concentrations of acid. The These simulations were made in a formation
effective etched-penetration cutoffs were selected with 1-md permeability, and CfD was increased
at a conductivity of 10 md-ft. from 0.2 to 0.4 and 12 to 28, with the equivalent
Increasing the volume of 15% HCl from 5,000 volume of 15% acid indicating a potential benefit
to 20,000 gal increases the etched penetration from from using the larger volume, lower concentration
120 to 190 ft (53%), and the average conductivity acid for this case. However, the costs and the logis-
is increased from 1,000 to 23,000 md-ft (about tics of the treatments should also be considered.
2300%). For the same volume changes with 28% The costs of various treating fluids are listed in
HCl, the penetration increases from 151 to 220 ft Table 10-5.
(46%), and the conductivity increases from 1650 • Injection rate sensitivity
to 67,000 md-ft (4100%). This demonstrates the Etched penetration can also be increased by
large values of conductivity that can be obtained increasing the injection rate. The 10,000- and

Table 10-5. Costs (1997) of various treatment systems.
15% HCl 28% HCl Emulsified Crosslinked Water-Base Water-Base

Acid Acid Gelled Fluid Crosslinked Fluid
Corrosion inhibitor (gal) 1 3 3.5 2

Iron-reducing agent (gal) 10 28 10
Inhibitor aid (lbm) 5 5
Antisludge agent (gal) 4 4
Emulsifying agent (gal) 15
Diesel oil (gal) 300
20% HCl (gal) 700
Iron-reducing agent (lbm) 7
Chelating agent (gal) 7
Acid gelling agent (gal) 20
Nonemulsifying agent (gal) 2
KCl (lbm) 167 167
Slurried gelling agent (gal) 5.7 5.7
Surfactant (gal) 1 1
Crosslinker (lbm) 1.2
ph-control agent (gal) 4
Additive unit cost/1000 gal ($) 2239 4456 2844 3403 245 279
Note: Other unit costs: Sand $0.07/lbm, premium proppants $0.51/lbm–$0.89/lbm, hydraulic horsepower $7.40/hhp
20,000-gal volumes of 28% HCl injected at a rate effects of cooling and the pad filter cake, if there
of 10 bbl/min have CfD values from 32 to 347 and is no cake erosion, can be modeled. The results of
can benefit from increases in penetration at the simulations with a pad volume of 10,000 gal prior
expense of conductivity. Conductivity profiles from to injecting 20,000 gal of acid show an additional
simulations with the rate increased to 20 bbl/min 20% to 25% penetration (Fig. 10-30). Acid chan-
for 10,000 and 20,000 gal of 15% and 28% HCl neling was not considered in these simulations.
are shown in Fig. 10-29. The penetrations are Large pad volumes can cause more height
increased by about 60%. growth, resulting in decreased etched penetration.
• Pad volume sensitivity Simulations of this effect must be made with a P3D
acid geometry simulator, for which a stress profile
The use of a pad stage ahead of the acid stage was
is required. Maximum net pressure is usually
first performed in the 1960s. Its purposes were to
achieved in the pad stage; the acid stage causes the
cool the formation to slow the reaction rate and to
pressure to decrease because of its lower viscosity
deposit a filter cake to reduce the rate of acid fluid
and increased fluid-loss rate as well as some
loss, both of which should increase the penetration.
increase in width resulting from etching.
The technique’s results have made it almost a uni-
The optimum pad size for a given volume of
versal practice in acid fracturing, although the
acid achieves the maximum ROI, and, depending
effect of the pad filter cake on the acid fluid loss
on other parameters, acid volume may be economi-
and magnitude of the cooling have been disputed.
cally traded for pad fluid. The effect of the pad vis-
Some of the benefit may result from thinner acid
cosity depends on the combination of the other
channeling through the more viscous pad fluid.
parameters. Higher viscosity pads promote acid
The simulation of acid fracturing with multiple
channeling, which can greatly alter penetration
fluids such as pads and acids has only recently been
and etch patterns.
accomplished and is discussed in Section 6-9. The

100,000
20,000-gal 28% HCl, 10 bbl/min
10,000

1000
100
0
0 100 200 300 400
Etched fracture half-length (ft)
Figure 10-29. Conductivity profiles from simulations at a 20-bbl/min rate for 10,000 and 20,000 gal of 15% and 28% HCl.
100,000
20,000-gal 28% HCl
20,000-gal 15% HCl
10,000-gal 28% HCl
10,000
20,000-gal 15% HCl Pad + 20,000-gal 28% HCl
Pad + 20,000-gal 15% HCl

20,000-gal 28% HCl

1000
Pad + 10,000-gal 28% HCl
100
10 bbl/min
10 bbl/min
10 bbl/min
10 bbl/min
20 bbl/min
0
0 100 200 300 400
Etched fracture half-length (ft)
Figure 10-30. Effects of a 10,000-gal pad prior to injecting 20,000 gal of acid.

• Fluid-loss sensitivity on the compressibility-controlled leakoff coefficient

The fluid-loss rate has a significant effect on the Cc. The pad fluid is simulated using the laboratory-
etched penetration, not only from its effect on the measured Cw of 0.003 ft/min1/2 because its filter
total penetration of the fluid but also on the reaction cake controls its leakoff.
rate. The previous simulations were made for an Figure 10-32 shows the stage fronts of the pad
oil reservoir above the bubblepoint pressure and and acid stages for the simulations. Above bubble-
an equivalent fluid-loss coefficient of about point pressure, the fronts advance throughout the
0.002 ft/min1/2. The pad fluid, which had a labora- treatment. Below bubblepoint pressure, the pad
tory-measured fluid-loss coefficient through the front or open fracture length starts to recede after
wall filter cake Cw of 0.003 ft/min1/2 for a brine- the high-leakoff acid enters the fracture and causes
filled core, was also simulated with the reservoir- the total fluid-loss rate to exceed the injection rate.
controlled leakoff coefficient because it was less These simulations above and below bubblepoint
than the laboratory value. Figure 10-31 compares are for equivalent acid fluid-loss coefficients of
simulations of 20,000-gal 28% HCl with and with- 0.002 and 0.013 ft/min1/2, respectively. Most meth-
out a 10,000-gal pad injected at 20 bbl/min for ods of fluid-loss control discussed previously fall
leakoff above and below the bubblepoint pressure. within this range.
The effective combined fluid-loss coefficient Ct of • Reservoir temperature sensitivity
the acid below bubblepoint pressure is increased The reaction rate increases with temperature.
by about 1 order of magnitude, from 0.002 to Conductivity profile differences at 150° and 200°F
0.013 ft/min1/2, because of the effect of the [65° and 90°C] with and without a pad are shown
increased compressibility of the reservoir fluid in Fig. 10-33.
100,000
20,000-gal 28% HCl with and without 10,000-gal pad
20 bbl/min, 60-ft height, 3000-psi oil bubblepoint
fluid loss for 2500- and 5000-psi reservoir pressure
Viscosity Fluid Ct Spurt

density loss
10,000
5000 psi
Pad 2.0 0.003 0.0018 1.0
28% HCl 1.0 0.900 0.0018 0.0
Effects of fluid loss 2500 psi

Pad 2.0 0.003 0.003 1.0
28% HCl 1.0 0.900 0.013 0.0
1000
Effect
of pad Effect of pad
Below Above
bubble- bubblepoint
point
0
0 200 400 600 800
Figure 10-31. Simulation of 20,000-gal 28% HCl with and without a 10,000-gal pad and with leakoff above and below
bubblepoint pressure.

1,000,000
700

600 100,000 Land
Phosphoria
Front position (ft)
Above bubblepoint
500 Strassfurt
10,000
400 Arun Tor
Pad stage fronts 1000 Arab
Ekofish
300 Nierode Sakasar
Default
Austin
data chalk
Below bubblepoint
200 100 Smackover
San Andres
100
0
0 0 100 200 300 400
0 5 10 15 20 25 30 35 Fracture half-length (ft)
Figure 10-34. Simulation of conductivity profiles with the
Figure 10-32. Simulated stage fronts of the pad and acid same treatment in 11 formations with different reaction
stages above and below bubblepoint pressure. properties.
100,000 Formations can have other minerals that can affect

the reaction rate and therefore the penetration.
Reactivity should be determined in the laboratory.
10,000
200°F pad
No pad
1000 10-6.7. Propped or acid fracture decision
150°F pad
No pad For stimulating a carbonate or dolomite formation,
100
etched and propped fractures have both advantages
and limitations in comparison to each other; however,
10 acid fracturing should be the primary consideration.
0 100 200 300 400 500 600 700
In general, acid-etched fractures are limited in pene-
tration but can result in high conductivity whereas
propped fractures usually have a deeper penetration
Figure 10-33. Conductivity profiles at 200° and 150°F with
and without pad. but may be conductivity limited. The technology to
overcome the limitations of both techniques continues
to evolve.
The advantages of acid fracturing are
10-6.6. Formation reactivity properties • lower net pressure, minimizing fracture height
Proper simulation requires specific reaction rate para- growth
meters for the formation (see Chapter 6). Simulations • can achieve high conductivity
of conductivity profiles with the same treatment • no risk of screenout
schedule in various formations with different reaction
properties are compared in Fig. 10-34. • no proppant-flowback problems.
The formation with the greatest penetration has the The disadvantages of acid fracturing are
least conductivity. Improving its performance may • etched penetration potentially limited by higher
require more conductivity and a different strategy than fluid loss
for the formations with less penetration and higher • etched penetration potentially limited by the effect
conductivity. The reaction parameters are required for of temperature on the reaction rate
generating proper predictions. Even with the same
• potential emulsion and sludge problems in oil wells
percentage of minerals in two formations, it is possi-
ble that their reaction rates may not be the same • etched conductivity difficult to predict
because of different grain-surface mineralogies. • environmental considerations.

The optimization of acid treatments is more com- ered as a treatment option because although there
plex than that of propped hydraulic fractures. For is no significant change in length, the cost is still less
propped fractures, there is a ratio of the pad to the than that of the propped treatment. The higher con-
total volume for a specific penetration, and the conductivity combined with less etched penetration of
ductivity is directly related to the proppant permeabil- acid in formation B results in a CfD of 19 but with less
ity and concentration. The selection of the fracturing production because the dimensionless fracture time tD
fluid is usually based on compatibility and stability is about 12 times higher (see explanation in Section
characteristics or on field experience. 10-2.4). Increasing the injection rate, pad volume or
For comparison, propped fracturing treatments were both will not increase etched penetration in formation
considered for the reservoir described in Table 10-4 B because of the high fluid loss and fracture height
with the addition of upper and lower barriers with growth. The wellbore height history and net pressure
stress contrasts of 1000 psi and impermeable to fluid plots for the treatment are shown in Fig. 10-36.
loss. An optimum fracture penetration of 750 ft using If the reservoir were below bubblepoint pressure,
20/40-mesh ISP as the propping agent was deter- the fluid loss would be significantly higher, and other
mined. Acid fracturing was also simulated using the acid systems with better fluid-loss control, albeit
same injection rates and volumes. Conductivity pro- likely more expensive, would be required for the acid
files for these treatments are shown in Fig. 10-35. fracture treatment in formation A to compete with the
For these conditions, the acid fracturing treatment propped fracture.
provides about the same penetration and slightly more This comparison illustrates why there can be differ-
conductivity for the reactivity properties of formation ences in opinion on which treatment—proppant or
A. The material costs for the proppant treatment are acid—will produce the best results. Good data are
about twice that of the acid treatment. Performing the required to make the comparison.
treatments on formation B, with only the acid reac-
tivity on the formation differing, provides the same
propped penetration, but the acid-etched penetration 9600
is only 240 ft and of significantly higher conductivity.
Propped height
Acid height
2000
Costs (1997) for treating fluid systems with additives
Net pressure (psi)

Pay zone
are summarized in Table 10-5. 9800
Depth (ft)
The CfD for the propped fractures is 1.8 for both for-
Propped fracture net pressure
mations. The acid-etched fracture in formation A has 1000
10,000
a CfD of 2.4. Doubling the acid volume increases the Acid fracture net pressure
CfD from 2.4 to 11.2. This approach should be consid-
10,200 0
0 50 100 150
100,000 Treatment time (min)
Formation B
10,000
Acid etched Figure 10-36. Wellbore height history and net pressure
Formation A
plots for propped and acid treatments in Fig. 10-35.
1000
Propped
100
10
10-7. Deviated wellbore fracturing
0 200 400 600 800 Horizontal wells are often viewed as a replacement
for or an alternative to hydraulically fracturing vertical
wells. This view is common because in the idealized
Figure 10-35. Conductivity profiles for propped and acid situation of a homogeneous isotropic reservoir, hori-
treatments. Propped treatment: 12 bbl/min, 33,000-gal
25-lbm crosslinked pad, 29,000-gal 25-lbm crosslinked zontal well performance and vertically fractured well
fluid and 29,000-lbm 20/40-mesh ISP. Acid treatment: performance are similar. Unfortunately, reservoirs are
12 bbl/min, 33,000-gal 25-lbm gel pad and 29,000-gal neither homogeneous nor isotropic, and as a result, the
28% HCl.

performance of a horizontal well can differ greatly conductivity to a horizontal well with a 1000-ft lateral
from its performance in an ideal reservoir or from the extension. The fractured vertical well outperforms the
performance of a fractured vertical well. horizontal well except for the low-conductivity (CfD = 1)
Horizontal wells and fractured vertical wells have fracture example. This indicates that for accelerating
design considerations critical to their performance. production there is little reason to utilize horizontal
Fracture length and conductivity are the important wells in areas where fractured vertical wells are suc-
design parameters for a fractured vertical well, where- cessful, unless the horizontal lateral length can be
as in horizontal wells, the important factors are the ver- achieved at less cost or is in excess of the achievable
tical to horizontal permeability ratio, section thickness, tip-to-tip fracture length. The comparison in Fig. 10-37
lateral extension and wellbore damage. Insight into assumes that the horizontal well was placed in the
these critical design factors shows that coupling hori- middle of a homogeneous and isotropic reservoir with
zontal well and fractured vertical well technologies can a vertical to horizontal permeability ratio kV/kH of 1.
maximize reserve recovery in some reservoirs. For accelerated production, vertical placement in this
type of reservoir is inconsequential. However, place-
ment of the horizontal wellbore can be important in
10-7.1. Reservoir considerations mitigating unwanted water or gas production.
Comparison of horizontal well and fractured vertical The value of kV/kH is critical to the success of a hor-
well performance indicates that a fractured vertical izontal well’s performance. This is shown in the Fig.
well will outperform a horizontal well with a lateral 10-38 comparison of the PI of a fracture in a vertical
length less than or equal to the tip-to-tip fracture well to that in a horizontal well in reservoirs with
length. This performance variance occurs if CfD is kV/kH values of 1, 0.5 and 0.1. A horizontal well in a
greater than 3. The plot of cumulative recovery versus reservoir with poor vertical permeability can signifi-
time in Fig. 10-37 compares fractured vertical wells cantly underperform in comparison with a hydrauli-
with a 500-ft fracture half-length and varied fracture cally fractured vertical well. The extreme, but not
90,000
L = 1000 ft
xf = 500 ft, CfD = 1
80,000
xf = 500 ft, CfD = 5
xf = 500 ft, CfD = 10
70,000
Cumulative recovery (MMcf)
60,000
50,000
40,000
30,000
20,000
10,000
0
0 2.0E+3 4.0E+3 6.0E+3 8.0E+3 1.0E+4 1.2E+4 1.4E+4 1.6E+4
Production time (days)
Figure 10-37. Comparison of vertically fractured wells with different fracture conductivity values and 500-ft penetration
(1000 ft tip-to-tip) with a 1000-ft horizontal well.

9E+4
xf = 500 ft, CfD = 5
kV ⁄ k H = 1
8E+4
kV ⁄ kH = 0.5
kV ⁄ kH = 0.1
7E+4
Cumulative recovery (MMcf)
6E+4
5E+4
4E+4
3E+4
2E+4
1E+4
0
0 2.0E+3 4.0E+3 6.0E+03 8.0E+3 1.0E+4 1.2E+4 1.4E+4 1.6E+4
Production time (days)
Figure 10-38. Production comparison of vertical and horizontal wells with vertical to horizontal permeability ratios of
1, 0.5 and 0.1.
uncommon, case is an impermeable barrier that iso-

lates the well from a section of the reservoir. 1500
kH = 1 md Horizontal well
The rule for the permeability ratio is also shown h = 100 ft Hydraulically fractured
Tubing ID = 2.441 in. vertical well
in the Fig. 10-39 production comparison of horizontal
Flow rate (STB/D)
wells and fractured vertical wells with fracture con- 1000 kV/kH = 1.33
ductivities of 200 and 2000 md-ft in a 100-ft thick, kfw 1.0
160-acre reservoir with 1-md horizontal permeability. 2000 0.5
500
For a typical anisotropy of kV/kH = 0.33, the vertically 200
fractured well outperforms the horizontal well. How- 0.33
ever, for kV/kH = 1.33, the horizontal well provides
0
better performance for lengths greater than 400 ft. 0 1000 2000
This result links the successful use of horizontal wells Borehole length, 2xf (ft)
to naturally fractured formations. For this application,
large values of kV/kH are common. For the constant- Figure 10-39. Anisotropy effect on the performance of a
conductivity kf w of 2000 md-ft, CfD is 10 for a tip-to- horizontal oil well and comparison with a hydraulically frac-
tip length of 400 ft and decreases to 2 at a 2000-ft tured vertical well in a thick, low-permeability formation.
length. The relative performance of the vertically frac-
tured well is greater for lower horizontal permeability
and larger formation thickness. An advantage of a The coupling of horizontal wells and hydraulic frac-
horizontal well is that the desired azimuth of the turing technology in reservoirs with a large pay thick-
lateral extension can be achieved, whereas the azimuth ness, low values of kV/kH or both should result in
of a vertical fracture cannot be controlled. This control increased well performance. Drilling a long lateral and
can be important in complex geological settings (e.g., fracture stimulating with one or more orthogonal frac-
optimum alignment with natural fractures). tures provide the benefits of both technologies and

outperform the application of either individual tech- sections. It is not necessary that axial fractures meet.
nology. One effect of the orthogonal fractures or the If the horizontal wellbore is perpendicular to the frac-
role of longitudinal fractures is to introduce vertical ture orientation, multiple transverse fractures can
permeability. The extended reach serves the purpose improve the well productivity and drainage area.
of placing the fractures at specific locations and Cumulative gas recovery versus the number of trans-
replaces drilling vertical wells and fracturing. The verse fractures in horizontal wells with 2000-, 3000-
azimuth of the fractures is not changed. and 4000-ft lateral extensions is plotted in Fig. 10-41.
Because limited perforated intervals are required for The simulations were conducted on a gas reservoir
a successful treatment operation (see Section 11-3.2), with a permeability of 0.01 md. Production interfer-
production from a fractured horizontal well enters the ence between the fractures is the limiting factor for
wellbore only through the fractures. the number of fractures. The spacing between the same
number of fractures is greater with the increasing hori-
zontal extension. The figure shows that for a 2000-ft
10-7.2. Fracture spacing horizontal well, more than three transverse fractures
The successful application of multiple fractures in does not increase production significantly. This conclu-
horizontal wells is dependent on the relative orienta- sion is dependent on the reservoir and fluid properties.
tion of the wellbore with respect to the maximum The figure further indicates that for the 3000- and
principal stress (i.e., fracture orientation) as well as 4000-ft lateral extensions, the creation of more than
the number and spacing of the fractures. Longitudinal three or four multiple fractures must be evaluated with
(axial) and transverse fractures are shown schemati- the additional costs and risk associated for the opera-
cally in Fig. 10-40. In longitudinal fracturing scenar- tion. Three fractures in the 2000-ft lateral extension
ios where the horizontal wellbore is in line with the and four fractures in the 3000-ft extension will have
fracture direction, a fracture overcomes the adverse a fracture spacing of 1000 ft.
effects of poor vertical reservoir permeability and for- With multiple transverse fractures, the outer frac-
mation thickness. In this orientation, the wellbore area tures dominate production after some period of time.
can provide a significant contribution to the effective Figure 10-42 (Raghavan et al., 1994) shows the nor-
fracture conductivity (see Section 12-5.1). Multiple malized rate of five fractures with the same length and
longitudinal fractures originating from isolated sec- conductivity properties. The distance D between the
tions may have to be created to achieve the desired outermost fractures is 10 times the fracture half-length
geometry along the wellbore length in long horizontal xf ; i.e., D is equal to the sum of the fracture length.
When production from the intermediate fractures
declines significantly, perforating the section between
the outermost fractures can improve production.
Surface
10-7.3. Convergent flow
Axial
fractures Transverse
Production into a horizontal wellbore from an orthog-
Reservoir
fractures onal fracture or a longitudinal one with a short perfo-
rated interval will exhibit convergent flow as the flow
goes from linear far from the wellbore and converges
α = 0°
with accelerating velocity at the wellbore. The
σv increased velocity causes an additional pressure drop
α = 90°
and pressure loss. Under some circumstances the
higher velocity may cause turbulent or non-Darcy
σh,min
flow that can be substantially greater than the Darcy
σH,max component. Convergent flow behaves as a positive
skin effect and can have an offsetting effect to the
negative skin effect created by the fracture treatment.
Figure 10-40. Longitudinal (axial) and transverse This is another instance where the fracture conductiv-
fractures.

40,000
35,000
Cumulative recovery (MMscf)

30,000
25,000
20,000
15,000
10,000
Lateral extension
5000 4000 ft
3000 ft
2000 ft
0
0 2 4 6 8 10 12 14 16
Number of fractures
Figure 10-41. Cumulative gas recovery versus the number of fractures for a transverse fractured horizontal well with
2000-, 3000- and 4000-ft lateral extensions.
q j/q 1,5 2,4 3

Normalized rate, qj /q
10–1
1 2 3 4 5
n = 5, CfD,j = 10
D/xf = 10
10–2
10–4 10–2 1 10 2
Dimensionless time, tD
Figure 10-42. Normalized rate of five fractures with the same length and conductivity properties (Raghavan et al., 1994).
ity must be overdesigned. The effects of various para- drop because of Darcy flow, and the second term rep-
meters on the total pressure drop ∆pct(ro – ri) in the frac- resents the additional pressure drop because of non-
ture for a constant production rate q (Norris et al., Darcy flow. The fracture design parameters of the
1996) are fracture permeability kf and propped width w are in
the denominator of first term and the width squared
µq  r  βρq 2  1 1  in the denominator of the second term. Because the
∆pct = ⋅ ln  o  + 2 ⋅ − , (10-3)
2 πk f w  ri  (2 πw )  ri ro  greatest part of the pressure drop is near the wellbore,
a tail-in with a higher concentration could be benefi-
where µ is the viscosity, β is the non-Darcy coefficient cial to increasing the propped width. For the tail-in,
measured in the laboratory, ρ is the density, ri is the Eq. 10-3 is used in series, and ri could also be consid-
inner radius, and ro is the outer radius, equivalent to ered a tail-in radius if it has sufficient conductivity
one-half of the fracture height. The first term on the to be infinite acting (i.e., causing negligible pressure
right-hand side of the equation represents the pressure drop).

The relative effect of the parameters kf and w is 10-7.4. Fracturing execution in

shown in Fig. 10-43 for all cases where the pressure deviated and horizontal wells
drops for the Darcy and non-Darcy flows are initially
equal. The parameters are constant over the converg- Operational risks are magnified for placing numerous
ing area. The change in width affecting the non-Darcy fractures in a horizontal well. These and associated
component has a greater impact than a change in costs must be considered in the planning. Horizontal
proppant permeability affecting only the Darcy com- wells can be completed with an uncemented liner, but
ponent. The comparisons assume that β in Eq. 10-3 this completion practice is generally unacceptable for
does not change with permeability; however, β usu- fracture stimulation in deviated or horizontal wells.
ally decreases with increased permeability. If the per- Zonal isolation is essential for successful fracture
meability is doubled for a constant rate in Fig. 10-43, stimulation of high-angle wells and can be achieved
the Darcy pressure drop is reduced 50%. The total by cementing using good cementing practices (see
pressure drop is reduced only 25% because the non- Chapter 11) and selective perforating. The use of
Darcy part is not changed. Doubling the width external casing packers should also be considered.
reduces the pressure drop 63%. This example repre- Fracturing pressures can be much greater in hori-
sents only the pressure drop in the region of converg- zontal and deviated wells in comparison with their
ing flow in the fracture. vertical well counterparts. The excessive pressures
Non-Darcy flow is considered in Chapters 8 and result from restricted communication between the
12. The relationship with conductivity is discussed wellbore and the fracture (see Chapter 6). Laboratory
in detail in Chapter 8, including the important role and field evidence indicate that this poor wellbore and
of two-phase flow. fracture communication is caused by poor alignment
100
Width
10 Total flow
Relative pressure drop factor
Non-Darcy flow
Darcy flow
Proppant permeability
Total flow
Non-Darcy flow
1
Darcy flow
0.1
0.1 1 10
Propped width or proppant permeability factor
Figure 10-43. Effect of proppant permeability and width on the convergent pressure drop in the fracture.

among the perforations, wellbore and fracture orienta- substantial. Similar to the case of wellbore restric-
tion as well as by the creation and propagation of tions, this component of additional pressure drop
multiple fractures. In anticipation of this problem, disappears after the shut-in of pumping. An expected
contingency plans should be made to control exces- magnitude for the diverging-flow component can
sive pressures and establish good communication be obtained from a fracture model that considers
between the wellbore and main fracture. Some meth- this effect or Eq. 10-3 using the w2/12 substitution.
ods are discussed in the following sections. The diverging flow component within a perfectly
connected fracture can be substantial and may be
• Perforation strategy
misinterpreted as a restriction if this effect is not
Perforation plays an extremely important role in the considered.
execution of a hydraulic fracture from a horizontal
• Near-wellbore effects
or deviated well. The perforations represent the
only communication channel between the wellbore Excessive friction pressure is often cited as evi-
(i.e., casing) and the fracture. In horizontal and dence of multiple fractures, poor wellbore-fracture
deviated well fracturing, as with vertical wells, communication or both conditions. It is more com-
compromises can be made among the hole size, mon in horizontal or deviated wellbores, which
depth of penetration, shot density and perforation are prone to having high tortuosity resulting from
phasing (see Chapter 11). misalignment and multiple initiation fractures.
Perforation alignment with respect to the far-field Techniques used to reduce the excessive pressure
fracture azimuth is the most important aspect of are proppant slugs (Cleary et al., 1993; Stadulis,
perforation design. Because in many cases the 1995), viscous fluid slugs (Aud et al., 1994; Hainey
hydraulic fractures propagate from the base of the et al., 1995), pretreatment injections and shut-ins,
perforation tunnel (Behrmann and Elbel, 1992; and an extended initial stage of low proppant con-
Morales et al., 1993; Weijers, 1995), the depth of centration to erode restrictions.
penetration is a much less important parameter than Proppant slugs are small volumes of low-
the diameter of the perforation (see Chapter 11). concentration proppant slurries that are pumped
For highly deviated wells that are not in the pre- ahead of or during the main fracture stimulation
ferred fracture plane, minimizing the length of the treatment. The purpose of the slugs is to screen
perforated interval is used to minimize the creation out some of the multiple fractures so they will
of multiple fractures (see Chapter 11). One guide- cease accepting fluid and propagating. The suc-
line is that the perforated interval should be limited cessful application of proppant slugs enhances the
to 4 times the wellbore diameter. Thus, for horizon- propagation of a dominant fracture, which is of
tal and deviated wells, it is not uncommon to perfo- sufficient size (width) and provides the required
rate only a 2-ft interval in the wellbore. Extended pathway for the fracture treatment slurry. Proppant
perforation intervals can result in multiple frac- slugs are pumped at concentrations of up to 4 ppg
tures, putting the fracture stimulation and well per- and are typically overflushed through the perfora-
formance at risk. In the event that a larger interval tions. Their role to reduce fracture entry restriction
is perforated, steps should be taken to attempt to can also include erosion effects.
control the creation and propagation of multiple Because excessive friction is considered evi-
fractures as discussed in the following. Both prop- dence of a near-wellbore problem or multiple frac-
pant and viscous fluid slugs have been used for this tures, the success or failure of proppant slugs is
purpose. measured by the reduction of excess friction pres-
Correctly assessing excessive friction in highly sure. The number of proppant slugs pumped should
deviated wells also requires consideration of the depend on whether the excessive friction is being
pressure drop within the radial region of the frac- reduced. Therefore, the final design of this part of
ture. This is the injection analog of Eq. 10-3, with the fracturing operation must be performed on loca-
the channel permeability w2/12 substituted for kf tion and requires reliable downhole injection pres-
and only the laminar Darcy component applicable. sure data. As long as entry friction pressure
The radically diverging flow component during reduction occurs, the proppant slugs should be con-
injection at high rates and of viscous fluids can be tinued. Once no additional reduction of the entry

friction pressure is observed, the main fracture

stimulation can be initiated. 9000
Slug Bullhead Calibration
• Proppant transport in tubulars 8000 stages fracture fluid stage
Another design consideration for fracturing deviated 7000

wells is proppant transport in the horizontal or devi- 6000 40
ated section. There is a critical velocity, below which
5000

proppant is deposited at the bottom of the horizontal Horizontal 30
4000 section
section; after it is deposited an even greater velocity friction
pressure
is required to resuspend it (Shah and Lord, 1990). 3000 20
For normal fracturing fluids and injection rates, the 2000
velocities are generally high enough to keep the Sand 10
1000 slugs
proppant in suspension. However, the velocity
should be considered, particularly if low-viscosity 0 0
0 100 200 300 400 500
or delayed crosslinked fluids are used and for contin-
Time (min)
gency plans if the treatment is terminated early and
the proppant must be cleaned out using coiled tubing
or other means. These viscosity and velocity require- Figure 10-44. Bottomhole treating pressure versus time for
formation breakdown, three proppant slug stages and a
ments also ensure that the proppant plugs reach the calibration injection.
perforations.
mately 800-psi friction pressure, most of which

10-7.5. Horizontal well example occurred during the first stage without proppant.
The following is an example of a fracture treatment The flush fluid of the last slug injection was bull-
of a horizontal well in the Valhall field in the North headed with fracturing fluid. The fracture was allowed
Sea. Fracture stimulations in this chalk formation are to close to minimize the effect on the calibration
performed in both vertical and horizontal wellbores. analysis. The injection rate was reduced to 0.1 bbl/min
In addition to the normal fracture conductivity (instead of shutting down) near the end of displace-
requirements, propped widths for fracture treatments ment and while the fracture was closing. This reduc-
are increased to offset the effects of embedment into tion provided shear on the crosslinked fluid to prevent
the relatively soft reservoir rock (Smith et al., 1987). excessive gelation. After the pressure dropped below
To accomplish these objectives, TSO fracturing tech- fracture closure, a 30,000-gal calibration treatment
niques are used. Creating additional fracture width was injected into the formation at 35 bbl/min. Figure
beyond that to offset embedment minimizes the addi- 10-45 shows a plot of pressure versus the G-function.
tional pressure drop from convergent flow to the Analysis indicated a closure pressure of 5875 psi at
limited perforated section. a closure time of 19.9 min and a fluid efficiency of
Some of the information for this example was 0.4. For this efficiency, a TSO fracture treatment with
reported by Norris et al. (1996). The bottomhole a 43% pad fraction was designed and executed. An
treating pressure of a fracture treatment at the toe 8-min pad was pumped, and the initial proppant con-
of the horizontal section is plotted in Fig. 10-44. The centration of 2 ppg was injected. Analysis of the cali-
pressure gauge was near the heel of horizontal section, bration data indicated a higher fluid-loss rate and less
and therefore, considerable friction occurs over the efficiency than anticipated by the design. Therefore,
horizontal distance to the fracture. Also shown on the more fluid volume was required to obtain the desired
figure are the injection periods for formation break- TSO penetration, and a larger fraction of this volume
down, three proppant slug stages, a calibration injec- was required as a pad. The pad was increased from
tion and a step rate test. Proppant slug stages were 10,000 to 25,000 gal to induce the screenout at 120-ft
used because high injection pressures were noted dur- penetration. The initial slurry stage was at 2 ppg, and
ing the initial formation breakdown. The difference in the TSO occurred during the end of this stage, after
the maximum and final bottomhole injection pressures which the slurry proppant concentration was increased
after these stages indicates the reduction of approxi- to 3.5 ppg and then ramped up to 11 ppg during the

6600 800
Zone 1
Zone 2
Zone 3
pc3 = 6380 psi
6400 Derivative 1 700
Derivative 2
Derivative 3
Derivative d(bottomhole pressure)/dG(t) (psi)

6200 600
pc2 = 6220 psi
6000 500
pc1 = 5875 psi
5800 400
5600 300
5400 200
5200 100
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
G-function of time since shut-in
Figure 10-45. G-function versus time of post-mini-fracture treatment pressure decline data.
remainder of the treatment. The TSO occurred and total pressure drop of 1800 psi. The assumption of
stopped propagation at approximately 20 min; at this equivalent proppant embedment into the formation of
time the net pressure rose from 200 to more than 2 lbm/ft2 reduced the simulated width to 8 lbm/ft2 and
1000 psi during the next 70 min. The screenout and corresponds to a width of about 1 in. For this reservoir
pressure were matched with a geometry simulation. and fracture treatment, the 1-in. propped width gives a
The geometry simulation predicted an average nearly infinite-acting conductivity fracture. If the treat-
propped fracture concentration of 10 lbm/ft2. ment had been designed for only 4 lbm/ft2, the effec-
Production analysis showed a total drawdown from tive width would have been 2 lbm/ft2 and the pressure
the reservoir to the wellbore of about 1800 psi. The drop in the fracture 10 times greater (1000 versus 100
convergent flow pressure drop in the fracture was cal- psi) at the same production rate. The PI would have
culated to be about 100 psi from Eq. 10-3, of which been correspondingly reduced by more than 50%,
60 psi was due to non-Darcy effects. This pressure with 10% of the reduction resulting from the Darcy
drop is considered insignificant compared with the and 40% from non-Darcy convergent flows.

◆
Fracturing Operations
J. E. Brown and R. W. Thrasher, Schlumberger Dowell

L. A. Behrmann, Schlumberger Wireline & Testing
11-1. Introduction fracture orientation) and potential fracture geometry

under these conditions.
Hydraulic fracturing is a particularly complicated Fracture initiation from wellbores that are not
enterprise. The purpose of hydraulic fracturing is the aligned with one of the principal stress directions can
placement of an optimum fracture of a certain geom- be complex. In these cases, fractures may not be a
etry and conductivity to allow maximum incremental simple singular planar feature. Misalignment may
production (over that of the unstimulated well) at the lead to the initiation of either multiple near-wellbore
lowest cost. This process combines the interactions fractures or fractures that that are forced to twist and
of fluid pressure, viscosity and leakoff characteristics change direction to ultimately reorient and align with
with the elastic properties of the rock, which have the far-field minimum horizontal stress. Potential
been the subjects of the preceding chapters. problems can affect the geometry and productivity
Accomplishing this, while taking into account all of the fractures; there may be limited fracture com-
the presented technology, requires significant atten- munication with the wellbore, and the fracture width
tion to the treatment execution involving optimized profile at or near the wellbore may be restricted.
completion and perforating strategies, appropriate A restriction or reduced width can be sufficient to
treatment design, control and monitoring of rate, and cause proppant bridging during the treatment, poten-
pressure and fluid characteristics. tially resulting in an early screenout and restricted
production after treatment.
Planning for an S-shaped completion, where the
11-2. Completions extended reach wellbore is brought back to a more
Until recently, the majority of wellbores have been vertical or 90° attitude through the formation, is gen-
vertical and designed to produce from a single pro- erally advisable. The connectivity of a well and frac-
duction string. Other types of completions that can ture is always important, but it is critical in high-
be planned and performed are permeability fracturing. Although tortuosity is a
problem that affects all fractures during both execu-
• deviated wells tion and production, limited exposure is not a major
• horizontal wells issue in low-permeability reservoirs; it is in high-
• multiple completions permeability formations. Thus, deviated wellbores
with limited perforated intervals may provide attrac-
• multiple laterals.
tive fracture treatments (although S-shaped wells
The choice of the type of completion should are even better). However, for high-permeability
account for all considerations, ranging from drilling fracturing both S-shaped wellbores and 180° phased,
to production to abandonment. oriented perforations or slots cut along the well
trajectory are highly advisable. This approach is
intended to safeguard against further reducing a natu-
11-2.1. Deviated and S-shaped completions rally small fracture conductivity (Chen and Econo-
Precompletion and drilling planning should be given mides, 1998). Wellbore angles of less than 20° may
serious thought in cases where fracturing may be be necessary to achieve optimum fracture geometry
considered for highly deviated wellbores. It is impor- and performance. In general, production engineering
tant to understand the relationship between the well- considerations suggest that the wellbore orientation
bore angle, formation in-situ stress fields (dictating should coincide with the fracture orientation. The

wellbore orientation can be either vertical or horizontal. tion of stimulation. Because of the reduced wellbore
In the horizontal mode, the wellbore must be placed diameters, surface treating pressures are increased
to have either longitudinal or transverse fractures. If owing to excessive friction pressures. This limits the
transverse fractures are created, multiple treatments rate and size of treatments that can be pumped and
will have to be performed that will require proper therefore reduces the effectiveness of the stimulation
zonal isolation. treatment. Small tubulars also limit the application
of downhole tools; e.g., the stimulation of multiple
zones is prevented because zonal isolation tools can-
11-2.2.Horizontal and multilateral not be applied.
completions If it is considered essential to use slimhole com-
Effective stimulation of horizontal wellbores requires pletion techniques, the completion design should
specific completion procedures. In addition to the nor- specify heavy-walled tubing as casing. This allows
mal considerations for pressure and flow rates, hori- using higher pump rates for the higher treating pres-
zontal completions must address the problems associ- sure limitations.
ated with treatment placement into or along the zone
or zones of interest. In the case of executing several
hydraulic fracture treatments along a single horizontal
11-2.4. Zonal isolation
wellbore, the completion requirements can be compli- Usually there are several potential producing zones
cated and extensive. Zonal isolation for each individ- penetrated by a wellbore that must be hydraulically
ual treatment must be part of the completion design. fractured. To ensure that each zone is stimulated
This requires cementing the entire liner, using external effectively, these intervals must be isolated from one
casing packers, placing permanently installed selective- another. Several isolation methods have proved to
zone mechanical tools and/or using retrievable zonal be effective. These methods can be used only when
isolation tools. Coiled tubing is commonly used to the various formations and intervals are isolated from
place and operate zonal isolation tools in deviated and each other behind the casing with cement.
horizontal wells and multilateral completions.
• Importance of the cement sheath
The use of multilaterals can provide better reservoir
drainage and management. Laterals could conceivably The cement sheath must provide zonal isolation
replace the need for fracturing in some cases. In other during both production and stimulation operations.
cases, optimum reservoir drainage may require a com- For a producing well, the cement seal between the
bination of laterals and stimulation. For these compli- pipe and formation must be tight to prevent fluids
cated completions, proper planning is essential. See from flowing through the annular area (see Side-
Chapter 1 for an extensive discussion of appropriate bar 11A). The permeability of a set Portland
well paths to match specific geological flow units. cement of normal density is in the low microdarcy
range. If the cement does not bond perfectly to
either the pipe or the formation and a small channel
11-2.3. Slimhole and monobore completions remains, the effective cement permeability can be
significantly increased. Large permeabilities may
Slimhole and monobore completions are used to result from channel widths that are quite small
lower drilling and completion costs. Slimhole com- (Nelson, 1990). For example, a channel width of
pletions can result in production casing sizes as small only 1.4 × 10–4 in. is sufficient to create an effective
as 2 in. Tubulars, which are generally considered pro- cement permeability of 1000 md. Channel perme-
duction tubing, become both casing and production abilities of this order may allow significant cross-
tubulars. These small diameters can prohibit the use flow between zones.
of special stimulation tubulars, which in turn may During hydraulic fracturing these small channels,
limit treating pressures. or microannuli, are relatively insignificant. The
The need to drill and complete wells at less cost is effective cement permeability does not create a
evident, but there may be a trade-off when stimulation high leakoff risk for the fracturing fluid within the
must be considered. The small tubulars used in slim- annulus. A leakoff rate for a microannulus is less
hole and monobore completions can limit the applica- than 1 gal/min. If major channels within the cement
11-2 Fracturing Operations

11A. Factors influencing cement bond integrity This is accomplished by adding more or less water and fillers
with low or high specific gravity. Such changes in cement
E. B. Nelson, Schlumberger Dowell slurry composition can affect the compressive strength and
permeability significantly (Nelson et al., 1990). Although many
Primary cementing is the process of placing cement in the industry and regulatory standards exist regarding the mini-
annulus between the casing and the formations exposed to mum compressive strength of well cements for various pur-
the wellbore. The major objective of primary cementing is to poses, none pertains to cement matrix permeability. As a
provide zonal isolation in the wellbore. To achieve this objec- general rule, every effort is made to provide a cement sheath
tive, a hydraulic seal must be created at the cement/casing across a producing zone with at least 1000-psi compressive
and cement/formation interfaces, and the cement itself must strength and no more than 0.1-md permeability to water.
be sufficiently impermeable to prevent fluid movement
through the cement matrix. Without effective zonal isolation Volume changes
in the wellbore, the well may never attain its full production
potential. When Portland cement reacts with water, the total system
No shortchanging the quality of the cement or the cement/ undergoes a net volume diminution, usually ranging from 6%
casing and cement/formation bonds can ever be justified. to 8% (Lea, 1971). This absolute volume decrease occurs
Flow of fluids along the cement sheath is invariably an unde- because the absolute density of the hydrated material (cal-
sirable occurrence. For a producing well, this is manifested cium hydroxide and C-S-H gel) is greater than that of the
either by the loss of reservoir fluids through crossflow along initial reactants (cement and water). Despite the decrease
the cement sheath or by the influx of underground fluids from in absolute volume, the external dimensions of the set
other formations into the active layer. During hydraulic fractur- cement, or the bulk volume, remain the same or increase
ing, significant escape of fluids through an imperfect cement slightly (Fig. 11A-1). To accomplish this, the internal porosity
sheath may result in either undesirable fracture height migra- of the system must increase.
tion or screenout of the intended fracture in the targeted for- Cement systems that expand after setting are recognized
mation because of the fracturing fluid loss (Economides, 1990). as a means of sealing microannuli and improving primary
Many factors can influence the degree of zonal isolation pro- cementing results. The improved bonding is the result of
vided by the cement sheath, including the removal of drilling mechanical resistance or tightening of the cement against the
fluid from the annulus prior to cementing, permeability of the pipe and formation, and improved bonding can be obtained
cement matrix, dimensional stability of the cement sheath and even if mud is left on the casing or formation surfaces. In
annular fluid migration. addition, once the cement has expanded to eliminate void
spaces, further expansion is translated into a reduction of
Mud removal internal cement porosity. However, excessive expansion
(greater than 1%) should be avoided, as buckling of casing
To meet the objective of zonal isolation, the drilling fluid must or damage to the formation may occur.
be fully removed from the annulus and the annular space Several cement additives are currently used to induce
must be completely filled with cement. Incomplete mud dis- expansion, including sodium chloride, calcium sulfate hemi-
placement can result in the formation of channels (or micro- hydrate and calcined magnesium oxide. All of these rely upon
annuli) across producing zones and allow interzonal commu- the crystallization or formation after the cement has set of
nication. Without complete mud removal, the benefits of all compounds that have a lower bulk density than their precur-
subsequent efforts to improve cement sheath integrity, and sors. For example, when magnesium oxide reacts with water
the effectiveness of stimulation treatments, are in jeopardy. to form magnesium hydroxide, the hydrated material occupies
Many primary cementing techniques have been developed more space than the original ingredients. A comparison of
to maximize the probability of achieving complete mud dis- unconfined linear expansion between a neat Portland cement
placement. From this work, a program of good cementing system and a typical expansive cement system is shown in
practices has evolved that is acknowledged by most operators Fig. 11A-1.
today. Such techniques include casing centralization, mud
conditioning (chemical and physical) prior to cementing, cas-
ing movement during mud displacement, preflushes and
spacer fluids prior to cementing and careful optimization 0.25
of the fluid flow regime by controlling the fluid rheology and
pump rate (Smith, 1984).
0.20
Linear expansion (%)
Ettringite-based expansive
Cement permeability cement system (14.8 lbm/gal)
0.15
Portland cement is used in nearly all well cementing opera-
tions. It is composed mainly of anhydrous calcium silicate and
calcium aluminate compounds that, when brought in contact 0.10
with water, react to form hydrated compounds. The principal
reaction products are calcium hydroxide and a quasi-amor- 0.05 Neat Portland
phous calcium silicate hydrate (C-S-H gel). C-S-H gel is
cement (15.8 lbm/gal)
responsible for the development and maintenance of com-
pressive strength. The normal density of Portland cement 0
slurries used in primary cementing operations is about 0 5 10 15 20 25 30
16 lbm/gal. This corresponds to a water content of about 40% Time (days)
by weight of cement (that is, 40-lbm water added to 100-lbm
cement). Within 7 days the compressive strength is expected
to reach about 4000 psi and the permeability to water will be Figure 11A-1. Comparison of expansion between neat
less than 0.001 md. Depending on the wellbore environment, Portland cement and a typical expansive cement.
the optimum cement density may vary from 9 to 22 lbm/gal.

11A. Factors influencing cement bond integrity (continued)
Annular fluid migration Several physical and chemical techniques have been devel-
oped that can prevent or reduce the severity of annular fluid
During a cement job and for a short period afterward, the
migration. Foremost are the good cementing practices
slurry is liquid and capable of transmitting sufficient hydrostatic
described here. Packers have been applied to seal the annular
pressure to prevent the invasion of formation fluids into the
space above and below a producing zone. However, the most
annulus. Before the cement slurry develops structural integrity,
common methods involve the addition of chemicals to modify
it undergoes a transition period. During the transition period,
the behavior of the cement slurry during the transition period.
absolute volume reduction and gelation result in lowering of
Such systems include slurries with very low fluid loss, com-
the hydrostatic pressure exerted on the formation (Fig. 11A-2).
pressible cements (e.g., foams or slurries containing gas-
If the hydrostatic pressure falls below the formation pressure,
generating agents), expanding cements, thixotropic cements,
annular fluid migration may occur. This can severely compro-
right-angle-set cements (the duration of the transition period
mise zonal isolation (Parcevaux et al., 1990), particularly for
is minimized) and “impermeable” cements (containing latices,
gas zones.
polymers and surfactants).
10
Fresh water
20
8.3
4-lb
m/g
Depth (ft)
al w
ate
r
150°F bath 30
.. p1
. .
.. . .
16
..
.4
.
-lb
. . . Pressure
m
. 40
/g
Tubing
. ..
al
. transducers and
ce
m
. .. .. thermocouples
en
t
. . .
. ..
. p2
0 hr
4.0 hr
3.0 hr
2.5 hr
2.0 hr
1.5 hr
1.0 hr
0.5 hr
0.2 hr
Porous plate 50
Reg.
Meter
Nitrogen
0 4 8 12 16 20 24 28
Pressure (psi)
Figure 11A-2. Hydrostatic pressure transmission of a cement slurry versus slurry consistency
(Levine et al., 1979).
can be avoided, containment of the fracturing treat- Consequently, the resulting condition of the cement
ment should be possible. sheath following hydraulic fracturing is difficult to
The effects of fracturing pressures on the adhe- predict. Sonic logs run after fracturing treatments
sion tension between the cement and casing or the typically indicate that the cement bond (hydraulic
cement and formation are not clearly understood. seal) across the fractured interval is destroyed, but

the bond farther uphole remains intact. The loss of a bridge plug immediately above the perforated in-
the cement bond across the fractured interval prob- terval. The next zone is then perforated and fractured.
ably does not affect the placement or containment • Sand plugs
of the fracturing treatment. Cement of relatively
A similar method of isolation can be achieved by
low compressive strength should prevent the frac-
using sand plugs after the fracturing treatment. The
ture from migrating between the casing and the
volume of sand necessary to cover the perforated
formation; however, alteration of the cement bond
interval is added to the casing. The sand plug is
may occur during the treatment. Any failure of the
tested by applying pressure to the casing, and then
hydraulic seal may result in a microannulus that
the next zone is perforated and stimulated. Once all
will lead to the crossflow of reservoir fluids.
zones have been fracture stimulated, the sand can
• Fracture placement control be circulated out of the wellbore by using either
The most reliable method of controlling the place- conventional or coiled tubing. The amount of sand
ment of fracturing fluids is to limit perforations to required above the top perforations is generally
a single zone. When several zones of a well are to small and can be calculated by applying Darcy’s
be stimulated, the individual zones can be isolated law to linear flow:
from one another and stimulated individually. This
 Ak∆p 
can be accomplished through progressive perfora- qi = 7.827 × 10 −4  , (11-1)
tion and stimulation. After a fracturing treatment  µL 
has been placed in the first zone, it is isolated, and where qi is the rate in bbl/min, A is area in ft2, k is
another zone of interest is perforated and treated in permeability in darcies, ∆p is the pressure drop in
another single stage. Of course, this methodology psi, µ is viscosity in cp, and L is the length in ft.
works best when the deepest zone is completed first Using this expression, 10 ft of 20/40-mesh sand in
and subsequent zones are individually stimulated 51⁄2-in. casing will create a pressure drop of more
by working uphole. than 6000 psi for a 40-cp linear gel leaking through
Another effective multizone stimulation tech- the sand pack at 0.5 bbl/min. A mixture of sand
nique is coiled tubing–conveyed treatment (see meshes can be used if the permeability of the sand
Sidebar 11B). Other multilayer fracturing techniques pack must be reduced to prevent flow through the
that can be used are discussed in Chapter 10. pack.
• Mechanical bridge plugs • Frac baffles
Several mechanical methods are available to pro- Mechanical diversion can also be accomplished by
vide adequate isolation between perforated zones. using frac baffles. Frac baffles are run as part of the
The most reliable method is the use of mechanical casing string and are placed between individual
bridge plugs, which can be run on tubulars or wire- producing zones. After the lowest interval is perfo-
line. Bridge plugs that are run on tubulars are rated and fractured, a ball is dropped down the cas-
retrievable and can be moved and reset several ing. The ball seats on the baffle and prevents fluid
times. Wireline bridge plugs cannot be moved once flow below this point. The next zone can then be
set, and their removal typically requires milling perforated and fracture stimulated. When multiple
after the treatment. Wireline bridge plugs are used zones are isolated with baffles, care must be taken
when several treatments are attempted in one day to place the baffles so that the baffle openings pro-
or when a rig is not over the hole during the treat- gressively decrease in size from top to bottom; i.e.,
ment. They can be run in the hole quickly, and the bottom frac baffle has the smallest diameter
cleanout trips are not required between stages. The opening. The first ball must be able to pass through
retrievable bridge plugs are used when zones are the upper baffles and still seat in the bottom baffle.
individually tested before another zone is opened. Similarly, the second ball must pass through the
Any excess proppant must be circulated out of the upper baffles and seat in the second baffle from
hole before the tool is moved to prevent the prop- the bottom, and so on.
pant from sticking the tool. A typical treatment All baffles have an inherent weakness owing to
involves perforation of the bottom zone, hydraulic the limited area available in the casing coupling.
fracture treatment and zonal isolation by setting The pressure differential across the baffle should

11B. Coiled tubing–conveyed fracture treatments • For multistage treatments, isolate the lowest perforated
interval. For bypassed pay zones, isolate the interval of
Sunil N. Gulrajani, Schlumberger Dowell interest.
Wellbores typically intersect several pay intervals or zones. • Perform the fracture treatment as designed.
Usually all the pay zones require fracture stimulation for eco- • Clean up any residual proppant in the coiled tubing work
nomic hydrocarbon production. A combination of economics, string.
operational and reservoir management considerations dic- • For multistage treatments, repeat the steps for isolation,
tates the stimulation strategy for these multizone completions. fracturing and proppant cleaning for each perforated inter-
A common approach for these conditions is to stimulate as val within the wellbore.
many producing intervals as possible with one fracture, either
through height propagation of a single hydraulic fracture or Coiled tubing fracturing operations rely on two technological
with the limited-entry technique (see Sidebar 11C). The effec- developments:
tiveness of these approaches is reduced because of the
uncertainty in predicting the fracture height and the inability of • bottomhole assembly (BHA) that enables zonal isolation
perforations to provide adequate diversion during slurry injec- and cleanup
tion, respectively. Similarly, multistage treatments that isolate • fracturing fluid that reduces the friction pressure so that the
and stimulate each pay interval require additional workover injection rates required for hydraulic fracturing may be
operations, such as those discussed in Section 11-2.4. These attained within the significantly smaller cross-sectional area
operations reduce the economic benefits of the stimulation of coiled tubing.
program, often to the extent that fracture stimulation can no
longer be justified for marginally producing intervals. Con- Zemlak et al. (1999) described a straddle BHA that is well-
sequently, several pay zones may remain bypassed for hydro- suited for the objectives of coiled tubing fracturing. The elas-
carbon production, either inadvertently because of the uncer- tomer-based packers perform in a manner similar to conven-
tainties and limitations of the fracture design process or tional swab cups. The packers also enable injecting fluid
deliberately because of economics. within the annular region and flowing it back through the
The primary reason that multizone reservoirs are ineffec- coiled tubing without requiring intermediate mechanical opera-
tively stimulated is the inability of conventional fracturing oper- tions. This reverse circulation procedure removes any residual
ations to control the placement of a hydraulic fracture. This proppant from the coiled tubing, which can then be raised and
limitation can be overcome by using a coiled tubing work positioned against the next perforation interval for its fracture
string for slurry transport within the wellbore. Its inclusion stimulation. The effectiveness of the isolation and cleanup
adds versatility to the fracturing operation wherein the mobility operations has been demonstrated by several successful frac-
of the coiled tubing can be applied to individually fracture turing operations for a range of reservoir depths, wellbore
stimulate each producing interval within the wellbore in a sin- sizes and other operating conditions.
gle coiled tubing trip. This operational modification, however, Most of the friction constraints of coiled tubing fracturing
does not limit the effectiveness of the fracture treatment. are overcome by incorporating a viscoelastic surfactant (VES)
Coiled tubing deployment provides the desired fracture pene- fracturing fluid. VES fluids exhibit a friction pressure that is as
tration and conductivity that would be expected with conven- low as 1⁄2 to 1⁄3 that of the polymer fluid with the least friction
tional multistage fracture treatments. pressure, delayed crosslinked, low-guar borate fluid (Fig.
The economic advantages that coiled tubing fracturing 11B-2). The nondamaging nature of VES fluids provides the
brings to multistage fracture treatments can also be realized additional advantage of ensuring that the desired fracture
through the fracture stimulation of previously bypassed pay conductivity is attained at lower proppant concentrations. This
zones (Olejniczak et al., 1999). In this application, coiled tub- capability further benefits coiled tubing fracturing because the
ing fracturing prolongs the economic life of the wellbore and slurry friction exponentially increases over that of the base
therefore increases the available hydrocarbon reserves. fluid with increasing proppant concentration. The friction-
The general operational procedure for coiled tubing frac- reducing, nondamaging properties of VES fluids are primarily
turing treatments is illustrated in Fig. 11B-1. The application responsible for making coiled tubing fracturing a feasible
methodology is as follows: operation for typical petroleum reservoir depths.
For both multistage fracturing treatments and the fracture
• Perforate the pay intervals targeted for fracture stimulation stimulation of single-zone bypassed pay intervals, the coiled
1 day before the scheduled fracturing operation. tubing work string protects the wellbore tubulars from exces-
• Run a gauge ring to ensure that there are no restrictions in sive pressures encountered during fracturing. Coiled tubing
the wellbore caused by scale deposition, damaged tubulars deployment may achieve fracture stimulation where the com-
or perforation burrs. promised nature of the existing wellbore would otherwise pre-
vent effective fracture stimulation.
10,000
Pay zone 1
Delayed
Rate (bbl/min)
low-guar
borate
1000
Pay zone 2
VES
Pay zone 3
100
1 10 100
Friction pressure (psi/1000 ft)
Figure 11B-1. Coiled tubing fracturing procedure.
Figure 11B-2. Friction pressure of viscoelastic surfac-
tant and delayed low-guar fracturing fluids.

be limited to the manufacturer’s specification. The cosity and create an equilibrium proppant bank.
use of baffles in pipe sizes greater than 51⁄2 in. is not With this type of proppant transport, the proppant
recommended because the total force applied to the bank is not destroyed during the overflush of the
cross-sectional area of the baffle may exceed the diverter.
baffle strength. Using conventional ball sealers to divert frac-
• Bridge plugs and packers turing stages has many of the same inadequacies
as bridging materials. It is extremely difficult to pre-
When completion practices do not allow the pro-
dict the seating efficiency of ball sealers. This prob-
gressive order of fracturing to proceed from the
lem is even more pronounced after proppant has
lowest zone of interest up to higher intervals,
eroded the perforations. Also, the ball sealers must
bridge plugs must be used in conjunction with
be introduced to the fluid while proppant is being
packers. Using the combination of a bridge plug
added. The presence of proppant reduces the seating
and packer to straddle an interval provides an
efficiencies, but it is impossible to predict by how
extremely reliable method of isolation. These
much. If the ball sealers are introduced to a clean
retrievable tools can easily be moved to cover any
fluid stage immediately following the proppant
interval, provided the unperforated casing is suffi-
stages, the clean fluid overflushes proppant away
ciently long to provide a packer seat. However,
from the perforations until the balls finally seat.
caution should be taken when open perforations are
Designing a schedule that ensures precise prop-
present above a packer because of the possibility of
pant placement into multiple zones by using
proppant entering into the annular area if the frac-
diverter stages is almost impossible. When several
ture reaches the open perforations. Small quantities
zones are open to the wellbore it is extremely diffi-
of proppant on top of a retrievable packer can stick
cult to calculate which zone will fracture first. The
the tool string.
zones are almost surely different in size and have
• Diversion slightly different rock properties. Because it is not
Diverting techniques are used in some cases to con- practically possible to know which zone will frac-
trol the placement of fluid and slurry into the zones ture at a given time, most hydraulic fracturing
of interest. A diversion treatment is advantageous schedules using diverters are simply divided into
over separately isolating individual zones because even stages. The uneven in-situ parameters cause
all the treatments can be pumped continually and slurry placement in the separate zones to create
are therefore more economical and time efficient. fractures of uneven geometries and conductivities.
Although initially attractive, controlling placement It is also difficult to design and size the diverter
through diversion carries many inherent risks. At stage so that all perforations in the zone being frac-
a minimum, drilling damage can be bypassed; how- tured become plugged and the other zones remain
ever, optimum stimulation of the formation cannot unaffected by the diverter. A diverter stage that is
be achieved using diverting materials. too large may plug the unfractured intervals before
The use of bridging materials, such as rock salt the fracturing slurry designed for that stage has
and benzoic acid flakes, as the diverting medium been pumped. If the diverter stage is too small, the
usually results in an overflushed fracture. Some first zone may not be adequately plugged and the
of the bridging material enters the fracture and dis- original fracture may continue to accept fluid.
places the near-wellbore proppant before diversion Portions of the pad fluid intended for the second
is achieved at the perforations. Conductivity near interval will overflush the proppant pack away
the wellbore may be minimized, resulting in a from the immediate wellbore. At the same time,
choked fracture with limited production capabilities. the second zone is losing critical volumes of pad
This problem is pronounced when high-viscosity fluid, which may result in an early screenout.
crosslinked fluids are used for fracturing. These • Limited entry
fluids are highly efficient at proppant transport and
Limited-entry treatments are designed to place frac-
carry the proppant away from the perforations as
turing fluids into multiple zones simultaneously.
they are displaced by the diverter slurry. The
The limited-entry technique uses the pressure drop
diverter overflush may not be a significant problem
created across the perforations during pumping to
if the fluids used in the treatment are of low vis-

divert the fracturing fluid into several different per- 500 psi in net pressure. An imbalance in net pres-
forated intervals (see Sidebar 11C and Chapter 10). sures between zones can effectively negate the per-
Generally, a 500- to 1000-psi pressure differential foration pressure drop. Fracture height and Young’s
is considered necessary to provide adequate control modulus are two parameters with a major effect
over fluid placement. on net pressure. Both parameters should be closely
The total flow of fluid entering a given zone is evaluated prior to the design of a limited-entry
restricted by controlling the size and number of treatment.
perforations in that interval. The high pressure drop The net pressure in the fracture is inversely pro-
at the perforations forces fluid to go to another portional to the gross fracture height. Large zones
zone. This diversion technique has proved popular have smaller net pressures and therefore tend to
because of its simplicity and economics. The diver- accept a disproportionate amount of fracturing
sion does not require expensive tools, and it does fluid. Very small zones most likely remain unstimu-
not require running and retrieving tools or making lated because they rapidly build high net pressures
cleanout trips. The only cost for applying this type and do not accept significant volumes of fracturing
of diversion is the excess hydraulic horsepower fluids. The global Young’s modulus of the zone
required to pump the treatment at higher pressure. has a similar effect: the larger the Young’s modu-
If a limited-entry treatment is not applied cor- lus, the narrower the fracture and the higher the
rectly, each producing zone may not receive ade- net pressure.
quate treatment. Several factors must be considered The final parameter with significant impact on
when designing a limited-entry treatment. The the successful placement of fractures using the
number and size of perforations are calculated to limited-entry technique is fluid leakoff. The size
divert the pad fluid. Smaller zones do not require of the zone and the rate the fluid is pumped into
as much fluid or proppant and therefore require the zone directly impact the fluid efficiency and
fewer perforations. Some zones may require fewer hence the fracture penetration. With several zones
than five perforations to control flow into that sec- accepting fluid at one time, the total pump rate into
tion. With a limited number of perforations avail- any one interval may be quite low. Zones with the
able, the importance of the breakdown procedure lowest pump rates generally have poor fluid effi-
becomes obvious. The loss of one or two perfora- ciency, which may result in an early screenout.
tions can significantly alter the flow distribution Accurately placing proppant into multiple zones
into all the zones. by using limited entry is extremely difficult. Frac-
Introducing sand into the fracturing fluid quickly ture penetration and width are most likely highly
erodes the perforations and changes the corre- irregular among zones. Smaller zones may not
sponding flow coefficient for each perforation. accept any significant amount of fluid. The
After only 10,000 lbm of proppant, the pressure increased producing capabilities of several stimu-
drop across the perforations will be greatly reduced. lated zones resulting from isolation and separate
Therefore, diversion of the pad fluids may be suc- treatments should be carefully examined and
cessful, but diversion of the proppant-laden stages weighed against the economic advantages of a
cannot be presumed to be successful. After the per- limited-entry treatment before the fracturing strat-
forations have been eroded, one zone is most likely egy is selected.
to accept most of the fluid.
An accurate stress profile of the wellbore is neces-
sary to design a successful limited-entry treatment. 11-3. Perforating
Each zone has a different fracture gradient and there-
fore they break down and fracture at different pres- 11-3.1. Background
sures. If great contrast exists among the fracture gra-
Perforating provides the means of communication
dients of individual zones, the perforation scheme
between the wellbore and the reservoir, and during a
must be designed to reflect this difference.
stimulation treatment, the perforation is the fluid con-
Limited-entry designs usually do not consider the
duit between the fracture and the wellbore (see the
net pressure effects of the fracture. It is not uncom-
Appendix to this chapter). Within this section, frac-
mon for a fracturing treatment to create more than

11C. Estimating multizone injection profiles Nolte (1991) expanded this expression to account for fluid loss
during hydraulic fracturing and non-Newtonian fluids. It can be shown that the injection
rate qi is proportional to the formation and fracture parameters:
[h (E ′ η )],
J. L. Elbel, Schlumberger Dowell
q i ∝ ( pf − σ min )
( 2 n ′+ 3 ) / ( n ′+1) 1/ ( n +1)
f
3 2
(11C-2)
Wells completed in low-permeability formations typically have
a number of layers scattered over a few hundred feet open to where n´ is the power law exponent and η is the efficiency.
the wellbore. This completion is necessary to make the wells Assuming that injection into each layer begins at the same
economical where no individual layer has sufficient production time, the ratio of qiA /qiB at any time can be expressed as
to warrant individual completion. These wells usually require
fracturing, and obtaining the desired stimulation for each layer ( 2 n ′+ 3 ) / ( n ′+1) 3 2 1/ ( n ′+1)
q iA  p f − σ minA   hfA   E B′   ηB 
is difficult. The best way to ensure adequate stimulation of each = h  E′   η  . (11C-3)
layer is to isolate and treat each layer individually; however, this q iB  p f − σ minB   fB   A   A 
technique may be uneconomical considering the production
potential of the individual layers. A proper economic analysis is The efficiency η can be determined by
impossible unless the injection rate and proppant penetration of
each layer can be determined (see Section 10-5). Fb ( pf − σ min )hf2 / E ′
η= , (11C-4)
A common means of stimulating a number of layers is the Fb ( pf − σ min )hf2 / E ′ + h (1.9CLt 1/ 2 + 0.0017VS )
limited-entry technique of perforating (Lagrone and Rasmussen,
1963). This method relies on high perforation friction, because which is of particular significance because it can be determined
of the limited number of perforations, to ensure that the well- independently of the injection rate. Equation 11C-4 requires the
bore pressure during injection is greater than the highest clo- properties used in Eq. 11C-3 plus the parameters of the ratio of
sure stress for all the layers. The number of perforations the average to maximum width Fb, fluid-loss height h, fluid-
assigned to each layer is prorated in an attempt to distribute leakoff coefficient CL and spurt loss VS.
the injection into each layer at some desired rate. If the stress Equation 11C-3 shows the relative importance of the various
contrasts are low, it is usually assumed that the rate into each parameter ratios; the gross height ratio is the most significant.
zone is governed by the number of perforations allocated to The desired injection rates for zones are commonly normalized
that zone. This assumption is false because other formation by the zone height. For the same qi /hf, the relation becomes
parameters also affect the injection rate into each zone. These
parameters are the fracture height, formation modulus and cur-
rent efficiency, which depends on the height, modulus, fracture (q i / hf ) A  p − σ minA 
= f
( 2 n ′+ 3 ) / ( n ′+1) 2
 hfA   E B′   ηB 
2 1/ ( n ′+1)
(11C-5)
 h  E′   η  ,
fluid-loss rate and fluid-leakoff height. They affect the net pres- (q i / hf )B  p f − σ minB   fB   A   A 
sure, which in turn must be added to the layer’s closure stress
when the desired perforation pressure drop is calculated. Even and for this condition, the net pressure ratio becomes more sig-
if one zone initially accepts all the fluid, a second zone may nificant than hf . The net pressure ratio can be controlled to
begin to accept a portion of the treatment later in the proce- some extent by changing the perforation pressure drop with the
dure. As net pressure in the fracture increases, so does the limited-entry technique. Increasing the perforation friction in one
pressure in the wellbore. At some point the net pressure layer will increase the net pressure in the other layer.
becomes sufficiently high to allow a second fracture system to For more than two layers, the ratio of the injection rate of an
accept fluid. If this scenario happens during a pad stage, prop- individual layer qii to the total rate qij for N layers is
[h (E ′ η )] ,
pant placement may be unsuccessful because of insufficient
pad. If a second zone opens during the proppant stages, the q ii (p f − σ min )i (
2 n ′+ 3 ) / ( n ′+1)
f
3 2 1/ ( n +1)
second zone will quickly screen out. Fracture initiation in the = i

(11C-6)
[h (E ′ η )]
N
∑ (p − σ min ) j ( 2n ′+ 3 ) / (n ′+1)
screened-out second zone may not occur, even when a new q ij 3 2 1/ ( n +1)
f f
pad fluid is started after the diverter stage. j =1
j
Using diverter stages to control the placement of fracturing

fluids usually results in uneven fracture geometry, poor conduc- where pfi = pwi – ppfi and pfj = pwi + ∆ph – ppfj (pw is the pressure
tivity near the wellbore and overall poor well performance. in the cased wellbore, ppf is the perforation friction pressure,
Relying on diverters to place multiple fracturing stages should and ph is the hydrostatic pressure). The perforation friction pres-
be avoided unless other isolation methods are not practically sure is estimated with
feasible. q i2 ρ s
An expression for determining injection ratios during the frac- p pi = 2
, (11C-7)
0.323n perfi d pfi4
turing of individually confined layers assuming PKN geometry
follows. This expression is combined with previously published where ρs is the slurry density, nperf is the number of perforations,
equations for determining fracture geometry, efficiency, pad vol- and dpf is the perforation diameter. The casing friction pressure
ume and perforation friction to allow simulation of the fracture between the layers can be neglected for most cases, where it is
geometry for any number of layers (see Section 6-10). small compared with the net pressures in the fractures.
Parameter relations Equations 11C-1 through 11C-6 are limited to constant-height
fractures; however, it has been reported (Nolte, 1982, 1991)
The relation of the net pressure (pf – σmin) to the parameters that if the net pressure in the fracture is less than 80% of the
used in simulating the fracture geometry is stress contrast of the barrier and growth into the barrier is a
distance equal to the zone height, then the fracture penetration
pf − σ min = 4.46(µE ′ 4q i2t / hf6 ) ,
15
(11C-1) and net pressure are reduced by only about 10%. This small
variation would not change the preceding equations signifi-
where pf is the fracture pressure, σmin is the minimum stress, µ cantly, and two-dimensional simulation would be within the
is the viscosity, E´ is the plane strain modulus, qi is the injection required engineering accuracy in cases with moderate height
rate, t is the time, and hf is the fracture height. growth.

turing implies using proppant; however, in general the more than 1000 psi. The existence of comminuted
presentation also applies to acid fracturing. The choice sand in the perforation cavity limits injectivity and
of the perforating parameters can have a significant increases the fracture initiation pressure. High pump
effect on the quality of the subsequent fracturing or rates and high fluid viscosity enhance these effects,
matrix-stimulated treatment (Daneshy, 1973; Warp- which are more important for extreme overbalance
inski, 1983). Perforating parameters are as follows: stimulation.
A microannulus is normally present after perforat-
• size and type of gun
ing, immediately after pumping begins or at both
• type of charge times. Maintaining a good bond during the breakdown
• shot density phase can be problematic because a hydraulically
• shot phasing propagated microannulus is analogous to hydraulic
• interval length fracturing, as discussed in Sidebar 11D. Fracturing
then proceeds as though from an openhole with some
• gun orientation. defects (perforations) that may be near the preferred
For the combination of gravel packing and fractur- hydraulic fracture plane (PFP). Most laboratory frac-
ing (frac and pack), the perforating practices are gov- turing studies take extraordinary measures to avoid a
erned by the gravel-packing considerations. These microannulus by epoxying the casing to the rock,
considerations are discussed at the end of this section. using O-rings around the perforations, etc. Thus, the
The objective of perforating for fracturing is to choose generality of these laboratory findings must be viewed
perforating parameters that minimize near-wellbore with caution. The magnitude of the microannulus is
pressure drops during both the fracturing operation dependent on the wellbore fluid, wellbore size and
and production. Some of these near-wellbore effects type of perforating gun (Table 11-1).
are perforation friction, microannulus pinch points Except when gas is the wellbore fluid, perforating
from gun phasing misalignment (Nolte, 1982, 1988a, debonds a portion of the cement/sandface hydraulic
1988c), multiple competing fractures and fracture tor- bond. This is a result of the loading of the wellbore
tuosity caused by a curved fracture path (Romero et fluid from the gun swell (charge/explosive coupling
al., 1995). These near-wellbore effects are discussed for capsule charges), passage of the perforating jet
in Chapter 6. For any type of well treatment, there are through the wellbore fluid and expulsion of the explo-
two additional perforation-related parameters that may sive detonation gases into the wellbore fluid. For hol-
also affect the choice of perforating system: low carrier guns, the debonding may be a function of
the gun phasing. Figures 11-1 and 11-2 are examples
• integrity of the cement/sandface hydraulic bond
of cement debonding observed in large-scale block
(microannulus) after perforating
tests (Mason et al., 1994; Behrmann and Elbel, 1992).
• residual fractured sand grains in the perforation An ideal perforation for fracture initiation would
cavity, particularly for a matrix treatment. have a minimum injection pressure drop, initiate only
Effective matrix treatments require communication a single bi-wing fracture and generate a fracture with
through most of the perforations. This can be achieved minimum tortuosity (turning from the initiated frac-
by effective underbalance (see Appendix to this chap-
ter), extreme overbalance (see Section 11-3.6) or the Table 11-1. Perforating parameters that
use of ball sealers. affect a microannulus.
If a reservoir is perforated with insufficient under-
Parameter Promotes Microannulus
balance to remove most of the perforation sand debris,
then fluid injection may cause this comminuted sand Capsule gun Yes
to create an external filter cake on the perforation cav- Hollow carrier gun Modestly
ity during fluid injection. This was first observed on a
Small gun-to-casing clearance Yes
water injector and later on extreme overbalance tests
(Behrmann and McDonald, 1996). Two unique char- Liquid in wellbore Yes
acteristics observed in these tests are that productivity Low shot density No
was not affected and that the “filter cake” was also a
Gas in wellbore No
pressure barrier with an estimated pressure drop of

11D. Propagating a microannulus during 2. The flaw initially extends in a radial geometry from the per-
formation breakdown foration tunnel. Actually, extending annular flaws would origi-
nate from most of the perforations. The mechanics describ-
K. G. Nolte, Schlumberger Dowell ing the flaw’s deformation are relatively complex during this
early stage. The complexity results from the multiple regions
For normal completion practices, the creation of a microannu- of localized pressure acting on the curved surfaces of the
lus should be anticipated during the breakdown process. The casing and borehole.
microannulus results from the same mechanics that govern the
propagation of a hydraulic fracture, but on a smaller scale and 3. After some period, the individual radial patterns coalesce
confined to the annular circumference of the cement’s inter- into one microannulus around the complete circumference
face with the well. The affected annular interface can be that of the cement sheath. At this stage, the annulus is analo-
of either the cement/casing or the cement/formation. The for- gous to a confined-height PKN fracture (i.e., height equals
mation interface is more prospective because of the mudcake the circumference of the wellbore) and can begin to extend
remaining between the cement and formation. up, down or both directions along the wellbore until break-
A hydraulic fracture or microannulus can propagate when down of the formation by a hydraulic fracture.
fluid of sufficient pressure energizes a prospective flaw and the For the last stage, the mechanics governing the annular
flaw is embedded in a deformable environment. Flaws of these width in the perforated section become relatively simple. The
types exist in the cement interfaces and around perforation simplicity comes from several sources. The expressions for the
tunnels, and they are in communication with the wellbore fluid. change in radii of the casing and wellbore can be determined
These flaws can originate in the mudcake, which can rehy- from specialized cases for the elastic deformation of a thick-
drate by capillary action from the wellbore fluid, or in a region walled cylinder. The multiple-connecting perforations provide
of mechanical alteration around the perforation tunnel. As the fluid with minimal pressure gradients in the annulus. As a
wellbore is pressurized during breakdown, the fluid in the result, the pressures inside and outside of the casing become
prospective flaw is also pressurized, increasing the width of essentially equal with no radial change of the casing, and the
the flaw by compressing the material surrounding the flaw and annular width depends only on the change in radius for the
allowing more fluid to enter and extend the flaw. For the borehole:
microannulus, the relevant surrounding material is the rock
containing the wellbore and the casing confining the cement.
Increased fluid pressure in the annulus compresses the rock ( )
w = (1+ ν)D p m − p p / 2E , (11D-1)
and enlarges the wellbore radius. Similarly, the casing radius
changes as the internal pressure and the external annular where w is the microannulus width in in., ν is Poisson’s ratio,
pressure change. The deformation of the cement sheath is rel- D is the wellbore diameter in in., E is Young’s modulus in psi,
atively small and could be considered part of the rock contain- pm is the microannulus pressure in psi, and pp is the far-field
ing the wellbore. The annular width results from the combined reservoir pressure in psi.
radial deformation of the borehole and casing. As an example, consider Eq. 11D-1 for a bottomhole pres-
The evolution of the flaw’s geometry is similar to that sure increase of 2000 psi over the initial hydrostatic pressure,
described in Chapter 9 for a hydraulic fracture originating from borehole diameter of 7 in., rock modulus of 1E+6 psi and
a point source of injected fluid. Poisson’s ratio of 0.2. For these conditions, the hole radius
and microannulus width will increase by about 0.009 in. (or the
1. The flaw opens and propagates when the energizing pres- thickness of four sheets of standard writing paper). The
sure exceeds a closure pressure equaling the stress tending hydraulic conductivity of such a microannulus is significant
to close the flaw. Sidebar 11A implies that for a typical when coupled with a large pressure differential (e.g., 2000 psi)
cement composition and prior to breakdown, the stress in and a low-viscosity fluid (e.g., <0.4 cp for completion brine at
the cement sheath is essentially hydrostatic and about equal representative bottomhole temperatures).
to the reservoir pressure of the formation. Therefore, for The conductivity microannulus created during the break-
normally pressured conditions, the flaw’s closure pressure down pressurization provides the same pressure and fracture
is about the same as the hydrostatic pressure of the com- initiation environment as an openhole, and for a vertical well
pletion fluid within the casing, and the flaw can begin to it enhances the creation of a single fracture in the preferred
propagate as breakdown begins. fracture plane.
ture into the PFP) at an achievable fracture initiation a significant distance for a vertical well. When the
pressure. The following sections provide recommen- PFP is not vertical, as can occur near a significant
dations on how best to achieve this ideal perforation. fault, the PFP deviates from the axis of a vertical
In the following sections, a vertical well is one with a well. For this case, “Open microannulus, deviated
deviation less than 30°. and horizontal wellbores/vertical fractures” subse-
quently in this section should be consulted.
Dry gas wells, wells swabbed of liquid prior to
11-3.2. Perforation phasing for hard-rock perforating, wells shot with small hollow carrier
hydraulic fracturing guns and wells shot with low-shot-density (1 or
• No microannulus, vertical wellbore 2 shots per foot [spf]) hollow carrier guns are
potential candidates for maintaining good cement/
For the following discussion, the PFP is assumed to
sandface bonds. With a perfect cement bond, frac-
be vertical and therefore can connect directly over
tures are forced to initiate from the perforations,

which should eliminate additional fracture initiation

sites around the sandface. Most laboratory
hydraulic fracturing experiments have been con-
ducted with a sealed annulus with the only fluid
entry to the reservoir through artificial (drilled or
cast) perforations. Thus, these experiments should
provide insight on how to perforate wells with a
good hydraulic bond, but the results must be used
with caution because a microannulus is possible
after perforating or during breakdown of the
formation.
One of the papers describing laboratory fracture
experiments (Abass et al., 1994) shows that 180°
phased perforations oriented within 30° of the
PFP provide good communication between the
perforations and the fracture. The good connection
minimizes multiple fracture overlap and turning
tortuosity and therefore minimizes any restriction
of fracture width. As the perforation-to-PFP angle
increases, the fracture initiation pressure increases
Figure 11-1. Disruption of the cement/sandface bond as a result of the horizontal stress difference. Also,
(underbalance test) (Mason et al., 1994). Intact cement
sections are light gray. when the fracture initiates at the perforations, it
must turn to eventually align with the PFP, and
the near-wellbore fracture width decreases. This
work suggests that if a 180° phased gun cannot
be oriented within 30° of the PFP, then the use of
a 60° phased gun is recommended for a good frac-
ture connection. It is assumed that only those perfo-
rations closest to the PFP will initiate a fracture and
that the shot density of the 60° phased gun must be
3 times that of a 180° phased gun to achieve the
same number of holes directly linked to the frac-
ture. These assumptions also imply that multiple
parallel fractures will not initiate. However, the
increased shot density of the 60° phased gun proba-
bly negates the assumption of no microannulus.
Assuming that equal perforation areas open in
direct communication with the primary fracture,
Table 11-2 lists the trade-offs for different non-
oriented phased guns.
Selection of the optimum gun depends on assign-
ing weighting factors to the different parameters,
which becomes subjective and dependent on per-
sonal experience. For equal weighting, all guns are
equivalent; however, to minimize the initiation of
Figure 11-2. Disruption of the cement/sandface bond (frac- multiple fractures and if the possibility of higher
ture test) (Behrmann and Elbel, 1992).The dark areas are
fracturing fluid that flowed around the wellbore annulus.
fracture initiation pressure is acceptable, then a
180° phased gun can be used. In all cases, the cas-
ing hole diameter should be chosen to provide an

Table 11-2. Perforating gun trade-offs for a vertical well, no microannulus and not oriented.
Gun Fracture Initiation Multiple Fracture Tortuosity Cement Bond

Pressure Initiation Destroyed
0°, 1 spf 3 1 3 1
180°, 1 spf 3 1 3 1
120°, 1.5 spf 2 2 2 2
60°, 3 spf 1 3 1 3
Note: 1 = best, 3 = worst
acceptable injection pressure drop (see Section of the perforations, unless the perforations are
11-3.3). Because injection rates are generally within about 10° of the PFP (Behrmann and Elbel,
greater than production rates and proppant enlarges 1992). Fractures can also be initiated from perfora-
the perforations and erodes near-wellbore restrictions that are within about 30° of the PFP. Multiple
tions, the production area open to flow should gen- fractures are encouraged from perforations between
erally be adequate. 10° and 30° from the PFP if sufficient fluid is
The literature indicates that other perforating allowed to move in the microannulus; however,
strategies can be applied. For example, Stadulis maintaining significant flow, and hence width, in
(1995) discussed the use of 0° phased guns at 1 spf more than one fracture is inherently unstable
with proppant slugs to prevent the initiation and because of the increased pressure requirement. The
propagation of competing multiple fractures and pressure drop across the multiple fractures
near-wellbore screenouts. (It is not clear if the increases by the square root of the number of frac-
observed success was from the use of low shot tures (Nolte, 1987).
density, closed microannulus or proppant slugs.) When the fracture does not originate from the
It is not known if a dominant single- or bi-wing perforations, the flow path connects through the
fracture propagates with 0° phased guns. A bi-wing microannulus. The original microannulus separates
fracture must initiate, but the wing opposite the further from the sandface to allow displacement
perforations can have a limited flow rate and may continuity with the fracture width; however, geo-
screen out as a result of the restricted flow around metric effects result in pinch points at the fracture
the microannulus. The asymmetry for a dominant entrance that can cause large pressure drops for
single-wing fracture offsets the drainage pattern fluid flow and near-wellbore proppant bridging.
from the well location. To minimize multiple frac- These points are subject to enhanced erosion, with
tures when using 0° phasing, the lower shot density their endurance depending on the rock hardness.
helps maintain the cement/sandface integrity; Fracture tortuosity should not exist for a vertical
whereas the use of a 60° phased gun at 6 spf (pro- wellbore in a normally stressed environment.
viding the same 1-ft spacing between perforations Table 11-3 lists the fracture trade-offs for differ-
along any azimuthal plane) is more detrimental to ent gun phasings for a normally stressed vertical
cement debonding and increases the potential for well (deviation less than 30°). The shot densities
the initiation of multiple fractures. in Table 11-3 are relative only. Determination of the
Because deviated (deviation greater than 30°) required shot density is provided in Section 11-3.3.
and horizontal wells typically have an open For equal weighting, either 120° or 60° phased
microannulus because of gravity (i.e., independent guns should be used. If pinch points are a greater
of the perforator), the perforating requirements for concern than multiple fractures, a 60° phased gun
these wells are discussed in the following section. should be used; however, the 60° phased gun may
• Open microannulus, vertical wellbore/vertical potentially create more multiple fractures and
fractures requires twice the shot density of a 120° phased
gun because only one of three perforations will
The presence of a microannulus (see Sidebar 11D)
connect to the fracture.
promotes fractures from the sandface, independent

Table 11-3. Perforating gun trade-offs for a vertical well, with microannulus and not oriented.
Gun Fracture Initiation Pressure Microannulus Pinch Points Multiple Fracture Initiation
0°, 1 spf 3 4 1
180°, 1 spf 3 3 2
90°, 2 spf 2 3 3
120°, 1.5 spf 1 2 3
60°, 3 spf 1 1 4
Note: 1 = best, 4 = worst
Other strategies are reported in the literature. 10-ft perforated intervals would be reasonable to
For example, modifications of the pad with a high minimize the initiation of nonlinking multiple frac-
pump rate, use of high-viscosity fluid and use of tures. When the PFP approaches a 90° intersection
proppant slugs (Aud, 1994; Cleary et al., 1993; with the wellbore for a wellbore deviation greater
Stadulis, 1995) have been used to theoretically con- than about 75° (horizontal well), perforations
trol near-wellbore screenouts by restricting fluid should be clustered in a short length of less than
communication around the microannulus to reduce 3 ft with maximum shot density and multiple phase
pinch points, tortuosity and multiple fractures. angles to maximize perforation communication
• Open microannulus, deviated and horizontal with the fracture (Abass et al., 1992, 1994). This
wellbores/vertical fractures extremely limited interval, with sufficient zonal
isolation from the cement, enhances the propagation
The desired fracture geometry for an arbitrarily ori-
of only one dominant fracture. Staged multiple
ented deviated well is to initiate a single bi-wing
fractures have been successful in horizontal wells
fracture along the perforated length of the wellbore
drilled perpendicular to the PFP (Baumgartner et
that then gradually turns into the PFP. If the well-
al., 1993; Chambers et al., 1995; Abou-Sayed et
bore is in the PFP, then the fracture will initiate
al., 1995).
from the perforations at the top and/or bottom of
the casing, and thus 180° phased guns oriented up
and/or down, respectively, are recommended. The 11-3.3. Other perforating considerations
use of oriented 180° phased guns has successfully for fracturing
been used on deviations up to 65° (Pearson et al.,
1992; Pospisil and Pearson, 1995; Vincent and • Penetration depth
Pearson, 1995). The guns were aligned in the plane Perforation penetration beyond 4 to 6 in. into the
of minimum tangential compressive stress (Yew formation is not required for fracturing because
and Li, 1988; Yew et al., 1989, 1993). Recent labo- fracture initiation from a perforation generally
ratory experiments by van de Ketterij (1996) con- begins near the sandface and propagates toward the
firm these field observations. If the stress direction preferred fracture plane (Behrmann and Elbel, 1992).
is not known, then a vertical, up/down, orientation Gun performance for penetration should be com-
is suggested. See Section 11-3.3 to ensure that the promised in favor of casing hole size. Size require-
casing hole diameter on top meets the required size. ments have been adopted from gravel packing
Wellbore rotation azimuthally around the PFP (Gruesbeck and Collins, 1982) and are discussed
causes the length between the PFP and the wellbore in Chapter 5. The general requirement is that the
to decrease, with a minimum occurring at a 90° minimum casing hole diameter exceeds 6 times the
rotation (i.e., the PFP and the plane through the proppant diameter. A ratio of 8 to 10 times larger
top–bottom of the wellbore are at 90°). The perfo- than the average proppant diameter should be gen-
rated interval should be continually decreased as erally used because of variance between the nomi-
the combination of well deviation and azimuth nal and actual hole diameters, gun positioning and
becomes less favorable and decreases the length variation in proppant size. Manufactured proppant
of this intersection. For the most extreme case, is highly biased toward larger diameters (lower

mesh range) to maximize its permeability. For plus an additional pressure drop for crosslinked
manufactured proppants, the minimum casing hole gels. See Shah et al. (1996) for further correlations
diameter should be sized for the lower mesh size with the viscosity for linear polymer solutions,
(e.g., 16 mesh for 16/30 mesh proppant). crosslinked gels and fracturing slurries. Figure 11-3
• Perforated interval shows the injection pressure drop versus casing
hole diameter for water, where Cd is
Limiting the perforated interval was previously
discussed for deviated wells. Even when the perfo-
( )
0.4
Cd = 1 − e −2.2 D/ µ
0.1
, (11-3)
rated portion of the well is nominally aligned with
the PFP, consideration should be given to limiting where µ is the apparent viscosity in cp.
the perforated interval, particularly for relatively
thick sections that probably will be covered by the
propped fracture. For example, a 6° deviation 10
Flow rate/perforation (bbl/min)

between the well and PFP over 100 ft provides 9 25-psi pressure drop
50-psi pressure drop
a 10-ft offset and the potential for more than one 8
100-psi pressure drop
7
dominant fracture. Multiple fractures are detrimental 200-psi pressure drop
6
when they overlap and decrease their width in the 5
overlapped region, which is usually the center of 4
the perforated zone. Assuming vertical coverage of 3
2
a zone by the propped fracture, the limiting effects
1
for reducing the interval are achieving sufficient 0
hole density and the resulting converging flow for 0 0.2 0.4 0.6 0.8 1.0
the subsequent production (see Chapter 10). Perforation diameter (in.)
Another consideration for limiting the perforated
section near the center of a zone is to assist the ver- Figure 11-3. Injection rate versus perforation diameter.
tical confinement of a tip-screenout (TSO) treat-
ment. The limited section increases the vertical Unless a perforating gun is centralized, the perfo-
exposure of the slurry to fluid loss, which increases ration casing hole diameter is a function of gun
the concentration and promotes competent bridging phasing. This means that the injection rate is differ-
during the increased pressure portion of the TSO. ent for different perforation diameters. For example,
• Large stress contrasts a cross-casing perforation diameter equal to 0.7 of
Large horizontal stress contrasts favor using 60° the near-casing perforation diameter has 0.49 times
phased guns to minimize the perforation-to-PFP the injection rate of the near-casing perforation. An
alignment. Lack of alignment increases the fracture average perforation diameter can be calculated
initiation pressure and enhances the microannulus- using
pinching effect. 0.5
 i=n 
• Shot density and hole diameter D = ∑ Di2 n  , (11-4)
 i =1 
Minimum shot density is determined by the diame-
ter of the perforation casing hole, injection rate per where n is the number of effective casing holes.
perforation, desired perforation friction and fluid Figure 11-4 provides the typical hole size varia-
properties. The perforation friction ∆ppf in psi for tion caused by the gun-to-casing clearance. For
noncrosslinked fluids is specific gun and casing systems, the service com-
pany should provide data on the variation in casing
[
∆p pf = 0.237ρ qi (Cd × D2 ) ,]
2
(11-2) entrance hole size. Also, during the fracture treat-
ment the hole entrance becomes rounded, increasing
where ρ is fluid density in lbm/gal, qi is injection
the discharge coefficient, and the hole size may also
rate in bbl/min/perforation, Cd is the dimensionless
increase from erosion by the proppant. For further
discharge coefficient, and D is the perforation cas-
discussion, see Chapter 6 and Shah et al. (1996).
ing diameter in in. Lord (1994) provided tables of
The hydraulic horsepower and surface treating
Cd for different perforation sizes and fluid types
pressure limit determine the maximum permissible

an IGP would give the maximum area open to flow

0.7 through the casing with the minimum hole diameter
0.6 required for gravel placement. A gun with shots
Casing hole diameter (in.)
phased every 60° or 45° is desired. Depending on the

0.5
expected flow rate per perforation (i.e., the required
0.4 minimum pressure drop), guns using big hole charges
at 12, 16 and 21 spf could be used. Because of the
0.3
low modulus and strength of sand-producing forma-
0.2 tions, large fracture widths with a correspondingly
0.1 21⁄2-in. deep penetrator large displacement of the microannulus are created
2 1⁄2-in. big hole
3 3⁄8-in. deep penetrator
to minimize pinch points.
0
0 0.5 1.0 1.5 2.0 A treatment design and execution objective should
Water clearance (in.) be a successful TSO fracture that packs back into the
gravel pack to ensure that the large microannulus is
Figure 11-4. Casing hole diameter versus water clearance. also packed and creates an external gravel pack. The
external pack provides a highly conductive path
between the fracture and the perforations, which is
treating rate. The number of perforations in contact a primary benefit of frac and pack operations. The
with the fracture determines the average injection issues of multiple fractures and tortuosity do not arise
rate per perforation. For 0° and 180° phased guns, because of the erosive nature of the pumped fluid and
all perforations should contribute to the fracture. gravel on the weak formation. To minimize the likeli-
For a 120° phased gun, only two-thirds of the hood of a void within the IGP, the frac and pack inter-
perforations will probably communicate with the val should not exceed about 100 ft. The use of alter-
fracture, and for a 60° phased gun, only one-third nate path screens extends the frac and pack interval to
of the perforations will likely be effective. Sidebar hundreds of feet (Jones et al., 1997). Consistent with
11E provides an example calculation to determine the goal of achieving a competent IGP if the fracture
the minimum required number of shots and thus placement is not completed successfully, standard
shots per foot for a given pump rate and perforated practice is to perforate the complete gravel-packed
interval. For an acceptable perforation friction, section.
the casing hole diameter and shot density can be
traded off for a given total injection rate. Except
for limited-entry treatments (discussed in Chapter 11-3.5.Fracturing for sand control
10), perforation friction should be minimized (e.g., without gravel-pack screens
25 psi) to reduce unnecessary fluid shear and prop-
Fracturing for sand control without gravel-pack
pant damage.
screens can be accomplished by pretreating the forma-
tion or post-treating the proppant with resin, pumping
11-3.4.Frac and packs and high-rate curable-resin-coated proppant or pumping chopped
water packs fibers with the proppant. Except for the resin pretreat-
ment, these techniques fix or control proppant flow-
Perforating requirements for frac and packs and high- back plus provide a filter to prevent sand production.
rate water packs are the same as for an internal gravel Because no gravel is in the casing/cement tunnel to
pack (IGP). This similarity ensures a good gravel pack restrict flow, the perforation requirements are different
if the planned fracture placement is not completed than for frac and pack operations. The perforating
successfully. Perforating requirements for gravel objective, beyond the hole size for the proppant, is
packs have been driven by the required minimum pro- to eliminate any nonessential perforations that could
duction pressure drop through the casing/cement tun- produce formation sand. Therefore, for all well devia-
nel that contains the packed gravel. As fines from the tions a limited perforated section (e.g., 20 ft or less)
perforation, formation or both move into this tunnel and either 0° or 180° phased guns, ideally aligned
during production, the gravel permeability is reduced with the PFP determined before the treatment, are
and the pressure drop increases. An optimum gun for

11E. Calculation of minimum shot density

for fracture stimulation Table 11E-1. Casing hole diameters.
Phase Entrance Hole (in.)

The following example illustrates the process and importance
of calculating the perforation friction pressure drop and using 21⁄2-in. BH 33⁄8-in. DP
actual downhole perforation casing hole diameters.
API data 0.61 0.40
Problem statement
0° 0.58 0.38
Given the maximum injection rate and length of the perforated
interval, calculate the required shot density in shots per foot 60° and 300° 0.58 0.37
(spf) for two-gun systems for both 180° and 60° phasings.
Calculate the pressure drop if entrance hole (EH) data from 120° and 240° 0.49 0.30
the American Petroleum Institute (API) are used instead of
downhole data. 180° 0.36 0.27
Maximum injection rate = 20 bbl/min
Perforated interval = 20 ft
Table 11E-2. Average casing hole diameters.
Gun 1 = 21⁄2-in. big hole (BH) with API EH = 0.61 in.
Gun 2 = 33⁄8-in. deep penetrator (DP) with API EH = 0.40 in. Phase Pair 21⁄2-in. BH (in.) 33⁄8-in. DP (in.)
Casing size = 51⁄2 in., 21 lbm/ft and Q125
0°/180° 0.482 0.33
Maximum perforation friction pressure drop ∆ppf = 25 psi
60°/240° 0.537 0.34
Solution
1. Obtain the perforated casing hole diameters versus phase
angle from the service company (Table 11E-1).
2. Calculate the average hole diameter for 180° and 60°
Table 11E-3. Minimum gun shot density.
phasings using Eq. 11-4. Because the effective shots are
always pairs of holes at 180° phasing, n = 2. For example, Guns and Spf Using Actual Minimum
the average hole diameter for the 21⁄2-in. BH for the 0°/180° Phase API EH Pressure Spf Using
phase is Drop Table 11E-2
Using EH and
[ ]
D = (0.58 2 + 0.36 2 ) / 2 ∆ppf = 25 psi
0.5
= 0.482 . API EH
(psi)
Table 11E-2 provides the average hole diameters. 21⁄2-in. BH, 180° 0.86 70.5 1.5
3. For both the 180° and 60° phased guns, use the average 2 ⁄2-in. BH, 60°
1
2.57 70.5 4.5
hole diameters for the 0°/180° pairs. Because the guns are
not oriented, it is not known if the preferred fracture plane 33⁄8-in. DP, 180° 2.18 58.5 3.4
will be closer to the 0°/180° or 60°/240° perforation pairs,
and thus a worst-case condition is used (i.e., smaller aver- 33⁄8-in. DP, 60° 6.53 58.5 10.0
age holes). Use Eq. 11-2 to calculate the average flow rate
per perforation: 0.70 bbl/min for EH = 0.482 in. and 0.30
bbl/min for EH = 0.33 in. The flow rates for the API holes
are 1.17 and 0.46 bbl/min, respectively, for EH = 0.61 and
= 0.40 in. Divide these flow rates into the total injection rate
of 20 bbl/min to obtain the minimum number of effective
holes: 29 for the 21⁄2-in. BH and 67 for the 33⁄8-in. DP. To
obtain the total number of holes, multiply the effective holes
by the phasing deficiency, which is 1 for 180°, 1.5 for 120°
and 3 for 60° phased guns, and then divide by the 20-ft per-
forated interval to obtain the minimum required shot density.
Table 11E-3 summarizes these final numbers.
4. Check the required shot density and phasing against the
available guns. Most 21⁄2- and 33⁄8-in. guns are built with
6 spf at 60° phasing and can be downloaded to 2 spf at
180° and 3 spf at 120°. The 10-spf 33⁄8-in. gun at 60° phas-
ing does not exist with the DP charge used in this exam-
ple; the maximum shot density is 6 spf. This would result
in a pressure drop of 70.8 psi. Other combinations of shot
density and phasing would be special orders requiring lead
time and planning.

recommended. As with a frac and pack, one of the • Residual “crushed” sand debris in the perforation
design and execution objectives should be to achieve tunnel restricts both fluid injection and pressuriza-
a TSO that packs back to fill the expanded microannu- tion of the perforation.
lus with proppant. For the screenless case, the prop- • Lack of a microannulus and the perforations not
pant is treated for flowback control to create a compe- aligned with the PFP initiate against a larger stress.
tent external pack that controls the formation sand.
Again, as a result of the weak rock, the pumped fluids These effects result in required pressures that are 2
wash away any near-wellbore restrictions. Zero- to 3 times greater than those in conventional hydraulic
degree phased guns should be used when the guns fracturing.
cannot be aligned with the PFP to eliminate nonessen- Because experiments show no microannulus effect
tial perforations. or parallel multiple fractures, only tortuosity must be
considered. For a vertical well, a misaligned 0° phased
gun is the least acceptable whereas 60° or 120° phased
11-3.6. Extreme overbalance stimulation guns provide the least tortuous path (Table 11-2).
However, Petitjean et al. (1995) reported that they
Extreme overbalance (EOB) has been defined as mitigated tortuosity by minimizing the use of liquid
either the application of a very high overbalance pres- to that necessary for fracture initiation and using nitro-
sure during the perforating process (called extreme gen gas to extend the fracture and erode any near-
overbalance perforating, or EOP) or the very high wellbore tortuosity. This approach should be applica-
pressure “surging” of existing perforations. Other ble for both vertical and deviated wells. Also, Snider
names for this procedure are rapid overpressured (1996) reported the use of a proppant carrier to erode
perforation extension and high-energy stimulation. near-wellbore tortuosity and improve the near-well-
EOB utilizes pressurized gas (usually nitrogen [N2]) bore fracture conductivity, but the fracture width is
to inject various fluid systems into the formation at generally insufficient for proppant entry (Petitjean
pressure gradients from 1.4 to 2.0 psi/ft. The primary et al., 1995).
objective is to create fractures either as a pre–hydraulic In general, perforating considerations for EOB in
fracture treatment or as a dynamic fluid diversion vertical wells are similar to those in “No microannulus,
(Handren et al., 1993; Dees and Handren, 1993). vertical wellbore” in Section 11-3.2. Similarly, for
Handren et al.’s (1993) early publication suggests deviated and horizontal wells, the perforating recom-
that effective multiple fractures were created from all mendations in “Open microannulus, vertical wellbore/
perforations. However, additional full-scale laboratory vertical fractures” in Section 11-3.2 can be used.
fracture initiation experiments (Behrmann and A detailed discussion of extreme overbalance perfo-
McDonald, 1996; Willson, 1995) plus field tests rating was provided by Behrmann et al. (1996).
(Snider, 1996) confirm that although fractures may
initiate from many perforations, only a single bi-wing
fracture is propagated from the perforations nearest 11-3.7. Well and fracture connectivity
the PFP. Furthermore, there is no evidence of the initi-
ation of parallel multiple fractures. All fractures initi- The tortuosity between the initiation of a fracture from
ate from the perforations, with the primary fracture a well and the plane of its ultimate propagation has
from the perforations nearest the PFP. It is assumed been recognized by many as an important issue during
that the sudden pressurization of the wellbore casing fracture execution, resulting at least in an undesirable
closes any microannulus prior to hydraulic communi- increase in fracturing pressure or, worse, in a screen-
cation away from the perforations. The high fracture out and even the creation of only one fracture wing.
pressure gradients required are a result of at least four Well deviation and unacceptable perforation phas-
events: ing have already been addressed in this chapter. Well
deviation can be remedied with an S-shaped well
• Dynamic fracture initiation is greater than static (i.e., turning a deviated well to become vertical upon
fracture initiation. entering the target formation), and problems with per-
• Near-wellbore pore pressure does not increase as foration phasing can be addressed by decreasing the
much as in a static injection. phase angle among the perforation planes. To prevent

the similarly undesirable problem of multiple fracture ing equipment for small and large fracturing treat-
initiation, the number of perforations can be reduced, ments, respectively.
leading to “point source” fracturing.
These solutions, although they usually prove suc-
cessful in fracture execution in all types of reservoirs 11-4.1. Wellhead isolation
and will not affect well performance in low-permeability Specialized isolation tools, or tree savers, can protect
formations, may prove inadequate in high-permeability a Christmas tree at the wellhead from damage and the
applications. possible failure that results from exposure to high
Chen and Economides (1998) demonstrated the pressure, corrosive fluids or abrasive proppant-laden
effect of the near-well choke, resulting from inade- fluids. The pressure rating of a wellhead is usually
quate fracture and well connectivity in low- and high- less than the pressure required to pump a stimulation
permeability reservoirs. Although the well perfor- treatment. Replacing the existing tree with one that
mance reduction from the ideal value in low-permeability has a higher pressure rating is expensive and requires
reservoirs is insignificant, the reduction in high-per- killing the well with potentially damaging fluids. Even
meability reservoirs can be substantial (50% to 75% if the existing tree has a sufficient pressure rating,
from the ideal) and the controlling factor in the suc- exposure to high pressures and treating fluids may
cess of the treatment. This effect suggests that for leave it in an unsafe condition.
high-permeability fracturing, S-shaped wells are The tree saver is mounted on the existing Christmas
highly desirable and that in treating these vertical tree. A mandrel is extended through the valves on the
holes the indicated perforation phasing is 180° but tree and into the tubing. The mandrel has a rubber cup
with the guns oriented to align the perforation plane assembly that seals to the walls of the tubing and pre-
with the fracture plane. This assumes, of course, that vents fluid or pressure from directly reaching the tree.
the fracture is vertical. Analogous ideas may apply to Once set, a tree saver can extend the working pressure
horizontal wells for shallow applications where the of a wellhead up to 20,000 psi. Once the stimulation
fractures may be horizontal. treatment is completed, the mandrel is pumped back
An additional idea, investigated at the time of this out of the Christmas tree, and the wellhead valves can
writing, is to forego perforating altogether for fractur- be closed.
ing and consider the creation along the well of vertical
notches parallel to the fracture plane by using a jet-
cutting technique. 11-4.2. Treating iron
The size of the high-pressure pipe called treating iron
used on a treatment is dictated by both the anticipated
11-4. Surface equipment for rates and pressures. Smaller lines have a higher maxi-
fracturing operations mum treating pressure limitation than the larger sizes.
Assembling the surface equipment in a safe, organized The velocity of the fluid should be limited to 45 ft/s
and efficient manner is extremely important for the to minimize excessive erosion of the iron. Pumping
success of a fracturing treatment. Thorough pretreat- above these rates for any prolonged period of time can
ment planning is essential to the organizational process erode the treating iron and thereby lower the effective
of coordinating equipment hookup. An inspection of working pressure that the iron could be exposed to
the location prior to the treatment allows making dia- before a catastrophic failure would occur. If the design
grams to optimize the use of available space. treating rates exceed the rate limits of the iron’s size,
Many steps of the organizational procedures are dri- then either a larger iron must be used or multiple lines
ven by common sense. Yet, small problems can easily must be laid to the wellhead.
be overlooked in the rush to get things ready. Pretreat- The treating iron should not have welded seams
ment planning can eliminate many small problems that or exposed threaded connections. To eliminate these
have the potential to develop into larger problems that seams and exposed threads, the iron and associated
may ultimately jeopardize the success of the treatment. connections should be machined from single pieces of
Figure 11-5 shows equipment positioning for a frac- metal. The connections between two pieces of treating
turing treatment. Figures 11-6 and 11-7 show fractur- iron should have nonpressure unions. This style of

Fracture tanks
Liquid transport
trailer
Transfer
pump Proppant feeders
Fluid blending unit
Job
mo
n
uni itorin Proppant Proppant conveyer
t g blender
Hopper
er Pu
mp
mp er
Pu
Manifold trailer
Preferred Pu
wind er mp
mp er
direction Pu
Pu
er mp
mp er
Pu
Pu
rt
mp
po
er er
mp
rt
ns
po
Pu
ort
tra
n s
nsp
2 tra
2
CO
2 tra
CO
CO
N2 pumper
N2 pumper CO2 booster
N2 pumper Flowmeter
CO2 pumper
Densitometer
Pressure transducer Flowmeter

Pressure transducer
Pumper Well Bleed line
Figure 11-5. Equipment positions for a typical fracturing treatment.

Figure 11-6. Fracturing equipment on a small fracturing treatment.
Figure 11-7. Fracturing equipment on a large fracturing treatment.

connection prevents exposure of the threaded portion well fluids. A permanent flowback line should be
of the connection to the treating pressures. laid for the long-term flow of well fluids. This line
During pumping operations the treating iron tends should be placed between the check valve and the
to move and vibrate slightly. To prevent exposure of wellhead control valve. If the bleedoff arrangement
the iron to stress from this movement, the line is laid is misplaced, pressure will be trapped between the
to provide some flexibility. Swivel joints allow the wellhead and the check valve, and a pressure haz-
iron some freedom of movement. These same connec- ard will result when the treating line is rigged
tions also provide a means for the iron to make cor- down.
ners and change directions. To ensure that straight sec- A choke and double-valve arrangement should
tions of the iron are completely free to move, it is be teed off of the main treating line to start the
recommended to have three swivel joints between any bleedoff line. One valve serves as a master valve
two fixed points in the line. A swivel joint is required that is always fully open or fully closed and is
at each end of a straight section, and another three- opened first and closed last. The second valve is
way swivel joint (i.e., the connection swivels in the slowly opened and closed to control the flow of
center plus each end is free to rotate) is required at fluid. Swivel joints must never be used in a bleed-
one of the ends. off line. Where a turn in the line is necessary, teed
connections should be used. The bleedoff line must
• Check valves
be restrained every 15 to 20 ft to prevent it from
A check valve should be placed in the treating line, moving. Care should be taken when flowing back
on the ground, as close to the wellhead as practical. fluids that could potentially be carrying ball sealers.
The valve allows flow in one direction; therefore, it A ball sealer flow diverter can be included on the
can be pumped through but automatically closes wellhead side of the choke to catch the balls and
once pumping stops. This isolates well pressures prevent them from interfering with the bleedoff
to the closed side of the check valve and prevents procedure.
the flow of well fluids. This type of valve is essen- The bleedoff line should tee off of the main treat-
tial for controlling the well when there is a sudden ing line and be staked. Lateral, or Y, connections
loss of pump pressure, such as when a treating line should be avoided when laying the bleedoff line.
fails. In these emergency situations there is no time A bleedoff line with a lateral is difficult to stake
to physically close the wellhead valves, but the in place and may move when the bleed valves are
check valve can automatically close almost instan- opened and there is high pressure at the wellhead.
taneously.
Two common types of check valves are used
in high-pressure treating lines. Flapper-type check 11-4.3. High-pressure pumps
valves are usually used in the main treating line.
Proppants, solid diverting agents and ball sealers High-pressure pumps should be spotted close enough
can all be pumped through this style of valve with- to the blender so that the discharge pumps on the
out fear of plugging or destroying its reliability. blender can easily feed slurry at a sufficiently high
Dart check valves are used in nitrogen and carbon net-positive-suction head to the intake manifolds on
dioxide (CO2) treating lines. These valves use a the pumps. On large treatments with many pumps,
spring to seat a dart and close the valve when injec- a manifold trailer may be used to consolidate the
tion stops. Because dart check valves are easily cut hookup. The manifold trailer helps organize both the
by proppants and plugged by diverting agents, they low-pressure suction hookup and the high-pressure
are not recommended for use in the main treating discharge hookup.
line. The number of suction hoses between the blender
and the pumps is determined by the pump rate.
• Bleedoff lines Standard 4-in. suction hoses in 25-ft lengths or less
A bleedoff line is used to relieve pressure from the allow about 12 bbl/min of fluid flow to the pump. If
system of high-pressure treating lines once the higher rates are attempted through one hose, insuffi-
wellhead valves have been closed. The bleedoff cient net-positive-suction head may result and cause
line is not intended for the extended flowback of the pump to cavitate and run roughly. If rates by one

pump are expected to exceed 12 bbl/min, another suc- or disposal costs are incurred. In addition, a more
tion hose should be used to provide fluid to the intake predictable and consistent viscosity is obtainable for
manifold. large treatments, where bacteria can degrade the gel
For low-rate treatments, the hose diameter may viscosity of a batch-mixed fluid before pumping
have to be decreased to maintain a sufficiently high begins. Personnel time and costs can also be greatly
fluid velocity inside the hose, especially on high- reduced. The continuous-mix process eliminates the
proppant-concentration treatments such as a foam need to have gelling crews precede fracturing opera-
fracture treatment. Alternatively, a recirculation line tions, resulting in direct savings in time for personnel
from the pump suction manifold to the blender can and equipment. Finally, viscosities can be easily
be used. If the fluid velocity in a hose drops to a point changed throughout the treatment. This allows taper-
where proppant settling is severe, the hose may plug ing the polymer loading so that fluid damage to prop-
off and starve the pump for fluid. pant conductivities can be minimized or a net pressure
Each pump truck should have an isolation valve limitation can be met.
where it is tied into the main treating line to facilitate To ensure that a continuous-mix operation goes
making minor repairs during pumping operations. smoothly, several requirements must be observed. The
Without this valve the pump would always be polymers should be of a liquid or slurried variety to
exposed to the treating pressure. Behind the isolation ensure that they can be added at precise concentra-
valve a bleedoff valve should be installed so that the tions. Liquid or slurried additives can be pumped and
pressure on the pump can be safely bled off any time monitored much more accurately than dry powdered
the pump is brought off line. materials. These polymers produce an improved,
The size of the iron on the pump should be compat- quicker hydration especially when mixed with process-
ible with the rate and pressure capabilities of the controlled equipment. Specialized mixing and hydra-
pump. If the pump and iron are not performance tion units provide the metering capabilities, proper
matched, the effective efficiency of the pump is mini- shear environment and sufficient residence time for
mized. proper hydration. The hydration process related to
Recirculation lines between the blender and the time and shear has proved to be extremely important
suction manifold of the high-pressure pump may be for continuous-mix treatments. If the base fluid has
necessary when high proppant concentrations are not progressed sufficiently in the hydration process
pumped. At high concentrations the proppant may before the fluid is crosslinked, the fluid may experi-
settle out of the slurry within this manifold. Settling ence stability problems.
problems are more likely to occur at low pump rates Process-controlled proppant blenders use computers
and for low-viscosity fluids. The recirculation line to meter precise proppant-to-fluid ratios throughout
keeps fluid moving within the suction manifold and the treatment. This precision-blending capability is
prevents proppant from settling out. Fracturing treat- perfect for ramping proppant, which is considered the
ments using foamed fluids usually require a recircu- ideal for optimum proppant placement. The blenders
lation line. accurately mix and meter proppant, dry additives,
liquid additives and fracturing fluid together at a spec-
ified density in a preprogrammed, automatic mode.
11-4.4. Blending equipment The proppant concentration can be precisely and safely
Process-controlled blending equipment that meters controlled at concentrations higher than 22 lbm/gal
and continuously mixes polymer slurry, concentrated added for sand or 32 lbm/gal added for high-strength
potassium chloride (KCl) solution and liquid additives proppant. The overall proppant/fluid ratio is con-
has made continuous-mix operations a viable alterna- stantly monitored and controlled in a range of ±0.5%.
tive to batch-mix operations. There are several advan-
tages to performing a fracture treatment in continuous-
mix mode. Environmental concerns are greatly
11-4.5. Proppant storage and delivery
reduced because only freshwater residuals remain in The total volume of proppant, rate of proppant deliv-
the fracture tanks after a treatment. Besides eliminat- ery and number of different proppants to be used in
ing the cost of gelled tank bottoms, no tank cleaning the treatment must be considered when choosing the

proper vessel for storing proppant on location. Small umes of proppant are required for a treatment, con-
treatments may require only dump trucks to deliver veyor-equipped sand bins can be positioned to offload
the proppant to the blender. Bobtail dumps and trailer proppant onto a second conveyor that feeds the
dumps are routinely used throughout the industry. blender. Spotting sand bins in this arrangement allows
These units are spotted directly over the blender hop- millions of pounds of proppant to be easily stored and
per and gravity feed the proppant at a controlled rate pumped.
through a chute. Care should be taken to ensure that Regardless of the type of bulk proppant storage
the hydraulic lift mechanism is in good working order, used, great care and planning must go into treatments
allowing the dump to be completely raised. If the in which proppant types and sizes will be changed
dump is not completely raised, the proppant delivery during the treatment. The time and procedures
rate may not be sufficient, especially as the dump required to make proppant changes during the treat-
nears empty. This can be critical at high proppant con- ment must be considered in the pretreatment planning.
centrations, which require the highest delivery rates.
To prevent problems with the proppant delivery rate,
it is a good idea to bring extra proppant if capacities 11-4.6. Vital signs from sensors
allow. Bobtail units generally haul between 250 and The monitoring of hydraulic fracture treatments has
450 ft3 of proppant, and trailer units haul up to 600 ft3. evolved from simple pressure strip charts to sophisti-
The maximum delivery rate of these units depends on cated computer recording and display. The informa-
the chute size and placement but is generally between tion displayed by these instruments provides the
3000 and 7000 lbm/min of proppant. supervising engineers with diagnostics on how the
Moving proppant dumps during a treatment should treatment is proceeding. Real-time execution decisions
be minimized. Even with good operators, it takes 3 to are made during the treatment based on this information.
5 min to move a dump into place and raise it for prop- Sensors acquire input data to track and account for
pant delivery. Excessive movement of the proppant the numerous operations taking place on location.
dumps increases the chance of not pumping the proper Most of the parameters required for evaluating a frac-
proppant concentration. If a treatment is designed with turing treatment can be followed with sensors.
large proppant volumes requiring movement of the Pressure, density, rate, temperature, pH value and vis-
dump trucks during pumping procedures, other prop- cosity are some of the more common parameters dis-
pant storage equipment should be considered. played and recorded.
Upright storage silos can hold up to 4000 ft3 of
proppant. The silos rely on gravity to feed the blender • Pressure measurement
and do not require hydraulic power to operate. This Pressure transducers measure the deformation of
type of storage eliminates moving trucks for most a sensing material to provide a pressure reading.
treatments because of the quantity of proppant it holds. Specially designed strain gauges are bonded to
However, this type of proppant storage requires con- precision-machined metals that subtly deform as
siderable planning because the units are not mobile they are exposed to pressure. The quality of pres-
once they are spotted. Specialized cranes are required sure transducers varies both in accuracy and reso-
to raise these units into place and to lower them for lution. The accuracy of a gauge is determined by
transportation. how closely it measures pressures over an entire
The conveyor-equipped sand bin is the most com- pressure span. The resolution indicates the size of
monly used unit for delivering proppants to the the pressure increment required to affect the mea-
blender. These units have several compartments for surement.
storing proppant. Each compartment has a set of It is imperative that treating pressures are accu-
hydraulically controlled doors at the bottom. When rately known during a treatment. The main sensor
the doors are opened, proppant falls from the con- used to measure the pressure should be placed as
tainer onto a conveyor belt that leads to the blender. near to the wellhead as practically possible. Care
The storage capacity of these units ranges between should be taken to locate the pressure sensor on the
2500 and 4000 ft3. Hydraulically powered conveyors wellhead side of the check valve to ensure that an
can unload these units at proppant delivery rates accurate pressure is available even after pumping
approaching 25,000 lbm/min. If extremely large vol- operations have stopped. Should the transducer

inadvertently be placed on the upstream side of the A good densitometer is accurate to within
check valve, pressure readings could indicate that 0.1 lbm/gal over a density range of 8.0 to 25 lbm/gal.
pressures have bled off while actual wellhead pres- This remarkable accuracy can still lead to errors,
sures remain dangerously high. especially when a densitometer is used to detect
Generally, several pressure sensors are available proppant totals and the treatment goes to a high
for monitoring purposes in addition to the primary proppant concentration. If the fluid and proppant
transducer at the wellhead. A second transducer densities vary by 0.1 lbm/gal and the densitometer
should be identified and calibrated to the primary has an accuracy level of 0.1 lbm/gal added, an error
sensor. Pressure measurements from this transducer of nearly 10% can occur at proppant concentrations
should be used if the primary sensor fails during greater than 10 lbm/gal added because the densito-
the treatment. All pump trucks should have a trans- meter is actually measuring radioactive adsorption,
ducer on each high-pressure pump. However, these which is proportional to the density. The value of
sensors should not be used as a backup to the pri- pounds per gallon must be calculated from the den-
mary pressure transducer because pressure mea- sity reading. This calculation at higher densities is
surements will be lost if the pumps are shut in and less accurate because a small change in absolute
isolated during the treatment. The pump sensors density means a larger change in pounds per gallon.
can be used to indicate problems associated with As an example, the density readings comparing no
each unit. A pressure that deviates significantly proppant with 3-lbm/gal added sand are 8.43
from the other sensors indicates that the pump lbm/gal and 10.06 lbm/gal, respectively (a change
should be shut in and isolated. Once the source of in density of 1.63 lbm/gal). However, a similar
the abnormal pressure is located and repaired, the change in density (1.52 lbm/gal) from 13.96 to
pump can be brought back on line. 15.48 lbm/gal corresponds to a change from 15.0-
• Density measurement to 23.5-lbm/gal added.
For years radioactive densitometers have been used • Rate measurement
to measure density. This technique provides a non- Several styles of rate sensors are used for monitor-
intrusive, continuous density measurement for any ing high-pressure treatments. Commonly used rate
fluid flowing in a pipe. The technique is based on sensors include pump stroke counters, turbine
the absorption of gamma rays by the measured flowmeters, magnetic flowmeters and venturi
fluid. flowmeters for gaseous fluids. Pump stroke coun-
A densitometer consists of ters generally measure the rotational speed of the
– radioactive source on one side of the pipe pump. Each full rotation of the pump moves the
pump plunger through one complete cycle. From
– gamma ray detector on the other side of the pipe
the volume of each pump stroke, number of pump
– electronic panel to provide a signal reading. strokes and the pump’s efficiency, a rate can be eas-
As fluid passes through the pipe, gamma rays ily calculated. This type of measurement is advan-
emitted by the source are attenuated in proportion tageous for high-pressure pumps that are pumping
to the fluid density. The detector senses the gamma high volumes of solids, as is the case in pumping
rays transmitted through the fluid and converts this proppants during a fracturing treatment. The pres-
signal into an electrical signal. The electronic panel ence of solids or varying fluid rheology does not
processes the electrical signal into a density indica- affect this type of measurement. However, if the
tion. Denser materials absorb more radiation, pumps are not fully primed, the actual rate may be
resulting in the detection of fewer gamma rays. significantly lower than that indicated by a stroke
Thus, the signal output of the detector varies counter.
inversely with respect to density. Turbine flowmeters are designed primarily for
Most densitometers use a radioactive isotope measuring clean fluids. As fluid is pumped through
with an extended half-life. A densitometer using the flowmeter housing, turbine blades are forced to
137
cesium can function accurately for nearly 30 rotate. Each of the blades on the rotor creates a
years if the electronic components are maintained small pulse as it passes beneath a magnetic pickup.
properly. Each revolution of the rotor generates a distinct

pattern of pulses equal to the number of turbine Process control of blending and pumping opera-
blades. The rotational speed of the rotor defines the tions provides a major advancement in treatment
flow rate, and each pulse defines a volume of liquid execution. Proppant schedules can be ramped
passing through the meter. Most turbine meters are rather than added in stages. The viscosity of the
calibrated with fresh water. As the viscosity of the fracturing fluids can also be changed by process
measured fluid increases, the rotor speed decreases control. Polymer concentrations are tailored
proportionally. Therefore, a turbine flowmeter throughout the job to deliver the desired fluid prop-
should be calibrated with the actual fluid to be erties. The computer-controlled execution helps
measured to ensure accuracy. These meters are use- eliminate human-induced errors that can be detri-
ful for measuring liquid additive rates but have lim- mental to the treatment.
ited value for proppant slurries. Fracturing fluid
viscosities often vary throughout a treatment, espe-
cially when proppant is introduced. The proppant 11-4.7. Equipment placement
also tends to erode the turbine blades, further
Several considerations must be made when planning
reducing the accuracy of the flow measurement.
the placement of pumping equipment. Equipment
Magnetic flowmeters are popular for measuring
should be placed upwind and at least 50 ft from the
treating fluid rates. These flowmeters measure the
wellhead. If a service rig is on location, equipment
rate of fluid flow in a line, similar to a turbine
should be spotted out of the fall line of the rig mast.
flowmeter, but they do not have intruding parts that
Care should also be taken to place the high-pressure
can be affected by changing viscosities or proppant
pumping equipment where personnel will not be
addition. The electromagnetic flowmeter operates
exposed to the fluid end of the pump. If the equipment
on Faraday’s law of electromagnetic induction.
on location is to be operated remotely, personnel
When a conductor is moved across a magnetic
should set up in an area protected from potential prob-
field, an electromagnetic force is induced in the
lems with high-pressure lines. If equipment for N2 or
conductor. The electromagnetic force is orthogonal
CO2 is on location, it should be spotted at least 60 ft
to both the direction of movement and the magnetic
from other equipment and the wellhead.
field itself. The electromagnetic flowmeter uses a
pair of coils mounted on the outside of a nonmag-
netic pipe. An electrical current flows through the
coils to produce the magnetic field. The electro- 11-5. Bottomhole pressure
magnetic force is created when fluid (a conductive measurement and analysis
material) passes through the coils. The biggest limi-
tation of the electromagnetic flowmeter is that it Various modes of fracture propagation can be inferred
cannot be used with oil-base fluids. from straight-line periods that develop on log-log
To accurately measure N2 or CO2 rates, a venturi plots of net pressure versus time. These modes charac-
flowmeter should be used. The velocity of the gas terize different fracture shapes. Fracture closure and
through the venturi is determined by measuring the fluid leakoff parameters can be determined from the
pressure drop across a reduced area. The line pres- analysis of fracturing pressures resulting from closure
sure and temperature must also be measured to (pump-in/flowback [PI/FB]) tests. Chapter 9 provides
compensate for changes in density. additional fracturing analysis information.
• Data acquisition and process control Collection and analysis of fracturing pressures are
facilitated by the monitoring vehicles provided by
Computer monitoring systems can track and record
service companies. These vehicles are equipped with
numerous sensor inputs, making them ideal for
computer hardware and software that can digitize,
monitoring complex treatments. These systems can
display, analyze and record relevant data.
also use data to create a presentation that makes
The successful application of any method of frac-
interpretation easy. Real-time processing of data
turing pressure analysis relies on the accuracy of the
is essential when monitoring complex treatments.
fracturing pressure data. For a quantitative analysis
These systems not only monitor but also control
of the data, a sensitivity level on the order of 5 psi and
additive rates based on a predetermined schedule.
an accuracy level on the order of 25 psi are generally

sufficient. In particular, the bottomhole fracturing mission and decoding rates are much slower than for
pressure and magnitude of the least principal stress electric line transmission, wireless telemetry is not
(fracture closure pressure) must be known before the desirable for fracturing operations.
net pressure can be calculated. An error in either of
• Computed bottomhole pressures from surface
these can result in an incorrect net pressure and incor-
measurements
rect fracturing pressure analysis.
The sampling rate affects the amount of data that Bottomhole pressures can be estimated by extrapo-
must be acquired. The fastest time is required for col- lating surface pressure measurements. However,
lecting microfracturing pressures because the pump these extrapolations are of little value unless the
times are typically short and closure can occur perforation friction, fluid/slurry friction, near-well-
quickly. Single injection periods can be as short as bore friction and hydrostatic pressures are known
3 min; however, multiple PI/FB tests may be required. and accounted for. The degree to which computed
In the microfracturing case, a sampling period of 2 s bottomhole pressures resemble actual bottomhole
is desirable; 15 to 30 s may be acceptable. pressures depends primarily on the accuracy of the
The execution and evaluation phases of a fracturing fluid/slurry friction data. Fluid friction and its effect
treatment tend to last for a significant period of time, on computed bottomhole pressure increase as the
and the sampling period can be longer than a pipe diameter decreases; therefore, accurate fluid
microfracturing sampling period. Fracturing treat- friction data are especially necessary for treatments
ments typically last 0.5 to 12 hr. One minute per sam- pumped down tubing.
ple is acceptable; however, if adequate memory is Improved engineering estimates of closure pres-
available, the sampling period should be shorter. sure, perforation friction and fluid/slurry friction
Three types of pressure transducers are used in bot- for determining the bottomhole pressure and net
tomhole pressure gauges: pressure can be computed using a methodology
consisting of calibration and application phases.
• strain gauge The calibration phase is devoted primarily to mea-
• capacitance gauge suring the pipe friction of the fracturing fluids (fluid
• quartz crystal. with and without proppant). This phase can be
completed in just one fracture treatment, provided
Each of these pressure transducers uses mechanical that fluid friction data are required for only one set
displacement to generate a pressure signal. When pres- of fluids in one pipe size. During the application
sure is applied to the transducer or sensor, an element phase, the results of the calibration phase are cou-
is elastically deformed and the displacement converted pled with a brief on-site pump test to produce
to an electrical signal. The electrical signal is measured meaningful, real-time net pressure plots.
and a pressure is inferred. All bottomhole pressure Surface pressure gauges are present at all frac-
transducers are packaged with a temperature sensor turing treatments. It is standard practice to use sur-
to enable temperature correction of the transducer. face measurements to infer bottomhole pressure.
Several methods are used to measure bottomhole Pressure gauges used by service companies are
pressure: modeling from surface parameters, non-real- accurate only to within ±50 psi. This does not
time downhole pressure devices (memory gauges) and include temperature-induced errors and offsets
real-time gauges conveyed by wireline or coiled tub- caused by clamping forces and drift, which can be
ing. The downhole pressure measured includes any significant. If surface pressure measurements will
perforation friction pressure and tortuosity effects, be frequently collected for calculating bottomhole
which must be accounted for during analysis (see pressure data, then the pressure gauge should be
Chapter 9). of high accuracy. This is particularly true when
Devices are available that transmit real-time data a packerless completion is used to monitor bottom-
but do not require a physical connection to the surface hole data (i.e., tubing or annulus used as a “dead-
equipment. Prior to the fracturing treatment, the trans- string”).
mitter is set below the perforations using wireline. Significant errors can occur when inferring bot-
Pressure data are transmitted or telemetered through tomhole pressure from surface pressure measure-
the rock to a receiver at the surface. Because the trans- ments of deep, hot wells. The change in fluid den-

sity as the cooler fracturing fluid heats can be sig- – The instrument is potentially subject to loss.
nificant, particularly if the fracturing treatment is – The instrument must be retrieved using wireline
pumped down tubing (faster temperature changes). or tubing after the fracture treatment.
Figure 11-8 shows the effect temperature has on
• Wireline-conveyed gauges
hydrostatic pressure using water.
Real-time bottomhole pressure data can be gener-
• Downhole memory gauges
ated using wireline-conveyed gauges. In the case
Pressure gauges can be placed downhole to record of a tubing/packer combination, the wireline is
pressure during the fracturing treatment. Placement banded to the outside of the tubing (a slow pro-
can be in a gas-lift mandrel, in a closed-end joint at cess), and the gauge is on a side-port mandrel
the bottom of a tubing joint with perforations above placed just above the packer. In situations where
the gauge or in a packer below the zone to be frac- the fracture treatment is pumped down the
tured. After the fracturing treatment, the gauge is casing/tubing annulus and the tubing is open-
retrieved, and the data are downloaded and pro- ended, a wireline-conveyed pressure gauge can be
cessed for analysis. run in the tubing to measure the pressure near the
Memory gauges have several disadvantages: bottom of the well. This configuration is generally
– Data are stored in gauge memory, so they cannot not used because there are no direct advantages
be used immediately. over surface recordings through a static fluid col-
– Because the gauge is downhole, if a treatment or umn in either the tubing or annulus for applications
sensor anomaly occurs, it will not be discovered in which the density of the column remains essen-
until after the treatment. tially constant.
Wireline-conveyed gauges run on unprotected
– Treatment delays can affect the recording of data,
wireline can be used for calibration treatments
for which there is a time limit. (Pressure-tripped
without proppant if certain precautions are taken:
recorders can be used to overcome the time
limit.) – Determine the increased wire tension caused by
fluid drag. Equation 11-5 can be used to calculate
an estimate of increased wire tension. A safety
factor is used to compensate for dynamic effects
Pressure Decline
Surface
and errors in the estimate of pipe friction:
(initial head
reference) Depth (ft) T = Ts − Td > 1.1Dp Dw p pipe friction , (11-5)
Pressure
350 20,000
pw ∆ head where T is the wire tension in lbf/ft, Ts is the
Head change from 75°F to static gradient (psi)
18,000
300 Time wire strength in lbf/ft (for a free-end case that
16,000 allows rotation and wire twist), and Td is the
14,000
dead weight in lbf. Prior to pumping, Dp is the
250
inner diameter of the pipe in in., Dw is the wire
12,000 diameter in in., and ppipe friction is the estimated
200 pipe friction in psi (total pressure drop) for the
10,000
fracturing fluid at the indicated pumping rate.
150 – Monitor the actual fluid friction, relative to the
estimated value, to ensure a reasonable prediction.
100 – Protect the wire at the fluid-entry location by
either mechanical isolation or equal opposing
50 fluid streams.
225 250 275 300 325 350
The several limitations to Eq. 11-5 include the
Bottomhole static temperature (°F)
following assumptions:
Figure 11-8. Hydrostatic head changes caused by temper-
– turbulent flow of a non-friction-reducing fluid
ature. pw = wellbore pressure. (e.g., water)

– equal wire and pipe roughness fracture width, closure stress was shown to affect the
– no dynamic loading effects on the wire (no oscil- occurrence of flowback in the modeling study and
lations). experimentally.
Proppant flowback usually occurs over a cleanup
In situations where the fracture treatment is
period of several weeks after the fracture treatment,
pumped down the casing (no tubing in the well),
but it can also occur throughout the economic life of
the wireline and gauge should never be exposed to
the well. Up to 20% of the proppant placed in the
fluid containing proppant, but should be protected
fracture can return during the cleanup period. The
for the entire length inside tubing.
proppant that flows back has a detrimental wear effect
• Coiled tubing–conveyed gauges on the production equipment and requires the use of
Gauges run on the end of coiled tubing can be used separators in the production line. Concern about prop-
to measure bottomhole pressure for real-time frac- pant flowback can limit fluid flow rates during
turing pressure analysis. Measurements of other cleanup and production.
parameters such as bottomhole temperature and In most cases, proppant flowback does not reduce
casing collar locator data can also be made; bot- well production. It can therefore be concluded that the
tomhole temperature can be used to calculate tubu- fracture does not close completely as proppant is pro-
lar stresses resulting from temperature changes. duced. Also, the production rate can be reduced to
The bottomhole temperature gauge can also be a point where proppant is not flowed back.
used for postfracturing temperature logging. Proppant flowback is more likely to occur with
• Measurement options lower closure stress or wider fracture widths. Another
Table 11-4 lists techniques for various well config- possible cause is that the closure stress probably
urations for fracturing treatments. varies from point to point in the fracture between the
maximum and zero as a result of uneven settling of
proppant in the fracture. The resulting stress variation
can allow proppant to be carried out of the fracture
11-6. Proppant flowback control from regions of lower closure stress.
A problem with fractures in some applications is the Several techniques have been used to control prop-
back production of proppant (proppant flowback) with pant flowback: forced closure, resin flush, the use of
the oil or gas. Fluid drag forces dislodge and carry curable-resin-coated proppants and fiber technology.
proppant out of the fracture. Numerical modeling Like sand control (except for wide proppant packs,
indicates that arch formation in the pack is important. as with TSO designs) the prediction of proppant flow-
Places in the fracture wider than 5.5 grain diameters back is specific for a site, field, formation and fracture
are inherently unstable, independent of the effective and depends on field observations and correlations
proppant stress. In these cases, fluid flow serves to developed for an area.
sweep proppant out of the fracture. In addition to
Table 11-4. Real-time measurement options.
Design Execute Evaluate

Casing Calculation
Packerless completion Packerless completion Packerless completion
Electric line Electric line
Existing production tubing Calculation

New production tubing Calculation

Permanent gauge Permanent gauge Permanent gauge
Temporary work string Calculation

Permanent gauge Permanent gauge Permanent gauge

11-6.1. Forced closure enables increased polymer cleanup early in the flow-
back and can result in an increase in the total polymer
Forced closure is a procedure in which fluid flowback returned. Because no curing reactions occur, wellbore
begins immediately at the end of pumping. The theo- cleanout is similar to that for normal proppant.
rized benefits of forced closure are that a “reverse” Additional information on fiber technology is in
screenout takes place at the perforations (i.e., the frac- Sidebar 11F.
ture width closes to below that required for a stable
arch) and that the fracture closes before the proppant
has a chance to settle in the fracture.
11-7. Flowback strategies
Proper flowback of a fracturing treatment is designed
11-6.2. Resin flush to recover a maximum amount of the injected fluids
The resin flush technique involves pumping a curable while removing a minimum amount of proppant from
resin into the fracture at the end of the job. In theory, the fracture. When to start the flowback and the flow-
the resin coats the proppant in the fracture near the back rate are the key issues. Variables such as fracture
wellbore and cures through a polymer crosslinking closure time, fluid break times (influenced by breaker
reaction. Additional postflushes are used to ensure that schedule), reservoir energy available to produce the
the resin does not fill the pores in the proppant pack. fracturing fluid, formation strength and proppant con-
The disadvantages of this technology are the difficul- centration in the near-wellbore area affect the flow-
ties in covering the entire interval with resin and then back strategy.
pumping the postflush through the entire treated vol- The overall goal of a fracturing treatment is to
ume and the requirement to drill the excess resin out achieve maximum fracture conductivity within the
of the wellbore after it cures. limits of the designed treatment. The flowback effi-
ciency of the stimulation fluid can have a significant
effect on the resultant fracture conductivity. In all
11-6.3. Resin-coated proppants cases and especially in the event that early-time prop-
pant flowback control techniques are not used, fractur-
The use of curable phenolic resins with proppants is ing fluid flowback efficiency should be considered
a popular method for controlling proppant flowback. during the design process.
They are used as all or some (tail-in) of the proppant Typically the fracture is allowed to close before
placed in the fracture. Under sufficient closure stress, flowback is initiated. Fracture closure is required to
shut-in time and temperature, the resin coating is sup- trap the proppant within the fracture and to prevent
posed to bind the proppant together in the fracture and subsequent proppant movement through the applica-
form a dense aggregate around the perforations. A dis- tion of fracture closure stress on the proppant. How-
cussion of resin-coated proppants is in Chapter 7. ever, even the application of closure stress may not be
sufficient to prevent some degree of proppant flow-
back during treatment cleanup and post-treatment pro-
11-6.4. Fiber technology
duction. In these cases, the application of proppant
Fiber technology was developed to hold the proppant flowback control techniques may be necessary.
in the fracture during the production of oil, gas or both Some fracturing fluids contain gases to foam or
and to allow more flexibility in flowback design than energize the fluids. The gases most commonly used in
possible with curable-resin-coated proppants. These energized fluids are N2 and CO2. During flowback fol-
additives work by the physical mechanism of random lowing a treatment, they provide an efficient source of
fiber reinforcement; therefore, chemical curing reac- concentrated energy to aid in rapid and more complete
tions are not necessary to hold the proppant in place. post-treatment cleanup, especially in low-bottomhole-
No combination of temperature, pressure or shut-in pressure wells. Compared with long-term swabbing or
time is required. Wells can be flowed back at high pumping operations, energized fluids can be cost
rates (dependent on the number of perforations). Also, effective. Entrained gas in the fracturing fluid is also
flowback is possible immediately after the fracturing beneficial for fluid-loss control.
treatment is completed. The rapid flowback rate

11F. Fiber technology damaging the reinforcing structures (Fig. 11F-1; Howard et al.,
1999).
Simon James, Schlumberger Dowell The fibers are essentially inert materials that do not interact
with fluids, breakers or additives (Card et al., 1995) and have
The inclusion of fibers in an intimate mixture with proppant is been successfully applied in all fluid types—high- and low-tem-
a recent innovation for controlling proppant flowback following perature crosslinked guar, derivatized guar, gelled oil and vis-
hydraulic fracturing. The fibers stabilize the proppant pack by coelastic surfactant. The inert nature of the fibers provides
physical rather than chemical means. Several reinforcement chemical stability while the physical reinforcement mechanism
mechanisms may contribute to the stabilization of the proppant provides proppant pack stabilization from low to high tempera-
arches (Card et al., 1995); the relative importance of each tures. Fibers have been successfully applied in initially low tem-
mechanism depends on the fiber properties. Fibers at the face perature wells that were then subject to cyclic steam injection
of the fracture may increase the force necessary to remove a (Jones and Soler, 1999).
proppant grain from the pack (Romero and Féraud, 1996)—the Fibers can also be used in conjunction with RCP to provide
dominant mechanism for high-modulus fibers. Longer fibers maximum proppant flowback control for extreme conditions.
also tie the proppant particles together, extending the stabilizing The RCP provides resistance to very high drawdowns. The
forces deeper into the pack—the dominant mechanism for fibers enable quicker cleanup times (Bartko et al., 1997) and
longer, flexible fibers. Large-diameter (>50-µm) fibers may also reinforce the RCP, providing resistance to failure through clo-
stabilize proppant packs by acting as wedges between grains sure stress cycling.
to prevent relative movement. An additional benefit of fibers observed in field applications
Physical reinforcement leads to the following significant is the reduction of treating pressure by as much as 8 psi/100 ft
advantages over curable resin-coated proppants (RCP): (James et al., 1998). The effect is most noticeable in delayed
• no shut-in required before flowback crosslinked fluids. The presence of fibers is believed to inhibit
the formation of turbulent eddies in the tubing (i.e., boundary
• resistant to closure stress cycling layer effects).
• no fluid interactions Whatever type of proppant flowback material (fiber or RCP)
• effective from very low to very high temperatures. is used, it can be effective only if it covers all the perforations
attached to the fracture. If only a portion of the proppant is
Fracturing treatments including fibers can be flowed back treated with proppant flowback control material (e.g., a tail-in
rapidly and immediately after the end of the job (Howard et al., treatment), there is a risk that nontreated proppant will remain
1995). Rapid flowback can allow the removal of more polymer in the near-wellbore region (Fig. 11F-2). Proppant placement is
from the fracture, leading to higher productivity (Anderson et al., affected by settling, near-wellbore recirculation just above and
1996; Willberg et al., 1998), earlier gas to sales and reduced below the perforations and post-shut-in redistribution (Smith et
flowback time compared with RCP (Howard et al., 1995). al., 1997). The design of tail-in proppant flowback control treat-
The advantages of rapid flowback cannot easily be simulated ments, particularly in multilayer and large intervals or where
in laboratory conductivity tests, in which fibers generally provide successful proppant flowback control is essential, should con-
a 20%–30% reduction in retained permeability (Card et al., sider treating the entire proppant volume.
1995). However, at low closure stresses (≤3000 psi) the addi-
tion of fibers can increase retained permeabilities above 100%
of the base proppant by disrupting the close packing of the
proppant grains.
The stability of proppant packs containing fibers is not
affected by closure stress cycling. The lack of permanent bonds
between the fibers and proppant allows relative movement
between the particles during closure stress cycling without
10.0 2.5
Proppant production (g) per 3 cycles
9.9
Proppant pack width (mm)
9.8 2.0
9.7
9.6 1.5
9.5 Proppant pack width
9.4 1.0
9.3
Proppant production
9.2 0.5
9.1
9.0 0
0 5 10 15 20 25 30 35
Number of stress cycles
Figure 11F-1. Proppant production and proppant pack Figure 11F-2. Proppant placement following a fracturing
width as a function of closure stress cycling (Howard et treatment determined by radioactive tracers. Approx-
al., 1999). Each closure stress cycle was 4000-1000- imately 15 ft of the 50-ft perforated interval has proppant
4000 psi over 1 hr, with flow to 85 psi/ft every third cycle. from the early stages close to the wellbore. If only the
After initial cleanup, the proppant pack remained stable last stage had been treated with proppant flowback con-
throughout the test. trol material, proppant production might have occurred.

N2 is inert and relatively immiscible in the fluid • Personal safety equipment

when added in small amounts and without a surfac- Each person on location should wear appropriate
tant. CO2 is soluble in water and becomes a reactive safety equipment to minimize the risk of personal
component in the fracturing fluid. CO2 can be con- injury. Hard hats, hard-toed shoes and safety
verted to carbonic acid, which lowers fluid pH values glasses should be the minimum level of safety
and can be incompatible with fracturing fluids. equipment worn on location. Other equipment such
A large percentage of CO2 goes into solution at typ- as hearing protection, goggles, fire-retardant fabrics
ical reservoir conditions, which can be advantageous. and filter masks should be worn if exposure to the
As a dissolved gas, CO2 does not easily dissipate into conditions they protect against is a possibility.
the formation. During flowback, the dissolved gas Wearing safety equipment is a simple step that
evolves from the mixture and imparts a solution-gas creates a positive safety atmosphere on location.
drive to the fracturing fluid.
• Safety meeting
Energized fluids should be flowed back as soon as
possible while the fluid retains high pressure; how- Holding a pretreatment safety meeting ensures that
ever, shut-in times of up to 4 hr can be tolerated with- all personnel on location are aware of specific dan-
out a great loss in the energizing medium. gers and required procedures relative to the treat-
ment. Each person on location should clearly
understand his or her role during the treatment as
well as individual responsibility during emergency
11-8. Quality assurance and situations. A head count must be taken to account
quality control for everyone on location. An escape route and
The principal factor in the successful performance of meeting place should be agreed upon where all per-
a fracturing operation is the incorporation of quality sonnel will gather in the event of an emergency sit-
assurance and quality control (QA/QC) into all phases uation. Personnel who are not directly involved in
of the fracturing operation. This can be achieved by the treatment should have limited location access
the implementation of a quality management system during the actual pumping operations.
that encompasses all personnel and activities. Everyone should be aware of the unique dangers
A significant effort is made in the design process to of each treatment. Some locations may be in an
determine an optimum fracturing treatment. An equal area with hydrogen sulfide (H2S) or possibly the
effort in QA ensures that the treatment is executed as fluids being pumped are highly flammable. As
designed. Simple QA steps can greatly increase the many of the potential safety problems or concerns
odds of success for a hydraulic fracturing treatment. as possible should be brought to the attention of
everyone.
Maximum pressure limits should be set at this
11-9. Health, safety and time, and every high-pressure pump operator must
be aware of these limits. Instructions for pressure
environment testing the treating iron must also be covered. The
high-pressure treating line, up to the wellhead
11-9.1. Safety considerations valve, should be tested to slightly above the antici-
At no time should the safety aspects of a treatment be pated fracturing pressure. A properly tested line
compromised. Safety guidelines have been developed includes tests of each pump in addition to the main
from experience derived from previous incidents. treating line. The pressure rating of the wellhead
Many of these incidents have had great potential to should be checked to make sure it exceeds the
seriously injure personnel or destroy valuable equip- treating pressure. If the wellhead has a lower pres-
ment. The inherent risk of dealing with high pressures sure rating than the anticipated treating pressure,
can be greatly minimized by following simple safety a wellhead isolation tool will be necessary to iso-
procedures. Hydraulic fracturing treatments can never late the wellhead from this pressure level.
be considered a success if an accident results in the The pretreatment safety meeting is the principal
destruction of equipment or injury to personnel. means of communication for giving final instruc-

tions to all personnel. A well-organized safety ally lighting up. Finally, fire-fighting equipment
meeting helps ensure that the treatment is an opera- should be on location and ready to be operated. In
tional success without being a threat to human this way, a small fire may be contained before it has
safety. a chance to spread and become a major disaster.
• Well control at the wellhead • Precautions for energized fluids
To ensure that well control is always maintained, N2 and CO2 are the gases most commonly used in
the valve arrangement at the wellhead should con- foamed and energized fluids. During flowback fol-
sist of at least two valves. A frac or master valve lowing a treatment, they provide an efficient source
should be installed above the main wellhead valve. of concentrated energy to aid rapid, more complete
If one valve fails to hold the pressure, the other post-treatment cleanup. There are potential hazards
valve can quickly be closed to control the well. It is associated with the use of N2 and CO2. As the fluid
preferable to have the main wellhead valve flanged exits the flowline during flowback, the gaseous
to the casing head, rather than using a threaded phase expands rapidly. This rapid release of energy
connection. If a threaded connection is necessary, must be controlled to avoid a loss of flowback effi-
the condition of the threads must be thoroughly ciency and to ensure personnel safety. Service com-
inspected for thread wear and proper taper. panies have recommended procedures for the flow-
• Precautions for flammable fluids back of energized fluids.
Another potential hazard that is often overlooked
Oil-base fluids should be tested for volatility before
is asphyxiation. N2 and CO2 can collect in low
they are accepted as a fracturing fluid. An oil is
areas, displacing breathable air. Personnel should
generally considered safe to pump if it has a Reid
avoid these areas and remain upwind at all times.
vapor pressure less than 1, API gravity less than
Only one person should be in the vicinity of the
50° and open-cup flash point of 10°F [–12°C]. How-
well during flowback operations. The use of
ever, even if the fluid is considered safe to pump,
remotely operated valves will increase the margin
several additional safety rules should be followed
of safety.
when pumping an oil.
Storage tanks for flammable fluids should be
diked and spotted at least 150 ft from the wellhead.
Spotting the fluids in this manner helps minimize 11-9.2. Environmental considerations
exposing the wellhead to fire if problems occur Fracturing operations should be conducted using
during pumping. Also, all low-pressure hoses sound environmental practices to minimize the poten-
should be enclosed in a hose cover to prevent oil tial for contamination of air, water and soil. All opera-
from spraying on hot engine components of the tions should comply with all applicable environmental
trucks, should a hose leak. laws and regulations.
Care must be taken to ensure that there is no Hazardous material spills should be cleaned up
smoking on location. It is a good idea to have all quickly in accordance with a spill plan. All waste and
personnel check matches and lighters when they unused materials should be handled and disposed of
arrive on location to prevent them from unintention- in an environmentally safe manner.


◆
Appendix: Understanding
Perforator Penetration
and Flow Performance
Phillip M. Halleck, Pennsylvania State University
Introduction Section 11-3 and the excellent monograph on perfo-

rating by Bell et al. (1995).
The productivity of a cased, perforated well can Underbalance perforating has evolved from its
depend to a large extent on perforation practices. early purpose of preventing the infiltration of well-
Modern shaped charge perforators can penetrate for bore fluids to a method for removing shock-induced
some distance into the formation, overcoming both permeability damage. Surge flow from the reservoir
drilling damage and restricted flow through the perfo- into the wellbore is thought to clean loose debris
rated casing. However, perforated completions fre- from the perforation and flush the rock around the
quently do not flow as predicted. Penetration differs perforation to restore permeability. Although effective
from that in certification and quality control tests, for certain specific applications, questions remain
owing to both the character of the formation rock and regarding the limits of its usefulness and the mini-
the in-situ overburden stress. In addition, laboratory mum underbalance and surge flow volume required
tests (Allen and Worzel, 1957) demonstrate that perfo- to obtain maximum well performance. Answers to
ration tunnels are commonly completely filled with these questions can be sought from combined field
debris, and even debris-free perforations may not flow experience, laboratory measurements and models
as predicted from pretest permeability measurements. of perforation cleanup mechanics.
Although most relevant to natural completions, the The review provided by this Appendix consists of
perforating procedure and the condition of the result- a brief discussion of penetration depth, experimental
ing perforations may affect the performance of post- observations of the surrounding rock, models for the
perforation operations such as fracturing or gravel mechanism of underbalance surge cleanup, descrip-
packing. In hydraulic fracturing, acid can be used to tion of laboratory experiments that reveal the time
lower breakdown pressures, presumably because of scale and flow volume required for cleanup and
some form of perforation damage. The orientation of a closing discussion of ancillary effects on perfora-
perforations has also been shown to affect breakdown tion performance.
pressure and well performance. For fracture treatment
design, the interpretation of pretreatment well tests
requires knowledge of the skin resulting from pene- Review of penetration depth
tration and perforation damage effects. During
gravel-packing operations, excessive underbalance in prediction
unconsolidated formations may actually reduce flow Penetration depends on the character of the target
efficiency by damaging the rock (Behrmann et al., 1992). rock formation, effective overburden stress (Saucier
The factors involved in perforation damage may and Lands, 1978) and, in extreme cases, wellbore
include mechanical shock damage to the formation, pressure (Behrmann and Halleck, 1988). In Berea
deep filtration of fines from the wellbore or the forma- sandstone, penetration can be reduced by as much
tion, rock and charge debris in the perforation tunnel as 50% with the application of 3000-psi effective
and multiphase flow effects. Two perforating tech- overburden stress (Halleck et al., 1988). More typical
niques have evolved to avoid or bypass perforation reductions are on the order of 20%. Little additional
damage. These are underbalance perforating and reduction occurs for effective stresses higher than
extreme overbalance perforating (EOP). For more 3000 psi. The nature of the rock affects both penetra-
comprehensive discussions of the field application tion and the extent of the stress effect. Higher
of these techniques, the reader is referred to strength leads to lower penetration. Conversely, stress

affects penetration less in harder rocks. Thus, penetra- Permeability damage and flow
tion data obtained from a single target material under
atmospheric conditions do not reflect actual hole pene- performance
tration. The API standard test (Halleck, 1987, 1988;
Section 2 of Recommended Practices for Evaluation of Idealized view
Well Perforators, [RP 43] 1991) in Berea sandstone at
Most discussions of perforation damage refer to debris
3000-psi effective stress provides a better estimate.
in the perforation tunnel and a crushed zone surround-
Uniaxial compressive strength has long been used
ing the tunnel. The tunnel debris consists of broken
(Thompson, 1962) to correlate ambient-stress pene-
rock, jet metal, explosive products (mostly carbon)
tration. This approach is complicated by the lack of
and other charge debris. If sufficient underbalance
a comprehensive penetration theory incorporating
or production flow occurs, this debris is removed,
target strength, by the change in rock strength with
leaving a clean tunnel. In well productivity models
effective stress (Halleck and Behrmann, 1990) and
(Locke, 1981; Tariq, 1984; Ahmed et al., 1990; Behie
by active target effects in some rocks (Aseltine, 1985),
and Settari, 1993), the perforation is usually treated
which are not explained by conventional penetration
as a circular cylinder surrounded by a zone of reduced
theory. An attempt was made (Ott et al., 1994a) to use
permeability variously called the crushed or compacted
laboratory data to predict downhole penetration from
zone. The zone is usually given (Bell et al., 1972) as
API RP 43 Section 1 certification data. The resulting
0.4 to 0.5 in. thick with permeability on the order of
nomogram provides a simple method for estimating
0.1 to 0.2 times the original rock permeability.
downhole penetration in lieu of costly tests in stressed
In reality, the diameter varies along its length, taper-
reservoir rock. The method relies on a series of empir-
ing toward the tip, and depends on the rock properties
ical correlations among compressive strength, porosity
and charge design. The damaged zone is easily visible
and penetration in rock and concrete targets at atmos-
in perforated samples as lighter colored material
pheric pressure, combined with data in stressed rock
around the tunnel. Microscopic examination reveals
targets. Normalization of the data at 3000 psi, com-
considerable grain breakage in this zone, resulting in
bined with statistical scatter in each of the successive
reduced permeability as the grain-size distribution
correlations used, may lead to substantial errors (Colle
broadens and shifts to smaller diameters. Reduced
et al., 1994; Ott et al., 1994b), particularly at effective
porosity is also thought to play a role.
stresses less than 3000 psi.
Pucknell and Behrmann (1991) confirmed earlier
An alternative for penetration estimates is acoustic
observations made with optical microscopes that
log data based on direct relations between penetration
reveal broken grains in this zone. Asadi and Preston
and dynamic elastic moduli in stressed rock. Origi-
(1991) used scanning electron microscopy (SEM) to
nally proposed by Venghiattis (1963), subsequent data
estimate permeability damage from the size distribu-
(Halleck et al., 1991) largely support the idea that dif-
tion of the broken grains. Appendix Fig. 1 illustrates
ferences in acoustic velocity resulting from rock com-
this type of data with a comparison of SEM photos
position and overburden stress mimic differences in
at 0.1 and 0.4 in. from the center of a perforation
shaped charge penetration.
in Berea sandstone performed using underbalanced,
Because of symmetry requirements, manufacturing
liquid-saturated conditions.
tolerances for shaped charges are necessarily strict at
Attempts to predict the extent of damage have also
all stages. Small variations in dimensions and hetero-
been made. Yew and Zhang (1993) applied the
geneities in the explosive or liner density can cause
method of characteristics to obtain a solution for the
disproportionate variations in penetration performance.
amplitude of the shock waves resulting from the
As a result, each charge type displays a statistical
impact of a shaped charge jet. They assumed cylin-
variation in measured penetration depth. Sukup et al.
drical radial loading and included poroelastic effects.
(1989) reviewed quality control testing methods that
They confirmed that stress wave amplitude falls off
can be used to ensure good reliability. King et al.
rapidly away from the perforation tunnel, although
(1986) investigated additional factors that can affect
they did not address the specific amplitudes expected.
charge performance, including storage duration and
Papamichos et al. (1993) used this solution to estimate
environment.
the extent of grain breakage expected from such stress
A11-2 Chapter 11 Appendix: Understanding Perforator Penetration and Flow Performance

Pucknell and Behrmann (1991) measured the radial

pressure profile in 2-in. long sections of perforated
core during steady-state radial flow. They used five
probes at one end of each 4-in. diameter core.
Although the spatial resolution was limited, evidence
for greatly reduced permeability near the perforation
was not found consistently. They also observed no
qualitative evidence of compaction in their X-ray
computer-aided tomography (CT) images.
Rochon et al. (1995) used pressure transient analy-
sis in a similar short section of perforated core to mea-
sure radial variations in permeability as k(r), where
k is the matrix permeability in md and r is the distance
from the center of the perforation in cm. They injected
light oil into the perforation and followed it with a
constant-flow-rate injection of a more viscous oil.
Analysis leads to an expression for k(r) in terms of the
viscosities of the fluids µ1 and µ2, the flow rate q and
Appendix Figure 1. Scanning electron microscope images the ratio of the differential pressure to time ∆p/∆t. The
of perforated Berea sandstone at 0.1 (top) and 0.4 in. (bot- position of the front between the two fluids must be
tom) from the center of the perforation. calculated as a function of time from the flow rate and
previously measured porosity. They found that the
region near the perforation tunnel had permeability
waves, taking into account the stress amplification that
reduced only 20% from the original. Permeability was
occurs at contacts between sand grains. They noted
reduced by about 70% in a zone from 0.2 to 1.2 in.
that grain breakage also decreases sharply away from
from the perforation tunnel wall. Appendix Fig. 2 is
the perforation and that this can theoretically be
a reproduction of one of their permeability profiles.
related to permeability. A different approach is to pre-
Ramlakhan (1994) combined this method with
dict grain breakage using a smooth particle numerical
X-ray CT to study a complete perforation. Using
model.
X-ray CT to map the position of the interface between
the two fluids eliminates the need to calculate the
Laboratory view interface position from the volume of tracer injected.
Measurements can be taken over the entire length of
Careful laboratory examination has redefined our
understanding of perforation damage. As discussed
in the following, several lines of evidence show that 400
permeability damage is not necessarily related only
Radial permeability (md)
to the visible crushed zone adjacent to the tunnel wall. Perforation

300 radius
The material adjacent to an open perforation is not
highly compacted, at least in liquid-saturated sand- 200
stones, and shows little consistent permeability reduc-
tion. Instead, reduced permeability is observed outside
100
this crushed zone extending 0.1 in. or more from the
tunnel. Permeability damage is not uniform along the
length of the perforation but is thicker and more 0
0 1 2 3 4 5 6
severe near the entrance hole. Mechanical damage
Radial distance from perforation axis (cm)
patterns also reflect this observation. There is some
qualitative evidence for the mobilization of clay parti-
Appendix Figure 2. Permeability profile obtained using vis-
cles, which may plug the formation as they are swept cous flow transients (Rochon et al.,1995).
toward the perforation.

the perforation, even where permeability varies along

its length. X-ray CT data were also used to determine 0.40
Weighted porosity (CT) = 0.172
the local porosity. This provides both a quantitative Undamaged porosity (lab) = 0.171
map of shock-induced compaction and data to calcu-
0.30
late local flow rate and flow velocity. Methods similar
to those described by Withjack (1987) were applied
after processing the images for beam hardening
Porosity
0.20
caused by the pressure vessel.
Appendix Fig. 3 shows porosity results for a Berea
sandstone core perforated at 1500-psi underbalance 0.10
and 3000-psi effective stress with a 3.2-g charge.
Although the porosity varies considerably, there is no
evidence of a well-defined compacted zone near the 0
tunnel wall. Appendix Fig. 4 shows a porosity profile 0 20 40
at a point near the tip of the perforation. There is no Slice 3 radius (mm)
open tunnel, and parts of the metal jet remain sur-
rounded by shock-damaged rock. The radial extent of Appendix Figure 4. Porosity profile near the tip of a perfo-
ration containing charge and rock debris.
this damaged rock is consistent with the radius of the
tunnel at shallower depths, and it is probable that suf-
compressibility of the pore fluid in this case allows the
ficient underbalance would have loosened and
porosity near the perforation to be largely destroyed.
removed this material to form an open tunnel. The
X-ray CT analysis has also demonstrated that per-
porosity profile shows that the material is compacted,
meability damage is not distributed uniformly along
with porosity reduced from 0.18 to about 0.12. Rock
the length of the perforation. Halleck et al. (1992)
beyond this is at the original porosity.
reported results of flow experiments performed in an
The extent of porosity reduction is a function of the
X-ray-transparent vessel (Appendix Fig. 5). A 6.5-g
pore fluid present at shot time. The vast majority of
charge was used to perforate a Berea sandstone core
perforation flow tests have been performed in liquid-
at 1500-psi effective stress and 500-psi underbalance.
saturated rocks. As noted previously, compaction is
After flushing the perforation with 20 L of odorless
limited to the volume of rock and charge debris that
mineral spirits (OMS) at 20 cm3/s, a di-iododecane
would be swept out if sufficient surge flow occurred.
tracer was substituted for the OMS. The entire sample
However, experiments by Bird and Dunmore (1995)
was scanned at 0.25-in. intervals after 100 cm3 of
in gas-saturated rock samples indicate that the high
tracer and after an additional 50 cm3 of tracer. The
first 100 cm3 of tracer saturated the porous packing
0.40 around the core and started to flow into the sample.
Weighted porosity (CT) = 0.150
Undamaged porosity (lab) = 0.171 The permeability of the porous packing (20/40-mesh
bauxite proppant) is on the order of 100 darcies at
0.30 these confining stresses. At this low flow rate the pres-
sure drop through the porous packing is negligible,
and the fluid pressure applied to the exterior of the
Porosity
0.20
core is assumed to be uniform along its length.
Appendix Fig. 6 illustrates the profile of the radial
advance of the flow found along the length of the
0.10
core. The profile was obtained by measuring the radial
advance of the flow front that took place during the
0
second (50-cm3) tracer injection. The data have been
0 20 40 corrected for decreasing circumference as the front
Slice 2 radius (mm) moves inward to obtain an average local flow rate.
Very little advance is seen near the entrance hole. The
Appendix Figure 3. Porosity profile adjacent to an open majority of flow takes place along the center of the
perforation tunnel was obtained from X-ray CT data.

Aluminium vessel
Accumulator Gun
Jacket and packing
Flow distributor
Sample
Simulated wellbore
Cement and casing
Appendix Figure 5. Apparatus for X-ray CT observations of perforation flow.
0.25 15
Distance from perforation center
r = 1.5 cm
0.20
r = 2 cm
Fraction of total flow
r = 3 cm
Brinell hardness (kg/mm)
0.15 r = 4 cm
10
Perforation
0.10
0.05
5
0
0 5 10 15 20 25
Distance from entrance hole (cm) Perforation depth
0
0 5 10 15 20
Appendix Figure 6. Profile of flow rates along the length of
a perforation was obtained by X-ray CT observations. Distance from entrance hole end (cm)
Appendix Figure 7. Map of indentation hardness around a

open perforation tunnel. Little flow is observed near perforation in a weak sandstone.
the perforation tip, where deformed rock and liner
metal have not been removed. The permeability distri-
bution appears to depend on specific factors such as damage is more severe near the entrance hole than
the character of the rock and the underbalance condi- near the tip. Although there is no direct evidence link-
tions. Other (unpublished) tests show more flow near ing mechanical weakening and permeability damage,
the entrance hole. the pattern observed is consistent with the radial and
Evidence for damage at some distance from the axial permeability distributions illustrated previously.
perforation comes from indentation hardness data.
Halleck et al. (1995) presented data based on the
method of Santarelli et al. (1991) for mapping hard- Role of underbalance in improved
ness around a perforation in a weak sandstone.
Appendix Fig. 7 shows that mechanical weakening
perforation flow
extends to the boundary of the 4-in. core. Similar The flow efficiency of a perforated core has been
unpublished data from a 151⁄2-in. diameter core show shown to depend on the difference between the reser-
that damage extends several inches and is not caused voir and wellbore pressures at shot time. Under-
by boundary effects. Appendix Fig. 7 also shows that balance perforating has become a standard practice

based on the work of Bell (1984) and others. Under- a finite-element simulation of wellbore flow with
balance surge flow is thought to improve perforation idealized perforations to show that the actual flow
flow efficiency in three ways: depends on shot spacing, phasing and degree of
drilling and perforation damage. Typical cases lie
• Surge flow prevents the invasion of wellbore fluids
between the radial and axial boundary condition tests.
and fines during the period between perforating and
The modified test procedure thus allows for either
drawing down the well to start production.
condition and can be changed to produce intermediate
• Surge flow removes some or all of the debris from conditions.
the tunnel. Using a form of this procedure, Hsia and Behrmann
• Surge flow removes some or all of the matrix per- (1991) found a generally monotonic increase in core
meability damage from the rock surrounding the flow with increasing underbalance. They confirmed
tunnel. that for 200-md Berea sandstone, approximately
The procedure has been proved generally effective 1500 psi was required to obtain a free-flowing perfo-
in the field. For specific applications, questions remain ration. This is consistent with the earlier axial bound-
as to what degree of underbalance is necessary, what ary-condition data of Halleck and Deo (1989), sug-
surge flow volume is required and whether the proce- gesting that the applied pore pressure boundary
dure is effective in a particular situation. Techniques conditions have little effect on surge cleanup.
to address these questions are provided in the follow- King et al. (1986) compiled field data to correlate
ing section. the effectiveness of acid stimulation in wells with for-
mation permeability and perforating underbalance.
They found a strong relationship among formation
Laboratory and field observations permeability, underbalance used and effectiveness of
later acid treatment. It is assumed that if acid has no
Early laboratory studies (Bell et al., 1972) suggested effect, the underbalance used was sufficient to clean
that 200-psi underbalance was sufficient to clean a the perforation. Although the cause and effect rela-
perforation in 200-md Berea sandstone, and test pro- tionship is not clear, the data strongly suggest that
cedures in the fourth edition of API RP 43 (1985) lower permeability rock requires higher underbalance.
were based on this number. Later, Halleck and Deo The minimum underbalance predicted by these data is
(1989) showed that the initial flow performance was less than for laboratory tests. The laboratory tests seek
only 40% of that expected until an underbalance in conditions for zero skin whereas the field data are
excess of 500 psi was applied. About 1500-psi under- based on conditions for which acid did not improve
balance was required to ensure a perforation with the well. The latter may not actually represent ideal
100% of its expected flow. They also showed that pro- perforation flow.
duction flow, even at high rates, could not completely
clean the perforation if insufficient underbalance was
used initially. Models
As part of an extensive API-sponsored project
(Halleck, 1989, 1990), a different test procedure To extend these results to a wider range of formations
(RP 43 Section 4, 1991) was developed that better and to make them reliable as practical guidelines, the
simulates in-situ conditions. This procedure provides mechanics behind underbalance perforating must be
for radial flow into the perforation and is performed placed in a firmer theoretical foundation. Tariq (1990)
under hydrostatic pressure to simulate overburden began this effort by hypothesizing that cleanup occurs
rock stress. An accumulator is attached to the well- when the velocity of transient flow in the rock sur-
bore chamber to simulated wellbore storage effects. rounding the perforation reaches a critical value. He
The radial flow boundary condition is based on the suggested that the Reynold’s number (as defined for
ambient stress tests of Regalbutto and Riggs (1988). porous media) is the appropriate parameter. Using
They suggested that in-situ perforations experience a finite-element model for perforation flow, he con-
this type of flow rather than that associated with the cluded that the critical Reynold’s number is about 0.05.
earlier (RP 43, 1985) tests, which applied pressure To illustrate how the basic model functions, con-
only to the rear of the test core. Deo et al. (1989) used sider a 4-in. diameter test core with a debris-free per-
foration. If the pressure in the perforation is instanta-

neously lowered to simulate an underbalance surge, permeability lowers the transient velocity vc from the
the pressure profile in the core can be calculated as matrix velocity vm.
a function of time. This is illustrated using a simple The model checks at each time step to determine
radial flow numerical model in Appendix Fig. 8 for whether flow in a particular element has exceeded
the particular case of a 200-md sample containing a predetermined critical value. If so, permeability for
1.8-cp fluid with 1000-psi underbalance. that cell is increased back to the undamaged value.
The pressure gradients at early times are steep and The progress of this cleanup can be followed with
associated with equally high flow rates. However, time. When each cell is cleaned up, the pressure gradi-
these high rates persist briefly and are present only ent and flow rate in the adjacent cell are increased.
near the walls of the perforation. Appendix Fig. 8 If the resulting velocity is high enough, this cell may
shows the maximum Darcy velocity (ignoring inertial then also be cleaned up. The process stops either
effects) as a function of radial position. Only rock when all cells have been cleaned or at the radius for
within 4 cm of the perforation experiences rates above which velocity no longer reaches the critical value.
0.005 m/s. The series of plots in Appendix Fig. 9 illustrates the
How do these transient flows clean a perforation? progress of the cleanup with the critical flow rate
Rerunning the model with a zone of reduced perme- based on a Reynold’s number of 0.1. In this case the
ability kc at the tunnel wall illustrates this. For this cleanup process terminates with some of the damage
case kc was reduced to 0.2k, as is commonly assumed. remaining. The plots show the maximum flow rate
This would result in about 0.4 times the ideal flow and the corresponding permeability profile with time.
rate if the damage were not removed. The reduced Higher underbalance would have resulted in cleanup
of the remaining damaged zone.
The underbalance for cleanup thus depends on the
0.010 required critical velocity, permeability of the damaged
1E–6 s
5E–5 s
and undamaged rock and thickness of the damaged
5E–4 s zone. A series of test cases for different rock perme-
Steady state
abilities was performed using the model, varying the
Velocity (m/s)
underbalance to find the underbalance required for

0.005 complete cleanup in each case. The results are plotted
in Appendix Fig. 10 with the original data of King et
al. (1986). The trends are similar, although the model
does not coincide with the field data. This may be due
to a number of factors:
0
• Lack of flow improvement after acid treatment in
Perforation radius the field may not actually indicate that perforations
were clean prior to the treatment.
4.0
• Range of fluid viscosities in the field may differ
from that used in laboratory experiments.
3.0 • Critical velocity was incorrectly assigned.
Pressure (MPa)
• Unrealistic damaged zone properties and geometry

were assumed by the model.
2.0 The model assumes instantaneous pressure drop in
the tunnel, implying that charge debris and damaged
rock have already been removed. In reality, the time
1.0 frame for reduced pressure in the perforation must
0 0.02 0.04 0.06 0.08 0.10 be on the order of a millisecond. The penetration itself
Radial position (m)
takes on the order of 10–4 s, and flow through a debris-
filled tunnel cannot lower the pressure much faster
Appendix Figure 8. Transient flow rates and pressures for than that. An improved model must also take into
cylindrical inward flow into an ideal perforation.

103
1.0 Not improved by acid
Initial Improved by acid
k/k c and vm /vc
0.5 102
Formation permeability (md)

Maximum velocity
Permeability
0 101
0 0.1
1.0
100
k/k c and vm /vc
0.5
t = 1E–6 s 10–1
100 1000 10,000
Total underbalance (psi)
0
0 0.1
Appendix Figure 10. Minimum underbalance required for

1.0 complete cleanup based on numerical model and Tariq’s
(1990) Reynold’s number criterion.
k/k c and vm /vc
0.5
There are a number of additional difficulties with
t = 5E–5 s
this simple model. First and foremost is whether
Reynold’s number should be used to establish critical
0 flow velocity. To begin with, the definition of Rey-
0 0.1
nold’s number is controversial. The most commonly
accepted method is to use the ratio β/α to describe
1.0 the length dimension:
k/k c and vm /vc
β ρv
N Re = , (1)
α µ
0.5
Steady state
where α and β are the coefficients in the Forscheimer
equation, ρ is the density, v is the flow velocity, and µ
0 is the viscosity. Using this equation requires knowl-
0 0.1
edge of both k(1/α) and β for a given rock. Because
Radial position (m) these data are seldom available, empirical correlations
between β and k are used. Such correlations are gener-
Appendix Figure 9. Cleanup of the reduced-permeability ally of the form β proportional to k–m. The value of
zone around a perforation by high-velocity transient flow. m varies considerably in the literature. Tariq (1990)
used a value of 1.65, but a more widely accepted
account the effects of rock remaining in the tunnel value of 1.2 results in a much steeper curve in
as well as its removal. Additional model runs have Appendix Fig. 10, which fails to reproduce the field
been made in which pressure in the tunnel is lowered data. In addition, these correlations are based on intact
during a 1-ms period (Appendix Fig. 11). Although rock and may not apply to the damaged rock around
there is some effect on cleanup, the basic results are the perforation. Finally, the use of Reynold’s number
unaffected. implies that minimum underbalance should also

In an effort to resolve problems associated with

the use of Reynold’s number, Behrmann (1995) intro-
1.0
k/k c and vm /vc duced a cleanup criterion based on the drag force on
a particle in a stream of fluid. By using an adjustable
shape factor, Behrmann was able to match the
0.5 observed field data and eliminate the dependence
t = 1E–4 s
on viscosity. In addition, model improvements such
Maximum velocity as fully three-dimensional finite-difference simulations
Permeability
0 (Y. Dogulu, pers. comm., 1996) allow the study of
0 0.1
anisotropic, non-Darcy transient flow, including that
through the tunnel debris, as well as comparison of
1.0 different cleanup criteria to investigate the effects of
nonuniform damage.
k/k c and vm /vc
0.5
Synergy of models and experiments
t = 5E–4 s
One implication from modeling is that if cleanup is to
0 occur, it must be during the short duration of high-rate
0 0.1 transient flow. Evidence comes from the data of
Bartusiak et al. (1997), who used a modified RP 43
1.0 test procedure to include simulation of the fluid capac-
ity and the impedance of the reservoir itself. In con-
k/k c and vm /vc
ventional testing, the surge flow volume is limited to

0.5
decompression of the fluids in the pore structure of the
t = 1E–3 s
sample and associated pore pressure plumbing. In the
revised system, an accumulator attached to the sample
maintains constant far-field reservoir pressure during
0
0 0.1 surge flow. This pressure is applied to the sample
through a series of auxiliary cores that simulate the
flow impedance of the rock surrounding the well.
1.0 The result is a more realistic simulation of downhole
conditions during perforating.
k/k c and vm /vc
The simulated reservoir also allows an indirect

0.5 measurement of flow rate through the perforation
Steady state
during the surge flow process. By measuring the tran-
sient pressure drop across the simulated reservoir, the
0 flow rate through it can be determined from its known
0 0.1 impedance. After pseudosteady-state conditions are
Radial position (m) attained, the flow rates through the two are approxi-
mately equal. With proper calibration, this leads to
Appendix Figure 11. Cleanup while the pressure in the tun- a real-time measurement of the flow rate starting
nel is lowered during a 1-ms period.
about 0.3 s after detonation. The measurement in turn
can be converted to time-resolved values of core flow
depend on fluid viscosity. Although this is true in the efficiency (CFE) by dividing the result by the pre-
model, it is not supported by King et al.’s field data. dicted ideal flow rate. Appendix Fig. 12 illustrates
Data from laboratory tests (L. A. Behrmann, pers. the development of CFE with time for a test in Berea
comm., 1997) also suggest that viscosity is not an sandstone at 1500-psi underbalance. The CFE rises
important factor in determining minimum under- sharply in the first second to 0.8, increases to about
balance for fluids between 1.5 and 400 cp. 0.85 over the next few seconds and then remains

as they move radially away from the perforation.

1.0 Rock near the perforation is stressed to failure and
flows with pseudo-plasticity, whereas rock farther
away remains elastic. After the shock dissipates, the
failed material returns to an elastic state, although
with a lower strength. The net effect of the permanent
plastic displacement is a zone of rock that remains
under compression even after the shock dissipates.
This phenomenon has been frequently observed in the
Core flow efficiency
laboratory. Perforated cores encased in steel sleeves

0.5 are observed to fracture when the steel sleeve is cut
open. Similarly, in a pressurized core, no fractures are
observed after perforating at 6000-psi overburden
pressure. The core subsequently fractures when the
stress is reduced to about 250 psi. Radial fractures are
frequently observed in samples recovered from such
pressurized tests. Although they are usually assumed
to be hydrofractures, closer examination reveals that
the widest part of the fracture is at the outside of the
0 core rather than at the perforation, suggesting that the
0 50 100
fractures grew inward when the overburden pressure
Time (s)
was released.
The residual stress may be partly responsible for
Appendix Figure 12. CFE development with time for a test reduced flow efficiency as well as for the excessive
in Berea sandstone at 1500-psi underbalance.
breakdown pressures observed during some hydro-
fracture treatments. Warpinski (1983) found such
constant. This value is consistent with quasi-static effects in field tests in a mineback experiment in
flow tests performed later with conventional proce- welded tuff and suggested that residual stresses may
dures. Another test in the series shows that reducing be responsible. The effects are not well understood,
the total flow volume, by reducing the wellbore stor- and further experimental and theoretical studies are
age volume, had little effect on this process. These needed.
results support the idea that cleanup occurs at early
times, when transient flow rates are highest. The prac-
tical result is that it should not be necessary to design Damage to natural fractures
the completion procedure for a minimum surge vol- Ideally, if a perforation intersects a natural fracture
ume if sufficient underbalance is applied. and a connection is formed between the fracture sys-
tem and the wellbore, permeability damage to the
walls of the perforation would not matter because
Other phenomena the flow would bypass any damage via the fracture
Several additional perforating phenomena are worthy system. However, laboratory data (Halleck, 1996a,
of mention. These are residual stresses, damage to nat- 1996b; Halleck et al., 1996) show that perforating can
ural fissures or fractures, injection of detonation gases damage fracture permeability in two ways. Under
into the formation and wettability effects. some circumstances, jet metal can invade the fracture
and reduce its conductivity. More generally, shock
stresses deform the surrounding rock and close the
Residual stresses fracture where it intersects the perforation. Appendix
Fig. 13 is a photograph of such an intersection. The
Perforations in porous rocks are left with a compres- fracture was injected with epoxy after perforating.
sive residual stress surrounding the tunnel. This is Closure of the fracture near the perforation is circled.
caused by shock waves that decrease in amplitude

solely with a jet of detonation gases. Little has been

published regarding the associated gas jet.
In most completions, the phenomenon may not be
of importance. However, in unconsolidated rock, gas
injection may lead to high transient pore pressures,
causing fluidization of the sand bed. The result would
be that a large zone of sand around the perforation
would act temporarily like a liquid. Such a phenome-
non observed in unpublished laboratory experiments
resulted in the immediate production of about a kilo-
gram of sand. There was no evidence of a perforation
tunnel in the recovered sample, although charge debris
was found about 20 cm deep in the sand.
Appendix Figure 13. Fracture closure where it intersects
with a perforation. The closure is a result of shock stresses.
Wettability phenomena
Finally, the effect of wettability on perforation cleanup
Injection of detonation gases should be mentioned. The presence or absence of
water has a large effect on both perforation cleanup
Another common laboratory observation is the pro- and sand production. Halleck and Deo (1987) com-
duction of gas from test cores that were initially liquid pared results in kerosene-saturated samples with those
saturated. Flow tests performed using initial labora- in samples that contained irreducible brine. They found
tory procedures had the outlet of the perforation at that the brine-containing samples had higher values
some elevated pressure to simulate in-situ bottomhole of CFE and cleaned up at lower underbalance condi-
pressure. In a subsequently modified API test proce- tions. Injection of brine into a previously water-free
dure, this requirement was eliminated to simplify the perforation resulted in rapid cleanup during subsequent
test. Wellbore pressure can be lowered to the ambient kerosene flow. Halleck and Damasena (1989) observed
pressure so that outlet flow rates are simpler to mea- large increases in sand production rates when simu-
sure. Early production flow in such tests contains a lating water breakthrough in a weak sandstone.
high gas cut. Furthermore, the CFE is observed to first All these observations indicate that wettability, sur-
decrease and then increase slowly to a final value that face tension and capillary pressure effects are involved
can be associated with the absence of further gas pro- in the transport of solids from the perforation damage
duction. This phenomenon suggests relative perme- zone. These effects should not be discounted in pre-
ability effects resulting from gas in the pore structure. dicting the effects of underbalance surge cleanup.
The phenomenon was not observed in earlier tests
because the applied backpressure kept the gases in
solution. It is therefore recommended that the modi-
fied procedure be performed with backpressure or that
Conclusions
the flow be sufficient to elute all the gas before the Permeability damage is distributed much differently
flow rates are measured. than assumed by idealized models. Permeability
There are two possible mechanisms for gas injec- varies substantially with radius. Reduced permeability
tion even when substantial underbalance is applied. is not necessarily associated with the visible crushed
One is that detonation gases may cause temporary zone. This observation suggests that permeability
overbalance in the wellbore before they cool. reduction may be more a function of fines migration
Laboratory tests with transient pressure gauges show than grain breakage. Mechanical damage in weaker
short-lived wellbore pressure increases but not to the rocks extends several inches around the perforation.
point of overbalance. The more likely explanation The extent of the damage is not uniform along the
(Bell, 1984) is that the metallic jet has a jet of detona- perforation but is more severe near the entrance hole.
tion gas associated with it. Early shaped charges, The so-called crushed zone surrounding a clean
called Munroe jets, had no metal liners and penetrated perforation is not significantly compacted in liquid-

saturated rocks, but it may be in gas-saturated ones. velocity through the rock around the perforation and
Compacted material exists in the rock near the center occurs at early times, when transient rates are highest.
of the perforation, but it is seen only at the perforation Large surge flow volumes should not be necessary
tip or in tests with insufficient underbalance to remove where sufficient underbalance is applied.
it. Perforation cleanup is related to transient flow

◆
Post-Treatment Evaluation and
Fractured Well Performance
B. D. Poe, Jr., Schlumberger Dowell
12-1. Introduction 8800

Evaluation of the fracture geometry, well deliver-
ability and reservoir performance of fractured wells 9000
with post-treatment well and reservoir responses has
Thermal Prefracture
been extensively investigated. Post-treatment mea- conductivity profile
surements of the created fracture geometry have 9200 effects
been obtained with various logging and micro- Static

log
seismic fracture mapping techniques. 9400
Pressure transient analysis has been used as a post-
Hole depth (ft)

treatment evaluation procedure for estimating the Postfracture
fracture extension into the reservoir as well as for 9600 profile
obtaining estimates of the fracture conductivity and

reservoir properties. With the reservoir and fracture 9800
parameter estimates obtained with these and various
other post-treatment analysis techniques, a production
systems analysis can be performed to determine the 10,000
post-treatment deliverability of the fractured well. Fracture top
Profiles
Perforations separate
10,200
12-1.1. Fracture mapping techniques

10,400
The post-treatment evaluation of fracture geometry 175 200 225 250 275
using logging techniques is usually limited to near- Temperature (°F)
borehole effects, such as estimation of the created
fracture height with temperature surveys, gamma ray Figure 12-1. Pre- and postfracture temperature surveys
logs and spinner surveys in cased holes. Because (Dobkins, 1981).
these are near-well measurements and the wellbore
stress concentration (see Chapter 3) encourages the
fracture to leave the wellbore, the measurements are fracturing effects. Radioactive tracers in the fractur-
lower-bound estimates of height. ing fluid, proppant or both can be monitored with
Temperature surveys are useful in estimating the gamma ray logs to determine which zones took fluid
wellbore fracture height that results from cooling of or proppant during the fracture treatment. Various
the formations that take the fracturing fluid during the stages may be identified by using different radioiso-
treatment. An example of prefracture static (after cir- topes. Clearly the radioisotopes used should not natu-
culating the wellbore) and postfracture temperature rally occur extensively. Post-treatment spinner sur-
surveys of a well are presented in Fig. 12-1. The surveys simply indicate which sets of perforations in a
vey after circulating the well, but before perforating, cased hole are producing fluid. A wellbore density
enables identifying thermal conductivity anomalies log (Fig. 12-2) is used to determine what types of flu-
caused by various formation layers and variation in ids are being produced. However, in an openhole the
the cement thickness. The anomalies can be ignored spinner survey may be used effectively to estimate
or compared with the postfracture survey to identify which layers of the reservoir were stimulated.

Gamma ray log Depth Production log SP

(ft)
9200
temperature profile
5000
Postfracture
Gas flow
Water flow 9300
Postfracture
Depth (ft)
gamma ray
9400 Base
gamma
Post- ray
treatment
gamma
ray Fracture
5100 zone
9500
Pre-
treatment Density Perforations
gamma
ray
Temperature
9600
Figure 12-3. Gamma ray measurements match the tem-

perature survey (Dobkins, 1981).
5200
leaving the wellbore. In wellbores containing opti-

Figure 12-2. Production log and gamma ray log informa-
tion from a fractured East Texas well.
cally clear fluids, opening and closing of the created
fracture can be visually observed to determine its
height at the wellbore.
An example of using production logs with radioac- Zemanek et al. (1969) investigated the use of a
tive tracers to identify fracture height and the zones sonic borehole televiewer to obtain estimates of the
that were fractured is in Fig. 12-2. The correspon- wellbore fracture height in openholes. Although the
dence between the gamma ray log and the spinner sur- acoustic principles upon which this tool is based
vey indicates which intervals were fractured. The cor- should result in a robust, accurate fracture identifica-
respondence between gamma ray logging of a fraction procedure, many borehole conditions limit its
tured well and temperature logging is presented in reliability in practice. Some of these conditions are
Fig. 12-3. borehole ellipticity, wellbore deviation resulting in
In openholes, the many measuring devices that can noncentralization of the tool and very narrow frac-
be used to estimate the wellbore fracture height tures that have widths at the wellbore that are signifi-
include the sonic borehole televiewer, Formation cantly less than the acoustic wavelengths emitted by
MicroScanner* tool, impression packers and down- the tool.
hole closed-circuit television. The direct method of The Formation MicroScanner tool is generally of
measuring fracture width and height in an openhole comparable accuracy to the sonic borehole televiewer
with a downhole closed-circuit television was inves- in an openhole. The resistivity buttons on the Forma-
tigated by Smith et al. (1982) and may provide the tion MicroScanner tool image the inside of the well-
most reliable estimate of fracture height at the wellbore wall to allow determination of the created frac-
bore. The results clearly show the fracture plane ture height. As for the prefracture (but postcirculating)
temperature survey (Fig. 12-1), a complementary pre-
12-2 Post-Treatment Evaluation and Fractured Well Performance

fracture survey enhances the interpretation of postfrac- shows typical techniques for surface placement (a
ture imaging surveys. circle with a diameter equal to fracture depth) and
Impression packers are probably the oldest direct telemetry analysis (tilt vectors toward and normal
measurement devices for estimating the created frac- to a vertical fracture) for the use of tiltmeter arrays.
ture azimuth and orientation at the wellbore. The Davis (1983) reported analysis procedures for the
earliest reported investigation concerning the use of surface displacement of horizontal and vertical frac-
impression packers for this purpose is by Fraser and tures from tiltmeter measurements. The accuracy and
Petitt (1962). Inflatable packers with deformable rub- reliability of tiltmeter analysis for fracture mapping
ber elements are used to take an impression of the purposes are limited primarily by environmental
borehole surface and record the created fracture fea- conditions. The sensitivity of the devices is such that
tures. Wellbore deviation surveys are commonly run the effects of vehicular traffic, wind, rain and any
in conjunction with impression packers to obtain an number of other environmental influences may be
estimate of the fracture azimuth. greater than the surface tilts caused by a subter-
Active mapping techniques have also been used ranean hydraulic fracture.
during fracture injection to estimate the growth of Triaxial borehole seismics have also been used
the fracture. Among the techniques that have proved successfully to map the growth of subterranean
the most useful for mapping the growth of hydraulic hydraulic fractures (Albright and Pearson, 1982;
fractures are tiltmeter and triaxial borehole seismic Dobecki, 1983; Batchelor et al., 1983). The fracture
analyses. In addition to the previously discussed tele- azimuth, lateral extension from the wellbore and
vision tool, tiltmeter arrays measure the change of growth into adjacent layers above and below the pre-
the earth’s surface tilt caused by the creation of sumed target zone can be obtained using acoustic
hydraulic fractures at depth in the earth. Where tilt- and seismic emissions recorded from the fracture
meter analyses can be successfully employed, esti- during injection and shut-in. Subsurface acoustic
mates of the fracture azimuth and potentially the telemetry is used to estimate the distance to recorded
extent from the wellbore can be obtained. The main microseismic events from a variety of geophone
limitation on the effective use of tiltmeters is gener- receivers by using the difference in arrival time
ally a function of the depth of the reservoir being between the compressional and shear waves of the
hydraulically fractured relative to the volume of the microseismic events. Section and plan views of basic
fracture. Volume directly increases the response, and principles for the use of borehole seismics are shown
depth attenuates the surface response. Figure 12-4 in Fig. 12-5.
• Example of microseismic imaging
12 Although microseismic imaging of hydraulic frac-
11 1 tures has been used since the 1970s (Albright and
x-axis
14 Pearson, 1982; Thorne and Morris, 1987; Hart et
18
y-axis 13
al., 1984; Vinegar et al., 1992; Truby et al., 1994;
10 2
17 Warpinski et al., 1995), the technique was used pri-
marily for large-scale research experiments because
of receiver limitations and processing difficulties.
3
However, improvements in downhole receivers and
9 15 16
Fracture the advent of high-speed portable computers have
made this technology available for more general
application. In addition, the technology has reached
4 the stage where it is possible not only to image the
8
final size and shape of the fracture but also observe
the manner in which fracture growth occurs.
5
One large-scale research experiment where
7 10 nrad
6
microseismic imaging was successfully used is the
M-Site project in the Piceance basin of Colorado,
Figure 12-4. Vertical hydraulic fracture tiltmeter array which was jointly funded by the Gas Research
design and computed vectors.

6000 50
B sand propped fracture
5000

Bottomhole 40
Pressure (psia)
pressure
4000
Closure pressure
30
Seismic 3000 Stress (psi)
source 3000 5000 20
2000 Rate
Fracture 30 ft B sand 10
S 1000
P
0 0
0 30 60 90
Time (min)
Section
Figure 12-6. Treatment data and stress configuration for
microseismic experiment (Warpinski et al., 1996).
Source
S
4300
P Monitor well
ne First time period MWX-2
e pla 4400
MWX-3 (five level)
tur
ac
Fr
Plan 4500
B sand
Figure 12-5. Section and plan views of borehole seis- 4600
mometer monitoring events in a fractured well (Dobecki,
1983). 4700
–500 –400–300–200–100 0 100 200 300 400 500
4300
Institute and the U.S. Department of Energy Second time period MWX-2 Monitor well
MWX-3 (five level)
(Peterson et al., 1996). Multilevel triaxial receiver 4400
Depth (ft)
arrays in two offset monitoring wells were used

4500
to develop time-growth images of hydraulic frac- B sand
tures induced in a nearby well (Warpinski et al., 4600
1996). Figure 12-6 shows treatment and formation
data for one of the fractures conducted in a sand- 4700
–500 –300 –100 0 100 300 500
stone at this location. In this fracture, approxi-
4300
mately 600 bbl of 40 lbm/1000 gal of crosslinked Monitor well
Final image MWX-2
gel and 78,000 lbm of sand were injected at MWX-3 (five level)
4400
20 bbl/min.
Figure 12-7 shows side-view images at three 4500
times for the microseismic locations associated B sand
4600
with the fracturing process. In the first time period,
linear gel in the wellbore and a crosslinked-gel pad 4700
have been injected, but little height growth has –500 –300 –100 0 100 300 500
occurred because of the stress contrasts in the abut- Along fracture length (ft)
ting shales. In the second time period, sand concen- Figure 12-7. Side-view microseismic images for three time
trations of up to 4 ppg have been injected, and periods.
height growth is just beginning to occur. The final
microseismic image shows that as net pressure
increased, fracture height growth became markedly

asymmetric, with little growth downward and • Example of downhole tiltmeter imaging
upward growth on each wing. A plan view of the The M-Site experiments from which the microseis-
microseismic images in Fig. 12-8 yields an azimuth mic images were produced is also the first site where
of about N75°W. As can be seen most clearly in the downhole tiltmeters were employed to obtain frac-
plan view, the microseismic image is an envelope ture diagnostic information (Branagan et al., 1996).
of points in which the fracture is generally con- An array of six biaxial tiltmeters was emplaced
tained, as the microseisms usually occur some dis- approximately 300 ft normal to the hydraulically
tance away from the actual fracture. fractured interval (Fig. 12-9). Measurement of the
downhole tilt as a function of time provided data on
the rotational deformation of the formation induced
600
Monitor well by the fracture opening. In the example data shown,
MWX-3 Five level
500 Regression (N75°W)
the tiltmeter amplitudes deviate from zero as the frac-
ture opens, with the maximum tilt occurring at shut-
400 in. The tilt decline during shut-in can also be observed
North–south (ft)
as the fracture slowly closes on the proppant.

300 Figure 12-10 shows the measured tilt at each
tiltmeter level just prior to shut-in, as well as
200 some modeling results for this treatment. As noted
MWX-2 in Chapter 6, the tilts adjacent to a hydraulic frac-
100 ture take on a characteristic S shape, but the lower
Monitor well part of the S is missing because no instruments
0
were placed at those depths. The available data
–500 –400 –300 –200 –100 0 100 200 300 400
are clearly sufficient to estimate the height of the
East–west (ft)
fracture and the width to produce the measured
Figure 12-8. Plan view of a microseismic image. distribution. Simple homogeneous models bracket
the height and give a good initial estimate but do
Downhole Tiltmeter Data

Instrument
10
Well
Level 3
5
Tilt (µrad)
Level 5 Level 4
Level 6
0
Level 2
Level 1 Level 6
–5
0 50 100
Time (min) 5
Tiltmeters
4
Shale 3
Sandstone 2
Hydraulic Shale 1
fracture
Figure 12-9. Tiltmeter experiment and example data.

4250
3D finite element, h = 67 ft
Fracture 5B tiltmeter
Homogeneous, h = 60 ft
Homogeneous, h = 90 ft
4350
Depth (ft)
4450
4550
4650
–8 –4 0 4 8 12
Tilt (µrad)
Figure 12-10. Tiltmeter and modeling results of a fracture treatment.
not entirely match the character of the measured tions to the body of knowledge concerning the
data because of the effects of layering. In this behavior of fractured wells include investigations by
example, the 60-ft fracture matches the tail-off, van Poollen et al. (1958) and Prats (1961). Steady-
whereas the 90-ft fracture matches the peak state conditions are also assumed in the van Poollen
amplitude. Both models, however, are offset from et al. and Prats models, which consider the response
the measured data, most likely because of vertical of both finite- and infinite-conductivity fractures.
asymmetric width effects. Finite-element analyses Prats introduced the concepts of dimensionless frac-
give a much better match of the character, as they ture conductivity and effective wellbore radius for
can include layering effects (as well as they are vertically fractured wells, in which the effective
known), but they are much more difficult to pro- wellbore radius is demonstrated to be a function of
duce. Nevertheless, the modeled height of the the fracture length and dimensionless fracture con-
fracture (67 ft) closely matches the microseismic ductivity. For infinite-conductivity fractures, the
image of the fracture and yields a width in good effective wellbore radius is demonstrated to be equal
agreement with the elastic deformation commonly to one-half of the fracture half-length in a fully
used in geometry models. developed flow pattern (Fig. 12-11).
1
12-1.2. Pressure transient analysis 0.5
Pressure transient analysis has been used success-

fully to obtain estimates of reservoir and fracture 0.1
properties. This section provides a review of devel-

r w´/x f
opments related to fractures.

0.01
The earliest investigation reported in the literature
concerning the pressure transient behavior of verti-
cally fractured wells is by Muskat (1937). Muskat 0.001
used an analytic fractured well model that assumes 0.01 0.1 1 10 100 1000
steady-state flow conditions to investigate the pres- Relative capacity parameter, a = π/2CfD
sure distributions and fluid entry patterns in the
vicinity of a vertical fracture. Other early contribu- Figure 12-11. Concept of effective wellbore radius for ver-
tically fractured wells (Prats, 1961).

The performance of vertically fractured wells

12
under pseudosteady-state flow conditions was inves-
tigated by McGuire and Sikora (1960) using a poten-
tiometric analog model. The results of that study 10
provide the basis for evaluation of the stabilized Raymond and Binder (1967)
Flow efficiency, J F /J
post-treatment performance of vertically fractured 8
McGuire and Sikora (1960)
wells. The stabilized productivity index ratio
increase from the fracture stimulation of a well 6
is presented in Fig. 12-12. Prats (1961)
The earliest work reported in the literature that
4 Tinsley et al. (1969)
considers unsteady-state flow conditions on the tran-
sient behavior of vertically fractured wells is by
Dyes et al. (1958). Prats et al. (1962) and Russell 2
and Truitt (1964) also investigated the transient 103 104 105
Relative conductivity, k f w/k (ft)
behavior of vertically fractured wells under transient
flow conditions. Russell and Truitt used an explicit
Figure 12-13. Flow efficiency ratios from various studies.
finite-difference simulator to investigate the transient
behavior of an infinite-conductivity vertical fracture
and found that the classic semilog analysis tech- The investigations reported by Clark (1968) and
niques developed for unfractured wells could be Milheim and Cichowicz (1968) apply linear flow
applied for the evaluation of transient behavior of concepts to analysis of the pressure transient behav-
vertically fractured wells during the pseudoradial ior of infinite-conductivity vertically fractured wells
flow regime. Figure 12-13 compares the results of and demonstrate that a Cartesian graph of bottom-
various early studies estimating the stabilized flow hole pressure versus the square root of time results in
efficiency improvement resulting from fracturing a a linear transient behavior, thus providing an analysis
well. The discrepancies observed between the results procedure for the early transient formation linear
of the various studies are due primarily to the vari- flow regime.
ous assumptions associated with the development of Gringarten and Ramey (1973, 1974) first introduced
the models used. Of the correlations presented in application of the instantaneous Green’s function and
Fig. 12-13, McGuire and Sikora’s (1960) results Newman’s product method for the development of
have been found to generally represent the postfrac- solutions to a wide variety of problems of fluid flow
ture pseudosteady-state production increase better in porous media. Among the solutions developed by
than the other correlations. those authors are analytic solutions for the pressure
transient behavior of uniform flux and infinite-conduc-
tivity vertical fractures and a uniform flux horizontal
14
L/re = 1
0.9
fracture. The instantaneous Green’s function and
)
12 0.8
0.7
Newman’s product approach introduced by Gringarten
ln 0.472 rw
re
10 0.6 and Ramey and Gringarten et al. (1974) has subse-

7.13
0.5
8 quently been used extensively to develop solutions
0.4
6 to investigate the transient behavior of a wide variety
0.3
(
0.2
of well types and reservoir configurations. An addi-
4
(J/Jo)
0.1 tional result of Gringarten et al.’s study is the demon-

2
stration that the infinite-conductivity vertical fracture
0 response could be obtained with the uniform flux ver-
102 103 104 105 106
tical fracture solution, in which the pressure transient
kfw 40
behavior is evaluated at a dimensionless spatial posi-
Relative conductivity,
k √A tion in the fracture from the wellbore equal to 0.732.
Using the Source and Green’s function approach
Figure 12-12. Productivity index ratio for vertically frac- reported by Gringarten and Ramey and Gringarten
tured wells (McGuire and Sikora, 1960).

et al., numerous studies have been reported in the

Wellbore
literature concerning particular aspects of the perfor- kf
mance of vertically fractured wells. Among these
are Cinco-Ley et al.’s (1975b) study of the transient w
behavior of inclined wells and fractures and
Raghavan’s (1976) study concerning the transient
k
behavior of partially penetrating infinite-conductivity
vertical fractures, which uses the Gringarten et al. Fracture
relationship between uniform flux and infinite-conduc- Impermeable h
tivity vertical fractures. A similar study on partially boundaries
penetrating infinite-conductivity vertical fractures
was later made by Rodriquez et al. (1984). pe
In the classic work by Cinco-Ley et al. (1978), the
development of an analytic solution for the pressure
transient behavior of a finite-conductivity vertical xf
fracture provides a sound theoretical basis for inter-
pretation of the transient behavior of vertically frac- re
tured wells. An idealized fracture used in that study is
presented in Fig. 12-14. Cinco-Ley and Samaniego-V. Figure 12-14. Geometry of an ideal vertical hydraulic
(1981b) later provided a detailed analysis of the tran- fracture.
sient behavior of finite-conductivity vertical fractures
and identified the various flow regimes that can be
exhibited. and formation linear flow regimes are in Figs. 12-15,
Cinco-Ley and Samaniego-V. (1981b) identified four 12-16 and 12-19, respectively. The specialized analyses
flow regimes that may exist: fracture storage linear developed in that study have become the theoretical
flow, bilinear flow, formation linear flow and pseudo- basis for many of the pressure transient analysis tech-
radial flow regimes. The reservoir and fracture fluid niques for fractured wells. Details of the procedures
flow patterns of a finite-conductivity vertical fracture involved in the Cinco-Ley and Samaniego-V. analysis
for the fracture linear flow (fracture storage), bilinear are presented in Section 12-2.
q
1
Figure 12-15. Fracture linear flow in a finite-conductivity vertical fracture.
Figure 12-16. Bilinear flow in a finite-conductivity vertical fracture.

Agarwal et al. (1979) investigated the rate transient multilayer evaluation procedure of Bennett et al. and
response of a finite-conductivity vertical fracture with Camacho-V. et al. for commingled systems and elim-
a finite-difference reservoir simulation model. The inates the requirement of linear flow to the fracture.
finite-conductivity vertical fracture type curve devel- This commingled reservoir analysis is amenable for
oped by Cinco-Ley et al. (1978) is extended in this the analysis of all flow regimes and combinations of
study so that it is more applicable for the production reservoir layer completion types.
response of massively fractured wells in low-perme- The transient behavior of vertically fractured wells
ability reservoirs. in dual-porosity reservoirs was studied by Houzé et
Cooke (1973) demonstrated the importance of a al. (1988), Ben Naceur and Economides (1988),
propped fracture and non-Darcy flow in the fracture. Cinco-Ley and Meng (1988) and van Kruysdijk
Guppy et al. (1982b) investigated the effect of iner- (1988). The studies of Houzé et al. and Ben Naceur
tial-turbulent (non-Darcy) flow on the pressure tran- and Economides consider infinite-conductivity verti-
sient behavior of finite-conductivity vertical fractures cal fractures, and the models of Cinco-Ley and
using an analytic model. Although Wattenbarger and Meng and van Kruysdijk were developed to study
Ramey (1968) and Holditch and Morse (1976) pro- the behavior of finite-conductivity vertical fractures
vided detailed qualitative insight into the effects of in dual-porosity reservoirs.
non-Darcy reservoir and fracture flow, respectively, The study by van Kruysdijk also considers the
the investigation by Guppy et al. provided the first effects of fracture face skin effect damage. This fol-
quantitative method for determining the reduction in lows earlier work by Cinco-Ley and Samaniego-V.
apparent fracture conductivity resulting from non- (1977, 1981a) in which the effects of flow restric-
Darcy fracture flow. tions in and along the face of vertical fractures are
The application of derivative analyses to the inter- compared with the corresponding vertical fracture
pretation of the transient behavior of infinite-conduc- cases without damage.
tivity fractures was reported by Tiab and Kumar Ozkan and Raghavan (1991) developed readily
(1980). Tiab and Puthigai (1988) developed analytic computable analytic pressure transient solutions for
expressions for the derivative behavior of infinite- uniform-flux vertical fractures in a variety of reser-
conductivity fractures and applied those relationships voir configurations. A major contribution of the
to analysis examples. Pressure derivative analyses study by Ozkan and Raghavan, in addition to the
for finite-conductivity vertical fractures were devel- development of the various solutions presented, is
oped by Wong et al. (1984). The Wong et al. analy- improved computational procedures for evaluating
sis procedure involves both pressure and pressure the transient response of vertically fractured wells.
derivative analyses to reduce some of the uniqueness A review of the limitations and uncertainties con-
problems that commonly occur with interpretation of cerning the design and evaluation of hydraulic fracture
the behavior of finite-conductivity vertical fractures. treatments was presented by Nolte and Economides
The transient behavior of finite-conductivity verti- (1991). The uncertainty of the reservoir parameter
cal fractures in multilayer reservoirs was studied by values used in the design of hydraulic fractures is
Bennett et al. (1985, 1986). An important contribu- demonstrated to be of significant importance. Further-
tion of the Bennett et al. studies is introduction of more, analysis of the effectiveness of hydraulic frac-
the dimensionless reservoir conductivity concept. ture treatments using conventional transient analysis
This analysis procedure provides a means of corre- procedures is shown to be affected significantly by
lating the multilayer commingled reservoir response heterogeneous reservoir properties and to generally
with that of an equivalent single-layer reservoir dur- indicate a shorter fracture than the actual half-length
ing the early transient behavior of fractured wells, contributing to production (i.e., why this volume dis-
when there is essentially linear flow normal to the tinguishes between the reservoir inferred length xf and
plane of the fracture. The dimensionless reservoir geometric length L).
conductivity concept was extended by Camacho-V. With the advent of fracturing horizontal wells,
et al. (1987) to include the effects of unequal frac- intensive research efforts began to develop transient
ture lengths in each of the layers of a multilayer analysis and well performance models for describing
reservoir system. A later study by Spath et al. (1994) the behavior of horizontal wellbores that have been
presented an analysis procedure that extends the hydraulically fractured to improve productivity.

Transient interpretation analyses have been developed 0.000264 kt

tD = . (12-1)
for horizontal wellbores intersected by orthogonal and φµct x 2f
colinear fractures. Soliman et al. (1990) studied the
productivity improvement of vertically fractured hori- Dimensionless time is related to dimensional time t
zontal wells. Models for interpretation of the perfor- in hr by the system properties (permeability k in md,
mance of hydraulically fractured horizontal wells porosity φ, fluid viscosity µ in cp and total system
were also reported by van Kruysdijk and Dullaert compressibility ct in psi–1) and the characteristic
(1989), Economides et al. (1989, 1991), Roberts et length of the system, which is the fracture half-
al. (1991) and Raghavan and Joshi (1993). length x f in ft.
Specific interpretation principles for hydraulically Dimensionless wellbore pressure is defined by
( ),
fractured horizontal well behavior are not addressed
in this chapter because of the length and complexity kh pi − pwfs
pwD = (12-2)
of the subject. However, most of the basic principles 141.2 qµB
for the interpretation of finite-conductivity vertical
fracture transient behavior discussed in this chapter where h is the net pay thickness of the reservoir in ft,
are directly applicable to horizontal wells that have q is the fluid flow rate in STB/D, B is the fluid for-
been vertically fractured. Section 12-5 presents a mation volume factor in RB/STB, and the initial
model for longitudinally fractured horizontal wells reservoir pressure pi and sandface flowing pressure
that demonstrates the important impact on fracture pwfs are in psia.
conductivity. The dimensionless fracture permeability kfD , width
bfD , height hfD , hydraulic diffusivity η fD and conduc-
tivity CfD are, respectively,
12-2. Post-treatment fracture kf
evaluation k fD = (12-3)
k
The post-treatment fracture evaluation procedures that
w
have found the widest general use for obtaining esti- b fD = (12-4)
mates of the reservoir and propped fracture dimensions xf
and properties are pressure transient well testing and
long-term history matching. Because these techniques hf
are essentially equivalent, the well testing method is h fD = (12-5)
h
emphasized in the following sections. Typical examples
of history and dimensionless rate relations are provided k f φct
by Agarwal et al. (1979), Lee and Holditch (1982) and η fD = (12-6)
Bennett et al. (1983). kφ f ctf
A detailed discussion of the transient behavior of
kf w
finite-conductivity vertical fractures is most readily pre- C fD = k fD b fD = . (12-7)
sented using conventional definitions of the dimension- kx f
less variables that are used for the transient behavior of
fractured wells and in developing appropriate analysis In Eqs. 12-3 through 12-7, the fracture permeabili-
procedures. In the following discussion, it is assumed ty k f, width w, height hf, compressibility ctf and
that the reservoir contains a slightly compressible liquid porosity φf are in consistent units.
with constant viscosity and compressibility. Fluid flow Analysis procedures for the transient behavior of
in the reservoir and fracture is assumed to obey Darcy’s finite-conductivity vertical fractures, pioneered by
law. The reservoir is assumed to be a homogeneous, Cinco-Ley and Samaniego-V. (1981b), have become
isotropic horizontal layer with uniform thickness, the industry standard. These procedures require
porosity and permeability. Specific exceptions to these proper identification of the flow regimes that are
assumptions are addressed separately in Section 12-3. exhibited in the transient response of the well. The
Dimensionless time tD is defined for a system with flow regimes that can be exhibited by finite-conduc-
a vertically fractured well by tivity vertical fractures are wellbore storage domi-

nated, fracture storage dominated, bilinear flow, for- conductivity vertical fracture results in production at
mation linear flow, pseudoradial flow and boundary the wellbore that is due primarily to fluid expansion
dominated flow regimes. The transient behavior also in the fracture.
includes transition regimes that separate each of the This linear flow regime is characterized by a one-
previously identified flow regimes. half slope on a log-log graph of the change in pres-
sure ∆p versus the change in time ∆t. The fracture
storage linear flow regime is generally of such short
12-2.1. Wellbore storage dominated duration that an analysis of this transient behavior is
flow regime usually not possible. In cases where wellbore storage
A shut-in (pressure buildup or falloff), in which the effects are not negligible, they generally distort or
well is shut in at the surface and the wellbore contains mask the fracture storage linear behavior to an extent
a compressible fluid, may be distorted for a significant that it may not be readily observed in the transient
portion of the early response by the storage effects of data. Fluid flow patterns in a fracture for the fracture
the wellbore. The duration of wellbore storage effects linear flow regime are illustrated in Fig. 12-15.
is governed primarily by the volume of the wellbore During the fracture linear flow regime, the behavior
exposed to the completed reservoir and the compress- of dimensionless wellbore pressure pwD is a function
ibility of the fluids contained in the wellbore. Thus, of the dimensionless fracture conductivity, hydraulic
downhole shut-in for pressure buildup can reduce diffusivity and height, and dimensionless time:
wellbore effects considerably. Pressure drawdown or 2
injection transients may also result in domination of pwD = πη fDt D . (12-8)
C fD h fD
a significant portion of the transient behavior by well-
bore storage distortion if the well is opened for pro- The approximate dimensionless time t Defl at which
duction or injected into at the surface and the wellbore the fracture linear flow behavior ends is also a func-
contains compressible fluids. tion of the dimensionless fracture properties:
Wellbore storage dominated flow is characterized
( )
2
by a unit slope for both the pressure and pressure 0.01 C fD h fD
derivative behaviors in log-log coordinates. The pres- t Defl = . (12-9)
η2fD
sure and pressure derivative response of a system with
a positive steady-state skin effect and constant well-
bore storage is a single unit-slope line (see Chapter 2
for a detailed discussion of this issue). Nonideal con- 12-2.3. Bilinear flow regime
ditions may exist in the system in which the wellbore A bilinear flow regime may be exhibited in finite-
storage cannot be characterized by a constant value. conductivity vertical fractures as a result of two linear
These conditions are common in wells that simultane- flow patterns existing in the system simultaneously
ously produce both oil and gas and in which momen- and the fracture tip effects not influencing the tran-
tum, density and thermal effects result in afterflow sient behavior of the well. The flow in the reservoir
and a changing liquid level in the wellbore. Changing is primarily compressible linear flow to the plane of
wellbore storage effects are also referred to as well- the fracture. Once the fluid crosses the fracture face
bore phase redistribution effects (Fair, 1981). into the fracture, the fluid is conducted down the frac-
ture to the wellbore, under essentially incompressible
linear flow conditions. Flow patterns in the reservoir
12-2.2. Fracture storage linear flow regime and fracture during the bilinear flow regime are pre-
The initial pressure transient behavior of a finite- sented in Fig. 12-16. This fracture flow regime is gov-
conductivity vertical fracture (with negligible well- erned by the fracture conductivity, and bilinear flow
bore storage effects) is dominated by the fluid stor- provides the optimum regime for defining conductivity.
age of the fracture itself. In the case of negligible Wellbore pressure transient behavior during the
wellbore storage effects (in practice, this generally bilinear flow regime is expressed by
involves only wells with downhole isolation tools), 2.45083 1/ 4
the constant-rate drawdown of a well with a finite- pwD = tD . (12-10)
C fD h fD

The transient behavior of a finite-conductivity frac-

ture under bilinear flow is presented in Fig. 12-17 as
a Cartesian graph of ∆p versus the fourth root of time.
For a finite-conductivity fracture with CfD > 1.6, the
tail of the transient data graph is concave upward.
Similarly, for a fracture with a very low dimensionless
fracture conductivity (CfD < 1.6), the transient data CfD > 1.6
< 1.6
tails off concave downward. This quick-look diag-
nostic feature helps to identify the conductivity range
∆p
of the fracture. Another diagnostic provided by this
graphic is the nonzero ∆p intercept: a positive value mbf
indicates near-well conductivity damage (e.g., a
choked fracture from overdisplaced proppant or kill 1
fluid damage) and a negative value indicates enhanced

near-well conductivity. The slope of the graph also is
used to calculate the fracture conductivity as implied
by Eq. 12-10, which does not depend on fracture
length. The log-log pressure and pressure derivative 0
behavior during the bilinear flow regime are separated 4
t
by a multiple c of log 4 (Fig. 12-18). The bilinear flow
regime ends at the dimensionless time approximated by Figure 12-17. Cartesian graph of bilinear flow regime
behavior.
 0.1
 2 , C fD h fD ≥ 3
( )
 C fD h fD

[( ) ]
−1.53
t Debf = 0.0205 C fD h fD − 1.5 , 1.6 ≤ C fD h fD < 3 ∆p = c 4 t
 −4
 4.55 
d∆p
 C h − 2.5 , C fD h fD < 1.6 .

dt
  Log 4
Log ∆p, log t
fD fD
(12-11)
Slope = 1⁄4
Although the correlations given by Eq. 12-11 were
originally developed for h fD = 1, they have also been
found adequate for values of h fD ≤ 2.
• Example calculation for the end of bilinear flow Log t
regime
Figure 12-18. Pressure and derivative response of bilinear
Assume that CfD = 12.5, h = 50 ft, h f = 80 ft, flow regime in log-log coordinates.
k = 1 md, φ = 0.15, µ = 1 cp, ct = 10–5 psi–1 and
xf = 500 ft. Calculate the time at which the bilin-
ear flow regime will end. What is this time if and from the definition of dimensionless time
CfD = 1.25? (Eq. 12-1) and rearrangement:
Solution
From Eq. 12-5, h fD = h f /h = 1.6. Then, the product
t=
(2.5 × 10 )(0.15)(1)(10 )(500 ) = 0.36 hr.
−4 −5 2
CfD h fD = 20 and the top expression of Eq. 12-11 is (0.000264)(1)

in effect.
This leads to If CfD = 1.25, then CfD h fD = 2 and the middle
expression in Eq. 12-11 is indicated, leading to
0.1
t Debf = 2
= 2.5 × 10 −4 , tDebf = 5.9 × 10–2 and a much larger t = 84 hr.
20

12-2.4. Formation linear flow regime 100

t Dblf = (12-13)
(C )
2
For finite-conductivity fractures with dimensionless fD h fD
fracture conductivity in excess of approximately 80,
a second linear flow period may be exhibited in
t Delf = 0.016 . (12-14)
which the fracture conductivity is sufficiently high
that the pressure loss caused by flow in the fracture For CfD h fD < 80, Eqs. 12-13 and 12-14 clearly
is negligible. In this case, the pressure transient demonstrate that a formation linear flow regime does
behavior at the well is governed by reservoir com- not exist.
pressible linear flow normal to the plane of the frac- The flow pattern in the reservoir for a vertical
ture. This flow regime is commonly referred to as fracture under formation linear flow is presented in
the formation linear flow regime. The wellbore pres- Fig. 12-19. The transient pressure behavior illustrat-
sure transient behavior of formation linear flow is ed in Fig. 12-20 is in log-log coordinates with the
governed by derivative offset by the multiple of log 2. The Car-
pwD = πt D . (12-12) tesian representation of the pressure and pressure
derivative behavior is in Fig. 12-21, and the slope
The start and end of the formation linear flow determines xf (see Eqs. 12-68 and 12-69). The effect
regime are determined, respectively, by of fracture conductivity on the early transient behav-
ior of fractured wells is illustrated in Fig. 12-22.
Figure 12-19. Formation linear flow to the plane of a finite-conductivity vertical fracture.
∆p = c1 + c2 t telf
d∆p
dt
Log 2
Log ∆p, log t
1⁄2
∆p or ∆m(p)
1
mlf ∝ 1 2
kxf
tblf
Log t
xf
blf ∝
Figure 12-20. Pressure and derivative response of the for- kf w
mation linear flow regime in log-log coordinates.
Figure 12-21. Cartesian graph of formation linear flow ( lf)

behavior.

(0.016)(0.10)(0.02)(4 × 10 −5 )(190 2 )
telf = = 0.35 hr.
15
10π
(0.000264)(0.5)
CfD = 100π
It is therefore determined that the formation lin-
5π
ear flow regime would be exhibited in the pressure
10 buildup transient behavior for about 17 min:
pwD (CfD)
2π
telf − tblf = 0.35 − 0.06 = 0.29 hr = 17.4 min.

π
5
12-2.5. Pseudoradial flow regime

0 All vertically fractured wells (regardless of the CfD
0 5 10 value) may exhibit a pseudoradial flow behavior at
tDxf (CfD)2 late times before the effects of boundaries are
observed. During the infinite-acting pseudoradial
Figure 12-22. Effect of fracture conductivity on bilinear and flow regime, the flux distribution in the fracture has
formation linear flow behavior (Cinco-Ley et al., 1982). stabilized and the transient behavior of the well can
be equated to that of an unfractured well with an
enlarged effective wellbore radius rw´. The radial flow
• Example calculation for start and end of formation steady-state skin effect resulting from this flow
linear flow regime is a function of CfD only. Prior to the pseudo-
A reservoir has h = 100 ft, φ = 10%, ct = 4 × 10–5 radial flow regime, an apparent radial flow steady-
psi–1, k = 0.5 md and µ = 0.02 cp. The well is state skin effect resulting from a vertical fracture may
intersected by a vertical fracture with h f = 150 ft, also be considered. However, during the early tran-
k f w = 12,000 md-ft and xf = 190 ft. What would sient behavior of a vertical fracture, the flux distribu-
be the expected duration of the formation linear tion in the fracture has not stabilized and the apparent
flow regime in a pressure buildup transient test? radial flow steady-state skin effect is a function of
both CfD and time. A composite illustration of the var-
Solution
ious flow regimes that may be exhibited by a finite-
The dimensionless conductivity-height product conductivity fracture is presented in Fig. 12-23.
CfD h fD is determined with Eqs. 12-5 and 12-7:
12, 000 150

C fD h fD = = 189.
(0.5)(190) 100 10 4
Linear
flow
The dimensionless time at which the formation lin-
pwD (CfD)
10 2
ear flow would start is determined with Eq. 12-13: 1⁄ 2
Pseudo-
radial
flow
100 10 0
t Dblf = 2 = 2.8 × 10 .
−3 1⁄ 4
(189) Bilinear Transition

flow flow periods
The dimensionless times that correspond to the 10 –2
start and end of the formation linear flow regime are 10 –4 10 –2 10 0 10 2 10 4 10 6 10 8
determined with Eqs. 12-1 and 12-14, respectively: tDxf (CfD)2
(2.8 × 10 )(0.10)(0.02)(4 × 10 )(190 ) = 0.06 hr

−3 −5 2 Figure 12-23. Log-log response of a finite-conductivity ver-
tblf = tical fracture exhibiting various flow regimes.
(0.000264)(0.5)

The wellbore pressure transient behavior of a ver- • Example calculation of apparent pseudoradial skin
tically fractured well during the pseudoradial flow effect and folds of increase
regime is Using the variables of “Example calculation for
[ ]
1 the end of bilinear flow regime” (page 12-12) and
pwD = ln t Dr ′ + 0.8091 . (12-15) CfD = 1, calculate the folds of increase of a frac-
2 w
tured well over an unfractured, undamaged well,

The corresponding dimensionless time referenced with the reservoir radius re = 3000 ft and rw =
to the effective wellbore radius rw´ is 0.328 ft.
0.000264 kt Solution
t Dr ′ = , (12-16) With CfD = 1 from Fig. 12-17, s + ln(xf /rw) = 1.5
w
φµct rw′2
and thus s = 1.5 – ln(500/0.328) = –5.8. The pro-
and the effective wellbore radius is ductivity index ratio is
rw′ = rw e − s , ln(re / rw ) ln(3000 / 0.328)

(12-17) JF J = = ≈ 2.8,
ln(re / rw ) + s ln(3000 / 0.328) − 5.8
where rw is the wellbore radius in ft and s is the
apparent steady-state skin effect caused by the frac- where JF is the fractured well productivity index
ture. The equivalent radial flow steady-state skin and J is the pseudosteady-state productivity index.
effect resulting from a fracture is presented in Fig. The effective wellbore radius is rw′ = 0.328 ×
12-24. A very high conductivity fracture (C fD > 500, exp(+5.8) = 108 ft.
and for practical purposes C fD > 50) has rw´ = xf /2,
whereas the effective wellbore radius of a very low
conductivity fracture (CfD < π/20, and for practical 12-2.6. Pseudosteady-state flow regime
purposes C fD < 0.6) is 0.25x f C f D = 0.25kf w/k and
independent of xf. In finite reservoirs at late times (fully developed
The onset of fully developed pseudoradial flow boundary dominated flow), the actual type of well
occurs at approximately a dimensionless time (Eq. 12-1) completion does not solely govern the pressure tran-
equal to 3. The dimensionless time at which pseudo- sient behavior of the well. Rather, the dimensionless
radial flow begins is actually a function of CfD, vary- pressure in a closed system is also a function of the
ing from a dimensionless time of 2 for CfD = π/10 to reservoir drainage area size and shape, well location,
a dimensionless time of 5 for CfD = 100π. Once fully formation properties and time. The fully developed
developed pseudoradial flow behavior is exhibited, the pseudosteady-state transient behavior (i.e., the tran-
classic semilog analysis procedures presented in sient has reached all boundaries) of a well is charac-
Chapter 2 are applicable. terized by pressure and pressure derivative behavior
following the same unit-slope line on a log-log
graph. The dimensionless wellbore pressure p wD
of a vertically fractured well under fully developed
pseudosteady-state flow conditions is
2
 x  1  2.25 
sf + ln (xf /rw)
pwD = 2 πt DA + ln e  + ln . (12-18)

 x f  2  CA 
1
It is assumed in this relationship that the well
is centrally located in a square drainage area with
dimensions of 2x e on each side (Earlougher, 1977).
0
The reservoir geometric shape factor CA in Eq. 12-18
10 –1 10 0 101 10 2 10 3
CfD
is a function of the reservoir drainage area shape, well
location, well type and ratio of the drainage area
Figure 12-24. Equivalent radial flow steady-state skin extent to fracture half-length xe/xf. Table 12-1 sum-
effect resulting from a finite-conductivity vertical fracture marizes the geometric shape factors for vertically
(Cinco-Ley and Samaniego-V., 1981a).

Table 12-1. Geometric shape factors for (0.175)(0.10)(0.02)(4 × 10 −5 )(6, 969, 600)
t= = 739 hr ≈ 31 days.
pseudosteady-state flow. (0.000264)(0.5)
xe /xf CA Within 1% error, Eq. 12-18 would therefore be
applicable for describing the dimensionless well-
1 0.7887
bore pressure transient behavior at
2 1.6620
3 1.9607 (0.09)(0.10)(0.02)(4 × 10 −5 )(6, 969, 600)
t= = 380 hr ≈ 16 days.
4 2.0296 (0.000264)(0.5)
5 2.0348
6 2.0760
7 2.1665 12-3. Factors affecting fractured
8 2.2984 well performance
9 2.4636 The dimensionless wellbore pressure solutions pre-
10 2.6541 sented in the previous section are applicable for linear
reservoir flow problems. Linear fluid flow problems
fractured wells that are centrally located in square are those in which the reservoir and fracture can be
drainage areas equal to 4x e 2. assumed to be uniform, homogeneous and isotropic.
The dimensionless time tDA referenced to the The reservoir is assumed to contain a slightly com-
drainage area of the reservoir is pressible liquid of constant viscosity and compress-
ibility. The fracture is assumed to be a rectangular
0.000264 kt
t DA = , (12-19) vertical slab of uniform width, height, length and
φµct A conductivity. The fluid flow throughout the system
where A is the reservoir drainage area in ft2. is also assumed to obey Darcy’s law.
Many nonideal conditions exist that result in a tran-
• Example calculation of time to start of sient behavior that does not strictly follow the ideal
pseudosteady-state flow reservoir behavior previously discussed. Some of the
For a vertically fractured well centrally located nonideal conditions that must be considered are non-
in a square drainage area, the value of tDA at which Darcy fluid flow in the fracture and reservoir, non-
fully developed pseudosteady-state flow occurs linear fluid properties, spatially varying fracture and
is approximately equal to 0.175 (for all xe /xf ratios reservoir material properties and geometry, and het-
between 1 and 10). The onset of fully developed erogeneous reservoir systems that include layered
pseudosteady-state flow can be approximated for reservoirs and dual-porosity systems.
this same range of xe/xf values within 1% error
for tDA ≈ 0.09. What would be the transient time
required for fully developed pseudosteady-state 12-3.1. Non-Darcy flow behavior
flow behavior to occur on a constant-rate draw- High-velocity fluid flow in porous media results in
down of a 160-acre reservoir with the properties pressure losses that generally cannot be described by
in “Example calculation for start and end of for- Darcy’s law. Evidence of this has been available for
mation linear flow” (page 12-14 )? At what time many years (Forchheimer, 1901). The effect of high-
could the pseudosteady-state approximation velocity fluid flow in a reservoir is that an additional
tDA ≈ 0.09 be used? pressure loss is exhibited that is believed to be due
Solution primarily to inertial slip effects resulting from the
The drainage area of the reservoir is rapid acceleration and deceleration of the fluid parti-
cles as the fluid moves along the tortuous flow path
A = ( 43, 560)(160) = 6, 969, 600 ft 2. of the interconnected pore space of the reservoir rock
or fracture proppant (Geertsma, 1974). Although the
The time at which fully developed psedosteady-
additional pressure loss resulting from high-velocity
state flow occurs is obtained by rearranging
fluid flow in the reservoir is often referred to as turbu-
Eq. 12-19 with tDA = 0.175:

lent flow, the proper classification for this flow behav- Wattenbarger and Ramey and Holditch and Morse pro-
ior is inertial-turbulent, because in many cases the duced qualitative estimates of the effects of non-Darcy
pore throats in the reservoir rock or fracture proppant fracture flow, the study by Guppy et al. (1982b) was
are too small for a given production rate level to allow the first investigation to provide a means of quantita-
fully developed turbulence similar to that commonly tively determining the effects of non-Darcy flow on
observed in pipe flow. the pressure transient behavior of a well intersected
Non-Darcy flow pressure losses in the fracture are by a finite-conductivity vertical fracture (Fig. 12-25).
of primary interest in the analysis of relatively high The reduction in apparent fracture conductivity result-
rate gas wells, but they can also be important for ing from non-Darcy flow in a constant-rate drawdown
very high rate oil wells (Bale et al., 1994b) and con- is demonstrated in Fig. 12-26.
verging fracture flow because of a limited wellbore Guppy et al. (1982a) also provided a correlation
connection (e.g., orthogonal fracture in a horizontal for computing the apparent reduction in dimension-
well). Wattenbarger and Ramey (1969) investigated less fracture conductivity C fDapp of a vertically frac-
the effects of turbulent (or inertial) flow in both the tured well that has been produced at a constant rate.
reservoir and fracture and found that often inertial The correlation resulting from non-Darcy fracture
flow in the fracture is more significant than high- flow for a constant-rate drawdown is
velocity flow effects in the reservoir.
C fDapp 1
Holditch and Morse (1976) investigated the effects = , (12-20)
of non-Darcy fluid flow in finite-conductivity fractures C fDtrue 1 + 0.31q DND
using a finite-difference numerical simulator. Although
10 –2
Guppy et al. (1982a)
Cinco-Ley et al. (1978)
10 –1
k fD b fD = 5
pD
π
qDND = 15 k fD b fD =
5
4
10 0 0
k fD b fD = 100π
10 –1
10 –2 10 –1 10 0 10 1 10 2 10 3
tDx
f
Figure 12-25. Effect of non-Darcy flow on the pressure transient response of finite-conductivity vertical fractures.

(k fD b fD )true
7 9 50 70 90 150
20 1 1.5 2 3 4 5 6 8 10 15 20 30 40 60 80 100 200 300 400
16
12
q DND
0
0.1 1 10 100 1000
(k fD b fD )app
Figure 12-26. Apparent fracture conductivity reduction resulting from non-Darcy fracture flow.
and the dimensionless non-Darcy flow rate qDND for where γg is the gas specific gravity (relative to air)
oil and gas, respectively, is and the constant is the molecular weight of air.
The non-Darcy flow coefficient β in Eq. 12-21
1.03 × 10 −15 k f ρoβqo is in ft–1. A simple expression for estimating the non-
q DND = (12-21)
whµ o Darcy flow coefficient was specifically developed for
4.64 × 10 −16 k f βMWqg fracture flow by Cooke (1973):
q DND = .
whµ gi 30.88 × 10 6 b ′
β= . (12-23)
k fa′
The ranges of applicability for the Eq. 12-20
correlation are The coefficient b´ and exponent a´ in Eq. 12-23 are
specific to the type and mesh size of the fracture prop-
• all values of qDND that yield CfDapp > 2, for CfDtrue < 10 pant and are determined experimentally. Table 12-2
• for 1 ≤ qDND ≤ 20 and 10 ≤ CfDtrue ≤ 100 lists examples of coefficients and exponents for
• for qDND ≤ 10 and 100 ≤ CfDtrue ≤ 500. Hickory sand (Cooke, 1973) for which the fracture
permeability kf in Eq. 12-23 is in md.
The fracture permeability kf in Eq. 12-21 is in md,
fracture width w is in ft, reservoir net pay thickness
h and average fracture height hf are in ft, fluid den- Table 12-2. Non-Darcy coefficients for fracture flow.
sity ρo is in lbm/ft3, fluid viscosities µo and µgi are Mesh a′ b′
in cp, oil flow rate qo is in STB/D, gas flow rate qg
8/12 1.24 17,424
is in Mscf/D, and the gas molecular weight MW is
in lbm/lbm-mol. The molecular weight of natural 10/20 1.34 27,539
gas can be computed with 20/40 1.54 110,470
40/60 1.60 69,405
MW = 28.96 γ g , (12-22)

Extension of the Guppy et al. (1982b) correlation 20/40-mesh Hickory fracture sand and has
relating the true dimensionless fracture conductivity kf = 85,000 md at reservoir confining conditions.
CfDtrue and the apparent (resulting from non-Darcy The well is produced at 10 MMscf/D, and
flow) dimensionless fracture conductivity CfDapp γg = 0.6. Determine the percent apparent fracture
developed for constant-rate drawdown transients conductivity reduction resulting from non-Darcy
(Eq. 12-20) to other inner boundary condition tran- fracture flow for the reservoir described in
sients (e.g., constant pressure drawdown or shut-in “Example calculation for start and end of forma-
transients) is not directly possible because the non- tion linear flow” (page 12-14) using both the
Darcy fracture flow problem is nonlinear. However, Guppy et al. and Gidley evaluation procedures.
Guppy et al. (1982a) also developed a similar corre- Solution
lation for the interpretation of pressure buildup The molecular weight of the reservoir gas is deter-
responses of fractured wells that have been produced mined with Eq. 12-22:
under non-Darcy fracture flow conditions:
MW = 28.96(0.6) = 17.4 lbm/lbm- mol.
C fDapp 1
= . (12-24) The inertial flow coefficient is determined with
C fDtrue 1 + 0.55q DND
Eq. 12-23 and the values in Table 12-2:
The dimensionless non-Darcy flow rate qDND is eval-
30.88 × 10 6 (110, 470)
uated using the production rate of the drawdown β= = 87, 421 ft –1 .
transient preceding the pressure buildup transient 85, 0001.54
and Eq. 12-21. The range of applicability of Eq. 12-24
The dimensionless non-Darcy fracture flow rate
is defined as
qDND is determined with Eq. 12-21 for gas:
• for qDND ≤ 2 and CfDtrue < 10
q DND =
(4.64 × 10 )(85, 000)(87, 421)(17.4)(10, 000) = 14.4 .
−16
• for 1 ≤ qDND ≤ 10 and CfDtrue ≥ 10.

(0.25 / 12)(100)(0.02)
Gidley (1990) also developed a simple correlation
The true dimensionless fracture conductivity
procedure for estimating the reduction in CfDapp
CfDtrue of the fracture system is determined with
resulting from non-Darcy fracture flow. This expres-
Eq. 12-7:
sion is correlated in terms of a fracture flow Reynold’s
number NRe:
C fDtrue =
(85, 000)(0.25 / 12) = 18.6 .
C fDapp 1
(0.5)(190)
= . (12-25)
C fDtrue 1 + NRe Therefore, the second condition governing the
range of applicability of Eq. 12-20 is satisfied.
The fracture flow Reynold’s numbers for oil and gas, The apparent dimensionless fracture conductivity
respectively, are CfDapp is determined for the Guppy et al. analysis
5.53 × 10 −16 k f ρoβqo with Eq. 12-20:
NRe = (12-26)
whµ o 18.6
C fDapp = = 3.4 .
7.00 × 10 k f γ gβqg
−15 1 + (0.31)(14.4)
NRe = .
whµ gi The percent reduction in apparent fracture con-
ductivity using the Guppy et al. analysis proce-
Additional consideration of non-Darcy fracture flow dure is thus
is in Chapter 10.
(18.6 − 3.4)(100)
• Example calculation of reduction in apparent % reduction = = 81.7%.
18.6
fracture conductivity resulting from non-Darcy
fracture flow The vertically fractured well would therefore
behave as if the fracture had a conductivity kf w
A vertically fractured gas well has hf = 100 ft of only
and w = 0.25 in. The fracture is propped with

The gas viscosity µg and the real gas law deviation

(k w)
f app
= (85, 000)
0.25 
 12 
(1 − 0.817) = 324 md-ft, factor Z in Eq. 12-27 are pressure-dependent fluid
properties.
when the actual conductivity of the propped frac-
A similar integral transformation was presented by
ture was 1771 md-ft.
Agarwal et al. (1979) to effectively complete the lin-
For the Gidley analysis procedure, the fracture
earization of the gas diffusivity equation. This inte-
flow NRe is determined with Eq. 12-26 for gas:
gral transformation, to account for variation in reser-
N Re =
(7.00 × 10 )(85, 000)(0.6)(87, 421)(10, 000) = 7.49 .
−15 voir fluid properties with respect to time, is com-
(0.25 / 12)(100)(0.02) monly referred to as pseudotime:
The apparent dimensionless fracture conductiv-  ∂t 

pr ( t )   dp
 ∂p 
t
ity is determined with Eq. 12-25: dt
t a (t ) = ∫ = ∫ . (12-28)
18.6 µ g ct µ g ct
= = 2.19 .
0 pi
C fDapp
1 + 7.49 Agarwal et al. introduced the pseudotime integral
The percent reduction in apparent fracture con- transformation on an intuitive basis. Lee and
ductivity is therefore Holditch (1982) later provided theoretical justifica-
tion for use of the pseudotime integral transforma-
(18.6 − 2.19)(100)
% reduction = = 88.2%. tion in analysis of the transient behavior of gas
18.6 wells. The evaluation of pseudopressure and pseudo-
time integral transformations for well performance
Both analysis procedures demonstrate that a
prediction with analytic solutions is discussed in
significant reduction in CfDapp would be observed
greater detail by Poe et al. (1995). Evaluation of the
for non-Darcy fracture flow in this example and
pseudotime integral transformation for long-term
that fracture design changes are warranted, as dis-
well performance prediction includes consideration
cussed in Chapter 10.
of the variation in fluid properties over the entire
reservoir volume with respect to time to effectively
12-3.2. Nonlinear fluid properties linearize the gas diffusivity equation.
The corresponding definitions of dimensionless
A common factor that may complicate analysis or pseudotime and pseudopressure for gas reservoir
prediction of the transient behavior of a well is the analyses, respectively, are
nonlinear behavior of reservoir fluids. The variation
0.000264 kta (t )
of reservoir fluid physical properties with time and taD = (12-29)
pressure must be accounted for to evaluate the tran- φx 2f
sient behavior of a well and obtain reliable estimates
of the fracture and reservoir properties for long-term
production forecasts. All gas reservoirs and some oil p pwD =
[
khTsc m( pi ) − m pwf ( )] . (12-30)
solution-gas reservoirs require integral transforma- 50, 300 psc Tqg
tions to effectively linearize the resulting nonlinear
The dimensionless pseudotime and pseudopressure
reservoir flow problem.
definitions in these equations are used in gas reservoir
Integral transformations are used to account for
analyses in place of Eqs. 12-1 and 12-2, respectively.
variation in reservoir fluid properties with the pres-
The standard condition temperature Tsc and pres-
sure level and time. For gas reservoirs, Al-Hussainy
sure psc in Eq. 12-30 are specified by the applicable
et al. (1966) introduced the real gas pseudopressure
governmental regulatory agency. A typical value for
integral transformation to effectively linearize the
psc is 14.7 psia, but it may range between 14.65 and
diffusive part of the diffusivity equation governing
14.73 psia. The most common Tsc used is 60°F
real gas flow in the reservoir:
[520°R].
p′dp′
p
The multiphase flow analysis procedure developed
m( p ) = ∫ . (12-27)
by Perrine (1957) and Martin (1959) for radial flow
po
µgZ

has been successfully applied to the analysis of pres- analysis of the reservoir performance is usually
sure buildup data for fractures. The linearization of preferable to a single-phase flow transient analysis.
the multiphase flow nonlinear problem using Perrine There are no set standards for selecting multiphase
and Martin’s approach assumes that the fluid satura- analysis over single-phase analysis. However, there
tion distributions in the reservoir are essentially uni- are qualitative guidelines that classify which type
form over the reservoir drainage area. The lineariza- of analysis is appropriate:
tion involves replacing the single-phase mobility of
• If the producing gas/liquid ratio exceeds approxi-
the system with a comparable multiphase mobility:
mately 45,000 scf/STB, there is essentially single-
k k k  phase gas flow in the reservoir and a gas reservoir
λ t = k  ro + rw + rg  . (12-31) analysis should be used.
 µo µw µg 
• If the gas/oil ratio is less than 20,000 scf/STB
The products of the absolute permeability k and the and the water production rate is significantly less
relative permeabilities kro, krw and krg of each of the (1 order of magnitude or more) than the oil pro-
fluid phases are the respective effective permeabilities duction rate, a single-phase oil reservoir analysis
ko, kw and kg. The total system compressibility is iden- is generally appropriate.
tical to that used in the single-phase flow analysis • If the water production rate is comparable to or
procedures: exceeds the oil production rate and the gas/liquid
ratio is less than 45,000 scf/STB, a multiphase
ct = So co + Sw cw + Sg cg + c f , (12-32)
analysis is generally appropriate.
where the fluid saturations So, Sw and Sg of the oil, • Wells that cannot be classified in the preceding
water and gas phases, respectively, are expressed as three steps may require both a single-phase analysis
fractions of the pore volume, and the compressibilities (of the principal produced-fluid phase) and a multi-
co, cw and cg of the fluids and pore compressibility cf phase analysis. The effective permeabilities to the
are in psia–1. principal fluid phase from both analyses should be
The total reservoir production rate qRT represents in reasonable agreement. If significant disagreement
the contribution of each of the fluid phases: exists, the multiphase analysis estimates of effective
permeabilities and steady-state skin effect are usually
q RT = qo Bo + qw Bw +  qg − o So − w Sw  Bg ,
qR q R the more reliable of the two analyses.
 1000 1000 
(12-33)
12-3.3. Fracture damage and spatially
where qRT is in RB/D and the fluid-phase production varying fracture properties
rates qo, qw and qg are in STB/D, STB/D and Mscf/D,
respectively. The fluid formation volume factors Bo, The effects of flow impairments along the fracture
Bw and Bg are in RB/STB, RB/STB and RB/Mscf, face and in the fracture near the wellbore on the tran-
respectively. The solution-gas/oil ratio Rso and solu- sient behavior of finite-conductivity vertical fractures
tion-gas/water ratio Rsw are in scf/STB. were investigated by Cinco-Ley and Samaniego-V.
Substitution of these relationships into Eqs. 12-1 and (1977, 1981a). Fluid-loss flow impairment along the
12-2 results in definitions of the multiphase flow analy- fracture surface in the reservoir is commonly
sis dimensionless time and pressure, respectively: referred to as a fracture face skin effect. Flow
impairment caused by reduced conductivity in the
0.000264λt t fracture near the wellbore is commonly described
tD = (12-34)
φct x 2f as a choked fracture. Both of these types of flow
impairments in fractured wells result in a lowered
pwD =
(
λ t h pi − pwf ). (12-35)
productivity than would be obtained if flow impair-
ments were not present.
141.2 q RT Fluid-loss damage in the reservoir adjacent to the
fracture is illustrated in Fig. 12-27. A choked frac-
For wells that produce significant amounts of oil,
ture with a significant fracture conductivity reduction
gas and water simultaneously, a multiphase flow

Well Fracture
kf w
xf
k
Damaged zone
Figure 12-27. Fracture face skin effect damage flow impairment.
in the vicinity of the wellbore is shown in Fig. 12-28. The fracture damage skin effect sfs from Eq. 12-36
As noted in Section 12-2.3, the bilinear flow graph is not readily equated to a radial flow steady-state
of ∆p versus the fourth root of time provides a diag- skin effect s because these two quantities are refer-
nostic of a choked fracture and the ∆p intercept enced to different system characteristic lengths and
quantifies the choke’s pressure loss. The effect on the different flow patterns in the reservoir are de-
the transient behavior of finite-conductivity fractures scribed by the two steady-state skin effects. Typical
resulting from fracture damage skin effects is illus- values of fracture face skin effect damage from
trated in Fig. 12-29. The effects on the effective clean-breaking fracturing fluids are generally low,
wellbore radius of choked and damaged infinite- on the order of 0.05 or less.
conductivity fractures in the pseudoradial regime Holditch (1979b) also studied the effect of fluid-
are compared in Fig. 12-30. loss-induced impairment in low-permeability reser-
Cinco-Ley and Samaniego-V. (1977, 1981a) intro- voirs. A finite-difference reservoir simulation model
duced a relationship for quantifying fracture damage was used to investigate water blockage and post-
skin effects in terms of the fracture half-length xf, treatment fracture fluid cleanup of vertical fractures,
extent of the damaged region into the reservoir nor- including the effects of relative permeability and
mal to the fracture plane bs and undamaged-to-dam- capillary pressure changes in the vicinity of the frac-
aged permeability ratio k/ks: ture. The study found water blockage (i.e., relative
permeability effects) in low-permeability gas reser-
πbs k  voirs could result in significant production impair-
s fs =  − 1 . (12-36)
xf  ks  ment of a fractured well.
Damaged zone Fracture
Well
w kf k fs
xs xf
Figure 12-28. Choked-fracture flow impairment. kfs = fracture permeability in the near-well skin effect zone, xs = half-length
of the skin effect zone.

10 1
End of linear flow
10 0 s fs = 1
Beginning of
semilog
straight line
pwD
s fs = 0.2
s fs = 0.1
10 –1
0.02
s fs =
= 0.01
s fs
= 0
s fs
10 –2
10 –4 10 –3 10 –2 10 –1 10 –0 10 1
t Dx f
Figure 12-29. Damaged fracture pressure response (Cinco-Ley and Samaniego-V., 1981a).
10 0 10 3
F4 (s fs)
104
10 2
0
10 1
F1 (pwD)
Fluid-loss damage
10 –1
r´w
xf
10 0
Choked fracture
10 –1
10 –2
10 –2
10 –1 10 0 10 1 10 2 10 –2 10 –1 10 0 10 1 10 2 10 3 10 4 10 5 10 6
F2 (tDxf)
sfs–1 or sfs–1 choked
Figure 12-30. Effect of damaged fractures on effective Figure 12-31. Pressure response of fluid-loss-damaged
wellbore radius. fractures (Wong et al., 1984).
Wong et al. (1984) presented a correlation tech- 1/ 3

 C2 
nique for evaluating the pressure response of fluid- F1 ( pwD ) = pwD  fD  (12-37)
loss-damaged fractures. Figure 12-31 represents the  CD 
effect of fluid-loss damage on the bilinear flow tran- 2/3
C 
F2 (t D ) = t D  fD2 
sient behavior of finite-conductivity fractures and is
(12-38)
correlated in terms of three functional groups:  CD 

( )
s fs conductivity, for which the fracture storage effects are
F4 s fs = . (12-39)
(C )
1/ 3
CD2 assumed negligible.
fD
• Example calculation of fracture face skin effect
The dimensionless wellbore storage coefficient CD A vertical well is hydraulically fractured with
used in these functional groups is 300 bbl of proppant-laden slurry comprising
0.89C water-base gel fracturing fluid and proppant. The
CD = , (12-40) fracture is created using a tip screenout procedure
φct hx 2f
(i.e., proppant is distributed over the entire created
where the dimensional wellbore storage C is in hydraulic fracture length), which results in a sym-
bbl/psi. Cinco-Ley and Samaniego-V. (1977, 1981a) metrical fracture with xf = 200 ft, hf = 120 ft (also
investigated the transient behavior of a choked frac- the permeable section for fluid-loss thickness) and
ture with a conductivity reduction in the fracture an average propped w = 0.25 in. The reservoir
near the wellbore. This is one subset of a more gen- initially has ko = 5 md, kro = 0.8 and φ = 10% with
eral set of problems involving evaluation of the tran- an initial oil saturation Soi = 80%. The flushed
sient behavior of vertical fractures with spatially zone is assumed to have an oil saturation of 25%,
varying conductivity. for which the relative permeability to oil is 0.1.
Bennett et al. (1983) studied the transient behavior What would be the steady-state fracture face skin
of a vertically fractured well with spatially varying effect observed in a single-phase pressure buildup
fracture conductivity. A finite-difference reservoir analysis of this fractured well? For simplicity,
simulator was used to investigate the transient per- assume that all damage is caused by fluid satura-
formance of a fractured well with a stepwise-varying tion changes only.
conductivity distribution. It was found that the initial Solution
transient behavior of a fractured well with this type The closed fracture volume is estimated as
of conductivity distribution is governed by the con-
ductivity of the fracture nearest the wellbore. At later Vf = 2 x f wh f
transient times, the transient behavior of a well with
= 2(200)
0.25 
a monotonically decreasing fracture conductivity  12 
(120)
with distance from the wellbore can be correlated
with the transient behavior of an equivalent uniform- = 1000 ft 3 ≈ 178 bbl.
conductivity fracture by using an average of the The total fracturing fluid lost into the formation
dimensionless fracture conductivity distribution: is therefore
1 x VL = 300 − 178 ≈ 122 bbl.
kx 2f ∫o
C fD = k f wdx ′. (12-41)
The distance into the reservoir at which the
Soliman (1986a, 1986b) also investigated the effect fracturing fluid has penetrated through each face
of spatially varying fracture conductivity on the pres- of the fracture is estimated volumetrically as
sure transient behavior of a finite-conductivity frac- VL
bs =
ture. The results of those studies provide a qualitative 4φ( Soi − Sor )hx f
reference for the effects of spatially varying fracture
conductivity on the transient behavior of vertical frac- 122(5.615)
=
tures. The analytic solutions for the pressure transient 4(0.1)(0.8 − 0.25)(120)(200)
behavior of finite-conductivity vertical fractures with = 0.13 ft ≈ 1.6 in.,
arbitrarily varying fracture conductivity and height
were later developed by Poe et al. (1992). The solu- where Sor is the residual oil saturation.
tions presented by Poe et al. for finite-conductivity The flushed zone effective permeability to oil
vertical fractures include fracture storage effects, as is therefore
well as an extension of the model by Cinco-Ley and  k 
ks = kr ( Sor ) o  = (0.1)
5 
= 0.625 md,
 kr ( Soi )   0.8 
Meng (1988) to include spatially varying fracture

which is an 87.5% reduction in the effective per-

10
meability to oil. b2/xf = 0.20 Skin effect = 20
b2/xf = 0.08 Skin effect = 0
The steady-state fracture face skin effect that
Composite damage
1
skin effect, sd
b2/xf = 0.02
would be exhibited in the pressure transient analy-
sis as a result of fracture fluid leakoff is deter- 0.1
mined with Eq. 12-36:
0.01
π(0.13)  5
s fs = = − 1 = 0.014 .
200  0.625  0.001
0.01 0.1 1
From Eq. 12-30, sfs–1 = 1/0.014 = 70. This Fracture face permeability impairment ratio, k 2 /k r
amount of fluid-loss impairment would have a
negligible effect on the effective wellbore radius. Figure 12-32. Variation of composite damage skin effect
with the fracture face permeability impairment ratio (Mathur
et al., 1995).
12-3.4. Damage in high-permeability

fracturing for low fracture face permeability impairment, but
in this case the composite skin effect is itself small.
Fracture face damage is most likely to have minor The total pseudoradial skin effect (sd plus the
effects on the performance of fractured wells in low- Cinco-Ley et al. [1978] fracture skin effect; Fig.
permeability reservoirs. Long lengths, characteristic 12-24) is plotted in an example calculation in Fig.
of fractures in these reservoirs, result in small values 12-33 for CfD = 1, xf = 25 ft and kr = 10 md. The first
of the skin effect determined with Eq. 12-36. This is conclusion is that the total skin effect st can be posi-
not the case in high-permeability wells, where frac- tive, as reported by several investigations of high-
ture lengths are smaller, by as much as 2 orders of permeability fracturing.
magnitude. Furthermore, high reservoir permeability A more important conclusion, however, is that the
results in considerable penetration of fluid-loss dam- total skin effect is favorable (negative) if the extent
age, unless extraordinary leakoff control measures of fracture face permeability impairment is con-
are taken. trolled by reducing either the damage ratio k2/kr or
Mathur et al. (1995) studied various types of dam- the penetration ratio b2/xf. Impairment control for
age in high-permeability fractured wells and intro- relatively short fractures in higher permeability for-
duced a composite skin effect accounting for radial mations has major implications in fracturing fluid
and fracture face permeability effects: selection (for damage) and the importance of leakoff
sd =
π

b2 kr
+
(b1 − b2 )kr − b1  , control (fluid penetration).
 ( ) ( )
2  b1k3 + x f − b1 k2 b1k1 + x f − b1 kr x f 

4
(12-42) C fD = 1 b2/xf = 0.20 k1 = kr, k2 = k3
3 x f = 25 ft k3 = k1 = 1 md
k r = 10 md
Total skin effect, st
where, for convenience, b1 and b2 are the extents of 2

the radial and fracture face damages, respectively, 1
and kr, k1, k2 and k3 are the virgin radial, damage 0
radial, fracture face and composite radial/fracture b2/xf = 0.08
–1
face permeabilities, respectively. If k1 = kr, k3 = k2
–2
and b1 = b2, Eq. 12-42 reduces to Eq. 12-36. b2/xf = 0.02
Figure 12-32 is a graph of the composite skin effect –3 Total ideal skin effect
0.01 0.1 1
as a function of the fracture face permeability impair- Fracture face permeability impairment ratio, k 2 /k r
ment k2/kr, ratio of the extent of the fracture face dam-
age to the fracture length b2/xf and radial skin effect.
Figure 12-33. Variation of total skin effect with the fracture
It is apparent that radial damage is largely insignifi- face permeability impairment ratio (Mathur et al., 1995).
cant, even in very high permeability reservoirs, except

12-3.5. Heterogeneous systems layer systems. They determined that fracture half-
length is underestimated in a layered commingled
The transient behavior discussed in the preceding sec- reservoir if the layered nature of the formation is not
tions of this chapter is presented with the assumption properly considered. Bennett et al. introduced the
that the reservoir is homogeneous and isotropic. concept of dimensionless reservoir conductivity CRD
Spatial variation in reservoir thickness, porosity, per- to correlate single-layer and multilayer responses:
meability and fluid saturations can be readily consid-
n
∑h k j (φct ) j ,
ered with finite-difference or finite-element simulation 1
CRD = j (12-44)
models. However, using analytic solutions to evaluate h k φct j =1
the transient behavior of vertically fractured wells, or
their long-term production behavior, may require nor- where CRD is always less than or equal to unity for
malization procedures to properly account for spatial a multilayer system. For post-treatment analysis of
variation in the reservoir properties. Geostatistical fractured well transient behavior, the dimensionless
techniques have been used to normalize spatial varia- time scale is tD/CRD2, which shifts the fracture half-
tion in reservoir properties and to correlate transient length scale by a factor of CRD. Ignoring the layered
analysis and reservoir simulation results. reservoir nature may result in shorter than actual
Reservoir permeability anisotropy is another com- apparent fracture lengths from the analysis.
mon factor that complicates analysis of the transient Camacho-V. et al. (1987) later extended the work
behavior of finite-conductivity fractures. Directional by Bennett et al. on the concept of dimensionless
permeability anisotropic effects on the transient reservoir conductivity CRD to include the effects of
behavior of finite-conductivity fractures were studied differing fracture half-lengths in multilayer reservoir
—
by Ben Naceur and Economides (1988). That study systems. The total system net pay thickness h is
found that the anisotropy case believed most com- n
mon is the one in which the direction of maximum h = ∑ hj , (12-45)

permeability is perpendicular to the direction of min- j =1
—
imum in-situ horizontal stress (i.e., parallel to the average system reference permeability k is
fracture plane). For this case, reservoir permeability
1 n
anisotropy has an unfavorable effect on the produc- k= ∑ k j hj ,
h j =1
(12-46)
tivity improvement obtained by fracturing the well.
——
This unfavorable anisotropy ratio in effect results in and average system storage φ ct is
a shorter apparent fracture half-length than that actu-
1 n
ally obtained.
For radial flow in anisotropic reservoirs, it is com-
φ ct = ∑ φ j ctj hj .
h j =1
(12-47)
mon practice to define an average effective horizon- The individual layer and composite system average
tal permeability for the system that is the geometric hydraulic diffusivities are, respectively,
mean of the directional permeabilities kx and ky:
kj
ηj = (12-48)
k = kx ky . (12-43) φ j ctj µ
For kx = 10ky, the radial response from a prefracture k
well test yields k = 3.1kx, whereas the fracture η= . (12-49)
φct µ
responds approximately as if k were only 1.8kx.
—
Another reservoir heterogeneity that must be con- An equivalent system average fracture half-length x f
sidered in analysis of the transient behavior of verti- is, therefore,
cally fractured wells and their future production per-
η n
k j h j x fj
formance is the effect of multiple reservoir layers, xf =
kh
∑
j =1 ηj
. (12-50)
which may have significantly different formation
properties, pore pressure levels, fracture half-lengths An average fracture conductivity k f w is
——
and drainage areal extents.

2
Bennett et al. (1986) investigated the transient 1 n 
response of hydraulically fractured wells in multi- k f w =  ∑ k fj w j h j CRDj  , (12-51)
h  j =1 

where the individual fracture layer dimensionless The uniqueness problem in interpretation of the
reservoir conductivity CRDj is transient behavior of finite-conductivity fractures is
due to the fact that the transient performance of a
k j hj η
CRDj = . (12-52) fractured well in each of the flow regimes considered
kh ηj is a function of both reservoir and fracture proper-
ties. An independent estimate of formation perme-
The dimensionless reservoir conductivity CRD con-
ability is required to quantify the fracture parameter
cept of Bennett et al. and Camacho-V. et al. is
estimates separately from the combined response of
applicable only for analysis of the early transient
the reservoir and fracture.
behavior of finite-conductivity fractures (dimension-
Evaluation of the drawdown (or injection) tran-
less fracture time tDxf < 1) for which there is essen-
sient behavior of a fractured well involves graphical
tially linear flow in the reservoir normal to the plane
analysis of the pressure drawdown ∆p = pi – pwf or
of the fracture. This correlation technique is not
the pressure rise for injection ∆p = pwf – pi, where pi
applicable to pseudoradial or pseudosteady-state
is the initial pressure and pwf is the bottomhole flow-
flow regimes.
ing pressure, for liquid or multiphase flow analysis
An entirely general commingled reservoir solution
as a function of transient time ∆t. Gas reservoir
developed by Spath et al. (1994) using Laplace trans-
analyses are most properly evaluated for drawdown
form solutions of the reservoir performance of each
transients using the appropriate pseudopressure dif-
reservoir layer eliminates the linear flow restriction
ferences of ∆m(p ) = m(p i ) – m(p wf ) for a draw-
of the Bennett et al. and Camacho-V. et al. analyses.
down and ∆m(p ) = m(p wf ) – m(p i ) for an injection
transient and the elapsed transient pseudotime ∆ta.
For shut-in transients, similar pressure graphing
12-4. Well test analysis of vertically functions are used. For pressure buildup transients in
fractured wells oil or multiphase analyses, the pressure buildup rise
is ∆p = pws – pwf (∆t = 0), where pws is the bottomhole
Analysis of the transient behavior of a finite-conductivity
shut-in pressure. For pressure falloff transients (fol-
fracture to obtain reliable estimates of the reservoir and
lowing an injection transient), the appropriate graph
fracture properties requires proper identification of the
function is ∆p = pwf (∆t = 0) – pws. For gas reservoir
flow regimes exhibited in the transient data. The analy-
analyses, the pseudopressure equivalents of these
sis procedure includes (1) log-log diagnostic analyses
functions are used.
of both the pressure and pressure derivative responses,
Shut-in transients may be analyzed using draw-
(2) specialized Cartesian analysis and verification of
down solution type curves and Agarwal et al.’s
identified flow regimes, (3) simulation of the entire
(1979) “effective” time correlation parameter:
transient history using the determined reservoir and
fracture parameter estimates and (4) a final validation t p ∆t
∆te = . (12-53)
check of the obtained parameter estimates in conjunc- t p + ∆t
tion with the flow regime demarcation limits.
A prefracture estimate of the formation average per- The producing time tp prior to the shut-in transient
meability k is necessary for the interpretation of frac- and shut-in transient elapsed time ∆t correlate the
ture properties in the postfracture transient analysis, shut-in transient behavior with the drawdown solu-
unless the transient data contain at least some pseudo- tion. This correlation parameter is derived directly
radial flow, for which an estimate of formation perme- from the superposition-in-time principle for a two-
ability can be computed directly from the transient rate transient history in which the last rate step is
data. Depending on the formation permeability, frac- the shut-in transient. For a multirate production his-
ture half-length and conductivity, the pseudoradial tory prior to the final shut-in transient, it is generally
flow regime may not be exhibited for an extensive preferred to use continuous superposition of the pro-
period of time, making it impractical to obtain post- duction history in the analysis instead of Agarwal et
fracture estimates of formation permeability with well al.’s effective time.
tests. However, the synergy of history matching long- For gas reservoirs, the analysis may be performed
term production with well test data (e.g., Agarwal et either with dimensional time or preferably with the
al., 1979) may overcome this limitation.

pseudotime equivalent of the time-scale function. 14.8 psc Tct qg 14.8 psc Tct qg
C= ∆tausl = ∆tausl .
The effective pseudotime equivalent of Eq. 12-53 is Tsc ∆m( p)usl  ∂∆m( p) 
Tsc  ∆ta 
∆tae =
( )
ta t p ta ( ∆t )
(12-54)
 ∂∆ta  usl
t (t ) + t ( ∆t )
.
a p a
(12-59)
Pressure buildup analysis of a pseudoradial flow

regime using classic semilog analysis procedures may
12-4.2. Fracture storage linear flow
be performed with Horner’s (1951) approximation:
analysis
Qp
tp = , (12-55) The fracture storage linear flow regime is generally
qlast exhibited clearly in the transient behavior of finite-
where Qp is the cumulative production of the principal conductivity fractures only when wellbore storage
fluid phase of the analysis and qlast is the production effects are minimized, typically with downhole isola-
rate of the flowing transient immediately prior to the tion devices. When this behavior is exhibited in the
shut-in transient. transient response, an analysis can be performed that
provides an estimate of the fracture diffusivity/con-
ductivity relationship √η f /kf w. If only fracture stor-
12-4.1. Wellbore storage dominated age dominated flow behavior is used, the hydraulic
flow analysis diffusivity and conductivity cannot be determined
separately. Separation of the quantities is obtained in
The wellbore storage dominated transient behavior
conjunction with analysis of the other fracture flow
discussed in Section 12-2.1 (unit-slope behavior) is
regimes exhibited in the data to determine fracture
described mathematically by
conductivity.
tD The log-log fracture storage linear flow behavior
pwD = . (12-56)
CD of an oil reservoir analysis can be evaluated using
the pressure derivative response ∆t(∂∆p/∂∆t). The
In dimensional units, the log linear behavior of the fracture hydraulic diffusivity can be determined (for
∆p function versus transient time is a known fracture conductivity) with
qB  ∂∆p 
2
∆p = ∆t. (12-57) 2 ∆t
24C  k wh   ∂∆t  fsl
ηf =  f f
 , (12-60)
The wellbore storage coefficient C can therefore be  4.06qo µ o Bo  ∆t fsl
computed directly using the coordinates of any point
on the unit-slope ∆p line (∆tusl, ∆pusl, where the sub- where the subscript fsl denotes fracture storage linear
script usl denotes unit-slope line) or from the deriva- flow. Similarly, fracture hydraulic diffusivity may be
tive ∂∆p/∂∆t with obtained in a gas reservoir analysis using pseudo-
pressure and pseudotime with
qB qB
C= ∆tusl = ∆tusl ∂∆m( p) 
2

24 ∆pusl  ∂∆p   ∆ta 
24  ∆t  k wh T 
2
∂∆ta  fsl
 ∂∆t  usl η f =  f f sc 

. (12-61)
1448qg Tpsc  ∆tafsl µ gi ct
(12-58)
for liquid flow analyses. Either the pressure or pres- Verification graphs can then be constructed by
sure derivative responses can be used because they graphing ∆p versus ∆t1/2 (or ∆m(p) versus ∆ta1/2 for
are identical (assuming constant wellbore storage). gas) on a Cartesian scale, for which the fracture stor-
A similar relationship for determining the wellbore age linear flow behavior is characterized by a straight
storage coefficient for gas wells is line passing through the origin for oil and gas, respectively:
8.13qo µ o Bo
( )
1/ 2
∆p = η f ∆t (12-62)
k f wh f

and near-wellbore conductivity alternations after the

( )
2895qg Tpsc
∆m( p) =
1/ 2
η f µ gi ct ∆ta . (12-63) end of the bilinear flow regime clearly indicates
k f wh f Tsc
whether the dimensionless fracture conductivity is
less than, equal to or greater than 1.6 (Fig. 12-17).
If the transient data tail off upward above the bilin-
12-4.3. Bilinear flow analysis ear flow straight line, CfD > 1.6; if the transient data
trend below the bilinear flow straight line, CfD < 1.6.
The analysis of transient data that exhibits bilinear
flow behavior is performed by a log-log diagnostic
analysis of the pressure and pressure derivative 12-4.4. Formation linear flow analysis
behavior, with a Cartesian analysis and verification
of the identified flow regime. On a log-log graph of The formation linear flow regime is exhibited in the
∆p versus ∆t (or ∆m(p) versus ∆ta for gas reservoirs), transient behavior of a finite-conductivity fracture
the bilinear flow regime is identified as that with only if CfD > ~80. The fracture half-length can be
quarter-slope behavior on both the pressure and pres- determined from formation linear flow transient data
sure derivative responses. using log-log pressure derivative behavior for oil and
The fracture conductivity kf w can be determined gas reservoir analyses, respectively:
using any point on the log-log derivative quarter- 1/ 2
2.03qo Bo  µ o ∆tlf 
slope line. The interpretation relationships for oil and xf =   (12-68)
gas reservoir analyses, respectively, are  ∂∆p   φct k 
h  ∆t
2  ∂∆t  lf
 
   ∆tbf 
1/ 2
 ∆talf 
1/ 2
11.025qo µ o Bo 724 qg Tpsc
kf w =     (12-64) xf =   , (12-69)
  ∂∆p    φµ o ct k   ∂∆m( p)   kφ 
 hh f  ∆t ∂∆t   Tsc h  ∆ta
 bf   ∂∆ta  lf
2 where the subscript lf denotes formation linear flow.
 
  Cartesian graphs of ∆p versus ∆t1/2 for oil reser-
  ∆tabf  , (12-65)
1/ 2
voirs or ∆m(p) versus ∆ta1/2 for gas reservoirs are
k f w = 
3927.3qg Tpsc
 ∂∆m( p)    φk  used to verify that the transient data selected in the
 hh f Tsc  ∆ta    
log-log derivative analysis are exhibiting formation
  ∂∆ta  bf 
linear flow. The Cartesian graph relations describing
where the subscript bf denotes bilinear flow. the formation linear flow for oil and gas reservoirs,
Verification graphs are generated on Cartesian respectively, are
scales using ∆p versus ∆t1/4 for oil reservoir analyses 1/ 2
and ∆m(p) versus ∆ta1/4 for gas reservoir analyses. 4.06qo Bo  µ o ∆t 

∆p =   (12-70)
The straight line is characterized by the relationships hx f  kφct 
for oil and gas reservoir analyses, respectively, as 1/ 2
1448 psc Tqg  ∆ta 
44.1qo µ o Bo  ∆t 
1/ 4
∆m( p) =   . (12-71)
∆p = Tsc hx f  kφ 
  (12-66)
hh f k f w  φµ o ct k 
1/ 4
15, 710 qg Tpsc  ∆ta  12-4.5. Pseudoradial flow analysis
∆m( p) =   . (12-67)
Tsc hh f k f w  φk  Analysis of the pressure transient behavior of a finite-
conductivity fracture that exhibits pseudoradial flow
The Cartesian graphs for the analysis and verifica-
behavior involves log-log analysis of the transient data
tion of bilinear flow also provide additional informa-
and classic semilog analysis procedures. The transient
tion concerning the fracture conductivity. As dis-
behavior of a finite-conductivity fractured well during
cussed in Section 12-2.3, the deviation of the tran-
the pseudoradial flow regime is discussed in Section
sient data from the straight-line flow of the bilinear
12-2.5.

The formation permeability can be determined the pressure and derivative responses at late times in the
directly using any point on the log-log derivative shut-in transient. Production time effects appear in the
curve during the pseudoradial flow regime (where transient data as a steep increase in the log-log deriva-
the dimensionless derivative is a horizontal line tive response and a generally smaller (although it can
equal to 0.5). The dimensional derivative is also be significantly greater) increase in the pressure be-
a horizontal line, and the permeability is evaluated havior. Commonly, the derivative response caused
using the coordinates of a point on this straight line by production time effects is greater in magnitude than
for oil and gas reservoirs, respectively: the ∆p response.
Production time effects are due to the simple physi-
70.6qo µ o Bo
k= (12-72) cal principle that each time there is a unit step change
 ∂∆p 
h  ∆t in the well production rate, a new pressure pulse is
 ∂∆t  pr propagated away from the well into the reservoir. The
velocity at which the pressure pulse propagates in the
25,150 qg Tpsc
k= , (12-73) reservoir is a function of the formation properties and
 ∂∆m( p)  time. In the case of vertically fractured wells, this also
hTsc  ∆ta 
 ∂∆ta  pr includes the effect of fracture length.
If a given reservoir feature (boundaries, reservoir het-
where the subscript pr denotes pseudoradial flow. erogeneities, interference from offset wells, etc.) or
The apparent steady-state skin effect caused by the fracture feature that requires the end of bilinear flow
fracture can also be determined from log-log analysis (length or conductivity near the top) is of interest, the
during the pseudoradial flow regime using a point flow period immediately prior to the shut-in transient
on the pressure curve (∆tpr, ∆ppr) for oil reservoir must be of sufficient duration that the distance of inves-
analyses or a point on the pseudopressure curve tigation from the well for the flow period exceeds that
(∆tapr, ∆m(p)pr) for gas reservoir analyses for oil required to see the same reservoir or fracture feature in
and gas reservoirs, respectively: the interpreted shut-in transient. The post-treatment
cleanup process must also be completed, as this period
kh∆p pr 1   0.000264 k∆t pr   can be of substantial duration in lower permeability
s= − ln   + 0.8091
141.2 qo µ o Bo 2   φµct rw
2
  reservoirs (Holditch, 1979).
The use of continuous superposition-in-time in the
(12-74) transient analysis does not remove or eliminate pro-
duction time effects in the transient data. They are the
khTsc ∆m( p) pr 1   0.000264 k∆tapr   actual response of the reservoir, and any interpretation
s= − ln   + 0.8091 .
50, 300 psc Tqg 2   φrw
2
  that includes the identification of a reservoir or frac-
ture feature, such as a boundary or fracture length,
(12-75)
from data distorted by production time effects is
Semilog analysis procedures may also be used to incorrect or highly questionable at best.
obtain estimates of the formation permeability and To distinguish between what may actually be the
apparent steady-state skin effect of vertically frac- effect of reservoir or fracture limits and production
tured wells (see Chapter 2). time effects, the duration of the flow period immedi-
ately prior to the pressure buildup transient can be
easily checked. If the flow period is of greater dura-
12-4.6. Well test design considerations tion than the pressure buildup, then the response
Analysis of the pressure buildup behavior of a fractured observed in the transient data is actually a reservoir
well to obtain meaningful and reliable reservoir and or fracture feature. However, if the elapsed shut-in
fracture parameter estimates requires an accurate record time at which an anomalous behavior in the deriva-
of the well flow history. The stabilized flow period tive response occurs corresponds to the duration of
immediately prior to the pressure buildup should be the stabilized flow period prior to the buildup, the
of equal or greater duration than the buildup transient. transient response can be strongly suspected as
Otherwise, production time effects are observed in both resulting from production time effects.

12-4.7. Example well test analyses assume that the effective hydraulic flowing height
hf in the fracture is equal to the net pay thickness h
To demonstrate the appropriate well test analysis of the reservoir.
procedures that should be used in the interpretation Table 12-3 is a summary of the recorded pressure
of post-treatment fractured well buildup behavior, buildup response of the well that was propped with
examples are presented in this section for two adja- bauxite (Well A). The corresponding pressure buildup
cent wells in the same reservoir. Both wells were behavior of the sand-propped fracture (Well B) is pre-
fracture treated with identical fluid volumes. The sented in Table 12-4. Both wells were produced at a
only difference between the fracture treatments of relatively constant production rate of 100 STB/D
the two wells is that one treatment used bauxite for using temporary production facilities for 2 months
proppant and the other well was propped with con- to effectively clean up the fracture treatment from the
ventional fracture sand. wells. The wells were then shut in for 2 months to
The reservoir is relatively uniform in thickness (h = monitor pressure buildups and evaluate the formation
100 ft), homogeneous and isotropic, and it is assumed potential and fracturing treatment effectiveness.
that all reservoir and fluid properties are initially iden- During this time, the wells were connected to perma-
tical for the two analyses: φ = 10%, So = 90% and nent production facilities and a pipeline connection.
Sw = 10%, pw = ~6000 psia, Bo = 1.0588 RB/STB, Analysis of the pressure buildup behavior of the
µo = 5 cp and ct = 3 × 10–6 psi. The wellbore radius bauxite-propped fracture in Well A begins with prepa-
for both wells is rw = 0.33 ft. For these analyses, ration of a log-log diagnostic plot of the ∆p and pres-
Table 12-3. Pressure buildup data for high-conductivity fracture (Well A).
Transient time, Bottomhole shut-in Transient time, Bottomhole shut-in

∆t (hr) pressure, pws (psia) ∆t (hr) pressure, pws (psia)
0.0216 2393.15 4.2101 3788.99

0.0280 2412.85 4.8118 3863.52
0.0365 2435.41 5.5939 3950.32
0.0474 2461.43 6.3363 4025.48
0.0616 2491.65 7.1509 4101.00
0.0801 2526.88 8.0420 4176.81
0.1041 2567.82 9.0155 4252.88
0.1354 2614.81 10.0780 4329.15
0.1760 2667.76 11.2378 4405.50
0.2288 2726.14 12.5053 4481.89
0.2975 2789.26 13.8932 4558.21
0.3867 2856.57 15.4175 4634.33
0.5027 2928.02 17.0986 4710.19
0.6535 3004.36 19.2842 4797.23
0.8199 3075.96 21.3978 4872.09
1.0363 3155.61 24.1454 4956.36
1.2632 3228.99 26.8935 5029.66
1.5583 3312.25 30.4660 5110.06
1.8529 3386.12 34.2129 5181.19
2.2358 3470.35 39.0838 5256.56
2.6141 3544.49 45.4160 5333.58
3.1057 3629.27 52.3480 5399.85
3.5859 3703.47 60.0000 5457.39

Table 12-4. Pressure buildup data for moderate-conductivity fracture (Well B).
Transient time, Bottomhole shut-in Transient time, Bottomhole shut-in

∆t (hr) pressure, pws (psia) ∆t (hr) pressure, pws (psia)
0 2182.56 0.8188 3068.21

0.0001 2219.87 1.0336 3147.92
0.0002 2232.96 1.2586 3221.26
0.0003 2239.68 1.5510 3304.52
0.0004 2243.52 1.8430 3378.40
0.0005 2247.82 2.2225 3462.65
0.0007 2252.65 2.5971 3536.81
0.0009 2258.07 3.0842 3621.62
0.0012 2264.08 3.5599 3695.85
0.0016 2270.65 4.1783 3781.39
0.0020 2277.74 4.7746 3855.95
0.0026 2285.32 5.5497 3942.76
0.0034 2293.39 6.2857 4017.94
0.0045 2302.05 7.0934 4093.48
0.0058 2311.46 7.9772 4169.31
0.0076 2321.91 8.9427 4245.39
0.0098 2333.71 9.9966 4321.68
0.0128 2347.19 11.1468 4398.06
0.0166 2362.62 12.4038 4474.48
0.0216 2380.23 13.7799 4550.83
0.0280 2400.29 15.2908 4627.00
0.0365 2423.21 16.9567 4702.92
0.0474 2449.59 19.1223 4790.07
0.0616 2480.19 21.2142 4865.00
0.0801 2515.81 23.9336 4949.43
0.1041 2557.15 26.6491 5022.83
0.1354 2604.54 30.1792 5103.44
0.1760 2657.88 33.8730 5174.70
0.2288 2716.67 38.6749 5250.32
0.2975 2780.18 44.9170 5327.74
0.3867 2847.89 51.7180 5394.18
0.5027 2919.78 60.0000 5456.69
0.6535 2996.59
sure derivative behavior versus the equivalent shut-in transient was conducted long enough to observe some
time (Fig. 12-34). By observation of the character and pseudoradial flow behavior.
shape of pressure and derivative behavior, it is readily Selecting the best-fit half-slope linear derivative
determined that the well is intersected by a high-con- behavior of the data results in a computed estimate
ductivity fracture because a significant portion of the of the formation permeability and fracture half-
data exhibits linear flow (half-slope behavior) and the length squared kxf2 by rearrangement of Eq. 12-68.

10,000
1000
++++++++++++++
+++++++
+
++
++++
+++
∆p, derivative (psia)
++
+ ++
+ + ++
100 ++
+ ++
++
++
+ ++
++ + +
+ ++
10 + + ++
+
0
1E–3 1E–2 1E–1 1E+0 1E+1 1E+2 1E+3
Equivalent time (hr)
Figure 12-34. Diagnostic graph of a high-conductivity fracture.
Selecting a point on the linear flow portion of the 1500

derivative curve (1.14 hr, 107.5 psia):
4.1292 qo2 µ o Bo2 ∆tlf +
kx 2f = 1000 +
 δ∆p 
2
+
∆p (psia)
h 2  ∆t φct +
 δ∆t 
+
+
+
+
(4.1292)(100)2 (5)(1.0588)2 (1.14) 500 +
+
= = 7607 md - ft 2 . +
(100)2 (107.527)2 (0.1)(3 × 10 −6 )
+
+
++
++++
(12-76) ++
+++++
+++++
0
The verification graph of the selected linear flow 0 1 2 3 4 5 6 7
regime is prepared by graphing the buildup pressure ∆t1/2 (hr1/2)
rise versus the square root of the shut-in time (Fig. Figure 12-35. Formation linear flow verification graph.
12-35). The equation of the straight line is
4.06qo Bo µ o ∆t
∆p = The next step in the analysis of the transient be-
h kx 2
φct
f
havior of Well A is to perform the log-log type-curve
(4.06)(100)(1.0588) 5 match of the pressure buildup behavior using a finite-
= ∆t = 201.2 ∆t .
(100) 7607 (0.1)(3 × 10 −6 ) conductivity type curve (such as Cinco-Ley et al.,
1978) or a well test analysis computer model that can
(12-77)
generate the corresponding type-curve dimensionless
Because the same shut-in time range of the data solutions (Fig. 12-36). The matching parameters of
that lies on the best-fit derivative linear flow agrees this analysis indicate an effective k = 0.442 md,
with the data following the linear flow behavior on xf = 130 ft and 10,000-md-ft fracture conductivity.
the verification graph, the linear flow regime of the A quick check of the agreement between the results of
diagnostic graph has been properly selected and kxf2 type-curve matching and the diagnostic graph analysis
is approximately 7600 md-ft2.

10
++
++
++
1.0
+++
++
++
+++
++
++
++++++++++++
++
+++++
++
++
++
+
++++
++
+++
pD, tD • pD´
++
++
+++
++
0.1 +++
++
+
++
++
++
++
++
+
++
++
++
++
++
++ +++
++
+
+
++++
++
0.01
+
++
++
+
+ ++
+
0.001
1E–6 1E–5 1E–4 1E–3 1E–2 1E–1 1E+0 1E+1 1E+2
tD
Figure 12-36. Type-curve match of a high-conductivity fracture.
is obtained by computing the corresponding kxf2 value 70.6qo µ o Bo

k = (12-78)
of the type-curve match:  δ∆p 
h  ∆t
 δ∆t  pr
kx 2f = (0.442)(130) = 7470 md-ft.
(70.6)(100)(5)(1.0588)
= = 0.4 md.
This indicates relatively good agreement between the (100)(980)
two analyses.
Because pseudoradial flow was indicated on the The apparent radial flow steady-state skin effect
diagnostic graph (i.e., the derivative trends to one-half the well is computed using the last pressure point
at the end), a classic semilog analysis is valid for this (720 hr, 3222.4 psia):
transient interpretation. The representation of the clas-
(0.4)(100)(3222.39) 1   0.000264(0.4)(720)  
sic Horner semilog analysis of these transient data in s = − ln + 0.8091
141.2(100)(5)(1.0588) 2   0.1(5)(3 × 10 −6 )(0.33) 
2

Fig. 12-37 exhibits a semilog straight-line slope of
1880 psi/log cycle, extrapolated initial pressure esti- = −5.21 .
mate p* = 6023 psia, effective ko = 0.458 md and
These results are in agreement with the permeabil-
apparent radial flow steady-state skin effect of –5.02
ity and skin effect estimates obtained for Well A
owing to the presence of the vertical fracture. The
using the semilog analysis procedure. In summary, it
interpretation equations for the classic semilog analy-
is concluded that the interpretation indicates that the
sis procedure are presented in Chapter 2 and are not
reservoir associated with Well A has ko = ~0.45 md,
duplicated in this example. However, an additional
the effective propped half-length is approximately
method for determining the effective permeability of
130 ft, and the fracture conductivity is very high,
the reservoir and the apparent radial flow steady-state
about 10,000 md-ft. This corresponds to CfD = 171.
skin effect of the well is provided by Eqs. 12-78 and
As a final check on the validity of the analysis, the
12-74, respectively. The effective permeability is
start and end times for the formation linear flow
regime can be computed using Eqs. 12-13 and 12-14:

5500
++
+
+
+
+
5000 +
+
+
+
+
+
4500 +
+
+
+
+
+
4000
+
+
+
pws (psia)
+
+
+
+
3500 +
+
+
+
+
+
3000 +
+
+
++
++
++
2500 ++
+++
++++++
+ + + + + + + + ++
+ + +
2000
1E+0 1E+1 1E+2 1E+3 1E+4 1E+5 1E+6
Horner time ratio
Figure 12-37. Semilog graph of a high-conductivity fracture.
100 nostic graph of the data in Fig. 12-38, it is readily

t Dblf = = 0.00342
(171)2 observed that bilinear flow is exhibited at early time,
followed by a relatively long transition period and
finally by a short pseudoradial flow regime near the
t Delf = 0.016 .
end of the transient. No linear flow regime is exhib-
These dimensionless times correspond to dimen- ited in the pressure buildup behavior of this well.
sional times of A quarter-slope linear best-fit portion of the deriva-
t Dblf φµ o ct x 2f tive curve is selected and an estimate of the reservoir
tblf = permeability and fracture conductivity squared k(kf w)2
0.000264 k
is obtained by rearranging Eq. 12-64 and using a point
(0.00342)(0.1)(5)(3 × 10 −6 )(130)2 on the best-fit derivative line (0.0012 hr, 17.3 psia):
= = 0.73 hr
0.000264(0.45) 4
 
11.025q µ B  ∆tbf
t Delf φµ o ct x 2f
( )
k kf w =  o o o 
2
telf =   δ∆p   φµ o ct h 2 h 2f
  ∆t δ∆t  
0.000264 k
 bf 
(0.016)(0.1)(5)(3 × 10 −6 )(130)2
= = 3.41 hr . 11.025(100)(5)(1.0588) 
4
0.012
0.000264(0.45) =  (0.1)(5)(3 × 10 −6 )(100)4
 17.3102 
These linear flow period demarcation times agree = 1.034 × 10 6 md 3 - ft 2 .
with the transient behavior exhibited on the diagnos-
tic graph. (12-79)
The interpretation of the transient behavior of Well Verification of the selected bilinear flow behavior
B, which was propped with sand, is performed in a is shown in Fig. 12-39, and the equation of the
manner similar to that used for Well A. In the diag- straight line is

10,000
+
++
++
++
++
++
+++
++
++
++
++
1000 +++++++++++++
++
+++++
++
++++
+
+
+
+
++
+++
+
+
+++
+
+
+
+
+++
∆p, derivative (psia)
+
+
++
+
++
+
+
++
+
+
+
+
+
+
+
++
+
100
+
+
++
++
+
+
+
+
+ ++
+
+
+
+
++
+
++ + +
++
+ + ++ +
+
10
0
1E–3 1E–2 1E–1 1E+0 1E+1 1E+2 1E+3
Equivalent time (hr)
Figure 12-38. Diagnostic graph of a moderate-conductivity fracture.
3500
+
+ +
3000
+
+
+
+
+
+
+
2500 +
+
+
+
+
+
2000 +
+
+
∆p (psia)
+
+
+
+
1500 +
+
+
+
+
+
1000 +
+
+
+
+
+
+
500
+++
++
++++++
+
+++++
0 +++++++++++
0 1 2 3 4 5 6 7
∆t1/2 (hr1/2)
Figure 12-39. Bilinear flow verification graph.

1/ 4
  0.1 0.1
44.1qo µ o Bo ∆t t Debf = = 2 = 8.45 × 10 .
−5
∆p =  2
(C ) ( )
2
hh f  o t f( )
 φµ c k k w 
 fD h fD 344
1/ 4
44.1(100)(5)(1.0588)  ∆t  In terms of dimensional time, this becomes
=  (0.1)(5) 3 × 10 −6 1.034 × 10 6 
(100)(100)  ( )( ) t Debf φµ o ct x 2f
tebf = (12-81)
= 209.2( ∆t ) .
1/ 4 0.000264 k
(12-80) =
(8.45 × 10 )(0.1)(5)(3 × 10 )(124)
−5 −6 2
0.000264(0.409)
The same pressure buildup data points that lie on the
diagnostic plot quarter-slope linear behavior corre- = 0.0181 hr .
spond to the data points that are on the verification
graph line, indicating that the proper bilinear flow This result agrees with selection of the bilinear
regime data were selected on the diagnostic graph. flow regime range on the diagnostic graph because
The type-curve matching procedure is identical to the chosen data on the graph are prior to this time.
that discussed for Well A, with the resulting type Another qualitative validity check that can be made
curve match for Well B shown in Fig. 12-40. The to ensure the internal consistency of the interpreta-
determined values are ko = 0.409 md and xf = 124 ft; tion results is to compare the type-curve match esti-
the fracture conductivity is 1746 md-ft. The match mate of the permeability and fracture conductivity
results in CfD = 34.4. By referring to the expressions squared k(kf w)2 with the result of the diagnostic
for obtaining estimates for the end of the bilinear analysis:
( )
flow regime (Eq. 12-11), a check of the validity of
k k f w = (0.409)(1746) = 1.5 × 10 6 md 3- ft 2.
2 2
the range of data selected for the bilinear flow
regime on the diagnostic plot can also be obtained:
10
++
++
++
1.0
+++
++
+++
+++
++
++
++ +++
++
+++++
++
++++
++
+
+
pD tD • pD´ (psia)
+
++
++++
++
+++
++
+++
++
0.1 +++
++
++
++
+
++
+
++
++
++
+
++
++
++
++
++
+
++
+
+
+
++
+
++++
0.01 ++
+ ++ +
+ +
0.001
1E–6 1E–5 1E–4 1E–3 1E–2 1E–1 1E+0 1E+1 1E+2
tDxf
Figure 12-40. Type-curve match of a moderate-conductivity fracture.

Again, good agreement is obtained between the The apparent radial flow steady-state skin effect of
results of the two analyses. the fractured well is evaluated using Eq. 12-74:
Finally, because the pseudoradial flow regime is
also exhibited in the transient behavior of Well B, kh∆p pr 1   0.000264 k∆tpr  
a semilog analysis is valid (Fig. 12-41). The results s = − ln  + 0.8091
141.2 qo µ o Bo 2   φµct rw
2
 
of the semilog analysis indicate a straight-line slope
of 1870 psi/log cycle, extrapolated initial pressure
(0.44)(100)(3273.69)
estimate p* of 6019 psia, effective permeability of =
0.461 md and apparent radial flow steady-state skin 141.2(100)(5)(1.0588)
effect of –4.98.
1   0.000264(0.44)(720)  
By performing the analysis of the pseudoradial − ln 2  + 0.8091
flow regime as discussed previously for Well A, the 2   (0.1)(5)(3 × 10 )(0.33) 
−6

effective permeability is computed from the deriva- = −5.05.
tive value at the end of the transient (843.2 psia) and
Eq. 12-72:
The transient behavior of these two wells, which
70.6qo µ o Bo 70.6(100)(5)(1.0588) were stimulated using different types of proppants
k= = = 0.44 md . with different conductivities, demonstrates the various
 δ∆p 
h  ∆t
(100)(843.21) flow regimes and analysis procedures that are com-
 δ∆t  pr monly employed in postfracture formation evaluation.
5500
5000
4500
4000
p ws (psia)
3500
3000
2500
2000
1E+0 1E+1 1E+2 1E+3 1E+4 1E+5 1E+6
Horner time ratio
Figure 12-41. Semilog graph of a moderate-conductivity fracture.

12-5. Prediction of fractured well wellhead pressure production condition that requires
simultaneous solution of the reservoir inflow perfor-
performance mance and wellbore outflow analysis to estimate the
Simulation of the future performance of a vertically productivity of the well.
fractured well can be performed with a variety of dif- Meng et al. (1982) applied the production systems
ferent types of reservoir models. The most general analysis approach specifically to vertically fractured
type of model for estimating the future production wells and developed a correlation technique for esti-
performance of a fractured well is a conventional mating the inflow performance of finite-conductivity
finite-difference (or finite-element) reservoir simulator. fractures in gas reservoirs. This analysis procedure
Other types of predictive models for estimating the was utilized by Hunt (1986) to develop a systems
performance of vertically fractured wells are numeri- analysis procedure for evaluating the rate-transient
cal models developed using analytic solutions to the performance of fractured gas wells using analytic
diffusivity equation governing fluid flow in the system solutions.
and the appropriate initial and boundary conditions. A more general approach reported by Poe et al.
Reservoir heterogeneities and anisotropy, bound- (1995) for estimating the future performance of ver-
aries, multiple wells and changing inner boundary tically fractured wells incorporates reservoir voiding
conditions can be readily specified with finite-differ- effects in the production system analyses. The com-
ence reservoir simulation models to simulate rela- mingled reservoir evaluation procedure of Spath et
tively realistic production scenarios. Finite-difference al. (1994) was used in that study to evaluate the pro-
models are also ideally suited for solving nonlinear duction performance of multilayer reservoirs using
problems such as those associated with nonlinear Laplace space analytic solutions for the transient
fluid pressure-volume-temperature (PVT) properties, behavior of finite-conductivity fractures. The com-
multiphase flow, non-Darcy fracture and reservoir mingled reservoir algorithm allows mixed reservoir
flow, and stress-dependent fracture conductivity. completion types, such as a combination of fractured
These effects are important for the long-term (e.g., and unfractured reservoir layers, in the prediction of
20-year) economic analyses commonly used for tight the future performance of the well.
gas wells. In almost all cases of interest, hydraulic fractures
Finite-difference reservoir simulation models are are vertical and normal to the minimum horizontal
generally the preferred method for estimating the stress direction. Horizontal wells can thus be drilled
future performance of vertically fractured wells. either normal or longitudinal to the fracture azimuth.
Limitations related to the use of a finite-difference The first configuration results in orthogonal fractures
model are and has been found applicable for relatively low-per-
meability formations (Brown and Economides, 1992)
• quantity and quality of reservoir and fracture
and is discussed in Chapter 10 with multiple orthog-
information that is required as input data for the
onal fractures.
model
The other configuration, involving longitudinally
• generally greater computational expense associ- fractured horizontal wells, was first studied by Larsen
ated with the use of finite-difference reservoir and Hegre (1991). Their work neglects the horizontal
models over analytic techniques. flow component and the nonuniformity of the vertical
Numerical models developed using analytic solutions pressure distribution in the fracture. A more general
for the transient behavior of vertically fractured wells approach that is free of these restrictions was intro-
have been used for many years, largely because of duced by Valkó and Economides (1996a). They con-
these limitations for finite-difference simulation models. sidered the same reservoir-fracture system as the one
Brown et al. (1984) demonstrated the utility of studied by Cinco-Ley and Meng (1988), including
analytic models for future well performance predic- the possibility of double porosity. In the Valkó and
tion. Brown et al. introduced the use of a systems Economides case, the well is horizontal and located
analysis approach for estimating the production in the middle of the fracture (Fig. 12-42). The frac-
behavior of a well for which the system inner bound- ture is longitudinal with respect to the well, and com-
ary condition may be specified at a system location munication between the fracture and the well is estab-
other than the sandface. An example is the constant lished along the total fracture length. The height of

2xf
Figure 12-42. Horizontal well intersected by a longitudinal vertical fracture.
the fracture coincides with the thickness of the hydro- is irrelevant (i.e., the fracture behaves nearly as an
carbon-bearing stratum. The performance of the frac- infinite-conductivity fracture), at lower CfD values
ture depends on both the dimensionless fracture con- the production increase is significant. At a given CfD
ductivity and the ratio of the height to the half-length, the increase is more pronounced at lower hD values,
called the dimensionless height hD and providing the which provide increased vertical dimensionless
dimensionless conductivity for vertical flow. conductivity (Fig. 12-44). On both Figs. 12-43 and
The results show that CfD = 1 is no longer neces- 12-44 the highest production increase is already near
sary if the fracture is intersected by a horizontal well the theoretical maximum production increase,
instead of a vertical well and the dimensionless assuming an idealized infinite-conductivity fracture.
height is small or moderate. Figure 12-43 shows the The reason for this effect is that the horizontal
folds of increase in production from the same frac- well acts as a high-conductivity streak in an other-
ture intersected by a horizontal well instead of a wise limited-conductivity flow conduit. The lower
vertical well with a small dimensionless height. the dimensionless height and fracture conductivity,
Although at large CfD values the location of the well the more pronounced the effect of the high-conduc-
6 6
hD = 0.25 CfD = 0.1
5 5 hD
CfD
0.25
0.1 0.5
4 1 4 1
NDH / NDV
NDH / NDV
10 2
3 3
2 2
1 1
0 0
10 –1 10 0 10 1 10 2 10 –1 10 0 10 1 10 2
tD tD
Figure 12-43. Ratio of dimensionless cumulative produc- Figure 12-44. Ratio of dimensionless cumulative produc-
tion (longitudinally fractured horizontal well to a fractured tion (longitudinally fractured horizontal well to a fractured
vertical well) for dimensionless height hD = 0.25 (Valkó vertical well) for dimensionless fracture conductivity
and Economides, 1996a). CfD = 0.1 (Valkó and Economides, 1996a).

tivity streak. The model predicts an almost infinite The key advantages of a horizontal well combined
conductivity fracture behavior for CfD = 0.1 and with a longitudinal fracture is that a much smaller frac-
hD = 0.25, in contrast to the finite-conductivity ture width (than that suggested for the fractured vertical
behavior of a vertical well intersected by a similar well) may be sufficient to achieve a certain production
fracture. For the vertical well, a CfD at least 2 orders goal, reducing the necessity of very wide fractures,
of magnitude higher is required to see the same which are generally obtained only with the tip screen-
nearly infinite conductivity behavior. In addition, out technique. The fracture also overcomes two limita-
a large CfD value is difficult to achieve in high per- tions of a conventional horizontal well: removing
meability for a long fracture from a vertical well. drilling-induced damage and mitigating the adverse
effect of the relatively low vertical permeability.


◆
Introduction to
Matrix Treatments
R. L. Thomas, Schlumberger Dowell
L. N. Morgenthaler, Shell E&P Technology Company
This chapter is dedicated to Alfred R. (Al) mination, treatment design, job execution and treat-
Hendrickson, the father of modern-day acidizing. ment evaluation. Although many matrix stimulation
Al made numerous unique discoveries during his treatments are performed in an unsystematic manner,
career and is best known for his work in sandstone the success of matrix stimulation can be enhanced
acidizing. The knowledge he offered the industry is when each of these activities is performed properly.
the cornerstone of acidizing technology today. His Proper performance of these steps requires the inter-
energetic personality and technical contributions will action of numerous individuals with expertise in geo-
forever be appreciated. science, engineering and operations. The remaining
chapters in this volume provide detailed discussions
of the state of the art in matrix stimulation. The pur-
13-1. Introduction pose of this chapter is to provide an integrated
overview of the process of engineering successful
Matrix stimulation is a technique that has been used matrix stimulation treatments.
extensively since the 1930s to improve production Engineering a matrix treatment includes many
from oil and gas wells and to improve injection into tasks and a methodology that are accomplished on
injection wells. Sidebar 13A discusses the history of the basis of the best available data and knowledge
matrix stimulation. Matrix stimulation is accom- at hand, which is usually incomplete. This does not
plished by injecting a fluid (e.g., acid or solvent) to mean a treatment will be unsuccessful. The process is
dissolve and/or disperse materials that impair well a continuous cycle, starting with the diagnostic phase
production in sandstones or to create new, unimpaired of the design process and progressing through the
flow channels between the wellbore and a carbonate execution and evaluation phases to develop improve-
formation. In matrix stimulation, fluids are injected ments. In addition, computer systems with “advisors”
below the fracturing pressure of the formation are available to assist the process. Advisors are expert
(McLeod, 1984). At the time of this writing, it is esti- systems with a knowledge base derived from current
mated that matrix treatments constitute 75% to 80% technology.
of all stimulation treatments (matrix and fracturing)
worldwide, but the total expenditure for matrix treat-
ments is only 20% to 25% of the total for all stimula- 13-1.1. Candidate selection
tion treatments. However, because the payout time
for matrix treatments is normally days rather than Candidate selection for matrix stimulation is based on
months as it is for conventional fracturing treatments, finding wells with impaired productivity and diagnos-
engineers should master this technique. Many opera- ing the cause of the impairment. Failure of a well to
tors around the world have indicated that an average obtain economic objectives alone is not evidence of
of 40% to 50% of their wells have significant dam- impaired productivity. Failure to achieve economic
age, but routinely only 1% to 2% of their wells are objectives may be the result of limitations of the
treated every year. reservoir (e.g., pressure, permeability) or wellbore
Substantial production improvements can be (e.g., artificial lift, inadequate tubing size). Matrix
achieved with matrix stimulation if treatments are stimulation cannot solve these problems.
engineered properly. A success rate greater than 90% Candidate selection requires an accurate assess-
is reasonable. The systematic approach to stimulation ment of what a well can produce without impairment
treatments consists of candidate selection, formation and the current productivity of the well. Techniques
damage characterization, stimulation technique deter- for making these assessments rely heavily on knowl-

When this was published in World Petroleum the price of oil

13A. The history of matrix stimulation was $1.00 per barrel and acidizing was directed at limestone
formations. Since then acidizing technology has expanded,
“The results gained by the use of acid to increase oil and gas driven by oil and gas discoveries in a variety of formations, all
production are no longer hypothetical. Wherever production with unique problems.
comes from essentially limestone reservoirs, its use is indi- Sandstone acidizing with hydrofluoric acid (HF) was prac-
cated and recommended.” This quote is from “The Value of ticed in Texas in 1933 following the issuance of a patent to
Repeated Acid Treatments” by P. E. Fitzgerald of Dowell Inc. the Standard Oil Company; however, the field tests were not
in 1934. Fitzgerald went on to write, “It is recognized now that successful because of plugging of the formation. Commercial
every well is a problem in itself, and must be analyzed individ- application of HF acidizing of sandstones occurred in the Gulf
ually in order to obtain the best results.” It is interesting to Coast of Mexico in 1940 when Dowell introduced mud acid, a
note that the forefathers of matrix stimulation realized the mixture of HCl and HF. Dowell research indicated that the HCl
importance of proper well diagnosis for treatment success, helped maintain a low pH and decreased the precipitation of
because it is a key step in the matrix stimulation engineering damaging precipitates. Following this event, the application of
process used today. In addition, it is interesting that the sandstone acidizing grew rapidly.
“acidizing era” began more than 30 years following develop- As the application of acidizing expanded, several chemical
ment of the concept. It took that long to develop the technol- and mechanical problems were addressed. Numerous acid
ogy required to overcome the fundamental problem of acid additives and systems were developed to solve the problems
corrosion of the tubing and casing. As history indicates, a of acid sludging, acid-induced emulsions, spent acid cleanup,
chemical company with a need to stimulate brine wells devel- live acid penetration and fines migration. Parallel to this devel-
oped the solution of an acid corrosion inhibitor. With this key opment was the development of methods to improve zone
invention the matrix stimulation era began in full force. coverage during acidizing.
The history of acidizing dates back to 1895, when the Acidizing has progressed through the following eras:
Ohio Oil Company acidized both oil and gas wells with signifi- • 1950s and 1960s—Emphasis was on the development of
cant increases in production; however, the casing was additives to address emulsions, sludge, spent acid return
severely corroded and the process became unpopular. One and zone coverage. In addition, work on the physics of
year later, a patent was issued to Herman Frasch of Standard acidizing in limestones and the secondary reactions of
Oil Company. His patent described the use of hydrochloric sandstone acidizing was performed. The emphasis for
acid (HCl) in wells with limestone pay formations but did not clay problems shifted from clay swelling to clay migration,
address the corrosion problem. and numerous clay control agents were developed. Oil-
Not until 1928 was the use of acid again attempted. This soluble resins were introduced as diverting agents for
was when the problem of brine disposal, as well as increased improved zone coverage.
production of natural brine, became important. Dr. Herbert
Dow, early in his career, lowered bottles of acid into brine • 1970s—A need for deeper penetration of live HF acid was
wells for the purpose of increasing their production. However, addressed with various systems, including alternating
the results were not satisfactory, largely on account of the cor- stages of HCl and HF, fluoboric acid and a mixture of
rosion incurred and the expensive materials required to pro- methyformate and HF.
tect the metal equipment. Thus, the Dow Chemical Company • 1980s—Foam diversion and coiled tubing placement were
initiated a project to develop the first acid corrosion inhibitor. introduced to improve zone coverage. Production system
In 1931, Dr. John Grebe of Dow discovered that arsenic analysis became a common tool of the design engineer.
acid acted as a corrosion inhibitor. Later, copper salts were Computers were used to assist in all phases of the matrix
used with arsenic to avoid the formation of calcium arsenate process, including candidate selection, treatment design,
precipitate, and soon organic inhibitors were found to be far monitoring of execution (real-time evaluation of skin effect
superior. Soon after the discovery of the arsenic inhibitor, it evolution) and post-treatment evaluation.
was applied in the field by Dow and Pure Oil Company to • 1990s—Computers continued to evolve to faster and more
successfully treat the latter’s Fox No. 6 well in the Greendale user-friendly programs that incorporate improved produc-
Pool, Michigan, in February 1932. Five hundred gallons of tion prediction capability, economics software, geochemi-
HCl were siphoned into the well, resulting in a previously cal models and on-site evaluation techniques. Environ-
“dead” well flowing 16 BOPD. Thus, acidizing was reborn, mentally friendly additives were introduced to meet
and Dow formed the Dow Well Services Group that soon government regulations along with the development of
evolved into Dowell. Three years later, the small company a better understanding of sandstone acidizing chemistry.
Halliburton Oil Well Cementing Co. in Duncan, Oklahoma, Emphasis was placed on the entire matrix process through
began commercial acidizing service, matrix stimulation engineering.
In 1935 Drs. Grebe and Stoesser of Dowell wrote,
“Commercial acidization of oil and gas wells—although non- Great advances have been made since the first acidizing
existent four years ago, now is practiced over the entire coun- treatment was performed. As wells become deeper, with
try. Approximately 6000 oil and gas wells have been treated higher temperatures and harsher conditions along with longer
by Dowell Incorporated, to give an average increase in pro- zones as in horizontal wells, matrix technology will expand to
duction of 412 percent. In central Michigan alone, one-sixth meet operators’ needs. Matrix stimulation will continue to be
of the total oil production is the result of acid treatments, thus a useful, economical tool for production enhancement in the
indicating a net gain of $5,000,000 to the oil companies.” years to come.
edge of the formation geology and reservoir proper- voirs and well performance analysis have been pub-
ties. Methods for assessment of production system lished (Earlougher, 1977). These methods enable the
performance have been developed and are in wide engineer to quantify the extent of formation damage
use. Methods for pressure transient testing of reser- and the potential for productivity improvement.
13-2 Introduction to Matrix Treatments

13-1.2. Formation damage characterization laboratory testing and can be derived from an expert
system.
Once it has been established that a well is producing In cases where it is impossible to remove or bypass
below its potential, an assessment of the cause and damage chemically, a small-volume fracturing treat-
location of the impairment must be made (Krueger, ment may be indicated (Fletcher et al., 1995). This is
1986, 1988). Diagnosis can be based on a review of particularly true today, with the tip-screenout (TSO)
the well and field history, samples of plugging material technique applied to high-permeability reservoirs.
recovered from the field and analyzed, knowledge of Chemical additives, as well as pre- and postflushes,
formation mineral and fluid (e.g., water and oil) proper- are selected to enhance the action of the main stimula-
ties, as well as pressure testing and logging evaluation. tion fluid, prevent acid corrosion or prevent productiv-
The process of searching and sifting through the mass ity impairment from by-products of the stimulation
of data that may provide clues to the problem of a par- process (see Chapter 15).
ticular well is facilitated by databases and expert sys-
tems. In the end, the ingenuity and training of the
person analyzing the data and the application of engi- 13-1.5. Pumping schedule generation
neering tools are critical to a successful diagnosis. The and simulation
challenge is to recognize the cause of the well impair-
ment from the information available. In many cases, it The volume of each material pumped is based on an
is not possible to characterize the formation damage assessment of the amount of damage or the required
completely. If the diagnosis is uncertain, it is recom- depth of treatment and must address the potential inef-
mended to prioritize the probable causes and design a ficiency of the placement process. The location of the
treatment for the most probable scenarios. Thus, multi- damage dictates the placement technique. Both
ple damage types may be suspected, and all should be mechanical methods using tools and fluid mechanical
considered in designing the treatment. methods (e.g., particulate suspensions and foams) can
be used to ensure that the stimulation fluids contact
the formation damage.
13-1.3. Stimulation technique determination Once the volume, composition and sequence of a
treatment are established, a treating schedule can be
Selection of the stimulation technique is based on the designed on the basis of the well and reservoir proper-
productivity target, lithology, operational limitations and ties. However, the effect of acidizing on formation
various other considerations. Normally the productivity strength may impose an upper limit on the acid
target dictates the stimulation technique. For example, if strength and volume in wells that are not gravel
a 90% reduction in skin effect in a sandstone yields the packed. The pumping schedule includes the treating
target production, then matrix stimulation will probably fluid and diverter sequence and injection rate for each
be the most cost-effective technique. If matrix stimula- stage. It can be generated using empirical rules based
tion cannot be accomplished, then the feasibility of on previous field experience. In addition, the schedule
using propped fracturing should be evaluated. can be optimized to meet specific objectives for each
In carbonates, acid fracturing, propped fracturing and fluid type by using a single-phase reservoir model
matrix acidizing techniques are applicable. However, if (Perthuis et al., 1989).
matrix acidizing yields a final skin effect of –2 to –3, it A field-validated stimulation simulator should be
will probably be the most cost-effective technique. used for the systematic engineering of matrix stimu-
lation treatments. A numerical simulator can be a
valuable tool for predicting damage removal and eval-
13-1.4. Fluid and additive selection
uating skin effect evolution, flow profile and wellhead
Identification of the cause and location of the well and bottomhole pressures versus injection rate for the
impairment drives the treatment design process. The proposed pumping schedule. The simulator should
chemistry of the stimulation fluids is chosen either to take into consideration both the chemistry and damage
dissolve or disperse the impairment or in the case of removal along with the placement strategy, which is
carbonate reservoirs to create high-permeability chan- another important part of the design process. This step
nels through the damage zone. The treating fluid sys- allows the design engineer to fine tune or optimize the
tem is selected on the basis of field experience and schedule to obtain the desired results in the most cost-

effective manner. Because many key parameters in 13-2. Candidate selection

the simulation may not be accurately assessed in the
laboratory, the simulator is most useful for predicting 13-2.1.Identifying low-productivity wells
trends. Comparison to field data is critical to fully and stimulation candidates
validate the simulation model for future design work
(Bartko et al., 1996). The process of candidate selection consists of identi-
fying wells with low productivity relative to what they
are capable of producing and then evaluating possible
13-1.6. Economic evaluation mechanical problems in these wells (Thomas and
Milne, 1995).
Although a preliminary economic analysis is performed Geology, petrophysics and reservoir engineering play
during the candidate selection process, the final analysis important roles in quantifying the productive potential
should be conducted once the treatment design is final- of a given well. Ideally, a thorough understanding of
ized. Production performance is predicted by using a the reservoir geology and reservoir drive mechanics is
production forecast module and is based on initial required to quantify production potential. In many cases
and/or final skin effect calculations. Payout time, net such data are incomplete, and the engineer must rely
present value (NPV), cash flow or other financial indi- on comparison to similar wells or the field history to
cators may be determined to evaluate economic justifi- identify performance norms. The productivity of each
cations (Bartko et al., 1996). well can be mapped using a three-dimensional (3D)
surface graph to assist the identification of underpro-
ducers (i.e., candidates). For oil wells, the productivity
13-1.7. Execution
index (PI), productivity index per net pay thickness
Execution of the treatment is an often neglected step PI /h, production rate in barrels of oil per day (BOPD),
by the design engineer but is obviously critical to the production rate per reservoir porosity thickness product
process. Materials must be monitored to ensure that (BOPD/porosity-ft) or skin effect can be plotted for
they meet the specifications of the design, equipment each location. The variable selected is a function of the
must be maintained to perform properly, and personnel data available. For example, if limited data are available
on site must understand and execute their assigned (e.g., permeability-height kh is unknown) BOPD/poros-
roles. Quality control (QC) testing and training should ity-ft or h can be plotted. This process allows the engi-
be documented as standard practices (Brannon et al., neer to pinpoint likely underproducers and determine
1987). Modern data acquisition and communication their relationship with other wells in the field.
equipment make it possible to obtain and analyze Once the production potential of a well is established
detailed data during the well treatment, on and off the (to the highest certainty possible with the available
wellsite. data), it can be compared with the actual production.
This is a deceptively simple statement because it hides
the difficulty in measuring actual production in some
13-1.8. Evaluation cases. A facility’s engineers and operations staff can
Treatment evaluation consists of pretreatment, real- help ensure that an accurate assessment of current well
time and post-treatment phases. Each is an important performance is obtained.
link to treatment success and the economic impact of Once a well is diagnosed as underperforming, the
the stimulation treatment. In addition, technical evalu- reason or reasons must be determined. Appropriate
ation is required to validate and calibrate the models remedial action must be taken, including the determi-
and assumptions used in candidate selection, treatment nation of whether artificial lift is required or if the
design and execution. Treatment evaluation uses the existing lift is adequate and functioning properly. In
same well evaluation tools and knowledge of potential some cases, production is constrained by tubing size,
well productivity as candidate selection. Consistently downhole equipment restrictions or other mechanical
following the engineering concepts outlined here reasons. Stimulation will not help in these cases. Once
makes stimulation a learning process and drives mechanical reasons are eliminated as a potential cause
improvement in production performance in the field. of poor production, the remaining wells become stim-
ulation candidates.

Well screening should be based on the potential no systems analysis). Comparison of the actual rate
production increase and incremental economics. to the computed theoretical rate is made. The engineer
Obviously, wells with the greatest potential should then determines whether further evaluation is war-
be selected as candidates. This process should include ranted. One rule-of-thumb cutoff suggests that the well
determination of the maximum allowable drawdown should be stimulated if the actual rate is less than 75%
pressure before formation or sand production occurs of the theoretical. This cutoff should be used as a gross
(i.e., critical drawdown). The critical drawdown is indicator only because the theoretical rate does not
used to predict expected production and is important include effects resulting from production tubulars or
in evaluating the economic potential of the treatment separation equipment. By using this process to screen
(Weingarten and Perkins, 1992). several wells, the engineer can identify underproducers,
An expert system can be used to screen potential followed by refinement using systems analysis.
candidate wells. Its knowledge base uses a series of Once the analysis has been completed, wells with
rules to determine the suitability of a well for matrix low permeability or pressure can be identified. In
stimulation. The determination is based on the results these cases, matrix stimulation will not provide an
of pressure transient analysis (PTA) of the well or on economic improvement in performance, but fracture
the ratio of actual to theoretical flow rates. Consid- stimulation may be appropriate. Sidebar 13B dis-
eration of possible mechanical problems is also built cusses the candidate selection process applied to a
into the rules. Alternatively, a systems analysis can be highly damaged well located in the Gulf of Mexico.
used to determine the production potential of a well A useful and quick indicator of well performance
immediately following treatment. Matrix acidizing is the productivity index:
simulators and field experience indicate that a 90% q
PI = , (13-1)
reduction in skin effect in sandstones and skin effect pr − pwf
values of –2 in carbonates are conservative assump-
tions for a properly designed and executed treatment where q is the flow rate in BOPD, — p r is the average
unless local experience indicates otherwise. reservoir pressure in psi, and pwf is the bottomhole
In Fig. 13-1, the knowledge base uses a simplified flowing pressure in psi.
method to estimate the theoretical reservoir flow (i.e.,
Input reservoir Calculate Compare

No well data q theoretical q actual versus
q theoretical
PTA
Start available Rule of thumb: evaluate
if q actual < 75% q theoretical
Evaluate
Yes stimulation
No
Sandstone Carbonate
Lithology Yes
End
Proceed
No No Yes Mechanical No to formation
Skin > 0 Skin > –4 (problems exist) damage
advisor
Consider
Yes evaluating Yes
fracture potential
Additional
diagnosis
necessary
Figure 13-1. Candidate selection advisor (CSA) flowchart.

veys are required to measure the two pressures accu-

13B. Candidate selection field case history
rately. These can also provide data on near-wellbore
Schiable et al. (1986) reported the identification and suc- damage, as discussed in the following. An ideal PI
cessful removal of severe damage in a newly completed oil
well. can be calculated from reservoir data by using the
Well logs and completion records documented that a uni- equations described in Chapter 1. The ratio of the
form sandstone with 71 ft of net pay was completed using an
inside-casing gravel pack. All wells on the platform were pro- actual PI to the ideal PI can be used as an indicator
ducing from a similar zone. of well performance.
Mapping indicated that the wells on the platform completed
in this formation had similar production with the exception of Well tests are required to quantify permeability and
the new well. A knowledge-base expert system indicated that current reservoir pressure, but usually they are not
the well was producing at less than 5% of its theoretical flow
rate and should have a skin effect of more than 200 on the
available. When this is the case, systems analysis can
basis of the following equations. Assumptions were made for be performed to match current production or bottom-
k from previous well test results on offset wells, and the cal- hole flowing pressure and calibrate the reservoir net
culation did not consider tubing or completion pressure drops
(i.e., systems analysis). pay, permeability, average reservoir pressure, skin
The following equations incorporate assumptions for cer- effect, etc., for an underproducer. The key is to cali-
tain variables.
For oil: brate the system to forecast production on the basis
7.082 × 10 −3 kh ( p r − pwf ) of various stimulation scenarios.
q= , (13B-1)
µ(7.5 + s )
where k is the permeability in md, h is the net pay in ft, pr

is the reservoir pressure in psi, pwf is the flowing pressure in
13-2.2. Impact of formation damage
psi, µ is the viscosity in cp, Bo is the formation volume factor on productivity
for oil (default value of 1), ln(re /rw) is the ratio of the radial
distance to the external reservoir boundary to the radius of Matrix stimulation is successful because the near-
the wellbore (estimated value of 7.5), and s is the skin effect.
For gas:
wellbore region controls well productivity. Damage
in this area can significantly decrease production by
5.39 × 10 −2 kh ( p r2 − pwf2 )
q= , (13B-2) restricting flow in the formation (Krueger, 1986, 1988).
(7.5 + s )
A knowledge of the inflow relationship and Haw-
where the variables are as described for oil, with default val-
ues of viscosity µ = 0.02 cp, standard pressure psc = 14.5
kin’s equation is essential to understand the effects of
psi, standard temperature Tsc = 487.3°R and flowing temper- near-wellbore formation damage on well production.
ature Tf = 620°R. The steady-state equation for an oil well is:
On the basis of this information, a well test was performed
to determine the permeability and skin effect. As suspected,
the pressure transient test indicated that the well had a signifi-
q=
(
kh pe − pwf ) , (13-2)
cant skin effect of 209 and permeability of 526 md. The time
 r 
to pseudosteady state was 9 hr. Production system analysis
141.2 Bµ ln e + s
indicated that the well could increase from 1200 to 5000
BOPD at a skin effect of 20 and 1450-psi wellhead pressure.
 rw 
In addition, the well was tubing limited, and the 34,479-BOPD
rate predicted by the knowledge-base expert system was not where k is the permeability in md, h is the reservoir
possible through the 31⁄2-in. tubing. Because all wells completed thickness in ft, pe is the constant outer reservoir pres-
in this zone were unconsolidated and gravel packed, a critical
computer-aided analysis of drawdown was not performed.
sure in psi, pwf is the bottomhole flowing pressure in
Other considerations were that the well was located in an psi, B is the formation volume factor in RB/STB, µ is
easily accessible area with good chemical supplies, availability the viscosity in cp, re is the drainage radius in ft, rw is
of service companies and no environmental hazards.
A preliminary economic evaluation based on an estimate the wellbore radius in ft, and s is the skin effect.
for the entire treatment and a 90% reduction in skin effect The total skin effect s is a dimensionless term used
yielded excellent economics. Assuming a net gain of 3800
BOPD at $15 per barrel net yielded $57,000 revenue per to account for the additional pressure drop in the well-
day. Payout for the treatment and associated costs would be bore area that results from formation damage and
less than a week, with an NPV of several million dollars in 3
to 4 months. other factors (e.g., inadequate perforations, partial
completion). Skin effect values, determined from
PTA, typically range from 0 to more than 100. Skin
The variable PI/h takes into account the effect of effect is positive if an additional pressure drop is pre-
drawdown on the production rate. The key parameters sent and negative if the actual pwf is lower than the
required (and not always well known) are — p r and pwf. ideal pwf. For example, natural fissures in a reservoir
Although q is not always accurately known, it is usu- or a deviated well contribute a negative skin effect to
ally an accessible number. Pressure testing and sur- the total skin effect.

For a given well, Eq. 13-2 can be simplified to Figure 13-2 shows the productivity index ratio
q 7.5 + so PIs /PI versus the permeability ratio ks /k, where the
= , (13-3) subscript s indicates damage. PI and k represent the
qo 7.5 + s
nondamaged (ideal) well productivity and reservoir
where q and qo are the final and initial flow rates, permeability, respectively (Muskat, 1949).
respectively, in BOPD, and s and so are the final and PIs /PI is equal to qs /q for a constant-pressure draw-
initial skin effects, respectively. down. If all damage is removed (i.e., the natural per-
A realistic value for ln re /rw is 7.5 because this term meability is restored), the well will produce at its
is relatively insensitive to the actual values of re and rw natural flow capacity. A 90% permeability reduction
for oilfield conditions. Thus, for a well with total and extending radially 0.25 and 1.0 ft from the wellbore
damage skin effects equal to 100, a reduction to 10 can can decrease production by 35% and 50%, respective-
exhibit a sixfold increase in productivity. A skin effect ly (Fig. 13-2). However, a 90% permeability increase
reduction from 100 to 0 leads to a 14-fold increase in that extends less than 2 ft has little impact on produc-
productivity. Although reduction to 0 is usually unreal- tivity. Unlike in carbonates, in sandstones it is difficult
istic in a sandstone, reduction to values less than 10 is to increase the permeability above the natural state
a reasonable expectation. The actual value would have because of reaction kinetics limitations, reaction stoi-
to be quantified using a valid systems analysis to con- chiometry and economics.
sider tubing flow, etc. In the same well, a reduction of At this point in the candidate selection process, the
skin effect from 10 to 0 would yield a 2.3-fold increase engineer knows how the productivity of the candidate
in production. Chapter 1 provides the productivity well compares with what it would be if the well were
equations for gas wells and horizontal wells. undamaged and if artificial lift, tubing restriction, etc.,
Hawkin’s equation relates permeability and the are not the problem. An estimate of the skin effect
thickness of the damaged zone to the skin effect for may be available from production system analysis, or
vertical wells: the skin effect may have been determined directly by
k  r PTA. If the skin effect is greater than zero, there is
s =  − 1 ln s , (13-4) potential benefit from matrix stimulation.
 ks  rw
where ks is the damaged permeability in md and rs is
the damage penetration beyond the wellbore in ft, and
100
Percent of original flow
for horizontal wells:

0.75
(PIs ⁄ PI × 100)
 kH kV  r rs =
0.25 rs =
s= − 1 ln s , (13-5) 10 rs =
2
 kHs kVs  rw 5
rs =
where kH is the horizontal permeability, kV is the verti-
cal permeability, kHs is the horizontal damaged perme- 1
ability, and kVs is the vertical damaged permeability, 1 10 100 1000
Percent of original natural permeability
with all the permeabilities in md. (ks/k × 100)
Hawkin’s equation can be used to determine the
skin effect when assumptions are made for the dam- Figure 13-2. Matrix productivity improvement.
age radius and permeability. These variables cannot
be absolutely quantified, but in combination with well
analysis data and/or bottomhole treating pressure
response they may indicate trends and define limits. 13-2.3. Preliminary economic evaluation
Combining Eqs. 13-2 and 13-4 yields
Having established the production potential of the
ks r
log e well as a function of the skin effect and what it is
PIs k rw
= . (13-6) actually producing, the engineer can evaluate the eco-
PI log rs + ks + log re nomic value of improved production and the required
rw k rs investment in well work. Economic evaluation
requires a good production forecast for the current

well condition and a forecast for the stimulated well, 13-3. Formation damage
possibly as a function of how successful the stimula-
tion is (e.g., post-treatment skin effect). A brief dis- characterization
cussion of the methodology follows. Damage characterization is the next step in the matrix
The preliminary economic evaluation requires a stimulation engineering process. It is an essential task
model for production (revenue) and job cost (expense) prior to treating fluid selection and treatment design.
as a function of skin effect. The engineer must deter- Damage is characterized using laboratory tests, logging
mine the effect of the treatment on well economics on techniques and well history. Detailed study is necessary
the basis of the net gain in cumulative oil or gas over to develop a list of suspected damages from the avail-
a given period of time and the cost associated with the able data. Multiple types of damage are normally
treatment. The production prediction model should suspected and are all considered when designing the
include material balance, allow analysis under different treatment. Tables 14-3 through 14-5 list damage types,
outer boundary conditions (infinite acting, no flow and diagnostic clues and remedial recommendations for
constant pressure) and forecast on the basis of constant comparison to the condition and characteristics of the
sandface (bottomhole pressure or constant wellhead candidate well.
pressure). The sensitivity of economic performance to The following sections focus on problems that com-
skin effect is quantified with this model. The range of monly occur in the field and discuss how expert sys-
skin effects input should be based on the analysis of tems can guide an engineer through the damage
current (presumably damaged) performance and characterization process. Expert and advisor systems
expected post-stimulation behavior. Post-stimulation have proved especially useful in the damage charac-
skin effect expectations and job costs can be based on terization process (Krueger, 1988).
field experience or analysis from a numerical stimula- All available information on the well such as well
tion simulator. The numerical stimulation simulator can logs and records, reservoir characteristics and infor-
be used to evaluate potential skin effect removal as a mation on the completion and previous workovers
function of job design (e.g., volume pumped). should be collected. Samples of produced fluids and
The economics portion of the model must enable the any solid materials from the well should be analyzed.
evaluation of cost versus expected revenue. Estimated In some cases, it may be useful to conduct laboratory
costs should include treatment cost, operation costs, compatibility tests of the completion or drilling fluids
cost of money, applicable taxes and any auxiliary costs. and the formation fluid or rocks. Such tests can help
Produced product price, price adjustment factors and in developing understanding of the problem in the
treatment chance (risk) factors should be included to current well and lead to corrective action.
compute various financial indicators such as the rate of Chemical analysis of solid and liquid samples
return (ROR), return on investment (ROI), NPV, payout retrieved from the well can provide valuable insights
(days) for investment only, payout (days) for invest- into damage mechanisms and characteristics. Field test-
ment plus interest and unit cost (investment divided ing can quantify certain species (e.g., carbonates and
by incremental production). oxides are soluble in hydrochloric acid [HCl], paraffins
Several important decisions can be made at this and asphaltenes float in water, an oil-external-phase
point. For example, if the skin effect required to pro- emulsion disperses in diesel). Water analyses are com-
vide an acceptable return is greater than 0 in a sand- monly performed to assist the determination of scaling
stone or –4 in a carbonate, then matrix stimulation is potential. The ionic composition of one or two samples
practical; i.e., fracture stimulation is not required. can be used to calculate potential scales based on a
In conclusion, from a knowledge of the reservoir minimum Gibb’s free energy or a Stiff-Davis calcula-
and well history coupled with the use of diagnostic tion. In addition, gamma logs can be run to comple-
tools, the engineer should select candidate wells with ment the water analyses and detect barium and
low risk and significant potential for economic return. strontium sulfate scale buildups. Because of trace
This requires the evaluation of numerous wells and is amounts of the naturally occurring radioactive materials
an important link to matrix stimulation success. (NORM) radon and radium, barium and strontium sul-
fate deposits may exhibit an abnormally high gamma
count.

The history of the well should be evaluated in detail Production logs can be run to quantify the flow per
to assist damage characterization. All phases of well layer. This information is input into computer pro-
operations—including drilling, cementing, perforating, grams to determine the formation damage skin effect
gravel packing, workover operations, stimulation per layer. The use of a production log is reviewed in
operations and production or injection—are potential the field case history in Sidebar 13D.
originators of damage (Krueger, 1986). Damage characterization is the basis of treatment
Expert and advisor systems have been developed design. Chemicals will be selected to remove the sus-
to assist damage characterization (Bartko et al., 1996). pect cause or causes of damage. Treating procedures
They provide a guide through a series of questions will be designed to access the suspected damage with
directed at relating the type of damage to the well adequate amounts of chemicals to remove enough
behavior and available data. They can pinpoint key damage to achieve the well productivity goals. Correct
gaps in the available data and recommend additional damage characterization is critical to matrix stimula-
required diagnostics. A brief discussion of the most tion success.
common questions the engineer should ask in using
an advisor are in Sidebar 13C.
13C. Formation damage characterization Well logging and testing

field case history • Are the results of a pressure transient test available?
High skin effect values indicate there is severe formation
Schiable et al. (1986) performed a detailed laboratory and damage.
field study to characterize formation damage. Some of the
• Does a production log show nonuniform production rates?
questions and answers resulting from the study follow.
Although the porosity logs indicate that the pay zone is a
Well history homogeneous sandstone, it is not damaged uniformly.
• Is the well newly completed, an old well, huff and puff, • Is there fill, a reduction of tubing inner diameter or an
a recent workover and/or a recent stimulation treatment? increase in gamma count?
This is a newly completed gravel-packed well that has There is no indication of any of these.
been tested. • Is there water production from a specific zone?
• Is it an oil, gas or water producer or injector with water or There is no water production.
hydrogen sulfide (H2S) production?
This oil well produces 29.5°API gravity crude. Laboratory and field testing
• What is the mineralogy of the pay zone?
• Was the well perforated under- or overbalance?
Laboratory core testing indicates that the formation is
The well was perforated underbalance at 12 shots per foot
composed of approximately 75% quartz, 10% feldspar and
(spf).
10% clay with kaolinite, illite, chlorite and illite-smectite
• What type of completion fluid was used? present with a 4% to 5% HCl solubility.
Filtered calcium chloride and potassium chloride comple-
• Was a water analysis performed?
tion fluids were used to provide temporary clay control.
No, because the well did not produce water.
• Was the well produced at high drawdown rates?
• What is the composition of samples from the well?
Although newly completed, the well was produced at high
No samples other than cores and oil were retrieved from
drawdown rates during the well test.
the well.
• Has the well responded to previous treatments?
• Is the salinity of the completion fluid lower than that of the
No previous treatments were performed on this well, but
formation water?
offset wells have responded to mud acid treatments.
Not applicable
• Did the production decline slowly or rapidly?
• Does the oil have an emulsion or sludging tendency?
Not applicable
Laboratory testing indicates that the crude will not form an
• Did the well respond positively to a pump-in test? emulsion or sludge.
A pump-in test was not performed because of the associ-
• Do core tests using the completion or stimulation fluids
ated cost and availability of other data (see “Laboratory
indicate damage?
and field testing” section).
Laboratory testing indicates that the 2% KCl fluid used
• Was there excessive drilling mud loss to the pay zone? during gravel packing created a 62% reduction in perme-
No unusual loss of drilling mud occurred when drilling ability to oil (101 md damaged to 45 md).
through the pay zone.
• Do core flow tests using stimulation fluids improve the
• Is there fill in the wellbore, produced solids or solids in sur- permeability?
face equipment? (Obviously, this would be a major prob- Core testing using fluoboric acid stimulation fluid indicates
lem indicative of gravel-pack failure.) that the formation damage can be removed, with 280% of
There is no water production to create scale or solids. the initial permeability to oil obtained.

tem had been previously used successfully in the forma-

13D. Fluid and additive selection field case history tion, laboratory core tests were performed to determine
the damage removal efficiency of the system. A 101-md
The well was severely damaged by the completion fluid (clay core was damaged with completion fluid by flowing 20
swelling in mixed-layer clays) and had possible fines migra- pore volumes of water containing 2% KCl to yield 45.6-md
tion during the well test. Production could potentially increase permeability to oil. The core was then treated with a 7.5%
from 1200 to 5000 BOPD. Schiable et al. (1986) reported the HCl preflush followed by 12% HCl–3% HF mud acid over-
laboratory testing performed to quantify the fluid and additive flushed with fluoboric acid. The fluoboric acid was used to
selection. decrease silt and clay migration. The resulting final perme-
ability to oil was 284 md (>500% increase). On the basis
• Mineralogy of laboratory results, the mud acid–fluoboric acid system
X-ray diffraction and scanning electron microscopy of for- was selected to restore permeability in the damaged well.
mation cores indicated that the formation contained 8% to • Treating and reservoir fluid compatibility
11% clay and 8% to 12% feldspar. The clay was com-
posed of 7% to 31% kaolinite, 9% to 18% chlorite, 37% The 29.5°API gravity crude did not exhibit sludging or
to 46% illite and 22% to 29% mixed-layer clays. emulsion tendencies in laboratory testing.
• Petrophysics • Damage and permeability profiles
The 71-ft zone was not fissured and consisted of a rela- A log recorded by the PLT* Production Logging Tool indi-
tively homogeneous sandstone with approximately 29% cated that the homogeneous sand was not damaged uni-
porosity and 526-md permeability. formly. The skin effect per layer based on the production
The formation was borderline, between the first two per layer and completion are listed in Table 13D-1.
classes of the mud acid fluid selection guide in Table 18-7. An oil-soluble diverter was selected to provide uniform
Assuming that the sandstone was less than 10% silt and flow per layer (i.e., uniform volume of each fluid per foot of
less than 10% clay, a 12% HCl–3% HF system would be zone). Other additives to the acids were an iron control agent,
used, whereas a sandstone with high clay (greater than acid corrosion inhibitor and surfactant. An ammonium chloride
10%) and low silt (less than 10%) would require a 12% flush with a mutual solvent ahead of the HCl preflush was rec-
HCl–2% HF system. Because the 12% HCl–3% HF sys- ommended to remove oil from the wellbore area.
Table 13D-1. Skin effect per layer based on production per layer and completion.
Layer Skin Effect
Development and Perforation Damage Total Production Porosity

Partial Completion and Gravel Pack (%) (%)
1 0 0.1 48.5 48.6 50 29

2 0 0.1 277.4 277.5 10 29
3 0 0.1 75.0 75.1 35 29
4 11.5 0.3 2534.2 2546.0 5 29
13-4. Stimulation technique effect. The impact of skin effect on the economic limit
and reserve recovery must also be considered.
determination Figure 13-3 is a decision tree to help the process
At this point, the well has been identified as an under- of candidate selection and stimulation technique. This
performer. The monetary value of improving well type of flowchart can be incorporated into computer
productivity and the possible cause or causes of for- software. As shown in the figure, the productivity tar-
mation damage have been determined. Next, the engi- get dictates the stimulation technique.
neer must determine the remedial action. If the productivity target can be reached with a skin
The entire production system must be considered effect of 10% of the original damage skin effect in
in making this decision. If the problem is in the well sandstones and –2 to –3 in carbonates, matrix stimula-
design or operation (e.g., tubing size or artificial lift), tion will be adequate and probably cost effective. In
then stimulation is not indicated and the equipment sandstone reservoirs the only stimulation alternative
should be upgraded or repaired. The target well per- is hydraulic fracturing. In carbonate reservoirs (lime-
formance must be balanced; i.e., there is no need to stones or dolomites) acid fracturing can be a cost-
produce more than the tubing or lift will transport or effective way to increase productivity. In both cases,
than the facilities will process. This may affect the a hydraulic fracture is induced in the reservoir. In con-
economics of incremental improvements in skin ventional fracturing, the conductivity of the fracture is

Well is valid candidate for stimulation
Negative or nearly zero skin effect Positive skin effect

yields adequate production yields adequate production
Sandstone or carbonate? Matrix treatment
Sandstone: Carbonate:
Sandstone: Carbonate:
propped matrix treatment
treatment to remove treatment to bypass
fracture propped fracture
damage damage
acid fracture
Mechanical Mechanical
limitations limitations Mechanical
limitations
Economics evaluation Economics evaluation

for Economics evaluation
Disqualified:
evaluate Matrix Propped Acid
matrix treatment treatment fracture fracture
Disqualified: Disqualified: Disqualified:

evaluate evaluate evaluate
propped fracture matrix treatment matrix treatment
acid fracture acid fracture propped fracture
Figure 13-3. Stimulation decision tree.
maintained by propping it with high-permeability 13-5. Treatment design

proppant. In acid fracturing, conductivity is created by
differential (nonuniform) etching of the rock surface. 13-5.1. Matrix stimulation techniques
Other factors may influence selection of the stim-
ulation technique. In unconsolidated or friable sands, Two types of nonfracture treatments are used to
it is advisable to examine the maximum pressure draw- improve production in oil and gas wells. Wellbore
down allowed before formation (sand) production. cleanup uses chemical and/or mechanical methods to
This drawdown limit may lead to the selection of frac- clean the wellbore. For matrix stimulation, fluids are
ture stimulation to allow obtaining target rates at a injected into the formation to treat the near-wellbore
lower drawdown. On the other hand, concern about region.
zonal isolation may preclude fracture stimulation. If • Wellbore cleanup
vertical fracture growth into an aquifer or gas cap can- Wellbore cleanup is commonly used to remove
not be controlled, matrix stimulation may be indicated. scale, paraffin, bacteria or other materials from the
The engineer must also make some decisions about tubing, casing or gravel-pack screen. These treat-
data collection before designing the treatment. As dis- ments normally use acid or solvent systems that are
cussed in the following, many technologies are avail- placed in the wellbore area to soak. Key parameters
able to improve job monitoring and evaluation. in treatment design are the placement technique,
Among the options to be considered are real-time data chemical selection and soak time.
acquisition, use of downhole pressure gauges (mem- Mechanical assemblies such as packers, bridge
ory or surface readout) and use of wireline surveys plugs, spring-loaded “spot control” valves and coiled
and tracers to determine treatment placement. For tubing can be used to ensure proper placement. This
some wireline evaluations, baseline surveys must is critical in minimizing the volume of treating fluid.
be run prior to the treatment.

Density differences between treating fluids and dis- and to allow soluble products or solids to flow out of
placement or well control fluids must also be consid- the well (Smith and Hendrickson, 1965). In the case
ered when designing a treatment that will stay in of carbonate formations, the goal is to bypass the
contact with the damage for the required soak time. damage with acid or dissolve the damage with sol-
The well cannot be assumed to remain static (no vents (Hendrickson et al., 1960). The main treating
cross flow) during the soak time. chemicals fall into the following categories:
Chemicals should be selected on the basis of their – solvents to remove organic deposits (such as
effectiveness at dissolving the suspected damage. paraffin)
Laboratory tests at bottomhole temperature and, if
– oxidizers to remove damage from polymers
possible, pressure are desirable to determine neces-
sary soak times. Because of the small surface area – scale removers to remove sulfate or oxide scales
of the exposed damage, the soak time can be hours – acids to remove carbonate and oxide scales,
to days depending on the type of damage, tempera- break polymer residues or stimulate carbonate
ture and damage removal fluid. Agitation or jetting formations
with coiled tubing can accelerate damage removal. – hydrofluoric acid (HF) to remove aluminosilicate
In low-pressure wells, nitrogen (N2) is recommended damage (primarily clays) from sandstone forma-
to assist the removal of reacted treating fluid. tions.
• Matrix stimulation The main treating fluid is chosen to bypass,
Matrix stimulation treatments injected below frac- dissolve or remove the main damage. If multiple
turing pressure down tubing, drillpipe or coiled damages are suspected, it may be necessary to use
tubing usually include a sequence of several fluids, several main fluids or to combine the listed functions
referred to as stages. A minimal treatment consists into a single fluid; however, in combining fluids and
of a preflush stage with a nondamaging, nonreac- functions, care must be taken to maintain the effec-
tive fluid to establish an injection rate, a stage of tiveness of each and avoid incompatibilities.
the main treating fluid and an overflush stage to Solvents are selected when organic deposits are
clear the main treatment fluid out of the tubing suspected. If possible, the solvent should be evalu-
and displace it into the near-wellbore area. In most ated in the laboratory with samples of the deposit.
treatments, other auxiliary stages are included to Various oxidizers have been reported for use in well
enhance the effectiveness of the treatment. The stimulation. Because the polymer is usually intro-
selection of chemicals for the stages and the design duced into the wellbore during drilling or completion
of the treating sequence (pumping schedule) are operations, its identity is well known. An effective
reviewed in the following sections. oxidizer can be evaluated in the laboratory using this
knowledge. In addition, the effect of other fluid com-
ponents, well materials and formation minerals on
13-5.2. Treatment fluid selection the oxidizer must be assessed. Scale removers can
Treatment fluid selection is an important step in the likewise be evaluated in the laboratory with samples
engineering process. Multiple fluids (fluid systems), of scale deposits. Soak time and chemical concen-
composed of base fluids and additives, are selected tration can be optimized with these tests. One draw-
on the basis of lithology, damage mechanism and well back to these forms of stimulation is that they are
condition. Each fluid in the treating schedule serves a effective against a limited range of damage materi-
special purpose. Although the process of fluid selec- als. If the damage diagnosis is uncertain, a more
tion is complex because many parameters are broad-spectrum treatment may be indicated.
involved (e.g., damage type, mineralogy), guidelines Acid stimulation is performed to remove or
have been developed to simplify the process and bypass a variety of damages. When used to remove
improve treatment success. carbonate scale or polymer residues, acids act simi-
larly to the treating chemicals discussed previously.
• Main treating fluid selection Acids are also used to stimulate the near-wellbore
The main treating fluid is selected to dissolve or dis- region of the reservoir. In carbonates, HCl or
perse the principal damage in sandstone formations organic acids (formic or acetic) are used to etch

conductive paths between the wellbore and the for- sludging and wettability problems. To remove dam-
mation. In sandstones, mixtures of HCl and HF age, the treating fluid must be in intimate contact
(mud acids) are used to remove drilling mud, for- with the damage. This requires a water-wet forma-
mation fines, fines generated during drilling and tion and oil displacement from the pore throats.
perforating residue. These materials are usually Thus, preflushes used in oil wells may include an
more difficult to define or sample than other forms organic solvent or ammonium chloride with surfac-
of damage. tants and/or a mutual solvent to remove heavy
Because acids are effective against several types hydrocarbons from the wellbore area and ensure
of damage and are inexpensive, they are used in the a water-wet environment.
vast majority of matrix stimulations. As a result, The main acid formulation is based on the type
more effort has been expended in the engineering of formation to be stimulated. Formulation guide-
of acid treatments. In the remainder of this chapter, lines are based on studies of acid-mineral reaction
the focus is on engineering matrix acidizing treat- chemistry. A brief review of the main considera-
ments, but the concepts discussed apply to all types tions in acid formulations follows.
of matrix stimulation. • Fundamentals of acid-rock chemistry
• Fluid formulation for matrix acid stimulation Chapter 16 presents a detailed discussion of the
Formulating fluids for matrix acid stimulation chemistry and physics associated with acidizing.
includes selection of the main acid and identification An appreciation of the key points in the chapter is
of the need for preflushes and overflushes. Figure necessary to understand the fluid selection process.
13-4 shows a decision tree for fluid selection in – Acid types and kinetics for carbonates
sandstones and carbonates. Fluid selection depends
HCl is used for carbonate acidizing because it
on the damage type, lithology, mineralogy and type
readily dissolves calcite and dolomite.
of well. It is also based on field and laboratory expe-
The reaction of limestone (calcium carbonate,
rience and can be derived from an expert system
or CaCO3) with HCl is
(Perthuis et al., 1989). Oil wells are more difficult
to treat than gas wells because of potential emulsion, CaCO3 + 2HCl CaCl2 + H2O + CO2
Minimum
permeability test No
Start gas wells k > 1 md Consider fracture
oil wells k > 10 md or other treatment
No: mud acid treament

Yes
Carbonate Sandstone
Lithology Calcite
≥ 20% HCI preflush: see Table 18-6
HCI/HF: see Table 18-7
Yes HCI postflush: same as preflush
NH4Cl overflush
Perforated Perforated
interval interval
Fines migration problem exists
No Chlorite > 0
Glauconite > 0
Small problem: Severe problem:
Add clay control agent Use fluoboric acid,
Bottomhole in preflush and preflush and
temperature Yes
overflush fluids overflush
300°F < T
≤ 300°F ≤ 400°F > 400°F Bottomhole
temperature
15% HCI or 15% HCI or 10% HCI or
Minimum permeability test
28% HCI HCI-acetic HCI-acetic
≤ 200°F > 200°F
All volume recommendations
10% HCI 10% acetic
Figure 13-4. Fluid selection flowchart.

For dolomite, the reaction becomes wormholes are not created in sandstones. In a
CaMg(CO3)2 + 4HCl typical sandstone, 80% of the radial flow can be
CaCl2 + MgCl2 + 2H2O + 2CO2 through 20% of the pores (the larger pores). The
damage is removed from the larger pores, result-
Normally, formation damage is not dissolved
ing in a skin effect reduction but possibly not
but rather bypassed to yield new flow channels
restoration of the natural permeability of the
(wormholes) and/or an etched surface on fissures,
formation.
resulting in a reduced pressure drop (decrease in
The primary reaction of HF with quartz grains
skin effect). The wormhole pattern is created
(pure silica) is expressed in the two following
because the highly reactive acid enters the largest
equilibria:
pore throats, vugs and/or fissures and essentially
enlarges them. The number of wormholes is a SiO2 + 4HF SiF4 + 2H2O
function of the pore-size distribution (Schechter SiF4 + 2F – SiF62–
and Gidley, 1969). In the latter, silicon tetrafluoride combines with
Acid reaction with carbonate reservoirs is gov- fluoride to produce silicon hexafluoride.
erned by three mechanisms: wormholing, compact The conventional mud acid formulations (12%
dissolution and radial flow. Each mode occurs HCl–3% HF and 6% HCl–1% HF) used in
under certain conditions. For example, at low matrix acidizing dissolve little quartz. The intent
injection rates, compact dissolution occurs when of sandstone matrix acidizing is to dissolve clays
the formation face is dissolved to enlarge the well- and other damaging materials, leaving the sand-
bore. If the flow rate is increased to where the stone matrix undisturbed.
Peclet number (a function of injection rate, acid Although the primary reaction of HF with
concentration and diffusion rate) is approximately silica is simple, the reaction with silt and clay
1, then wormholing initiates. If the rate is increased is more complex.
significantly, radial flow dominates in a manner For aluminosilicates (clays and feldspars), the
similar to sandstones. Wormholing probably primary reaction is
occurs in most treatments owing to the heteroge-
MzAlxSiyO(z/2 + 3x/2 + 2y) + 6(x + y)HF →
neous nature of carbonates; i.e., normally there
xAlF63– + yH2SiF6 + (z/2 + 3x/2 + 2y)H2O +
are thief pores where wormholing initiates.
(3x – z)H+ + zM+
In cases where the temperature exceeds 400°F
[205°C] (where corrosion inhibitors are ineffec- where M is a metal atom (e.g., Na or K).
tive in HCl), organic acids (acetic or formic acid) The reaction products AlF63– and H2SiF6 con-
are used. tinue to react with the aluminosilicates to form
The reaction of HCl with calcite is diffusion silica gel on the surface of the clay (Crowe, 1986).
limited (i.e., mass-transport limited). The limiting An example of such a secondary reaction is
step in the dissolution reaction is the diffusion of 2yH2O + (x + z)H+ + x/3H2SiF6 +
acid (hydronium ion, H3O+) to the surface of the MzAlxSiyO(z/2 + 3x/2 + 2y) →
calcite. Once the H3O+ contacts the surface of (y + x/3)Si(OH)4 + xAlF2+ + zM+ +
the calcite, the reaction occurs very fast. The (x/6 + z/2)H2O
reaction of HCl with dolomite is diffusion limited Theoretically, other secondary reactions can
at temperatures greater than 150°F [65°C] and produce aluminum fluorosilicate solids, which
limited by the surface reaction rate (i.e., diffusion may be damaging. Thus, there is a potential for
is not the slowest step) below this temperature. the generation of damaging reaction products,
Thus, because of the diffusion-limited kinetics although the effects are normally minimal.
of HCl on calcite, wormholes can normally be Although HCl accelerates the reaction of HF
easily formed through the damaged zone (2 to with siliceous minerals (silt and clay) the reac-
3 ft radially). tion is far slower compared with that of HCl
– Acid types and kinetics for sandstones reacting with calcite. The reaction is reaction rate
Mud acid mixtures of HF and HCl are used for limited (i.e., numerous collisions of F– with the
sandstone acidizing. Unlike carbonate acidizing, surface of the mineral must occur before a reac-

tion occurs) and yet results in most of the HF In addition, acids selected for carbonate acidiz-
reacting within the first foot of the wellbore. This ing must be compatible with the formation fluids.
occurs because of the large surface area of the Ensuring oil-acid compatibility is discussed later
silt and clay. (Five grams of smectite clay has the and in Chapter 15.
surface area of a basketball court.) Retarded – Sandstones
acids have been developed to yield deeper live
The sensitivity of a sandstone formation to a
acid penetration (Thomas and Crowe, 1981).
matrix treatment fluid depends on the mineralogy
• Acid and formation compatibility of the formation, damage type, reaction products,
– Iron hydroxide temperature and permeability. As a rule of thumb,
Precipitation of iron (ferric, Fe3+) hydroxide the more silt, feldspar and clay and the lower the
(Fe(OH)3) can occur during sandstone and carbon- permeability, the greater the sensitivity of the for-
ate acidizing when the pH value is greater than mation. The amount of precipitate formed is pro-
2.2. Sources of iron in the formation and tubulars portional to the amount of rapidly HF-soluble
are chlorite, siderite, hematite and rust. Ferrous material present, and the impact of the precipitate
iron (Fe2+) is not a problem in sandstone acidizing on plugging is greater at lower permeability. In
because ferrous hydroxide does not precipitate addition, fines migration and the formation of pre-
until the pH value exceeds 7.7. Although fluoride cipitates during a treatment create more problems
complexation of iron increases the pH required in low-permeability formations. Thus, sensitivity
for precipitation, iron control agents are recom- is related to chemical and mechanical problems
mended in all acid stages during sandstone and in sandstones.
carbonate acidizing (Crowe, 1985). Although sensitivity cannot be eliminated com-
Iron also increases the tendency of oil to form pletely, the goal is to minimize it by selecting flu-
rigid films and emulsions with acids. These prob- ids with the greatest possible compatibility with
lems can result in formation damage and facility the formation. The mineralogy of the formation
upsets when producing wells after acidizing. plays a large role in formation sensitivity and fluid
Live acid or partially to completely spent acid selection. The solubility of the minerals in a sand-
occasionally forms an emulsion with crude oil stone depends on their location. Authigenic clays
or condensate, resulting in two-phase flow and are pore lining and come into contact with the
a production decrease. In addition, sludging can treating fluid, whereas detrital clays were depos-
occur because of the coagulation of asphaltenes ited with the original sand and are not completely
plugging pore throats. Both emulsions and sludg- exposed to the treating fluids. A discussion of the
ing can be avoided by adding the appropriate most important reactive minerals in a sandstone
chemical determined from laboratory testing. In follows.
severe cases, organic acids replace mineral acids The solubility of the formation in HCl is nor-
to decrease emulsion and sludging tendencies. mally considered to represent the carbonate con-
Preflushes of solvents or potassium or ammon- tent of the sandstone. This assumption is incorrect
ium chloride with mutual solvents are used to where other acid-soluble species are present (e.g.,
separate the crude oil from the acidic fluids. chlorite, zeolites, oxides and sulfides). If the acid
solubility is greater than 20%, HF is not recom-
– Carbonates
mended. Excessive carbonate can react with the
Formulating acids for carbonates is relatively HF to precipitate calcium fluoride, and spent HF
simple because the majority of the reaction prod- can react to form calcium hexafluosilicate.
ucts is simply calcium chloride, water and car- Chlorite clay and some zeolites (hydrous calcium/
bon dioxide, which do not undergo further sodium/potassium aluminum silicates) are partially
reactions. However, there are numerous minerals soluble in HCl and can cause severe plugging
(silt and clay) in most carbonates that are insolu- from the migration of their residue. Special pre-
ble in HCl. The insoluble material normally cautions must be taken when treating formations
flows back through large wormholes or fissures with these minerals (see Chapter 18). Thus, in
and is not a problem. sandstone acidizing, an HCl (or organic acid)
preflush is used ahead of the mud acid.

The silt and clay content can be determined precipitates. However, adequate HCl must be
in the laboratory by dissolving a core sample in present to account for this process because it
mud acid. The difference between the mud acid consumes acid that may result in the precipita-
and HCl solubilities is the approximate silt and tion of other species.
clay content. Although it is important to know • Acid selection guidelines for carbonates
the amount and type of silt, feldspar and clay
– Temperature—Acid selection depends greatly on
present, it may be equally important to know
temperature because of corrosion inhibition limita-
their position in the formation.
tions. For example, at temperatures greater than
Secondary reactions can occur to form precipi-
300°F [150°C], 28% HCl cannot be used because
tates in the sandstone if precautions are not taken
of excessive corrosion whereas 15% HCl can be
(see the “Design types” discussion in Section
used. Above 400°F, organic acid or fluids con-
13-5.3). Potential precipitates are as follows.
taining ethylenediaminetetraacetic acid (EDTA,
Fluosilicates (SiF62–) are the most detrimental
a non-acid-reactive solvent) can be used. In addi-
precipitates produced in sandstone matrix acid-
tion, emulsified acid or HCl with a retarder is used
izing. Sodium, potassium and calcium fluosili-
to increase penetration.
cates (Na2SiF62–, K2SiF62– and CaSiF62–,
respectively) are extremely insoluble and form – Mineralogy—Calcite and dolomite react differ-
when formation water or any sodium/potassium/ ently with acids. Acid formulation should be
calcium brine contacts spent mud acid. They are optimized to achieve the desired reaction and
avoided by preflushing with a compatible fluid penetration characteristics in each formation.
such as ammonium chloride or HCl and then Carbonates commonly contain insoluble minerals
overflushing. (silt and clay) that are released during acidizing.
Calcium fluoride forms from the reaction of When this occurs, a silt-suspending agent is
HF with calcite if adequate HCl (or organic acid) required along with N2 in wells where spent acid
is not injected ahead of the mud acid. It is not will not flow back naturally (less than 0.46 psi/ft
necessary to dissolve all the calcite. As reported pressure gradient).
by Walsh et al. (1982), as much as 6% calcite – Petrophysics—Acid penetration and the amount
can remain and not cause the precipitation of of damage depend on the type and distribution
calcium fluoride or aluminum fluoride. of porosity. For example, a highly vugular, high-
Precipitation of the aluminum salts aluminum permeability or fissured carbonate usually is
fluoride (AlF3) and aluminum hydroxide damaged severely during drilling and completion
(Al(OH)3) (Walsh et al., 1982) can be avoided because of solids. Attempts to obtain deep pene-
by maintaining a low pH (less than 2). In addi- tration (greater than 5 ft) may be difficult, with
tion, the tertiary reaction of aluminum bifluoride short but wide wormholes formed. An emulsi-
cations with silt and clay can occur very slowly fied, retarded or foamed acid is required to
at temperatures greater than 200°F [95°C] or achieve deep penetration. Because much of the
with retarded mud acid formulations containing porosity (hydrocarbons in place) is located in
AlCl3. The addition of more Al to the mud acid the vugular or fissure network, it is essential to
formulation results in Al saturation and precipita- remove the damage and restore communication
tion. This process consumes acid, resulting in a with the wellbore. As stated previously, a silt-
pH increase and the precipitation of aluminum suspending agent in a gelled or nongelled acid
silicate and/or fluoride species. Citric acetic acid and N2 are recommended to remove drilling mud
has been used successfully in treatments to com- and completion solids.
plex aluminum and buffer the pH below 2, thus • Acid selection guidelines for sandstones
avoiding precipitation (Rogers et al., 1998). – Temperature—Acid selection depends only
Hydrated silica precipitate forms in all mud slightly on temperature because lower acid con-
acid treatments but is normally not a problem centrations are normally used (i.e., corrosion is
because it occurs during the topochemical reac- not a major issue). The sensitivity of a sandstone
tion of spent mud acid with the surface of the silt formation increases with temperature because of
and clay. The precipitate is not mobile like other

the rapid spending of HCl and mud acid and near-wellbore region and leave the minerals and
generation of mobile fines in the spent acid; damage in a water-wet condition. This enhances
however, the sensitivity resulting from precipita- the rate of acid attack. A solvent or solutions con-
tion reaction products decreases with temperature taining mutual solvents can be injected for this pur-
because of their increased solubility. These pose. Gidley et al. (1996) advocated using carbon
processes tend to counterbalance one another, dioxide as a preflush in oil wells to remove oil and
resulting in little temperature dependence. increase acid effectiveness. When acidizing with
– Mineralogy—Sandstone mineralogy is the most mud acid, an acid (HCl or organic acid) preflush is
important factor to consider in fluid selection used to remove calcium carbonate and iron carbon-
because of the potential formation of precipitates. ate or oxide from the near-wellbore region. This
Because of the large surface area of silts and helps to eliminate calcium fluoride– and iron-relat-
clays, most mud acid reacts rapidly to generate ed impairment problems. Once injection is accept-
a complex mixture of compounds. Numerous able and the formation is in the proper condition,
guidelines have been developed from field expe- the main treating fluid can be injected. The main
rience and laboratory data (Fogler et al., 1976; treating fluid normally contains HF and is displaced
Walsh et al., 1982; McLeod, 1984; Ayorinde et into the reservoir by an overflush fluid.
al., 1992; Gdanski, 1996), and their use has • Diverting agent selection
improved field results. In addition, acid penetra- Nonuniform damage and permeability heterogene-
tion is highly dependent on mineralogy because ity in targeted zones cause treating fluids to prefer-
of the large surface area of silts and clays. entially enter thief zones. The result is nonuniform
– Petrophysics—As in carbonates, acid penetration damage removal and economic failure in most
and the amount of damage depend on the type cases. Chapter 19 addresses treating fluid place-
and distribution of porosity. Fissured sandstones ment (diversion) in detail. Numerical simulators
are not common in most matrix candidates but can model this process and assist treatment opti-
can be present in basement formations and geo- mization (Bartko et al., 1996). The heterogeneity
thermal areas. A silt-suspending agent in acid of pay zones requires that the diversion and treating
and N2 are recommended to remove drilling mud fluids must be compatible with the chemical sys-
and completion solids. Although HCl can be tems used in the process.
used exclusively in this operation, mud acid is Diverters are used to help control the distribution
recommended to assist damage removal and dis- of acid in the wellbore. The four general types are
persion. bridging agents (60 mm), particulates (4 to 60 mm),
The sensitivity of a sandstone depends on the viscous solutions (gels) and foams. Bridging agents
permeability of the formation. Because permea- are relatively large particles that are used to bridge
bility is a function of pore size, low-permeability at the face of fissures in carbonate formations; partic-
sandstones are more sensitive than high-perme- ulates are smaller particles that are used in sand-
ability sandstones for a given mineralogy. Thus, stones. For particulates, size and composition are
the permeability must be considered in fluid selec- the key parameters (Crowe and Cryar, 1975). Parti-
tion. Guidelines for mud acid selection and HCl culates should be sized to form an external filter
selection are in Chapter 18 (Ayorinde et al., 1992). cake on the formation and should be soluble in the
Although they provide useful general guidance, produced fluids (King and Hollingsworth, 1979).
acid formulations should be optimized on the Viscosifiers should be stable enough to provide
basis of a detailed formation evaluation (Davies significant resistance to flow, but they must degrade
et al., 1992; Nitters and Hagelaars, 1990). fast and completely so that they do not impede pro-
• Preflush fluid selection duction. Viscosifiers are normally used in carbonate
formations. Foams are made by adding a foaming
The preflush has three important functions. A pre-
surfactant to the acid or brine and mixing it with N2.
flush of a nonreactive fluid is pumped initially to
The foaming material should be stable in the fluid
ensure that injection can occur at an acceptable rate
it is mixed in, and sufficient N2 should be available
and pressure. In some oil wells, it is advisable to
to ensure a foam with a 65% to 70% (gas volume)
inject a preflush formulated to remove oil from the
quality downhole (Thomas et al., 1998).

All types of diverters can be injected continu- mixed with the treating fluid to modify a property
ously with the acid or staged in discrete segments (e.g., corrosion, precipitation, emulsification, sludg-
of the injection sequence. The concentration can ing, scaling, fines migration, clay swelling tendency,
also be varied as required. Experience can provide surface tension, flow per layer, friction pressure).
guidance as to which approach works best, or a The brief discussion that follows emphasizes the
numerical model can be used to optimize the diver- optimization of treatment results. Most additives do
sion process. not depend on the formation type but rather on the
• Postflush fluid selection reservoir conditions (i.e., hydrocarbon type and tem-
perature).
A postflush is almost always used to remove the
All additives should be tested for compatibility
reactive (and corrosive) fluid from the tubing and
to ensure that they are chemically compatible with
maximize the contact of main fluid with the near-
the other additives used in a particular fluid stage.
wellbore area. The decision to inject a postflush fluid
In addition, stages that will be in contact with each
depends on the type of stimulation. If the dissolved
other in the wellbore should be tested for compati-
or dispersed damage has the potential to damage the
bility. Compatible spacer stages can be used if
formation if displaced radially into the formation,
incompatible fluids must be pumped, but eliminat-
injection into the reservoir must be avoided (e.g.,
ing incompatibilities is preferred if possible.
paraffin solvent, mud/silt dispersant fluids would not
Additive types are as follows:
be overflushed). At the other extreme is sandstone
HF acidizing in which the overflush is essential to – Acid corrosion inhibitors—Different corrosion
ensure that secondary precipitation occurs deep in inhibitors are required for all inorganic and
the reservoir, where the impact on productivity is organic acids. In selecting an inhibitor and con-
greatly reduced. Damage in the near-wellbore area centration level, it is important to realize that
can significantly decrease production by restricting partially spent HF may still be highly corrosive.
flow in the formation; however, a 6-in. collar with Recommended practices for testing acid corrosion
80% damage will reduce production by only 10% inhibitors and methods for handling returned acid
from ideal if located more than 3 ft radially from the production have been published (NACE Inter-
wellbore (Fig. 13-5). national, 1995).
An overflush is also commonly used to displace – Solvents—General purpose aromatic solvents
HCl into a carbonate formation to improve the live based on xylene are used widely as preflushes.
acid penetration. Following the treatment, the well In some cases, treating data show a significant
is cleaned up by flowing the reacted fluids to the reduction in injection pressure with the use of
surface along with the undissolved damage (i.e., a solvent. Solvent preflushes may also help to
drilling fluid, scale, paraffin and asphaltenes). prevent interaction with the formation oil by sep-
• Matrix treatment fluid additives arating the crude oil and aqueous treating fluids.
Chapter 15 provides a detailed discussion of the – Iron stabilizers—Additives to control iron are
additives required in matrix fluids. Additives are required in all acid treatments. They can be
grouped in three categories: buffers keep the pH
value less than 2.2, and the most common buffer
100 is acetic acid, which helps suppress oxide precip-
Percent of original productivity
3-in. collar
80
6-in. collar
12-in. collar
itation at temperatures less than 125°F [50°C];
chelating or complexing agents bond to the iron
60 rc – rx = collar thickness
rc
and suppress other reactions and are used to pre-
Damage
40 collar rx vent precipitation and sludging (Crowe, 1985);
and reducing agents prevent oxidation of the iron
20
rw from ferrous to ferric (Crowe, 1985). Ferrous
0 Wellbore iron is less likely to precipitate and form sludges.
0 1 2 3 4 5 6
Inner radius of damage (ft)
– Surfactants—Surfactants are used to reduce
oil/water surface tension, ensure water wetness,
Figure 13-5. The effect of shifting an 80% damage collar. prevent sludge and stabilize foams. They are

recommended in most acid treatments to solve – Aluminum stabilizer—This class of chemicals

one or more potential problems. Small concen- was introduced in recognition of the role of fluo-
trations (e.g., 0.2% to 1%) of surfactants are suf- roaluminates in secondary precipitation in wells
ficient. Surfactants are also used as dispersants at temperatures above 200°F in sandstones
to keep materials (oil or solids) dispersed in the (Gdanski, 1996). Although new chemicals have
stimulation fluid and as nonemulsifiers to help been introduced to address this problem, citric
prevent emulsions from oil-acid interaction. acid works well and is economical (Rogers et
– Mutual solvents—Mutual solvents are used to al., 1998).
ensure that the formation remains water-wet and – Retarders—Numerous systems for delaying acid
to lower surface/interfacial tension. They are rec- reaction in carbonates and allowing deeper pene-
ommended in most acid treatments. tration into the reservoir have been developed.
– Diverters—Diversion is recommended for all Simulations should be performed to quantify
matrix treatments. The bridging agents benzoic their benefit.
acid and rock salt are used to bridge in perfora- – Nitrogen—N2 is commonly added to treating
tions and/or fissures in carbonate formations and fluids in low-pressure wells to assist in cleanup
injection wells. Oil-soluble resins (OSRs) sized or to create a foam diverter.
to form an external filter cake on the formation Once the proper treating fluid formulations have
face are typically used as particulate diverters. been selected, the pumping schedule must be
Small benzoic acid particles can be used in injec- designed.
tion wells completed in sandstone and nonfissured
carbonate reservoirs. Water-soluble polymers used
as gel diverters must be carefully chosen to have 13-5.3. Pumping schedule generation
the right combination of stability during the treat- and simulation
ment yet break sufficiently to prevent formation
The pumping schedule includes the treating fluid and
damage. They should not be used in sandstones.
diverter sequence and the injection rate of each stage.
Foams are formulated with surfactants and a gas
It is generated using empirical rules based on previous
phase (usually N2). Surfactants must be compati-
field experience or computers.
ble with other additives included in the foaming
A numerical simulator can be used to simulate dam-
solution (Zerhboub et al., 1991).
age removal and evaluate skin effect evolution, flow
– Scale inhibitors—Scale inhibitors are materials profile and wellhead or bottomhole pressure versus
that suppress the precipitation of inorganic scales injection rate for the proposed pumping schedule. The
from produced fluids. In general, they are retained simulator takes into consideration the placement strat-
on the formation and are more effective at neutral egy, an important part of the design process. The
pH values. They are normally applied as an over- schedule can be optimized using a single-phase reser-
flush to an acid treatment or mixed with a brine voir model to meet specific objectives for each fluid
(with or without N2) and displaced into the reser- type (Bartko et al., 1996). This step allows the design
voir when scale formation is a problem. engineer to optimize the schedule to obtain the desired
– Clay stabilizers—In treatments of clay-bearing results in the most cost-effective manner. A treatment
sand formations, these polymeric cationic materi- design is not systematically engineered until it is run
als decrease clay migration. They are temporarily using a field-validated simulator. Most treatments are
effective at low concentrations. Clay stabilizers do currently based on empirical rules of thumb.
not prevent silt (e.g., feldspar, mica, chert) migra- The matrix treatment design process includes sev-
tion (Ezeukwu et al., 1998). They are available in eral steps. The manual process that is discussed in
different molecular weight ranges for higher and detail in Chapters 17 through 19 should be understood
lower permeability formations, and they must be before using computer-aided design programs. This
tailored to the formation to avoid causing damage section focuses on computer-generated design, with
from the physical plugging of pore throats. Clay emphasis on the required input and interpretation of
stabilizers are recommended where experience the output. A field case history for a sandstone is used
indicates that clay migration is a problem. to illustrate the utility of the procedure.

• Design types lines in Fig. 13-4, 50 gal/ft of HCl and 75 gal/ft

The four levels of matrix design are photocopy of mud acid were recommended. Normally, the
design, advisor design, empirical/kinetic-based fluoboric volume is equal to the mud acid vol-
design and geochemical-based design. The first two ume. Thus, the pumping schedule in Table 13-1
are used for 98% of all matrix treatments. The high was developed.
failure rate is partially attributed to this nonengi- Simulation of the advisor design using the
neered approach. A discussion of each follows. kinetic-based numerical simulator described sub-
sequently yielded a decrease in the damage skin
– Photocopy design
effect from 206 to 37. Figure 13-6 shows adequate
A photocopy design is based on a previous treat- skin effect evolution in all layers except layer 4.
ment pulled from the well file or an offset well Based on the negative slope of skin effect evolu-
file. The well name and date are changed, and tion in layers 2 and 4, it appears that more mud
the treatment is submitted for approval or simply acid should be pumped. The total damage skin
processed and pumped. effect versus volume for the same treatment is
– Advisor design shown in Fig. 13-7. The change in slope during
An advisor treatment design is developed on the the mud acid and fluoboric acid stages is attrib-
basis of guidelines (rules of thumb) or experi- uted to retardation of the fluoboric acid and the
ence. The design may be a refinement over the
photocopy design, or it may be similar to it. Table 13-1. Advisor design for sandstone
field case.
Obviously, because a treatment design should be
reservoir- and formation-damage specific, appli- Fluid Stage Description Rate Volume
(bbl/min) (gal)
cation of the advisor design is usually not opti-
mum, as shown for the sandstone case history. Preflush 3% NH4Cl brine 1.0 1000
Sandstone field case history—The well was Preflush 7.5% HCl 1.0 2000
previously selected as a matrix stimulation candi- Main fluid 12% HCl–
date, with clay swelling and/or fines migration 3% HF mud 1.0 3000
damage. Laboratory testing indicated that the Overflush 3% NH4Cl brine 1.0 250
well should be treated with mud acid and over- Main fluid Fluoboric acid 1.0 3000
flushed with fluoboric acid. Following the guide-
3000
1 2 3 1 4 Treatment fluids
1. 3% NH4CI brine
2. 7.5% HCl
2500
3. 12% HCI–3% HF
mud acid
4. Fluoboric acid
2000
Damage skin effect
1500
Layer 4
1000
Layer 1 Layer 3
500
Layer 2
0
0 2000 4000 6000 8000 10,000
Volume (gal)
Figure 13-6. Acid placement: skin effect versus volume by layer (advisor design).

1. 3% NH4CI brine
2. 7.5% HCl
200 3. 12% HCI–3% HF
mud acid
Damage skin effect 4. Fluoboric acid
100
Total skin
0
0 2000 4000 6000 8000 10,000
Volume (gal)
Figure 13-7. Acid placement: skin effect versus volume (advisor design).
change in available HF. The fluoboric acid gener- design engineer to fine tune or optimize the sched-
ates 2.2% HF whereas the mud acid contains 3% ule to obtain the desired results in the most cost-
free HF (Thomas and Crowe, 1981). effective manner. The simulator should be
– Empirical/kinetic-based design validated using field data to remove uncertainties
in the design parameters (Bartko et al., 1996;
An empirical/kinetic-based design is produced
Thomas et al., 1998).
using a numerical simulator. At a minimum, the
Before using an empirical/kinetic-based simu-
simulator should be a two-dimensional (2D), two-
lator, a preliminary pumping schedule should be
phase, finite-difference simulator that allows a
generated with a numerical simulator for matrix
multilayer configuration for computing pressure
sandstone and carbonate acidizing. This advisor
and skin effect evolution during the matrix acidiz-
recommends treatment volumes based on the
ing of sandstones and carbonates. Mineral dissolu-
damage penetration. The flowchart in Fig. 13-8
tion should be simulated using the most common
is incorporated in the expert system to assist
minerals and acids along with the appropriate
diversion selection. The pumping schedule
reaction kinetics (reaction rate limited in sand-
includes the treating fluid and diverter sequence
stones and mass-transfer limited in carbonates
and injection rate of each stage. It is generated
with wormholing). This type of simulator corre-
using empirical rules based on previous field
lates the local porosity change during acidizing to
experience or computers. The schedule can be
a local permeability modification and finally to an
optimized with a single-phase reservoir model
overall damage skin effect evolution per layer.
to meet specific objectives for each fluid type
Currently, precipitation is not considered in the
(Perthuis et al., 1989).
empirical/kinetic-based simulator; however, if the
Sandstone field case history—The pumping
acids are formulated properly this should not
schedule shown in Table 13-2 was generated
affect treatment results.
using the numerical simulator described in this
The empirical/kinetic-based simulator can
section. The objective input to the model was
model damage removal and evaluate skin effect
a target damage skin effect per layer of approxi-
evolution, flow profile and wellhead or bottom-
mately 10% of the original for the mud acid sys-
hole pressure versus injection rate for the pro-
tem. This design is approximately twice the
posed pumping schedule. This step allows the
volume of the preceding advisor design.

Start Interval
> 20 ft
Yes
Yes Not Single-

Yes Yes <100 Yes Well
Foam N2 required gravel fluid
holes perforated
packed treatment
No No No No
No No
Match Yes Consult

density T > 300°F local
ball sealers expert
No
Yes N2 ≤0.475 psi/ft Gradient >0.475 psi/ft

Foam required pressure/
TVD
No
Sandstone Carbonate
Sandstone Carbonate Lithology
Lithology
Sodium Inject Gas Gas Inject Gas Inject Sodium

Well type Foam Well type Foam Well type
benzoate benzoate
Benzoic acid
Oil Oil flakes/rock
Oil
salt
Oil-soluble
Oil-soluble Benzoic acid resin
resin flakes/rock salt
Figure 13-8. Diversion selection flowchart.
Table 13-2. Pumping schedule generated by

To check the validity of the simulator, the
simulator for sandstone field case. actual treatment performed on the well was sim-
ulated. The actual pumping schedule is shown in
Fluid Stage Description Rate Volume Table 13-3. The empirical/kinetic-based simula-
(bbl/min) (gal)
tor yielded a final damage skin effect of 14.2; the
Preflush 3% NH4Cl brine 1.0 1500 actual damage skin effect was 11.4 from a post-
Preflush 7.5% HCl 1.0 5350 treatment pressure transient test. Thus, the simu-
Main fluid 12% HCl– lator predicted a 93% reduction in damage skin
3% HF mud 1.0 6660 effect whereas the actual skin effect reduction
Overflush 3% NH4Cl brine 0.8 840 was 94%, indicative of model validation. The
Main fluid Fluoboric acid 0.8 5050 good agreement of the simulated skin effect evo-
lution of the treatment with the actual pressure
transient test results is shown in Fig. 13-11.
The empirical/kinetic-based numerical simula- Figure 13-12 shows the excellent skin effect
tion pumping schedule yields a total damage skin evolution in each layer for the actual treatment.
effect of 16.6 (8% of the original damage skin Figure 13-13 shows the flow rate into each layer
effect) with good skin effect evolution in layer versus the volume injected. The flow rate into all
4 (Fig. 13-9). layers changed during the treatment; i.e., the rate
The bottomhole pressure and pumping rate are into the thief zones decreased whereas the rate
shown in Fig. 13-10. The overall increase in into the highly damaged zones increased.
pressure during the treatment results from diverter – Geochemical-based design
deposition. This especially dominates in the
stages with the least amount of damage removal A numerical simulator similar to that discussed
(i.e., HCl and fluoboric acid stages). for the empirical/kinetic-based design is used for

3000
1. 3% NH4CI brine
2. 7.5% HCl
2500 3. 12% HCI–3% HF
mud acid
Layer 4 4. Fluoboric acid
2000
Skin effect
1500
1000
Layer 3
500
Layer 1 Layer 2
0
0 5000 10,000 15,000 20,000
Volume (gal)
Figure 13-9. Acid placement: skin effect versus volume (empirical/kinetic-based design).
7200 1.6
1 2 3 1 4
Fracture pressure
1.5
Total rate
6800
1.4
Total
1.3 Rate (bbl/min)
bottomhole
pressure
6400
1.2
1.1
6000
Treatment fluids
1. 3% NH4CI brine 1.0
2. 7.5% HCl
3. 12% HCI–3% HF mud acid
4. Fluoboric acid
5600 0.9
0 5000 10,000 15,000 20,000
Volume (gal)
Figure 13-10. Bottomhole pressure and pumping rate versus volume (empirical/kinetic-based design).

Table 13-3. Actual pumping schedule for • Placement strategy

sandstone field case. Placement strategy is an important step in the
design of a matrix treatment (see Chapter 19). The
Fluid Stage Description Rate Volume
(bbl/min) (gal) goal is how to obtain uniform penetration of the
Preflush 7.5% HCl 1.0 4000
treating fluid throughout the entire section and/or
into each natural fracture system. If complete zone
Main fluid 12% HCl–
3% HF mud 1.0 8000 coverage is not achieved, full production potential
Overflush 3% NH4Cl brine 1.0 500
cannot be realized.
Simulation of the actual sandstone design with-
Main fluid Fluoboric acid 1.0 5000
out a diverter indicates that layers 1 and 3 remain
thief zones throughout the entire treatment (Fig.
the geochemical-based design; however, the pre- 13-14). In addition, skin effect values for the
cipitation of reaction products is considered and severely damaged bottom layer reduce to 75 for
reported. This is accomplished by incorporating treatments with diverter and 1250 for treatments
acidizing equilibrium chemistry for up to 14 ele- without diverter (Figs. 13-12 and 13-14, respec-
ments and 100 reaction products. tively). The poor fluid distribution can result in an
excessive acid volume per foot of layer in the case
• Calculations
without a diverter and inefficient formation damage
Equations to calculate the maximum surface treat- removal. Figure 13-15 shows that the final damage
ing pressure and injection rate are provided in skin effect is 20 in the case without a diverter.
Chapter 16 for vertical and horizontal wells. The Although this represents a 90% reduction in skin
maximum injection rate in a horizontal well usually effect and would yield approximately 4000 BOPD,
exceeds the available pumping equipment capabil- the actual treatment with a particulate diverter
ity; however, many jobs pumped into horizontal yielded a 93% reduction in skin effect and more
wells employ coiled tubing, which limits the rate than 5000 BOPD. Thus, the small amount of
far below that necessary to fracture.
1. 3% NH4CI brine
2. 7.5% HCl
3. 12% HCI–3% HF
200 mud acid
4. Fluoboric acid
Skin effect
100
Total skin effect
Final predicted
skin effect = 14.2
+ actual skin effect = 11.4
0
0 5000 10,000 15,000 20,000
Volume (gal)
Figure 13-11. Skin evolution of the simulation of the actual treatment.

3000
1. 3% NH4CI brine
2. 7.5% HCl
2500 3. 12% HCI–3% HF
mud acid
4. Fluoboric acid
2000
Skin effect
1500
Layer 4
1000
Layer 1 Layer 2
500
Layer 3
0
0 5000 10,000 15,000 20,000
Volume (gal)
Figure 13-12. Skin effect evolution per layer for the actual treatment.
0.6
1. 3% NH4CI brine
2. 7.5% HCl
0.5 3. 12% HCI–3% HF
mud acid
4. Fluoboric acid
Layer 1
0.4
Rate (bbl/min)
0.3
Layer 2
0.2
Layer 3
0.1
Layer 4
0
0 5000 10,000 15,000 20,000
Volume (gal)
Figure 13-13. Flow rate into each layer versus volume injected.

3000
1. 3% NH4CI brine
2. 7.5% HCl
3. 12% HCI–3% HF
mud acid
4. Fluoboric acid
2000
Layer 4
Damage skin effect
1000
Layer 3 Layer 2
Layer 1
–1000
0 5000 10,000 15,000 20,000
Volume (gal)
Figure 13-14. Damage skin effect per layer versus volume for a treatment incorporating the actual treatment volume with-
out a diverter.
1. 3% NH4CI brine
2. 7.5% HCl
3. 12% HCI–3% HF
200 mud acid
4. Fluoboric acid
Total damage skin effect
100
Final predicted
skin effect = 20
0
0 5000 10,000 15,000 20,000
Volume (gal)
Figure 13-15. Total damage skin effect evolution versus volume.

diverter added to the treating fluids improved the zone. This process can be repeated as required on
efficiency of the acid and the resulting well perfor- the basis of information from production and fis-
mance (i.e., added approximately 1000 BOPD). sure location logs. The coiled tubing can subse-
The importance of the placement strategy is mag- quently be run to total depth to start the treatment.
nified in a horizontal well because of the long inter- Drilling breaks and mud logs can be used for the
val. The placement strategy must address the type same purpose. If the data are inadequate, the entire
of tubing used to inject the fluid and diversion of the section is normally treated by alternating acid and
fluid from thief zones to the damaged sections. (The diverter stages as the coiled tubing is retrieved.
term zone is used here to describe longitudinal sec- Diversion must be achieved to ensure that the
tions of a horizontal well, from the heel to the toe.) treating fluid is continuously removing damage
Although conventional tubing or casing can be used rather than simply being injected into a thief zone.
to place the treating fluid, this process can be time Mechanical techniques (straddle packers or ball seal-
consuming or inefficient (Fig. 13-16). ers) are not practical in many wells because they are
Coiled tubing is generally used to overcome these completed openhole, with slotted liners or gravel
problems. Unlike conventional tubing, coiled tubing packed. In cased hole completions, a straddle packer
can easily be run in and out of the hole, and treating can be used to selectively place the treating fluid. In
fluid and diverters can be injected during movement carbonate reservoirs, a chemical diverter (i.e., ben-
(Thomas and Milne, 1995.) Figure 13-17 shows the zoic acid, rock salt, polymer systems, wax beads)
good skin effect evolution obtained with this tech- can be used.
nique. Injection down the annulus of the coiled tub- OSR or foam is normally used in sandstone for-
ing and tubing may be required to optimize fluid mations for diversion. Solid diverting agents such as
diversion and should be modeled during simulations. benzoic acid flakes have been used for decades but
The placement technique is based on the infor- sometimes do not clean up well following the treat-
mation available for the well. For example, if a ment. This especially occurs in carbonates, where
production log or spinner survey, mud logs or a log large quantities of inefficient diverters are commonly
to locate fissures is available, this information can required to create an increase in bottomhole pres-
be used to aid the treatment design. If a spinner sure, resulting in diversion. Although OSRs perform
survey indicates that a thief zone exists in the cen- well in sandstones, numerous operators use foam
ter of the horizontal length, the coiled tubing can be diversion techniques to avoid any chance of plug-
run to that depth, followed by injection of a diverter ging by the diverter. A special application of foam
slug. The diverter does not completely plug the diversion is in high-water-cut wells, where the foam
thief zone but greatly decreases fluid flow into the preferentially plugs the water zone, allowing acid to
10
1 2 1 2 1 2 1 2
Zone 3
8
Treatment fluids Zone 2
1. 15% HCl
6 2. Salt, benzoic
Thief at heel
Skin effect
20 bbl/min bullhead
4 with diverter
Zone 1
0
–2
0 5000 10,000 15,000 20,000
Volume (gal)
Figure 13-16. Skin effect evolution in a horizontal well treated using the bullhead technique.

10
1 2 1 2
8
Treatment fluids
1. Temporarily
crosslinked acid
6 2. 15% HCl
Skin effect
Thief at heel
4 Coiled tubing
Zone 2 with diverter
2
Zone 1 Zone 3
–2
0 5000 10,000 15,000 20,000 25,000
Volume (gal)
Figure 13-17. Skin effect evolution in a horizontal well treated using a coiled tubing placement technique with a tempor-
arily crosslinked gelled acid diverter.
flow into the damaged oil zones (Zerhboub et al., zone followed by running the coiled tubing to total
1991). Similarly, the use of solids for diversion in depth. Acid is then pumped as the coiled tubing is
carbonates has decreased in favor of foam and tem- withdrawn, and a diverter slug is injected every
porarily crosslinked polymer (Saxon et al., 1997). 100–200 ft as required. The result is uniform pene-
This approach eliminates the slow cleanup experi- tration over the damaged sections with a small
enced following conventional treatments and results amount of acid injection into the thief zone located
in good zone coverage in vertical and horizontal at the heel. Figure 13-17 shows the results of a
wells. numerical simulation, indicating diversion from the
Modeling of matrix acidizing of a horizontal well heel (i.e., the thief zone) is accomplished using
with the 2D simulator previously discussed (Thomas coiled tubing and a diverter. Improved diversion is
and Milne, 1995; Jones and Davies, 1996) indicates achieved when annular flow is optimized during the
that bullhead matrix treatments in high-permeability coiled tubing treatment. Figure 13-18 shows a poor
or highly fractured limestones may be inefficient. distribution of acid per zone when the bullheading
Bullheading acid with a diverter normally results in technique is used, whereas the treatment goal is
poor coverage beyond 200 to 300 ft. Apparently the achieved when coiled tubing placement is used (i.e.,
acid rapidly creates a thief zone at the entrance to the proper stimulation of damaged zones 2 and 3). The
zone, and conventional chemical diversion tech- case histories reported in Sidebar 13E support the
niques are ineffective (Fig. 13-16). If coiled tubing results of the simulations.
is used, a diverter can be placed across a known thief

50
45 Zone 1 (heel)
Zone 2 (middle)
40 Zone 3 (toe)
15% HCl per foot of zone (gal)

35
30
25
20
15
10
0
Bullhead with Bullhead with Coiled tubing Coiled tubing
no diversion, diversion, with diverter, with diverter,
20 bbl/min 20 bbl/min 2 bbl/min 2 bbl/min
down tubing down tubing
and annulus
Figure 13-18. Simulation of acidizing a horizontal well with a thief zone at the heel.
13E. Placement study case histories
The following placement studies are the results of work reported by

Thomas and Milne (1995). 0 BWPD 20,000
0
Horizontal oil Well 1: bullhead technique
Horizontal injection Well 1 was completed in a limestone formation
with 1800 ft of 51⁄2-in. slotted liner inside an 81⁄2-in. openhole. The well
initially accepted 1000 BWPD. Because drilling mud damage was
suspected, the well was acidized with 90,000 gal of 15% HCl at 25
bbl/min using the bullhead technique. Diversion was not used in the 50
treatment. After the treatment, injection increased to 16,000 BWPD.
The post-treatment PLT log indicated that 80% of the flow was into
the first 100 ft below the casing shoe (Fig. 13E-1).
Although no diverter was used in this treatment, the addition of
conventional benzoic acid or rock salt diverters was predicted to not
improve zone coverage significantly. This was based on similar
Depth (ft)
results observed in vertical wells completed in long limestone zones 100

acidized using a bullhead technique with diverter. Thus, the acidizing
treatment using the bullheading technique was not effective in obtain-
ing zonal coverage as predicted by the modeling work.
Injectivity into Well 1 declined over the following 2-month period to
12,000 BWPD. The well was subsequently treated with 26,000 gal of
HCl (15 gal/ft) using coiled tubing without a diverter, and injection was
restored to 15,000 BWPD. Although injection was not restored to the
150
original level, it was maintained for 1 year, which was indicative of a
larger surface area accepting fluid.
Horizontal oil Well 2: coiled tubing with foam diversion
Well 2 was completed with a 41⁄2-in. slotted liner in two sections of the
Arab D limestone to yield 2378 ft of horizontal hole. The upper sec-
tion was approximately 1500 ft long with a 100-ft barrier leading down
200
to the lower Arab D limestone with approximately 780 ft of hole. The
well was acidized using 25 and 20 gal/ft of 15% HCl in the upper and
lower sections, respectively. The larger volume in the upper section
was used because of the extended exposure time to drilling mud and
suspected higher damage.
The treatment was performed by running 11⁄2-in. coiled tubing in Figure 13E-1. Post-treatment injection log of a hori-
the hole to total depth, followed by loading the hole with diesel. HCl zontal well acidized using the bullhead technique.

13E. Placement study case histories (continued)
was then injected at 2.5 bbl/min while the coiled tubing was withdrawn to yield the appropriate gallons per foot. A pressure of 1200
psi was maintained on the coiled tubing annulus to minimize acid flow up the backside. After it was pulled 50 ft, the coiled tubing was
stopped and 65% quality foam was injected as a diverter. This sequence was repeated 15 times. Subsequently, the barrier section
was loaded with 65% quality foam.
Once the coiled tubing was in the upper section of the Arab D limestone, it was acidized in 30 stages using the same process of
alternating stages of acid and foam. Following the last acid stage, the coiled tubing was run to total depth and the horizontal section was
displaced with diesel.
Prior to acidizing the well would not flow, yet post-treatment production was 2484 BOPD at 1318-psi flowing tubing pressure.
Post-treatment production from the well is more than that from similar horizontal wells treated with up to 4 times more acid (100 gal
of 15% HCl per foot of the horizontal section). The key to successful horizontal well acidizing was apparently not the volume of acid
but the placement and diversion techniques.
Horizontal oil Well 3: coiled tubing with temporarily crosslinked acid diversion
Well 3 was completed openhole in the Arab D limestone with 1000 ft of horizontal hole. Upon initial completion a bullhead acid treat-
ment was performed to remove drilling mud and cuttings damage. However, the results were unsatisfactory. The goal of the treatment
was to provide uniform injection along the entire wellbore. Subsequently, the well was acidized using coiled tubing placement to inject
10 gal/ft of 15% HCl with silt dispersants and 4 gal/ft of temporarily crosslinked HCl (Saxon et al., 1997). The treatment was performed
by running coiled tubing in the hole to total depth and then injecting 6.3 gal/ft of HCl while withdrawing the coiled tubing from 10,553 to
10,353 ft. This process was repeated while running to total depth (10,553 ft). Next a temporarily crosslinked HCl diverter was injected
while pulling out of the hole to 10,353 ft, which was then repeated 3 times to obtain good coverage. Once the coiled tubing reached
9753 ft, it was withdrawn while injecting HCl.
The post-treatment production log in Fig. 13E-2 shows an 800-ft section from the heel with significant flow near the toe of the
well. (The production logging tool could not go beyond 800 ft.) This favorable flow profile is completely opposite that of Fig. 13E-1,
showing a majority of the flow coming from near the heel of the well. Coiled tubing placement complemented by a temporarily
crosslinked acid diverter apparently yielded improved coverage.
Figure 13E-2.
100 Post-treatment
production log
Percent of total oil rate
80 of a horizontal
well acidized
60 using coiled
tubing and
40
temporarily
crosslinked
HCl diversion.
20
0
9500 9650 9750 9850 9950 10,050 10,150 10,250 10,350
Measured depth (ft)
Horizontal oil Wells 4, 5 and 6: coiled tubing with temporarily crosslinked acid diversion
Wells 4, 5 and 6 were treated similarly to Well 3 with coiled tubing and a temporarily crosslinked HCl diverter. Eight to 15 gal of HCl with
mud dispersants per foot of the horizontal zone was used in combination with 4 to 5 gal/ft of temporarily crosslinked HCl diverter.
The treatment was performed by running coiled tubing in the hole to total depth and subsequently reciprocating across a 100- to
500-ft section while pumping HCl. Next the coiled tubing was withdrawn while the diverter was injected. This process was repeated
as required to obtain coverage over the entire horizontal section. The horizontal length, permeability and production results for the
wells are summarized in Table 13E-1. In all cases, production increased significantly, with treatment payout in less than a month.
The average increase in production was 1630 BOPD.
Vertical wells: bullhead technique

Long (150- to 200-ft) vertical openhole sections of the Arab D formation were historically treated with large volumes of HCl using the
bullhead technique. Normally, 100 gal/ft of 15% HCl was injected down the tubing with diverter stages consisting of rock salt and
benzoic acid flakes. As reported in the preceding horizontal well case histories, the authors observed that the acid went primarily into
the high-permeability zones and the upper sections. This observation is based on the evaluation of pre- and post-treatment flow-
meter injection surveys, which indicated a highly nonuniform injection profile.
Laboratory tests indicated that the diverter system was inefficient. Although a low-permeability filter cake was formed with the
diverter, it was readily penetrated and destroyed by live acid. This situation explained the poor injection profile following the bullhead
technique.

Table 13E-1. Production in horizontal oil Wells 4, 5 and 6 increased 440, 1750 and 2700 BOPD, respectively,
following HCl treatment placed with coiled tubing and diverted with temporarily crosslinked HCl.
Pretreatment Post-Treatment
Well Horizontal kH kV pe Wellhead q Wellhead q

Length (ft) (md) (md) (psi) Pressure (STB/D) Pressure (STB/D)
(psi) (psi)
4 3600 6 3 2089 500 970 600 1410

5 1500 10 3 2950 750 950 875 2700
6 2000 25 12 2800 320 1000 260 3700
As indicated in Table 13E-2, the injectivity index increased although injection into the lower permeability zones in the six studied
wells was not achieved. The average post-treatment injectivity index was 61 BWPD/psi.
Table 13E-2. Pre- and post-treatment injectivity index of vertical wells acidized
using the bullhead technique.
Well Pretreatment Pretreatment Post-Treatment Post-Treatment Pretreatment Post-Treatment

Injection Injection Injection Injection Injectivity Index Injectivity Index
Rate (BWPD) Pressure (psig) Rate (BWPD) Pressure (psig) (BWPD/psi) (BWPD/psi)
A 21,200 1950 12,000 550 24 65

B 41,700 2750 55,300 2200 20 48
C 27,700 2950 52,800 2100 21 70
D 44,200 2800 44,800 1950 30 72
E 33,600 2900 25,700 1200 20 62
F 33,500 2200 41,600 1815 32 47
Average injectivity index = 61 BWPD/psi
Vertical wells: coiled tubing with foam diversion

Five wells were completed in the same field and formation as the six vertical wells discussed previously. Coiled tubing and foam
diverter stages were used to improve placement of the acid. The goal was to improve the injection profile by effectively acidizing
both the high- and low-permeability zones in each well.
Approximately 50 gal/ft of 15% HCl was injected via coiled tubing at 1 to 1.5 bbl/min in stages. The coiled tubing was run to total
depth as the foam was injected. Once at total depth, the coiled tubing was withdrawn as acid was pumped to yield 50 gal/ft. After
withdrawing 20 to 50 ft, the coiled tubing was stopped and a stabilized foam was pumped. This viscous foam was designed to fill the
wormholes created by the previous acid stage to yield improved zone coverage. The staging process was repeated until the top of
the zone was reached.
From Table 13E-3, the average injectivity index for the studied wells was 120 BWPD/psi, nearly double that obtained with the
bullhead technique. Injection surveys also show that the coiled tubing and foam diversion treatment yielded a more uniform injection
profile, indicating that both the high- and low-permeability zones were acidized. This placement technique resulted in excellent over-
all results and cost 20% less than the conventional bullhead technique.

Table 13E-3. Pre- and post-treatment injectivity index of vertical wells acidized
using the foam diversion technique.
Well Pretreatment Pretreatment Post-Treatment Post-Treatment Pretreatment Post-Treatment

Injection Injection Injection Injection Injectivity Index Injectivity Index
Rate (BWPD) Pressure (psig) Rate (BWPD) Pressure (psig) (BWPD/psi) (BWPD/psi)
G 5,000 1,300 67,800 1240 4 119

H 26,200 1,550 42,400 1200 26 112
I 10,000 24,000 59,000 1100 5 139
J 0 38,500 934 65
K 5,000 2,000 73,000 1200 3 165
Average injectivity index = 120 BWPD/psi, a 97% increase over the results obtained with the bullhead technique
Summary
These case histories illustrate that the key to successful matrix acidizing in carbonate reservoirs is not the amount of acid injected
but how it is injected. This is also true for sandstone reservoirs. Coiled tubing placement in combination with foam diversion in sand-
stones or carbonates appears to be an improved technique. Chapter 17 discusses the self-diverting acid system, which has been
used successfully in long carbonate sections, including horizontal wells. This system is advantageous over foam diversion because
it does not require nitrogen.
13-6. Final economic evaluation 13-7. Execution

As discussed for the preliminary economic evaluation The execution (pumping operation) must be per-
in Section 13-2.3, an economic analysis is important. formed as specified by the treatment schedule or oper-
A final economic evaluation should be made on the ator on location. During this process, QC and data
basis of various treatment scenarios (i.e., advisor- collection are important. The operational objective is
based and empirical/kinetic-based designs). An evalu- defined by the design. QC is the process of ensuring
ation of the sandstone field case history designs in the that the materials and equipment meet specifications
previous section follows. and are delivered in proper condition to the wellsite.
Properly trained personnel are the key to success.
• Sandstone field case history
The wellsite personnel must understand the funda-
The advisor-based design decreased the skin effect mentals of the stimulation techniques that will be used
from 206 to 37 (Fig. 13-7), but the empirical/kinetic- and must know how to use the equipment, software
based design decreased the skin effect from 206 to and techniques.
16.6 (Fig. 13-9) to yield an additional 1373 BOPD
(3406 versus 4779) and nearly $1.9 million in NPV
(almost $2.7 million versus $4.5 million). The actual 13-7.1. Quality control
treatment resulted in more than 5000 BOPD at a
lower drawdown, with more than 3 million bbl of Two keys to effective QC are communication and
oil produced over the life of the well. Figures 3-19 documentation. The designer must communicate the
and 3-20 show the economic indicators for the design expectations to the material suppliers. Because
advisor- and empirical/kinetic-based designs, these expectations must be consistent with the supplier
respectively. or manufacturer’s specifications, the specifications

Figure 13-19. Economic indicators for the advisor design sandstone case history. Skin effect decreased from 206 to 37.
Figure 13-20. Economic indicators for the empirical/kinetic-based design sandstone case history. Skin effect decreased
from 206 to 16.6.

should be available to the designer. The designer review procedures in detail to ensure that the available
should request and receive test data on the perfor- equipment is capable of conducting all required opera-
mance of materials (e.g., corrosion inhibitors) as tions. All equipment to be used in stimulation opera-
necessary to ensure that the materials will perform tions should be properly maintained to perform reli-
adequately. Numerous standardized test methods ably and accurately. As discussed in Section 13-5 for
developed by the American Petroleum Institute (API), treatment design, adherence to the design is required
American Institute of Mechanical Engineers (AIME) to increase the chance of success.
and National Association of Corrosion Engineers Calibration of all measuring devices, such as trans-
(NACE) can be used to evaluate materials. Additional ducers and flowmeters, should be a regular part of
specialized testing may also be required in critical maintenance procedures. Calibration conditions should
applications. mimic operating conditions to the extent required to
QC documentation should follow materials properly calibrate the equipment. Sufficient inventories
throughout their life cycle and be available if needed of spare parts should be available to make maintenance
by the designer and operating personnel. Suppliers repairs quickly. Calibration tests should be conducted
and vendors should ensure that materials shipped to routinely and the results documented.
the wellsite meet specifications. Given the time, dis-
tance and expense of delivering materials to the well-
site, nobody benefits if materials not meeting spec- 13-7.2. Data collection
ifications arrive on location. In many cases, a material Careful recording of events during the treatment should
passes through several hands before being delivered be made, including records of unusual observations by
to the wellsite. The QC testing routine and required operations personnel (McLeod and Coulter, 1969).
documentation should be agreed on by vendors and Over the past decade, the emphasis on improved moni-
customers as part of the commercial arrangement toring and recording equipment and QC rather than
under which services are supplied. Testing should be pumping and mixing equipment has resulted in better
sufficient to ensure that materials will perform their records for postjob treatment evaluation and improved
required functions in the field. Chronic failures in matrix success.
testing are a sign that a more reliable material or better The basic information available from a stimulation
handling procedures are required. Testing methods treatment minimally consists of a record of pressure
and options are discussed in Chapter 14. and rate and a log of operations prepared by opera-
The degree of attention paid to QC will vary with tions personnel. All pressure charts can be analyzed;
the critical nature of the operation. For stimulation however, their usefulness may be limited by their
operations, useful QC measures include imprecision (e.g., pressure gauge accuracy may be
• on-site titration of acids to verify concentration ±150 psi) and the difficulty and time required to put
• regular QC testing of each batch or lot of corrosion the data in a more useable form.
inhibitors by the service company The modern approach is to provide continuous
digital monitoring of the surface rate and pressure
• verification that surface-active agents are supplied with either an on-site computer or digital data logger.
in the specified concentration Combined with a detailed log of operations, this infor-
• on-site testing of gel viscosity for diverters (carbon- mation can be analyzed in real time or processed after
ates) the procedure with the same type of software. The key
• regular testing of particle size and solubility of par- advantages of continuous monitoring are higher preci-
ticulate diverters sion of the data and easier manipulation of the digital
• sampling of fluids pumped during stimulation treat- data files using computers. The chief drawback is that
ments, with the samples retained until the treatment bottomhole treating pressure must be calculated from
has been evaluated. surface treating pressure. Reliable means for this cal-
culation are available for Newtonian fluids (e.g.,
The equipment must be able to execute the design brines and acids), but the calculation for complex flu-
required. The designer and operations personnel must ids (e.g., gels, slurries, foams) is not always reliable.

A relatively simple solution to this problem is to 13-8. Treatment evaluation

allow pressure communication between the injected
fluid column and a static fluid column if the reservoir 13-8.1. Pretreatment evaluation
pressure is greater than the hydrostatic pressure result-
ing from a column of fluid. The surface pressure of A step-rate test can be performed prior to the stimu-
the static fluid is measured, and the bottomhole treat- lation treatment to quantify reservoir pressure, perme-
ing pressure is obtained by subtracting the hydrostatic ability and skin effect. The benefit to the operator is
pressure of the static fluid column. The static column improved real-time evaluation. The test requires injec-
is normally an annulus, either tubing/casing or coiled tion of an inert fluid into the zone of interest and bot-
tubing/tubing. Packer assemblies can be modified to tomhole pressure calculation and recording. The first
allow pressure communication. The main operational diagnostic performed determines the reservoir pres-
drawback is that stimulation fluids may leak into the sure using a plot of the bottomhole pressure versus
tubing casing annulus, which necessitates additional rate. Theoretically, the y-axis intercept at zero rate
operations to remove them. is the reservoir pressure (1815 psi in Fig. 13-21).
Memory gauges can also be run on treating strings The second diagnostic incorporates PTA of the data
to record the bottomhole temperature and pressure using the Odeh-Jones methodology. Figure 13-22
during treatment. They are retrieved after the treat- shows the analysis plot used in the process. This water
ment and analyzed directly. Their obvious drawback injection well exhibits permeability of approximately
is that real-time analysis is not possible. In running 12.8 md and skin effect of 0.1. The permeability and
memory gauges, it is critical to select a sufficiently reservoir pressure determined from the step rate test
long time interval to record the operation. If similar should be used in the subsequent real-time evaluation.
treatments using gel, foam or slurry diverters are to be
performed on a series of wells, a memory gauge can
be run on the first one and the data used to develop a
13-8.2. Real-time evaluation
friction pressure correlation for the diverters. Surface In recent years, technology to determine real-time skin
pressure can then be used with acceptable accuracy effect evolution during a treatment has been developed.
on subsequent treatments. Although this technology is not practiced routinely, it
Surface-readout bottomhole pressure recording can be a useful diagnostic tool (see Chapter 20). For
devices are also available. In general, the devices are example, if the skin effect is decreasing during an HCl
expensive and the additional operational difficulties stage, an acid-soluble species (e.g., calcium carbonate/
associated with running them further increase their cost. oxide, iron carbonate/oxide) created damage. This
However, they are invaluable for stimulation evaluation information should be complemented by the well his-
and also for reservoir evaluation and management. tory, laboratory testing, etc., to improve understanding
In addition, where coiled tubing is run, sensor pack- of the present problem and assist future work.
ages to monitor pressure and temperature are used to One method available to the industry is based on
determine bottomhole treating pressure for calculation the steady-state design and evaluation method devel-
of the skin effect. The temperature profile is determined oped by Paccaloni (1979b). Data can be displayed on
prior to a treatment following injection of an inert fluid a graph of the bottomhole pressure versus injection
(e.g., water containing ammonium chloride) to estimate rate. Comparison is made with standard curves calcu-
the flow profile. Another temperature profile can be run lated for fixed values of skin effect to evaluate skin
after the treatment to quantify zone coverage. effect evolution.
Technology exists to transmit the job data to the office More advanced programs calculate skin effect evo-
from most locations. This efficient technique allows an lution in real time, taking into account transient effects
engineer to monitor numerous jobs, participate in deci- (Prouvost and Economides, 1989). The reservoir pres-
sion making and direct operations from off site. sure response during pumping is computed assuming

Figure 13-21. Determination of reservoir pressure from a step rate test.
Figure 13-22. Pressure transient analysis from a step rate test.

zero constant skin effect. One value of skin effect is 13-8.3. Post-treatment evaluation
provided, combining the effects of damage, comple-
tion and diverters. Using these tools, a more quantita- The evaluation of stimulation effectiveness is a pro-
tive assessment of each component of the stimulation cess similar to well performance evaluation. In this
design can be made. section, the process is applied to returning wells to
Figure 13-23 shows the real-time skin effect evolu- production following stimulation. Details on the tech-
tion of the injection well analyzed in the step rate test. niques are provided elsewhere in this volume.
The initial skin effect observed when water containing Careful recording should be made while bringing a
ammonium chloride was injected is close to the skin well on production following stimulation. Early indi-
effect obtained in the previous step rate test. The cations of success can be found in the ease of initiat-
increase in skin effect when HCl was injected into the ing production and higher flow rates and flowing
formation indicates that the HCl was incompatible tubing pressures. Return fluids should also be sam-
with the formation or was incompatible with some- pled. Early production problems may indicate an
thing that was injected in the leading portion of the incompatibility between treating fluids and the pro-
acid. In this example, the latter is suspected because duced oil. The analysis of water samples may indicate
the tubing was not cleaned (pickled) prior to the injec- other problems with the treating fluid selection (e.g.,
tion of acid down the pipe. The initial acid dissolved precipitation problems with HF acidizing).
magnetite and/or rust on the walls of the tubing and Wells should be subjected to pressure buildup and
precipitated ferric oxide before the acid reached the PTA following a stimulation treatment. These data are
zone. Thus, the pressure increased when the solid the basis for a quantitative assessment of the well and
material (ferric oxide) filtered out on the formation reservoir characteristics. Comparison of these results
face with an increase in skin effect to approximately to pretreatment buildups can provide the best assess-
2. As shown in Fig. 13-23, unspent acid removed the ment of the success of the stimulation treatment.
damage and the skin effect decreased to a value of
approximately 1.
1 2 4 3 2
2
Skin effect
–1
3% NH4Cl HCl Mud acid 3% NH4Cl
1 26 51 76 101 126 151

Time (min)
Figure 13-23. Real-time skin effect evolution in the injection well analyzed in the step rate test.

In the final analysis, stimulation treatments are eval- in the well for pressure buildup is less attractive—and
uated at the sales meter. If the treatment results in sus- may not be acceptable—after the well has responded.
tained oil production at a higher rate than before and Using field data, the engineer can evaluate differ-
if the revenue generated by the increased production ences between the design and actual treatment. If a
represents an acceptable return on the cost of the stim- numerical simulator can be rerun with the actual treat-
ulation, the treatment is considered a success. The ment parameters, the model can be calibrated by adjust-
designer can learn valuable information from the eval- ing reservoir parameters such as the damage radius,
uation of well performance, regardless of whether the permeability per layer, skin effect per layer, damage
stimulation is an economic success. mineralogy and diverter efficiency until a match with
In addition to examining oil production response, the actual treatment profiles is obtained. If a post-treat-
changes in the total fluid production (oil, water and ment well test is performed and evaluated, the resulting
gas) in reservoir volumes, gas/oil ratio and water/oil data can be used to further calibrate the simulator.
ratio must be reviewed. Well productivity must also Wireline formation logs and the combination of
be examined. An increase in the productivity of total radioactive tracers and gamma ray or gamma ray spec-
fluids may not be economic, but it may indicate that troscopy surveys (where the energy and intensity of
the target reservoirs were not stimulated or were the gamma rays are measured to enable the discrimina-
watered-out in the candidate well. This can result in tion of multiple tracers) can be used to quantify zone
a change in treatment design philosophy (i.e., to sac- coverage. Of course, the decision to run tracers and
rifice total stimulation effectiveness for the selective baseline logs must be made while planning the stimu-
treatment of a limited part of the reservoir). lation. A pressure transient test (e.g., buildup, four-
Correlation of the unwanted results from stimula- point) to quantify permeability and skin effect can be
tion can provide data for improved performance in the performed. Thus, adequate evaluation tools exist today
future. They may also identify other opportunities to to significantly improve matrix treatment success.
improve field performance, such as water or gas shut- Future treatment designs for the well or field can be
off. In the end, the data acquired from evaluating optimized with the calibrated model. All results should
recently stimulated wells can lead to improved reser- be compiled in a report with recommendations for all
voir management. phases of the design, execution and evaluation (i.e.,
Ultimately, a treatment should be evaluated on the continuous improvement).
basis of the well performance and economic parame- Stimulation operations present the engineer with an
ters used to justify the treatment. Factors such as rate, opportunity to significantly improve the economic per-
flowing bottomhole pressure, reservoir characteriza- formance of the assets under his or her stewardship. If
tion, artificial lift performance and equipment perfor- properly designed, executed and evaluated, stimulation
mance have been discussed. Several weeks or even operations can teach the engineer about the current con-
months may be required for production to stabilize and dition of the well and reservoir and identify other
establish a representative trend. In addition, shutting opportunities to improve economic performance.

◆
Formation Damage: Origin,
Diagnosis and Treatment Strategy
;;
Donald G. Hill, Olivier M. Liétard and Bernard M. Piot, Schlumberger Dowell
George E. King, BP Amoco
; ;
14-1. Introduction
Formation damage reduces the well production
;;
or injection capacity, and the removal of damage is
one of the major goals of petroleum engineers. This
chapter identifies and quantifies formation damage
and includes ideas on treatment strategy. It is impor-
tant to note that not all types of formation damage
require a removal treatment. Some types of damage Tubing Gravel pack Perforations Formation
Scales
will clean up during production, and others can be
Organic deposits
removed by changes in operating practices. In addi-
Silicates,
tion, some producible impairment is misconceived aluminosilicates
as “damage,” when it is actually poor well design Emulsions
that can be remedied with operational changes. Water blocks
Classifying damage correctly requires more than Wettability changes
experience in the chemistry or physics of damage.
A thorough knowledge of field operating conditions Figure 14-1. Location of various types of damage.
is essential, and correct identification is critical to
successful removal.
The terms formation damage and skin effect dam- many publications have appeared on the subject
age have been applied to describe many well pro- (Allen, 1973; Hurst, 1973; Leon, 1973; Sands,
ductivity impairments (Krueger, 1986; Porter, 1973; Christian and Ayres, 1974; Bruist, 1974;
1989). Damage can be anything that obstructs the Shaw and Rugg, 1974; Black and Rike, 1976; Maly,
normal flow of fluids to the surface; it may be in the 1976; Sparlin and Hagen, 1983; Krueger, 1988;
formation, perforations, lift system, tubulars or Amaefule et al., 1998; Adair and Smith, 1994;
restrictions along the flow path. Formation damage Beadie, 1995; Reid, 1996).
specifically refers to obstructions occurring in the The goal of this chapter is to give a broad view
near-wellbore region of the rock matrix. Other types of formation damage. Damage characterization is the
of damage can be identified by location. Figure 14-1 key to proper design of removal treatments. A general
shows some common types of damage; these pro- description of the various damage types and mecha-
duction impairments can occur anywhere in the pro- nisms is presented, followed by a discussion of the
duction system, from the wellbore to perforations origins of damage resulting from natural causes and
and into the formation. Such a distinction is not usu- well operations. The testing required to determine the
ally made because seldom are most of the plugging presence of formation damage and its characteriza-
phenomena located in only one part of the flow sys- tions are also discussed. Treatment strategies for
tem. The importance of determining the causes of removing formation damage are presented.
the observed damage cannot be understated. Only
by knowing the damage mechanism, its location and
how it is affecting flow can an effective treatment
strategy be developed. There have been significant
14-2. Damage characterization
improvements over the past few years in recogniz- Damage characterization is the “history” in damage
ing and describing the various types of damage, and removal. The search for the identity of the damage

begins in the production and development history 14-2.1. Pseudodamage

of the well and even neighboring wells. Drilling
records, completion design, offset well performance Chapter 1 demonstrates the impact of skin effect
and/or operator experiences and past treatment reduction on well performance. Chapter 2 describes
records are all sources of information. The objective techniques for estimation of the total well skin
is to identify the location and types of damage that effect, which are accomplished primarily through
may be a problem. Although damage is usually con- well testing. As discussed in Chapter 1, not all skin
sidered a singular problem, multiple occurrences of effect is due to damage. There are other contribu-
damage are common, some with interfering removal tions that are not related to formation damage. These
treatments. pseudoskin effects are generally mechanical, result-
Formation damage identification and investigation ing from obstructions to flow or because of rate- and
include phase-dependent effects. Their values must be sub-
tracted from the total skin effect to estimate the skin
• types of damage effect associated with formation damage (Petersen et
• location of damage al., 1984). One way to accomplish this is to use
• extent and screening of damage NODAL production system analysis to develop an
inflow performance relationship (IPR) curve specific
• effect of damage on well production or injection.
to the well. NODAL analysis allows optimizing pro-
duction conditions for a given well and thus optimiz-
ing the well completion (Fig. 14-2).
∆p8 = ptf – psep

Sales line
ptf ∆p6 = pdsc – psep Gas
Separator
pdsc psep
Surface choke Liquid
Stock
∆p5 = ptf – pdsc tank
pdsv
∆p4 = pusv – pdsv
∆p 7 = p wf – ptf pusv ∆p1 = p – pwfs = Loss in porous medium
∆p 2 = pwfs – pwf = Loss across completion
Bottomhole ∆p 3 = pur – pdr = Loss across restriction
restriction
pdr ∆p4 = pusv – pdsv = Loss across safety valve
∆p 5 = ptf – pdsc = Loss across surface choke
∆p3 = pur – pdr ∆p 6 = pdsc – psep = Loss in flowline
∆p 7 = pwf – ptf = Total loss in tubing
∆p 8 = ptf – psep = Total loss in flowline
pur
pwf p wfs p pe
∆p2 = p wfs – pwf ∆p1 = p – p wfs
Figure 14-2. Pressure losses in the producing system of a flowing well. pwf = bottomhole flowing pressure, p – = average
pressure, pe = reservoir pressure, pdr = downstream restriction pressure, pdsc = pressure downstream of the surface
choke, pdsv = pressure downstream of the safety valve, psep = separator pressure, ptf = tubing flowing pressure,
pur = upstream restriction pressure, pusv = pressure upstream of the safety valve, pwfs = wellbore sandface pressure.
14-2 Formation Damage: Origin, Diagnosis and Treatment Strategy

14-2.2. Pseudoskin effects and well 0.9

completion and configuration Without turbulence
With turbulence
Positive pseudoskin effects can result from the well
completion design or well configuration. Problems 0.8
k c/k = 1
include k c/ k = 1
• limited entry to flow (Odeh, 1968; Jones and 0.7

Watts, 1971; Saidikowski, 1979)
• off-center wells (Denson et al., 1976; Fetkovitch
k c/k = 0.1
and Vienot, 1984)
Perforation ratio
0.6
• low perforation density, short perforations or
incorrect phasing (Hong, 1975; Locke, 1981;
McLeod, 1983) 0.5 k c/k = 0.1
• mechanical flow restrictions

• mismatched or inadequate fluid-lift systems
0.4
• laminated reservoirs (shale streaks).
A negative pseudoskin effect always exists in 0˚ phasing
4 spf
deviated wells; it is a function of the deviation angle 0.3 0.4-in. perforation diameter
and the formation thickness (Cinco-Ley et al., 1975). No damaged zone
0.7-cp viscosity
• Pseudoskin and rate-and-phase effects 35˚ API velocity
150-md permeability
Operating pressures and production conditions 0.2
0 3 6 9 12 15
may induce additional pressure drops or pseudo- Perforation length (in.)
skin effects. Producing a well at a high flow rate
can cause turbulent flow in perforations and some- Figure 14-3. Effect of turbulence and a compacted zone
times in the formation (Fig. 14-3) (Tariq, 1984; (90% permeability reduction) on the productivity ratio
Himmatramka, 1981). The corresponding positive (Tariq, 1984). k = permeability, kc = damaged zone
permeability.
pseudoskin effect is proportional to the flow rate
above a minimum threshold value. Below this crit-
ical value, this pseudoskin effect is suppressed
(Jones et al., 1976). The problem increases with around the wellbore causing a reduction to the
gravel-packed perforations (for sand control) and relative permeability to oil. A similar but more
for high-production-rate gas wells in general. The severe problem can happen when producing gas
problem becomes particularly acute for the frac- condensate wells below the dewpoint. Both phe-
ture-to-wellbore connection in deviated wells with nomena manifest as positive pseudoskin effects.
fractures at an angle to the wellbore. In unconsolidated sand reservoirs, a flow-rate-
Scale precipitation in and around the perforations dependent skin effect can be caused by modifica-
can progressively modify the reservoir flow conditions of the sand arches around the perforations.
tions, increasing the pressure drop and changing the Abrupt variations of skin effect and concomitant
flow regime from Darcy to non-Darcy and creating sand releases occur above the threshold flow rate
a turbulence pseudoskin effect (Meehan and Schell, value (Tippie and Kohlhaas, 1974).
1983). This is in addition to the actual damage skin • Other pseudodamages
effect.
Other mechanical causes of production impairment
Producing a well below the bubblepoint pres-
are
sure may cause a positive pseudoskin effect as liq-
uids condense around the wellbore, impeding flow – tubing collapse or restriction by lost objects or
(Blacker, 1982; Hinchman and Barree, 1985; any adhering deposit
Economides et al., 1989). This is a relative perme- – collapsed perforations in formations where the
ability phenomenon, with free gas concentrating formation competence has been overestimated

(Antheunis et al., 1976; Chenevert and • acid reactions

Thompson, 1985) • acid by-products
– poor isolation between zones resulting from • iron precipitation
poor cementation of the annulus—e.g., oil inva-
• iron-catalyzed sludges
sion of a gas cap significantly reduces the rela-
tive gas permeability, mixing oils of two differ- • bacteria
ent zones may cause paraffin and asphaltene • water blocks
precipitation, or water invading an oil-bearing • incompatibility with drilling fluids.
zone reduces the relative permeability to oil
(water block), may create emulsions and may Each of these mechanisms is addressed in detail in
cause clay and scale problems. Stimulation treat- the following sections. Damage from iron precipita-
ments that do not achieve required results and tion and iron-catalyzed sludges is discussed else-
that cause additional damage may have been where in this volume.
sabotaged by bad isolation (Abdel-Mota’al,
1983).
14-3.1. Fines migration
– poor design of gas lift systems—small-diameter
tubing, unadapted operating gas lift pressure Formation damage can occur as a result of particle
(Blann and Williams, 1994), improper valve migration in the produced fluid. The particles can
design and high surface backpressure (Jones and bridge across the pore throats in the near-wellbore
Brown, 1971). region and reduce the well productivity. When the
damaging particles come from the reservoir rock,
they are usually referred to as fines. Migrating fines
can be a variety of different materials, including
14-3. Formation damage descriptions clays (phyllosilicates with a typical size less than
Once mechanical pseudoskin effects are identified, 4 µm) and silts (silicates or aluminosilicates with
positive skin effects can be attributed to formation sizes ranging from 4 to 64 µm). Kaolinite platelets
damage. Formation damage is typically categorized (Fig. 14-4) are thought to be some of the more com-
by the mechanism of its creation as either natural mon migratory clays. Table 14-1 lists the major com-
or induced. Natural damages are those that occur ponents of various clays and fines particles. The
primarily as a result of producing the reservoir fluid. table also lists the surface area of the clays, one of
Induced damages are the result of an external opera- the indicators of how quickly the clay can react with
tion that was performed on the well such as a drilling, a reactive fluid (Davies, 1978). Damage from fines is
well completion, repair, stimulation treatment or located in the near-wellbore area, within a 3- to 5-ft
injection operation. In addition, some completion radius. Damage can also occur in a gravel pack (e.g.,
operations, induced damages or design problems silicates and aluminosilicates in Fig. 14-1).
may trigger natural damage mechanisms. The distinction between types of clays depends
Natural damages include more on the arrangement of the atoms in their crys-
• fines migration talline structure rather than any major difference in
• swelling clays Table 14-1. Major components of various clay
• water-formed scales and fines particles.
• organic deposits such as paraffins or asphaltenes Particle Major Surface
• mixed organic/inorganic deposits Mineralogy Components Area (m2/g)
• emulsions. Quartz Si, O 0.000015

Kaolinite Al, Si, O, H 22
Induced damages include
Chlorite Mg, Fe, Al, Si, O, H 60
• plugging by entrained particles such as solids or Illite K, Al, Si, O, H 113
polymers in injected fluids
Smectite Na, Mg, Ca, Al, Si, O, H 82
• wettability changes caused by injected fluids or (montmorillonite)
oil-base drilling fluids

Figure 14-4. Photomicrographs of (a) pore-filing smectite sheets, (b) “books” of kaolinite platelets in a pore space,
(c) honeycomb growth of chlorite on a sand grain and (d) hairs of illite extending from a sand grain.
chemical formula. These structural differences deter- holding the matrix grains together. As a binder or
mine the surface area exposed to the reservoir fluids cement, clay may react with fluids such as acid and
for each clay. Clay reactivity is a function of this sur- water to disaggregate the formation. If the clay
face area. The location of the clay is also critical to cement is shielded by a quartz overgrowth, as is
its reactivity. Authigenic clay is in a pore throat as common in many sandstones, the clay will not be
fill or as a lining (i.e., grown in the pore from miner- reactive.
als in the connate water) (Wilson and Pittman, Only authigenic clays, unprotected clay cements
1977). Authigenic clays have a large amount of sur- and the few detrital clays on the pore boundary are
face area exposed in the pore and can be reactive. worth consideration as potential damage mecha-
Detrital clay is part of the building material in the nisms. Scanning electron microscopy (SEM) is gen-
original matrix. Detrital clays are usually less reac- erally used to determine clay type; however, recogni-
tive than authigenic clays because they have less sur- tion of the type of clay should not be staked entirely
face area in contact with the fluids in the pore. The on an SEM analysis. Focused dispersive X-ray
vast majority of detrital clays usually cannot be con- analysis is much more accurate. Even after identifi-
tacted by sufficient volumes of reactive fluids to cation of the clay, laboratory core flow tests are typi-
cause problems. Clay may also act as a cement, cally required to determine if the clays within the

flow channels are reactive with a given fluid. The present in the tubing, perforations and formation
common clays that account for most of the real and (Fig. 14-1). The most common oilfield scales are cal-
perceived clay problems are kaolinite, smectite cium carbonate, calcium sulfate and barium sulfate.
(montmorillonite), illite and chlorite. The structures Water-formed scale deposits are among the most
of kaolinite, smectite, illite and chlorite are shown in troublesome damage problems (Cowen and
Fig. 14-4. Simply because the clay is in the rock Weintritt, 1976). Scale usually consists of precipi-
does not mean that the clay is reactive. tates formed from mixing incompatible waters or
upsetting the solution equilibrium of produced
waters. A water that may be stable under reservoir
14-3.2. Swelling clays conditions may become supersaturated with an ion
Clays may change volume as the salinity of the fluid when the pressure decreases, which allows carbon
flowing through the formation changes. Several dioxide (CO2) outgassing, or the temperature
authors have dealt with clay swelling in sandstones, changes. The supersaturated solutions react by pre-
showing either ion exchange, movement or critical cipitating a compound from solution. The deposition
salt concentration triggering clay dispersion (Azari of scale is influenced by pressure drop, temperature,
and Leimkuhler, 1990b; Jones, 1964; Khilar and dissolved gases, flow viscosity, nucleation sites and
Fogler, 1983; Mungan, 1968; Sharma et al., 1985; metal type—in short, anything that upsets the solu-
Priisholm et al., 1987). Changes in formation perme- tion equilibrium.
ability resulting from the alteration of clay are due to The following scales are among the most trouble-
the amount, location and type of clay minerals with- some.
in the formation. The total quantity of clay inside the • Calcium carbonate or calcite (CaCO3)
formation is a misleading indication of potential CaCO3 is usually formed when the pressure is
changes to permeability. It is the arrangement of the reduced on waters that are rich in calcium and
clay, its chemical state at the moment of contact and bicarbonate ions. The deposition can be affected
the location of the clay with respect to the flowing by CO2 outgassing, which raises the pH value and
fluids that are responsible for the changes. Predicting makes the high concentrations of calcium unstable.
the response of a clay to water flow is almost impos-
sible without testing. • Gypsum (“gyp”)
The most common swelling clays are smectite and Gypsum may be the most common sulfate scale
smectite mixtures. Smectite swells by taking water in the oil industry (Cowen and Weintritt, 1976).
into its structure. It can increase its volume up to With a chemical structure of CaSO4 ⋅ 2H2O, it
600%, significantly reducing permeability. If smec- shares a similar composition to the hemihydrate
tite clay occupies only the smaller pore throats and CaSO4 ⋅ 1⁄2H2O, commonly called plaster of paris
passages, it will not be a serious problem; however, or by its mineral name, bassonite. It is also formu-
if it occupies the larger pores and especially the pore laically similar to the evaporite mineral anhydrite
throats, then it is capable of creating an almost (CaSO4).
impermeable barrier to flow if it swells.
• Barium sulfate (BaSO4)
Clays or other solids from drilling, completion or
workover fluids can invade the formation when these BaSO4 is a less common form of sulfate deposit,
particles are smaller than the pore throat openings. but it causes extensive problems. Almost any com-
Any subsequent increase in flow rate through the bination of barium and sulfate ions causes precipi-
invaded zone will force a high concentration of parti- tation. It is difficult to remove, as it is not signifi-
cles into the rock matrix. cantly soluble in acids and solvents unless it is
finely ground or the structure is interrupted with
impurities such as carbonate scale. Like calcium
14-3.3. Scales sulfate, barium sulfate is usually thought to be a
product of mixing incompatible waters, with pre-
Scales are water-soluble chemicals that precipitate
cipitation accelerated by pressure drop, outgassing
out of solution in response to changes in conditions
or turbulence. Some barium sulfate is radioactive;
or the mixing of incompatible waters. They can be
this is part of naturally occurring radioactive

material (NORM) scales. The radioactivity results 14-3.4. Organic deposits

from a concentration of uranium in the lattice of
the scale. The buildup of radioactive scale can be Organic deposits are heavy hydrocarbons (paraffins
monitored using a gamma ray logging tool. Care or asphaltenes) that precipitate as the pressure or
must be exercised when analyzing well debris to temperature is reduced. This is a form of distillation.
avoid mislabeling barite (BaSO4) from drilling They are typically located in the tubing, perforations
mud residue as barium sulfate scale. or formation (Fig. 14-1). Although the formation
Strontium sulfate or celestite (SrSO4) is a com- mechanisms of organic deposits are numerous and
mon substitute in the barium sulfate crystal lattice. complex (Houchin and Hudson, 1986), the main
Strontium scale can be associated with radioactive mechanism is a change in temperature or pressure
scale (NORM). It may be more soluble than bar- in the flowing system. Cooling of the wellbore or the
ium sulfate in chemical remover systems. injection of cold treating fluids has a much more pro-
nounced effect.
• Iron scales Organic deposits must not be confused with anoth-
Iron scales such as iron carbonate and iron sulfide er type of deposit called sludge. Sludges are viscous
can be extremely difficult to remove. They are emulsions produced by the reactions between certain
usually seen in wells that have both a high back- crude oils and strong inorganic acids or some brines.
ground iron count and a tendency to precipitate Sludges cannot be easily dissolved.
calcium carbonate. Iron sulfide scales react accord- • Paraffins
ing to their structure. Seven different forms of iron
sulfide scale have been identified. Only two of Paraffins are the simplest of hydrocarbons. They
these iron sulfide forms are readily soluble in are composed of only carbon and hydrogen atoms,
hydrochloric acid (HCl). The remaining iron sul- and the carbons occur as an unbranched chain.
fide scales are either slowly soluble or not signifi- Carbon chain length associated with formation of
cantly soluble. solid paraffin deposits has a minimum of 16 car-
bon atoms per molecule and may have up to 60 or
• Chloride scales more. The precipitation of paraffins is triggered by
Chloride scales, such as sodium chloride precipita- a loss of pressure, loss of temperature or loss of
tion from water caused by temperature decrease or short-chain hydrocarbon compounds (i.e., the light
evaporation of the water, are common. There is no ends). The temperature at which the first solid
effective way to prevent salt precipitation, and paraffin crystal forms from an all-liquid solution
cleanup has been accomplished using water only. is called the cloud point. Designing the completion
Salt has a limited solubility in acid (1⁄4 lbm/gal in so that produced fluid surface temperatures are
28% HCl), so using acid is not generally consid- above the cloud point and modifying the cloud
ered. Redesigning the mechanical system to avoid point using chemical methods are accepted prac-
temperature loss and water evaporation is also a tices to prevent paraffin deposition in the tubing.
possibility. Melting points increase as the length of the
paraffin chain increases. The hardness of the solid
• Silica scales
paraffin structure also increases with molecular
Silica scales generally occur as finely crystallized size. Table 14-2 lists several paraffin chain lengths
deposits of chalcedony or as amorphous opal. and their melting points. Impurities may cause the
They are associated with alkaline or steamflood melting point of a field sample to vary slightly.
projects and stem from the dissolution of siliceous Paraffins can form anywhere in the producing
formation minerals by high-pH fluids (Lieu et al., system when conditions become favorable for pre-
1985) or high-temperature steam condensates cipitation (Cole and Jessen, 1960; Burger et al.,
(Reed, 1980; McCorriston et al., 1981; Amaefule 1981; Newberry et al., 1986; Thomas, 1988;
et al., 1984). This dissolution can cause poorly Newberry, 1981; Sutton and Roberts, 1974).
consolidated sandstones to collapse or silica to Paraffins are normally found in the tubing near the
reprecipitate at a distance from the wellbore where surface, where the temperature and pressure drops
the alkalinity, temperature or both of the floods has are highest. In cases such as reservoirs that are
decreased.

Table 14-2. Melting points of paraffins. to the quantity of asphaltic materials. Ratios larger
than 1:1 (resins to asphaltenes) are more stable,
Carbon Atoms (no.) Melting Point (°F) whereas ratios less than 1:1 are unstable and may
16 64 precipitate during production. Ratios of more than
17 72 10:1 are known and are much less likely to cause
18 82 significant problems. Although asphaltene contents
20 100 up to 60% have been found, major problems occur
23 122 with oils with a 1% to 3% asphaltene range.
25 129 Asphaltene precipitation can be influenced by
32 158 pressure drop, shear (turbulence), acids, solution
42 181
CO2 (lowers the pH value), outgassing of CO2 and
49 196
other gases (turbulence), injected condensate, gas,
60 211
commingling with other (incompatible) oils and
charged metal surfaces (Danesh et al., 1988;
nearly pressure depleted or formations that have
Monger and Trujillo, 1991; Kawanaka et al., 1991;
experienced dry gas cycling (which removes the
Monger and Fu, 1987; Pittaway et al., 1987).
light hydrocarbon ends), the paraffins can form at
Anything that takes away the resins or breaks the
the perforations or in the formation. Paraffins may
stability of the aggregate particle can lead to a pre-
also be precipitated by the injection of a cool fluid.
cipitation of asphaltene. Iron ions in solution (usu-
Although not typically considered, this latter cause
ally during an acid job) compound and favor the
can be the reason for the slow cleanup of many
formation of asphaltene deposition. As noted pre-
wells after stimulation.
viously, the concentration of asphaltenes is not a
• Asphaltenes good indicator of potential problems. Only the
Asphaltenes are organic materials consisting of con- treatment history and well response examination
densed aromatic and naphthenic ring compounds can suggest the potential for asphaltene problems.
with molecular weights of several hundred to sev- Asphaltene deposition on pore walls may not
eral thousand (Leontaritis, 1989; Leontaritis and significantly decrease the formation porosity and
Mansoori, 1987; Tuttle, 1983; Newberry and absolute permeability. However, through this
Barker, 1985; Addison, 1989; Bunger, 1979; process, the rock tends to become oil-wet, which
Thawer et al., 1990). They are characterized by the reduces the relative permeability to oil (Clementz,
nitrogen, sulfur and oxygen molecules they contain 1982; Collins and Melrose, 1983) and, under cer-
and are defined as the organic part of the oil that is tain conditions, favors the buildup of emulsion
not soluble in a straight-chain solvent such as pen- blocks if water is simultaneously produced.
tane or heptane. Asphaltenes are generally found in • Tar
one of three distinctive forms:
Tar is simply an asphaltene or other heavy-oil
– hard coal-like substance deposit. It cannot be removed by acid or mutual
– blackened sludge or rigid-film emulsion (usually solvents. Removal requires dispersion in an aro-
triggered by iron in solution) matic solvent, and energy is typically necessary
to achieve removal.
– in combination with paraffins.
In “solution,” they usually exist as a colloidal
suspension, forming particles 30 to 65 Å in diame- 14-3.5. Mixed deposits
ter and stabilized by maltene molecules in the oil. Mixed organic/inorganic deposits are a blend of
The volume of the maltene resins is the first key to organic compounds and either scales or fines and
the stability of the asphaltene in suspension. The clays. When migrating, fines associated with an
actual quantity of asphaltenes in the oil is much increase in water production in a sandstone reservoir
less important. The stability of asphaltic disper- become oil-wet, and they act as a nucleation site for
sions depends on the ratio of the quantity of resins organic deposits (Houchin and Hudson, 1986).

Figure 14-5 shows such a mixed deposit in which fication of the surface film strength at the interface
clear (white) crystals of sodium chloride are dis- by chemical reaction, precipitation or the addition of
persed in a dark organic matrix. partially wetted fine particles, electric charge, or high
viscosity of the components or resultant fluid viscosity.
These forces may act singly or in combination.
Natural surfactants help stabilize emulsions by
stiffening the film around the droplet or by partially
wetting small solid particles. Natural surfactants are
present in many waters and most crude oils. They
may be of several chemical formulas and may be a
by-product of bacteria or originate as part of the oil-
generation process. Like other surfactants, they have
an oil-soluble end and a water-soluble end (usually
possessing a small electric charge) and congregate
at the oil/water interface.
Micron-size solids in the liquid may stabilize an
Figure 14-5. Thin section of a layered matrix deposit. The emulsion by increasing the toughness of the surface
black layers are organic deposits, and the clear (white) lay- film around the droplets or by acting as an emulsifier
ers are inorganic scales of mainly halite (NaCl).
and binding droplets of the dispersed liquid with an
electrical charge. Almost any solid can be a stabiliz-
ing agent if it is sufficiently small. For a solid to be
14-3.6. Emulsions effective in stabilizing an emulsion, it must be pres-
Emulsions are combinations of two or more immis- ent at the interface of the drop and the continuous
cible fluids (including gas) that will not disperse phase. The more common solid materials that stabi-
molecularly into each other (Hoover, 1970; Sherman, lize oilfield emulsions are iron sulfide, paraffin, sand,
1968; Lissant, 1974; Lissant and Mayhan, 1973; silt, clay, asphalt, scale, metal flakes (from pipe
Bandbach, 1970; Hausler, 1978; Bikerman, 1964; dope), cuttings and corrosion products.
Ogino and Onishi, 1981; Gidley and Hanson, 1973; Changes in the pH value can affect emulsion sta-
Coppel, 1975). Emulsions are composed of an exter- bility. Most free-water knockouts and treaters oper-
nal phase (also called nondispersed or continuous) ate efficiently at a pH value of 6 to 7, depending on
and an internal phase (also called dispersed or dis- individual well conditions. Following an acid treat-
continuous). The internal phase consists of droplets ment, the pH value may drop below 4 and emulsions
suspended in the external phase. Almost all emul- may be created. Emulsions created in this manner
sions found in the field are produced by the addition are stable until the pH value rises above 6 or 7.
of some form of energy that produces mixing. Most When acid treating a well where the crude is an
emulsions break rapidly when the source of energy is emulsifier or a sludge former, the wellhead may
removed. The breaking mechanism of these unstable be equipped with a chemical injection port just
emulsions is by droplet contact and growth and then upstream of the choke or a chemical injection valve
by fluid density separation. As the droplets draw may be placed in a gas-lift mandrel at some point in
near and touch, the surface film around the drop may the treatment string. These ports are used to inject an
thin and rupture, forming large drops in a process emulsion breaker or de-emulsifier.
called coalescence. The larger droplets settle rapidly
owing to density differences between the liquids
forming separate layers. Only a portion of the drops
14-3.7. Induced particle plugging
that touch will coalesce. When minimum coales- In addition to naturally occurring migrating particles
cence occurs, the emulsion is stable. such as clays and fines, many foreign particles are
If separation of the emulsion does not occur, there introduced into the formation during normal well
is a stabilizing force acting to keep the fluids emulsi- operations. Drilling, completion, workover, stimula-
fied. The most common stabilizing forces are a modi- tion, and secondary or tertiary production operations

can cause the injection of extraneous particles into models cannot predict the damaging potential of
the formation. these products because they do not include any phys-
Particle damage from injected fluids happens in the ical description of the way they are precipitated.
near-wellbore area, plugging formation pore throats. Hydrated silica (Crowe, 1985) may precipitate on
Problems include bridging of the pores, packing of clay surfaces but is not necessarily damaging. Com-
perforations and the loss of large amounts of high- pounds such as borosilicates (Thomas and Crowe,
solids fluid into natural fractures or propped fracture 1978) and fluoborates (Bertaux, 1989) can even be
systems. The best method of avoiding this type of beneficial, probably because they precipitate as films
damage is to use a clean fluid in a clean flow system that bind fines to the sand grains. This phenomenon,
with a controlled range of particle sizes that will stop when purposely produced, leads to efficient fines-
fluid loss quickly by bridging at the wellbore. stabilization treatments (Thomas and Crowe, 1978;
Induced particles can be composed of a wide Sharma and Sharma, 1994). However, gelatinous
range of materials. Particulate materials in drilling precipitates, such as ferric hydroxide, can completely
fluids that are potentially damaging are clays, cut- plug pores and can be particularly difficult to remove
tings, weighting agents and loss-control materials, (Crowe, 1985). Another class of by-products consists
including polymers (Barna and Patton, 1972; Fisher of species such as fluorsilicates (Bertaux, 1989),
et al., 1973; Sloan et al., 1975). These materials can which precipitate in the form of individual crystals
be a problem independent of mud type (oil, water or that can migrate toward pore throats and produce a
polymer base). Workover and stimulation fluids can “log jam.” Iron sulfide, which precipitates even at low
also contain suspended solids (Rike, 1980; Rike and pH values during the acid treatment of sour wells, is
Pledger, 1981) including bacteria and polymer another compound belonging to this category.
residues (Tuttle and Barkman, 1974). Kill fluids in
particular use various polymers as weighting agents
or for fluid-loss control. Typical kill-fluid solids 14-3.8. Wettability alteration
include salt pills with polymer, crosslinked poly- Formation plugging can be caused by liquid (or gas)
mers, hydroxyethylcellulose (HEC) polymers, lost- changing the relative permeability of the formation
circulation pills and CaCO3 with polymer. rock. Relative permeability can reduce the effective
Particles in stimulation fluids are a result of poor permeability of a formation to a particular fluid by as
water quality, tank coatings, tank residuals, and pip- much as 80% to 90%. The wettability and related
ing and tubing debris (e.g., dried mud, scale and pipe relative permeability of a formation are determined
dope). The problem with stimulation fluids is that by the flowing-phase quantity and by coatings of nat-
they can contain effective cleaners and acids that dis- ural and injected surfactants and oils.
perse and partially dissolve the debris inside of the If a drop of a liquid is placed on the surface of
tanks and piping on their way to the formation. another immiscible liquid or on the surface of a solid
Particle damage after stimulation may also occur that it cannot dissolve, it may spread out into a thin
when partially dissolved parts of the formation or the film or it may remain in the form of a drop or a thick
damage materials come back, through either the lens (Hausler, 1978). If the drop of liquid spreads, it
pores or the natural fracture or propped fracture sys- wets the surface; if the drop of liquid does not
tem. The release of fines from the formation is usu- spread, it does not wet the surface. The surface free
ally brought about by cleaners and mutual solvents. energy of the two phases and the interfacial tension
Acid treatments may also cause formation damage between them determine whether the liquid spreads
because of the precipitation of secondary acid reac- or remains in a deformed drop. Wettability is mea-
tion products. Precipitation products include iron sured by the contact angle that a droplet of fluid
hydroxide, calcium fluoride, potassium fluorsilicate forms on a particular surface. If the angle of contact
and silica. θ is less than 90°, the drop spreads from the initial
Geochemical models can predict the chemical ball shape and the surface is said to be wetted by the
nature of these by-products, depending on formation liquid. The smaller the angle, the greater the water
rock and treatment fluid compositions on one hand wettability. If the angle is more than 90°, the surface
and pressure, temperature and contact time on the is not wetted by the liquid. Wettability can be mea-
other (Walsh et al., 1982; Dria et al., 1988). The

sured with a liquid surrounded by gas or a liquid sur- • sludges produced by the reaction between acids
rounded by an immiscible liquid. Simple wetting has and asphaltenes, especially in the presence of
little meaning—it is the resultant effect of wetting on some additives (particularly surfactants) or dis-
fluid flow that is important. solved iron
In the natural state, formations may be water-wet, • by-products precipitated by the reaction of acids
oil-wet or neutral, depending both on the surface with formation materials. Geochemical simulators
exposed to the fluid and the natural surfactants in the can predict the chemical nature of the by-products,
fluid. There are some instances, such as the Cardium depending on the formation rock and treatment
reservoir in Alberta, Canada, where both oil- and fluid compositions and the pressure and tempera-
water-wetting behavior have been reported in differ- ture. Simulators cannot predict the damaging
ent sections of the reservoir. potential of the by-products. Hydrated silica may
When a surface of a pore passage is oil-wet, more precipitate on clay surfaces and is not necessarily
of the passage is occupied by the bound oil (thicker damaging. Compounds such as borosilicates and
monomolecular layer), and less of the pore is open fluoborates can even be beneficial. Gelatinous pre-
to flow than in a water-wet pore. Naturally, to get as cipitates, such as ferric oxide, can completely plug
much flow capacity as possible in a formation, it is pores and be particularly difficult to remove.
desirable to change the wettability to water-wet (in Another class of by-products consists of species
most cases). Unfortunately, it is impossible to change such as fluorsilicates precipitating in the form of
most naturally oil-wet surfaces for long. Wettability individual crystals that can migrate toward pore
can be modified by preflushing the formation with a throats and then bridge in the throats. Iron sulfide
wetting surfactant or a solvent that establishes a new that precipitates, even at very low pH values dur-
coating on the face of the formation or cleans the ing the acidization of sour wells, is another com-
current coating from the formation. Regardless of the pound belonging to this category.
altered condition of a surface, the wettability is even-
• precipitates formed by the addition of certain
tually decided by the surfactants in the produced
sequestering agents to acids to prevent iron prob-
fluid. Thus, the water-wet condition of a formation
lems when the acid is spent and no iron is present
following an acid job can revert to an oil-wet condi-
tion after a sufficient volume of strongly oil-wetting • permeability impairment by residues present in
crude is produced. corrosion inhibitors or produced through the ther-
mal degradation of polymers, such as friction
reducers.
14-3.9. Acid reactions and acid reaction
by-products
14-3.10. Bacteria
Numerous problems that may occur during acidizing
treatments include Although many microorganisms can be present in
the unsterile world of the oilfield, only a handful
• damaging material from the tubing entering the
produce widespread problems (Shuler et al., 1995;
formation
Clementz et al., 1982; Crowe, 1968; Carlson et al.,
• oil-wetting of the reservoir by surfactants, espe- 1961; Raleigh and Flock, 1965). Bacteria can be a
cially corrosion inhibitors, which can create emul- serious problem in production operations because of
sion blocks what they consume and their by-products. Bacteria
• water blocks can grow in many different environments and condi-
• asphaltene or paraffin deposition when large vol- tions: temperatures ranging from 12°F to greater
umes of acid are injected. than 250°F [–11° to >120°C], pH values ranging
from 1 to 11, salinities to 30% and pressures to
In addition to these common damaging processes, 25,000 psi.
production impairment can result from poor design Bacteria are classified as follows:
of an acidizing treatment. Impairments include the
• Aerobic bacteria are bacteria that require oxygen.
following:

• Anaerobic bacteria do not need oxygen (in fact, The bacteria that attack polymers are various aero-
their growth is inhibited by oxygen). bic types and a few of the anaerobic bacteria. Most
• Facultative bacteria can grow either with or with- polymers are excellent carbon sources that are read-
out oxygen because their metabolism changes to ily consumed to support rapid bacterial growth rates.
suit the environment. They usually grow about 5 The resulting large quantities of biomass contribute
times faster in the presence of oxygen. to formation plugging. All these bacteria can be con-
trolled by the application of various biocides.
The bacteria most troublesome in the oilfield are
sulfate-reducing bacteria, slime formers, iron-oxidizing
bacteria and bacteria that attack polymers in fracturing 14-3.11. Water blocks
fluids and secondary recovery fluids.
Sulfate-reducing bacteria cause the most problems Water can cause blocking in low-permeability rocks
in a reservoir. Sulfate-reducing bacteria reduce the (Fig. 14-6) (Keelan and Koepf, 1977). Water blocks
sulfate or sulfite in the water to sulfide and produce are a special case of relative permeability problems.
hydrogen sulfide (H2S) as a by-product. The reduc- In a water block, water usually occupies the flowing
tion process provides the energy for bacterial growth. spaces (either pores or natural fractures) that are typ-
Biomass accumulation can lead to pitting of the steel ically used by hydrocarbons to flow to the wellbore.
under large colonies. The H2S increases the corrosiv- Because of the mobility and viscosity differences,
ity of the water and creates the possibility of blister- the hydrocarbon fluid may not be capable of displac-
ing of carbon steels and sulfide cracking. The by- ing the water. The most severe cases of water blocks
product of an H2S attack on steel is iron sulfide (the are usually observed in low-pressure, low-permeabil-
general form is FeS). Sulfate-reducing bacteria are ity, gas-producing formations after treatment with
anaerobic bacteria with slow growth rates when oxy- water that has a high surface tension.
gen is present. Sulfate-reducing bacteria occur natu-
rally in surface waters, including seawater. The 1.0
growth of the sulfate-reducing bacteria is controlled
0.9
by temperature and limiting their access to nutrients.
The primary nutrients are carbon, nitrogen, phospho- 0.8
rus and dissolved iron. Bactericides are also used
0.7
commonly to control these bacteria.
Relative permeability
Irreducible Residual
water oil
Iron-oxidizing bacteria are aerobic and convert 0.6
iron from the ferrous (Fe2+) to the ferric (Fe3+) state. kro =
ko
0.5 k
They produce gelatinous ferric hydroxide, which is
highly insoluble and precipitates out of water. The 0.4
bacteria metabolize dissolved iron in the water.
0.3
Ferrous iron is soluble only at low pH values (i.e.,
when the water is acidic). Therefore, FeOH3 is typi- 0.2 krw =
kw
k
cally considered an acid-reaction product. The iron-
0.1
oxidizing bacteria produce some corrosion, but they
usually cover sulfate-reducing bacteria colonies and 0
0 10 20 30 40 50 60 70 80 90 100
protect them from attack.
Water saturation (% pore space)
Slime-forming bacteria are facultative and produce
mats of high-density slime that cover surfaces. Their
primary detrimental effects are the protection of col- Figure 14-6. Water blocks: increasing the water saturation
from 20% to 35% decreases the relative oil permeability
onies of sulfate-reducing bacteria and pore plugging. from 90% to 30%, respectively (Keelan and Koepf, 1977).
ko = oil effective permeability, kro = oil relative permeability,
kw = water effective permeability, krw = water relative per-
meability.

14-3.12. Oil-base drilling fluids these materials to bridge and severely decrease the
permeability of the near-wellbore area.
Oil-base mud (OBM) is the drilling fluid of choice Such damaging processes are usually limited to
for the lubricity required in many highly deviated the first few inches around the wellbore (an aver-
wells and for formations that are extremely sensitive age value of 3 in. is commonly used), but the
to water-base mud (WBM). Most OBMs, and partic- resultant permeability reduction can be as high as
ularly those with densities greater than 14 lbm/gal, 90%. Invasion of formation rock by drilling fluid
contain sufficient solids to create silt-stabilized solids is favored by
emulsions when mixed with high-salinity brines or
acids. These emulsions are viscous and resist break- – large pore size of the formation rock (Brownson
ing. Some of these emulsions have been shown to be et al., 1980)
stable for several months, both in the laboratory and – presence of fissures and natural fractures in the
in the wellbore. The level of damage caused by these reservoir
emulsions can be so severe that an entire pay zone
can be missed. For example, in a South Texas well – small particle size of the solid components
the OBM emulsion created damage so severe that of the drilling fluid (the initial particle size of
almost no flow from the well could be measured. weighting agents and lost-circulation preventers
When the damage was removed, the well tested at is usually coarse but can be fragmented by the
more than 12 MMscf/D. drill bit) (Abrams, 1977)
A related problem with OBM is the relative per- – low drilling rate resulting in mudcake destruc-
meability effects commonly created by the powerful tion (mud-loss increase) and long formation-to-
wetting surfactants used for creating stable OBM. mud contact time
When these materials coat or adsorb onto the forma-
tion, the wettability of the formation is altered, and – high drilling fluid circulation rate (mudcake
permeabilities may be only 10% to 20% of what erosion)
they were initially. The most severe problems usually – high drilling fluid density causing large over-
occur with muds weighing more than 14 lbm/gal. balance pressure (Givens, 1976)
The main cause of problems is oil-wetting of the
fines from weighting and viscosifying agents and – scraping of mudcake, provoking pressure surges
from cuttings. and increasing formation-to-mud contact time
during bit trips (Records, 1976).
Using clear brines (containing no particulate
14-4. Origins of formation damage materials) as drilling fluids minimizes formation
invasion by fines but may create a large loss of flu-
This section describes the origins of formation dam-
ids in the rock matrix.
age and reviews typical well operations, including
When drilling a formation with natural fractures,
drilling, cementing, completion, gravel packing, pro-
some mud loss is expected in the natural fracture
duction, stimulation and injection for enhanced oil
system. Because natural fractures are important to
recovery. All are potential sources of damage.
reservoir flow, avoiding loss of mud to the fracture
Damage is also commonly categorized by its associ-
system by using a high-quality fluid-loss control
ated well operation (Tables 14-3, 14-4 and 14-5).
system is crucial. If the natural fractures are
already damaged by mud, the success of the
14-4.1. Drilling cleanout will depend on how much mud was lost
and the type and condition of the mud when it was
• Mud solids invasion lost. If a low-solids mud was used in a system
Mud solids can progressively fill the porosity with minimal overbalance, little damage may have
of the reservoir rock if forced into the pay zone. occurred. If a high-weight mud system with a large
Subsequent attempts to start production or injec- amount of fines was used or if the drilling over-
tion at moderate or high flow rates may cause balance was high (more than 2 lbm/gal overbalance
equivalent), damage may be severe.

Table 14-3. Initial damage during drilling or completion.
Condition or Type of Damage Diagnostic Clues Remedial Operation
Unstable formation May occur in any formation that is poorly Gravel packing, fracture packing, plastic
consolidated or that will fail under pressure consolidation or production rate limits
May occur with onset of water production

or loss of pressure from depletion
Problems include embedment of proppant,

closing of acidized channels in acid fractures,
spalling of formation into perforations or
wellbore, or production of solids.
Oil-base mud emulsion damage Common in wells drilled with OBM Wash with aromatic solvent followed by
mutual solvent and acid
If treated with acid or brine before the
sulfonate emulsifiers are washed off the May require several treatments
cuttings by production or solvent treatments,
an emulsion can lock up the well. Cuttings removal is important.
The first one or two treatments may be short-

lived as more mud and mud filtrate move
back toward the wellbore, as is often the
case in naturally fractured formations.
Fracture plugging from mud Large whole-mud losses in naturally fractured Acid useful if damage is shallow
formations
Fracture if damage is deep
Intermittent production at low and moderate
rates Prevent by improving solids recovery
Infrequent recovery of whole mud and mud In severe cases, a sidetrack drill of the
fines pay may be necessary.
Some emulsions, especially after acid

treatments
May also occur if the hole is poorly cleaned

during drilling
Particle damage from drilling Skin effect on buildup test HCl or HCl-HF in matrix acid job and
and completions solvent wash followed with acid in wells
Injection difficult with OBM
May show emulsions in oil wells Foam or jetting cleanups can be useful.
Poor mud conditioning before cement
Common in openhole completions and

horizontal wells
Poor perforations Shows up as damage on a buildup test but Reperforate

cannot be cured with acid or mechanical
changes in the well Problems with a well that cannot be broken
down or even pumped into should always
Common problems are screenouts of fracture be approached by reperforating the well.
downhole scale occurrence, unstable emulsions,
and downhole paraffin and asphaltene deposits.
Cement in natural fractures Poor well response following completion when Small fracture treatment or sidetrack drill
possible perforation problems have been
eliminated

Table 14-4. Damage during and after stimulation.
Migrating fines Kaolinite or fibrous illite clay or some Clay control for prevention
feldspars (nonclays)
Retarded acid for removal
Brine changes may trigger fines movement.
May require limiting rate in extreme case
Sporadic reductions in flow rate or fracture treatment to spread out
draindown
Variable production rate tests
Fines in produced fluids
Emulsions rare but possible
Particle damage after stimulation May also occur following acid or fracture Filter treatment fluids
stimulations or workover fluids where dirty
water was used or the water was hauled Use clean tanks
or stored in a dirty tank
Particles in waterfloods Reduced injection rate Better water filtering
Higher injection pressures
Backflow shows particles and oil carryover return.
Wettability problems Commonly occurs after an acid job where the Treat with mutual solvent wash over the
corrosion inhibitor was not mixed in the acid pay, displace and soak
just before injection
Emulsions and reduced flow are common,

particularly after inhibitor loss or OBM contact.
May be permanent, but usually cleans up slowly
with time and flow
Attempts to reverse natural wettability are

usually short-lived.
Natural wettability is determined by the natural

surfactants in the produced fields.
Relative permeability problems May occur when oil is injected into a gas zone Treat with a high-API-gravity solvent such
or gas is injected into an oil zone that is above as condensate or xylene (low flash point)
the bubblepoint
Squeeze and produce back
Poor load-fluid recovery Usually formation dependent Avoid or minimize water contact and
lower the surface tension of the water
May decrease production rate in severe cases to prevent
or cause a long cleanup time
Removal with alcohols and some
Most common in formations with small pores surfactants
and in microporous clays
Water blocks Usually in gas wells with small pore throats, Matrix treat with alcohol or surface-
untreated water and low-pressure formation tension-lowering surfactant
If low pressure (less than about 0.2 psi/ft), the Inject gas in gas reservoir to a distance
pore throat size has no effect of 10 ft
Swelling clay Smectite clay, some illite and smectite Acidize with HCl-HF if damage is shallow
interbedded clay
Fracture if damage is deeper than 12 in.
Permeability is sensitive to change in water
salinity or brine type.
Microporosity (water trapping) Caused by some forms of clay No treatment necessary
May bind water and make high Sw readings

on log without water production

Table 14-5. Damage during production.
Retrograde condensate This special case of the relative permeability Control drawdowns and repressure
effect is a condensate (liquid phase) that forms reservoir
from a rich gas.
Redesign tubing if forming in tubulars
If the condensate forms in the tubing or casing,
heading may occur.
If the condensate forms in the formation, creation

of another phase will reduce permeability to gas.
Usually occurs near the wellbore as pressure

drops there
Paraffins in tubulars Pressure restriction in tubulars Scraping and cutting for mechanical
removal
Soft to hard mass found at pressure drops
Hot oil useful if the deposit is less than
Reddish brown to black in color, with white or 100 ft from surface
gray also possible
Solvent soaks on deeper deposits
Also commonly occurs in tubing near the surface
as oil cools and the cloud point is reached Inhibitors available for pipelines and
some problem wells
May increase as a problem as a field ages
Some wells require continuous downhole
Most paraffin deposits melt at less than treatment through “macaroni string.”
150°F [65°C].
Special bacteria are useful.
Paraffins in flowlines Soft to hard deposits (not a scale) in surface Mechanical or solvent removal or pigging
flowlines and equipment
Inhibitors can be used
Paraffin melts when exposed to sufficient heat .
(usually about 150°F).
Paraffins after stimulation Injection of a cool stimulation fluid may precip- Allow the well to clean up on its own
itate paraffin in the reservoir on contact.
Where this problem is known to occur,
The well may be cleaned up slowly (1 to 4 prevent by using xylene preflush ahead
weeks) after stimulation, although load fluid of acid
is recovered faster.
May have decreasing skin if multiple buildup

tests are performed
Paraffins in formation Seen as skin on test Treat with downhole heat-generating

processes if the well is a good producer
May disappear if well is shut in for several days
Solvent soaks also used
Cloud point of oil is near the reservoir
temperature. Some inhibitors can be used with a
fracture treatment.
Pressure drop may trigger paraffin drop out.
Asphaltenes Black, soft to hard mass that may occur as Treatment with aromatic (cyclic ring)
flakes, sludge, marble-size balls and a sticky solvents such as xylene or toluene
buildup that occurs with paraffins
Some surfactants are also useful for
Precipitation is triggered by destabilization of dispersion of the asphaltic mass.
maltene resins caused by acid contact, outgas-
sing, shear in pumps, electrically charged metal Use antisludge additive or xylene with
surfaces, temperature reduction and CO2. acid in reservoirs with more than 0.5%
asphalt to prevent sludges
Asphaltenes soften with increasing temperature
(>150°F) but do not melt.

Table 14-5. Damage during production (continued).
Tar Flows slowly into perforations during Solvent soak as required

production of oil
Test solvent with sample of tar before job
Usually associated with the presence of a tar
deposit near pay and typically highly asphaltic Heat often helps.
May contain some water that is tied up as

droplets or “pockets” in the high-viscosity mass
Emulsion Unstable emulsions that break on standing No downhole treatment suggested
Created in tubing at pressure drop points in Treat on surface if string redesign is

the piping system from pumps to choke impractical
Silt-stabilized emulsion Stable emulsion with partially wetted fines Treat with mutual solvent and acid
at interface
Remove downhole source of solids if
Common after drilling mud dispersal or possible
cleanup of mud or cement fines by acid
May also occur on polymer cleanup
Common in production from unconsolidated

formations, especially after acid or gravel-pack
operations
Surfactant-stabilized emulsion Stable to highly stable emulsion Treat on surface if temporary
Common to severe after acid treatment Use mutual solvent or surfactant to

prevent emulsion with next acid treatment
Stabilized skin may be seen at the drop interface.
Sludge (iron/asphaltic) Sludge is an emulsion that is nearly solid. Prevention is the best cure.
May be triggered by acid, OBM, asphaltenes or Use nonsludging acid systems, and test
iron compounds at the iron content expected in the well
Disperse the sludge in xylene and analyze for

components, particularly iron
Bacterial infestation This difficult problem to predict is more common Treat with sodium hypochlorite followed
in injection wells where surface or produced by HCl (do not allow contact of sodium
water is injected. hypochlorite and HCl)
If the colony is established in the water handling More than one treatment may be
system, it can occur with the injection of necessary.
any waters.
Alternative treatments are chlorine dioxide
Brown to black slimy masses or an H2S odor and bactericide slugs.
when tubing is pulled
Bacteria may cause slow reductions in the

injectivity of an entire field.
Complete removal of bacteria is rarely possible.
Treatment is usually on a remedial basis.
Untreated water in treatments
Drilling fluid or injection water can sour

reservoirs with sulfate-reducing bacteria.
Calcium carbonate scale May form at any pressure drop, either in the HCl to remove and inhibitor to prevent
formation or tubulars

Calcium carbonate scale May form quickly and can sharply limit pro- Inhibitor may be squeezed into the
(continued) duction, especially at gravel-pack interfaces or formation for longer lived protection.
near perforations in wells with high drawdown
across the perforations Some HCl jobs may trigger calcium
carbonate scale in rare cases; inhibit acid
May be more common in earlier stages in some or treat with EDTA if this is a problem.
fields when the pressure drop is more severe
Usually has no crystal pattern
Calcium sulfate scale Usually forms at pressure drop induced by Chemical converter or dissolver
turbulence followed by acid (do not contact converter
or dissolver with acid)
More common where high-sulfate waters
contact high-calcium waters and in CO2 floods Acid is not useful alone.
Scale is not acid soluble. Inhibitors placed by squeeze treatments

are useful for prevention.
May be found on the outside of pumps and
at intakes and gas expulsion ports and valves
downhole
Crystals are characteristic for this scale.
Barium sulfate scale Nonreactive scale that forms at pressure draw- Scraping, water blasting or other
downs or where outgassing occurs mechanical removal
No readily apparent crystal pattern in many Chemical treatment is usually not

deposits possible if scale occurs as a nearly
pure (greater than 90%) deposit or as
May occur as NORM scale in areas where thick (greater than 1⁄4 in.) deposits in pipe.
radioactive isotopes form in the lattice
Can be prevented by inhibitors
NORM scales are detectable with gamma ray
logging tools.
Iron carbonate scale Carbonate scale tendencies with large iron HCl for thin deposits or mechanical
content removal where possible
Molar acid reactivity on the scale
Brownish colored scale (cleaned of oil)
Iron sulfide scale Hard scale, dense and heavy Mechanical removal with mills or cutters
Many forms are not acid soluble. Water jets may not work.
Some forms are mildly magnetic.
Salt Precipitates as a white mass in the tubulars Freshwater or weak brine wash
or in the formation
Usually associated with a cooling of super-

saturated water, but can also be triggered by
a pressure drop
May be seen early in some wells but becomes

less of a problem as the water cut increases
Problems in formations with produced water

salinity that is near saturation
Hydrates (ice) in gas wells Gas well with intermittent flow to nearly total Glycol or alcohol injection below the
shutoff, followed by return to flow in a few hydrate formation point
minutes
Insulated risers or tubing
Produces a small amount of water

Hydrates (ice) in oil wells In oil wells, usually forms only near the Insulated risers
mudline in Arctic regions
May also form in drilling fluid
Fill in perforations High skin effect Clean out or reperforate
Well can be injected into at low rates.
Reperforating may show sharp increases.
Fill in casing High skin effect Reverse circulation or regular circulation
Difficult or impossible to inject into well
Partial fill shows skin effect on a buildup test,

but injection is possible at a reduced rate.
Confirm with a sinker bar on wireline
Water coning Onset of water production after extended Limit rate

production
Some treatments may be temporarily
Formation has no vertical permeability barriers useful.
and sufficient vertical permeability to allow
water to move toward drawdown. Most water control products are not useful
without natural reservoir barriers
Waterflood breakthrough Examine produced water analysis and compare High-permeability zones should be
through high-permeability zone with flood water for identification plugged deep (depth greater than 100 ft)
from producer and injector when oil
Watch for scale recovery from the zone is complete.
Commingled water production Initial production of water with oil in primary, Temporary or no treatment
with breakthrough in flood
Water cut increases in flood.
Collapsed pipe May show up as reduced rate or destruction If caused by earth-shift forces, use
of lift equipment heavier pipe or multiple strings
Check with a gauge ring on wireline or tubing Liners, cement and patches are used for
repair.
Most common causes are earth-shift loads
caused by subsidence of producing formations
with fluid and sand withdrawals, active faults
and formation movement near salt zones.
Other causes include severe corrosion, mal-

functioning perforating guns, pipe flaws and
wear of tubulars from drilling or lift system.
Tubing problem Well refuses to flow although pressure is Redesign string

sufficient with the expected assistance from
dissolved gas. A velocity string that fits inside the
existing tubing may help if tubing
Well may load up and die, or liquid slugs may is too large.
be produced if the string is too large.
Rate is restricted by friction backpressure if

tubing is too small.
Leaks Sudden changes in gas/oil ratio, water/oil ratio, Repair

pressure or chemical analysis of water
Consider corrosion control program

• Drilling fluid filtrate invasion Problems with clays are not detailed here.
For economic reasons, wells have to be drilled as However, any change in the salinity of the pore
fast as possible. To increase the penetration rate, fluids may affect the stability of clay particles in
it is tempting to reduce the fluid-loss control of the the porous medium. In particular, reducing the
mud (Black et al., 1985; Simpson, 1985a, 1985b; salinity or increasing the pH value of the water
Montogmery, 1985). About 600 bbl of fluid can be surrounding clay particles can promote dispersion
lost into a typical formation (Kutasov and Bizanti, of these materials. Destabilized formation fines act
1985) during the drilling of a 10,000-ft well. similarly to mud particles forced into formation
Higher values of filtrate invasion may result from rocks. As production or injection starts, fines
the deliberate choice of high penetration rates. migrate toward pore throats and either block them
Before this decision is made, the effect of the fil- or bridge over them, depending on the particle
trate and any associated solids should be known. size. The severity of the resulting permeability
The liquid phase of a drilling fluid also contains impairment varies with the fines concentration in
many potentially damaging compounds. Because the reservoir rock (Krueger et al., 1967; Droddy
filtrate invasion can be deep, as shown in Table et al., 1988; Jiao and Sharma, 1992).
14-6 (Simpson, 1974; Hassen, 1980), drilling fil- Factors that increase the probability of drilling
trate damage can be one of the most important fluid invasion include
causes of production impairment. The severity – high permeability of the mud filter cake (a result
of this damage depends on the sensitivity of the of either poor design of the drilling fluid or
formation to the filtrate. High-permeability clean detrimental drilling procedures)
sandstones undergo more invasion than low-
– high overbalance
permeability reservoirs (Phelps et al., 1984) but
are more likely to be less affected when their con- – long formation-to-drilling-fluid contact time.
nate water is chemically compatible with the fil- WBM filtrates may have a low salinity and a
trate. An average permeability reduction of 40% high pH value and may contain dispersants and
has been suggested; however, any value between polymers. Water is a cause of in-situ clay distur-
0% and 100% is possible, depending on the nature bance and water blocking in low-permeability
of the formation rock and fluids. rocks (Keelan and Koepf, 1977). The numerous
Sensitive formations contain clays that can be drawbacks of water-base drilling fluids led to the
dispersed or swollen, are low-permeability rocks development of OBM for drilling through sand-
in which saturation problems dominate or are stones containing clay (Methven and Kemick,
reservoirs producing almost saturated brines or 1969). The initial conclusion was that this new
high-content asphaltene or paraffin oils. mud was a safe, all-purpose drilling fluid. It is
Table 14-6. Depth of filtrate invasion
now recognized, however, that although the prob-
(Simpson, 1974). lems of OBM are less numerous than those of
WBM, they are commonly much more severe
Depth of Invasion (in.) (Goode et al., 1984). OBM filtrates contain addi-
Time (D) Oil-Base Low-Colloid Water-Base tives that cause emulsion and wettability problems
Mud Oil-Base Mud Mud
(Ballard and Dawe, 1988; McKinney and Azar,
1 1.2 3.3 7.7 1988; Sanner and Azar, 1994; Ventresca et al.,
5 4.6 11 12 1995). Polymer filtrates that are stable at circu-
10 7.7 17 18 lating temperatures, but already potentially dam-
15 10 21 23
aging, can decompose and form residues when
exposed to static reservoir temperatures for long
20 12 23 27
periods of time (Tuttle and Barkman, 1974), as
25 14 29 31
previously discussed.
30 16 32 34

14-4.2. Cementing – Loss of cement to the natural fracture system is

a catastrophic problem when using a cased and
• Washes and spacers perforated completion through formations with
The removal of drilling mud, while necessary for natural fracture systems. In side-by-side well
improving cement bonding, typically exacerbates comparisons, cement loss in natural fractures
formation damage through either increased fluid in the pay zone has been shown to decrease pro-
loss or incompatibility problems with cement duction to the point where the zone cannot flow
washes and spacers. effectively. Once cement is lost into the fracture
The duration of a cementing job is short com- system, hydraulic fracturing or sidetracking and
pared with the duration of drilling through a pay redrilling the well are the best alternatives. In
zone. The maximum depth of invasion by the fil- some carbonate formations, acid fracturing may
trate from either spacers or cement slurries is a be beneficial.
few inches, which is negligible compared to the
few feet of drilling mud filtrate invasion. This does
not mean that cement or spacer fluid loss should 14-4.3. Perforating
be neglected. Poor fluid-loss control can result in
Perforations are the entry point from the formation to
premature job failure by either complete loss of
the wellbore, and all flow in a cased, perforated com-
the preflush fluids (insufficient volumes) and sub-
pletion must pass through these tunnels. Although
sequent contamination (and strong gelling) of the
perforation job quality is at times overlooked in the
cement slurry by the drilling fluid or dehydration
search for reasons why a well does not produce as
of the cement slurry itself.
expected, any time that formation damage is sus-
• Cement slurries pected the perforations should be examined first.
The broad particle-size distribution of cement Perforating is always a cause of additional damage
grains, together with the use of high-efficiency (Suman, 1972). Extreme overbalance (EOB) perfo-
fluid-loss agents, results in limited particle and rating has been used expressly to reduce damage to
filtrate invasion of cement slurries (Jones et al., the perforation tunnels (see Chapter 11).
1991b). However, there are four cases where large • Perforating mildly overbalance always forces for-
permeability impairments can occur: mation and gun debris into perforation walls and
– The relatively high pH value of cement slurry decreases the permeability near the perforations
filtrates is particularly detrimental to formation (Keese and Oden, 1976).
clay minerals. Calcium ions liberated by cement • Perforating mildly overbalance in fluids that con-
particles are quickly exchanged on clays near tain particles produces a similar effect (Paul and
the wellbore, and the resulting modification of Plonka, 1973; Wendorff, 1974) and also builds a
the filtrate composition makes it a perfect desta- dense, impermeable cake on the perforation walls.
bilizing fluid in terms of dispersing ability
• Insufficient perforation penetration does not
(Cunningham and Smith, 1968).
bypass drilling damage (Klotz et al., 1974; Weeks,
– Cement filtrate that comes into contact with con- 1974). Penetration also decreases with formation
nate brines that contain high concentrations of effective stress (Saucier and Lands, 1978), a defi-
calcium can provoke the precipitation of calcium nite concern in deep wells.
carbonate, lime (Records and Ritter, 1978) or • If the underbalance pressure required to achieve
calcium silicate hydrate (Krueger, 1986). damage-free perforations is incorrectly estimated,
– Overdispersed slurries (with no yield value) pro- the insufficient pressure difference will limit dam-
mote the rapid separation of cement particles at age removal (Hsia and Behrmann, 1991;
the bottom and water at the top of the cement Behrmann, 1995; Bird and Dunmore, 1995),
column. A large invasion of free water will most whereas excessive pressure differences lead to
likely take place, and the resulting water block- sand influx in the wellbore (King et al., 1985;
age may be significant. Seanard, 1986).

• Low perforation density restricts flow. • screens with slots too large (do not retain gravel)
(Flanigan, 1980) or with slots too narrow that
The perforating guns and processes should leave
become plugged and reduce production.
adequate entrance holes for the amount of fluid flow-
ing into the wellbore. This can range from one shot
every other foot in low-rate, homogenous formations
(high vertical permeability) to as many as 12 to 16 14-4.5. Workovers
shots per foot (spf). Most formations differ in verti- The various types of damage from completion and
cal to horizontal permeability, with horizontal per- workover fluids are similar to the damage created
meability from 3 to more than 10 times the vertical by drilling fluids (Eaton and Smithey, 1971; Patton
permeability. This property makes perforation den- and Phelan, 1985):
sity critical, especially if there are shale laminations
• permeability impairment of formation rocks and
in the pay zone. If too few perforations are used in
productivity impairment of perforations by sus-
a laminated or highly structured zone (many vertical
pended solids (Rike, 1980; Rike and Pledger,
permeability barriers), then flow from the zone will
1981) including bacteria and polymer residues
be only a fraction of what an openhole completion
(Lissant, 1974)
could be.
• common problems resulting from filtrate invasion:
clay swelling and dispersion (Azari and Leim-
14-4.4. Gravel packing kuhler, 1990a, 1990b), water blocks (Oudeman et
al., 1994) and emulsion blocks, and scale precipi-
Formation damage mechanisms can affect gravel
tation.
packs. Gravel packs are sand-exclusion techniques,
essentially filters, in front of which formation fines The necessity of using clean, filtered workover flu-
are expected to bridge. It is almost universally true ids has long been recognized (Fig. 14-7). Techniques
that gravel packs deteriorate with time, causing a to place these fluids adequately without contamin-
progressive reduction of well performance. This is in
contrast to current methods of sand production control 500
and high-permeability fracturing (called frac and
pack), in which well performance improves with time.
Major sources of damage in gravel packs are 2.5 ppm
• improper placement of the gravel pack (perfora- 26 ppm

tions remain empty or the annulus between casing
and screen is incompletely filled), allowing perfo-
94 ppm
ration filling by formation sand, pack fluidization
Permeability (md)
100
and subsequent intermixing of sand and gravel in
the case of pressure surges (Stadalman et al.,
1985; Jones et al., 1991a; Chuah et al., 1994)
50
• damage by unbroken gels or formation particles
during placement as a result of incomplete perfo-
Bay water filtered through
ration cleaning (Sparlin, 1974) 436 ppm 2-µm cotton filter
• invasion by loss-control materials (LCM) Bay water filtered through
5-µm cotton filter
(Blanton, 1992; McLeod and Minarovic, 1994; Produced water, untreated
Hodge et al., 1995) Bay water, untreated
• thread dope, paint, rust and polymer residues 0
forced between formation sand and the gravel 0 0.02 0.04 0.06 0.08 0.10
pack during placement Volume injected (gal/perf)
• inadequate gravel size, leading to gravel-pack Figure 14-7. Apparent permeability reduction in Cypress
invasion by formation fines during production sandstone cores with the injection of various filtered and
(Gulati and Maly, 1975) unfiltered waters (Tuttle and Barkman, 1974).

ating them with the various particulate or gumlike – oil-wetting of the reservoir by surfactants, espe-
materials in the wellbore (including thread dope) cially corrosion inhibitors (Crowe and Minor,
have also been proposed (Krause, 1986). 1982), which can create emulsion blocks
There is a strong requirement for fluid-loss con-
– water blocks
trol, especially in depleted reservoirs, and additives
have been developed for this purpose (Crowe and – asphaltene/paraffin deposition when large vol-
Cryar, 1975; Mahajan and Barron, 1980). Another umes of acid are injected.
approach is to use foams, gases (Dahlgaard, 1983) or In addition to these common damaging pro-
mists as completion fluids (Millhone, 1983). Instead cesses, specific production impairment can result
of physically reducing the extent of filtrate invasion, from poor design of acidizing treatments. These
the fluid-loss agents can deliberately be removed impairments include the following:
from workover fluids. In this approach, the compati-
bility of the fluid with formation minerals and brine – sludges produced by reaction between acids and
must be carefully studied (Morgenthaler, 1986). asphaltenes (Moore et al., 1965; Houchin et al.,
Workover brines (especially high-density brines) 1990), especially in the presence of some addi-
usually require inhibitors to control corrosion. These tives (particularly surfactants) (Knobloch et al.,
products can contribute to emulsion-block problems 1978) or dissolved iron (Jacobs and Thorne,
through the wettability modification of formation 1986)
minerals and sometimes promote iron precipitation – deconsolidation of the formation rock caused by
in the reservoir (Potter, 1984). excessive dissolution of the cementing materials
by acids
14-4.6. Stimulation and remedial – precipitation of by-products from the reaction
treatments of acids with formation minerals (Boyer and
Wu, 1983; Gadiyar and Civan, 1994)
• Wellbore cleanup
When wells are cleaned to remove deposits or cor- – precipitation caused by some sequestering
rosion products from the tubing, high concentra- agents added to acids to prevent iron problems
tions of damaging materials may invade the pay if iron is suspected present (Smith et al., 1969)
zone. Extreme care should be devoted to prevent- – permeability impairment by residues from corro-
ing these suspensions from being forced into the sion inhibitors (Crowe and Minor, 1985) or pro-
porous medium. Particularly dangerous are com- duced through the thermal degradation of poly-
pounds that are soluble in the cleaning fluid mers, such as friction reducers (Woodroof and
because they cannot form impermeable cakes that Anderson, 1977)
prevent formation invasion. Rust in acid (Gougler
et al., 1985) or paraffins in hot oil (Newberry and – dissolution of pack sands, to a limited extent, in
Barker, 1985) are the two most typical redissolved hydraulically fractured and gravel-packed wells
wellbore compounds. They reprecipitate in the for- (Cheung, 1988; Yeager, 1990).
mation and cause extensive, severe and usually • Fracture treatments
permanent damage. Damage resulting from hydraulic fracturing takes
• Acid treatments two distinct forms: damage inside the fracture
Problems already encountered in other phases of itself (proppant-pack damage) and damage normal
the life of a well may also occur during acidizing to the fracture intruding into the reservoir (frac-
treatments, especially in cases of inappropriate ture-face damage). The first generally occurs
design. These include because of inadequate breaking of the fracturing
fluid polymer; the second occurs because of exces-
– damaging materials from the tubing entering the sive leakoff. Depending on the reservoir perme-
formation rock ability, the impact of these two damages varies.

For low reservoir permeability, neither one is 14-4.7. Normal production or injection
much of a factor. As the permeability increases, operations
proppant-pack damage (and its avoidance)
becomes increasingly important, whereas damage • Unconsolidated formations
to the reservoir face is relatively unimportant. At Formations that are capable of releasing parts of
high permeabilities, both are important, with frac- the matrix during production or after stimulation
ture-face damage dominating at very high perme- pose special treating problems. Although these sit-
abilities. uations are commonly thought of as sand-control
The selection of fracturing fluids, polymer con- problems rather than formation damage, the effect
centrations and breakers is critical in addressing of mobile sand and the pressure drop caused by
these issues. Incomplete breaking of the polymers collapsed formation tunnels closely resemble the
in fracturing fluid is the most obvious cause of effect of formation damage. Some reservoirs can-
damage within hydraulic fractures (Gidley et al., not be produced at high flow rates or large draw-
1992), as well as the poor selection of proppant downs without being affected adversely. Per-
fracturing fluids (Brannon and Pulsinelli, 1990) manent damage, which cannot be removed simply
and formation rock spalling or creeping into the by the reducing production rate, may be created.
proppant pack (Strickland, 1985). True damage in A major problem is the movement of fines in
the formation rock is the consequence of excessive the formation in response to either flow velocity
leakoff in high-permeability reservoirs when poly- or changes in the salinity of the flowing fluid. Al-
mer-base gels are used in combination with ineffi- though this subject was addressed in Section 14-3.1,
cient fluid-loss agents (Elbel et al., 1995; Parlar et it is worth mentioning again, because it is usually a
al., 1995). These damages are usually severe and significant factor in the behavior of unstable forma-
usually cannot be improved with matrix treat- tions. Native silts and clays loosely attached to pore
ments. To alleviate this problem, polymer-free, walls can be put into motion by high flow rates
surfactant-base fluids have been proposed (Stewart (Hower, 1974; Holub et al., 1974), especially
et al., 1995); however, they cannot be used in the when two or more immiscible fluids are produced
absence of crude oil, which is required for breaking at the same time (Muecke, 1979; Sengupta et al.,
the surfactant micelles. 1982). Depending on their particle size, they can
either block pore throats in the vicinity of their ini-
• Poor load-fluid recovery tial location or migrate toward the wellbore.
Typical load-fluid recovery on a stimulation treat- Whether migrating particles reach the wellbore
ment or remedial treatment may range from as little or bridge over pore throat entrances depends on
as 20% to 100%. Load fluids can invade and their original concentration in the formation, their
become trapped in the formation by entering the size, the extent of the increase of their concentra-
smaller capillary pores during higher injection pres- tion near the wellbore (because of radial flow) and
sures, or they may coat clays with high micro- the maximum flow rate (Vaidya and Fogler, 1990;
porosity (a condition in which a large surface area Gunter et al., 1992; Oyenenin et al., 1995).
exists for water coating or trapping). Once the Bridging is promoted when one or more of these
injection pressure is released, capillary pressure parameters are increased. Bridging is less detri-
effects will hold significant volumes of the fluid in mental than blocking because a short period of
the formation. In some formations, more than 50% reverse flow, followed by production at a flow rate
of the load fluid can be trapped after a treatment, lower than that at which bridging occurred, may
and yet the formation may not appear damaged. mechanically disperse bridges (Fig. 14-8). How-
Most of the flow is through the larger pore system ever, this cannot be achieved when the bridging
and natural fractures, and these passages usually agglomerates were previously cemented by precip-
clean out quickly. The smaller pore passages can itates (asphaltenes or scales) or chemically stabi-
trap fluid by imbibing or absorbing it, but they have lized through the injection of flocculants (such as
no real effect on the flow through the formation. clay stabilizers).

100 9000
80 8000
% of original permeability
1980 psig
50)
40)
60 7000
(∆p/L = 2 5 10 20
(∆p/L = 2 5 10 20
Low drawdown
Low drawdown
0 psig
High drawdown
6000
(∆p/L = 40)
40
High drawdown
(∆p/L = 50)
CaSO4 (mg/L)
5000
20
4000
0
Berea sandstone Oil well cores
(Miocene, Pliocene) 3000
Figure 14-8. Effect of drawdown pressure during cleanup

on the permeability regained by damaged sandstones 2000
(Krueger et al., 1967). ∆p = pressure difference,
L = length.
1000
Excessive drawdown can also decrease pore
pressure near the wellbore to such an extent that 0
the effective stress exceeds the formation rock 0 100 200 300
compressive strength. This phenomenon is obvious NaCl (g/L)
in poorly cemented sandstones, where wellbore
filling by formation sand is progressive (Stein and
Figure 14-9. Pressure effect on gypsum scale solubility at
Hilchie, 1972; Stein et al., 1974; Antheunis et al., 95°F [35°C] (Fulford, 1968).
1979). It is less apparent in chalks, where forma-
tion compaction and significant porosity reduction affect only the production string and surface
affect the near-wellbore region without any occur- equipment. However, they can reduce formation
rence of formation debris in the produced fluids permeability. Seeds, such as high-surface-area
(Morita et al., 1984; Blanton, 1978; Ben Marek, clays, promote the deposition of organic materials
1979; Van Ditzhuijzen and de Waal, 1984). Chalk (especially asphaltenes) (Rogers, 1976) or the pre-
compaction is four- to eightfold greater when soft cipitation of supersaturated salt solutions.
waters are injected during completion (Newman, Common scales are calcium carbonate (Vetter
1983). Formation breakage is particularly damag- and Kandarpa, 1980; Nancollas and Sawada, 1982;
ing in hydraulically fractured wells, where creeping Gudmundson and Oritz-R., 1984) and calcium sul-
inside the proppant pack results in large-scale fate (Vetter and Phillips, 1970; Oddo et al., 1991;
drops in production (Strickland, 1985). It has to be Lejon et al., 1995). Problems associated with the
emphasized that once created, this type of damage deposition of elemental sulfur (Kuo, 1972), sodi-
cannot be removed by matrix treatments. um chloride (Place and Smith, 1984) and barium
Reduction in the pore pressure during produc- sulfate (Wat et al., 1992) have also been described.
tion (Fig. 14-9) (Fulford, 1968), and sometimes Commingled precipitation of asphaltene and calci-
cooling resulting from gas expansion, results in um carbonate is common (Efthim et al., 1990).
the precipitation of organic (Hirschberg et al., Special cases of spalling or sloughing of parti-
1984; McClafin and Whitfill, 1983; Schantz and cles from brittle formations are also problems.
Stephenson, 1991; Singhal et al., 1991; Leontaritis These particles are derived from tensile failure cre-
et al., 1992; Takhar et al., 1995; Piro et al., 1995) ated during flow that results in chips of the forma-
or inorganic materials. Generally, these deposits

tion spalling into the perforations or the wellbore. understanding what the response of the rock (for-
Brittle formation behavior is seen typically in hard mation material) will be to any fluid. The relation
dolomites, brittle shales and a few shaly sand- between the rock and the treating fluid depends on
stones. Brittle formation failure is generally created the minerals present and the position of the miner-
by stresses in the formation and confining stresses als within the rock matrix. The analytical tech-
caused by depletion. Prevention of these stresses is niques used to characterize the mineralogy are
difficult without repressuring the reservoir. X-ray diffraction (XRD), SEM and thin-section
analysis.
• Retrograde condensation and two-phase flow
XRD analysis provides rapid and accurate iden-
Retrograde condensation and bubblepoint prob- tification of the crystalline material of the rock
lems are relative permeability blocking problems. matrix. Each crystalline material has a specific
Retrograde condensation is the condensation of a XRD pattern. The types and quantities of clays
liquid from gas. When this happens, the relative and feldspars can be qualitatively determined
permeability to gas can be reduced substantially. using XRD. Crystalline scale deposits can also
In oil reservoirs produced below the bubblepoint be identified using XRD.
pressure, free gas is formed, which reduces the rel- SEM provides information on mineralogy and
ative permeability to oil. morphology and the size of pore-lining materials.
Quantitative elemental analysis and mineral identi-
fication can be achieved by using this technique in
14-5. Laboratory identification and conjunction with energy-dispersive spectropho-
treatment selection tometry (EDS). The primary advantages of SEM-
EDS analysis over light microscopy are the depth
The objectives of laboratory experiments are to iden- of focus and magnification. The techniques are
tify potential damage and aid selection of the opti- useful for observing clay platelet structure and
mum treatment fluid and design. To achieve these analyses. The structures of smectite, kaolinite,
objectives, the formation material (cores), produced chlorite and illite are shown in Fig. 14-4.
fluids and damaged material must be analyzed. Thin-section analysis is used widely to study
Definitive core flow studies and solubility tests are rock structure and quantify minerals. In addition,
usually required to identify the source of damage and cementing minerals and the types and location of
to help determine the best procedure for damage pores can be identified. The rock is impregnated
removal. with a blue-colored resin to fill the interconnected
porosity. A thin (approximately 30 µm in thick-
ness) slice is cut perpendicular to the bedding
14-5.1. Damage identification
plane, and the surfaces are polished. Using a
• Core analysis polarized microscope, the minerals can be
The detailed analysis of formation cores is observed by transmitted light because they have
required to design the damage removal treatment. characteristic optical properties. The pore structure
It is difficult to determine formation mineralogy is easily identified by the blue resin.
without the use of cores (sidewall or conven- • Formation wettability
tional). Conventional cores are recommended to
Most formations (sandstone or carbonates) are
complete the analysis because sidewall cores can
water-wet. Occasionally, oil-wet formations are
be contaminated with drilling fluids and may not
encountered, especially when the produced oil is
be representative of the formation. If sidewall
a low-gravity oil. In some situations, the formation
cores are used, the analysis should be conducted
appears to be oil-wet because of the produced oil
on duplicate cores.
and the natural surfactants present in the oil; how-
• Formation mineralogy ever, when the oil is removed using appropriate
The formation mineralogy is an important param- solvents, the formation may be water-wet. When
eter affecting stimulation success. Knowledge of the oil adheres to the rock matrix strongly, it must
the petrography of the formation is essential to be removed prior to mineralogy or reactivity test-

ing. If the formation material is coated with oil, it The porosity of the rock sample can be deter-
should be cleaned with an aromatic solvent such mined using one of several techniques. The sim-
as xylene until all traces of oil have been removed plest technique for the determination of effective
and followed with a xylene- and water-miscible porosity uses Boyle’s law; the pressure of nitro-
solvent such as methanol. gen is determined in a constant-volume cell,
The simplest test to determine formation wetta- with and without the core. The total porosity is
bility is to take approximately 10 cm3 of formation derived by bulk and matrix density measure-
material and place it in the produced brine to equi- ments with a helium pycnometer. When required,
librate for approximately 30 min. The formation the pore-size distribution can also be measured
material is then placed in an oil (such as kerosene) using a mercury intrusion porosimeter. The size
and observed. To accentuate the test results, red and number of pores can be calculated and the
dye can be added to the clear oil to aid identifica- microporosity can be estimated. The micro-
tion of the oil adhering to the formation material. porosity can be more accurately determined
After it is allowed to equilibrate for an additional using the Brunauer, Emmett, Teller (BET) gas
30 min, the formation material is added to a fresh adsorption technique if required.
aqueous solution. Strongly water-wet formations
– Permeability
or other fines disperse readily in aqueous fluids but
agglomerate or clump together in the clear oil-base Permeability, an intrinsic characteristic of the
fluids. Conversely, oil-wet particles disperse in oil rock, is a measure of the rock’s capacity to
but agglomerate in water-base fluids. The surface transmit fluids. The measurement is usually
is water-wet if the contact angle of the fluid with made with gas (e.g., nitrogen [N2]) or liquids
the formation material is less than 90°; the surface (e.g., brines and oils).
is oil-wet if the contact angle is greater than or Permeabilities must be determined using sim-
equal to 90°. Wettability can exist in various degrees ulated downhole temperature and stress condi-
between extremely water-wet and extremely oil-wet. tions. In certain stress-sensitive formations, per-
Intermediate wettability is difficult to identify and meability determined under 1,000-psi confining
describe, with contact angles greater than 80° but stress may be 1 order of magnitude higher than
less than 100°. the permeability determined at 10,000-psi stress.
The wettability test can also be used to deter- • Formation fluid analysis
mine if the desired treatment fluid is water-wetting
or oil-wetting and how the treatment fluid may Analysis of the formation brine and oil can aid in
affect the desired natural wettability. The cleaned determining the types of damage that may be present.
formation material is treated as described previ- – Brine
ously except the formation material is placed in
Analysis of the formation brine can be used to
the desired treatment fluid instead of naturally pro-
predict scale formation. Common ions are listed
duced brine. Changes in wettability resulting from
in Table 14-7. Their presence can be determined
the selected treatment fluid can be detected using
using standard laboratory wet-chemical or
this method.
instrumentation techniques.
• Petrophysical characterization
Table 14-7. Common ions in formation brines.
Core porosity and permeability should be mea-
sured before performing a core flow evaluation. Cations Anions
– Porosity Sodium (Na+) Chloride (Cl–)

Potassium (K+) Bicarbonate (HCO3–)
Porosity is the ratio of the void space volume
to the bulk volume of the rock material. It is a Calcium (Ca2+) Carbonate (CO32–)
measure of the volume occupied by oil, gas and Magnesium (Mg2+) Sulfate (SO42–)
other fluids in the reservoir. Total, effective and Barium (Ba2+)
residual porosities are defined in Chapter 1. Strontium (Sr2+)
Iron (Fe2+ and Fe3+)

Calculations are available to predict scaling duced oil used for the analysis, the percentage of
tendencies. The intermixing of produced brines paraffin is determined.
with spent acid can result in the formation of
insoluble precipitates. Analysis of these brine
data aids in selecting the treatment fluids and 14-5.2. Treatment selection
their sequence.
The solubility of the formation or damage material,
– Oil treating fluid compatibility and core flow studies
Analysis and testing of the produced oil can also should be conducted to aid designing the best treat-
help stimulation treatment design. The use of ment for damage removal and to select the chemical
treating fluids that are incompatible with the products that are the most compatible.
produced oil can result in the formation of emul- • Solubility tests
sions and sludge that will subsequently result
– Formation material
in an unacceptable response to the stimulation
treatment. The oil may contain paraffins and Calcite, dolomite and ankerite are soluble in HCl
asphaltenes that can interact with the treatment and mud acid systems. Clay and silt are soluble
fluid. The quantity of various fractions of only in mud acid systems. Because of their high
asphaltenes and paraffins and their ratio to each surface area, clays and other fines are much
other are used to assess the possibility of organic more reactive with mud acid than sand grains
precipitation damage. It is also imperative that are. The total solubility of the formation material
the identity of deposits recovered from a well be is the sum of the solubility of each mineral in
determined. This is particularly true of organic the formation sample. Minerals other than car-
deposits because treatments for paraffin and bonates are also soluble in HCl; therefore, solu-
asphaltene removal are different. bilities should be used with caution. These min-
A test of the API gravity is performed on the erals include sulfates (e.g., anhydrite), iron
produced oil to evaluate sludging tendencies. oxides and halite. The solubilities of common
Certified hydrometers are used to measure the minerals are shown in Table 14-8.
specific gravity as a function of temperature. Solubility tests are performed under ideal lab-
For analysis of the asphaltene content, the oratory conditions and therefore exhibit the
produced oil is centrifuged to separate out emul- maximum formation solubility. The structure of
sified aqueous fluids. Mixing the crude oil with the rock and the position of each mineral in rela-
pentane and centrifuging the sample enables tion to the flow paths in the rock matrix may
preferential separation of asphaltenes from the result in different solubilities during actual
produced fluid. Repeated extraction is performed acidizing operations. Solubility determined in
until all oil has been removed. The precipitated the laboratory is not a definitive value for the
asphaltene material is collected, and standard maximum solubility that may result during the
laboratory techniques are used to obtain the con- acidizing process, but it provides guidelines as
tent by weight. The oil pentane fraction is saved to which treating fluids are most applicable.
and the pentane solvent is evaporated. The resid- A combination of solubility test results and
ual oil is then used to complete the analysis for XRD is commonly used to estimate the carbon-
the paraffin content. ate, silt and clay minerals and other mineral con-
For analysis of the paraffin content, a sample tents in the rock matrix. Determination of the
of vigorously mixed crude oil and acetone is cen- acid solubilities of the various materials in HCl
trifuged to enable preferential extraction of the and mud acid determines the total solubility.
paraffin material into acetone. Repeated extrac- When the results are used in conjunction with
tion and holding at temperatures less than 32°F XRD analysis, the composition can be deter-
[0°C] result in precipitation of the paraffins. The mined by the following procedure.
acetone mixture is filtered to remove the paraf-
1. The total solubility in HCl is typically used
fins, which are dried to remove residual acetone.
as an estimate of the total carbonate content.
On the basis of the total composition of the pro-

Table 14-8. Solubility of common minerals in acids.
Mineral Chemical Composition Solubility

HCl HF
Quartz SiO2 None Low
Feldspar
Microcline KAlSi3O8 None Moderate
Orthoclase KAlSi3O8 None Moderate
Albite NaAlSi3O8 Very low Moderate
Plagioclase (Na,Ca)Al(Si,Al)Si2O8 Very low Moderate
Mica
Biotite K(Mg,Fe2+)3(Al,Fe3+)Si3O10(OH)2 Low Moderate
Muscovite KAl2(AlSi3)O10(OH)2 Low Moderate
Clay
Kaolinite Al2Si2O5(OH)4 Low High
Illite (H3O,K) y (Al4 · Fe4 · Mg4 · Mg6)(Si8 – y · Aly)O20(OH)4 Low High
Chlorite (Mg,Fe2+,Fe3+)AlSi3O10(OH)8 Moderate High
Smectite (Ca0.5Na)0.7(Al,Mg,Fe)4(Si,Al)8O20(OH)4 · nH2O Low High
Mixed layers Kaolinite, illite or chlorite layered with smectite
Carbonate
Calcite CaCO3 High High †
Dolomite CaMg(CO3)2 High High †
Ankerite Ca(Fe,Mg,Mn)(CO3)2 High High †
Sulfate
Gypsum CaSO4 · 2H2O Moderate High
Anhydrite CaSO4 Moderate High
Other
Halite NaCl High High
Iron oxide Hematite (Fe2O3), goethite (α-FeO(OH)), High High
magnetite (Fe3O4), siderite (FeCO3)
†
Precipitation of CaF2
2. The solubility of the formation material in 5. Other minerals that are soluble in the acid
mud acid is used to determine the silt and should also be determined and confirmed
clay content (total fines) and the carbonate by XRD.
content.
– Scales
3. The difference between the solubilities in
mud acid and HCl is considered the approxi- The solubility of scale deposits depends on the
mate content of clay and fines. mineralogy of the rock. Tests similar to the tests
4. The total reactive silt content is calculated as performed to determine formation solubility can
the difference between the silt and clay con- be performed to determine the best solvent for
tent determined by solubility and the total scale removal. Identification of the deposit by
clay content determined by XRD. XRD prior to the solubility evaluation is recom-
mended to aid selecting the most active solvent.

Typically, acids, chelating agents or mechanical must be established. Additives such as corrosion
systems are used for removal. These tests should inhibitors, surfactants and mutual solvents must
be performed under simulated well conditions to also be included in the desired acid formulation
determine the optimum treatment. for compatibility testing. The concentrations of
these additives are established by core flow tests,
– Organic deposits
metallurgy and the well conditions that occur
Solubility tests at temperature and pressure can during treatment.
be performed to evaluate the most effective sol- Acid, or another aqueous treating fluid, is
vent to remove the plugging deposit. Typically, mixed with fresh crude oil (at low energy to
paraffin, asphaltene and tar deposits are soluble simulate downhole mixing) in a glass container
in aromatic solvents such as xylene or toluene; and observed at bottomhole temperature, if pos-
however, most deposits are a combined deposit sible. The characteristics and time for emulsion
and may require a combined treatment of an breakout are determined. The quanitity of aque-
organic solvent with acid and other surface- ous fluid breakout is determined at set time
active agents to improve the wettability of the intervals. The sample that provides the most
deposit and enhance the dissolving capacity of rapid emulsion breakout is desired. The acid sys-
the selected treatment fluid. Other solvents, such tem is modified, including the type and quantity
as mutual solvents and alcohols, may also aid of nonemulsifier, until the minimum aqueous
the removal of an organic deposit. breakout time is achieved. A clean oil/water
– Bacterial slime interface and rapid emulsion break (less than
10 min) are desired. The oil phase must also be
Bacterial slime is also removed by treating with clear, with no emulsified water draining freely
solvents. These deposits are soluble in water from the glass surface (signifying a water-wet
containing oxidizing agents such as sodium surface).
hypochlorite. If required, solubility tests can be Differences in emulsion stability are attributed
performed to evaluate the efficiency of the oxi- to more dominant factors such as stabilization
dizing agent. Tests should be performed using by solids precipitation (organic and inorganic),
simulated downhole conditions, if possible. ferric ion interaction with the asphaltene compo-
• Treating fluid–formation compatibility nent of crude oil and viscosity increase at the
interface. Results presented by Coppel (1975)
Preventing emulsion and sludge formation follow-
show that partially spent acids contain potentially
ing a damage-removal treatment requires an opti-
precipitable materials in solution. As the pH
mum fluid design to minimize or eliminate the
value of the spent acid increases during flowback
formation of precipitates. The potential for the for-
and mixes with other produced fluids, materials
mation of acid and produced crude oil emulsions
precipitate that can stabilize emulsions. Dunlap
and optimization of the de-emulsifier treatment are
and Houchin (1990) recommended using polar-
currently evaluated using American Petroleum
ized microscopy to evaluate return fluids to
Institute (API) Recommended Practices for
establish the cause for a stabilized emulsion.
Laboratory Testing of Surface Active Agents for
Specific damage mechanisms such as emulsion
Well Stimulation (RP 42) (1977).
stabilization by organic deposits, solids and iron
– Emulsion prevention testing may be detected and identified. Stabilization by
solids can be evaluated by the use of fine silica
Prior to a stimulation treatment, the compatibil-
or clay in the emulsion test described previously.
ity of the treatment fluid with the produced oil
Downhole emulsions can be attributed to
must be determined. Fine siliceous materials,
inadequate prejob compatibility testing because
ferric iron or both can be added to evaluate
the additives used in the acid system may prefer-
emulsion stability mechanisms. These fluid
entially adsorb onto the rock matrix or partition
interactions can produce downhole emulsion
to the oil phase during production. Alternate
problems. The quantity of additives used to con-
testing procedures have been recommended to
trol emulsion stability and sludge formation
simulate reservoir conditions more accurately.

Ali et al. (1994) recommended using an alter- tions. Therefore, the permeability changes depend
nate procedure that better simulates downhole on the dissolution and precipitation reactions that
intermixing of the stimulation fluids and the pro- occur. Observations that indicate what dissolves
duced crude. The formulated acid containing and what precipitates are extremely useful in
inhibitor, surfactants, mutual solvents and non- selecting the best treatment fluid.
emulsifiers is filtered through a sand pack con- The effects of sequential injection of the different
taining 10% silica flour, 10% clays (such as treatment fluids can also be observed. Fluid injec-
montmorillonite) and 80% 100-mesh sand prior tion rates range from 0.2 to 10 mL/min and pres-
to completion of the emulsion test. This proce- sures range up to 1500 psi. Flow rates should be
dure simulates the filtration of the acid system selected to ensure that the fluid movement has mini-
within the rock matrix. Surfactants, such as mal effect on the movement of fines contained
those used in nonemulsifiers, are adsorbed onto within the pore structure. The 1-in. diameter, 12-in.
the mineral surfaces. This adsorption minimizes long cores are placed in a core holder and confined
the availability of the surfactants in solution to under pressure to simulate reservoir stresses. To
prevent emulsions. ensure fluid flow through the core, the confinement
pressure should be greater than the pressure
– Sludge prevention testing
required to initiate flow through the core. Tests
Prior to the stimulation treatment, the compati- should also be performed with backpresssure. For
bility of the treating fluid with the produced oil core flow studies of acids, a backpressure of at least
must be determined for sludging tendencies. 1000 psi is required to maintain the CO2 produced
Interaction of the acid with crude oil can gener- by acid dissolution of carbonate deposits in solu-
ate the formation of solid precipitates, although tion. For scale and deposit removal, sufficient back-
the system does not form stable emulsions. pressure is required to prevent vaporization of the
These precipitates are sludge and are insoluble treating fluid. The diagram of the core flow appara-
in the hot formation oil or brine. The sludging tus is shown in Fig. 14-10.
tendencies are aggravated by intermixing with The results of these tests are shown by an acid
ferric iron. If not prevented, precipitated solids response curve (ARC), as illustrated in Fig. 14-11.
can result in decreased production following an The evolution of permeability versus the volume
acid treatment. of treating fluids is determined and the effect of
The sludging evaluation is similar to the emul- each fluid on the core permeability is calculated
sion test previously described. The oil should be and displayed. Core holders utilizing multiple pres-
free of solids and emulsified water. Following sure taps can be used to examine the effect of each
completion of the emulsion test, the acid and oil fluid as it penetrates deeper into the formation.
mixture is maintained at the bottomhole temper- Permeability changes depend on dissolution and
ature for 1 to 4 hr. The mixture is then filtered precipitation reactions. Tests to determine what
through a 100-mesh screen, and the precipitated dissolves and what precipitates are not used to
sludge on the filter is identified. determine treatment volumes, which depend on
To eliminate the formation of sludge, the acid the type and extent of damage. If formation cores
system must contain an antisludge additive. The are used for the study, they should be cleaned with
concentration of antisludge additive is increased aqueous alcohol or ethylene glycol monobutyl
in the acid system until formation of the sludge ether solutions to remove traces of oil and ensure
precipitate is eliminated. that the cores are water-wet.
• Core flow tests
Core flow tests are used to determine the effects of
treatment fluids on formation samples at simulated
14-6. Treatment strategies and
well treating conditions. The structure of the rock concerns
and the position of each mineral in relation to the Once the damage and its origin have been character-
flow channels in the rock matrix may result in dif- ized, the correct remedial action can be taken. Var-
ferent solubilities during actual acidizing opera- ious types of damage can coexist because almost

Reservoirs
Water CO2 N2 Vacuum N2 (air)
Confinement
pressure
regulator
V17 Differential pressure

cell
V7 V8
N2 pressure
V10
G2
Cell
V9
mode
V1 selector Backpressure
V6 V4
Fluid
G3 G4
selector Confinement
V16
Flow pressure V3 pressure
Separator
Pulsation dampener
V11
G1
Separator
V15 Rupture
V2 disk
Volumetric Back-
pump V14 pressure
regulator V12
Rupture
disk V13
V5
Filter Gas Temperature
flowmeter regulator
Figure 14-10. ARC core flow test apparatus.
2000
ARC test Sample: Test conditions Fluids used during the test
Field: Temperature (˚C): 65 • 3% NH4Cl
Well: Backpressure (b): 20 • Mud acid
Location: Formation: Calcareous sandstone
1600 Date: Depth:
3% NH4Cl
Permeability (md)
1200
800
Mud acid
400
3% NH4Cl
0
0 300 600 900 1200 1500
Volume (mL)
Figure 14-11. ARC of a carbonate-cemented sandstone to mud acid.

every operation performed on the well (drilling, the reservoir at “matrix” (subfracturing) rates and
completion, workover, production and stimulation) pressures. These relatively low rates and pressures
is a potential source of damage (Krueger, 1986). are necessary to remove the damage located in the
The efficiency of a matrix treatment in sandstones near-wellbore area. The flow rate is also limited to
depends primarily on removing the damage affecting prevent fracturing of the formation, which would
productivity or injectivity (Williams et al., 1979). result in the loss of treatment fluid behind the dam-
This restriction is usually shown by an overall lower aged zone.
level of or sharper than expected decline in produc- Inexpensive and readily available inorganic acids,
tion. The extent of damage is typically estimated such as HCl or hydrofluoric acid (HF), are used to
using pressure transient analysis. dissolve some of the damaging materials, rock con-
The physical characteristics, not the origin, of the stituents or both. A certain volume of acid is pumped
damage determine the treating fluid. A fluid can be into the formation to restore near-well permeability
used to treat occurrences of the same type of dam- (sandstones) or to increase rock permeability (car-
age, regardless of what caused the damage. Seven bonates).
basic types of damage are shown in Fig. 14-12. Treatment strategies for the various formation
When formation damage has reduced the produc- damages discussed previously are reviewed in the
tivity of a well, matrix acidizing is usually the appro- following sections. In addition, strategies for some
priate treatment, though reperforating with deeper common types of wellbore damage are discussed.
penetrating holes may be a ready alternative for shal-
low damage. Typically, formation damage is associ-
ated with partial plugging of the rock matrix around 14-6.1. Fines and clays
the wellbore. The objective of matrix acidizing is to • Migrating fines
remove the damage or bypass it by creating chan-
The treatment of moveable fines can be accom-
nels, such as wormholes. When matrix acidizing or
plished by either prevention (using a clay-control
reperforating is not possible, a short proppant frac-
process) or removal. Removal of migrating fines
turing treatment can be an alternative.
in sandstone formations is best accomplished by
Matrix acidizing treatments remove damage by
treatment with a fluid containing HF and HCl mix-
injecting reactive fluids into the natural porosity of
Types of damage
Emulsions Wettability change Water block Scales Organic deposits Mixed organic/ Silts and clays
Mutual solvent ± Mutual solvent ± Aromatic solvents inorganic deposits Depend on
de-emulsifier water-wetting Carbonates Solvent in acid mineralogy
surfactant CaCO3 dispersion
High temperature (>250°F): Aqueous acetic acid
Low temperature (<250°F): HCl
FeCO3 – HCl + reducing/sequestering agents
Oil well Gas well Sulfates

Acid + alcohols CaSO4 – EDTA
BaSO4 – EDTA
SrSO4 – EDTA
Fe scales
Low to medium temperature (≤250°F) High temperature (>250°F) FeS – HCl + reducing + sequestering
Acid + solvent Nonaqueous acetic acid Fe2O3 – HCl + EDTA
Acid + surfactant FeCO3 – HCl + reducing/sequestering
Chloride
NaCl – H2O, 1%–3% HCl
Hydroxide
Mg(OH)2 – HCl
Ca(OH)2 – HCl
Silica – mud acid
Figure 14-12. Treatment selection and the nature of damage.

tures—these are the commonly used mud acids. drawdown and control sand movement by both
Deeply penetrating acid systems, containing fluo- pressure reduction and use of the proppant at the
boric acid, show a good possibility for particle interface contacts of a gravel pack as an “in-
destruction and extend some potential for clay sta- formation” gravel pack.
bilization. Fracturing the formation is also a treat- The treatment of spalling problems is extremely
ment possibility because the effect from linear difficult. Propped fractures may help contain the
flow in the walls of the fracture has a less detri- formation and spread out the drawdown to reduce
mental effect on production than inward radial the spalling force, although totally halting spalling
flow in an unfractured well. The success of both may be impossible. One of the keys to treatment
clay control and fines removal depends on the selection is whether the spalling is caused by high
depth extent of the fines movement problem. In initial pressures that will quickly deplete or by
many cases, tip-screenout (TSO) fracture design cyclic mechanical loads that will recur. If high ini-
using a short fracture for damage bypass is a better tial pressure is the problem, a cleanout may suf-
alternative. fice. If cycling is the problem, a permanent control
HCl systems are typically used to remove fines method is the best solution. Control methods
damage in a carbonate formation. Because the include gravel packing, fracture packing, selective
fines are not dissolved, but are dispersed in natural perforating (along the fracture axis) and some
fractures or the wormholes that are created, N2 is plastic-bonding methods.
usually recommended to aid fines removal when
the well has a low bottomhole pressure.
14-6.2. Scales
• Swelling clays
The removal of smectite is usually accomplished Various solvents dissolve scales, depending on their
with HF or fluoboric acid, depending on the depth mineralogy. The most common treatments for the
of penetration. In the event of deep clay-swelling scales in a well are as follows:
problems (more than 2 ft), the best treatment is • Carbonate scale (CaCO3 and FeCO3)—HCl will
usually a fracture to bypass the damage. readily dissolve all carbonate scales if the acid can
• Unconsolidated formations penetrate to the scale location (Tyler et al., 1985).
Two basic problems determine the method of • Gypsum (CaSO4 ⋅ 2H2O) or anhydrite (CaSO4)—
treatment for unconsolidated formations. If the for- These calcium sulfate scales are removed with
mation moves as discrete large particles (i.e., the compounds that convert the sulfate to a hydroxide
building blocks of the formation are moving), then or other ion form followed by acid or by direct
the problem is a lack of cementation between the dissolvers such as ethylenediamenetetraacetic acid
grains for the applied production forces, and the (EDTA) or other types of agents. Following a cal-
formation is classified as a low-strength formation. cium sulfate dissolver with acid may double the
Treating low-strength formations can be difficult if amount of scale dissolved because most scales are
the cementing materials are reactive with the fluid mixtures of materials and HCl has some ability to
that is injected to remove formation damage or to dissolve the finest particles of calcium sulfate. The
improve permeability. Fortunately, the cementing tetrasodium salt of EDTA is preferred because its
materials in most formations have a small surface dissolution rate is greater at a slightly alkaline
area and are less reactive with acids than with pH value; the more acidic disodium salt has also
fines or clay particles in the pores of the rock. been used, as well as other strong sequestrants of
When formations expel large grains into the the same family, although they do not show a
wellbore, it may be beneficial to add additional marked difference from the EDTA performance.
perforations to reduce the velocity across the sand- Care must be used not to over-run the spent scale
face or to design a fracture to reduce the draw- dissolver or converter solutions with acid because
down. (It is common to fracture formations with massive reprecipitation of the scale will occur.
permeabilities higher than 100 md.) These frac- • Barite (BaSO4) or celestite (SrSO4)—These sulfate
tures are usually TSO designs that provide short, scales are much more difficult to remove, but their
highly conductive fractures that can reduce the occurrence is more predictable. Barium and stron-

tium sulfates can also be dissolved with EDTA if 14-6.3. Organic deposits
the temperature is high enough and contact times
are sufficient (typically a 24-hr minimum soaking Organic deposits are usually resolubilized by organic
time for a 12,000-ft well with a bottomhole tem- solvents. Blends of solvents can be tailored to a par-
perature of about 212°F [100°C]; Clemmit et al., ticular problem, but an aromatic solvent is an effi-
1985). Barium and strontium sulfate removal cient, general-purpose fluid. Environmental concerns
methods are usually mechanical. Most chemical have led to the development of alternative solvents
removers are only slightly reactive, especially in (Samuelson, 1992).
thick deposits, but mixtures of barium sulfate and Paraffin removal can be accomplished using heat,
other scales can usually be removed by properly mechanical scraping or solvents. Heating the tubing
formulated dissolvers with sufficient soak times. with a hot oiler may be the most common type of
Thick deposits should be removed by mechanical treatment. It may also be the most damaging and
or abrasive methods. Care must be exercised when least effective in some cases. Injection of hot oil
analyzing well debris to avoid mislabeling barite from the surface will melt the paraffin from the walls
from drilling mud residue as barium sulfate scale. of the pipe, but the depth to which the injected fluid
stays hot is a function of the well configuration. If
• Sodium chloride (NaCl)—Sodium chloride scale the well is allowed to circulate up the annulus while
is readily dissolved with fresh water or weak the hot oil is injected down the tubing, the heat will
acidic (HCl, acetic) solutions. Redesigning the not penetrate more than a few joints of tubing from
mechanical system to avoid heat loss and water the surface. The heat is quickly transferred through
drop-out are also treatment possibilities. the steel tubing to the fluids rising in the annulus and
• Iron scales, such as iron sulfide (FeS) or iron oxide little, if any, heat reaches deep in the well. As the hot
(Fe2O3)—HCl with reducing and sequestering oil cools, the paraffin picked up in the upper part of
(EDTA) agents dissolves these scales and prevents the well can precipitate. If hot oiling is required at
the reprecipitation of by-products, such as iron depths greater than 150 ft, an alternate method of
hydroxides and elemental sulfur (Crowe, 1985). placement must be used. Deeper application of heat
Soak times of 30 min to 4 hr are usually beneficial is available with other processes that feature heat
in removing these scales when using acid. Where generation as part of an exothermic chemical reac-
iron sulfide is a thick deposit, mechanical action such tion. The processes require close control and are
as milling is suggested. Water jetting typically will generally expensive.
not cut an iron sulfide scale except where it is dis- Mechanical scraping can be useful in cases where
persed with other scales or exists as a thin coating. extensive deposits of paraffin must be removed rou-
• Silica scales—Silica scales generally occur as tinely. Scraping is usually accomplished with slick-
finely crystallized deposits of chalcedony or as line and a cutter. In wells that utilize a rod string,
amorphous opal and are readily dissolved by HF. placing scrapers on the string may automatically
• Hydroxide scales: magnesium (Mg(OH)2) or calci- scrape the tubing walls.
um (Ca(OH)2) hydroxides—HCl or any acid that Solvent treating to remove paraffin may be based
can sufficiently lower the pH value and not precip- around a straight- or aromatic-chain solvent. The
itate calcium or magnesium salts can be used to most appropriate solvent depends on the specific
remove these deposits. paraffin and the location of the deposit. Heat (at least
to 130°F [55°C]) and agitation significantly increase
Contact time is an important factor in the design the rate of removal.
of a scale removal treatment. The major concern in Removal treatments for asphaltenes use aromatic
treating scale deposits is allowing sufficient time for solvents such as xylene and toluene or solvents con-
the treating fluid to reach and effectively dissolve the taining high percentages of aromatics. Naphtha is
bulk of the scale material. The treating fluid must usually not effective as a solvent. Some materials
dissolve most of the scale for the treatment to be being tested provide dispersant benefits without sta-
successful. bilizing the entire mass of the asphaltene. Solvent
soak time, heat and agitation are important consider-
ations for treatment.

14-6.4. Mixed deposits 14-6.7. Induced particle plugging

Combined deposits require a dual-solvent system, • Mud solids
such as dispersion of a hydrocarbon solvent (usually To remove shallow mud damage in natural frac-
aromatic) into an acid. tures, a solvent or cleaner that will disperse the
mud should be selected on the basis of tests of a
field sample of the mud. Energizing the fluid with
14-6.5. Emulsions N2 can assist in removing large masses of drilling
Emulsions are stabilized by surface-active materials mud from a fracture system. Experience with
(surfactants) and by foreign or native fines. Gener- drilling mud cleanup from natural fracture systems
ally, mutual solvents, with or without de-emulsifiers, shows that slugs of drilling mud may flow back
are used for treating emulsion problems. De-emulsion initial treatment, and damage can often reassert
fiers, which may work well in a laboratory or in a itself as mud moves from the outer reaches of the
separator or tank because of the large number of fracture system into the wellbore. This condition
droplets in contact per unit volume, may not work can require repeated treatments of the same high-
by themselves in a porous medium because of mass- efficiency cleaner, plus N2, to get good cleanup of
transport phenomena in getting the product to where the well. Acid may help, but tests of the acid’s
it should work. Another reason they may not work effect on the field mud sample are required.
alone is the mechanism involved in breaking emul- When extremely large volumes of heavyweight
sions, which should provoke the coalescence of mud are lost, it may be beneficial to sidetrack the
droplets and then phase separation. well and redrill the pay zone. Whenever possible,
Asphaltic iron-catalyzed sludges are the most dif- the drilling mud overbalance should be minimized,
ficult emulsions to break. These emulsions are cat- and the mud should be conditioned to reduce
alyzed by dissolved iron in the acid or water and solids before the pay zone is drilled. Experience
resemble a crosslinked oil polymer in some with drilling highly fractured formations has led
instances. Prevention is the best treatment. An effec- to experimentation with underbalance drilling in
tive antisludge treatment for the area and an iron- some zones. Underbalance drilling can result in
reducing agent in the acid are the best methods. only minimal damage in producing wells in com-
Removal of an existing asphaltene sludge is usually parison with the damage created by traditional
accomplished by dispersing it in a solvent and drilling methods. There are dangers, however, in
attacking the components of the sludge with addi- underbalance drilling, and the risk versus benefit
tives designed for cleanup and removal. must be evaluated carefully.
Mudcakes are usually damaging only in open-
hole completions without significant fractures
14-6.6. Bacteria (Burton, 1995). In vertical wells, they are usually
Prevention of polymer destruction by bacteria is usu- easily mechanically removed to a great extent by
ally handled with biocides and tank monitoring. pressure drawdown. In long horizontal wellbores,
Control of bacteria downhole is more difficult and the necessary drawdown is almost impossible to
involves scraping or treatments with sodium impose on any section other than the heel, particu-
hypochlorite or other oxidizers followed by acidizing larly when a compressible fluid is in the hole.
and then treatment with an effective biocide at a Circulations for mudcake removal should be con-
level at least 1.2 times the minimum kill level. ducted with minimum clearance between the wash
Frequent rotation of the type of biocide is also neces- pipe and the borehole to promote turbulence.
sary to prevent the development of biocide-resistant Residual mudcake in prepacked screens or slotted
strains of bacteria. liners completions is particularly problematic
because it can plug the screen (Browne et al.,
1995; Ryan et al., 1995).

• Dirty fluids cial. The treatment volumes depend on the fluid,

When particle damage is known to have occurred formation and amount of load fluid lost.
because of the use of unfiltered or poor-quality flu- • Waterfloods
ids, cleanup depends on finding a solvent or acid
The removal of particles injected during water-
that can either remove the particles or break the
flood operations depends on the identity of the
structure of the bridges formed in the formation or
material and use of a cleaner and an acid to dis-
fracture system. Surfactants, acids and mutual sol-
perse the material. One of the best techniques for
vents are usually the most beneficial materials.
cleaning up injection wells or disposal wells is to
The addition of N2 to provide a high-energy boost
backflow the well as hard as possible prior to the
may also be beneficial.
treatment. This usually removes enough mass
The decision of which surfactant or mutual sol-
from the wellbore to eliminate the need for stimu-
vent to use should be based on core tests or field
lation. However, if backflowing does not ade-
response. Including a gas such as N2 or CO2 is
quately clean the wellbore, acid and a mutual sol-
based on fluid and solids recovery requirements
vent in volumes ranging from 50 to 100 gal/ft are
and wellbore unloading ability. For designing
usually necessary. When large amounts of solids
cleanup operations for particulate damage, flowing
are expected, the well should be backflowed after
the well back quickly after the treatment helps in
acidizing. If oil carryover and emulsions are the
the removal of the particles. Lower pressure for-
problem, acid and a mutual solvent can be injected
mations may require a gas boost. In higher pres-
and displaced permanently with injection water
sure formations, natural flow is usually adequate
behind the acidizing job.
to unload these solids, especially when a properly
designed fluid has been used and the solids are no
deeper than the surface of the wellbore face. 14-6.8. Oil-base drilling fluids
Mechanical scraping and cleaning can exert influ-
ence only as far as the wellbore wall. The prevention of OBM emulsions is relatively easy.
Either a surfactant-base cleaner that is mixed after
• Acidizing specific OBM testing or a more general xylene wash
The leading edge of an otherwise effective mutual of the zone must be done before contact with either
solvent and acid system can be loaded with debris high-salinity brine or acid. After the cuttings and
cleaned off the walls of the tanks and tubing. For mud fines have been cleaned and totally water-
this reason, the leading edge of the acid job is usu- wetted, the remaining damage problems of wettability
ally circulated out of the well using a process can be reversed with a formation cleaner or mutual
called pickling the tubing. In this treatment, acid solvent. Acid is usually used as a following stage
and solvents are injected down the tubing to dis- after cleaning to remove mud particles and clean up
perse and dissolve iron, pipe dope, mud and other formation debris. Removal of known OBM emul-
debris from the tubing and are then circulated or sions resulting from mixing with high-salinity brine
reversed out of the well without being injected or acid usually requires an aromatic solvent wash or
into the formation. These jobs are extremely effec- a specialized surfactant treatment that targets the silt-
tive when the tubing has not been cleaned or its stabilized emulsion. Evaluation of any cleanup
condition is unknown. Volumes of both acids and mechanism or treatment using laboratory samples of
additive treatments range from 1 to 21⁄2 tubing vol- OBM should be avoided. Only field samples of the
umes depending on the condition of the tubulars. mud are appropriate for designing the removal treat-
Minimum acid and solvent volumes typically range ment. Treatment fluid volumes range from 15 to
from 250 to 500 gal. Coated tubing can reduce iron 50 gal/ft of aromatic solvent or surfactant mixture,
scale significantly, but other contaminants, such as and the agitation and soak times are critical to the
scale and pipe dope, may still be present. success of the treatment. Application difficulties
If load-fluid recovery influences well produc- include trapping the treating fluids across the pay in
tion, surfactants or mutual solvents that reduce a column of heavyweight fluids where density segre-
surface and interfacial tension are usually benefi gation may be rapid. Packers and gelled plugs are
the first line of isolation.

14-6.9. Water blocks ability impairment in this case is the result of rock
compaction, acidizing is ineffective. Short prop-
Removal of a water block can be accomplished pant fracturing treatments are apparently the only
using a surfactant or alcohol applied as a preflush to cure, though extreme overbalance perforating may
reduce surface tension, followed by a postflush of N2 give positive results in some cases (Petitjean et al.,
or CO2 to remove the water from the near-wellbore 1995).
area and reestablish gas saturation. Once the water
has been mixed with the surface-tension-lowering • Pipe problems
materials, removal is easier. The difficulties in this Whenever well production is reduced, the first
type of operation are placement of the fluid and get- determination should be to establish that the tub-
ting an even distribution of the fluid around the well- ing is open and the lift system is working.
bore. Repeated treatments are usually necessary, and Numerous pipe problems from leaks to collapsed
selective injection devices are beneficial. pipe can occur, and fill in the tubing is also a pos-
sibility. Well conditions change over time, and an
effective completion at the start of the well’s life
14-6.10. Wettability alteration may not be effective after several years of produc-
Wettability alteration damage is removed by inject- tion as the reservoir pressure declines.
ing (mutual) solvents to remove the oil-wetting • Poor perforations
hydrocarbon phase and then injecting strongly water-
The usual treatment for poor perforations is to add
wetting surfactants. Again, a surfactant by itself will
additional perforations. In zones that are extremely
not work. The oil phase, which is usually precipi-
laminated, such as the shaly sands of the U.S. Gulf
tated asphaltenes or paraffins, must first be removed
Coast and other areas, 8 to 12 spf is considered
with a solvent. (The same applies to an adsorbed
adequate, but perforation breakdowns (i.e., small
oleophilic surfactant.) Then, a strongly water-wetting
fractures) may be required for complete linking.
surfactant can be injected and adsorbed onto the rock
Lower perforation density is possible if the well
minerals. This reduces the tendency for new hydro-
will be fractured. Fracturing will cross the barriers
carbon precipitates to stick to the mineral surfaces
of laminations, and in many field cases has pro-
and oil-wet them again.
vided extensive productivity increases.
For retrograde condensation problems, the most
Adding perforations is easy, but the typical 0°
appropriate treatment technique is the injection of neat
phased, small through-tubing guns deliver only
natural gas in a periodic “huff and puff” operation.
small holes and short penetrations. The newer
Condensate is picked up by the gas and transported
downhole-deployable guns that provide minimum
into the reservoir. Reprecipition requires the retrograde
clearance and phasing are preferred, especially
of the process after several months of production.
when hydraulic fracturing will be performed.
• Hydrates
14-6.11. Wellbore damage Hydrates are mixtures of ice and other elements,
• Mechanical damage from drilling principally natural gas, that may resemble a dirty
The drilling process itself modifies the local ice deposit. Unlike ice, they can form at tempera-
stresses around the wellbore, generating a zone tures greater than 32°F. The formation of hydrates
of reduced permeability in the near-wellbore area is usually associated with a drop in temperature or
(Dusseault and Gray, 1992). It has been shown a reduction in pressure that may accompany the
that such damage affects primarily soft formations production of fluids. Hydrates may also form in
where the difference between the minimum and gas-cut drilling mud, particularly when the mud
maximum stresses orthogonal to the wellbore is is circulated near the seafloor in cold locations.
large. In the worst cases, the extent of the perme- Hydrate plugging of chokes and valves can be
ability decrease can be as large as 21⁄2 wellbore a serious problem. Hydrate particle abrasion of
diameters (Morales et al., 1995), and perforations equipment is also possible.
do not bypass the damaged zone. Because perme- The most common occurrence of hydrates is
in gas wells with a small amount of water produc-

tion. The quantity of water relative to the quantity pores of the rock it is difficult to displace with
of gas production is critical. As the water cut low-viscosity fluids such as gas. Shutting off water
increases, many hydrate problems disappear. (water control) is a special technique and is dis-
Hydrates are prevented by adding a freezing-point cussed elsewhere in the literature.
depressant such as alcohol or glycol below the
• Microporosity
hydrate formation point. They may also be con-
trolled by temperature preservation in the pro- Microporosity is created by a number of clays and
duced fluid or the elimination of severe pressure a few minerals. It is simply a condition where a
drops that allow expanding gas to chill the liquids large surface area exists for water coating or trap-
to their freezing points. ping. Microporosity rarely presents a problem
except when it occupies the pore throat area of the
• Fill formation. In these cases, it may trap either debris
Debris from formation spalling into the perforation or water and obstruct flow. The removal of micro-
or wellbore can be one of the most serious detri- porosity can generally be accomplished with HF,
ments to production. Fill in the wellbore is easily or deep problems can be bypassed by fracturing.
identified with a sinker bar on the wireline and is
usually easily removed using tubing or coiled tub-
ing and N2 or foam unloading practices. Fill in the 14-7. Conclusions
perforations is more difficult to identify and much
more difficult to remove. When fill in the perfora- To maximize well performance, the paths from the
tions is suspected, reperforating the well is gener- formation to the pipeline must present the lowest
ally the most direct method of testing the theory pressure impedance possible. Achieving this condi-
and restoring the well to productivity. Where the tion requires both a well-designed completion and
fill is acid soluble, acid injection may be useful; the elimination of formation damage. The tools are
however, injecting acid into a perforation that is an array of stimulation and damage removal tech-
filled with small debris is usually difficult. niques and chemicals that are readily available. Still,
although formation damage removal seems easily
• Water problems achievable, the goal of a damage-free completion can
Water production is not only a major economic be elusive. The problem is not so much one of find-
problem in surface separation, but it also causes a ing a tool, but one of finding the right tool. A little
major reduction in the relative permeability of oil experience shows that the right tool can be selected
and gas. Water production from the well can lead only when the problem has been identified. The effort
to significant problems such as corrosion, back- and expense to understand the nature and identity of
pressure, emulsions and movement of the forma- the problem can be the wisest investments.
tion or fines. Water may flow from the bottom The most common damage mechanisms and sug-
(coning), rise through fractures or flow from the gested methods of removal are listed in Tables 14-3,
edge in fractures through the matrix or in high- 14-4 and 14-5. It bears repeating that many damage
permeability streaks. Because of its low viscosity, conditions have similar symptoms and that there are
water flows much easier than oil, and once in the no universal treatments.


◆
Additives in Acidizing Fluids
Syed A. Ali, Chevron U.S.A. Production Co.

Jerald J. Hinkel, Schlumberger Dowell
15-1. Introduction and their interactions with fluids. Clay minerals are
important because of their large surface area and
Proper fluid selection is critical to the success of because the surface carries an electrical charge.
a matrix treatment. Figure 15-1 shows the types of Although relatively minor, some surface charge
damage that may occur and what types of treatment results from ionic substitution and structural imper-
would be required. The treatment may be a failure if fections. For instance, the substitution of Al3+ for
the proper additives are not used. The treating fluid Si4+ in tetrahedral sheets or Mg2+ or Fe2+ for Al3+ in
is designed to effectively remove or bypass the dam- octahedral sheets is the origin of negative charges in
age, whereas additives are used to prevent excessive the lattice of smectites. These are unaltered by
corrosion, prevent sludging and emulsions, prevent changes in the solutions that contact them.
iron precipitation, improve cleanup, improve cover- Silicon and aluminum oxides, the principle ingre-
age of the zone and prevent precipitation of reaction dients of sandstone and clays, exhibit amphoteric
products. Additives are also used in preflushes and behavior. Therefore, the surface charges of these
overflushes to stabilize clays and disperse paraffins minerals depend on the pH of the solutions in con-
and asphaltenes. The functions of some of the addi- tact with them. The surface charge changes from
tives are discussed briefly in this chapter. positive to negative as the pH increases. The point
Throughout this chapter references are made at which the surface charge is neutral is known as
to the interactions of additives with the formation. the point of zero charge (pzc). Table 15-1 provides
Schechter (1992) provided an excellent discussion the approximate values of the pzc for several com-
of the chemical properties of formation materials mon minerals.
Types of damage
Emulsions Wettability change Water block Scales Organic deposits Mixed organic/ Silts and clays
Mutual solvent ± Mutual solvent ± Aromatic solvents inorganic deposits Depend on
de-emulsifier water-wetting Carbonates Solvent in acid mineralogy
surfactant CaCO3 dispersion
High temperature (>250°F): Aqueous acetic acid
Low temperature (<250°F): HCl
FeCO3 – HCl + reducing/sequestering agents
Oil well Gas well Sulfates

Acid + alcohols CaSO4 – EDTA
BaSO4 – EDTA
SrSO4 – EDTA
Fe scales
Low to medium temperature (≤250°F) High temperature (>250°F) FeS – HCl + reducing + sequestering
Acid + solvent Nonaqueous acetic acid Fe2O3 – HCl + EDTA
Acid + surfactant FeCO3 – HCl + reducing/sequestering
Chloride
NaCl – H2O, 1%–3% HCl
Hydroxide
Mg(OH)2 – HCl
Ca(OH)2 – HCl
Silica – mud acid
Figure 15-1. Types of damage.

Table 15-1. Point of zero charge of some

Fe2+ H2 H+
representative reservoir materials.
Material Point of Zero Charge
Quartz (SiO2) 2.7

Corundum (Al2O3) 9.0
Hematite (Fe2O3) 5.0 al
et
Calcite (CaCO3) 9.5 M
Kaolinite (AI4Si4O10(OH)8) 4.6

Montmorillonite
(Al,Mg,Fe)4(Al, Si)8O20(OH)4 · nH2O 2.0 Figure 15-2. A metal surface composed of both anodic
Albite (NaAlSi3O8) 2.0 and cathodic sites (Schechter, 1992).
Thus, the way in which additives interact with the on the metal surfaces, along with the reduction of
reservoir actually depends on the pH. Under the con- hydrogen ions and formation of hydrogen at the
ditions normally associated with the matrix acidiza- cathodic sites.
tion of sandstones, the pH of the acid is less than the The equations for an anodic reaction (oxidation):
pzc of the reservoir constituents, which causes the
formation materials to behave as positively charged. Fe → Fe2+ + 2e–
It is important to keep this concept in mind. metallic iron ionic iron electrons
and a cathodic reaction (reduction):
15-2. Corrosion inhibitors 2H+ + 2e– → H2

ionic hydrogen electrons hydrogen
The most important acid additives are corrosion
inhibitors. A corrosion inhibitor is a chemical that can be combined to show the overall reaction:
slows the attack of acid corrosion on drillpipe, tubing Fe + 2H+ → Fe2+ + H2
or any other metal that the acid contacts during treat- metallic iron ionic hydrogen ionic iron hydrogen
ment. Brief explanations of corrosion mechanisms,
The unit of measurement used to determine the
corrosion inhibition and techniques for evaluating
amount of corrosion is lbm/ft2. Acceptable corrosion
inhibitor performance are presented in this section.
limits for low-alloy steel are shown in Table 15-2.
The acceptable range is from 0.02 to 0.09 lbm/ft2
15-2.1. Corrosion of metals of metal surface, depending on the temperature and
exposure time. Acids corrode more evenly, with less
Any metal surface is a composite of electrodes elec- pitting, at higher temperatures. At elevated tempera-
trically short-circuited through the body of the metal tures, more corrosion may occur, but the chance of
itself (Fig. 15-2). As long as the metal remains dry, forming a hole in the drillpipe or tubing string is
local currents and corrosion are not observed. How- actually less.
ever, on exposure of the metal to aqueous salt, alkali The effectiveness of a given corrosion inhibitor
or acid solutions, local action cells are able to func- depends on the metal. McDougall (1969) reported
tion, leading to chemical conversion of the metal to tests of five different commercial corrosion inhibitors
corrosion products. on N80, J55 and P105 tubulars. These results clearly
Table 15-2. Corrosion limits for low-alloy steel.

15-2.2. Acid corrosion on steel
Temperature Corrosion Limit (lbm/ft2)
All uninhibited acid solutions corrode steel. The
200°F [95°C] 0.02
attack of acid on steel occurs through the dissociated
hydrogen ions in the acid solution. This results in the 200° to 275°F [95° to 135°C] 0.05
oxidation and dissolution of iron at the anodic sites 275° to 350°F [135° to 175°C] 0.09
15-2 Additives in Acidizing Fluids

show the necessity of matching the inhibitor to the relative to the surrounding steel structure and thus
steel, because a wide range of responses was promote acid attack.
obtained for a given inhibitor. Smith et al. (1978)
pointed out that American Petroleum Institute (API)
designations for steels are inadequate to distinguish 15-2.4. Hydrogen embrittlement
corrosion rates because they identify yield strength,
whereas corrosion characteristics are determined pri- Hydrogen embrittlement results from the cathodic
marily by chemical composition. Because yield reaction described previously in which hydrogen
strength may be achieved either by physical or ions are reduced to elemental hydrogen. Elemental
chemical modifications in manufacturing, it is essen- hydrogen results from acidizing operations. Ele-
tial that the metal coupons tested are representative mental hydrogen adsorbed by a metal can lower
of the tubulars to be protected. ductility to the point that the metal becomes brittle.
Figure 15-3 shows the effect of hydrogen content
and tensile strength on the loss of ductility of steel.
15-2.3. Pitting types of acid corrosion The N80, P110 and Q125 grades of tubing and cas-
ing are prone to embrittlement as a result of acidiz-
In uninhibited acid solutions, the corrosion of steel is
ing. Consequently, inhibitor selection must take this
usually uniform. The constant shifting of anodic and
into account.
cathodic areas spreads corrosion fairly evenly over
the entire surface of the metal. In inhibited acids, pit-
ting of the surface occurs in some situations as a 60
result of inhibitor breakdown, insufficient inhibitor Tensile ductility (% reduction in area)

50 Utimate tensile
or metal impurities. strength = 130 ksi
40 156
• Inhibitor breakdown
Pitting-type corrosion is not uncommon on steel 30
containing an inferior corrosion inhibitor that is 20

exposed to acid solutions. All corrosion inhibitors
eventually break down after some period of time, 10
270 240 196
depending on various factors including tempera- 0
ture, acid strength and metal type. When this point 0 0.5 1.0 1.5 2.0 2.5
is reached, an inferior inhibitor may actually pro- Hydrogen content (ppm)
mote pitting by desorbing from the metal surface
in localized areas. Figure 15-3. Reduction in ductility with increasing hydro-
gen content and strength (Tetelman, 1973). 1 ksi = 100
• Insufficient inhibitor lbf/in.2.
Regardless of the quality of the corrosion inhibi-

tor, pitting may also occur if there is an insuf-
ficient amount of inhibitor to effectively coat the 15-2.5. Corrosion by different acid types
steel surfaces. Unprotected steel surfaces are The degree of dissociation of hydrogen ions from the
rapidly attacked by the acid solution and pitting acid molecule, along with the acid concentration,
occurs. determines the hydrogen ion activity, which is di-
• Metal impurities rectly proportional to its corrosivity on steel. The
Another condition that promotes pitting is the relative degree of dissociation for some common
presence of impurities or inclusions within the acids is hydrochloric > formic > acetic. Therefore,
steel. For example, small pieces of slag may hydrochloric acid (HCl) is more corrosive on steel
become trapped during the forming of the steel, than formic acid, which is more corrosive than acetic
or improper heat treating or quenching of the steel acid. Quite logically, the more aggressive an acid is
may produce discontinuities in its grain structure. in its attack on steel, the more difficult it is to inhibit.
These imperfections may, in turn, become anodic However, the mechanism of attack is the same for all
acid types.

15-2.6. Inhibitor types – can be used in the presence of H2S without the
precipitation of salts such as arsenic sulfide
Inhibitors function by interfering with the chemical (which can plug the wellbore)
reactions that occur at the anode or cathode of the
corrosion cell. The two basic types of corrosion – do not poison refinery catalysts
inhibitors are inorganic and organic. There are also – work effectively in all acid concentrations.
inhibitor aids. Disadvantages of organic inhibitors are that they
• Inorganic corrosion inhibitors – chemically degrade with time in the presence of
This class of inhibitors includes the salts of zinc, acid and thus do not readily provide long-term
nickel, copper, arsenic, antimony and various other protection at temperatures above 200°F [95°C]
metals. Of these, the most widely used are the (Table 15-3)
arsenic compounds. When these arsenic com- – cost more than the inorganic corrosion inhib-
pounds are added to an acid solution, they “plate itors.
out” at cathodic sites of exposed steel surfaces.
• Inhibitor aids
The plating decreases the rate of hydrogen ion
exchange, because the iron sulfide that forms Although these additives do not function as inhib-
between the steel and the acid acts as a barrier. It itors, they increase the effectiveness of organic
is a dynamic process in which the acid reacts with inhibitors. The common inhibitor aids are potas-
iron sulfide, rather than the metal. sium iodide, cuprous iodide, cuprous chloride and
Some advantages of inorganic inhibitors are that formic acid. The addition of these materials to
they existing organic inhibitor formulations greatly
extends the range of their effectiveness, particularly
– work effectively at high temperatures for long
in higher temperature applications.
contact times
– cost less than organic inhibitors.
Table 15-3. Effectiveness of corrosion inhibitors at
Disadvantages of inorganic inhibitors are that high temperatures in 15% HCl.
they
Inhibitor Inhibitor Temperature Protection
– tend to lose their effectiveness in acid solutions Concentration (°F) Time†
stronger than about 17% HCl (%) (hr)
– react with hydrogen sulfide (H2S) through the Organic 0.6 200 24
iron sulfide that may be present to form an insol- 1.0 250 10
uble precipitate such as arsenic sulfide 2.0 300 2
– poison refinery catalysts (such as platinum) Inorganic 0.4 200 24
– may liberate toxic arsine gas as a corrosion by- 1.2 250 24
product 2.0 300 12
– are difficult to mix and unsafe to handle. † Time required for 15% HCl to remove 0.05 lbm/ft2 of exposed metal area
• Organic corrosion inhibitors
Organic corrosion inhibitors are composed of polar
organic compounds capable of adsorbing onto the 15-2.7. Compatibility with other additives
metal surface, thereby establishing a protective film
that acts as a barrier between the metal and the acid Any additive that alters the tendency of the corrosion
solution. They usually serve as a cathodic polarizer inhibitor to adsorb will also change its effectiveness.
by limiting hydrogen ion mobility at cathodic sites. For example, surfactants added to acid for various
Organic inhibitors are composed of rather complex purposes may form micelles that solubilize the
compounds, with one or more polar groups made inhibitor, thereby decreasing the tendency for the
of sulfur, oxygen or nitrogen. inhibitor to adsorb on the metal surface. Inorganic
Some advantages of organic inhibitors are that salts and mutual solvents can also interfere with
they inhibitor adsorption. If possible, additives that reduce

the effectiveness of inhibitors should be included in a ment if no pitting occurs. For some cases an amount
preflush or overflush rather than in the acid solution. as high as 0.095 lbm/ft2 is allowable. If these metal
losses cannot be sustained without adverse effects,
a more effective inhibitor must be found.
15-2.8. Laboratory evaluation of inhibitors If inhibitor cost becomes prohibitive, it may
The laboratory evaluation of a corrosion inhibitor be possible to reduce the cost by careful treatment
involves subjecting a coupon of the metal to be pro- design. Precooling the tubing by injecting a large
tected to the acid to be used. The coupon is normally water preflush will be helpful. Using formic acid
inserted into a mixture of the acid and corrosion rather than HCl helps reduce the corrosion problem.
inhibitor to be evaluated and placed in a heated pres- Also, minimizing the contact time reduces inhibitor
sure vessel. The replication of downhole conditions requirements.
is important. The amount of corrosion is determined The most important aspect is the proper testing of
by weighing the coupon before and after the test. representative metal samples using the precise acid
Inhibitor effectiveness is expressed in terms of metal formulation to be applied at the most adverse condi-
loss per unit area of metal exposed per unit of time. tions of temperature and pressure. The test should be
The typical unit is lbm/ft2/day. conducted in an oxygen-free environment. Only
Also to be noted is pitting, which is a localized under these conditions can a decision about the ade-
attack. If appreciable attack is confined to a rela- quacy of the corrosion inhibitor be made.
tively small area of the metal acting as an anode,
then the resultant pits may be deep. A pitting factor,
defined as the ratio of the deepest metal penetration 15-3. Surfactants
to the average penetration, as determined by the Surfactants, or surface-active agents, are used in
weight loss of the specimen, is sometimes reported. acidizing to break undesirable emulsions, reduce
A pitting factor of unity represents uniform attack. surface and/or interfacial tension, alter wettability,
Many factors influence the corrosion rate measured speed cleanup, disperse additives and prevent sludge
in such tests. Of major importance are formation. The use of surfactants requires careful
• degree of agitation selection of an appropriate molecule. Remarkably,
in the design of most well treatments, surfactants are
• metal type selected with little or no laboratory data to support
• exposure time the choice and sometimes without full knowledge of
their properties at the conditions in which they will
• temperature
be applied. Improper surfactant selection can lead to
• acid type and concentration results contrary to those intended and may be detri-
mental to the success of the treatment.
• inhibitor type and concentration
Surfactants owe their properties to their “dipolar”
• metal-to-acid volume ratio composition (Fig. 15-4). The surfactant molecule
• gas composition (if an interface is present) consists of a water-soluble (hydrophilic) group and
• pressure
• presence of other additives such as surfactants.
15-2.9. Suggestions for inhibitor selection

Some metal loss from wellbore tubular goods must
be expected when acidizing, and the primary issue Lipophilic group
Hydrophilic group
is what level can be tolerated. Most service company (water soluble) (oil soluble)
information is based on the assumption that a metal
loss of 0.02 lbm/ft2 can be tolerated during the treat- Figure 15-4. Dipolar surfactant composition.

an oil-soluble (lipophilic) group, which are separated 15-3.2. Cationic surfactants

from each other although linked by a strong covalent
chemical bond. The molecules are classified into five Cationic surfactants carry a positive charge when
groups according to the ionic charge carried by the they ionize in aqueous solutions. There are two gen-
water-soluble group: eral categories of cationic surfactants. The first cate-
gory consists of long-chain primary, secondary and
• anionic tertiary amines. These are soluble only in acidic
• cationic solutions, where they ionize to form a long-chain
cation and simple anion salt. Figure 15-5 shows vari-
• nonionic ous molecular cationic structures.
• amphoteric The second important category of cationic surfac-
tants is the quaternary ammonium compounds. These
• fluorocarbons. ionize to form long-chain cations over a wide range
The hydrophilic-lipophilic balance (HLB) depends of solution pH. Cationic surfactants experience the
on the composition of the organic chain, which can same sensitivity to multivalent ions or increased con-
be 100% oleophilic (e.g., an alkane chain) or contain centrations of dissolved solids as anionic surfactants;
some oxygen atoms (e.g., an ethylene oxide chain) therefore, the same care must be exercised in their
and become more hydrophilic. The HLB value of a application as with anionic surfactants.
surfactant indicates how much its chains can oil wet Cationic and anionic surfactants are generally
a mineral. incompatible. When mixed, they tend to precipitate
in aqueous solutions.
15-3.1. Anionic surfactants

Primary amine R — NH+3
Anionic surfactants are commonly added to well
treatment fluids. These surfactants carry a negative R2
charge when they ionize in an aqueous solution.
Because most reservoir minerals are also negatively Tertiary amine R1 — NH+
charged at nearly neutral and higher pH values, ani- R3
onic surfactants exhibit minimal adsorption. Some
examples of anionic surfactants are +
R2
Sulfates R—O—SO3–
Quaternary amine R1 —N — R3
Sulfonates R—SO3–
R4
Phosphates R—O—PO3–
Phosphonates R—PO3–
Figure 15-5. Various types of cationic surfactants.
where R is an oil-soluble organic group.
Of these, the most common anionic surfactants are
sulfates and sulfonates. Anionic surfactants are sen- 15-3.3. Nonionic surfactants
sitive to multivalent ions such as Ca2+ and Mg2+. Nonionic surfactants have no charge at all in the
These tend to precipitate anionic surfactants, hydrophilic group and a long-chain organic (R) for
although this tendency can be overcome to some the lipophilic group. The water-soluble group is a
extent by increasing the surfactant concentration. polymer made from either ethylene oxide or propy-
Anionic surfactants are used primarily as nonemulsi- lene oxide. Other types include alkanol amine con-
fying agents, retarding agents and cleaning agents. densates and amine oxides. The general formulas for
these products are as follows:

Ethylene oxide polymer R—O—(CH2CH2O)xH to a greater extent than hydrocarbon surfactants.

Surface tensions as low as 30 dynes/cm can be
Propylene oxide polymer R—O—(CHCH2O)yH
obtained using surfactants with a hydrocarbon tail.

Values as low as 17 dynes/cm have been reported
CH3
using fluorocarbon surfactants. Fluorocarbons are
Alkanol amine commercially available in anionic, cationic and non-
condensate R—C—N—(CH2CH2OH)2 ionic forms.

O
15-3.6. Properties affected by surfactants
Amine oxides CH2CH2OH
 The main properties of a fluid or a mineral affected
COCO—N—O by surfactants are the surface and interfacial ten-
 sions, emulsification tendency, wettability, micelle
CH2CH2OH formation and dispersibility.
Nonionic surfactants obtain their water solubility • Surface and interfacial tensions
by attaching the long hydrocarbon chain to a highly Rosen (1989) presented a good discussion of how
soluble molecule such as polyhydric alcohol or by surface and interfacial tensions are reduced by sur-
reacting it with ethylene oxide. Most of the com- factants. Figure 15-6 is a simplified diagram of the
pounds in this classification are esters, ethers and interface between two condensed phases. A mole-
ether-esters. The lipophilic group may be derived cule in the bulk of a liquid is surrounded by mole-
from natural oils and fats, petroleum oils or synthe- cules like itself and the forces are balanced. At the
sized hydrocarbons. The hydrophilic group is usually interface, however, the forces acting on the mole-
a polyhydric alcohol or an alkyd oxide polymer. cules in the two liquids are different from the
These surfactants are used as nonemulsifiers and forces acting on the molecules in the bulk liquids.
foaming agents. The terms Aaa and Abb in Fig. 15-6 represent the
interaction energy between molecules at the inter-
face and similar molecules in the bulk of the liquid
15-3.4. Amphoteric surfactants whereas Aab represents the interaction energy
Amphoteric surfactants have a hydrophilic group between dissimilar molecules at the interface.
that changes from cationic to nonionic to anionic The increase in potential energy of the mole-
with increasing pH. In other words, if the solution is cules at the interface compared with that of the
acidic, the amphoteric surfactant acts like a cationic molecules in the bulk liquid is simply Aaa – Aab
surfactant; if the solution is neutral, it acts like a and Abb – Aab. Therefore, the total increase in
nonionic surfactant; and if the solution is basic, it
acts like an anionic surfactant. These properties are Aaa Aab
derived from the two groups of opposite charge on
the surfactant head. The amphoterics are usually
either amine sulfonates or amine phosphates. The Liquid a
general formulas are as follows:
Amine sulfonates RNH—(CH2)xSO3H
Amine phosphates RNH—(CH2)yOPO3H Liquid b
15-3.5. Fluorocarbon surfactants Aab Abb

Fluorocarbons form surfaces of lower free energy
than hydrocarbon surfaces. Consequently, fluorocar- Figure 15-6. The interface between two liquids (Rosen,
bon surfactants lower the surface tension of solutions 1989).

potential energy at the interface becomes Aaa + Abb

– 2Aab. The interfacial free energy per unit of sur- Gas
face area, which is also known as the interfacial
tension γI, becomes
γI = γa + γb – 2γab , (15-1)
where γa and γb are the surface tensions of pure Liquid b
liquid a and liquid b, respectively, and γab is the
interaction energy per unit of surface area between
liquids a and b. Abb
The value of γab tends to be large when similar
liquids are in contact; this causes γI to be small. Figure 15-7. The interface between a liquid and a gas.
Conversely, if liquid a is quite different from liq-
uid b, then the interaction energy is small and the
interfacial tension between the two liquids is large. • Emulsification tendency
Surface tensions of some liquids are listed in
An emulsion is a mixture of two fluids in which
Table 15-4.
fine droplets of one fluid are suspended in the
other. Emulsions may be oil external or water
Table 15-4. Surface tension of some liquids.
external. In the first instance, oil is the continuous
Liquid Surface Tension phase with the water droplets dispersed through-
(dynes/cm) out. This is the most common emulsion found in
Water 72.0 wellbores. A water-external emulsion has an aque-
15% HCl 72.0 ous external phase with oil droplets distributed
Spent 15% HCl 76.9
throughout.
Many crudes contain naturally occurring surfac-
Octane 21.8
tants that reduce the surface tension between oil
and formation water and thus promote the devel-
Figure 15-7 illustrates the interaction between opment of emulsions. Treatment of the formation
a liquid and a gas. In this case the interface is the with certain surfactants can also lead to the devel-
surface of the liquid. The interaction between a opment of emulsions. However, appropriate sur-
gas molecule and other gas molecules is negligi- factants can be used to treat wells with emulsion
ble, as is the interaction between gas molecules problems. Laboratory testing is required to deter-
and the liquid. Therefore, γI = γb; in other words, mine the appropriate surfactant.
the interfacial tension is simply the surface tension
• Wettability
of the liquid.
A surfactant, when added to two immiscible Oil and water are immiscible liquids that compete
phases, adsorbs at the interface. If one of the liq- for space on the formation surface. Which liquid
uids is water and the other is a hydrocarbon, the preferentially wets the rock surface is an important
hydrophobic end of the surfactant will be oriented factor in acidizing. When the formation surface is
toward the hydrocarbon, and the hydrophilic group completely covered by a film of oil, the formation
will be oriented toward the water phase. This causes is termed oil-wet. Conversely, when covered by
a change in the interface. Now the interaction is water, the formation is water-wet.
between the hydrophilic group of the surfactant and An electrochemical approach helps to explain
the water phase and the hydrophobic group of the the ability of a surfactant to adsorb at interfaces
surfactant and the hydrocarbon phase. These inter- between liquids and solids and alter the wettability
actions are much stronger than the original interac- of solids. Sand usually has a negative surface
tion between the water and hydrocarbon, which charge. When a cationic surfactant is present and
effectively leads to an increase in γab and a con- the pH rises above the pzc, the positive water-sol-
comitant decrease in γI. uble group is adsorbed by the negative silica parti-
cle. This leaves the oil-soluble group to influence

wettability. As a result, cationic surfactants gener-

ally create an oil-wet sand, although cationic sur-
factants can be chemically modified to yield a
water-wet sand. When anionic surfactants are
used, the sand and silicate minerals are generally Oil Water
left in their natural water-wet state. Wettability
characteristics exhibited by anionic and cationic
surfactants on a silicate particle are shown in
Fig. 15-8. Carbonates have a pzc at a pH of 9–10
and thus are normally oil-wet by cationic surfac- Hydrophilic Oleophilic
oleomicelle hydromicelle
tants above a pH of 9 whereas anionic surfactants
oil wet the rock below a pH of 9.
Silicate particle
Water
Oil-wet Water-wet
Lamellar micelle
Figure 15-9. Examples of surfactant micelles that can form

in liquids above a specific surfactant concentration (Hall,
1986).
Once the aggregate has been dispersed, a different

Cationic surfactant Anionic surfactant
surfactant may be required to prevent the dis-
Figure 15-8. Silicate particle wettability characteristics persed particles from rejoining.
(Hall, 1986).
• Micelle formation
15-3.7. Applications and types of surfactants
Surfactants can form micelles in liquids when pre- In recent years, the uses of surfactants have grown
sent above a specific concentration for each mole- quickly. This unique class of chemicals has found
cule, solvent and temperature. This concentration application in almost all phases of acidizing. A com-
is referred to as the critical micelle concentration prehensive review is beyond the scope of this chap-
(CMC). Below the CMC, many properties of the ter. However, a brief review follows.
system are concentration dependent. Some of these • De-emulsifiers
properties are surface tension, interfacial tension,
A de-emulsifier is used to break oil-water emul-
foam stability and emulsion tendencies. Figure
sions, which occur commonly. The action of a de-
15-9 shows examples of surfactant micelles.
emulsifier depends on how quickly it can concen-
• Dispersibility trate at the oil/water interface. The faster the con-
To separate a highly associated structure (e.g., centration at the interface, the more rapid will be
a sludge) into its particulate elements usually the rate of emulsion breaking. These surfactants
requires energy. If the dispersed phase is not in are usually oil-soluble chemicals that are blended
intimate contact with the continuous liquid phase, with other de-emulsifying surfactants to achieve a
then the energy transfer will not be very efficient. synergistic response.
A wetting surfactant that wets the dispersed phase • Nonemulsifiers
with the liquid phase greatly improves the dis-
These surfactants are added to well treating or
persibility. This is one aspect of aiding dispersion.
workover fluids to prevent emulsions. Chemical

suppliers and service companies have many differ- This normally renders the particles oil-wet and
ent surfactants that are classified as nonemulsifiers. causes them to act more like droplets of oil than
The chemicals are usually mixtures of surfactants solid particles. As long as they remain oil-wet, they
and solvents that are blended to obtain a final tend to remain suspended in the spent acid.
composition with broader applications. Normally, Another force working to suspend the silt parti-
the surfactants are chosen so that they will water cles is electrostatic repulsion. Because all the par-
wet the formation. They should also mix easily ticles carry the same charge, the particles repel
with treating or workover fluids and prevent the each other and tend to not agglomerate. This com-
formation of emulsions with the reservoir fluids. bination of oil-wetting and electrostatic repulsion
makes an effective silt-suspending system.
• Emulsifiers
Specialized emulsions are often used as treating flu- • Antisludge agents
ids. Many solvent systems concurrently use aqueous When acid contacts some crude oils, a sludge can
and hydrocarbon phases. Such emulsions may ex- form at the acid/oil interface. This is most severe
hibit higher viscosity than either base fluid. Emul- with high-strength acid systems (20% or higher).
sions also have the ability to isolate the internal Once formed, the sludge is difficult to redissolve
phase so that it is not as reactive. A common exam- into the oil. As a result, the sludge accumulates in
ple of this technology is the use of emulsified acid. the formation and decreases permeability.
With these properties, emulsions are efficient To combat the formation of sludge, cationic and
scale removal systems. When an inorganic scale is anionic surfactants are used to adsorb and provide
deposited in combination with a hydrocarbon scale a continuous layer of protection at the acid/oil
(paraffin or asphaltene), an efficient scale removal interface. Sludge development can often be pre-
system can be formulated by emulsifying a solvent vented by lowering the acid strength.
for the inorganic scale—usually HCl—with the
• Surface tension reducers
hydrocarbon. Usually, the emulsion has a low vis-
cosity with a stability of up to several hours at bot- Surfactants are commonly used in treating tight
tomhole conditions. The stability of the emulsion is gas wells and in scale removal treatments. Their
low to minimize the potential for damaging the well main function is to lower the surface and inter-
with the emulsion during workover. The hydrocar- facial tensions of the treating fluid. This reduces
bon phase ranges from 10% to 30% by volume with the capillary pressure, which makes it easier to
a water-base fluid as the external phase. Most emul- recover fluid from the capillary pore channels and
sifiers used in this application are nonionic. to prevent oil-wetting films. Lowering the capil-
lary pressure aids in cleaning up the well because
• Silt-suspending agents less differential pressure is required to move the
Surfactants can also be used to remove acid-insol- fluid through the capillary channels. The maxi-
uble clay and silt during an acidizing treatment. A mum differential pressure available to clean up a
surfactant is added that adsorbs onto the clay or well is the reservoir pressure less the lowest draw-
silicate particles and keeps them suspended by down that can be achieved at the perforations. If
electrostatic repulsion. With this surfactant pre- this maximum differential pressure is not high
sent, damaging insoluble residues can be removed enough to move the liquid out of a pore channel,
with the spent acid. These particles, if not a block exists and its removal may require a sur-
removed, can stabilize emulsions or settle out and factant that provides a lower surface tension (i.e.,
block the pore throats of the producing formation. lower capillary pressure).
The mechanism for suspending silt with a surfac-
• Corrosion inhibitors
tant can also be best understood by an electrochem-
ical approach. Silt particles are minerals, such as Corrosion inhibitors function by adsorbing onto
quartz, feldspars and clays, that, assuming the the steel and providing a protective barrier
treating fluid has a pH above the pzc, carry a nega- between it and the acid. Surfactants that are effec-
tive surface charge. Silt-suspending surfactants are tive as corrosion inhibitors include certain quater-
cationic, and the positively charged end of the mol- nary amines and acetylenic nonionics. Generally,
ecule adsorbs onto the surface of these minerals. these materials contain some cationic species.

• Bactericides – nonionic—stable foam with acid or brine; be-

When treating a formation, it is usually desirable cause of cloud point problems, nonionics nor-
to leave it water-wet to aid the production of oil. mally cannot be used above 200° to 250°F
This is done with an appropriate surfactant. It is [95° to 120°C]
also desirable to leave the formation free of conta- – anionic—stable foam with brine; can be used up
minating bacteria that may have been inadver- to 300°F [150°C]
tently pumped into the well with the treating fluid. – cationic—stable foam with acid or brine; can be
Many cationic surfactants possess this biocidal used up to 300°F
property and are commonly used in conjunction
– amphoteric—stable foam with acid or brine; can
with other wetting surfactants.
be used up to 350°F [175°C].
• Clay treaters
Foam stability usually can be improved by
Drilling muds commonly contain the highly water gelling the liquid.
swelling clay smectite (bentonite). While a well is
being drilled and completed, some of the clay may
invade the producing zone and decrease formation
productivity. Several chemical solutions are used 15-4. Clay stabilizers
for the removal of this invading clay. Chemicals used to stabilize clays and fines function
The swelling properties of smectite may be by being adsorbed, usually by electrostatic attraction
changed by chemical reactions, because all clays or ion exchange, on the minerals to be stabilized.
are negatively charged at pH values above their Because silicates above their pzc values have a nega-
pzc values. Thus, positively charged cations (cal- tive charge, the most effective stabilizer has a posi-
cium, sodium, potassium and hydrogen) may be tive charge (cationic). Common clay stabilizers are
attracted to the clay and are known as exchange- highly charged cations, quaternary surfactants, poly-
able cations. The particular cation and the quantity amines, polyquaternary amines and organosilane.
present in a clay system govern the degree of
water swelling or clay expansion. For example,
hydrogen smectite swells less in water than 15-4.1. Highly charged cations
sodium smectite. By reacting sodium smectite
(bentonite) with acid, hydrogen ions will replace Two highly charged cations that were once widely
the sodium ions by the cation-exchange process. used as clay stabilizers are hydroxyaluminum
Because hydrogen clay retains less water than (Al6(OH)12(H2O)126+) and zirconium (Zr4+) added as
sodium clay, the treatment of drilling mud with zirconium oxychloride (ZrOCl2). Solutions contain-
acid can shrink smectite and clean up mud filter ing either stabilizer are usually pumped after various
cake. However, acid may also cause clay particles preflushes. The stabilizer solution is then overflushed
to flocculate. This flocculation results in larger with a compatible fluid to remove excess clay stabi-
clay particles that may be more difficult to remove lizer from the near-wellbore region, and the well is
from the well. shut in. These systems do not appreciably affect for-
Surfactants have been developed to inhibit the mation wettability.
flocculation of clays, even in solutions of high The primary advantages of these systems are
ionic strength. This is accomplished by using a • inexpensive
surfactant that effectively disperses clay particles • treat for both migration and swelling damage
and minimizes the formation of aggregates. In
• can treat a large area of rock.
addition, smaller particles are easier to suspend,
resulting in more efficient removal from the well. The disadvantages are
• Foaming agents • hydroxyaluminum is not acid resistant
Surfactants are used to generate a stable foam. The • require shut-in to polymerize
following guidelines are usually applicable for sur- • can cause plugging
factant application:

• difficult to use in fracturing the polymers also promote water-wetting of the sili-
• require proper preflushes and overflushes. cate. In addition, polymers of sufficient length can
promote polymeric bridging between silicate particles.
Systems such as hydroxyaluminum can also be If a treated silicate is contacted by brines, the
used in treating water injection and disposal wells polyamine can lose its positive charge and be
to stabilize silicates and minimize future plugging washed off the silicate. When this occurs, that sili-
from colloidal fines in the injection water. cate is no longer stabilized.
The disadvantages of polyamines are that
15-4.2. Quaternary surfactants • treatment may not be permanent
Quaternary surfactants have been used as clay stabi- • they are expensive.
lizers for dry gas wells. At conditions above the pzc,
these surfactants are readily adsorbed by silicates
owing to electrostatic attraction between the posi- 15-4.4. Polyquaternary amines
tively charged surfactant and negatively charged
Polyquaternary amines can be used in any water-
clay. The resulting charge neutralization reduces the
base fluid, including acids and bases. The chemical
ion-exchange capacity of the clay. Therefore, the
structure of the two polyquaternary amines that have
clay is not as susceptible to swelling resulting from
been widely used is as follows:
the adsorption of hydrated cations.
Quaternary surfactants promote oil-wetting of the CH3 Condensation
silicate, which tends to minimize the adsorption of  products of
water by the silicate. However, if any liquid hydro- —CH2—CHOH—CH2—N(+)— dimethylamine
carbons are present, the silicate can readily become  and
oil-wet. This, of course, reduces the relative perme- CH3 epichlorohydrin
ability of the rock to hydrocarbons. Also, clays swell
by imbibing fluids into their lattice structure. —CH2—CH—CH2—CH— Dimethyl diallyl
  ammonium chloride
CH2 CH2
15-4.3. Polyamines
Polyamines are organic polymers that contain more N(+)
than one amine group. For the purpose of this dis-
cussion, only primary, secondary and tertiary amines CH3 CH3
are considered polyamines. Polyamines are posi-
tively charged in acidic fluids. The general structure Clays and fines are stabilized by charge neutraliza-
of a polyamine is as follows: tion, water-wetting and polymeric bridging. Silica
fines have a lower charge density than clays. There-
R´
fore, polyquaternary amines preferentially adsorb on

clays as opposed to silica. When acidizing a water-
CH3—[R—N—R]n—CH3
sensitive formation with hydrofluoric acid (HF), clay

stabilizer should be used if possible. If it is not pos-
H
sible to put clay stabilizer in all fluids, it should be
where R is a repeating hydrocarbon unit, R´ can be used in the overflush, which should be overdisplaced
either a hydrocarbon unit or a hydrogen, and n is the with fluid that contains no clay stabilizer to ensure
number of amine units in the polymer. that no unadsorbed clay stabilizer is left at the well-
Because the polymer has many amine units, it can bore. A normal concentration of polyquaternary amine
adsorb strongly on the silicate with many points of for applications in HF treatments is 5 gal/1000 gal of
attachment. The polymers can effectively neutralize active polymer in all fluids or 71⁄2 gal/1000 gal of clay
the negative charge of the silicate. By carefully con- stabilizer in 200 gal of overflush.
trolling the carbon:nitrogen ratio, usually 8:1 or less,

There has been concern regarding the distinction polysiloxane coating stabilizes fines by blocking ion-
between fines migration and clay migration. Fines exchange sites and increasing interparticle attractive
consist predominantly of silt-sized particles of quartz forces. Interparticle attractive forces can include van
and feldspars. Laboratory data indicate that the der Waal forces and hydrogen bonding. These forces,
polyquaternary amines are not as effective in con- in addition to the electrostatic forces, help maintain
trolling fines migration as compared with clays the fines in their aggregated state along the pore
(Hall, 1986; Ayorinde et al., 1992). wall. By blocking the ion-exchange sites, the poly-
siloxane coating decreases the magnitude of double-
layer repulsion.
15-4.5. Organosilane The polysiloxane formed can bind to minerals that
Kalfayan and Watkins (1990) proposed that an have low-cation-exchange capacities (quartz), as
organosilane compound can be used as an additive well as to clays that have high-cation-exchange
to HCl-HF mixtures to prevent poststimulation fines capacities. Therefore, the organosilane additive is
migration. Organosilane has the following general well suited for formations that contain nonclay fines
structure: as well as clay fines.
OR
 15-5. Mutual solvents
RO—Si—OR
 Mutual solvents, as the name implies, are chemicals
RŃH2 that are mutually soluble in both hydrocarbons and
water. The most efficient mutual solvents are glycol
where R and R´ are hydrolyzable organic groups. ethers, a reaction product of alcohols and ethylene
As an additive to acid, organosilane hydrolyzes oxide. These chemicals are relatively safe and easy to
to form silanols, which have the following structure: use in the field. The preferred glycol ethers contain at
OH least a butyl or higher molecular weight group.
 The use of mutual solvents in the acid stimulation
HO—Si—OH of a sandstone reservoir is a common practice.
 Mutual solvents are used in acid solutions and over-
RŃH2 flushes to
The silanols react with each other and with silanol • aid in reducing water saturation around the well-
(Si–OH) sites present on siliceous mineral surfaces to bore by lowering the surface tension of the water
form covalent siloxane (Si–O–Si) bonds by a conden- to prevent water blocks
sation/polymerization reaction mechanism. The reac- • solubilize a portion of the water into a hydrocar-
tion of the silanols with one another and with the bon phase to reduce the water saturation
silanol sites on siliceous mineral surfaces forms a non- • aid in providing a water-wet formation to maintain
oil-wetting polysiloxane coating on siliceous mineral the best relative permeability to oil
surfaces. The length of the polysiloxane chain formed
by the hydrolysis of the organosilane, and the subse- • help to prevent insoluble fines from becoming oil-
quent condensation/polymerization of silanols, is not wet and stabilizing emulsions
known but is believed to be short. Acid, which cat- • help to maintain the required concentration of sur-
alyzes the initial hydrolysis of the organosilane, also factants and inhibitors in solution by reducing
retards the growth of polymer chains. adsorption of these materials
The mechanism by which the polysiloxane coating • help to dissolve both the adsorbed inhibitor and
stabilizes fines is different from the ion-exchange acid-insoluble residue (certain acid corrosion
mechanism by which other clay stabilizers work. inhibitors contain acid-insoluble residues that can
However, it is similar to the coating process pro- cause formation plugging and inhibitor adsorption
posed for fluoboric acid (Boyer and Wu, 1983). The on formation minerals and change the wettability)
polysiloxane coating binds clay and other siliceous • dissolve any oil on the formation pore surface
fines in place through covalent siloxane bonds. The

• serve as a de-emulsifier 15-6. Iron control additives

• improve the cleanup of spent acid following
treatment. 15-6.1. Sources of iron
Commonly used mutual solvents and their normal When appreciable quantities of iron in the form of
concentrations are Fe3+ (ferric ions), rather than the usual Fe2+ (ferrous
• ethylene glycol monobutyl ether (EGMBE): 10% ions), are dissolved by the acid, iron precipitation
by volume and permeability reductions can occur after acidiz-
ing. The oxidation state of the iron governs precipi-
• ether/surfactant/alcohol blends: 5% by volume.
tation. Ferric iron precipitates at a pH of about 2,
whereas ferrous iron precipitates at a pH of about
7—the actual values depend on the concentrations of
15-5.1. Adsorption of mutual solvents the ferrous and ferric ions. Because spent acid solu-
King and Lee (1988) studied the adsorption proper- tions seldom rise to a pH above 6, precipitation of
ties of an alcohol-mixture mutual solvent, such as a ferrous iron is seldom a problem. Sources of iron
blend of isopropyl alcohol and isooctyl alcohol, and include
EGMBE. They found that adsorption of the mutual
• corrosion products found on the walls of the tubu-
solvent can be severe depending on the type of
lars
mutual solvent. For deep damage removal, products
that are lost during injection through the formation • mill scale
are not acceptable because the mutual solvent is • iron-bearing minerals.
removed from the leading edge of the acid. Although Iron occurs naturally in formation waters or in for-
there appears to be no direct damage caused by mation minerals (Table 15-5). However, the tubulars
adsorption of the mutual solvent, the acid is left after in the well are one of the most prominent sources of
the loss without any material to lower surface tension iron. Before the acid reaches the formation, it flows
or to break emulsions. Paktinat (1991) recommended for a relatively long time through the tubing. Tubing
the use of ethoxylated alcohol with EGMBE to mini- usually contains rust, which is dissolved by acid.
mize adsorption, thus resulting in deeper penetration Newly manufactured tubing has a crust of mill scale.
of the mutual solvent and reduced emulsion tenden- The thickness of the mill scale on oilfield tubulars
cies. varies considerably depending on the cooling rate
and how much the pipe was manipulated during the
straightening procedure. Mill scale is composed of
15-5.2. Chlorination of mutual solvents
two distinct layers—a hard, dense layer next to the
King and Lee (1988) presented data on the chlorina- pipe that is approximately 0.003 in. thick; on top of
tion of mutual solvents with 15% and 28% HCl at the dense mill scale is a layer of softer, flaky mill
temperatures from 70° to 250°F [20° to 120°C]. The scale that is approximately 0.007 in. thick. Most of
chlorination issue has long been of interest because the softer scale pops off the pipe during the straight-
chlorinated hydrocarbons poison refinery catalysts. ening procedure.
Their data show that the chlorination of mutual sol-
vents is a function of
Table 15-5. Oxidation state of iron
• type of mutual solvent in formation minerals.
• temperature Mineral Chemical Formula Oxidation State

• concentration of HCl (i.e., spending of HCl Hematite Fe2O3 Fe2+
reduces the possibility). Magnetite FeO–Fe2O3 Fe2+–Fe3+
At temperatures exceeding 200°F, especially with Pyrite FeS Fe2+
28% HCl, the chlorination of most mutual solvents Siderite FeCO3 Fe2+
was severe, if there were no carbonates with which the Chlorite clay – Fe2+
acid could react. The presence of carbonates resulted
Mixed-layer clay – Fe2+
in very low production of chlorinated hydrocarbons.

The theoretical quantity of 15% HCl required to • Sequestering agents

remove 0.073 lbm of mill scale from 27⁄8-in. tubing is Sequestering agents bond to the iron and hold it
69 gal/1000 ft of tubing. If a 10,000-ft well is acidized in solution so that it cannot precipitate. Citric acid,
for damage removal, approximately 690 gal of 15% ethylenediaminetetraacetic acid (EDTA) and
HCl could be neutralized by the mill scale. Of course, nitrilotriacetic acid (NTA) are some of the more
the acid contact time and temperature control the commonly used sequestering agents.
quantity of mill scale removed. If acid is spent on mill
scale, it will contain 85,900 ppm of total iron. Of this, • Reducing agents
57,300 ppm will be ferric iron (Fe3+). Thus, the lead- Reducing agents convert ferric (Fe3+) to ferrous
ing edge of the treating fluid may be partially spent (Fe2+) iron. Secondary precipitation of ferric iron
acid that contains a high concentration of iron. The occurs at a pH of about 2.0. The secondary precip-
ferric iron can be precipitated as ferric hydroxide itation of ferrous iron hydroxides does not occur
(Fe(OH)3), a dark brown gelatinous precipitate that until the pH is above 7.0. Returned spent acids
can be damaging to the formation. never have a pH this high. Erythorbic acid and
It should be emphasized that acid dissolves iron sodium erythorbate are commonly used as reduc-
compounds regardless of the presence of any type ing agents. Erythorbic acid is preferred over
of inhibitor used to protect the elemental iron in the sodium erythorbate in sandstone acidizing because
steel. Iron control additives help prevent the precipi- the addition of sodium salts of either sequestering
tation of iron hydroxide. or reducing agents to mud acid can lead to the pre-
Concern is often expressed about the dissolution cipitation of insoluble hexafluosilicate. Hall and
of iron-containing minerals (siderite, hematite and Dill (1988) reported that erythorbic acid is unsta-
chlorite) from the formation and subsequent precipi- ble in hot HCl and decomposes to form an insolu-
tation of the dissolved iron. Previous studies indicate ble precipitate. Although this is true, the decompo-
that most of the iron contained in these minerals sition process is slow and the acid normally
occurs in the Fe2+ oxidation state (Table 15-5) and spends long before precipitation can occur
does not present a precipitation problem. However, (Crowe, 1985).
numerous exceptions to this rule exist. Streaks of Most reducing agents also act as oxygen scav-
pure hematite (Fe2O3) occur in some sandstones, engers that remove dissolved oxygen from the
whereas others contain Fe3+-type minerals uniformly fluid. The scavengers prevent the oxidization of
distributed within the matrix of rock. Where these ferrous iron to ferric iron. This maintains iron in
conditions are present, greater amounts of an iron solution by preventing the precipitation of ferric
control additive are required to control ferric hydrox- iron. The amount of iron that can be reduced
ide precipitation. depends on the quantity of chemical added.
Aeration of the solution can introduce additional
oxygen.
15-6.2. Methods of iron control Table 15-6 compares iron control additives,
The three methods currently used to help keep iron showing their advantages and disadvantages and
in solution are pH control, sequestering agents and the amount of each required in 1000 gal of 15%
reducing agents (also effective as oxygen scaven- HCl to sequester 5000 ppm of ferric iron at 150°F
gers). These may be used individually or in combi- [65°C] for a minimum of 2 days.
nation, depending on the source and amount of iron The safest way to prevent damage to the reservoir
dissolution expected. from precipitated iron hydroxide is to clean or
pickle the pipe with acid before acidizing the for-
• pH control mation. The acid should contain large quantities of
pH control is accomplished by the addition of a iron control additives and should be circulated out
weak acid that is very slow to react so that a low of the well, not pumped into the formation. In con-
pH is maintained after the HCl has spent. Acetic junction with this treatment, a dispersed hydrocar-
acid is typically used for this purpose. A low pH bon phase should be incorporated or used as a pre-
aids in preventing the secondary precipitation of flush to remove pipe dope that could plug the perfo-
iron. rations.

Table 15-6. Comparison of various iron control agents.
Control Agent Advantages Disadvantages Amount (lbm)

Citric acid Effective at temperatures up to Precipitates as calcium citrate if excess 175
400°F [205°C] uncomplexed quantities are used (more
than 10 lbm/1000 gal)
Citric acid–acetic acid Very effective at lower temperatures Even for the indicated amount, calcium Citric: 50
mixture citrate precipitates unless at least 2000-ppm Acetic: 87
Fe3+ is present in spent acid.
Efficiency decreases rapidly at temperatures

above 150°F [65°C]
Lactic acid Little chance of calcium lactate Not very effective at temperatures above 190
precipitation if excessive quantities 100°F [40°C] (at 75°F [25°C])
are used
Acetic acid No problem from possible precipitation Effective only at temperatures at about 150°F 435
as calcium acetate
Gluconic acid Little chance of calcium gluconate Effective only at temperatures up to 150°F 350
precipitation
Expensive on a cost-performance basis
Tetrasodium salt Large quantities may be used without More expensive to use than many other agents 296
of EDTA precipitation of calcium salt.
Effective at temperatures up to 400°F
Nitrilotriacetic acid Effective at temperatures up to 400°F 150
More soluble in acid than EDTA—

higher concentrations can be used
Less expensive than EDTA
Sodium erythorbate Smaller quantities required Increased corrosion inhibitor concentration 23

required for certain applications
Effective at temperatures up to 400°F
Should not be used in HF—use erythorbic acid
The treatment of sour wells presents an entirely 15-7. Alcohols

different type of precipitation problem. In addition to
free sulfur precipitation, by the reaction of Fe3+ with Alcohols are used in acidizing fluids to remove water
H2S, the dissolved Fe2+ also precipitates as ferrous blocks, enhance fluid recovery, retard acid reactivity
sulfide (FeS) on spending of the acid. FeS precipi- and decrease water content. The most common alco-
tates at a pH of about 2. Laboratory and field data hols used in acidizing are isopropanol and methanol.
presented by Hall and Dill (1988) show that a com- Their physical and chemical properties are listed in
bination of NTA, EGMBE and a sulfide modifier is Table 15-7.
an effective system for controlling the precipitation
of FeS and free sulfur when acidizing sour wells. Table 15-7. Physical and chemical properties of
Crowe (1985) previously proposed a two-component isopropanol and methanol.
system to address sour well problems. The system Property Isopropanol Methanol
incorporates erythorbic acid to reduce ferric iron to
Density at 68°F [20°C] 0.785 0.792
ferrous iron and EDTA to chelate ferrous iron and
Weight (lbm/1000 gal) 6.54 6.60
eliminate the precipitation of FeS.
Flash point (closed up, °F [°C]) 54 [12] 54
Solubility in water Complete Complete

Isopropanol is normally used at a maximum of • Low flash point

20% by volume. Methanol is used at various concen- Both isopropanol and methanol, and even acid
trations, but a typical concentration may be 25% by solutions containing 20% or more by volume of
volume. Alcohol is used in acidizing fluids for the either, have low flash points.
following reasons.
• Increase in corrosiveness
• Removal of water blocks
Corrosion tests have shown that alcohol-acid mix-
One problem that can severely decrease produc- tures require a higher concentration of inhibitor
tion is blockage of the pore spaces by water; this than equivalent acid mixtures without alcohol.
is commonly known as a water block. Water
blocks may form where high capillary forces are • Adverse reactions
present in porous rocks. The most severe water Formation brines with a high concentration of dis-
block problems occur in formations with gas per- solved salts can “salt out” in the presence of alco-
meabilities less than 120 md. The alcohol in the hols. To help prevent the occurrence of salt precip-
treating fluid reduces the capillary forces within itation, treating solutions should not exceed 20%
the reservoir, thus enabling easier removal of the by volume of isopropanol or 40% by volume of
liquid phases. methanol.
• Enhancement of fluid recovery • Incompatibility
Another problem that occurs in treating oil or gas Some crude oils are incompatible with both
wells is the recovery of treating fluids, especially methanol and isopropanol. Compatibility tests
in gas reservoirs. The high surface tension of should be conducted before acidizing a well with
water or acid solutions hinders their penetration a fluid containing alcohol. Some formation types
and recovery. Conventional surfactants help, may even be extremely sensitive to aqueous solu-
although they lose much of their activity by tions that contain high concentrations of alcohols.
adsorption. The addition of alcohol to acid solu- • Side reactions
tions reduces their surface tension. The concentra-
tion of alcohol normally used for this purpose is There are undesirable side reactions when alcohols
sufficient so that loss by adsorption is not a prob- are used in acidizing. Even under moderate temper-
lem. ature conditions, alcohols react readily with acid.
In the case of organic acids (e.g., acetic or formic
• Retardation of acid reactivity acid), these reactions result in ester formation, with
Alcohol has a retarding effect on acid reactivity. the only resulting problem being a possible loss of
The retardation rate is related to the type and per- available acid for the stimulation reaction. Even
centage of alcohol added. this may be of small consequence, because the
esterification reaction is reversible and regeneration
• Decrease of water in acids
of the organic acid for the desired stimulation reac-
Some formations contain a large amount of water- tion is possible (Keeney and Frost, 1975):
sensitive clays. To minimize the amount of water
contained in acidizing solutions, alcohols are used CH3COOH + CH3OH CH3COOCH3 + H2O
in place of the dilution water. acetic acid methanol methyl acetate
The major disadvantages of using alcohol in acid- However, in an alcoholic solution of HCl, a
izing fluids are as follows: reaction takes place that results in the formation
of the following organic chlorides:
• Effective concentration
CH3OH + HCl → CH3Cl + H2O
It takes a large amount of alcohol, 20% or more, to
methanol hydro- methyl
provide beneficial effects.
chloric acid chloride
• Cost
CH3CHOHCH3 + HCl → CH3CHClCH3 + H2O
Replacing water with alcohol in the acidizing solu- isopropanol hydro- isopropyl
tion makes the treatment more expensive. chloric chloride
acid

This poses a genuine problem in gas well stimula- Recent field trials by Shuchart and Ali (1992)
tion because the second and third of these reactions show that the use of 3% acetic acid in HF acidizing
are irreversible, and any HCl lost to this process can- sequences decreases the occurrence of scaling. The
not be regenerated, as is possible with organic acids use of acetic acid delays the precipitation of alumi-
in the first equation. Thus, these reactions reduce the nosilicate scale by two mechanisms—a buffering
HCl content. A second problem concerns the reac- effect and a chelating effect.
tion products themselves. Although the presence of
organic chlorides in natural gas might not present
any special problems, their presence in produced 15-9. Organic dispersants
crudes or distillates is considered serious contamin-
ation, because of the detrimental effect of chlorinated A relatively new method of removing organic
hydrocarbons on refinery catalysts. deposits is the use of a dispersant surfactant with
Figure 15-10 shows the effects of increasing tem- xylene. These surfactants can penetrate and loosen
perature on the reduction of acid strength for four organic deposits so that the xylene can effectively
concentrations of methanol. The data show that there dissolve or remove the deposit. Concentrations of
is no significant loss of acid strength with methanol 1% to 10% volume/volume are used depending on
concentrations up to 32% by volume until the tem- the deposit type, hardness and adhesion and the bot-
perature exceeds 175°F [80°C]. tomhole temperature. For matrix, wellbore and tub-
ing cleanup treatments, a soaking period of several
hours is recommended.
15
10 15-10. Organic solvents

HCl (%)
As Fig. 15-1 shows, organic solvents used alone or

5 16% MeOH
in combination with acid and other materials are use-
32% MeOH
48% MeOH
64% MeOH
ful for removing water blocks and solids. King
0 (1986) provided a good discussion of the uses of
0 50 100 150 200 250 300 these solvents alone and in combination with acid for
Temperature (˚F)
damage removal. Organic solvents include alcohols
in addition to more traditional solvents such as
Figure 15-10. Effect of various concentrations of methanol
on acid strength at increasing temperatures (courtesy of
xylene, toluene and diesel.
Halliburton Services). Organic solvents are particularly useful for the
removal of organic deposits. An organic solvent can
be combined with acid when the inorganic scales are
mixed with or coated by asphaltenes and paraffins.
An oil-external emulsion with the appropriate acid as
15-8. Acetic acid the internal phase is useful for treating mixed depos-
A common problem in HF acidizing is the precipita- its. The emulsion provides a further benefit of retard-
tion of reaction products. Many of the precipitation ing the acid, thus allowing the deeper penetration of
reactions occur almost immediately on spending the live acid.
acid solution. Most of these precipitation problems
can be overcome by the use of an adequate acid pre-
flush and ammonium chloride (NH4Cl) overflush. 15-11. Diversion
However, precipitation can also occur days, months
Diversion is often required to ensure that the treating
or even years after an HF acidizing treatment. This
fluid works effectively. Diverters function simply by
precipitation is often called aluminosilicate scaling.
equalizing the flow so that zones of differing perme-
Scales of this type are believed to form when the
abilities can be treated. Schechter (1992) provided a
spent HF mixes with formation fluid to slightly raise
good discussion of diversion principles. The empha-
the pH, thus causing precipitation.
sis in this section is on diversion through the use of

particulates added to the treating fluid and foams discharge regulations. On flowback, emulsion and
rather than the use of packers or ball sealers. oil/water separation problems that occur can poten-
Ideally the diverter should tially result in excess oil solubilized in the water by
the formation of microemulsions. Emulsion prob-
• be insoluble in the treating fluid
lems under such circumstances can result in thou-
• form an essentially impermeable layer on the for- sands of gallons of non-pipeline-quality oil and
mation face without penetrating deeply into the could shut down the separation unit and minimize
formation production from these wells. Dehydrated oil and
• be easily removed following the treatment. clean discharge water are required to maintain nor-
Generally all, or at least most, of these criteria can mal uninterrupted production.
be met using materials such as oil-soluble resins, Regulations for the disposal of produced water in
benzoic acid flakes and other particulates. Particulate offshore operations are becoming increasingly strin-
diverters may cause damage and can be difficult to gent. The U.S. National Pollutant Discharge
clean up as a result. Elimination System (NPDES) has established dis-
Foams may also provide effective diversion charge levels for oil and grease at a 29-mg/L
(Zerhboub et al., 1991). Foamed fluids have the monthly average with a grab sample maximum of
additional advantage of good cleanup with little or 42 mg/L. This is a reduction in the allowable levels
no potential for damaging the formation. The ratio from 48 to 29 mg/L. The penalty for noncompliance
of nitrogen to fluid depends on the bottomhole con- by the operator can be severe. These discharge levels
ditions during the treatment. Diversion and fluid are causing the reevaluation of acidizing treatments.
placement strategy are addressed in Chapter 19. During the flowback immediately following an
acid treatment, significantly higher levels of oil and
grease than those permitted for discharge have also
been recorded. During flowback, spikes in the basic
15-12. Additive compatibility sediment and water (BS&W) values occur immedi-
All additives should be tested in the laboratory. ately after an acid treatment. Microemulsions are sus-
Compatibility of additives with formation fluids pected of occurring during flowback. The produced
should be tested using a sandpack method developed oil is solubilized in the returned spent acid, and this
by Ali et al. (1994). This procedure is more repre- results in significantly higher levels of oil and grease.
sentative of downhole conditions and uses more sand The use of matrix additives is believed to contribute
than the test in API Recommended Practices for to the solubilization of the oil and additives. These
Laboratory Testing of Surface Active Agents for Well additives should be optimized to minimize emulsion
Stimulation (RP 42) (1977). The acid system with upsets and the levels of oil and grease.
additives is drawn through a pack of formation core Verification and optimization of the additives used
material or a mixture consisting of 10-wt% illite- in acid stimulation treatments have been requested
bentonite, 10-wt% silica flour and 80-wt% 100-mesh by operators. Obtaining an answer to treatment facil-
sand. There are no universal additives for all forma- ity upsets by additive optimization in matrix treat-
tion problems. Any potential incompatibilities ments can have a tremendous impact on acidizing
between additives and formation fluids must be iden- workover operations. Research by Bansal (1993)
tified and corrected before acidizing. determined the following maximum concentrations
of inhibitors, surfactants and mutual solvents that
may be present in the returned fluids to maintain oil
15-13. Facility upsets following and grease levels below 48 mg/L:
acid stimulation • inhibitors: 80 ppm
• surfactants: 10 ppm
15-13.1. Discharge requirements
• mutual solvents: 100 ppm.
The chemicals used in acidizing treatments have
The common denominator for each of these addi-
been linked with oil and water separation facility
tives is surface-active compounds (i.e., surfactants).
upsets, causing noncompliance with environmental

The surfactants provide acceptable fluid perfor- et al. (1994) recommended an alternate procedure
mance, such as corrosion inhibition, dispersion and that better simulates downhole intermixing of the
wetting properties, as they were initially designed for stimulation fluids and the produced crude. The pro-
the acid treatment. However, if the additives are used cedure involves filtering formulated acid, containing
in excessive concentrations, they may exacerbate the inhibitor, surfactants, mutual solvents and nonemul-
emulsification tendencies of fluids during flowback. sifiers, through a sandpack containing 10-wt% silica
The difference in emulsion stability is attributed to flour, 10-wt% clays (such as montmorillonite) and
other, more dominant factors (e.g., stabilization by 80-wt% 100-mesh sand prior to completion of the
solids precipitation [organic and inorganic], ferric emulsion test. This procedure simulates the filtration
ion interaction with the crude oil and viscosity of the acid system within the rock matrix. Surfact-
increase at the interface). Results presented by ants, such as those used in nonemulsifiers, are
Coppel (1975) show that acids partially spent on for- adsorbed onto mineral surfaces. The adsorption
mations contain potentially precipitable materials in minimizes the availability of the surfactants in solu-
solution. As the pH of the produced acid increases tion to prevent emulsions.
during flowback and mixes with other produced flu- Durham et al. (1995) reported that oil and water
ids, precipitates form that can stabilize emulsions. emulsion upsets could be minimized by proper selec-
Dunlap and Houchin (1990) recommended polarized tion of the acidizing additives and surface treating
microscopy to evaluate return fluids to establish the chemicals. When acid flowbacks are introduced into
cause for a stabilized emulsion. Specific damage a production system, the control of emulsions by
mechanisms, such as emulsion stabilization by chemical treatment can significantly change in com-
organic deposits, solids and iron, could be detected parison with control during production of the well
and identified. prior to treatment. These authors completed exten-
sive emulsion breaker and water clarification tests
with the desired acid/additive systems and fresh
15-13.2. Prevention of facility upsets crude oil to duplicate the commingling of the acid
Preventing upsets requires optimizing the fluid design with the oil during flowback. The low pH of the
to minimize or eliminate the formation of precipitates water in the emulsion rendered the current treating
and using the optimum de-emulsifier. The potential program ineffective until the acid flowback was
for the formation of emulsions of acid and produced completed. Nonionic de-emulsifiers were found to be
crude oil while still providing the required downhole more effective than sulfonate-type additives for treat-
function optimization of the de-emulsifier treatment ing these emulsions. On-site testing is recommended
is evaluated currently by the procedures recommend- to select an effective chemical treatment.
ed in API RP 42 (1977). Vigorous mixing of acid Ali et al. (1997) reported the optimization of
with fresh crude oil in a glass container is observed acidizing additives to reduce production facility
under bottomhole temperatures (hot-water bath) to upsets. The authors combined additive treatment
determine the characteristics and time for emulsion optimization with a new absorption/filtration process.
breakout. Fine siliceous materials, ferric iron or both The optimized treatments involved the evaluation of
can be added to evaluate emulsion stability. surfactants and mutual solvents, which are consid-
These fluid interactions can simulate the downhole ered to be the additives primarily responsible for
emulsion problems encountered; however, these tests emulsion problems during acid flowback. Elimina-
fail to accurately predict the interaction of the treating tion of some additives, minimizing the concentra-
fluids with the acid-sensitive crudes during flowback. tions of other additives and using a dual-purpose
As a result, marginal treatment response is observed. mutual breakout solvent reduced facility upsets. All
Downhole emulsions and facility upsets may be recommendations were based on laboratory testing
attributed to inadequate prejob compatibility testing, using the sandpack method and on-site testing with a
because the additives employed in the acid system laboratory unit. The filtration process used a special-
may preferentially adsorb out onto the rock matrix or ized absorption medium to aid in lowering oil and
partition to the oil phase during production. grease levels. This combination of treatment opti-
Alternate testing procedures have been developed mization with filtration/absorption reduced facility
to more accurately simulate reservoir conditions. Ali upsets while not interfering with the efficacy of the
treatment.

◆
Fundamentals of
Acid Stimulation
A. Daniel Hill and Robert S. Schechter, University of Texas at Austin
16-1. Introduction or other mechanical aspects of the completion. Two

exceptions to this rule may occur. First, in highly pro-
Matrix stimulation is a technique in which a solvent ductive wells, the productivity improvement of about
is injected into the formation to dissolve some of the 20% that is possible with matrix stimulation of an
materials present and hence recover or increase the undamaged well may be economic. Second, in natu-
permeability in the near-wellbore region. Such treat- rally fractured or highly vugular carbonate reservoirs,
ments are called “matrix” treatments because the sol- live acid may penetrate to a sufficient distance to yield
vent is injected at pressures below the parting pressure a productivity enhancement greater than that normally
of the formation so that fractures are not created. The expected from a true matrix treatment.
objective is to greatly enhance or recover the perme- An ideal matrix treatment restores the permeability
ability near the wellbore, rather than affect a large in the near-wellbore region to a value at least as high
portion of the reservoir. as the original undamaged permeability; it accom-
The most common matrix stimulation treatment plishes this over the entire completed interval and it
is acidizing, in which an acidic solution is injected leaves the formation in the treated region with high
to dissolve minerals in the formation. However, other relative permeability to the oil and/or gas phase.
solvents are also used. The next most common fluids Designing a treatment should strive to achieve this
are organic solvents aimed at dissolving waxes, paraf- ideal at the lowest possible cost, which requires con-
fins, asphaltenes or other organic damaging materials. sideration of the many physical and chemical interac-
Nonacid matrix stimulation is addressed in Chapter 14. tions taking place between the injected fluids and the
This chapter focuses on matrix acidizing; however, the reservoir minerals and fluids. The most important of
reader should keep in mind that many of the theories these phenomena are the following:
and calculation procedures presented here can also be
applied to nonacid solvent treatments. • mass transfer of acid molecules to the mineral sur-
The most common acids are hydrochloric acid face and subsequent reaction at the surface—This
(HCl), used primarily to dissolve carbonate minerals, fundamental process of acidizing is illustrated in
and mixtures of HCl and hydrofluoric acid (HF), used Fig. 16-1. Acid reactions with minerals are termed
to attack silicate minerals such as clays and feldspars. heterogeneous reactions because they occur at a
Other acids, particularly some weak organic acids, boundary between the solid and the liquid rather
are used in special applications, such as high-temper- than in the bulk phases. Before the reaction can
ature wells. Matrix acidizing is a near-wellbore treat- occur, acid must be transported to the mineral sur-
ment, with all the acid reacting within about 1 ft of face by convection or diffusion. The overall reac-
the wellbore in sandstone formations and within a tion rate (i.e., the rate of change of the concentra-
few inches to perhaps as much as 10 ft from the well- tion of one component in the bulk liquid phase)
bore in carbonates. may depend on both the rate of mass transfer and
Matrix acidizing can significantly enhance the pro- the rate of surface reaction. Many times, however,
ductivity of a well when near-wellbore formation one of these processes is much slower than the
damage is present and, conversely, is of limited bene- other and controls the overall rate, in which case
fit in an undamaged well, as shown in Chapter 1. the faster process can be ignored.
Thus, matrix acidizing generally should be applied • changing pore structure—The physical change in
only when a well has a high skin factor that cannot be the pore structure caused by dissolution of some of
attributed to partial penetration, perforation efficiency the minerals by acid is the mechanism by which

• fluid selection—acid type, concentration and volume

• injection schedule—planned rate schedule and
sequence of injected fluids
Bulk solution
• acid coverage and diversion—special steps taken to
improve acid contact with the formation
Acid transport Product transport • real-time monitoring—methods to evaluate the
by convection by convection acidizing process as it occurs
or diffusion or diffusion
• additives—other chemicals included in the acid
Acid
solution to enhance the process or to protect tubular
concentration goods.
near the
Heterogeneous surface Treatment design is considered in detail in other
reaction at Reactive mineral
solid/liquid chapters. This chapter lays the foundation for the
interface design methods aimed at optimizing a matrix acidizing
treatment by reviewing the underlying chemistry and
Figure 16-1. Acid reaction occurring in a system. physics of the acidizing process and introducing the
latest models of the processes involved. First, the inter-
action of acids with reservoir minerals is addressed.
matrix acidizing increases permeability. The manner Then, current models of the matrix acidizing process
in which the pore structure changes is fundamental- in sandstones and in carbonates are presented.
ly different in sandstones and carbonates, which
leads to radically different approaches to modeling
the acidizing process in these two mineralogies. 16-2. Acid-mineral interactions
• precipitation of reaction products—Secondary reac-
tions occur in acidizing, particularly in sandstones, 16-2.1. Acid-mineral reaction stoichiometry
that can result in the precipitation of reaction prod-
The amount of acid required to dissolve a given
ucts from the bulk liquid phase. Obviously, precipi-
amount of mineral is determined by the stoichiometry
tated solids may block pore spaces and work
of the chemical reaction, which describes the number
against the goal of matrix acidizing.
of moles of each species involved in the reaction. For
• acid fluid–reservoir fluid interactions—The acid example, the simple reaction between HCl and calcite
solution injected in matrix acidizing may interact (CaCO3) can be written as
physically and/or chemically with the reservoir flu-
ids as well as with the minerals. These interactions 2HCl + CaCO3 → CaCl2 + CO2 + H2O
can result in changes in wettability, phase saturation which shows that 2 moles of HCl are required to dis-
distribution, precipitation of solids or emulsification. solve 1 mole of CaCO3. The numerals 2 and 1 multi-
• variations in reservoir permeability or the distribu- plying the species HCl and CaCO3 are the stoichio-
tion of damage—A successful acidizing treatment metric coefficients νHCl and νCaCO 3 for HCl and
requires contacting all damaged regions around the CaCO3, respectively.
well with acid. This is usually complicated by vari- When HF reacts with silicate minerals, numerous
ations in the injectivity to acid along the wellbore, secondary reactions may occur that influence the over-
which leads to the use of techniques to affect good all stoichiometry of the reaction. For example, when
acid coverage (acid diversion). HF reacts with quartz (SiO2), the primary reaction is
In considering the many aspects of the matrix 4HF + SiO2 SiF4 + 2H2O
acidizing process, the focus is on the key design vari- producing silicon tetrafluoride (SiF4) and water. The
ables; to be useful, any model of the process must aid stoichiometry of this reaction shows that 4 moles of
in optimizing the design. The primary design consid- HF are required to consume 1 mole of SiO2. However,
erations are the SiF4 produced may also react with HF to form flu-
osilicic acid (H2SiF6) according to
16-2 Fundamentals of Acid Stimulation

SiF4 + 2HF H2SiF6 In modeling the acidizing process, it is therefore

crucial to determine the local solution composition
If this secondary reaction goes to completion, 6 moles
as the process progresses. To satisfy this requirement,
of HF, rather than 4 moles, will be consumed to dis-
Sevougian et al. (1992) developed a geochemical
solve 1 mole of quartz. A complication is that the flu-
model that considers a local partial equilibrium among
osilicates may exist in various forms, so that the total
certain reactions involved in the acidizing process.
amount of HF required to dissolve a given amount of
This is an important modeling capability; for example,
quartz depends on the solution concentration.
Shuchart and Gdanski (1996) observed that the
Typical reactions involved in HF acidizing are sum-
AlF2+/AlF 2+ ratio obeys a pseudoequilibrium relation
marized in Table 16-1. Only the primary reaction prod-
over a range of reaction conditions. This implies local
ucts, AlF2+ and SiF4, are shown although many other
equilibrium among the various aluminum fluoride
products are possible (Labrid, 1975; Hekim and Fogler,
species in the aqueous phase. To utilize the full capabil-
1982; Walsh et al., 1982; Sevougian et al., 1992;
ity of the Sevougian et al. model, the rates of both the
Shuchart and Gdanski, 1996). How many moles of
homogeneous and heterogeneous reactions that are not
mineral are dissolved by a mole of acid is an important
fully at equilibrium are required. Not all the important
consideration in the selection of an acid treatment vol-
reaction rates are known.
ume that may differ substantially from the values
Thus, the chemistry of the reactions of HF with the
shown in Table 16-1. As an example of the uncertainty,
minerals found in sandstone formations is complex
Schechter (1992) suggested that for conservative design
and difficult to model. An additional complexity is the
purposes, about 20 moles of acid are required to dis-
tendency for some of the reaction products to precipi-
solve 1 mole of feldspar rather than the 14 moles
tate as the acid reactions go to completion (see
shown in Table 16-1. Depending on the composition of
Section 16-2.3). For practical purposes, it is conve-
the acid solution, the actual value may be substantially
nient to express the stoichiometry in terms of the
less. This is especially the case if the silicon extracted
approximate “dissolving power,” as introduced by
from the feldspar crystals appears finally as the precipi-
Williams et al. (1979).
tate Si(OH)4 rather than as soluble SiF4 as shown in
The dissolving power expresses the amount of min-
Table 16-1. If the final silicon product is, in fact,
eral that can be consumed by a given amount of acid
Si(OH)4, then only 1 mole of HF would be required
on a mass or volume basis. First, the gravimetric dis-
to dissolve a single mole of albite.
solving power β, which is the mass of mineral con-
sumed by a given mass of acid, is defined as
Table 16-1. Primary chemical reactions in acidizing.
ν mineral MWmineral
β= , (16-1)
HCl ν acid MWacid
Calcite 2HCl + CaCO3 → CaCl2 + CO2 + H2O
where the ν terms are the stoichiometric coefficients
Dolomite 4HCl + CaMg(CO3)2 → CaCl2 + MgCl2
+ 2CO2 + 2H2O and MWmineral and MWacid are the molecular weights of
Siderite 2HCl + FeCO3 → FeCl2 + CO2 + H2O
the mineral and the acid, respectively. Thus, for the
reaction between 100% HCl and CaCO3,
HCl-HF
(1)(100.1)
Quartz 4HF + SiO2 SiF4 (silicon tetrafluoride)
β100 = = 1.37 lbm CaCO3 /lbm HCl, (16-2)
+ 2H2O
(2)(36.5)
SiF4 + 2HF H2SiF6 (fluosilicic acid)
where the subscript 100 denotes 100% HCl. The dis-
Albite solving power of any other concentration of acid is
(sodium feldspar) NaAlSi3O8 + 14HF + 2H+ +
Na + AlF +
β100 times the weight fraction of acid in the acid solu-

2
+ 3SiF4 + 8H2O
Orthoclase tion. For the commonly used 15% HCl, β15 = 0.15 ×
(potassium β100 = 0.21 lbm CaCO3/lbm HCl. The stoichiometric
feldspar) KAlSi3O8 + 14HF + 2H+ K+ + AlF2+
+ 3SiF4 + 8H2O
coefficients for common acidizing reactions are found
from the reaction equations in Table 16-1.
Kaolinite Al4Si4O10(OH)8 + 24HF + 4H+ 4AlF2+
+ 4SiF4 + 18H2O The volumetric dissolving power X, which is simi-
Montmorillonite Al4Si8O20(OH)4 + 40HF + 4H+ 4AlF2+
larly defined as the volume of mineral dissolved by a
+ 8SiF4 + 24H2O given volume of acid, is related to the gravimetric dis-
solving power by

ρacid solution 16-2.2. Acid-mineral reaction kinetics

X =β . (16-3)
ρmineral
The reaction between an acid and a mineral occurs
A 15% HCl solution has a specific gravity of about when acid reaches the surface of the mineral by diffu-
1.07 and CaCO3 has a density ρ of 169 lbm/ft3. For sion or convection from the bulk solution. The overall
the reaction of these species, the volumetric dissolving rate of acid consumption or mineral dissolution depends
power is on two distinct phenomena—the rate of transport of
acid to the mineral surface by diffusion or convection
 lbm 15% HCl 
lbm CaCO3   (
1.07)(62.4) 3 and the actual reaction rate on the mineral surface.
X15 = 0.21 ft 15% HCl  Usually, one of these processes is much slower than
 lbm 15% HCl   lbm 15% CaCO3 
 169 3  the other. In this case, the fast process can be ignored,
 ft 15% CaCO3  because it can be thought of as occurring in an insignif-
ft 3 CaCO3 icant amount of time compared with the slow process.
= 0.082 .
ft 3 15% HCl (16-4) For example, the HCl-CaCO3 reaction rate is
extremely high, so the overall rate of this reaction is
The dissolving powers of various acids with limestone usually controlled by the rate of acid transport to the
and dolomite and of HF with quartz and albite are surface, the slower of the two processes. On the other
given in Tables 16-2 and 16-3, respectively (Schechter, hand, the surface reaction rates for many HF-mineral
1992). Sidebar 16A is an example calculation using reactions are slow compared with the acid transport
dissolving power. rate, and the overall rate of acid consumption or min-
eral dissolution is reaction-rate controlled. The “kinetics”
Table 16-2. Dissolving power of various acids (Schechter, 1992).
X (%)
Formation Acid β100 5 10 15 30
Limestone (CaCO3) Hydrochloric (HCl) 1.37 0.026 0.053 0.082 0.175

Formic (HCOOH) 1.09 0.020 0.041 0.062 0.129
Acetic (CH3COOH) 0.83 0.016 0.031 0.047 0.096
Dolomite (CaMg(CO3)2) Hydrochloric 1.27 0.023 0.046 0.071 0.152

Formic 1.00 0.018 0.036 0.054 0.112
Acetic 0.77 0.014 0.027 0.041 0.083
Notes: ρCaCO 3 = 2.71 g/cm3, ρCaMg(CO 3)2 = 2.87 g/cm3
Table 16-3. Dissolving power of hydrofluoric acid (Schechter, 1992).
Acid concentration (wt%) Quartz (SiO2) Albite (NaAlSi2O8)

β X β X
2 0.015 0.006 0.019 0.008

3 0.023 0.010 0.028 0.011
4 0.030 0.018 0.037 0.015
6 0.045 0.019 0.056 0.023
8 0.060 0.025 0.075 0.030
Notes: β = mass of rock dissolved/mass of acid reacted, X = volume of rock dissolved/volume of acid reacted

16A. Calculating minimum acid volume using dissolving power
The volume of acid required for a matrix acidizing treatment can be estimated using the concept of dissolving power. Because the vol-
umetric dissolving power X is the volume of a particular mineral that is dissolved by a given volume of a particular acid solution, the
minimum acid requirement to remove that mineral can be calculated with little information other than the dissolving power. Consider
the following problem:
A sandstone formation with a porosity of 0.2 contains 5-vol% albite (sodium feldspar). What is the minimum volume of 3% HF solu-
tion required to dissolve all the albite a distance of 6 in. beyond a 6-in. diameter wellbore?
Solution
The minimum acid volume is the amount VHF required to dissolve all the feldspar plus the amount Vp required to fill the pore space in
the region of feldspar dissolution. These volumes are
Vfeldspar = π(rHF2 − rw2 )(1− φ)x feldspar = π(0.75 2 − 0.25 2 )(1− 0.2)(0.05)
ft 3 (16A-1)
= 0.063 feldspar
ft
Vfeldspar 0.063 ft 3 HF
VHF = = = 5.7 . (16A-2)
X3 0.011 ft
In these equations, rHF is the radial penetration distance of HF, rw is the wellbore radius, φ is the porosity, and xfeldspar is the volume frac-
tion of the sandstone that is feldspar. The volume of pore space within 6 in. of the wellbore after removal of the feldspar is
( ) (
Vp = π(rHF2 − rw2 ) φ + x feldspar (1− φ) = π(0.75 2 − 0.25 2 ) 0.2 + 0.05(1− 0.2) )
ft 3 (16A-3)
= 0.38 ,
ft
so the total volume of HF required is
ft 3  7.48 gal  gal

VHF ,T = VHF + V p = (5.7 + 0.38)   = 46 . (16A-4)
ft  ft 3  ft
Thus, the minimum volume of 3% HF solution required to remove all feldspar in a radial region extending 6 in. beyond the wellbore
is 46 gal/ft of reservoir thickness. In an actual acidizing treatment, the injected acid does not react with feldspar only, and as shown by
examining models of the acidizing process, the acid is not spent uniformly, as tacitly assumed in this calculation. Nevertheless, this
simple calculation provides a ballpark figure for acid requirements and is a handy check of more complex models of the process.
of a reaction is a description of the rate at which the eral RB, which is related to the acid consumption rate
chemical reaction takes place, once the reacting through the stoichiometry of the reaction
species have been brought into contact. ν
A reaction rate is generally defined as the rate of RA = A RB , (16-6)
νB
appearance in the solution of the species of interest in
units of moles per second (mol/s). A surface reaction where νA and νB are the stoichiometric coefficients for
rate depends on the amount of surface exposed to acid A and mineral B.
reaction, so these reactions are expressed per unit of The reaction rate rA generally depends on the con-
surface area. In general, the surface reaction rate of an centrations of the reacting species. However, in the
aqueous species of acid A reacting with mineral B is reaction between an aqueous species and a solid, the
concentration of the solid can be ignored, because it
RA = rA SB , (16-5) remains essentially constant. For example, a grain of
quartz has a fixed number of moles of quartz per unit
where RA is the rate of appearance of acid A in mol/s,
volume of quartz, irrespective of reactions that may
rA is the surface area-specific reaction rate of A in
be occurring on the surface of the grain. Incorporating
mol/s-m2, and SB is the surface area of mineral B.
concentration dependence into the rate expression
When A is being consumed, the reaction rates rA and
yields
RA are negative. Acid-mineral reaction rates are typi-
cally expressed as the rate of dissolution of the min- − RA = E f CAα SB , (16-7)

where Ef is a reaction rate constant in mol A/[m2-s- The constants α, Efo and ∆E/R are listed in
(mol A/m3)α], CA is the concentration of species A at Table 16-4. SI units are used in these expressions,
the reactive surface, and α is the order of the reaction so CHCl is in kg-mol/m3 and temperature T is in
(i.e., a measure of how strongly the reaction rate degrees Kelvin. The reaction rate rHCl is expressed
depends on the concentration of A). The reaction rate as kg-mol HCl reacted/m2-s.
constant depends on temperature and sometimes on
• Reactions of hydrochloric and weak acids with
the concentration of chemical species other than A.
carbonates
Finally, Eq. 16-7 is written in the conventional manner
for a species of acid that is being consumed from HCl is a strong acid, meaning that when HCl is
solution, by placing a minus sign with RA so that dissolved in water, the acid molecules almost com-
Ef is a positive number. pletely dissociate to form hydrogen ions (H+) and
chloride ions (Cl–). The reaction between HCl and
• Laboratory measurement of reaction kinetics carbonate minerals is actually a reaction of the H+
To measure the surface reaction rate of acid-mineral with the mineral. With weak acids, such as acetic
reactions, it is necessary to maintain a constant or formic acid, the reaction is also between H+ and
mineral surface area or measure its change during the mineral, with the added complication that the
reaction and to ensure that the rate of acid transport acid is not completely dissociated, thus limiting the
to the mineral surface is fast relative to the reaction supply of H+ available for reaction. Because H+ is
rate. The two most common methods of obtaining the reactive species, the kinetics of the HCl reac-
these conditions are with a well-stirred slurry of tion can also be used for weak acids by considering
mineral particles suspended in an acid solution (a the acid dissociation equilibrium.
stirred reactor) or with a rotating disk apparatus The kinetics of a weak acid–carbonate mineral
(Fogler et al., 1976.) In the rotating disk apparatus, reaction may therefore be obtained from Eq. 16-8
a disk of the mineral is placed in a large container as follows (Schechter, 1992):
holding the acid solution. The disk is rotated rapidly,
− rweak acid = E f K dα / 2 Cweak
α/2
acid , (16-10)
so that the acid mass-transfer rate is high relative
to the surface reaction rate. A third, more indirect where Kd is the dissociation constant of the weak
method is by matching the coreflood response to acid and Ef is the reaction rate constant for the HCl-
acidizing with a model of the process of flow with mineral reaction.
reaction.
Lund et al. (1975, 1973) measured the kinetics • Reactions of hydrofluoric acid with sandstone
of the HCl-calcite and HCl-dolomite reactions, minerals
respectively. Their results may be summarized as HF reacts with virtually all of the many mineral
constituents of sandstone. Reaction kinetics have
− rHCl = E f CHCl
α
(16-8)
been reported for the reactions of HF with quartz
(Bergman, 1963; Hill et al., 1977), feldspars
∆E 
E f = E of exp − . (16-9) (Fogler et al., 1975) and clays (Kline and Fogler,
 RT 
Table 16-4. Constants in HCl-mineral reaction kinetics models.
∆E
Mineral α Ef o (K)
R
 
 kg - mol HCl 
 α

 2  kg - mol HCl  
 m -s 3  
 m acid solution  
Calcite (CaCO3) 0.63 7.291 × 107 7.55 × 103
Dolomite (CaMg(CO3)2) 6.18 × 10 −4T 9.4 × 1011 11.32 × 103

1− 2 × 10 −3T 1000 α

1981a). These kinetic expressions can all be repre- tact with the acidic solution. Therefore, the mor-
sented by phology of the mineral assemblage becomes
[ ]
an important issue.
− rmineral = E f 1 + K (CHCl ) CHF
β α
, (16-11) Comparison of the reaction rates of various min-
erals requires placing the rates on the basis of a unit
for which the parameters α, β, Ef and the empirical of reactive area. On this basis, montmorillonites
kinetic constant K are listed in Table 16-5. and kaolinites react about 2 orders of magnitude
These expressions show that the dependence on slower than feldspars, and illites react at least 1
HF concentration is approximately first order (α = order of magnitude slower than kaolinite. Viewing
1). For feldspar reactions, the reaction rate increases thin sections of rocks following acid treatment with
with increasing HCl concentration, although HCl is HCl-HF mixtures shows that the feldspars are usu-
not consumed in the reaction. Thus, HCl catalyzes ally removed because of their high specific reaction
HF-feldspar reactions. Kline and Fogler (1981a) rates. Authogenic clays also appear to react rapidly
showed that the reactive area depends on the crystal- because of their intimate exposure to the acidic
line structure of the clay reacting with an HCl-HF solution. On the other hand, clastic clays are com-
mixture and is generally only a small fraction of monly found in thin sections following acid treat-
the total surface area of clays as determined by ment (Hill et al., 1977). Thus, it is not only the
traditional methods of measurement. Thus, the specific reaction rate but also the area in contact
surface area of montmorillonite as determined by with the acid that determines the rate of removal of
nitrogen (N2) adsorption may be as high as 5 × 105 a specific mineral. An example calculation of rela-
m2/kg, whereas the reactive surface area is approxi- tive reaction rates of sandstone minerals is in
mately 104 m2/kg. The surface areas in Eq. 16-8 Sidebar 16B.
must be the reactive areas that are actually in con-
Table 16-5. Reaction rate constants for Eq. 16-11.
Mineral Ef K β α
 
 kg-mol minera l 
    kg-mol HF  −β 
α
 2  kg-mol HF      
 
 m - s    m3
 m3  
Potassium feldspar†
(orthoclase) 0.127 exp  − 4680  5.66 × 10–2 exp  956  0.4 1.2
 T   T 
Sodium feldspar†
(albite) 9.50 × 10–3 exp  − 3930  6.24 × 10–2 exp  554  1.0 1.0
 T   T 
α-quartz‡ 1.39 × 10–7 exp  − 1150  0 – –

 T 
Montmorillonite§ 1.1 × 10–2 exp  − 5200  0 – 1.0

 T 
Kaolinite§ 0.86 exp  − 6800  0 – 1.0

 T 
†
Fogler et al. (1973)
‡
Adapted from Hill et al. (1977)
§
Adapted from Kline and Fogler (1981)

16B. Relative reaction rates of sandstone minerals
A matrix acidizing treatment is aimed at overcoming the effects of near-wellbore formation damage. Ideally, the injected acid attacks
only the material causing the damage, which in most instances is clay particles or other fines. How efficiently the acid is being used
can be determined by calculating the reaction rates of all major mineral species present with the injected acid.
Consider a sandstone formation that has been damaged by the invasion of bentonite (montmorillonite) particles from drilling mud.
After the carbonate minerals have been removed by an HCl preflush, this clean sandstone contains 90% quartz, 5% albite (sodium
feldspar) and 5% montmorillonite by weight. The reactive surface areas of the minerals are 10 m2/kg for the quartz and albite and
8000 m2/kg for the montmorillonite. (A cube of quartz with a side of 1 mm has a surface area of 2.2 m2/kg; if it is 0.1 mm on a side, its
surface area is 22 m2/kg. Clays have a much larger surface area than detrital grains of quartz or feldspar.) Assume stoichiometric
ratios of 6 moles HF/mole quartz, 20 moles HF/mole feldspar and 40 moles HF/mole montmorillonite.
If this rock is contacted with a 12% HCl–3% HF solution at 125°F [50°C], what proportion of the HF will initially be consumed by
each of the three minerals? What are the mineral proportions of the rock dissolved?
Solution
Per unit mass of rock, the surface area of each mineral is its reactive surface area times the mass fraction of the mineral present in
the sandstone. For example, the reactive surface area of quartz per mass of sandstone Sq is (10 m2/kg)(0.9) = 9 m2/kg rock. Similarly,
the surface areas of feldspar and montmorillonite are 0.5 and 400 m2/kg rock, respectively. The acid concentrations in Eq. 16-11 are in
units of kg-mol/m3 solution (equivalent to gmol/L); the concentrations given as mass fractions are converted to these units by multiply-
ing by the solution density and the acid molecular weight, yielding 1.61 kg-mol HF/m3 solution and 3.53 kg-mol HCl/m3 solution. Quartz
is used to illustrate the calculation sequence to determine the reaction rates for each mineral.
First, the rate constant is calculated with the data from Table 16-5:
 1150  kg-mol quartz

E f = 1.39 × 10 −7 exp  −  = 3.95 × 10
−9
.
 273 + 50   kg-mol HF 
m2-s  3  (16B-1)
 m solution 
Then, the specific reaction rate for quartz from Eq. 16-11 is
kg-mol quartz
−rq = 6.59 × 10 −10 (1.61) = 1.06 × 10 −9 . (16B-2)
m 2-s
The overall reaction rate for quartz is the specific reaction rate multiplied by the reactive surface area:
kg-mol quartz
−R q = 1.06 × 10 −9 (9) = 9.54 × 10 −9 , (16B-3)
kg rock-s
which is multiplied by the molecular weight of quartz to put it on a mass basis:
kg quartz
−R q = 9.54 × 10 −9 (60.1) = 5.73 × 10 −7 . (16B-4)
kg rock-s
Finally, the rate of consumption of HF by the quartz reaction is obtained with Eq. 16-9, assuming 6 moles of HF are consumed for
each mole of quartz dissolved:
kg-mol HF
−R HF ,q = 9.54 × 10 −9 (6) = 3.44 × 10 −7 . (16B-5)
kg rock-s
The results of these calculations for all three minerals are summarized in Table 16B-1.
The fraction of HF expended in a particular reaction is the overall reaction rate for the mineral divided by the sum of the reaction
rates, which shows that 1.1% of the HF is reacting with quartz, 5.7% is reacting with feldspar and 93.2% is reacting with montmoril-
lonite. On the basis of the mass of mineral being dissolved, 95.1% of the rock dissolved is clay, 4.3% is feldspar, and less than 0.6% is
quartz. This is because of the high surface area of the authogenic clays (including, however, clay particles from drilling muds) and the
low reactivity of the quartz. Because clay and feldspar have relatively high reaction rates and generally form a small portion of the total
rock mass, they are dissolved first in sandstone acidizing. The quartz reaction becomes important in regions where most of the clay,
except clastic clays, and feldspar have already been removed.
Table 16B-1. Relative reaction rates of sandstone minerals.
Mineral –ri Si –Ri –RHF,i

(kg-mol i /m2-s) (m2/kg rock) (kg i /kg rock-s) (kg-mol HF/kg rock-s)
Quartz 5.73 × 10–8 9 3.44 × 10–6 3.44 × 10–7

Feldspar 1.77 × 10–7 0.5 2.32 × 10–5 1.77 × 10–6
Clay 1.81 × 10 –9
400 5.19 × 10 –4
2.89 × 10–5
Note: The subscript i denotes the mineral.

16B. Relative reaction rates of sandstone minerals (continued)
Notes for Sidebar 16B

Given 90% quartz, 5% albite and 5% montmorillonite by weight, the surface areas are 10 m2/kg quartz, 10 m2/kg albite and
8000 m2/kg montmorillonite. The reaction rates used are from Table 16-5.
• Quartz (MW = 60.1)

Because 1 kg rock has 0.9-kg quartz, the quartz area is (0.9)(10)90 m2/kg rock.
 1150 
E f = 1.39 × 10 −7 exp −  = 3.95 × 10
−9
 273 + 50 
kg-mol quartz
−rq = E f C HF = 6.36 × 10 −9
m 2-s
−R q = (6.36 × 10 −9 ) (9) = 5.72 × 10 −8

kg-mol quartz
kg rock-s
−R q = (5.72 × 10 −8 ) (60.1) = 3.44 × 10 −6

kg quartz
kg rock-s
 5.72 × 10 −8 kg-mol quartz  mol HF

R HF =   (6) = 3.43 × 10
−7
 kg rock-s  kg rock-s
• Albite (MW = 262, β = 1.0, α = 1.0)

 554 
K = 6.24 × 10 −2 exp  = 0.3468
 323 
CHCl = 3.53 kg mol/m3
 3930 
E f = 9.5 × 10 −3 exp −  = 4.938 × 10
−8
 323 
 103 m3   0.05 kg albite   −7 kg-mol albite  −8 kg-mol albite

RA =    1.77 × 10  = 8.84 × 10
 kg albite   kg rock   m2-s  kg rock-s
kg albite
R A = 8.84 × 10 −8 (262) = 2.32 × 10 −5
kg rock-s
kg-mol albite  kg-mol HF  −6 kg-mol albite

R A = 8.84 × 10 −8  20  = 1.77 × 10
kg rock-s  kg-mol albite  kg rock-s
• Montmorillonite (MW = 720)
 5200 
E f = 1.1× 10 −2 exp −  = 1.12 × 10
−9
 323 
−rmontmorillonite = (1.12 × 10 −9 )(1.61) = 1.81× 10 −9

kg-mol montmorillonite
m2-s
 m2   0.05 kg montmorillonite  m2
surface area =  8000   = 400
 kg montmorillonite   kg rock  kg rock
kg-mol montorillonite
R montmorillonite = 7.22 × 10 −7
kg rock-s
kg montorillonite
R montmorillonite = 5.20 × 10 −4
kg rock-s
kg montorillonite
R montmorillonite = 7.22 × 10 −7 (40) = 2.89 × 10 −5
kg rock-s

• Reactions of fluosilicic acid with sandstone minerals rates, so that the potential precipitation zone is rapidly
As discussed in Section 16-2.1, fluosilicic acid is displaced away from the wellbore. Also, spent acid
produced when HF dissolves silicate minerals, and should be produced back immediately after the com-
the fluosilicic acid itself may then react with alumi- pletion of injection, because shutting in the well for
nosilicates. From models of coreflood experiments, even a relatively short time may allow significant sil-
Bryant (1991) and da Motta et al. (1992a, 1992b) ica precipitation to occur in the near-well vicinity.
suggested that the reaction between fluosilicic acid When ferric ions (Fe3+) are present, they can precip-
and clays and feldspars is slow at room tempera- itate from spent acid solutions as Fe(OH)3 when the
ture, but that it is of the same order of magnitude pH is greater than about 2. Ferric ions may result from
as the HF reactions with these minerals at tempera- the dissolution of iron-bearing minerals in an oxida-
tures above 125°F [50°C]. These conclusions have tive environment or may derive from the dissolution
been substantiated by more direct experimentation of rust in the tubing by the acid solution. When a high
(see Shuchart and Gdanski, 1996). level of ferric irons is likely in the spent acid solution,
sequestering agents can be added to the acid solution
to prevent the precipitation of Fe(OH)3. However,
16-2.3. Precipitation of reaction products Smith et al. (1969) suggested using these sequestrants
with caution, as they may cause more damage through
A major concern in acidizing, particularly the acidiz-
their own precipitation than would have been caused
ing of sandstones, is damage caused by the precipita-
by the iron.
tion of acid-mineral reaction products. In acidizing
Finally, in some reservoirs, contact of the crude
sandstones with HF, the formation of some precipi-
oil by acid can cause the formation of asphaltenic
tates is probably unavoidable. However, the amount
sludges. Simple bottle tests in which a sample of
of damage they cause to the well productivity depends
crude oil is mixed with the acid can indicate whether
on the amount and location of the precipitates. These
the crude has a tendency for sludge formation when
factors can be controlled to some extent with proper
contacted by acid. When sludge formation is a prob-
job design.
lem, emulsions of acid in aromatic solvents or sur-
The most common damaging precipitates that may
face-active additives have been used to prevent
occur in sandstone acidizing are calcium fluoride
asphaltene precipitation (Moore et al., 1965).
(CaF2), colloidal silica (Si(OH)4), ferric hydroxide
The tendency for precipitation reactions to occur in
(Fe(OH)3) and asphaltene sludges. Calcium fluoride
acidizing is predicted with comprehensive geochemical
is usually the result of the reaction of calcite with HF,
models of the chemical reactions between aqueous
according to
species and the host of minerals present. The most
CaCO3 + 2HF CaF2 + H2O + CO2 common type of geochemical model used to study
sandstone acidizing is the local equilibrium model,
Calcium fluoride is highly insoluble, so the precipi-
such as described by Walsh et al. (1982) and Faber et
tation of CaF2 is likely if any calcite is available to
al. (1994). This type of model assumes that all reac-
react with the HF. Inclusion of an adequate HCl pre-
tions are in local equilibrium; i.e., all reaction rates are
flush ahead of the HCl-HF stage prevents the forma-
infinitely fast. A typical result from this model is
tion of CaF2.
shown in Fig. 16-2, a time-distance diagram for the
Production of some colloidal silica precipitate is
injection of 11% HCl–4% HF into a formation con-
probably unavoidable in sandstone acidizing. The
taining calcite, kaolinite and quartz. This plot shows
equilibrium calculations of Walsh et al. (1982) show
regions where amorphous silica and aluminum fluoride
that there are virtually always regions where the spent
will tend to precipitate. A vertical line on the plot rep-
acid solution has the tendency to precipitate colloidal
resents the mineral species present as a function of dis-
silica. However, laboratory corefloods suggest that the
tance if all reactions are in local equilibrium. By
precipitation is not instantaneous and in fact may
coupling this model with a model of the formation per-
occur at a fairly slow rate (Shaughnessy and Kunze,
meability response to both dissolution and precipita-
1981) that, however, increases with temperature. To
tion, predictions of the productivity improvement
minimize the damage caused by colloidal silica, it is
expected from particular acid formulations may be
probably advantageous to inject at relatively high
obtained, as illustrated in Fig. 16-3 (Faber et al., 1994).

1.0 6
I
II
5
2.4
2.2
0.8 Productivity
1.4
III
2.6
2.6
Kaolinite improvement
and factor contours
calcite
HF concentration (%)
4
Fractional distance
1.8
0.6
Si(OH)4
AIF3 3
IV
Optimum
0.4 Si(OH)4
2
2.6
V 1.6
0.2
1.
1 2.2
2.
Dissolved zone
4
1.4
2.0
0 0
0 2 4 6 8 10 0 5 10 15 20 25
Pore volumes injected HCl concentration (%)
Figure 16-2. Time-distance diagram showing regions of Figure 16-3. Productivity improvement plot (Faber et al.,
possible precipitation (Schechter, 1992). 1994).
Recently, Sevougian et al. (1992) and Quinn (1994) damage than a local equilibrium model because the
presented a geochemical model that includes kinetics finite rate of the reactions allows displacing the pre-
for both dissolution and precipitation reactions (see cipitate farther from the wellbore.
Sidebar 16C). This model predicts less permeability
16C. Geochemical model predictions
An example presented by Quinn (1994) illustrates how acid formulation can be evaluated with a comprehensive geochemical model.
A high-quartz-content sandstone will be acidized with 100 gal/ft of 12% HCl–3% HF solution (commonly referred to as full-strength mud
acid). The mineralogy is illustrated in Fig. 16C-1. The region to be studied includes a damaged zone extending 6 in. beyond the wellbore.
Figure 16C-1.
High quartz (>80 vol%) Low clay (<5 vol%), low feldspar (<10 vol%) A representative
sandstone used in
acidizing simulation.
Microcline 6%
Quartz 85%
Albite 3%
Calcite 2%
Kaolinite 4%
The geochemical model predicts the distribution of acid and minerals after injection of the acid, as shown in Fig. 16C-2. All the HF
is consumed near the wellbore; some precipitation occurs, but the amorphous silica precipitation occurs beyond the damage zone,
where its effect is small. From these results, the porosity distribution around the wellbore is determined. Then, a model of the perme-
ability response generates a prediction of the productivity improvement expected for this treatment.

16C. Geochemical model predictions (continued)

By repeating this modeling procedure for many different acid concentrations, the optimal acid formulation can be determined
(Fig. 16C-3). For this formation, 3% HF and high HCl concentrations are optimal.
Figure 16-C-2. Partial local equilibrium

Productivity improvement factor = 1.77
1.0 assumption (PLEA) model mineral profile
Reservoir minerals
of a high-quartz sandstone after acidiza-
Concentration (g-mol/L bulk volume)
0.9
0.8 tion with 100 gal/ft of 12% HCl–3% HF
0.7 injected at 0.1 bbl/min/ft at 125°F [50°C].
0.6 Kaolinite
0.5 Kaolinite (D)
Sodium feldspar
0.4
Potassium feldspar
0.3
0.2
0.1
0
0.25 0.75 1.25 1.75 2.25 2.75 3.25 3.75 4.25
3.0
Precipitates
2.5
2.0
Colloidal silica
1.5 AIF3 (s)
Na2SiF6
1.0 K2SiF6
0.5
0
0.25 0.75 1.25 1.75 2.25 2.75 3.25 3.75 4.25
4.5
Acid concentration
4.0
3.5
H+
3.0 HF
2.5
2.0
1.5 Original
damaged
1.0 zone
0.5
0
0.25 0.75 1.25 1.75 2.25 2.75 3.25 3.75 4.25
Radial distance (ft)
Figure 16C-3. Productivity improvement plot.

Volume of Reservoir mineralogy Injection
acid injected rate
High-quartz sandstone
100 gal/ft at 0.1 bbl/min/ft
6.0
1.8
1.5
HF concentration (%)
4.5 1.7 Productivity

1.6 improvement
factor (PIF)
PIF contours
1.8
1.7
1.5 1.6
1.5 1.5
1.4 1.4
1.3 1.3
0 1.2 1.2
0 6 12 18 24
HCl concentration (%)

16-3. Sandstone acidizing (Brannon et al., 1987). The benefits of lower concen-
tration HF solutions are a reduction in damaging pre-
16-3.1. Introduction cipitates from the spent acid and lessened risk of
unconsolidation of the formation around the wellbore.
A typical acid treatment in sandstones consists of the The selection of acidizing fluids should always begin
injection of an HCl preflush, with 50 gal/ft of forma- with an assessment of the formation damage
tion a common preflush volume, followed by the present—in general, the damaging material must be
injection of 50 to 200 gal/ft of HCl-HF mixture. A soluble in the treating fluids. Geochemical models can
postflush of diesel, brine or HCl then displaces the be used to guide acid selection, once the composition
HCl-HF from the tubing or wellbore. Once the treat- of the damaged formation is determined, as described
ment is completed, the spent acid should be immedi- in Section 16-2.3. Chapters 17 and 18 provide a com-
ately produced back to minimize damage by the prehensive treatment of acid selection for sandstone
precipitation of reaction products. and carbonate reservoirs, respectively.
A sandstone acidizing treatment design begins with
the selection of the type and concentration of acid to
be used. The volumes of preflush, HCl-HF mixture and 16-3.3. Sandstone acidizing models
postflush required and the desired injection rate(s) are
considered next. In virtually all acid treatments, the • Two-mineral model
placement of the acid is an important issue—a strategy Numerous efforts have been made over the years
to ensure that sufficient volumes of acid contact all to develop a comprehensive model of the sandstone
productive parts of the formation should be carefully acidizing process that could then be used as a
planned. Proper execution of the treatment is critical design aid. The most common model in use today is
to acidizing success, so the conduct of the treatment, the two-mineral model (Hill et al., 1977; Hekim et
including the mechanical arrangements for introducing al., 1982; Taha et al., 1989) that divides all minerals
the acid to the formation and the methods of treatment into two categories—fast-reacting and slow-reacting
monitoring, should be planned in detail. Finally, numer- species. Schechter (1992) categorizes feldspars,
ous additives are incorporated with acid solutions for authogenic clays and amorphous silica as fast react-
various purposes. The types and amounts of additives ing, and detrital clay particles and quartz grains are
to be used in the treatment must be determined on the the primary slow-reacting minerals. The model con-
basis of the completion, formation and reservoir fluids. sists of material balances applied to the HF acid and
These design factors are considered in detail in other reactive minerals, which for linear flow, such as in a
chapters: acid selection in Chapters 13 and 18, treat- coreflood, are
ment design (rate and volume) in Chapter 18, fluid δ(φCHF ) δC
placement and diversion in Chapter 19, treatment moni-
δt δx
{ }
+ u HF = − SF* VF E f ,F + SS*VS E f ,S (1 − φ)CHF
toring and evaluation in Chapter 20 and acid additives
(16-12)
in Chapter 15. This section presents models of the
sandstone acidizing process that provide a foundation
δ − MWHF SF* VFβ F E f ,F CHF
for the design methods used for field application.
δt
[
(1 − φ)VF =] ρF
(16-13)
16-3.2. Acid selection δ − MWHF SS*VSβ S E f ,S CHF

The type and strength (i.e., concentration) of acid used δt
[
(1 − φ)VS =] ρS
. (16-14)
in sandstones are selected primarily on the basis of
field experience with particular formations. For years, In these equations, CHF is the concentration of
the standard sandstone acidizing formulation consisted HF in solution and MWHF is its molecular weight,
of a 12% HCl–3% HF mixture, preceded by a 15% u is the acid flux, s is the distance, SF* and SS* are
HCl preflush. In fact, the 12% HCl–3% HF mixture the specific surface areas per unit volume of solids,
has been so common that it is referred to generically VF and VS are the volume fractions, Ef,F and Ef,S are
as mud acid. In recent years, however, the trend has the reaction rate constants (based on the rate of
been toward the use of lower strength HF solutions consumption of HF), MWF and MWS are the molec-

ular weights, βF and βS are the dissolving powers a unit volume of rock pore space to the amount of
of 100% HF, and ρF and ρS are the densities of the mineral present in the unit volume of rock, which
fast- and slow-reacting minerals, respectively, for the fast-reacting mineral is
denoted by the subscripts F and S. When made
φ oβ F CHF
o
MWHF
dimensionless, assuming porosity remains constant, Ac( F ) = . (16-24)
these equations become ( o ) FoρF
1 − φ V
δψ δψ The Damköhler and acid capacity numbers for

+ + {Da ( F ) Λ F + Da ( S ) Λ S }ψ = 0 (16-15) the slow-reacting minerals are similarly defined.
δθ δε As acid is injected into a sandstone, a reaction
δΛ F front is established by the reaction between the HF
= − Da ( F ) Ac( F ) ψΛ F (16-16) and the fast-reacting minerals. The shape of this
δθ front depends on Da (F ). For low values of Da, the
convection rate is high relative to the reaction rate
δΛ S
= − Da ( S ) Ac( S ) ψΛ S , (16-17) and the front is diffuse. With a high Da, the reaction
δθ front is relatively sharp because the reaction rate
where the dimensionless variables are defined as is high compared with the convection rate. Figure
16-4 (da Motta et al., 1992a) shows typical concen-
CHF
ψ= o (16-18) tration profiles for high and low values of Da (F ).
CHF
VF 1.0
ΛF = (16-19) Ψ
VFo 0.9
(F)
ΛF
Da = 15
0.8 Da(S) = 0.43
V θ = 100
Λ S = So (16-20) 0.7 A c(F) = 0.006
VS A c(S) = 0
0.6
Ψ or ΛF
x 0.5 Da (F) = 0.4

ε= (16-21)
L 0.4
ut 0.3
θ= , (16-22)
φo L 0.2
0.1
where ψ is the dimensionless HF concentration, 0
Λ is the dimensionless mineral composition, ε is 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
dimensionless distance, θ is dimensionless time ε
(pore volumes), and φ is the porosity. The super- Figure 16-4. Acid and fast-reacting mineral concentration
script o denotes initial values prior to acid treat- profiles (da Motta et al., 1992a).
ment. For a coreflood, L is the core length. In
Eqs. 16-15 through 16-17, two dimensionless Equations 16-15 through 16-17 can be solved only
groups appear for each mineral: the Damköhler numerically in their general form. Numerical models
number Da and the acid capacity number Ac. These providing solutions to these equations, such as that
two groups describe the kinetics and the stoichio- presented by Taha et al. (1989), are frequently used
metry of the HF-mineral reactions. The Damköhler for acidizing design. However, analytical solutions
number is the ratio of the rate of acid consumption are possible for certain simplified situations.
to the rate of acid convection, which for the fast- Schechter (1992) presented an approximate solution
reacting mineral is that is valid for relatively a high Da (Da (F) > 10).
Da (F)
=
(1 − φ )V
o F
.
o
E (f F ) SF* L
(16-23)
This solution approximates the HF/fast-reacting-
mineral front as a sharp front, behind which all the
u
fast-reacting minerals have been removed. Conversely,
The acid capacity number is the ratio of the ahead of the front, no dissolution has occurred. The
amount of mineral dissolved by the acid occupying reaction between slow-reacting minerals and HF

behind the front serves to diminish the HF concen-

8
tration reaching the front. The location of the front is
7 Fast-reacting minerals present
θ=
(
exp Da ε f − 1(S)
) +εf, (16-25)
6 Reaction
(F) (S) front
A Dac 5
Fast-reacting minerals removed
x (cm)
which relates dimensionless time (or equivalently 4
acid volume) to the dimensionless position of the 3

front ε f, defined as the position of the front divided 2
by the core length for linear flow. The dimension- 1
less acid concentration behind the front is Perforation tunnel
0
ψ = exp( − Da ε ). (S) 0 4 8 12 16 20 24 28
(16-26)
z (cm)
A particularly convenient feature of this approxi- Figure 16-5. Ellipsoidal flow around a perforation
mation is that it is applicable to linear, radial and (Schechter, 1992).
ellipsoidal flow fields with the appropriate definition
of dimensionless variables and groups. Radial flow
These two positions should be sufficient for design
represents the flow of acid from an openhole, gravel-
purposes; the reader is referred to Schechter (1992)
pack or slotted liner completion and may also be a
for methods to calculate the complete acid penetra-
reasonable approximation of the flow from a perfo-
tion profile in this geometry (see Sidebar 16D).
rated well with sufficient perforation density. The
The characteristic lengths referred to in Table 16-6
ellipsoidal flow geometry approximates the flow
are the length of a core L, wellbore radius rw and
around a perforation (Fig. 16-5). The proper dimen-
length of the perforation lp. Different measures of
sionless variables and groups for these three flow
acid flow are used in which u is the linear flux in a
fields are given in Table 16-6. For the perforation
core, qi /h is the volumetric rate of acid injection per
geometry, the position of the front ε f depends on
foot into an openhole, and qperf is the volume of acid
position along the perforation. In Table 16-6, expres-
per time entering a perforation. The definition of Da
sions are given for the front position of the acid
must correspond to the geometry considered, but Ψ,
extending directly from the tip of the perforation
ΛF and Ac(F) as defined by Eqs. 16-18, 16-19 and
and for acid penetration along the wellbore wall.
16-24, respectively, apply to all geometries.
Table 16-6. Dimensionless groups in sandstone acidizing models.
Flow geometry ε θ Da (S )
Linear x ut (1− φ )V
o S
o
E f(S )SS* L
L φoL u
Radial r2
−1
q it (1− φ )V
o S
o
E f(S )SS* πrw2h
rw2 πrw2hφo qi
Ellipsoidal
Penetration from the tip of the perforation
1 3 2 z q perf t 2π(1− φo )l p3SS*VSoE f(S )
z −z + ; z =
3 3 lp 2πl φo
3
p
q perf
Penetration adjacent to the wellbore

3
1 1  1 x
x+  − 3; x = l
3  x + x 2 + 1 p

16D. Comparison of acid volumes for radial 350

and perforation flow
300
The amount of acid required to remove damage beyond the Perforation
tip of a perforation is typically larger than the amount required 250
Acid volume (gal/ft)

to remove damage to the same distance in radial flow. This is
illustrated by using Schechter’s (1992) model to find the vol- 200
ume of acid required to penetrate a given distance for these
two geometries. 150
Consider a formation with 20% porosity containing 10%
fast-reacting minerals (feldspar, clay or both), 5% calcium car- 100
bonate and 85% quartz. A damage region extends 6 in. into
the formation (a radial region or 6 in. beyond the tip of a per- Radial
50
foration), and the initial skin effect resulting from the damage
is 10. The wellbore radius is 0.328 ft. In the perforated well 0
case, there are 4 shots per foot (spf) and the perforations are 0 1 2 3 4 5 6
6 in. long with a diameter at the wellbore of 0.5 in. A 12%
HCl–3% HF solution is injected at 0.1 bbl/min/ft after the injec- Acid penetration distance (in.)
tion of sufficient preflush of 15% HCl to remove the carbon-
ates from the region to be contacted with live HF. Determine Figure 16D-1. Acid penetration for radial and perforation
the acid volume required as a function of the penetration dis- flow.
tance of the acid and the skin effect evolution for radial and
perforation geometries. The downhole treating temperature
is 125°F [50°C]. For these conditions, the Damköhler number
Da for the slow-reacting mineral in radial flow is 0.013. 5
Solution 4
Equation 16-25 can be used with the appropriate definitions
of the dimensionless variables and groups for the two geome-
3
tries from Table 16-6. The acid capacity number Ac is the
Perforation
same for either geometry; the ratio of the Damköhler numbers
Skin effect
is 2
3
(2) 12 
6
4
Da perf 213perf (SPF ) 1
= = = 93, (16D-1)
(0.328)
2 2 Radial
Darad rw
0
where SPF is the perforation density in spf. Da for the slow-
mineral reaction is calculated as 0.12 for perforation flow. Ac
for the fast-mineral reaction is 0.021. (The values of Da and –1
Ac used in this example were obtained from laboratory core- 0 100 200 300 400 500
flood tests as described by Economides et al., 1994).
Using these values in Eq. 16-25 for acid penetration rang- Acid volume (gal/ft)
ing from 0 to 6 in. obtains the results shown in Figs. 16D-1
and 16D-2. For acid penetrations beyond 2 in., more acid is Figure 16D-2. Reduction in skin effect value for radial
required for the perforation geometry than for radial flow. The and perforation flow.
skin effect evolution reflects the larger volumes of acid
required to penetrate through the damaged region for the
perforation geometry compared with the radial geometry.
It is interesting to note that the slow-reacting min- is assumed, implying that Da(F) is infinite. This solu-
eral Da and the fast-reacting mineral Ac are the only tion can be used to estimate the volume of acid
dimensionless groups that appear in this solution. required to remove the fast-reacting minerals from
Da(S) regulates how much live HF reaches the front; a given region around a wellbore or perforation.
if the slow mineral reacts fast relative to the convec- The dimensionless groups Da(S) and Ac(F) can be
tion rate, little acid is available to propagate the fast- calculated with Eqs. 16-23 and 16-24, respectively,
mineral front. Ac for the slow-reacting mineral is not and Table 16-6 on the basis of the rock mineralogy
important because the supply of slow-reacting min- or can be obtained from experiments.
eral is almost constant behind the front. Ac(F) directly
• Two-acid, three-mineral model
affects the frontal propagation rate—the more fast-
reacting mineral present, the slower the front will Recently, Bryant (1991) and da Motta et al. (1992b)
move. Da(F) does not appear because a sharp front presented evidence that the sandstone acidizing

process is not well described by the two-mineral the fast-reacting minerals with a given volume of
model, particularly at elevated temperatures. These acid because the fluosilicic acid also reacts with
studies suggest that the reaction of fluosilicic acid these minerals and the reaction product of silica gel
with aluminosilicate (fast-reacting) minerals may (Si(OH)4) precipitates. This reaction allows live
be significant. Thus, an additional acid and mineral HF to penetrate farther into the formation; however,
must be considered to accommodate the following there is an added risk of a possibly damaging precip-
reaction, which is added to the two-mineral model: itate forming.
Sumotarto (1995) presented an example that illus-
H2SiF6 + fast-reacting mineral →
trates the improved performance predicted with the
νSi(OH)4 + Al fluorides
two-acid, three-mineral model compared with the
The practical implications of the significance of one-acid, two-mineral model. Figure 16-6 compares
this reaction are that less HF is required to consume the mineral concentration profiles predicted by these
One-acid, two-mineral model Two-acid, three-mineral model
1.2 Injection volume = 25 gal/ft 1.2 Injection volume = 25 gal/ft
1.0 1.0
0.8 0.8
0.6 0.6
C/Co
C/Co
0.4 0.4
0.2 0.2
0 0 Mineral 1
Mineral 1 Mineral 2
–0.2 Mineral 2 –0.2 Mineral 3
0 2 4 6 8 10 12 0 2 4 6 8 10 12
Distance (in.) Distance (in.)
1.0 1.0
0.8 0.8
0.6 0.6
C/Co
C/Co
0.4 0.4
0.2 0.2
0 0 Mineral 1
Mineral 1 Mineral 2
0 2 4 6 8 10 12 0 2 4 6 8 10 12
1.0 1.0
0.8 0.8
0.6 0.6
C/Co
C/Co
0.4 0.4
0.2 0.2
0 0 Mineral 1
Mineral 1 Mineral 2
0 2 4 6 8 10 12 0 2 4 6 8 10 12
Figure 16-6. Dimensionless mineral concentrations at various injection volumes.

two models for the injection of 12% HCl–3% HF precipitation is predicted by the two-acid, three-min-
into a damaged formation composed initially of 17% eral model. Using a model of the permeability
clays and feldspars (fast reacting) and 83% quartz response to both mineral dissolution and precipita-
(slow reacting). In this figure, mineral 1 is clay and tion, the permeability and skin effect response are
feldspar, mineral 2 is quartz, and mineral 3 is silica predicted for each model and compared in Figs. 16-7
gel. After 100 gal/ft of injection, the feldspars and and 16-8. Although some precipitation is indicated
clays have been dissolved in a region extending by the two-acid, three-mineral model, improved per-
about 6 in. beyond the wellbore according to the formance because of the fluosilicic acid reactions is
two-acid, three-mineral model, whereas only 2 in. predicted compared with the one-acid, two-mineral
of dissolution is predicted by the one-acid, two-min- model.
eral model. In addition, a significant zone of silica
One-acid, two-mineral model 1000

1000 Model 1
25 gal/ft Model 2
100 gal/ft
225 gal/ft 800
800
Permeability (md)
Permeability (md)
600
600
400
400
200
200
0 0
0 2 4 6 8 10 12 0 100 200 300 400 500
Distance (in.) Injection volume (gal/ft)
6
Two-acid, three-mineral model Model 1
Model 2
1000 5
25 gal/ft
100 gal/ft
800 225 gal/ft 4
Skin effect
Permeability (md)
3
600
2
400
1
200
0
0 100 200 300 400 500
0
Injection volume (gal/ft)
0 2 4 6 8 10 12
Distance (in.)
Figure 16-8. Top: Average permeability versus injection
Figure 16-7. Permeability at various injection volumes. volume obtained from running the simulator with models 1
(one-acid, two-mineral model) and 2 (two-acid, three-mineral
model). Bottom: Skin effect calculated using models 1 and 2
as a function of injection volume (time).

16-3.4. Permeability response The Lambert expression is identical to that of Lund

and Fogler when M/∆φmax = 45.7.
To predict the response of a formation to acidizing, it is Using the values of the constants suggested, the
necessary to predict the change in permeability as acid Labrid correlation predicts the smallest permeability
dissolves some of the formation minerals and other increase, followed by the Lambert and then the Lund
minerals precipitate. The permeability change as a and Fogler correlations. The best approach in using
result of acidizing is an extremely complicated process these correlations is to select the empirical constants
because it is affected by several different, sometimes on the basis of coreflood responses, if available. If
competing phenomena in the porous media. The per- data are lacking for a particular formation, the Labrid
meability increases as the pores and pore throats are equation will yield the most conservative design.
enlarged by mineral dissolution. At the same time,
small particles are released as cementing material is
dissolved, and some of these particles lodge (perhaps
temporarily) in pore throats, reducing the permeability.
16-4. Carbonate acidizing
Any precipitates formed also tend to decrease the per-
16-4.1. Distinctive features
meability. The formation of carbon dioxide (CO2) as
carbonate minerals are dissolved may also cause a tem- In this chapter, sandstone acidizing is distinguished
porary reduction in the relative permeability to liquids. from carbonate acidizing although sedimentary rocks
The result of these competing effects is that the perme- exhibit a spectrum of compositions ranging from
ability in corefloods usually decreases initially; with almost pure calcite or dolomite to very clean sands.
continued acid injection, the permeability eventually The fundamental distinguishing feature is the HCl-
increases to values significantly higher than the original soluble fraction. If the HCl solubility of a rock is less
permeability. than 20%, a sandstone treatment using an HCl-HF
The complex nature of the permeability response mixture (for a discussion of such rules of thumb, see
has made its theoretical prediction for real sandstones McLeod, 1984) would most likely be applied. For-
impractical, though some success has been achieved mations composed largely of calcite or dolomite,
for more ideal systems such as sintered disks (Guin et including chalks and marls, are largely soluble in HCl
al., 1971). As a result, empirical correlations relating and are candidates for carbonate acidizing using HCl
the permeability increase to the porosity change dur- without HF.
ing acidizing are used. The most common correlations Carbonate acidizing with HCl is not complicated by
are those of Labrid (1975), Lund and Fogler (1976) a tendency for precipitates to form, as is the case for
and Lambert (1981). The Labrid correlation is sandstone acidization. As shown by the typical reac-
n
tions in Table 16-1, the reaction products CO2 and
k  φ CaCl2 are both quite water soluble (for a discussion of
= M  , (16-27)
ko  φo  their solubilities, see Shaughnessy and Kunze, 1981;
Schechter, 1992). Therefore, the formation of a precipi-
where ko and φo are the initial permeability and porosity
tate or a separate CO2-rich phase is generally not a
and k and φ are the permeability and porosity after
problem. Even if CaCl2 precipitates or a CO2 phase sep-
acidizing, respectively. M and n are empirical constants,
arates, these phases are readily dissolved when oil (or
reported to be 1 and 3, respectively, for Fontainebleau
gas) and water production is resumed. Despite the sim-
sandstone.
plified chemistry, HCl acidizing is a difficult process to
The Lund and Fogler correlation is
model. The origin of the difficulty is the rate at which
k   φ − φo   the reactions take place as compared with those of HF
= exp  M   , (16-28) with the various minerals prevalent in sandstones.
ko   ∆φ max  
Reaction rates are discussed in Section 16-2.2, and it
where M = 7.5 and the difference in maximum por- is instructive to compare some of them. HCl reactions
osity is ∆φmax = 0.08 from best-fit data for Phacoides with carbonates are orders of magnitude faster than HF
sandstone. reactions with sand (quartz), clays, etc.
The Lambert correlation is Because of the high reaction rate, HCl tends to etch
preferred pathways in carbonate rocks, apparently fol-
k
ko
[ ]
= exp 45.7(φ − φ o ) . (16-29) lowing local high-permeability streaks (Wang, 1993),

rather than progressing through the formation as a uni-

form front, as is the case in sandstone acidizing. These
pathways are soon enlarged by acid reaction at the
walls into sizable holes that have a diameter much
larger than that of naturally occurring pores. The
process continues until a few large holes become so
dominant that essentially all the injected acid flows
through these pathways, both enlarging and extending
them.
It is this tendency for macroscopic pathways to form
that makes HCl acidizing difficult to model. Because
the holes that form are large, they become the most
important feature of the process. To model carbonate
acidizing, the formation of these holes must be taken
into account. In fact, it is believed that the success of
acid stimulation of carbonate formations is due to the
formation of these preferred flow paths extending out-
ward from the wellbore or perforation. If the pathways
extend through the damaged zone, the produced fluids
can flow into the wellbore through these flow paths
with relatively little pressure drop because the holes
are much larger than the natural pores.
Thus, the fundamental physics of carbonate acidiz-
ing is embodied in three topics discussed in this sec-
tion. The first concern is the characterization of the
holes or flow paths created by the acid. Second, the
conditions under which they form must be defined.
And third, the rate at which they are extended is an Figure 16-9. Wormholes created by acid dissolution of lime-
issue of considerable practical significance. stone (Hoefner and Fogler, 1988; reproduced with permis-
sion of the American Institute of Chemical Engineers.
Copyright 1998 AIChE. All rights reserved.).
16-4.2. Wormholes
It is not known who first described the acid-etched Daccord and Lenormand (1987) considered the
pathway as a “wormhole,” but this appellation is com- characterization of a wormhole in terms of its fractal
monly accepted by those familiar with the complex dimension. A fractal is a self-similar geometric pattern.
etch pattern produced by acidizing carbonate cores in This implies that under increasing magnification the
the laboratory. Perhaps it was A. R. Hendrickson of same pattern will continue to reappear. Thus, accord-
Dowell. Figure 16-9 shows that the terminology is apt. ing to this notion the structure of a large wormhole is
This is a photograph of a metal casting of a wormhole repeated with branches from the main trunk that are
created by forcing molten metal into a wormhole, smaller replicas of the larger one. This replication is
allowing the metal to solidify and then dissolving the repeated as the magnification is increased until the
remaining rock with HCl. The casting illustrates the pores of the native rock come into view. These do not
complex morphology of the etch pattern. This is typi- resemble acid etch patterns because they were created
cal of many castings, which have been produced under by different processes. The discovery by Daccord and
a variety of experimental conditions. The chaotic Lenormand that wormholes are fractals is a significant
nature of the pattern seemingly discourages any contribution.
attempt to characterize its structure. However, it has One manifestation of the repeating, or self-similar,
been suggested that there is an underlying regularity character is that the perimeter or the length of a worm-
that may be useful for modeling. hole increases as the degree of magnification used in

its measurement increases. This is because at higher flux entering the matrix is quite small, wormhole initi-
magnification, more of the detailed structure becomes ation is not prolific, thereby indicating a flow-rate
evident and is, therefore, susceptible to measurement. dependence of the initiation process. The proof of this
For self-similar systems, the length of a wormhole assertion is based primarily on metal casts of worm-
plotted against the length of a ruler used to measure holes, such as that shown by Fig. 16-9. However,
the length is a straight line on a log-log plot. The slope adopting the notion of a flow-rate-dependent initiation
of this line is related to the fractal dimension. In the process allows interpreting the results of laboratory
case of wormholes, Daccord and Lenormand reported acidizing experiments and understanding the origin of
that the fractal dimension is 1.6. This implies that the the fractals. Furthermore, this approach leads to pre-
length of a wormhole is proportional to L1.6 rather than diction of the optimum injection rate in linear core
L, where L is the macroscopic length. Daccord et al. experiments that has been experimentally observed.
(1989) utilized this fractal dimension in developing a Thus, the analysis presented here represents a founda-
field design method for carbonate acidizing. In their tion upon which the design of acid treatments can be
approach, the complexity of the wormhole is portrayed based, but further work is required to achieve the
by its fractal dimension. This is only a partial charac- desired goal, namely, the ability to predict the stimula-
terization, because it is not possible to describe the tion resulting from an acid treatment given the essen-
minute details of wormhole geometry. tial parameters of acid composition, injection rate,
Although a precise description of a wormhole is not formation temperature and rock properties.
attainable, it is desirable to have a model that provides The initiation of wormholes occurs when live acid
guidance in determining the best treatment parameters. penetrates into pores present in the native rock. These
What should be the injection rate? Should the injection pores are distributed in size and shape; therefore, the
rate be constant during the entire course of the treat- amount of acid flowing through each of the pores dif-
ment? What acid type and concentration are best? fers. The rate at which a given pore is enlarged by the
What additives should be included in the acidic solu- acid depends, of course, on the amount of acid enter-
tion? These questions relate to controllable variables ing that pore and the fraction of the acid reacted at the
and, therefore, must be addressed each time an acid walls of the pore before the acid exits and then enters
treatment is designed. Rather than depending solely on other pores located downstream. Thus, even at the
past experience, some theoretical help is welcome for pore level, the processes that contribute to the creation
developing the best strategy possible based on the of an etch pattern are complex, involving convection,
information available. diffusion and chemical reactions within each of the
The problem is approached in the following sec- invaded pores. It has not been proved practical to con-
tions by addressing two separate issues, both of which sider these processes in a single pore and then attempt
are relevant to the questions posed. The first concerns to consider the collective behavior to derive a macro-
the conditions requisite for the initiation of worm- scopic etch pattern. Schechter and Gidley (1969) used
holes, and the second deals with their growth or prop- this approach, but to make progress using their results
agation. Both of these studies provide information requires knowing in advance the entire distribution of
required for the design of carbonate acid treatments. pore sizes, permeability and porosity of the native
The Appendix to this chapter discusses advances rock to be acidized. Even armed with this knowledge,
in understanding and predicting wormhole formation. which is seldom available, prediction of the etch pat-
tern is not routine.
The prediction of wormhole initiation is, however,
16-4.3. Initiation of wormholes based on a result that emerged from considering the
The fractal, or self-similar, topology of a wormhole behavior of each pore in the medium. If a pore is rep-
structure implies that the mechanism for the initiation resented as a cylindrical hole with a radius R and a
of wormholes is a “local” phenomenon that occurs length l, then the rate at which the pore cross-sectional
continuously along its bounding surfaces as well as at area A increases as a result of acid reaction at the pore
its tip. Thus, tiny wormholes may be initiated when- walls may, in general, be written in the form
ever live acid enters the pores of the virgin rock irre- dA
spective of the etch pattern already in existence. = ψA n , (16-30)
dt
Experiments have shown that in cases where the acid

where ψ is a function of the fluid velocity in the pore, Equation 16-32 looks formidable, but the goal is to
reaction rate and other parameters that determine the reduce this expression to a form that reveals the expo-
rate at which rock is dissolved. The exponent n is also nent n. To accomplish this goal, two extreme condi-
a function of the many parameters that contribute to tions are considered. First, examine the limit
ψ. The advantage of representing the rate of acid reac-
tion within a pore in this form is that the stability of 2 πE f C0m−1l
the enlargement process depends on the value of the << 1, (16-34)
Av
exponent n (see Schechter, 1992). If n > 1 for a certain
few pores in the native rock, then these pores become which implies that only a small fraction of the acid
larger faster than all the other pores that have growth reacts within the pore. In this limit, which applies for
characterized by exponents smaller than unity. This large v–, the rate of pore enlargement reduces to
criterion for uncontrolled growth determines whether dA
a wormhole is initiated. The growth of each individual → 2 π E f C0m−1 XA1/ 2 = ψ 1 A1/ 2 . (16-35)
dt
pore is characterized by the value of the exponent n.
Wormholes form whenever one or more of the pores Thus, in this limit the exponent n is 1⁄2 (n < 1) and
grow at a rate determined by n > 1. pores with a cross-sectional area such that the inequal-
Thus, this criterion for wormhole initiation hinges ity of Eq. 16-34 is satisfied do not form wormholes.
on the value on the exponent n. To investigate the fac- The pores will enlarge rather uniformly, and the acid
tors determining n, recall that the rate of acid reaction front will remain sharp, progressing through the porous
at a pore wall is given by the empirical expression matrix also rather uniformly. Not all pores, however,
satisfy the inequality. A second limit for some of the
reaction rate = − E f CWm . (16-31) pores may be possible. This second limit occurs when
the pores are of such a size that the inequality of
The acid concentration CW in this equation is the Eq. 16-34 is reversed, implying almost complete acid
concentration near the solid/liquid interface. The acid reaction within these pores. In this second limit, which
that reacts at the pore surface must be replenished by applies for small v–, can be found
acid diffusing from the bulk solution to the pore wall.
If this diffusion rate is slow, the rate of pore area dA vXC0 2
→ A = ψ 2 A2 . (16-36)
enlargement may be limited by diffusion. However, dt 8πl
the native pores are generally small enough so that
Thus, pore areas satisfying the second limit are
diffusion is relatively fast, and the average velocity
unstable (n > 1). They grow more rapidly than the
within a pore v– is sufficient to maintain a rate of mass
neighboring pores. They become small wormholes and
transfer to the surface so that the acid reaction at the
continue to evolve into the macroscopic etch patterns
wall is the controlling factor. The rate at which a pore
shown by the metal casts.
is enlarged is, therefore, (see Schechter, 1992)
For a given reaction rate and acid flux, there are
dA   2 π E f C0m−1l   pores that are essentially too small to become worm-
l = vAXC0 1 − exp −  , (16-32) holes and perhaps others that are of sufficient size to
dt   Av   exhibit uncontrolled growth and eventually become
where l is the length of a pore defined as the distance macroscopic. Thus, for a given acid flux, the native
the acid in a pore travels before mixing with acid pores may fall into two different categories: candidates
emerging from other pores. The acid concentration for incipient wormhole formation and noncandidates.
C0 is the concentration at the pore entrance. There is, therefore, a critical (or transitional) pore size
The average velocity in a pore v– depends on the AT that may be estimated as the value where the two
local Darcy flux u and the pore cross-sectional area. limiting growth rates become equal:
The flow within a single pore is laminar, so that ψ 1 AT1/ 2 = ψ 2 AT2 . (16-37)
dp A uA
v= = , (16-33) Solving this expression for AT yields
dt 8πµ 8πk
AT ≅ 20 Da 2 / 3 ( kl ) ,
2/3
(16-38)
where k is the permeability of the rock matrix.

where Da = EfC0m – 1/u and k is the formation perme- 16-4.4. Acidizing experiments
ability (Wang, 1993; Wang et al., 1993). Thus, if all
Essentially two different types of results are found
the pores in the native rock have cross-sectional areas
on the basis of acidizing carbonate cores in the labora-
less than AT, wormhole initiation cannot occur until at
tory. One is the metal cast of a wormhole and the
least one of the pores has been enlarged by acid reac-
other is the pressure drop measured while acidizing
tion to a size sufficient to allow wormhole develop-
at a constant injection rate. Both types of experiments
ment. The critical, or transitional, area depends on
have proved instructive. The transition area defined by
both the reaction rate and the acid flux. This condition
Eq. 16-39 depends on the acid flux and reaction rate,
has, as discussed subsequently in this chapter, consid-
which in turn is a function of the acid concentration,
erable practical relevance.
reaction temperature and rock composition. The acid
The average length of a pore is a rather nebulous
flux is the easiest to control and is the most widely
quantity that, on the basis of a number of laboratory
studied variable.
experiments using two different limestones and a
dolomite, appears to be about 0.1 mm. If we use this • Acid flux
value, the criterion for the critical pore dimension The flow rate is expected to influence the acid etch
becomes pattern for reasons that may be best understood by
AT = 0.93[ Da k ] , considering the idealized depiction of a wormhole
2/3
(16-39)
shown by Fig. 16-10. It is a cylinder with fluid loss
where both k and AT must be expressed as cm2.† about the perimeter as well as at the tip. Depending
This equation is quite simple in appearance, but its on the external pressure field surrounding the worm-
implications are profound and these may be tested hole, the fluid-loss flux may vary from point to point
experimentally. Laboratory results that are seemingly about the surface of the cylinder. If the flux into the
counterintuitive may be satisfactorily explained by rock is small at some points, AT as determined by
invoking the concept embodied by Eq. 16-39. Eq. 16-38 may exceed the cross-sectional area of all
Wormhole
Injection pressure, pinj Wormhole radius, rwh
Total injection rate, qT

Flow at the tip, qe
Fluid loss, qL
Injection rate at core face, qc
Length to tip, Le
Figure 16-10. Single-wormhole model.
†
Given a permeability in md, multiply by 9.869 × 10–12 to obtain
the dimensions in cm2.

the native pores, and wormholes will not form at that rate. The left-hand core at the lowest flow rate is an
point. The walls of the wormhole will then be eroded example of a nearly uniform dissolution front, where
in a generally uniform manner. If, however, the flux the inlet face of the core has essentially been dis-
at the tip is large enough to initiate wormholes, a solved. The flow rate was evidently less everywhere
network of small wormholes will continuously form than the critical one. As predicted, the wormholes
at the tip, rapidly extending its length. that developed at the higher flow rates show substan-
In summary, it is expected that for injection rates tial branching, displaying the fractal structure postu-
that are very slow, wormholes will not form and the lated by Daccord and Lenormand (1987). At higher
face of the core will dissolve rather uniformly. At rates, much of the acid is expended in the creation
modest injection rates, large enough to initiate of the highly ramified structure shown by the casts.
wormholes at the tip of the primary wormhole, an Of primary practical interest are the wormholes
etch pattern is expected to develop that shows little created at intermediate injection rates. They develop
branching from the primary wormhole. Most of the a minimum of side branches extending from the
acid is then expended in extending the primary perimeter of the main channel, in agreement with
wormhole. If the rate of acid injection is then the etch pattern anticipated by consideration of a
increased, the acid fluid-loss flux into the rock critical pore area. Indeed, the series of casts in
matrix may be large enough everywhere—or at Fig. 16-11 confirms the existence of a transitional
least at many points—to allow the initiation of area. Although the casts provide strong evidence
wormholes along the boundary of the primary one. supporting the hypothesis set forth in the preceding
A highly ramified wormhole structure is expected section, they also suggest a means for further verifi-
at the higher injection rates. cation. It seems evident in considering the series
Hoefner and Fogler (1988) prepared metal casts of of casts that a wormhole formed with a minimum
wormholes that developed in calcite cores at various of side branching will penetrate through the core
rates of injection (Fig. 16-11). The casts are arrayed using a smaller quantity of acid than would be
from left to right to correspond to increasing flow required otherwise.
Figure 16-11. Metal casts of wormholes that developed in calcite cores at various rates of injection are arrayed from left to
right to correspond to increasing flow rate (Hoefner and Fogler, 1988).

Thus, Eq. 16-39 appears to be the key to the deter-

6
Indiana limestone mination of the effect of acid injection rates in field
Glen Rose limestone
operations. Ideally, a few dominant wormholes
5
1-in. diameter × 6-in. long should be initiated extending from the wellbore in
Pore volumes to breakthrough
15% HCl (4.4N)

Room temperature openhole completions or from the perforations in
4
cased holes. This would initially entail a modest acid
injection rate (calculated by Eq. 16-39). As the acid
3
treatment progresses and the lengths of the few dom-
inant wormholes increase, higher injection rates are
2
then desirable to continue to extend the dominant
wormholes, if possible. Nierode and Williams
1 (1971) were apparently the first to emphasize the pri-
mary importance of fluid loss from wormholes as a
0 limiting factor in extending them. Their hypothesis
0 100 200 300 400 500
Injection rate (mL/hr)
is in no way weakened by the experiments and
analysis presented here. Acid fluid loss from a
Figure 16-12. Coreflood results for Indiana and Glen Rose wormhole remains a limiting factor in the propaga-
limestones. tion of wormholes.
• Reaction temperature
Wang et al. (1993) measured the volume of acid Increasing the reaction temperature increases Ef
required to achieve wormhole penetration through exponentially (see Eq. 16-9). In accordance with
a core (breakthrough). Figure 16-12 is a plot of the Eq. 16-39, this should result in a corresponding
acid pore volumes to breakthrough as a function of increase in AT. Therefore, the optimum injection rate
the injection rate. As anticipated, an optimum must correspondingly be increased (see Sidebar 6E).
injection rate exists. Wang et al. calculated the acid This predicted trend is shown by Fig. 16-13. The
fluxes about the wormhole that develop during the optimum injection rate at a temperature of 125°F
experiment to find the flux at the tip as well as is almost twice as large as the optimum at room
along the sides of the wormhole. They calculated temperature. According to Wang et al. (1993), the
for the optimum case (i.e., the experiment requiring increased rate is predicted by Eq. 16-39. An impli-
the least volume of acid) that the flux at the tip is cation of this result is that, if possible, deep wells
well predicted by Eq. 16-39. The critical size AT should be acidized at higher rates than shallow
was determined by the capillary entry pressure and ones; but in either case, increasing the rate during
the reaction rate was determined by Eq. 16-8. the course of the treatment is apt to be beneficial in
A subsequent study by Bazin et al. (1995) using extending the wormholes. Initially, however, the
two other limestones (Lavoux and Savonnieres) acid flux should be restricted by the value deter-
shows the importance of core length in laboratory mined with Eq. 16-39 if this rate is possible without
studies. As the length of the dominant wormhole fracturing the formation. As shown by Figs. 16-12
increases, the amount of acid lost through the lateral and 16-13, the volume of acid required to achieve
boundaries increases, thereby reducing the volume breakthrough does not increase rapidly for injection
of acid reaching the tip and ultimately resulting in an rates in excess of the optimum, so maintaining the
acid flux at the tip that is too small to initiate worm- optimum injection rate is not thought to be critical.
holes there. In such a case, the wormhole can extend Stimulations conducted at rates somewhat in excess
only slowly while it is also being enlarged. Worm- of those demanded by Eq. 16-39 may not differ
hole growth is slowed, as observed by Bazin et al. greatly from those achieved at optimum. On the
T. Huang (pers. comm., 1996) studied the rates of other hand, the amount of acid to achieve break-
wormhole propagation reported by Bazin et al. and through does increase substantially for rates less
found them predictable on the basis of Eq. 16-39 than the optimum, so maintaining a sufficient rate
when fluid loss is properly taken into account. if possible is recommended.

16E. Optimum injection rate for initiating 7

carbonate treatment Room temperature
125°F
6
Consider an acid treatment of a well in a carbonate formation Indiana limestone

known to be composed of calcite with little dolomite and to 1-in. diameter × 6-in. long
5 3.4% HCl (1N)
have a permeability of 5 md. The treatment will be conducted
using 15% HCl injected initially, if possible, at a rate nearly
corresponding to the optimum one. The reservoir tempera- 4
ture is 125°F [50°C]. The thickness of the formation to be
treated is 30 ft. The well is completed openhole with a well-
bore diameter of 6 in. 3
To determine the optimum injection rate, the cross-
sectional area of the largest native pores must be estimated. 2
Generally, this cross-sectional area is determined from the
capillary entry pressure measured by mercury porosimetry, if
an adequate core sample is available. For the present exam- 1
ple, use AT = 1.45 × 10–7 cm2. If a sample of the formation is
not available, then AT may be estimated as a multiple of the 0
formation permeability (Dullien, 1979). What is the optimum 0 100 200 300 400 500
acid injection rate?
Solution
Based on the data provided in Table 16-4, Figure 16-13. Indiana limestone coreflood results at differ-
ent temperatures.
 7.55 × 10 3 
E f = 7.291× 10 7 exp −  = 5.14 × 10 ,
−3
(16E-1)
 (50 + 273) 
ramified structure that is inefficient in developing an
and because the acid is 15%, C0 = 4.4 kg-mol/m3
(= 4.4 mol/L). Therefore,
etch pattern consisting of a few dominant wormholes.
It is, therefore, expected that corresponding to a sub-
E f C 0m −1 =
(5.14 × 10 )(4.4)
−3 0.63
= 2.97 × 10 −3 m/s. (16E-2) stantial decrease in the reaction rate, there must be
4.4 an associated decrease in the acid injection rate to
In an appropriate set of units, k = (5 md)9.869 × 10–12 remain at optimum. Figure 16-14 shows the acid
cm2/md = 4.9345 × 10–11 cm2. Therefore, the optimum flux pore volumes to breakthrough as a function of the
from Eq. 16-39 is u = 2.38 × 10–3 m/s = 0.0078 ft/s. This
corresponds to an initial acid injection rate of 3.9 bbl/min injection rate for dolomite cores. The optimum rate
(0.13 bbl/min/ft). If this injection rate exceeds the formation at room temperature is not readily discernible, but
parting pressure, injection would be at the highest possible
matrix rate during the entire treatment. it is evident that slow rates are preferred to higher
The optimum rate decreases with formation temperature. ones. Hoefner and Fogler (1988) also studied the
Consider the same treatment when the temperature is 85°F
[30°C] rather than 125°F. Then, Ef = 1.1 × 10–3. This leads
acidization of dolomite cores and found results simi-
to an optimum flux u = 5.09 × 10–4 m/s corresponding to an lar to those shown by Fig. 16-14. The results for
injection rate of 0.8 bbl/min. This rate is usually sustainable dolomite represent a striking confirmation of predic-
without fracturing the matrix.
Thus, most calcite formations are treated using rates near tions based on Eq. 16-39.
the maximum that the matrix will accept, except perhaps in The field implication is that acidizing in shallow
cool, shallow formations. Once the treatment has been initi-
ated, it may be beneficial to increase the rate to propagate dolomite formations should be conducted at low
the wormholes created by the initial acid contact. rates. High rates result in a multiple wormhole pat-
tern that does not penetrate far into the formation and
appears as uniform acid invasion dissolving the face
• Formation composition of the wellbore, which is inefficient for removal of
The reaction rate of HCl with dolomite is much skin effect damage compared with producing a few
slower than that with calcite (see Table 16-4). This dominant wormholes that penetrate into the forma-
being the case, Eq. 16-39 indicates that unless the tion. In deeper dolomite formations, the rate may be
acid flux is greatly reduced, many of the native pores increased to some extent because the reaction tem-
are likely to exceed AT in size and be candidates for perature increases with depth. The increased opti-
wormhole initiation. Thus, closely spaced multiple mum rate with increasing temperature shown in
wormholes are likely to form, producing a highly Fig. 16-14 confirms predictions based on Eq. 16-39.

greater the rate of acid diffusion toward the wall, the

80
Room temperature
lower the concentration of acid at the tip of the worm-
125°F hole and the slower its rate of propagation. Increased
70 170°F
fluid-loss rates serve to convect acid to the wall and
Dolomite
1-in. diameter × 5-in. long at the same time reduce the acid flux reaching the tip,
60 3.4% HCl (1N) thereby decreasing the rate of propagation. Taking into
account both fluid loss and diffusion, Hung (1987)
50
found that for a constant injection rate, the rate of
extension of a wormhole begins to decrease as the
40 wormhole length increases. The length appears to ulti-
mately reach a plateau, as shown by Fig. 16-15 (Hung,
30 1987) but never actually ceases to grow. Hung attrib-
uted the diminishing growth rate entirely to fluid loss.
20 Thus, it is anticipated that wormhole penetration will
0 50 100 150 200 250 300 350
essentially cease after a certain length has been
attained as long as the injection rate is fixed. In long-
Figure 16-14. Effect of temperature on the variation of vol- core experiments, an ultimate length was observed by
umes to breakthrough with the injection rate for dolomite. Bazin et al. (1995).
Hung calculated that the wormhole evolves in shape
depending on the local rate of acid reaction and fluid
loss and the rate of fluid injection. Once the wormhole
16-4.5. Propagation of wormholes length stabilizes, the acid that is injected serves primar-
Once wormholes are initiated in the rock surrounding ily to increase the diameter. Because Hung’s model
the face of the wellbore or perforation, it is desirable to does not account for the meandering nature of worm-
extend them into the formation as far as possible with holes caused by small-scale heterogeneities in the rock
a given volume of acid. The skin effect should be or the creation of side branches, it tends to overpredict
reduced within the regions penetrated by wormholes. wormhole length.
To promote understanding of the factors governing the Daccord et al. (1989) recognized the importance of
rate of extension of a wormhole, Hung et al. (1989) propagating the wormhole to the fullest extent possible
modeled wormhole growth by considering it to be a and proposed a model based on laboratory experiments
cylinder with fluid loss as depicted by Fig. 16-10.
Hung et al. took into account a number of factors,
including the contributions of both acid diffusion and 125
Injection rate
convection resulting from fluid loss to the walls of the 0.005 cc/s
0.003 cc/s
wormhole where the acid reacts. These are important 0.001 cc/s
100
factors because the acid reactions in a wormhole are,
in general, limited by mass transfer as contrasted to
those in natural pores, which are controlled by the 75
Length (cm)
reaction rate.
The rate of wormhole extension is determined by
50
the amount of the acid arriving at the tip:
dL uCβ uC
= e e 100 = e e Ac , (16-40)
dt (1 − φ)ρrock φC0
25
where the subscript e refers to conditions evaluated at

0
the end or tip of the wormhole, ρrock is the density of 0 80 160 240 320 400
the rock, and L is the length of the wormhole. This Time (s)
equation shows the importance of diffusion, acid con-
vection and fluid loss on wormhole propagation. The Figure 16-15. Predicted wormhole length (Hung, 1987).

that differs from that proposed by Hung. Daccord et does not indicate a plateau value as the wormhole
al.’s expression for the rate of wormhole propagation lengthens. Thus, the equation is applicable to short
in linear systems is wormholes where fluid loss is not a factor, but it
2/3 should not be used for the prediction of wormhole
dL aAc  q 
= , (16-41) penetration length.
dt Aφ  D  Thus, none of the existing models for the rate of
where a is a constant determined experimentally, D is wormhole propagation is strictly correct. Because
the molecular diffusion coefficient, and A is the cross- wormhole length is thought to be a crucial factor in
sectional area of the wormhole. Daccord et al.’s model determining stimulation, better models incorporating
considers the influence of acid diffusion but does not the important features of the ones that have been pro-
take into account fluid loss; therefore, this equation posed are required.

◆
Appendix: Advances in
Understanding and Predicting
Wormhole Formation
Christopher N. Fredd, Schlumberger Dowell
Introduction 16 discusses the various transport and reaction processes

that influence carbonate dissolution in a variety of fluid
The transport and reaction of reactive fluids in car- systems including strong acids, weak acids and che-
bonate porous media results in the formation of highly lating agents. A brief review of the theories of worm-
conductive flow channels, or wormholes. Wormholes hole formation is provided, and wormhole formation
significantly influence the flow of reservoir fluids is discussed in terms of the Damköhler number for flow
because their conductivity is several orders of magni- and reaction. The existence of an optimum Damköhler
tude larger than that of the porous medium. There- number for effective wormhole formation is demon-
fore, the success of carbonate stimulation treatments strated by laboratory experiments, and a new extension
is highly dependent on wormhole formation (i.e., of the theory is introduced to predict optimum injection
wormhole formation during matrix acidizing treat- strategies for matrix acidizing treatments in the field.
ments and lack thereof during fracture acidizing treat-
ments). Wormhole formation is desirable during
matrix acidizing treatments because the wormholes Carbonate dissolution
are capable of bypassing near-wellbore damage. In
contrast, wormhole formation increases fluid leakoff Influence of transport and reaction
during fracture acidizing and, consequently, limits the
depth of acid penetration. The structure of the worm- The dissolution of carbonate minerals involves a het-
hole channel varies significantly with the flow condi- erogeneous reaction that may be influenced by a
tions and the properties of the fluid-mineral system. variety of transport and reaction processes. These
Because the structure ultimately controls the effec- processes include the transport of reactants to the
tiveness of the stimulation treatment, a fundamental surface, the reversible surface reaction and the trans-
understanding of the dissolution phenomenon is port of products away from the surface. The rate-limit-
required to design effective treatments. ing process depends on the fluid-mineral system and
Wormhole formation occurs because carbonate dis- may vary significantly in the presence of strong acids,
solution is influenced by mass-transfer processes. weak acids and chelating agents.
Many investigators have recognized the importance of This section demonstrates the interplay between
mass transfer to the phenomenon of wormhole forma- the various transport and reaction processes and sum-
tion in the hydrochloric acid (HCl)-limestone system marizes results from kinetic studies for a variety of fluid-
(Barron et al., 1962; Williams et al., 1970; Nierode mineral systems. The results are based on experimental
and Williams, 1971; Hoefner and Fogler, 1988; studies conducted using a rotating disk apparatus, which
Daccord et al., 1993; Wang et al., 1993; Frick et al., allows the differentiation of the surface reaction kinetics
1994b; Mostofizadeh and Economides, 1994; Bazin from the transport processes because of the well-defined
et al., 1995; Huang et al., 1997; Fredd and Fogler, hydrodynamics. From an understanding of the processes
1998a). Recent studies demonstrate that in the presence that influence the dissolution, as well as appropriate
of weak acids and chelating agents wormhole for- expressions for the kinetics of the surface reaction and
mation is influenced by a variety of transport and reac- rate of mass transfer, the overall rate of carbonate disso-
tion processes (Fredd and Fogler, 1998b). Thus, to lution can be determined as discussed in “Generalized
provide the reader with a full understanding of the description of carbonate dissolution.”
dissolution phenomenon, this Appendix to Chapter

• Strong acids ion attack (see the two preceding chemical reac-
The rate of calcite dissolution has been shown to tions). However, dissolution is complicated by the
be mass-transfer limited in a variety of acidic reversible surface reaction (Chatelain et al., 1976)
media (pH less than about 4) ranging from HCl to and the additional influence of the mass transfer of
pseudoseawater (Williams et al., 1970; Lund et al., products away from the surface. The dissolution of
1975; Plummer et al., 1978; Sjöberg and Rickard, calcite by acetic acid was found to be influenced by
1984; de Rozières et al., 1994). In strong acids the rate of transport of reactants to the surface and
such as HCl, calcite is readily dissolved and the the rate of transport of products away from the sur-
reaction may be considered irreversible: face at ambient temperature (Fredd and Fogler,
1998d). The interplay between the two transport
2H+ + CaCO3 → Ca2+ + CO2 + H2O processes causes transport limitations that are much
Because the reaction is essentially irreversible, more significant than either limitation independ-
the products do not influence the rate of dissolu- ently. This interplay results in an apparent effective
tion, and the observed mass-transfer limitations are diffusion coefficient that is over an order of magni-
due to the transport of hydrogen ions to the calcite tude lower than that of acetic acid. Similar results
surface. The mass-transfer limitations are more sig- were observed during the dissolution of calcite by
nificant when HCl is emulsified or gelled because formic and maleic acids (Takulpakdee, 1998).
of a reduction in the effective diffusivity (Hoefner The kinetics of the surface reaction for the disso-
and Fogler, 1985; de Rozières et al., 1994). The lution of calcite by acetic acid was shown to be a
dissolution of calcite by HCl becomes influenced significant limitation at pH values greater than
by the kinetics of the surface reaction at tempera- about 3.7 (Fredd and Fogler, 1998d). The rate
tures below about 32°F [0°C] (Lund et al., 1975). of the surface reaction can be expressed as
The following rate expression was found to
describe the rate of the surface reaction:
(
rD = kr Ci − Cpi Keq , ) (2)
where Ci and Cpi are the reactants and products con-
rD = kr Cin , (1) centrations at the solid/liquid interface, respectively,
where Ci is the concentration at the solid/liquid and Keq = Kc/CCOi is the effective equilibrium con-
interface, kr is the surface reaction rate constant, stant. The reactants concentration is the total con-
and n is the reaction order. The kinetic parameters centration of associated and dissociated hydrogen
are summarized in Table 16-4 (Lund et al., 1975). ions (i.e., acetic acid and H+), whereas the products
The dissolution of dolomite by HCl is mass- concentration represents the total concentration of
transfer limited at temperatures greater than about calcium-containing species. The total interface con-
200°F [95°C] (Lund et al., 1973) and reaction rate centration of carbonate species CCOi was found to
limited at temperatures less than about 125°F be independent of the rate of mass transfer (Fredd,
[50°C]. Between these two temperatures, the disso- 1998) and was, therefore, lumped into the effective
lution is influenced by both the rate of mass trans- equilibrium constant. The effective equilibrium
fer and the kinetics of the surface reaction. Like for constant also includes the conditional equilibrium
calcite, the reaction can be considered irreversible, constant Kc, which is a complex function of the
so the influence of mass transfer is due to the trans- equilibrium constants for the various equilibrium
port of hydrogen ions to the dolomite surface. reactions occurring near the interface.
Kinetic parameters evaluated at 72°F [22°C]
4H+ + CaMg(CO3)2 → Ca2+ + Mg2+ using a rotating disk (Fredd and Fogler, 1998d) are
+ 2CO2 + 2H2O listed in Appendix Table 1. Additional data for var-
The kinetics of the surface reaction is also described ious acid concentrations and higher temperatures
by Appendix Eq. 1, using the kinetic parameters sum- (125°F) are in the literature (Fredd and Fogler,
marized in Table 16-4 (Lund et al., 1973). 1998d). These kinetic parameters are influenced by
a variety of adsorption processes and equilibrium
• Weak acids reactions. Therefore, extrapolation of the parame-
In weak acids (i.e., partially dissociating acids), ters to significantly higher temperatures requires
carbonate dissolution still occurs through hydrogen careful evaluation of the surface chemistry
A16-2 Chapter 16 Appendix: Advances in Understanding and Predicting Wormhole Formation

Appendix Table 1. Parameters used to calculate the overall dissolution rate constant at 72ºF.
De Dpe kr Keq
(cm2/s) (cm2/s) ∆E [kcal/mol])
(∆
0.25M CDTA (pH = 4.4) 4.5 × 10–6† 3 × 10–6‡ 2.3 × 10–4§ 1 × 1010††
0.25M DTPA (pH = 3.3) 4 × 10–6† 3 × 10–6‡ 7.8 × 10–5§ 1 × 1010††
0.25M DTPA (pH = 4.3) 4 × 10–6† 3 × 10–6‡ 4.8 × 10–5§ 1 × 1010††
0.25M DTPA (pH = 12.5) 4 × 10–6† 3 × 10–6‡ 2.6 × 10–5§ 1 × 1010††
0.25M EDTA (pH = 4) 6 × 10 –6‡‡
4 × 10 –6§§
1.4 × 10 –4§
1 × 1010††
(12)§
0.25M EDTA (pH = 8 to 13) 6 × 10–6§ 4 × 10–6§§ 5.3 × 10–5§ 1 × 1010††
0.5M acetic acid (pH = 2.5) 1.1 × 10–5††† 8 × 10–6‡‡‡ 5 × 10–3§§§ 1.6 × 10–1§§§
0.5M acetic acid (pH = 4.6) 1.1 × 10 –5†††
8 × 10 –6‡‡‡
2.1 × 10 –4§§§
2.7 × 10–1§§§
0.5M HCl 3.6 × 10–5†††† 2 × 10–5‡‡‡ 2 × 10–1†††† 1 × 1010††
(15)††††
Note: De = effective diffusion coefficient, Dpe = effective diffusion coefficient for reaction products, kr = effective surface reaction rate constant,
∆E = activation energy, Keq = effective equilibrium constant
†
Estimated from the Stokes-Einstein equation
‡
Assumed on the basis of De and value for Ca-EDTA
§
Fredd and Fogler (1998c)
††
Assumed on the basis of the irreversibility of surface reactions
‡‡
Assumed equal to the diffusion coefficient of EDTA at pH = 13
§§
Assumed on the basis of the value for Ru-EDTA (Jiang and Anson, 1992)
†††
Vitagliano and Lyons (1956)
‡‡‡
Estimated from ionic diffusion coefficients (Cussler, 1984)
§§§
Fredd and Fogler (1998d)
††††
Lund et al. (1975)
involved in the dissolution. The method of approxi- the rate of transport of reactants to the surface and
mating the surface reaction rate reported by the kinetics of the surface reactions at ambient tem-
Schechter (1992) (see Eq. 16-10) provides an perature (Fredd and Fogler, 1998c). Although the
order-of-magnitude estimate of the reaction rate reactions are essentially irreversible because of the
constant but does not account for the significant formation of a stable calcium complex, the reac-
contribution of the reverse reaction. tions are influenced by the transport of products
The rate of dolomite dissolution by weak acids is away from the surface. This influence is due to the
currently not available in the literature. Because the blocking of surface sites involved in the dissolu-
dissolution occurs through hydrogen ion attack, the tion. The dissolution mechanism is different from
rate of the surface reaction can be estimated from conventional acids in that hydrogen ions are not
Appendix Eq. 2 by scaling the surface reaction rate required. However, the rate of dissolution is
constant by the ratio of the surface reaction rate enhanced at low pH as a result of the combined
constants of dolomite to calcite for HCl. This scal- influence of hydrogen ion attack and chelation.
ing should provide a reasonable order-of-magnitude The rate of calcite dissolution varies consider-
estimate from which the relative influences of ably with pH and the type of chelating agent
transport and reaction can be determined. because of changes in the ionic form of the chelat-
ing agent and the influence of hydrogen ion attack.
• Chelating agents
In general, the rate of calcite dissolution increases
In the presence of calcium chelating agents such as as the number of hydrogen ions associated with the
ethylenediaminetetraacetic acid (EDTA), the rate of chelating agent increases. As the pH is increased
calcite dissolution is influenced predominantly by from about 4.5 to 8.5 to 13, EDTA successively

deprotonates from H2Y2– to HY3– to Y4–. The corre- q

sponding overall surface reactions are
H2Y2– + CaCO3 → CaY2– + H2O + CO2 C Bulk fluid Cp
HY3– + CaCO3 → CaY2– + HCO3–

Y4– + CaCO3 → CaY2– + CO32– K1 K3
The rate of calcite dissolution by chelating agents

can be described by Appendix Eq. 2. The appropri-
ate kinetic parameters are listed in Appendix Table 1
for EDTA, 1,2-cyclo-hexanediaminetetraacetic acid
(CDTA) and diethylenetriaminepentaacetic acid Ci Cpi
(DTPA) (Fredd and Fogler, 1998c). Calcite dissolu- kr
tion requires the formation of stable calcium chelates. reactants + CaCO3 products
For example, nitrilotriacetic acid (NTA), which
forms a relatively weak calcium chelate, is ineffec-
tive for calcite dissolution (Fredd, 1998). Appendix Figure 1. Carbonate dissolution involves reac-
tants transport, reversible surface reaction and products
The dissolution of dolomite by chelating agents has transport.
not been thoroughly investigated. Preliminary experi-
ments with EDTA at ambient temperature reveal no
significant dolomite dissolution. The dissolution weak acids and chelating agents. It is a reasonable
mechanism is probably inhibited by the low stability approximation for strong acids when the concentration
of the magnesium chelate at that temperature. dependence is lumped into the reaction rate constant
(i.e., kr = kr′Cin – 1). This approximation does not sig-
nificantly affect the interplay between transport and
Generalized description of carbonate reaction for systems that are mass-transfer limited.
dissolution The expressions in Appendix Eq. 3 were solved
A generalized description of carbonate dissolution is simultaneously for the interface concentrations, which
required to account for the various transport and reac- were then substituted back into the equation. Making
tion processes that may influence the rate of dissolu- an additional substitution for Cp based on the stoi-
tion. This description is provided by modeling the chiometry of the reaction, the rate of reactant con-
overall carbonate dissolution mechanism as three sumption can then be expressed as
sequential processes (Fredd and Fogler, 1998b) of the
 Co 
mass transfer of reactants to the surface, reversible rA = κ C − , (4)
surface reactions and mass transfer of products away  1 + νKeq 
from the surface (Appendix Fig. 1). At steady state,
where κ is the overall dissolution rate constant and Co
the rates of the three sequential processes are equal
is the initial reactant concentration. The overall disso-
and the rate of reactant consumption rA is
lution rate constant depends on the sum of resistances
rA = νrD (3) in series:
= K1 (C − Ci ) 1
1+
(
= νkr Ci − Cpi Keq ) κ=
νKeq
. (5)
= νK (C
3 pi )
− Cp , 1
+
1
+
1
K1 νkr νKeq K3
where ν is the stoichiometric ratio of reactants con-
sumed to products produced and K1 and K3 are the Appendix Eqs. 4 and 5 have been used to describe
mass-transfer coefficients for the reactants and prod- the rate of dissolution observed during rotating disk
ucts, respectively. Appendix Eq. 3 assumes that the experiments (Fredd and Fogler, 1998c, 1998d) and
surface reaction can be expressed as a first-order during flow and reaction within carbonate porous
heterogeneous reaction. This expression is valid for media (Fredd and Fogler, 1998b). These equations can

be used to determine the rate of carbonate dissolution properties. The typical dependence on the injection
in any flow geometry, provided that an appropriate rate is demonstrated in Appendix Fig. 2 by neutron
expression for the rate of mass transfer is available. radiographs of dissolution structures formed during
the dissolution of limestone by 0.5M [1.7%] HCl at
72°F. At low injection rates (far left structure), the
Wormhole formation reactant is consumed on the inlet flow face of the
core, resulting in complete dissolution of the core
The transport and reaction of reactive fluids in carbon- starting from the inlet flow face. This face dissolution
ate porous media results in the formation of highly (also referred to as compact dissolution) consumes
conductive flow channels, or wormholes. Wormholes large volumes of reactant and provides negligible
form because of the natural heterogeneity of the increases in permeability. At slightly higher injection
porous medium and the rapid and almost complete rates, the reactant can penetrate into the porous matrix
dissolution of the mineral in the reactant fluid. During and enlarge flow channels. However, a significant
stimulation, the fluid preferentially flows to the amount of reactant is consumed on the walls of the
regions of highest permeability (i.e., the largest pores, flow channels. This consumption results in the forma-
vugs or natural fractures). The initial flow paths are tion of a conical-shaped dissolution channel and
enlarged by rapid dissolution of the matrix material, requires the injection of several pore volumes of fluid
causing these regions to receive even more of the for the channel to break through the porous medium.
flow. A dominant channel quickly forms and contin- At intermediate injection rates, unconsumed reactant
ues to propagate while diverting flow from other reaches the tip of the evolving flow channels. Sub-
regions. Once formed, the wormhole channels provide sequent consumption at the tip propagates the dissolu-
negligible resistance to flow and carry essentially all tion channels and eventually leads to the formation of
the injected fluid. a dominant wormhole. The wormhole provides signifi-
cant permeability increases and requires a minimum
pore volume of fluid to break through the rock matrix.
Dissolution structures
At high injection rates, the flow channels become more
The structure of the dissolution channel is highly highly branched or ramified (far right structure) as
dependent on the injection rate and fluid-mineral fluid is forced into smaller pores. Dissolution occurs
Appendix Figure 2. Neutron radiographs of dissolution structures formed during the dissolution of limestone by 0.5M HCl at
72°F (Fredd and Fogler, 1998a). PVinj = number of pore volumes injected, PVBT = number of pore volumes to breakthrough.

over a high surface area, which results in an increase The existence of an optimum injection rate is
in the number of pore volumes to breakthrough. At demonstrated in Appendix Fig. 3, which shows the
extremely high injection rates, all the pores are dis- dependence of the number of pore volumes to break-
solved uniformly as unconsumed reactant penetrates through on the injection rate for a variety of fluid-
deep into the porous medium. This uniform dissolution limestone systems (Fredd and Fogler, 1998b). The
is typically observed when dolomite is dissolved by figure includes data from linear coreflood experiments
HCl at ambient temperature (Hoefner and Fogler, 1988). with 0.25M CDTA (pH = 4.4), 0.25M DTPA (pH =
The flow rate at which the various dissolution struc- 4.3), 0.25M EDTA (pH = 4 and 13), 0.5M acetic acid
tures are formed is influenced by mass-transfer and and 0.5M HCl. All the fluids exhibit an optimum
reaction processes. Therefore, the structures depend on injection rate at which the number of pore volumes to
the temperature and fluid-mineral system. For example, breakthrough is minimized and dominant wormhole
the same trend of decreasing channel branching channels are formed. The number of pore volumes to
observed when the injection rate is decreased is also breakthrough increases to the left and right of the min-
observed when the temperature is increased (Fredd, imum owing to the formation of conical dissolution
1998). Increasing the temperature results in an increase channels and ramified wormholes, respectively.
in the mass-transfer limitations (because the activation The importance of mass transfer on the dissolution
energy for diffusion is less than the activation energy phenomenon is demonstrated by the influence of the
for reaction) and causes more reactant to be consumed diffusion coefficient on the optimum injection rate. As
on the walls of the dissolution channel. The result is the fluid type was varied, the optimum injection rate
an increase in the number of pore volumes to break- decreased with decreasing diffusion coefficient in the
through and an increase in the injection rate required order of HCl > acetic acid > chelating agents. (The
to form the dominant wormhole channels (Wang et al., diffusion coefficients are listed in Appendix Table 1.)
1993; Fredd, 1998). Thus, although the dissolution Because of this influence of diffusion, alternative fluid
must be influenced by mass transfer (i.e., the surface systems such as chelating agents and weak acids are
reaction rate must be rapid) for wormhole formation to more effective than HCl when injected at rates below
occur, ineffective wormhole formation will occur if the about 0.2 cm3/min in linear coreflood experiments.
system is too mass-transfer limited. Hence, there is an Therefore, in shallow or tight formations where low
optimum degree of transport and reaction limitations injection rates are required to prevent fracturing the
for effective wormhole formation (Fredd, 1998). formation or when injection rates are limited because
of frictional pressures, alternative fluids may be more
effective than HCl for matrix stimulation.
Optimum injection rate
+ 0.25M DTPA, pH = 4.3 0.25M CDTA, pH = 4.4
Several investigators have studied the phenomenon of
+
0.25M EDTA, pH = 13 0.5M acetic acid

wormhole formation in a variety of fluid-mineral sys- 0.25M EDTA, pH = 4 0.5M HCl
tems and reported the existence of an optimum injec-

tion rate (Hoefner and Fogler, 1988; Daccord et al., 100
1989, 1993; Wang et al., 1993; Frick et al., 1994b;
Mostofizadeh and Economides, 1994; Bazin et al.,

+
1995; Huang et al., 1997; Fredd and Fogler, 1998a, + +
1998b). The optimum injection rate represents the 10 + + ++
+
+
conditions at which a minimum volume of fluid is

+
required to obtain a given depth of wormhole penetra-

+
+
+
tion. The optimum injection rate corresponds to the

formation of dominant wormhole channels. Because 1 Length = 4 in.
no significant difference in the conductivity of the var- Diameter = 1.5 in.
ious dissolution structures is observed for the same 10–3 10–2 10–1 100 101 102
depth of penetration (C. N. Fredd, unpubl. data, 1998), Injection rate (cm3/min)
the injection rate at which the volume of fluid is mini-
mized represents the most effective condition for Appendix Figure 3. Optimum injection rates for the dissolu-
tion of limestone by various stimulation fluids at 72°F (Fredd
matrix stimulation. and Fogler, 1998b).

Fundamentals of wormhole where d is the pore diameter. The units of a vary from
Appendix Eq. 6 to 7 such that Da is dimensionless.
formation Hoefner and Fogler observed that a minimum volume
of fluid was required for channel breakthrough (i.e.,
Influence of transport and reaction optimum conditions for wormhole formation) when
The obvious importance of wormhole formation Da was varied over several orders of magnitude. This
on carbonate acidizing has led many investigators to observation is consistent with the existence of an opti-
study the dissolution phenomenon. Early investigators mum injection rate for constant fluid-mineral proper-
recognized the significant influence of mass transfer ties because the value of Da is inversely proportional
on the dissolution of limestone by HCl (Barron et al., to the injection rate.
1962; Williams et al., 1970; Nierode and Williams, Recently, a similar dependence on Da was demon-
1971). This influence has served as a basis for many strated for the flow and reaction of chelating agents and
of the theories describing wormhole formation. weak acids in carbonate porous media (Fredd and
Daccord et al. (1989) investigated a water–plaster Fogler, 1998a, 1998b). These alternative fluid systems
of paris system and reported wormhole formation to are influenced by both transport and reaction processes
depend on the Peclet number Pe. Pe is defined as the (as described previously in “Carbonate dissolution”)
ratio of transport by convection to transport by diffu- and therefore cannot be described by theories devel-
sion (see Chapter 17). A similar dependence on Pe oped in previous studies. To describe wormhole forma-
was observed for the HCl-limestone system (Daccord tion with these fluid systems, the various transport and
et al., 1993; Mostofizadeh and Economides, 1994; reaction processes were included in a generalized
Buijse, 1997). Daccord et al. (1993) and Frick et al. description of the dissolution phenomenon. A common
(1994b) combined the concepts of fractal geometry dependence on Da was observed when the combined
with the dependence on Pe to describe wormhole for- effects of transport and reaction were taken into
mation in the HCl-limestone system. Bazin et al. account. This common dependence on Da is described
(1995) studied the HCl-limestone system and reported in detail in the sections that follow.
efficient wormhole formation to occur at the transition
between convection-limited and mass-transfer-limited
regimes. In contrast, Wang et al. (1993) and Huang et Dependence on the Damköhler number
al. (1997) investigated HCl-carbonate systems and To simulate transport and reaction within a wormhole
proposed that the optimum injection rate occurred at a channel, the dissolution of a porous medium was
transition between reaction-rate-limited and fluid-loss- modeled as the dissolution of a representative cylin-
limited regimes (see Chapter 16). Despite the major drical tube (Fredd and Fogler, 1998b). The cylindrical
influence of mass transfer on wormhole formation, tube represents the dominant flow channels within the
diffusion plays only a minor role in their theory. porous medium (i.e., the wormholes). Convection and
Hoefner and Fogler (1988) investigated HCl- reaction in the tube were included in a reactant mass
carbonate systems and found that the phenomenon balance, where Appendix Eq. 4 was used for the rate
of wormhole formation is governed by the Damköhler of reactant consumption. Solution of the reactant
number Da for flow and reaction. Da is defined as the mass-balance equation resulted in an expression for
ratio of the overall rate of dissolution to the rate of the concentration profile along the length of the chan-
transport by convection. When the overall rate of dis- nel that is dependent on Da:
solution is mass-transfer limited,
C 1 + νKeq e
− Da
Damt = aD l q ,
2/3
(6) = , (8)
e Co 1 + νKeq
where De is the effective diffusion coefficient, q is the
flow rate, l is the pore length, and a is a constant that where Da is defined as
depends on the carbonate core. On the other hand, πdlκ
when the net rate of dissolution is reaction rate limited, Da = , (9)
q
Darxn = akr dl q , (7)
where d and l are the diameter and length of the capil-
lary tube, respectively. Because of the dependence on

the overall dissolution rate constant, this Da includes dence of wormhole formation on Da is observed when
the effects of reactants transport, reversible surface the various transport and reaction processes are taken
reactions and products transport. This Da reduces to into account. This dependence is shown in Appendix
those defined in Appendix Eqs. 6 and 7 for reactants- Fig. 4, where the number of pore volumes to break-
transport-limited (κ = K1) and reaction-rate-limited through is plotted as a function of the inverse of Da
(κ = kr) dissolution, respectively. Physically, Da pro- (as defined in Appendix Eq. 9). The curves for the
vides a measure of the amount of reactant being con- chelating agents and acetic acid are shifted to the right
sumed on the walls of the wormhole, as opposed to in comparison with Appendix Fig. 3 such that the mini-
being transported to the tip of the wormhole, where it mum number of pore volumes to breakthrough occurs
can be consumed efficiently. Unlike previous parame- at about the same value of Da for all the fluids. Thus,
ters that assume either mass-transfer-limited or reac- these fluids are all characterized by the same optimum
tion-rate-limited dissolution, this Da accounts for a Da. This optimum is observed for fluid-limestone sys-
variety of transport and reaction processes and is, tems that range from reactants transport limited (HCl)
therefore, able to describe alternative fluid systems to reactants transport and surface reaction influenced
such as chelating agents and weak acids. (chelating agents) to reactants and products transport
To determine the overall dissolution rate constant, limited (acetic acid). Also included in Appendix Fig. 4
the mass-transfer coefficients were obtained from are data from Daccord et al. (1989) for the dissolution
Levich’s (1962) solution of the convective diffusion of plaster of paris by water. This system, which is
equation for laminar flow in a cylindrical tube. The limited by the transport of products away from the
average mass-transfer coefficient Kmt along the length surface (Christofferson and Christofferson, 1976), also
of a tube is exhibits the same optimum Da. For this wide range of
1/ 3 fluid-mineral systems, the optimum Da occurs at a
K mt = 1.86 De2 / 3   ,
u
(10) value of approximately 0.29.
 dl 
The dependence of the wormhole structure on Da
where Kmt is for either reactants or products (K1 or K3), is shown in Appendix Fig. 5 for a variety of fluid-
depending on the value of the diffusion coefficient, limestone systems. The pairs of neutron radiographs
and u is the superficial velocity in the capillary tube. are at similar values of Da and represent a wide range
Because the diameter and length of the capillary tube, of transport and reaction limitations. The neutron radio-
or wormhole channel, change with time, the mass-
transfer coefficients and Da were evaluated on the + 0.25M DTPA, pH = 4.3 0.25M CDTA, pH = 4.4
+
basis of the final wormhole dimensions. The diameter 0.25M EDTA, pH = 13 0.5M acetic acid
Water-plaster (Daccord et al., 1989) 0.5M HCl
was measured from neutron radiographs. Typical 0.25M EDTA, pH = 4
diameters were of the order of 0.02 in. The length was
assigned a representative length of the wormhole
100
(one-half the core length). The effective surface reac-
tion rate constant and effective equilibrium constant

+
were obtained from independent kinetic studies using +

a rotating disk. The effective constants are listed in + +
10
++ +
Appendix Table 1 with values for the diffusion coeffi- + + +
+
cients for reactants and products. For the case of HCl

+
+
and the chelating agents, Keq is relatively high, so the

+
+
products transport and reverse reaction terms become

1
negligible.
100 101 102
Optimum Damköhler number 1/Da
Da has been shown to describe the phenomenon of Appendix Figure 4. The optimum Da for the dissolution
wormhole formation for a wide range of fluid-mineral of limestone by various stimulation fluids (Fredd and Fogler,
systems (Fredd and Fogler, 1998b). A common depen- 1998b).

Appendix Figure 5. Comparison of wormhole structures formed at similar Da values by a variety of fluid systems (Fredd,
1998).
graphs show wormhole structures formed at Da

values of approximately 1.1, 0.11 and 0.014. The 0.5M HCl 0.5M acetic acid
0.5M HCl, 50°C 0.5M acetic acid, pH = 4.6
wormhole structures exhibit similarities in the amount 0.5M HCl, 80°C Water-plaster
and type of channel branching at each of these values. 0.25M EDTA, pH = 4 (Daccord et al., 1989)
0.25M EDTA, pH = 4, 50°C 0.25M CDTA, pH = 4.4
+
As Da is decreased, the dissolution structures change 0.25M EDTA, pH = 4, 80°C + 0.25M DTPA, pH = 3.3
from conical-shaped channels to dominant wormholes 0.25M EDTA, pH = 8.8 + 0.25M DTPA, pH = 4.3
0.25M EDTA, pH = 13 0.25M DTPA, pH = 12.5
to highly ramified wormholes. These similarities
demonstrate that Da dictates the structure of the disso-
lution channels formed by systems with a wide range 100
of mass-transfer and reaction limitations.

A variety of fluid systems exhibit similar trends in
the number of pore volumes to breakthrough and the
+
corresponding wormhole structures. Because these 10

+
trends are a result of a common dependence on Da
and, therefore, the transport and reaction processes, +
+
a single description of the dissolution phenomenon +

+ +++
++
+
+
++
++
++
+
+
1 +
+
is possible. This single description is obtained by nor- +
malizing the number of pore volumes to breakthrough

by the minimum number of pore volumes to break- 100 101 102
1/Da
through for the respective fluids (Fredd, 1998). The
normalized number of pore volumes to breakthrough
Appendix Figure 6. Normalized number of pore volumes to
is plotted versus the inverse of Da in Appendix Fig. 6. breakthrough versus Da (Fredd, 1998).
The results reveal a single curve for a wide range of
fluid-mineral systems, including different fluid types,
pH values and temperatures. (Unless otherwise stated,

the data are for the dissolution of limestone at ambient The combined effects of transport and reaction have
temperature.) This single curve provides a means of been included in a three-dimensional physically repre-
estimating the entire curve for the number of pore vol- sentative network (PRN) model for flow and reaction
umes to breakthrough and the wormhole structures in porous media (Fredd and Fogler, 1998b). The PRN
from a single coreflood experiment near the optimum model is based on a packed-bed description of the
Da. Thus, the need for exhaustive coreflood studies to porous medium (Bryant et al., 1993; Thompson and
determine the effectiveness of a fluid under particular Fogler, 1997). Although the use of a packed-bed
reservoir conditions is eliminated. description of the medium limits the ability of the
model to represent consolidated carbonate porous
media, it is a major advancement over the capillary
Wormhole formation modeling tube representations typically used. The model pro-
In addition to involving a variety of transport and vides a complete topological description of the medi-
reaction processes, the dissolution of porous media is um and a fundamental description of the pore-scale
complicated by the stochastic nature in which the flow hydrodynamics. These descriptions serve as a physical
channels evolve. The difficulties associated with basis for determining macroscopic parameters, such as
accounting for all the complexities of wormhole for- the permeability. Dissolution is simulated by allowing
mation have led to many simplifications of the disso- the spherical particles that compose the bed to shrink
lution phenomenon. These simplifications typically as reactant is consumed. The pores naturally merge as
limit the models to either mass-transfer-limited or the spherical particles are dissolved. The simulations
reaction-rate-limited dissolution and include an ideal are in qualitative agreement with experimental obser-
representation of the porous medium. Early investiga- vations and demonstrate the common dependence of
tors simulated mass-transfer-limited dissolution using wormhole formation on Da. More importantly, the
models of the porous medium that were based on bun- existence of an optimum Da is substantiated by PRN
dles of capillary tubes (Nougaro and Labbé, 1955; model simulations.
Rowan, 1959). These simple models were extended
by Schechter and Gidley (1969) to include the effects
of pore merging and either mass-transfer-limited or Prediction of optimum field
reaction-rate-limited dissolution. Although the domi- conditions
nant pores were observed to grow more rapidly, these
models were unable to capture the branching charac- The ultimate goal of all laboratory studies is to aid the
teristics of the dissolution phenomenon. design of effective field treatments. Two main design
More recently, Hoefner and Fogler (1988) success- requirements are predicting the optimum injection rate
fully modeled the phenomenon of wormhole formation and selecting the most appropriate stimulation fluid.
using a network model. They simulated pore-scale flow Unfortunately, the direct application of laboratory data
and reaction by representing the porous medium as a to the field is not straightforward, as noted by several
two-dimensional network of nodes connected by cylin- investigators (Daccord et al., 1989; Frick et al., 1994b;
drical bonds. The model provided qualitative predic- Bazin, 1995; Buijse, 1997). Wormhole formation in the
tions of the dissolution structure (see Fig. 17-1) and field is complicated by the effects of fluid loss through
corresponding permeability response for both mass- the walls of the wormhole in the radial geometry and
transfer-limited and reaction-rate-limited dissolution. by competition among wormholes for the injected fluid
Similar results were observed by Daccord et al. (1989), (Appendix Fig. 7). The importance of these processes
who utilized a network model to simulate the mass- has been demonstrated theoretically (Nierode and
transfer-limited dissolution of plaster of paris by water. Williams, 1971; Hung et al., 1989; Buijse, 1997).
Although these network models do not include pore However, these studies were limited in their ability
merging, they were able to predict the experimentally to predict optimum injection conditions in the field
observed trends in wormhole formation because they because of a lack of consideration of the critical para-
capture the effects of pore-level heterogeneity on the meters required for effective wormhole formation.
macroscopic dissolution phenomenon. To account for the influence of fluid loss and worm-
hole competition on wormhole formation, two
approaches are the most obvious:

is through that single wormhole because the conduc-

Convection to tip
of wormhole tivity of the wormhole is several orders of magnitude
higher than that of the porous medium and fluid loss
through the walls of the wormhole is insignificant in
Fluid loss through the linear geometry. Therefore, these experiments
walls of wormhole
demonstrate the optimum Da within a single worm-
hole with negligible fluid loss. Applying the concept
of the optimum Da to the field requires knowledge
of both the relative amount of reactant lost from the
wormhole because of fluid loss and the number of
wormholes that will form per unit surface area of the
formation (i.e., wormhole density). To account for
the effects of fluid loss and wormhole competition on
wormhole formation, it is first necessary to understand
the conditions at which the dependence on Da is
affected most significantly by these processes.
Competition for • Fluid loss
injected fluid
Fluid loss through the walls of a wormhole has
Appendix Figure 7. Idealized representation of fluid loss three main effects on transport and reaction within
from the walls of wormholes and competition among worm- the wormhole. These effects are decreasing the
holes for injected fluid. flow rate along the length of the wormhole, affect-
ing the rates of mass transfer within the wormhole
(because of the combined influence of convection
• simulate wormhole formation using large network and diffusion normal to the walls of the wormhole)
models that account for the effects of pore-scale and increasing the amount of reactant lost through
transport and reaction the walls of the main wormhole channel (because
• account for the added complications in macroscopic of reactant leakoff). Each of these effects reduces
theories such as the existence of an optimum Da. the amount of reactant that is transported to the tip
Network models inherently include the effects of of the wormhole (because of either increased con-
fluid loss and wormhole competition because the pres- sumption or leakoff) and, consequently, reduces the
sure profile and flow rates through all pores in the net- rate of wormhole propagation.
work are continually updated during the simulation. The significance of these effects depends on the
Although the use of network models would eliminate rate of fluid loss and, therefore, depends on the per-
the need for macroscopic correlations, the ability to meability of the medium that is being stimulated.
simulate wormhole formation at the field scale is lim- Fluid-loss velocities reported in the literature range
ited by the excessive memory and computational time from about 1 × 10–4 to 1 × 10–2 cm/s at a perme-
required for such simulations. Thus, the most tractable ability of about 1 md (Hung et al., 1989; Wang
approach at this time is to extend macroscopic theo- et al., 1993; Settari, 1993). This range provides
ries. This section introduces an extension of the opti- a conservative estimate of fluid-loss velocities
mum Da to predict optimum injection strategies for expected in typical matrix stimulation treatments,
matrix stimulation treatments in the field. which are usually conducted in damaged forma-
tions with near-wellbore permeabilities of less than
1 md. This range of fluid-loss velocities serves as
Significance of the Damköhler number the basis for the discussion that follows. Although
there is a complex interplay between the three
The existence of an optimum Da was observed in lab- fluid-loss effects, they are discussed independently.
oratory experiments conducted in linear cores. In
– Flow rate
these experiments, only one wormhole was typically
observed because of the limited cross-sectional area Fluid loss results in a decrease in the flow rate
available for flow. In addition, essentially all the flow along the length of the wormhole and, hence,

leads to an increase in the value of Da. As Da decreases. For high vL values at 200°F, the mass-
increases, an increasing amount of reactant is transfer coefficients for HCl and EDTA are
consumed on the walls of the wormhole. Thus, increased by factors of about 2 and 5, respec-
fluid loss reduces the amount of reactant being tively. The mass-transfer coefficients for the
transported to the tip of the wormhole and leads products are decreased more significantly.
to ineffective wormhole propagation. Ineffective However, the decrease is more than offset by
propagation can be overcome by changing the products leaking off into the formation and,
injection conditions (i.e., increasing the flow rate hence, reducing their concentration near the
or decreasing the overall rate of dissolution) to interface. As a result of these corrections, Da is
maintain the optimum Da. Increasing the injec- increased by a factor of only about 2 if the flow
tion rate to maintain efficient wormhole forma- rate is increased to offset the effects of fluid loss.
tion is consistent with investigators reporting a – Reactant leakoff
higher optimum injection rate in radial experi-
Fluid loss decreases the amount of reactants
ments than in linear experiments (Frick et al.,
transported to the tip of the main wormhole
1994b; Mostofizadeh and Economides, 1994).
channel because of reactants leaking off into the
– Mass transfer formation. Although reactant leakoff leads to
Fluid loss affects the rates of mass transfer (and, dissolution of the rock matrix surrounding the
consequently, the overall rate of dissolution) wormhole, the dissolution is ineffective because
within the wormhole as a result of the combined it contributes to propagation of the dominant
influence of convection and diffusion normal to flow channel. To quantify the effects of fluid loss
the walls of the wormhole. Fluid loss increases on the concentration of reactants transported
the rate of reactants transport to the walls of the along the wormhole, fluid loss was added to the
wormhole, because mass transfer and fluid loss model for the dissolution of a representative
act in the same direction. Conversely, fluid loss cylindrical tube discussed in “Dependence on the
decreases the rate of products transport away Damköhler number.” Fluid loss (i.e., convection
from the walls of the wormhole to the bulk fluid. in the radial direction) was included in the reac-
The overall rate of products transport increases tant mass balance within the cylindrical tube. To
because of the added effect of products leaking obtain an analytical solution, the concentration
off into the formation. These changes lead to an of reactants leaking off into the formation was
increase in the amount of reactants consumed on assumed to equal the concentration at the solid/
the walls of the wormhole and a decrease in the liquid interface. This assumption is valid for
effectiveness of wormhole formation. three cases: zero fluid-loss velocity, reaction-
The effect of fluid loss on the rates of mass rate-limited dissolution and mass-transfer-limited
transfer can be taken into account by correcting dissolution. This assumption is commonly used
the mass-transfer coefficient Kmt used in the to model fracture acidizing and has been
absence of fluid loss. In considering rapid mass described in more detail in the literature (Settari,
transfer resulting from both diffusion and con- 1993). The resulting expression for the concen-
vection, the corrected mass-transfer coefficient tration profile along the length of the flow chan-
K is (Cussler, 1984) nel is similar to that in Appendix Eq. 8:
v C
K= = (1 − B)e ( ) + B ,
L − Da + ξ
( vL Kmt )
, (11) (12)
e −1 Co
where vL is the fluid-loss velocity. The value of where
vL is negative for flow in the direction of diffu-  vL 
sion (i.e., for the transport of reactants to the 1 − 
Da  K1 
surface). At typical reservoir conditions, the cor- B=
( )
. (13)
rected and uncorrected mass-transfer coefficients ( Da + ξ) 1 + νKeq
differ by less than an order of magnitude, a dif- The exponent in Appendix Eq. 12 includes Da
ference that becomes less significant as vL /Kmt as defined in Appendix Eq. 9 (with κ based on

the corrected mass-transfer coefficients, as more fluid than the shorter wormholes. As a result,
defined in Appendix Eq. 11) and an additional the longer wormholes propagate more rapidly, while
dimensionless fluid-loss term: the shorter wormholes eventually stop growing as
they receive an insufficient amount of reactant
πdlvL  κ
ξ= 1 −  . (14) (Hoefner and Fogler, 1988; Hung et al., 1989). The
q  K1  number of wormholes that are capable of penetrating
The obvious importance of vL is demonstrated to a given depth depends on the length of the com-
by the fluid-loss term ξ. As vL decreases, the peting wormholes and the distance between them.
influence of ξ decreases and Da eventually dom- For example, results from simulations in linear
inates the dissolution. Such is the case in linear geometry have shown that wormhole interactions
coreflood experiments in the laboratory. Not so significantly reduce the flow rate in a wormhole
obvious is the dependence of ξ on the degree of when the distance between neighboring wormholes
mass-transfer limitations. When dissolution is is less than the wormhole length (Hoefner and
limited by the transport of reactants to the sur- Fogler, 1988; Buijse, 1997). In general, the number
face (κ = K1), ξ becomes zero. Physically, this of dominant wormholes decreases as the depth of
trend is due to the interface concentration of penetration increases. In addition, the number of
reactants (and, consequently, the concentration dominant wormholes depends on Da within each
of reactants leaking off into the formation) wormhole. At the extremes of high and low Da val-
becoming negligible. The leakoff concentration ues, face dissolution and uniform dissolution result
can become negligible even at high rates of fluid in a single dissolution structure and an infinite num-
loss when the rate of surface reaction is much ber of dissolution channels, respectively. Between
more rapid than the rate of mass transfer and the these extremes, a gradual increase in the wormhole
rate of fluid loss—i.e., when the dissolution is density is expected as Da is decreased.
mass-transfer limited. Thus, for a reactants- To maintain efficient wormhole formation, the
transport-limited system such as HCl-limestone, optimum Da must be maintained in each of the
dissolution is dependent on transport and reac- competing wormholes. Therefore, the optimum
tion within the wormhole and, hence, Da dictates injection rate required for efficient wormhole for-
the reactant consumption (e.g., Da/(Da + ξ) = mation in laboratory experiments must be scaled by
0.999 for the HCl-limestone system at 200°F the number of wormholes that will form in the for-
with a high vL). On the other hand, for fluid- mation. Some investigators have proposed scaling
mineral systems that are influenced by the kinetics laboratory data by the relative surface areas (i.e.,
of the surface reaction (e.g., HCl-dolomite and maintaining the same superficial injection velocity)
EDTA-limestone at ambient temperature), ξ and (Buijse, 1997; Frick et al., 1994b). However, this
Da are both significant. Under these conditions, scaling inherently assumes that the same number
an optimum Da + ξ may exist. Hence, the depen- of wormholes will form per unit surface area in the
dence of wormhole formation on only Da is field as in the laboratory. In general, this assump-
valid for wormhole formation in all fluid-mineral tion overestimates the number of wormholes that
systems when the value of ξ is low (i.e., in dam- will form and, correspondingly, overpredicts the
aged formations with low permeability, when effect of wormhole competition on the optimum
fluid-loss additives are present or when the injection rate. The importance of maintaining an
wormholes are relatively short) or the dissolution optimum injection rate per wormhole (i.e., the opti-
is limited by the transport of reactants to the mum Da in each wormhole) is demonstrated by
surface. These conditions are consistent with the results from laboratory studies. Different wormhole
assumption used in deriving the analytical solu- structures were observed during linear coreflood
tion to this problem. experiments when the cross-sectional area for flow
was varied and the superficial velocity was held
• Wormhole competition constant (Buijse, 1997). In contrast, similar worm-
During field treatments, many wormholes form and hole structures were observed when the cross-
compete for the injected fluid. The longer worm- sectional area was changed and the injection rate
holes, which typically have larger diameters, accept was held constant (C. N. Fredd, unpubl. data, 1989).

Application of the optimum Damköhler where rAt is the rate of reactant consumption at the tip,
number MWA and ρA are the molecular weight and density of
the reactant, respectively, and Xt is the volumetric dis-
The previous section demonstrates the conditions at solving power (i.e., volume of mineral dissolved by
which wormhole formation is dominated by the depen- a given volume of reactant) at the tip. The rate of reac-
dence on Da. Because those conditions are typical of tant consumption at the tip is given by Appendix Eqs. 4
conditions often experienced during field treatments, and 5 with the mass-transfer coefficients replaced by
it is reasonable to assume that effective wormhole for- the superficial velocity in the wormhole at the tip,
mation will occur if Da is maintained near its optimum which is a function of the wormhole diameter. Appen-
value. Therefore, an optimum Da of 0.29 was used as dix Eq. 15 is similar to the growth rate expression intro-
the basis for predicting optimum injection conditions duced by Hung et al. (1989). These calculations require
during field treatments. To scale the laboratory obser- an iterative solution because Da and the growth rate are
vations to the field, additional information about the functions of the wormhole diameter. (If the system is
fluid loss and wormhole competition is required. mass-transfer limited, Da is independent of the worm-
Although the effects of fluid loss and wormhole com- hole diameter.) The injection rate determined in this
petition have not been rigorously investigated, this manner represents an average injection rate along the
information can be estimated from the literature. These length of the dominant wormhole that is required to
estimates, along with the method of calculating the maintain the optimum Da in that wormhole. To achieve
optimum injection conditions, are discussed here. that average injection rate, the injection rate into the
The predictions of the optimum injection conditions wormhole must be larger to compensate for fluid loss
are based on determining the injection rate required from the walls of the wormhole. The required injection
to maintain an optimum Da of 0.29. It was assumed rate qi is approximated as
necessary to maintain the optimum Da in only the
dominant wormhole channels (i.e., the longer worm- qi = q + q L , (16)
holes that carry most of the injected fluid, as shown
in Appendix Fig. 7). Therefore, the calculations where q is the average optimum flow rate within the
depend on the number of dominant wormholes that wormhole (such that the optimum Da is maintained)
can penetrate to a depth l. Because the number of and qL is the rate of fluid loss. The rate of fluid loss
wormholes, diameter of the wormholes, growth rate was estimated from predictions for fluid loss from
and fluid-loss rate change as the depth of penetration a single wormhole (Wang et al., 1993).
increases, the wormhole properties are evaluated for Once the optimum flow rate in a single wormhole
discrete increases in the depth of penetration. The has been calculated, the optimum injection rate in the
fluid-mineral system and temperature are specified, field is calculated by scaling by the number of domi-
and the relevant parameters are evaluated (i.e., diffu- nant wormholes that will penetrate to the given depth.
sion coefficients, surface reaction rates and effective Although no study has reported wormhole density as
equilibrium constants). The optimum injection rate in a function of injection conditions, two studies have
a single wormhole of length l is then determined from provided a basis for estimating the parameter. The
Appendix Eq. 9. The change in wormhole diameter is results of network model simulations indicate that the
determined from the volume of reactant consumed number of wormholes should scale roughly with the
within the wormhole channel, which is a function of ratio of the inlet area to the depth of penetration
Da and the wormhole growth rate. The growth rate is (Hoefner and Fogler, 1988). A similar observation was
calculated by assuming all reactant transported to the made from a capillary tube model (Buijse, 1997). Both
tip of the wormhole propagates the wormhole channel of these studies are based on wormhole structures con-
by dissolving the tip over the same cross-sectional sistent with the dominant wormhole channels. Because
area as the existing wormhole. The growth rate is the goal of this extension is to predict the growth and
competition of dominant wormhole channels, the
dl MWA
= rAt Xt , (15) wormhole density was estimated on the basis of the
dt ρA scaling observed by Hoefner and Fogler (1988).

Optimum injection strategies

10
Optimum injection strategies for field treatments were
Face dissolution
predicted for an optimum Da of 0.29. The predictions 8
are based on three main assumptions: the value of the Conical channels
optimum Da is not affected by fluid loss or wormhole 6
Skin effect
competition, the rate of fluid loss is consistent with 4
Ramified wormholes
that predicted for a single wormhole, and wormhole

density in the near-wellbore region is consistent with 2
that predicted by network models. Because these
0
assumptions must be verified experimentally, the opti- Wormholes
mum injection strategies presented in this section are –2
considered qualitative trends.
–4
Before optimum injection strategies for field treat-
Volume injected
ments can be presented, it is first necessary to under-
stand how the various dissolution structures, and Appendix Figure 8. Dependence of skin effect evolution on
hence Da, influence the effectiveness of matrix stimu- the dissolution structure.
lation treatments. Typical skin effect evolution curves
are shown in Appendix Fig. 8 for various dissolution
structures. The skin effect was calculated using the
three-zone model described by Frick et al. (1994a)
and the depth of stimulation as determined from the
dependence on Da. The results demonstrate that no
significant skin effect evolution is observed when face
dissolution occurs. As the value of Da decreases, the
dissolution structure changes to conical channels and
a slight decrease in skin effect is observed. Near the
optimum Da, wormhole formation results in effective
stimulation as evidenced by the negative skin effect.
Additional decreases in Da result in more ramified
wormholes and a less effective evolution of skin
effect. This later trend of decreasing effectiveness
with decreasing Da (below the optimum value) is
consistent with that predicted by Frick et al. for an
increasing injection rate.
Appendix Fig. 8 demonstrates that the formation
of dominant wormhole channels represents the most
effective mode of stimulation for a given volume of
fluid injected. This effectiveness is due to the domi-
nant wormhole channels providing the greatest depth
of penetration, as shown in Appendix Fig. 9. Obvi-
ously the other types of dissolution structures are also
capable of stimulation beyond the damaged zone if
a sufficient volume of fluid is injected. The trade-off
is the cost of the additional fluid injected and possible
loss of integrity of the near-wellbore matrix as a result
of excessive dissolution. Appendix Figure 9. Schematic of dissolution structures and
The influence of the dissolution structure on the relative depth of penetration obtained by injecting the same
volume of fluid at different Da values.
effectiveness of the matrix stimulation treatment
demonstrates the importance of maintaining the opti-

mum Da. The question that now must be addressed

8
is how to maintain that optimum injection condition HCl HF Acetic acid Emulsified
when dissolution is influenced by fluid loss and worm- HCl
Normalized optimum injection rate

hole competition. The optimum condition can be main- 6
tained by increasing the injection rate as the depth of
penetration increases. This approach is demonstrated
in Appendix Fig. 10, which shows the normalized opti- 4 Changing fluid
mum injection rate required to maintain the optimum type at constant
injection rate
Da as a function of the depth of penetration. The opti-
mum injection rate is normalized by the optimum 2
injection rate for HCl at 200ºF with zero fluid loss.
The curve represents injection rates at which efficient
wormhole formation occurs. At injection rates above 0
and below the optimum curve, ramified wormholes Depth of penetration
and face dissolution form, respectively. The curves

reveal that the injection rate must be increased signifi- Appendix Figure 11. Optimum injection strategies for vari-
ous acid systems at 200°F. The shaded area (Paccaloni,
cantly as the depth of penetration increases. This 1995) represents typical injection rates used in matrix acidiz-
increase in injection rate is necessary to offset the ing treatments.
effects of fluid loss from the wormhole channels. The
amount of fluid diverted to the dominant wormholes
as a result of wormhole competition is not sufficient rates used in conventional matrix stimulation treat-
to overcome the effects of fluid loss. When vL is low ments (Paccaloni, 1995). The results show that HCl
(such as in low-permeability damaged zones), the rate is unable to achieve significant penetration without
at which the injection rate must be increased is less requiring excessive injection rates that would fracture
significant. This effect of fluid loss is consistent with the formation. In contrast, alternative fluids such as
experimental data that demonstrate a decrease in stim- weak acids and emulsified HCl can stimulate to
ulation efficiency when the permeability is increased increasingly deeper depths without exceeding the
and the injection rate held constant (Frick et al., 1994b; same maximum injection rate. Under these conditions,
Mostofizadeh and Economides, 1994). emulsified HCl would be the most effective stimula-
The effect of fluid type on the injection strategy in tion fluid. Results for EDTA are similar to those of
limestone formations at 200ºF is shown in Appendix emulsified HCl. The data also demonstrate that weak
Fig. 11. The shaded area represents typical injection acids and emulsified HCl are more effective than
aqueous HCl when the treatments are limited to low
60 injection rates. This effectiveness at low injection
rates is consistent with results reported by previous
investigators (Hoefner and Fogler, 1985; Fredd and
50
Typical fluid loss Fogler, 1998a; Takulpakdee, 1998).
40 Appendix Fig. 11 reveals an alternative injection
strategy, as indicated by the dashed arrow. This strat-
30 egy involves maintaining a constant injection rate and
gradually changing the reactant type (thereby chang-
Low fluid loss
20 ing the overall rate of dissolution). For example, a
more effective stimulation could be achieved by
10 injecting an HCl–acetic acid blend that is gradually
changed from HCl to acetic acid as the depth of pene-
0 tration increases.
Depth of penetration The optimum injection rate is a strong function of
temperature, as shown in Appendix Fig. 12. To obtain
Appendix Figure 10. Effect of fluid loss on optimum injec- a particular depth of penetration, the optimum injection
tion strategies for effective wormhole formation with HCl at
200°F. rate must be increased as the temperature increases.

wormhole channel increases with increasing worm-

8 hole branching. Fortunately, effective wormhole for-
HCl HF mation tends to minimize fluid loss by minimizing
6
wormhole branching. Therefore, effective wormhole
formation provides the combined benefits of reducing
Acetic acid
fluid loss from the dominant channels and reducing
4 the volume of fluid required to achieve a given depth
of penetration. Thus, maintaining the optimum Da can
Emulsified HCl
result in significant improvements in the effectiveness
2
of matrix stimulation treatments.
0
0 100 200 300 400 500 Conclusions
Temperature (°F)
Wormhole formation in carbonate porous media
Appendix Figure 12. Effect of temperature on the optimum involves complex interplay between a variety of trans-
injection rate required to achieve a given depth of penetra- port and reaction processes. Stimulation fluids, such as
tion with various acid systems. The shaded area represents strong acids, weak acids and chelating agents, are influ-
typical injection rates used in matrix acidizing treatments.
enced by the effects of convection, reactants transport,
reversible surface reactions and products transport.
Thus, to adequately describe the dissolution phenome-
This trend is consistent with experimental results with non, a generalized description of carbonate dissolution
HCl where increasing the temperature resulted in an is required. When the combined effects of transport and
increase in the optimum injection rate (Wang et al., reaction are accounted for, a common dependence on
1993; Fredd, 1998). A much more significant depen- Da for flow and reaction is observed. The value of Da
dence of the optimum injection rate on temperature dictates the type of wormhole structures that is formed
was predicted by Huang et al. (1997) because of the by systems with various degrees of transport and reac-
emphasis they placed on the kinetics of the surface tion limitations. In addition, there exists an optimum
reaction. Appendix Fig. 12 demonstrates that optimal Da at which dominant wormhole channels are formed
stimulation with HCl is limited to low temperatures if and the number of pore volumes to breakthrough is
excessive injection rates and, consequently, fracturing minimized. This optimum Da occurs at approximately
the formation are to be avoided. The figure also reveals 0.29 for a wide range of fluid-mineral systems. The
that weak acids and emulsified HCl are more effective existence of an optimum Da was substantiated by net-
than aqueous HCl for stimulating high-temperature work model simulations.
limestone formations. The use of laboratory data to predict the optimum
This section has provided a means of predicting injection conditions in the field, such as fluid type and
optimum injection strategies based on laboratory data. injection strategy, is complicated by the effects of fluid
With the effective strategies demonstrated, it must be loss and wormhole competition. From estimates of the
emphasized that the effectiveness of a matrix stimula- effects of these processes on wormhole formation, the
tion treatment depends significantly on the depth of theory for the optimum Da has been extended to pre-
penetration, as shown in Appendix Fig. 8. The depth dict optimum field conditions. Direct extension of the
of penetration is typically limited by fluid loss from optimum Da to the field is valid for three cases: low
the wormholes (Nierode and Williams, 1971; Hung fluid-loss velocity, reaction-rate-limited dissolution and
et al., 1989). Therefore, limiting fluid loss is critical to mass-transfer-limited dissolution. Because these condi-
obtaining effective penetration depths, whereas main- tions often exist in the field, injection strategies required
taining an optimum Da is critical to maintaining effec- to maintain the optimum Da can be used to determined
tive wormhole growth. (Fluid loss from wormholes is optimum injection conditions in the field. Results pre-
inhibited when foamed acids are used for matrix dict that to maintain efficient wormhole formation,
acidization; Bernadiner et al., 1992.) These influences either the injection rate should be increased or the over-
are not independent because fluid loss from the main all dissolution rate decreased (by changing the fluid

type) as the depth of penetration increases. At the injec-

tion rates commonly used for matrix acidizing, conven-
tional treatments with HCl are optimal only in low-
temperature formations. Fluids such as weak acids and
emulsified HCl provide effective alternatives to HCl at
high temperatures and low injection rates. The results
discussed in this Appendix to Chapter 16 provide quali-
tative predictions of optimum injection strategies that
are in agreement with experimental observations. A
rigorous investigation of the effects of fluid loss and
wormhole competition on wormhole formation will
enable more quantitative predictions of the optimum
injection conditions.

◆
Carbonate Acidizing Design
J. A. Robert and C. W. Crowe, Schlumberger Dowell
17-1. Introduction microscopic fossils have low permeabilities in com-

parison with those of oolithic limestones, although
Carbonate rocks, by definition, contain more than they can have similar porosities.
50% carbonate minerals. The most common carbon- After deposition, carbonate sediments are subject
ate minerals are calcite (calcium carbonate, CaCO3) to chemical and physical transformation (diagenesis).
and dolomite, a single mineral associating 1 mole of Chemical modifications consist primarily of the evo-
CaCO3 with 1 mole of MgCO3. Carbonate rocks are lution of the sediments to stable forms such as cal-
typically classified by the calcite:dolomite ratio, and cite and dolomite. This process is called neomor-
those with a ratio higher than 50% are generally phism. Dolomites result from evolution under long-
called limestones. Carbonate rocks present singular term contact with fluids with a high magnesium con-
physical characteristics, such as double porosity or tent. Dolomitization usually increases porosity but
high permeability contrasts, which are inherited impairs permeability by precipitating dolomite crys-
from their process of formation. tals. When dolomites are in contact with rainwater,
Hydrochloric acid (HCl) is usually selected for the inverse process of dedolomitization can occur.
carbonate acidizing. It reacts readily with carbonate Modifications of permeability and porosity can
minerals and is available in large quantities at a rel- also result from other chemical transformations such
atively low price. Whereas the purpose of sandstone as dissolution, reprecipitation and cementation while
acidizing is to dissolve the damage, acid is used in in contact with subterranean water or molecular dif-
carbonate formations to dissolve the matrix and fusion during severe subsurface conditions. Stress
bypass the damage. For this reason, both damage and pressure modifications occurring upon burial
and rock characteristics must be taken into account of the sediments may also entail a lithologic change,
when designing the treatment. with porosity and permeability reductions. High in-
situ pressures result in grain compaction, with loss
of most of the interstitial water. Higher pressures
17-2. Rock and damage character- trigger physical dissolution with immediate repre-
istics in carbonate formations cipitation causing pore lining. If the interstitial water
can be expelled, grain joints are also dissolved and
17-2.1. Rock characteristics reprecipitated, decreasing the rock porosity to nearly
zero and creating fissures of residual circulation.
Carbonate rocks are sedimentary rocks resulting Carbonate sediments are much more sensitive to
mostly from organic activity. The vast majority of these phenomena than sandstones are. For example,
these sediments is composed of skeletons of marine a chalky mud becomes a compact calcitic rock at
organisms that vary in size from a few microns to depths greater than 6500 ft if all the interstitial
several centimeters. Some carbonate sediments are water escapes. Mechanical stress modifications also
formed by the chemical precipitation of CaCO3. induce fractures, which are important for the eco-
Because they consist of noneroded, homogeneous nomic viability of carbonate reservoirs.
materials, carbonate sediments usually have a high Carbonate reservoirs present a wide range of
initial porosity. Conversely, because the permeabil- porosities and permeabilities, depending on the
ity of sediments depends mainly on grain size, car- degree of reprecipitation and cementation. Many
bonate rocks present a wide range of permeabilities. carbonate reservoirs are fissured under the action of
For example, chalks formed from sediments of tectonic stresses and behave like homogeneous, pri-

mary porosity reservoirs, with an apparent perme- HCl can be retarded through the use of emulsions
ability several orders of magnitude higher than the or microemulsions to prevent rock deconsolidation.
rock permeability. In this case the wormholes are replaced by a more
uniform increase of pore size throughout the stimu-
lated zone. Deep penetration of live acid is obtained
17-2.2. Damage characteristics by reducing the contact area between the acid and
Damage identification is a prerequisite to the proper the rock.
design of a carbonate acidizing treatment. In forma-
tions that are highly sensitive to acid, nonacidic for-
mulations should be used, and the choice of the
17-3.2. Historical background
treating fluid is usually determined by the type of Acidizing was one of the earliest methods developed
damage. Furthermore, the volume of treating fluid for increasing well productivity, along with nitro
depends on the extent and location of the damage. shooting. The technique was first used in 1895, with
All the types of damage that occur in sandstone patents issued in 1896. The original Frasch (1896)
formations (see Chapter 14) can occur in carbonate patent describes a technique in which HCl is injected
formations, except those related to the presence of into a limestone formation, where it reacts to create
clay particles in the matrix. In addition, poorly channels within the rock. Frasch did not envision the
cemented chalks can be permanently impaired by use of corrosion inhibitors, and his process required
acidic water-base fluids, which can easily dissolve pipe lined with rubber or some other corrosion-resis-
the calcitic cement material, resulting in formation tant coating.
compaction from a loss of mechanical strength. Although the early acidizing treatments produced
Problems related to fluid surface tension (i.e., some impressive results, their actual use declined
water blocks) are not expected in fissured reservoirs within a few years, possibly because of corrosion
with low matrix permeabilities. In this type of for- problems resulting from the uninhibited acid. About
mation, induced damage is concentrated in the fis- 30 years later, the Gypsy Oil Company performed a
sures, and a greater invasion depth is expected than number of well treatments in sandstone formations
in a homogeneous reservoir. Pressure variation in the in which inhibited HCl was used in an attempt to
fissures can also result in the precipitation of mineral remove “gyp” deposits (calcium sulfate). The inhib-
(scales) and hydrocarbon (asphaltenes) deposits. itor used had been developed earlier in the steel
industry for the acid pickling of metals. The treat-
ment results were mostly unimpressive, and no
17-3. Carbonate acidizing with patents were filed on the process.
The modern era of acidizing began on February
hydrochloric acid 11, 1932, when the Dow Chemical Company
17-3.1. Introduction siphoned 500 gal of HCl containing 2 gal of an
arsenic inhibitor into a well owned by the Pure Oil
HCl is generally selected for carbonate acidizing. It Company and displaced it with an oil flush. This was
can be replaced by organic acids, mainly to minimize the first use of an inhibited acid on a limestone for-
corrosion problems at temperatures greater than mation. The previously dead well responded by pro-
400°F [205°C]. The purpose of acidizing with HCl is ducing 16 BOPD.
either to bypass the damage by creating high-conduc- The first treatments were apparently done in an
tivity channels (also called wormholes) or to etch par- attempt to dispose of surplus HCl. However, it was
tially plugged fissures in low-permeability fissured soon noted that these acid disposal wells accepted
formations. Although mere permeability restoration fluid at an increasing rate. Treatments performed
is usually targeted for sandstone formations (zero later on brine-producing wells at the Dow plant in
damage skin effect), carbonate acidizing commonly Midland, Michigan, resulted in increased brine flow,
results in negative skin effects. This is due to the prompting the idea that the process might also have
reopening of natural fissures and creation of high- application for oil wells.
permeability wormholes in the near-wellbore area. The use of inhibited acid to treat oil wells spread
quickly, and the Dow Well Service Group was
17-2 Carbonate Acidizing Design

formed to exploit this new process. The first two This solid-liquid reaction takes place at the rock
words of the company’s name were combined, surface. In excess of HCl, it is complete and irre-
becoming Dowell, Inc., in November 1932. Other versible. Table 17-1 lists the quantities of different
service companies soon followed. Within 3 years, by-products of the reaction of 15% HCl with calcite.
acidizing was practiced widely. Calcium chloride (CaCl2) and magnesium chloride
The first hydraulic fracturing treatments were (MgCl2) are highly soluble in spent acid and present
probably performed with acid, although they were no risk of reprecipitation.
not recognized at the time. Wells in tight carbonate
formations would usually not accept acid until a crit-
Table 17-1. Quantities of by-products created
ical pressure was reached. However, after this pres- by the reaction of 15% HCl and limestone.
sure was reached, acid could be easily injected at
high rates. It was later recognized that these wells HCl CaCO3 CO2 H2O CaCl2
had been hydraulically fractured. For this reason, 1000 gal 1846 lbm 6616 scf 343 lbm [41 gal] 2121 lbm
later hydraulic fracturing patents were never
enforced against acid fracturing treatments.
The reaction rate Kr, expressing the moles of acid
reacting per square meter of wetted surface area per
17-3.3. Reactivity of carbonate minerals second, depends primarily on the temperature and
with hydrochloric acid acid concentration:
Calcium carbonate reacts with HCl to produce car- K r = kr C n , (17-1)
bon dioxide, water and calcium chloride. This sys-
tem is governed by several chemical reactions where kr is the reaction rate constant, C is the HCl
(Garrels and Christ, 1965), listed as follows with concentration in mol/m3, and n is the reaction order.
their values of the equilibrium constant K: The terms kr and n have been determined experimen-
tally (Lund et al., 1973, 1975; Li et al., 1993). The
CaCO3 ↔ Ca 2+ + CO 32− KCaCO =
[Ca ][CO ] = 10
2+ 2−
3 −8.3 constant kr varies with temperature according to
3
[CaCO3 ] Arrhenius’ law, and the coefficient n varies with tem-
CO 32− + H + ↔ HCO 3− KHCO =

[H + ][CO32− ]
= 10 −10.3
perature for dolomite. The reaction of limestone with
[ 3]
−
3
HCO − HCl is fast and cannot be measured at room tempera-
ture. Lund et al. (1975) measured limestone reactiv-
HCO 3− + H + ↔ H 2 CO 3 KH CO =
[ H + ][HCO 3− ]
= 10 −6.4 ity with a rotating disk apparatus at a maximum tem-
2
[H 2 CO3 ]
3
perature of 28.4°F [–12°C]. Dolomite is less reac-
H 2 O ↔ H + OH+ −
KH O = [H + ][OH − ] = 10 −14 tive, and reaction rates can be measured at much
[H CO ]
2
higher temperatures (212°F [100°C]).

H 2 CO 3 ↔ H 2 O + CO 2 KCO = 2 3 = 10 −1.47 The reaction rates obtained by Lund et al. are
2
pCO 2
given in Chapter 16. The following example pro-
where the use of square brackets indicates the con- vides the order of magnitude of the reactivity of
centration in mol/L and p CO2 is the pressure of car- limestone and dolomite: assuming a reaction-rate-
bon dioxide. limited process (see Section 17-3.4) and using data
Because H2CO3 is a weak acid, and the concentra- from Lund et al. (1973, 1975), it can be calculated
tions of HCO3– and CO32– are negligible in presence that at 75°F [25°C], the thickness of a rotating disk
of HCl, the reaction of CaCO3 with HCl can be writ- of calcite reacting with excess 5% HCl decreases by
ten as 1.2 mm/min. This value drops to 1.4 µm/min for
dolomite at the same conditions.
CaCO3 + 2HCl → CaCl2 + H2O + CO2 Weak acids such as formic acid or acetic acid also
For dolomite, the reaction equation becomes react with carbonate rocks. However, the high con-
centration of CO2 produced by the reaction prevents
CaMg(CO3)2 + 4HCl → it from going to completion, even in the presence of
CaCl2 + MgCl2 + 2H2O + 2CO2 excess fresh acid.

17-3.4. Acidizing physics critical flow rate and that within the wormholing
region the volume required for breakthrough
• Reaction process increases with the rate at a power of 1⁄3. These
A solid-liquid reaction such as HCl with carbonate results were confirmed by Wang (1993) for calcite
minerals involves the transport of hydronium ions and dolomite at high temperatures.
(H3O+) to the rock surface, reaction of the ions Wormholing can be explained by the instability
with the rock and transport of the reaction prod- of the acidizing phenomenon: bigger pores tend to
ucts from the surface to solution. When one of the receive more acid, which increases both their area
steps is much slower than the others, it imposes and length. This eventually creates a macroscopic
itself on the reaction rate and is said to be the lim- channel, or wormhole, that accepts more acid than
iting step. Lund et al. (1973, 1975) studied the the surrounding pores and propagates through the
reaction rate of calcite and dolomite with HCl core. Wormhole branching depends on the injec-
using the rotating disk technique. Theory predicts tion rate. Near the wormholing threshold, a single,
that for a diffusion-limited process, the rate of dis- thick wormhole is formed. As the injection rate is
solution is proportional to the square root of the increased, a denser network of thinner channels is
rotation speed and to the bulk concentration. Lund created.
et al. found that the reaction with calcite at room Computer network simulations have been used
temperature and higher is transport limited, where- to replicate the wormholing phenomenon (Hoefner
as dolomite switches from reaction rate limited to and Fogler, 1988; Pichler et al., 1992; Lauritzen et
diffusion rate limited between 125° and 212°F al., 1992). The porous medium is represented by
[50° and 100°C]. These results were confirmed by a two-dimensional (2D) network of capillaries. In
de Rozières et al. (1994). On the other hand, Wang each pore, the flow obeys Kirchoff’s laws, and the
et al. (1993) argued that at the pore level the reac- growth of each pore is assumed to follow a given
tion rate determines the overall rate of acid con- function. These 2D models are extensions of the
sumption, thereby implying that the process is one-dimensional analytical model of Schechter
reaction rate limited. and Gidley (1969) and Guin and Schechter (1971).
Reaction rate limited and diffusion rate limited Similar results are obtained with regard to the
represent extreme cases in which one phenomenon development of unstable patterns. The main
is much slower than the others. There is also an advantage of the models is that the dissolution
intermediate case in which different processes function and flow parameters can be changed in a
have kinetics of the same order of magnitude and much wider range than experimentally possible. In
influence each other. Both the diffusion rate and these simulations, the occurrence of fractal behav-
the reaction rate depend on ion concentrations at ior is bounded by two extreme cases (Fig. 17-1):
the surface. Therefore, for surface concentrations
of the same range as the bulk concentration, trans- – At low velocity, molecular diffusion is predomi-
port by diffusion and the reaction rate are interde- nant, and the solution becomes saturated before
pendent phenomena. This is the case for dolomite any appreciable volume can enter the pore.
and HCl at intermediate temperatures and for cal- – At high velocity, the thickness of the boundary
cite and ethylenediaminetetraacetic acid (EDTA) layer becomes so small that the kinetics crosses
at room temperature (Fredd et al., 1997). over to surface reaction rate limited. The veloc-
• Wormholing phenomenon ity is sufficiently high compared with the disso-
lution rate to allow the invasion of all pores by
Limestone cores acidized with HCl show the for- fresh solution. Uniform etching occurs.
mation of macroscopic channels, called worm-
holes. Most of the published experiments were Daccord et al. (1993) presented a dimensionless
performed with linear cores (radial section, flow analysis in which the rate of acid diffusion toward
parallel to the longitudinal axis). They show that the pore walls is compared with the rate of acid con-
the acid injection rate affects the geometry of the vection into the pore. For a wormhole to form, the
channels and the amount of acid required for initial pore radius must be large enough to allow acid
breakthrough. Daccord et al. (1989) found that transport beyond the pore inlet (see Sidebar 17A).
wormholing occurs in limestone cores above a

mq k
Pe = , (17-2)
φAD
where m is a coefficient introduced for consistency
with earlier publications, q is the total injection
rate, k is the matrix permeability, φ is the porosity,
A is the area perpendicular to flow, and D is the
diffusion constant. The coefficient m is equal to
1 for a linear geometry and 2 for radial flow.
Table 17-2 shows the results of linear core tests
with various values of acid concentration, Pe and
temperature. The dimensionless volume to break-
through is the ratio of the pore volumes to break-
through to the volume required for complete
dissolution of the core. The limit for wormholing is
between Pe = 10–3 and Pe = 10–2. The tendency for
the critical Peclet number to increase with acid con-
centration has been explained by the onset of grav-
Figure 17-1. Dissolution patterns observed with a network ity phenomena caused by the difference between
model (Hoefner and Fogler, 1988). The width of the bonds the fresh and spent acid densities, which creates
is proportional to the amount of material removed for 6 additional currents near the pore walls that modify
orders of magnitude of the Peclet number (1 = lowest,
6 = highest). Flow is from top to bottom.
the effective acid diffusivity (Daccord et al., 1989).
Fredd et al. (1997) used the generalized
Damköhler number Da instead of Pe:
2πrLK
Wang et al. (1993) presented a different analysis but Da = , (17-3)
came also to the conclusion that a critical initial pore qc
size is required to initiate a wormhole. where r is the capillary radius, L is the capillary
According to Daccord et al. (1989, 1993), the length, K the overall reaction rate taking into
Peclet number Pe, which represents the ratio of account transport by diffusion and reaction, and
axial flow to radial transport in the pores, is the qc is the rate in the capillary. This method enables
dimensionless variable governing the transition extending the analysis to systems that are not
between compact dissolution at low rates and purely transport limited. Fredd et al. reported a
wormholing at higher rates for a transport-limited series of tests with linear calcite cores and differ-
reaction (i.e., calcite and high-temperature ent solvents (EDTA, acetic acid and 1.7% HCl
dolomite). The Peclet number is [0.5N]). For all systems the critical Damköhler
number is equal to 0.29.
Table 17-2. Acidizing results for Indiana limestone cores.
Acid Concentration Temperature Pe Dimensionless Volume Reference

(N) to Breakthrough
0.147 75°F [25°C] 5.5E–3 3.2E–3 Wang (1993)†

0.235 125°F [50°C] 8.9E–3 7.5E–3 G. Daccord (pers. comm., 1988)
0.47 75°F 4.6E–3 1.0E–2 G. Daccord (pers. comm., 1988)
0.47 175°F [80°C] 6.7E–3 1.1E–2 G. Daccord (pers. comm., 1988)
1 75°F 9.6E–3 3.11E–3 Wang (1993)†
1 125°F 2.2E–2 9.75E–3 Wang (1993)†
4.4 75°F 4.9E–2 6.47E–3 Wang (1993)†
† Data for the near-critical Peclet number

pore length and diameter, which promotes wormhole propa-

17A. Wormhole initiation and propagation gation. Once a wormhole is formed, it tends to capture the
whole flow because the pressure drop ∆p in a wormhole is
Acid transport inside a pore can be schematically represented negligible compared with ∆p in the matrix.
by two perpendicular fluxes: axial transport by convection and Assuming that a pore is a site for wormhole initiation if
transport to the pore walls by diffusion. only one-half of the fresh acid is spent within a length equal
The diffusion constant D expresses the ability of ions to to the pore radius:
migrate when submitted to a concentration gradient: 1/ 3
q cC 0 q r2 
= 2πC 0D  c  . (17A-4)
J =D
dC
, (17A-1) 2  πD 
dx
Equation 17A-4 leads to
where J is the acid flux (i.e., the number of moles passing
through a unit surface per unit time) and dC/dx is the concen- q ur
= = 8, (17A-5)
tration gradient. In a pore, because of acid reaction with the Dr π D
rock, the acid concentration at the walls is lower than the bulk
concentration. Levich (1962) used boundary layer theory to where u is the axial velocity in the capillary.
calculate the rate of diffusion in a capillary: The limit between a transport-limited and a convection-
limited regime is therefore defined by the Peclet number Pe
1/ 3 from Eq. 17A-6. For a low value of Pe (low rate or small
q l2  radius), all the acid is spent at the pore inlet. For a high value
I = 2πC 0D  c  , (17A-2)
 πD  of Pe, fresh acid is transported beyond the pore inlet and a
wormhole is created:
where I is the diffusion flux in the capillary in mol/s, C0 is the
acid concentration at the capillary inlet, qc is the rate in the ur
Pe = . (17A-6)
capillary, and l is the capillary length. D
Under laminar flow conditions, the rate in the capillary qc is
If a low injection rate is applied, all pores are below the
πr 4 ∆p critical rate for wormhole initiation and all the acid is con-
qc = , (17A-3) sumed at the wellbore face. This is compact dissolution.
8µ l
Compact dissolution produces poor stimulation and should
where r is the capillary radius, ∆p is the pressure drop in the be avoided. As the rate increases, the fraction of pores able
capillary, and µ is the dynamic viscosity. to initiate wormholing increases, which leads to a denser net-
Wormholing results from the instability of acid propagation: work of wormholes.
the rate in a wormhole (or a pore) is proportional to r to the A similar analysis can be performed for a reaction-limited
4th power (Eq. 17A-3), whereas the rate of acid consumption, process. In that case, the dimensionless number derived is
equal to the diffusion flux for transport-limited reactions, is the Damköhler number Da, expressing the ratio of axial flow
proportional to the flow rate to the 1⁄3 power (i.e., r 4/3, Eq. 17A-2). to the rate of consumption at the pore walls (Daccord et al.,
Therefore, bigger pores accept more and more acid, com- 1993).
pared with smaller pores, and the acid is used to increase the
Limestone cores acidized with acetic acid exhib- tivity of systems such as EDTA and acetic acid
it wormholes with more branching than acidizing allows fresh acid to penetrate into more pores and
with HCl at the same injection rate, and lower live causes the more homogeneous dissolution pattern.
acid penetration is obtained (Hendrickson, 1972; It also lowers the transition rate between compact
Fredd and Fogler, 1996). Rotating disk measure- dissolution and the wormholing regime (Fredd and
ments indicate that the dissolution of limestone by Fogler, 1996).
acetic acid is mass-transport limited (G. Daccord,
• Radial geometry
pers. comm., 1988; Fredd et al., 1997), and the
effective diffusivity of acetic acid is lower than Very few results have been published for acidizing
that of HCl (Table 17-3). The lower overall reac- tests using radial geometries. Daccord et al. (1989)
Table 17-3. Effective diffusivity coefficients (m2/s) for 0.5N acetic acid at different temperatures.
73°F [23°C] 165°F [75°C] 230°F [110°C] Reference
1.14 × 10–9† 2.7 × 10–9‡ 4.0 × 10–9 Dunn and Stokes (1965)
0.21 × 10–9 0.51 × 10–9 0.74 × 10–9 Nierode and Williams (1970)§
0.08 × 10–9 0.40 × 10–9 0.57 × 10–9 G. Daccord (pers. comm., 1988)
Note: For comparison, the diffusivity for 0.5N HCl at 73°F is 2.4 × 10–9 m2/s.
† Value at 75°F [25°C]
‡ Extrapolated from data at 75°F
§ Extrapolated from data for 2N acetic acid at 199°F [93°C]

used a plaster and water system to obtain casts of CVM

Ac = , (17-5)
wormhole patterns (Fig. 17-2). From the experi- Ω(1 − φ)
mental data, they derived a formula in Eq. 17-4
linking the apparent stimulation radius rac with the where C is the acid concentration, VM is the molar
injected acid volume Vac. The apparent stimulation volume, and Ω is the stoichiometric coefficient (2
radius is calculated from the pressure differential for calcite, 4 for dolomite).
across the core, assuming that the pressure differ- Wormholes in radial geometry exhibit a fractal
ential in the wormholes is null: pattern with a fractal dimension d equal to 1.7.
The constant a is introduced to fit experiments
ar 2 hPe
1/ 3
 r  d  with the model. Other experiments (Frick et al.,
Vac = w  ac  − 1 , (17-4) 1994b) on low-permeability (0.2-md) limestone
Ac  rw  
cores with various acid concentrations and temper-
where rw is the wellbore radius, h is the core atures confirm the trends expressed by Eq. 17-4:
height, Pe is the Peclet number as defined in – Rate of wormhole propagation increases with
Eq. 17-2, and Ac is the acid capacity number: acid strength.
– Rate of wormhole propagation increases as the
temperature is increased (acid diffusion increases).
– Within the wormholing regime, the volume for
breakthrough increases with the injection rate at
1
⁄3 power.
However, the constant a varies with the system
considered. Daccord et al. (1993) found that a =
1
⁄84 for the plaster and water system. Frick et al.’s
(1994b) data indicate that the constant a is approx-
imately 1⁄18 for low-concentration acid (4% HCl)
and 1⁄4 for high-strength acid (30% HCl). These
different values could be explained by gravity
effects or differences in the pores structures of the
systems studied. Discrepancies may also be due to
different test conditions; i.e., data from Frick et al.
were obtained with low-permeability rock and low
rates (maximum Pe of 6E–2), near the compact
dissolution/wormholing transition, whereas
Daccord et al. studied higher Pe values for which
the wormholing regime was fully established.
17-3.5. Application to field design

• Injection rate
To ensure wormhole propagation and successful
treatment, the acid velocity near the wellbore
should be sufficiently high to reach the wormhol-
ing regime. Examples of the rates required to
exceed the critical velocity at the wellbore are
in Table 17-4. The critical rate is calculated from
Eq. 17-2, with the critical Pe taken as 5E–2.
Figure 17-2. Cast of a wormhole pattern in radial geome-
try (Daccord and Lenormand, 1987).

Table 17-4 shows that pump rates applied for

17B. Acidizing case study
matrix acidizing are usually well above the critical
rate for wormholing. Field data generally confirm The well was acidized with 60 gal/ft of 20% HCl pumped
these results and show that good stimulation can be through coiled tubing over a 135-ft interval. The pumping rate
was limited to 2 bbl/min. Permeability varied from 15 to 150
obtained even with moderate pumping rates over md. Foamed diesel was pumped between the acid stages for
large intervals (see the case study in Sidebar 17B). diversion.
Figure 17B-1 compares the production profiles before
It is common practice to increase pumping rates and after stimulation. The new open interval between 31 and
as injectivity increases during a treatment. Apply- 37 ft corresponds to a high-permeability zone. The poorly
ing high rates ensures that all portions of the reser- producing zone between 37 and 79 ft corresponds to a zone
of low porosity, according to the results of a log survey. The
voir reach the wormholing regime, even in case of stimulation increased the well productivity index from 1.1 to
injectivity contrasts between different zones. It also 23.9 STB/D/psi. These results indicate that successful stimu-
lation can be achieved at moderate injection rates, provided
allows sustaining wormhole growth as the stimula- that coverage of the whole producing interval is ensured.
tion radius increases and the velocity at the acid
front decreases. Furthermore, in fissured reservoirs
Cumulatve productivity
where the purpose of the treatment is to clean up 25
index (STB/D/psi)
Before
fissures, applying high rates increases the live acid 20 After
penetration. 15
10
• Acid volumes 5
The results listed in Table 17-2 indicate that at 0
optimum conditions (i.e., the near-critical rate), 0 20 40 60 80 100 120 140
breakthrough is obtained when less than 1% of the Depth (ft)
total core has been dissolved. This indicates that

designing fluid volumes for 5% to 10% rock disso- Figure 17B-1. Flow profiles before and after treatment.
lution over the stimulated area should provide a
conservative design. For a radial geometry, the
volume of acid required for a given porosity using local conditions of mineralogy and acid con-
increase varies with the square of the treatment centration and velocity. Reaction parameters and
radius, assuming homogeneous dissolution. the rate of wormhole growth are correlated from
Figure 17-3 shows the volume of HCl required to experimental data obtained with linear cores for a
increase porosity by 10% for different values of broad range of flow and acid conditions. Tracking
the treatment radius. For example, 50 gal/ft of the wormhole propagation front allows calculating
15% HCl is required to increase the porosity by a skin effect factor s, assuming an infinite perme-
10% up to 2 ft from the wellbore. ability in the stimulated area:
For a more accurate design, use of a numerical r 
simulator is required (Bartko et al., 1997). Using s = − ln  ac  . (17-6)
 rw 
a finite-difference simulator enables tracking acid
velocity and mineralogy evolution. The amount of Table 17-5 lists examples of stimulation radii
rock dissolved as a function of time and acid loca- required for different values of skin effect. The
tion can then be calculated in each grid block completion skin effect resulting from partial com-
Table 17-4. Critical flow rates (bbl/min/10 ft) for transport-limited regimes.
15% HCl 28% HCl

Critical Flow Rate, at Sandface k = 5 md k = 100 md k = 5 md k = 100 md
Cased completion 0.015 0.003 0.031 0.006

Openhole, 0.8-ft diameter wellbore 0.14 0.03 0.3 0.063
Notes: Diffusivities were computed at 150°F [65°C]. For the perforated case, a density of 4 shots per foot (spf) was assumed, with
an 8-in. perforation length and 0.4-in. perforation diameter. Porosity = 15%. These are the critical rates at the wellbore. Higher rates
are required to sustain wormhole growth in the matrix.

300 Table 17-6. Maximum protection times for different

15% HCl acids and temperatures.
250 28% HCl
Volume (gal/ft)
200 Acid Temperature Maximum Protection Time (hr)
150 15% HCl 375°F [190°C] 8

100 400°F [205°C] 4
50 28% HCl 350°F [175°C] 4
0
0 1 2 3 4 5 10%
acetic acid 400°F 24
Distance beyond 0.4-ft wellbore radius (ft)
500°F [260°C] 16
Figure 17-3. Acid volume required to increase porosity
10% (absolute) as a function of the depth of live acid HA + H2O H3O+ + A–
reaction.
The equilibrium is characterized by the dissocia-
tion constant kd:
kd = [H 3O + ][ A − ] [HA ] .
Table 17-5. Stimulation radii for different values
of skin effect. (17-7)
Skin Effect Stimulation Radius (ft) The dissociation constant kd depends on the type
–1 1.1
of acid and varies with temperature. At 75°F, acetic
acid has a dissociation constant of 1.76 × 10–5 mol/L
–2 3
and formic acid has a dissociation constant of 1.77 ×
–3 8
10–4 mol/L. The value of kd increases slightly with
Note: Calculated using Eq. 17-6 for a 0.4-ft wellbore radius temperature up to 175°F [80°C] and then decreases
as the temperature increases further. This implies
that at usual reservoir conditions, organic acids are
pletion or well deviation must be added to the less reactive than at surface conditions.
stimulation skin effect to obtain the total formation Acetic and formic acids react with CaCO3 to form
skin effect. calcium acetate and formate, respectively:
Using a simulator provides a more accurate cal-
culation of skin effect evolution than Eq. 17-6 CaCO3 + 2CH3COOH
because factors such as the decrease of velocity Ca2+ + 2CH3COO– + H2O + CO2
as the acid front progresses into the formation and CaCO3 + 2HCOOH
heterogeneities in the reservoir can be taken into Ca2+ + 2HCOO– + H2O + CO2
account.
Calcium acetate is highly soluble in spent acid
(374 g/L at 75°F). High concentrations of acid, up to
20% to 25%, can be used without any precipitation
17-4. Other formulations problem, although concentrations above 10% are
generally not used (Table 17-7). Calcium formate
17-4.1. Organic acids and magnesium formate are much less soluble (162
Organic acids are used instead of HCl when high and 140 g/L at 75°F, respectively). Formic acid
bottomhole temperatures prevent efficient protection strength should be limited to 9% to 10% to avoid
against corrosion (above 400°F). The two main types calcium formate reprecipitation.
of organic acids used are acetic acid and formic acid.
Acetic acid is easier to inhibit than formic acid and Table 17-7. Quantities of calcite and dolomite
dissolved per volume of acid,
is used more often. Table 17-6 lists examples of cor- assuming complete spending.
rosion inhibition with organic acid and HCl.
Organic acids are weak acids, which do not totally Acid Calcite Dolomite
dissociate in water. The equilibrium reaction is writ- 10% formic 1000 gal 920 lbm [5.4 ft3] 854 lbm [4.8 ft3]
ten as 10% acetic 1000 gal 720 lbm [4.3 ft3] 663 lbm [3.7 ft3]

When weak acids attack CaCO3, an equilibrium 17-4.2. Gelled acids

is established between the chemical species pro-
duced by the reaction (see the first five reactions in Gelled acids were developed primarily for fracturing
Section 17-3.3) and the acid species (HA + H2O but have found some applications in matrix acidiz-
H3O+ + A–). In static conditions, the degree of coming. They are used in acid fracturing to increase the
pletion of the reaction depends on the concentration viscosity and decrease the leakoff rate. The same
of CO2 in solution. For instance, at high-pressure principle applies to matrix acidizing conditions in
conditions (typically above 1000 psi), only one-half fissured or vugular formations with low primary
of 10% acetic acid reacts with limestone at 150°F porosity. In this case, gelled acids are used mainly
[65°C]). In similar conditions, 80% of 10% formic to clean up the high-permeability channels and mini-
acid reacts. Although some live acid remains once mize fluid loss in the lower permeability matrix.
the reaction reaches equilibrium, a low pH is not Gelled acids can also be used as a carrier fluid for
maintained because of buffering by the reaction ball sealers or particulate diverters (flakes).
products. Acetic acid is also beneficial for preventing In the design of gelled acid treatments, the stabil-
ferric hydroxide precipitates, because it creates a ity of the gelling agent at bottomhole temperatures
weak complex with the iron in solution and thus must be checked carefully. Several types of gelling
increases the pH at which hydroxide precipitation agents are used. Xantham gums are adequate for
occurs (Crowe, 1985). This effect is particularly sig- moderate conditions (i.e., temperatures up to 230°F
nificant at low temperatures (i.e., below 125°F). At [110°C]), with the acid strength limited to 15%
higher temperatures, some delay of ferric hydroxide (Crowe et al., 1981). Under more severe conditions,
precipitation is expected. synthetic polymers are more appropriate for use up
Formic and acetic acids can be pumped together, to 400° to 450°F [205° to 230°C].
but usually only one acid is selected. Mixtures of Crowe et al. (1990) showed that under dynamic
organic acid and HCl are also used. The design vol- conditions gelled HCl exhibits the same reaction rate
ume depends on the suspected damage extent around with limestones as ungelled acid. In some cases,
the wellbore. Because organic acids react more reaction rates are accelerated. The reaction rate mea-
homogeneously than HCl, larger volumes are sured is the rate of calcite consumption. It is the
required. Good results have been reported with vol- overall reaction rate, determined by the limiting step,
umes of the order of 100 gal/ft of 20% acid mixture which is acid transport by diffusion. It is generally
if proper placement is ensured (Ridwan and Cannan, agreed (Muhr and Blanshard, 1982) that the rate of
1990). Organic acids are much more expensive than diffusion depends on the solvent viscosity and is not
HCl per unit volume of rock dissolved. The econom- modified by the presence of polymers, at least as
ics of the treatment must be taken into account for long as the distance between the polymer chains is
the design. large compared with the size of the ions in solution.
Acetic acid has other specific applications. The interaction between polymer chains and the
Combined either with an aromatic solvent and a rock surface can affect the overall reaction rate and
mutual solvent to obtain a clear solution or with live acid penetration. If the gel exhibits a non-
methanol, it is used to remove water blocks and Newtonian behavior, the shear rate at the rock sur-
break emulsions. Combined with a highly concen- face can be modified, which may increase the mass
trated corrosion inhibitor, it can be used as comple- transfer and result in a higher reaction rate. Aside
tion fluid to keep the pH low near the wellbore and from this effect, the polymer can plug the smaller
prevent clay swelling or as a perforating fluid. In the pores, acting as a fluid-loss agent. This effect was
latter case, organic acids are preferred to HCl at tem- studied by Nierode and Kruk (1973), who found that
peratures above 200°F [95°C] because their reduced the growth rate of wormholes is maximum for a
reactivity at higher temperatures enables good corro- small concentration of fluid-loss agent.
sion protection for several days. HCl is usually the acid component of gelled mix-
tures. Acid strength varies typically from 5% to

28%. The volume of acid depends on the suspected A recent trend involves adding nitrogen (N2) to the
depth of damage in the fissures and vugs and on emulsion to obtain a triphase system. Static tests
fluid placement efficiency. show that this further reduces the reactivity of the
acid (Guidry et al., 1989). The exact mechanism of
the retardation has not been fully studied. It is gener-
17-4.3. Emulsions ally admitted that N2 reduces the contact area of acid
Emulsions are obtained by mixing acid with a with the rock. With this type of system, the dissolu-
refined-oil-base fluid in the presence of a surfactant. tion pattern is expected to be more homogeneous (in
The stability of the emulsion depends on the temper- comparison with plain HCl at same pump rate), and
ature and the ionic strength of the aqueous phase. relatively large volumes are pumped to obtain large
Some emulsifying agents provide stable emulsions in stimulation radii. Liquid volumes as large as
temperatures up to 300°F [150°C]. Depending on the 500 gal/ft are usually pumped. This type of treatment
type of surfactant, a water-in-oil or an oil-in-water is economically advantageous because as much as
emulsion can be obtained (see Chapter 15). 50% of the volume pumped consists of nonacid flu-
In static conditions, emulsions have been found to ids. To increase the matrix injectivity prior to pump-
lower the overall reaction rate of the acid. Acid-in- ing the emulsion, a pretreatment with plain acid is
oil emulsions are more effective for reaction retarda- usually performed. Furthermore, to allow injection
tion. It is generally agreed that these emulsions build at higher rates than normally prescribed by the frac-
an oil barrier at the rock surface, preventing the acid turing limit, the wells are generally drawn down as
from reacting readily with the substrate. Acid-exter- much as possible, and the shut-in time prior to the
nal emulsions also provide some retardation, which treatment is reduced to the minimum technically
is generally attributed to physical interaction of the possible. Under these conditions, the near-wellbore
oil with the path of acid transport to the rock surface. pressure is much lower than the average reservoir
Few results have been published on core acidizing pressure, and higher matrix rates can be applied.
with emulsions under dynamic conditions. The initial Two- and triphase emulsions are recommended for
studies show that oil-external emulsions can treat the treatment of deep damage or if the purpose of
low-permeability cores more efficiently than plain the treatment is to stimulate the formation to obtain
acid (Horton et al., 1965). a highly negative skin effect.
Limestone cores acidized with emulsions display
a highly permeable network of microwormholes that
reflect significant modification of the process of acid
17-4.4. Microemulsions
transport and reaction. Emulsions tend to stabilize Microemulsions consist of a fine dispersion of oil
the acidizing process by reducing the contact area and acid, stabilized by proper surfactant and cosur-
of the acid with the matrix, thereby decreasing the factant additives. Depending on the concentration of
apparent reaction rate. Measurement of diffusion the different components, an oil-in-acid or acid-in-oil
coefficients in acid-in-oil emulsions by de Rozières emulsion can be obtained. The main difference from
et al. (1994) using the rotating disk technique found macroemulsions is the size of the droplets, which are
that effective diffusion coefficients in these systems reported to be in the range of 0.005 to 0.2 µm.
are as much as 3 orders of magnitude lower than dif- Oil-external microemulsions may behave as a single-
fusion coefficients in plain acid at the same tempera- phase fluid in porous media and sweep oil more easily
ture conditions. Therefore, good stimulation can be than plain acid, facilitating acid injection and flow-
obtained with emulsions at low rates corresponding back. Acid diffusivity in acid-in-oil microemulsions
to the compact dissolution regime with plain acid. is also greatly reduced compared with plain acid, by
Like gelled acids, emulsions are usually prepared at least 2 orders of magnitude (Hoefner and Fogler,
with HCl. Various acid strengths and volumes can be 1985). This results in a more homogeneous attack of
selected for the fractions, with 70:30 acid-in-oil the rock, as observed in core experiments (Hoefner et
emulsion a commonly used system. The viscosity of al., 1987), which presents two advantages. First, it
the emulsion is an important parameter because the should enable acidizing tight carbonate rocks at low
high viscosity of some mixtures used for fracturing rates that correspond to compact dissolution condi-
limits their application to matrix acidizing. tions with plain acid. Second, at higher rates it avoids

the formation of large wormholes, which can be particularly for naturally fissured reservoirs where
detrimental to the mechanical properties of soft for- completion materials have invaded the fissures (see
mations such as chalks. Sidebar 17C).
Despite their advantages, microemulsions are not For shallow damage caused by cake deposition
commonly used in field operations. Because of the in natural fissures, such formulations are highly effi-
high surfactant concentration required, these systems cient when acid is spotted through coiled tubing and
are more expensive and difficult to inhibit. Further- flowed back, possibly several times (Liétard et al.,
more, stability problems have been encountered. In 1995). If an oil-base mud was used, or if damage
most cases similar results can be achieved with material has been in contact with oil downhole, the
macroemulsions. The additional cost of microemul- use of a surfactant and mutual solvent is required to
sions is justified only for sensitive formations where water wet the cake surface and facilitate acid flow in
mechanical stability is a concern. the fissures.
17-4.5. Special treatments 17-4.6. Self-diverting acid

HCl can be used in combination with other chemi- Self-diverting acid, originally developed for fractur-
cals for specific treatments. This section briefly ing, has also been used to improve placement during
reviews the main formulations using HCl. More carbonate acidizing. It consists of HCl mixed with a
detailed description is in Chapter 15. gelling agent and a pH-sensitive crosslinker. Cross-
Blends of alcohol (mainly methanol) and HCl are linking occurs at intermediate values of pH (typically
used for gas well treatments. Alcohol lowers the sur- from 1 to 3.5) corresponding to partially spent acid.
face tension, but not as much as surfactants. How- The lower fresh acid viscosity allows penetration in
ever, because it does not adsorb on the rock, it pen- wormholes and fractures until acid reaction increases
etrates into the formation as deeply as the acid. It also the pH and causes crosslinking, thereby diverting the
increases the vapor pressure of the spent acid. This following acid stages to other portions of the reser-
facilitates cleanup of the spent acid and improves gas voir (Fig. 17-4). Because the gel breaks at a pH
permeability by reducing the residual water satura- above 3.5, flowback presents no problem once fresh
tion. The addition of alcohol slightly slows the reac- acid injection is stopped and the acid is allowed to
tion of acid with the rock and slightly accelerates the spend completely. Like particulate diverters or foams,
corrosion rate. The volume fraction of alcohol can self-diverting acids are pumped in several stages,
vary widely depending on the application, from 20% alternating with regular acid stages. Good results are
to 67%. Mixtures with 67% methanol are stable up reported in fractured formations and in long, open
to 250°F [120°C]. intervals where benzoic flakes or gelled acid has
Mixed with an aromatic solvent and a stabilizer, failed to provide fluid diversion (see Sidebar 17D).
HCl forms a solvent-in-acid emulsion. The volume
fraction of the acid varies typically from 50% to
50
90%. The emulsion stability depends on the solvent 2% 15% SDA 15%
KCl HCl HCl
concentration and temperature. This formulation is
Apparent permeability (md)
40
used to remove paraffin and mixed deposits. It is also
recommended for removing scale and treating wells 30
being converted from producers to injectors.
HCl can also be used in combination with a sur- 20
factant and a chelating agent to remove mud damage
in carbonate formations. The combined action of the 10
additives has been found to successfully disperse and
remove clays and mud damage. In formations pre- 0
0 10 20 30 40 50 60
senting a risk of deconsolidation, HCl can be Time (min)
replaced by a brine or calcium chelant solution such
as EDTA. This type of formulation is recommended Figure 17-4. Effect of self-diverting acid (SDA) on fluid
placement.

17C. Examples of special treatments head pressure decreased by 1200 psi while the first acid
stage was injected into the formation. According to Eq. 17C-1,
Well A this pressure falloff is equivalent to a skin effect decrease of
more than 40:
Well A was an oil producer from a fissured dolomitic reservoir
with an average matrix permeability of 10 md and a porosity 2πkh∆p
of 3%. Damage by mixed silts and organic deposits was sus- ∆s = . (17C-1)
µq
pected. The well was treated with 130 gal/ft of solvent-in-acid
emulsion, with a 90 gal/ft preflush of a mixture of 80% solvent Well C
with acetic acid and mutual solvent. The pumping schedule
was as follows: Well C was converted from an oil producer to a water injector.
1. Preflush: solvent + acetic acid (42 bbl) A well test indicated near-wellbore damage, presumably from
the presence of workover material. The average reservoir
2. Main fluid: solvent-in-acid emulsion (63 bbl) porosity was 16%, and permeability deduced from a porosity
3. Diverter: benzoic acid flakes (31 bbl) log varied from 5 to 500 md across the 200-ft open interval.
4. Repeat steps 1, 2 and 3 The well was treated with 30 gal/ft of 15% HCl laden with sus-
pending agents. The acid was pumped through coiled tubing,
5. Repeat steps 1 and 2 and four foam stages were used for diversion. The injection
6. Displacement: nitrogen profiles before and after the treatment are shown in Fig.
17C-1. The acid significantly improved injectivity in the middle
Production before the treatment had dropped to 1172 STB/D interval, corresponding to a lower permeability region (less
at a wellhead pressure 455 psi. Postjob production increased than 25 md). The porosity log indicated that the bottom part of
to 3580 STB/D at a wellhead pressure of 1179 psi. the interval (below 170 ft) corresponded to a very low perme-
ability zone. The total injectivity increased from 30,000 BWPD
Well B at a wellhead pressure of 1640 psi to 54,000 BWPD at
Well B was an oil producer from a calcitic reservoir containing 1420 psi at the wellhead.
5% clays. Reservoir porosity was 2% and the average matrix
permeability did not exceed a few millidarcies. The production 600
Cumulative injection (1000 STB/D)
rate prior to the treatment indicated a damage skin effect of Before
approximately 40. The treatment was executed with 75 gal/ft 500 After
of 15% HCl laden with suspending and sequestering agents,
preceded by a preflush of suspending agent-laden brine. The 400
schedule was as follows:
1. Preflush: suspending agent–laden brine (63 bbl) 300
2. Main fluid: 15% HCl with suspending agent + nitrogen
(107 bbl) 200
3. Diverter: benzoic acid flakes in gelled acid (31 bbl) 100
4. Repeat steps 1 and 2
5. Displacement: nitrogen 0
0 50 100 150 200
Production increased from 500 STB/D before the job to
Depth (ft)
3700 STB/D after the treatment at a wellhead pressure of
5900 psi (i.e., 500 psi below the wellhead shut-in pressure).
Striking evidence of the acid effect is provided by the pres- Figure 17C-1. Comparison of injection profiles before and
sure record during the treatment, which shows that the well after treatment.
17D. Placement using self-diverting acid
Three similar injector wells were treated with different placement techniques. The average reservoir permeability varied from 4 to
10 md. The formation comprised two zones of different injectivities. Treatment for the first well consisted of pumping 15% HCl through
coiled tubing. The second well was acidized with gelled acid pumped through tubing. For the third well, three stages of a blend of HCl
and suspending agents
were pumped through
coiled tubing. Two stages Zone B
of self-diverting acid were Well 3 — 15% HCl
used to separate the main Zone A Coiled tubing placement with SDA
stages. After
Figure 17D-1 compares Before
the injectivities of the three Zone B
Well 2 — Gelled 15% HCl
wells before and after Bullhead
Zone A
acidizing. Zone A was
treated successfully only
in the third well with the Well 1 — 15% HCl
use of self-diverting acid. Zone B Coiled tubing placement
Zone A
Figure 17D-1. Comparison
of acidizing results of three 0 500 1000 1500 2000 2500
placement techniques. (BWPD)

17-5. Treatment design 17-5.3. Additives

Additives must be added to the different stages to pro-
17-5.1. Candidate selection tect tubulars and ensure successful treatment of the
As explained in Chapter 13, candidate recognition formation. This topic is fully covered in Chapter 15.
utilizing a systems analysis approach is the first step For oil wells, adding a mutual solvent to the pre-
for the design of carbonate acidizing. Whereas sand- flush or acid stages helps to water wet the formation
stone acidizing is usually limited to damage removal, and provides good contact of the treating fluids with
carbonate acidizing is typically oriented toward res- the rock surface.
ervoir stimulation. Wells exhibiting a slightly nega- Acid corrosion inhibitors and inhibitor aids are
tive skin effect prior to the treatment are usually still required to protect tubulars. Organic acids are easier
considered good candidates. to inhibit than HCl. Stable emulsions are also rela-
Because of the ability of HCl to create channels, it tively easy to inhibit. However, in most corrosion
ensures excellent communication with the reservoir. tests, emulsion breakage occurs, and the corrosion
In higher permeability formations, acidizing can also rate is similar to or greater than that of the nonemul-
be used as an alternative to dense perforating. Field sified acid.
experience shows that cased completions with small The addition of surfactant and demulsifiers may
perforation densities (as low as one perforation every also be necessary to lower the interfacial tension
3 or 5 ft) can exhibit a negative skin effect after between treating fluids and the reservoir fluid and to
stimulation with HCl. prevent emulsions. Finally, antisludging agents, scale
inhibitors and iron control agents can be used to pre-
vent specific problems. When mixing additives, the
17-5.2. Pumping schedule compatibility of the different components with each
The second step in the design consists of choosing other and with the downhole conditions must be
the right acid formulation, depending on the damage checked thoroughly.
and formation characteristics. Examples of pumping
schedules are given in the sidebars to this chapter.
17-5.4. Placement
Generally, carbonate treatments consist of alternating
stages of the main fluid and diverter. A solvent pre- Proper placement of acid over the whole pay zone is
flush can be used ahead of the main fluid to clean up required for successful treatment. In thick formations
the formation and increase its receptivity to acid. An or multilayer reservoirs with different values of per-
overflush of brine or seawater can be used to displace meability or damage severity, acid tends to penetrate
the acid into the formation and ensure complete the more permeable zones and create high-injectivity
spending away from the wellbore. N2 can be used for streaks that prevent injection into the whole interval.
displacement at the end of the job or added to the Five main diversion techniques can be used to im-
treatment fluids to assist the flowback of spent acid prove fluid placement in carbonate acidizing: packers,
that may contain insoluble material or high-viscosity ball sealers, particulate diverters, foam diversion and
gel residuals remaining in the formation. Generally, self-diverting acid. The first four methods are dis-
N2 is recommended for low-pressure wells with a cussed in Chapter 19. The use of self-diverting acid
pressure gradient below 0.46 psi/ft. has increased recently (see Section 17-4.6). In large
The pumping rate is limited by the fracturing pres- intervals (e.g., horizontal wells) some of these tech-
sure. In tight formations, the rate must be sufficiently niques can be combined with the use of coiled tubing
high to prevent compact dissolution near the well- (Thomas and Milne, 1995).
bore if plain HCl is used. In naturally fractured for-
mations, good results have been obtained with high
flow rates. Such pump rates and high pressures do 17-6. Conclusions
not correspond to usual matrix conditions and are
Limestone and dolomite formations can easily be
likely to mechanically enlarge natural fractures and
stimulated with acid formulations. Unlike sandstone
increase their conductivity during the treatment.
acidizing, the goal of carbonate acidizing is usually

to bypass the damage rather than dissolve it. HCl stone acidizing, proper placement of the acid over
is typically used for carbonate acidizing. In case of the whole interval is necessary for successful treat-
incompatibility with the formation or the completion ment and usually requires employing placement
(i.e., risk of corrosion), other formulations are used, techniques. Foam diversion and self-diverting acid
such as emulsions or organic acids. Suspending are two methods that are increasingly used, with
agents or solvents can also be used if required by good results.
the type of damage.
The high reactivity of acid with limestones and
high-temperature dolomites results in the creation of Acknowledgments
wormholes, which considerably increase the apparent
permeability around the wellbore. When wormholes Some of the material in this chapter was previously
extend beyond the damaged zone or connect with authored by G. Daccord and published in the second
natural fissures in the formation, a negative skin edition of this volume. The authors also thank G.
effect is obtained. Daccord, C. Fredd, R. Marcinew, A. Saxon and R.
An engineering approach should be adopted to Thomas for reviewing this chapter.
design effective carbonate treatments. As in sand-


◆
Sandstone Acidizing
Harry O. McLeod, Conoco, USA

William David Norman, Schlumberger Dowell
18-1. Introduction species (e.g., scales such as calcium carbonates as

well as some clay species such as zeolites, illites,
Sandstone matrix acidizing is distinguished from car- kaolinites and smectites).
bonate acidizing in that it involves the dissolution of Various well operations can result in formation dam-
damage that is blocking or bridging the pore throats age (see Chapter 14). For example, drilling mud and
in the formation matrix, thus ideally recovering the completion fluid usually penetrate sandstone forma-
original reservoir permeability. Carbonate acidizing tions. This invasion of filtrate can introduce an entirely
dissolves the formation minerals around the damage, different chemical environment, which the acid treat-
creating new permeability. The mineral acids required ment must address. Additional formation damage may
to dissolve the damage are usually highly reactive occur during perforating, gravel packing, and normal
with the numerous formation minerals. The resulting production or injection operations. Acid dissolves a
chemical complexes can become insoluble in the variety of damaging materials along with most forma-
environment created and can precipitate, yielding tion minerals. An understanding of the chemistry is
gelatinous or solid particles. Because the formation basic to the selection of the acid type and concentration.
and the damage can have complicated crystalline This chapter includes the reaction chemistry of the
structures that can yield a variety of reaction prod- primary solvent used in sandstone acidizing, hydro-
ucts, sandstone acidizing success requires a signifi- fluoric acid (HF). Acid systems that contain mixtures
cantly better understanding of chemistry than does of hydrofluoric and hydrochloric acid (HCl) are com-
carbonate acidizing. As discussed in Chapter 13, monly called mud acids because they were first used
75% of well-engineered sandstone acid treatments to remove mud damage.
should be successful, resulting in significant produc-
tion enhancement.
The descriptor “sandstone” is derived from the 18-2. Treating fluids
geologic classification of rocks with a high quartz sil-
ica content. Besides the obvious quartz component, 18-2.1. Hydrochloric acid chemistry
they contain other minerals such as aluminosilicates,
metallic oxides, sulfates, chlorides, carbonates and HCl reactions are discussed in Chapter 17, and details
noncrystalline (amorphous) siliceous material. The of the reaction and by-products are omitted in this
minerals deposited in the original sediment are called chapter except for how they relate to sandstone min-
detrital species. Most have a high degree of associ- erals. The compatibility of the HF blends used in this
ated water. As fluids are produced through the matrix process is twofold; these mixtures must meet both
of the rock, the drag forces can move some of these compatibility standards for the formation mineralogy
minerals, clogging the pore throats. and dissolution of the damage mineralogy. HF mix-
Connate water in a sandstone contains many of the tures are preceded by HCl to avoid precipitation of
dissolved native mineral species. This is due to equi- the slightly soluble and insoluble reaction products
librium and partial pressures of gaseous solvents of HF with certain chemical species. The chemistry
(such as carbon dioxide [CO2]) and the presence of of HCl with carbonate minerals is discussed in a pre-
other ionic species. As fluids are produced, the asso- vious chapter, so the focus here is on the chemistry
ciated pressure drop can disturb this equilibrium and of the HF systems. Although the chemistry of the
the normal ionic content of the formation brines, reaction of HCl with carbonate or calcite is simple,
resulting in precipitation and possible pore-throat the chemistry of the reaction of HF and siliceous
restriction. This type of diagenesis yields authigenic minerals is complex.

Some complex reactions that occur with certain with zeolites is to recognize the presence of these
siliceous minerals have only recently been included minerals before a treatment is performed. The use of
in the reactions reported in mineralogy breakdowns. an organic acid as one of two preflush stages and fol-
These reactions involve HCl and the mineral family lowing the preflushes with a low-concentration HF
known as zeolites. Zeolite minerals are crystalline, but mixture that conforms with the remaining minerals
hydrated with active, porous channels in the crystalline in the formation has proved to be highly effective in
lattice. Zeolites are known in other industries as “mol- restoring permeability and removing damage. All flu-
ecular sieves” because their porosity allows the chem- ids that are injected should have an organic acid
ical extraction and filtering of selective materials. included to maintain a low-pH environment. Some
Zeolite minerals occur in nature as a by-product of operators have found the use of an all-organic-acid
volcanic activity and precipitate from water that is system followed by an organic acid–HF formulation
rich in silica. It is theorized that as zeolites are to be effective in high-temperature environments.
exposed to progressively higher pressures and temper-
atures they metamorphose from extremely loose,
hydrated crystalline structures to more dense and 18-2.2.Chemistry of hydrofluoric
compact structures. The results of this process have acid systems
different mineral names. The hierarchy of their struc- HF is the only common, inexpensive mineral acid
ture and crystalline nature is provided in Table 18-1. able to dissolve siliceous minerals. For any acid sys-
Because these minerals are precipitates, they are tem to be capable of damage removal, it should con-
always authigenic and located in pore spaces. tain HF in some form. The most common formulation
Zeolite minerals are sensitive to HCl and strong is simply ammonium bifluoride dissolved in HCl;
mineral acids. Several core studies have shown that another is by diluting concentrated HCl-HF formula-
the use of HCl alone causes significant damage, tions. The HCl:HF ratio is varied to accommodate the
whereas weak organic acid reduces the damage. The solubility of the dissolved mineral species present in
problem is that the weak organic acid does not neces- the formation. This can be augmented by both pre-
sarily remove the damaging mineralogy to restore per- flush and overflush acid formulations. Several poten-
meability. The solution to the problems associated tial precipitates can be addressed simply by the use
of appropriate HCl:HF ratios in the formulations.
Table 18-1. The zeolite family. Numerous mineral species react with HF, and they
all generate aluminum silica fluoride complexes
Mineral Description
(Table 18-2).
Stilbite Hydrous calcium aluminum silicate
Dissolves in contact with HCl; no gelatin formed • Reactions of hydrofluoric acid with formation
Occurs in shallow environments and may occur minerals
inside tubulars in silica-rich connate water
formations with high pressure drops
Details of HF reactions with formation minerals
have, for more than 60 years, been known and stud-
Heulandite Hydrous calcium/sodium/potassium aluminum ied. As early as 1965, it was quantified that 1000 gal
silicate
Dissolves in contact with HCl; no gelatin formed
of 2% HF can dissolve as much as 350 lbm of clay
Occurs in shallow environments (Smith et al., 1965).
An HCl preflush is always injected in sandstones
Chabazite Hydrous calcium/sodium/potassium aluminum
silicate
prior to the HF. This is done to avoid the possible
Dissolves in contact with HCl; no gelatin formed precipitation of insoluble or slightly soluble reaction
Occurs in medium-depth environments products. Typically, the insoluble species are calcium
fluoride (CaF2), which forms on reaction of HF with
Natrolite Hydrous sodium/potassium aluminum silicate
calcium carbonate (CaCO3), or sodium or potassium
Dissolves in contact with HCl; gelatin formed
hexafluosilicates (M2SiF6, where M = Na or K),
Occurs in deeper environments
which result from the reaction of cations in forma-
Analcime Hydrous sodium aluminum silicate tion brines with solubilized species. The dissolution
Dissolves in contact with HCl; gelatin formed of calcium carbonate or magnesium carbonate by
Occurs in deeper environments
reaction with HCl is discussed in detail in Chapter 17.
18-2 Sandstone Acidizing

Table 18-2. Chemical composition of typical sandstone minerals.
Classification Mineral Chemical Composition

Quartz SiO2
Feldspar Microcline KAlSi3O8

Orthoclase KAlSi3O8
Albite NaAlSi3O8
Plagioclase (Na,Ca)Al(Si,Al)Si2O8
Mica Biotite K(Mg,Fe2+)3(Al,Fe3+)Si3O10(OH)2

Muscovite KAl2(AlSi3)O10(OH)2
Chlorite (Mg,Fe2+,Fe3+)AlSi3O10(OH)8
Clay Kaolinite Al2Si2O5(OH)4

Illite (H3,O,K)y (Al4 ⋅ Fe4 ⋅ Mg4 ⋅ Mg6)(Si8 – y ⋅ Aly )O20(OH)4
Smectite (Ca0.5Na)0.7(Al,Mg,Fe)4(Si,Al)8O20(OH)4 ⋅ nH2O
Chlorite (Mg,Fe2+,Fe3+)AlSi3O10(OH)8
Carbonate Calcite CaCO3

Dolomite CaMg(CO3)2
Ankerite Ca(Fe,Mg,Mn)(CO3)2
Siderite FeCO3
Sulfate Gypsum CaSO4 ⋅ 2H2O

Anhydrite CaSO4
Chloride Halite NaCl
Metallic oxide Iron oxides FeO, Fe2O3, Fe3O4
• Stoichiometric equations The intermediate silicon complex, SiF 5–, which is

Reactions of mud acid with the aluminosilicate not stable in aqueous solution, is not considered. The
components of sandstones are those of HF; how- first step of silica dissolution consists of the chemi-
ever, HF is a weak acid and, because of the equili- sorption of the fluoride anion at the silica surface
brated reaction, is only slightly dissociated when (Iler, 1979). Kline and Fogler (1981b), on the con-
mixed with HCl: trary, showed that it is the molecular HF rather than
HF + H2O H3O+ + F– the fluoride anion that adsorbs (see Section 18-4).
Gaseous silicon tetrafluoride usually remains dis-
Ka = 10–3.2 at 75°F [25°C] solved in the liquid phase at bottomhole pressure, as
where Ka is the acid equilibrium constant. CO2 does in the case of carbonate acidization, so the
HF can also combine and form complexes, but equilibrium is shifted toward the formation of silicon
this reaction must be taken into account (Fogler et hexafluoride anions and the remaining SiF4 does not
al., 1976) only when the HF concentration is suffi- represent more than 1% of the total dissolved silicon
ciently high (less than 10M) to allow numerous (Labrid, 1971).
collisions to occur between the fluoride species. Silicon hexafluoride anions can be hydrolyzed fur-
This occurs only in the case of ultra mud acid ther into monosilicic acid with the evolution of heat:
(25% HCl–20% HF) formulations: SiF62– + 8H2O Si(OH)4 + 4H3O+ + 6F –
HF + F – HF2– K = 1.2 × 10–27 at 75°F
K = 3.86 at 75°F When the silicon concentration increases in the
The reaction of HF with quartz grains (pure sil- aqueous phase, part of the hexafluorosilicate anions
ica) is expressed in the following two equilibria: are also transformed into the acidic form of fluosili-
SiO2 + 4HF SiF4 + 2H2O cic acid according to the reaction
SiF4 + 2F –
SiF62– SiF62– + 2H3O+ H2SiF6 + 2H2O
K = 6.7 × 10 at 75°F
–4

This transformation is usually limited, because – Kaolinite clay

fluosilicic acid is a strong acid. Aluminosilicate Al4Si4O10(OH)8 + 4(n + m) HF
minerals generally have complex chemical compo- + (28 – 4(n + m))H3O+
sitions, such as those listed in Table 18-2. Their
4AlFn(3 – n )+ + 4SiFm(4 – m )– + (46 – 4(n + m))H2O
overall dissolution reactions thus involve many
simple equilibria. – Sodic or potassic feldspar
The disintegration of aluminosilicate minerals by MAlSi3O8 + (n + 3m)HF + (16 – n – 3m)H3O+
HF can be considered stoichiometric as a first step; M+ + AlFn(3 – n )+ + 3SiFm(4 – m )–
i.e., the Al:Si ratio is the same in the solution as in + (24 – n – 3m)H2O
the mineral. Silicon is solubilized by the same pro- where 0 ≤ n < 6 and m = 4 or 6.
cess mentioned for quartz, whereas aluminum is
involved in several fluorinated complexes:
AlFn(3 – n )+ AlFn(4 – n )+ + F– 18-3. Solubility of by-products
where 0 ≤ n ≤ 6.
When minerals are dissolved by HF, numerous by-
The prominent form of aluminum complexed
products can form. Some potential precipitates are
varies as a function of the free fluoride ion concen-
listed in Table 18-3. In many cases, the increase in the
tration: the average ratio of fluorine to aluminum
liquid-phase pH value resulting from acid mixture
decreases as the dissolution reaction progresses
spending constitutes the driving force for precipitate
(fewer fluoride anions are available), as shown
formation; therefore, precipitation can be predicted
in Fig. 18-1 (Labrid, 1971).
from consideration of the sole liquid phase. The extent
of precipitation should always be limited. If this is not
AIF5 possible, the potential precipitation zone that would
AIF4
cause a decrease in permeability should be diluted and
AIF concentration (%)
100 AIF3
AIF2
AIF displaced from the wellbore (Walsh et al., 1982).
Should precipitation occur, most of the calcium and
50 sodium complexes that precipitate in the field can be
redissolved by using boric acid. This is not true, how-
ever, for potassium and some of the magnesium com-
plexes. The very low solubility of potassium complexes
10–4 10–3 10–2 has been shown both in the laboratory and in the field.
Fluoride ion concentration (g ion ⁄ L) Colloidal silica precipitation cannot be avoided, as
it results partly from the greater affinity of fluorine for
Figure 18-1. Domains of existence of aluminum-fluorine
complexes (Labrid, 1971).
Table 18-3. Solubility in water at room temperature
of HF reaction by-products.
The dissolution reaction of all aluminosilicate
Secondary Product Solubility (g/100 cm3)
minerals in sandstones follows the previous equa-
tions for the basic lattice atoms (Si, Al) concerned. Orthosilicic acid (H4SiO4) 0.015
Other metallic ions, such as Na, K, Mg, Ca and Fe, Calcium fluoride (CaF2) 0.0016
which are in the minerals constituting the rock as Sodium fluosilicate (Na2SiF6) 0.65
substitution cations in the lattice or as exchangeable Sodium fluoaluminate (Na3AlF6) Slightly soluble
(adsorbed) cations, come into solution as free ions Potassium fluosilicate (K2SiF6) 0.12
during the reaction. In the case of iron, fluorinated
Ammonium fluosilicate ((NH4)2SiF6) 18.6
complexes (FeFz(3 – z )+, where 1 < z < 3) also are
Calcium fluosilicate (CaSiF6) Slightly soluble
formed through reactions similar to those for alu-
minum. Thus, different global reactions can be Aluminum fluoride (AlF3) 0.559
written as a function of the considered mineral: Aluminum hydroxide (Al(OH)3) Insoluble
Ferrous sulfide (FeS) 0.00062

aluminum than for silicon. This process accelerates 18-3.2. Alkali fluosilicates and
the hydrolysis of SiF6 because the released F– anions fluoaluminates
are further involved in aluminum complexes and more
monosilicic acid (Si(OH)4) is generated. Certain The aluminum or silicon fluorine complexes can react
authors (Labrid, 1971; Shaughnessy and Kunze, 1981; with alkali ions released in the solution from highly
Walsh et al., 1982) have emphasized the highly dam- substituted clays or alkali feldspars as soon as their
aging potential of the precipitation of colloidal silica concentration becomes sufficiently high to form insol-
in a porous medium; however, this damaging action uble alkali fluosilicates and, probably, fluoaluminates:
has never been demonstrated clearly and satisfactorily. 2Na+ + SiF62– Na2SiF6 Ks = 4.2 × 10–5
On the contrary, other authors (Crowe, 1986)
2K+ + SiF62– K2SiF6 Ks = 2 × 10–8
showed that such “precipitation” is actually the result
of a topochemical reaction (exchange of fluoride from 3Na+ + AlF3 + 3F– Na3AlF6 Ks = 8.7 × 10–18
the hexafluorosilicate anion occurs with aluminum on
2K+ + AlF4– + F– K2AlF5 Ks = 7.8 × 10–10
the surface of the silt and clay), and it does not induce
damage. where Ks is the solubility constant.
Precipitation begins earlier in the dissolution process Alkali fluosilicate precipitation is favored by a high
at higher temperatures (within 10 min at 200°F [95°C]) level of HF. Fluosilicate precipitates, which form from
because of the increased thermal agitation. It also the attack of mud acid on alkali feldspars or clays, are
occurs more quickly in montmorillonite-type clays than well crystallized and very damaging (Bertaux, 1989).
in kaolinite clays because of the different initial Al:Si These damaging precipitates also form when the vol-
ratios in these minerals (molar ratio of 1 for kaolinite ume of preflush is insufficient and HF contacts forma-
and less than 0.5, depending on the substitution extent, tion brine containing alkali ions.
for montmorillonite). Finally, aluminum can be totally
removed from clays, with a correlated silica deposition
at the surface (topochemical reaction). 18-3.3. Aluminum fluoride and hydroxide
Aluminum fluoride (AlF3) or aluminum hydroxide
(Al(OH)3) in the gibbsite form can precipitate upon
18-3.1. Calcium fluoride spending of the acid. The precipitation of AlF3 can be
Some carbonates may remain after preflushing, either reduced by maintaining a high proportion of HCl to
because of the initial amount of carbonate cementing HF (Walsh et al., 1982). These precipitates form
material in the sandstone or as a result of the carbon- according to the reactions
ates’ initial protective siliceous coating. Also, slightly Al3+ + 3F– AlF3
soluble, fine crystalline CaF2 readily forms when cal-
cite contacts HF. This can lead to substantial damage: Al3+ + 3OH– Al(OH)3 Ks = 10–32.5
CaCO3 + 2HF CaF2 + H2O + CO2
Where this precipitate has formed but has not com- 18-3.4. Ferric complexes
pletely blocked the porosity of the formation, it may
partially redissolve when HF is near complete spend- This mechanism of forming iron fluorine complexes
ing toward the end of the job. At this time, the con- applies only to relatively clean sandstones. In the pres-
centration of fluoride anions in solution is so low that ence of clays, the dissolved aluminum ions have a
aluminum is hardly complexed and appears mainly as greater affinity for fluorine than iron does. Therefore,
free Al3+ ions (Labrid, 1971). These aluminum ions the iron fluorine complexes do not form and iron
are then able to extract fluorine from the CaF2 precipi- hydroxide still precipitates at pH levels greater than 2.2.
tates, as they did for silicofluorides, and partly redis- The nature of the precipitate (crystalline or amor-
solve the CaF2 according to the reaction phous) varies as a function of the anions present
(Smith et al., 1969). Ferric hydroxide can be strongly
3CaF2 + 2Al3+ 3Ca2+ + 2AlF2+ bound to the quartz surface by electrostatic interac-
This reaction may be followed by subsequent equilibria tions because its point of isoelectric charge is above
between the different aluminum and fluorine complexes. a pH value of 7. In the presence of excess calcite, the

dissolved CO2 can also lead to the precipitation of In the case of a feldspar with the overall formula
insoluble ferric carbonates (siderite or ankerite). Na 0.72K 0.08Ca 0.2Al 1.2Si 2.8O8, the following
Chapter 15 provides additional information about iron expression has been determined (at 75°F under
control and solutions for problems. 275-kPa pressure) as the reaction rate (Fogler et
al., 1976):
18-4. Kinetics: factors affecting ( )

rfeldspar = 1.3 × 10 −9 1 + 0.4[H + ] [HF ] (18-2)
reaction rates in mol feldspar/cm2/s.

An elemental mechanism proposed to explain the
This section summarizes qualitatively the results
previous variation involves the adsorption of protons
described in detail in Chapter 16. Because theoretical
on the surface that weakens the siloxane bondings,
aspects are covered in Chapter 16, only the practical
which is followed by the reaction of HF molecules
implications are discussed here.
that creates unstable silicon-fluorine bonds at the sur-
Kinetically controlled reactions (surface reaction limit-
face, according to the scheme
ed) are effective during the acidization process of sand-
stones, and factors affecting reaction rates are discussed H
to complete previous thermodynamic considerations. 
–X–O–Si– +H+ → –X–O…Si+– + HF →
–X + FSi– + H2O
18-4.1. Hydrofluoric acid concentration
where X = Al or Si.
Dissolution reaction rates are proportional to the HF This is the acid (proton) catalysis mechanism pro-
concentration (Fogler et al., 1976; Kline and Fogler, posed by Kline (1980) for feldspar.
1981b) for most sandstone minerals, except smectite. The dissolution reaction is a first-order reaction
This explains why formations with low competence with respect to the HF concentration for most alumi-
(i.e., weak cementation, potentially mobile fine parti- nosilicate minerals. Nevertheless, dissolution kinetics
cles) should be treated with a reduced-strength mud is better represented by a Langmuir-Hinshelwood–
acid (1.5% HF) to avoid crumbling, especially at bot- type law in the case of sodium montmorillonite (Kline
tomhole temperatures greater than 200°F. Fluoboric and Fogler, 1981):
acid performs similarly because of the low concentra-
tion of HF present at any time. KK ads [HF ]
R= , (18-3)
1 + K ads [HF ]
18-4.2. Hydrochloric acid concentration where Kads is the equilibrium constant of the exother-
mic adsorption of HF molecules at surface-reactive
Dissolution reaction rates generally increase in a more sites. This adsorption constant is independent of the
acidic medium because the leaching of constitutive total acidity, whereas K increases with proton concen-
surface cations involves their replacement by protons, tration (acid catalysis). Kads is especially high for a
but the dependence on HCl concentration is not mineral with a high cation exchange capacity (CEC),
straightforward (Gdanski and Peavy, 1986). The prin- such as sodium montmorillonite. For most other clay
cipal role of HCl is to prevent secondary precipitation minerals, the value of this adsorption constant is
by maintaining a low pH value. The other main effect small. Therefore, when 1 >> Kads[HF] the expression
of HCl is to catalyze the attack of sandstone minerals can be simplified to the experimentally determined
by HF. The mechanism and degree of catalysis depend first-order kinetics law. An elemental mechanism dif-
on the type of mineral, as shown in the following. ferent from that mentioned for feldspars can be pro-
For example, the reaction rate measured at 95°F posed to explain the kinetics and to take into account
[35°C] for pure quartz has the following expression solely the HF adsorption:
(Fogler et al., 1976):
( )
rquartz = 9.2 × 10 −9 1 + 0.8[H + ] [HF ] (18-1)
in mol quartz/cm2/s.

F of each mineral to the total accessible surface area is

 considered, great discrepancies between the reaction
HO OH HO H OH HO OH rates of pure phases can be predicted and observed
       (Table 18-4).
–X–O–X– +HF –X–O–X– –X –F + HO–X–
Table 18-4. Relative surface areas
where X = Al or Si. of sandstone minerals.
Mineral Surface Area

18-4.3. Temperature Quartz <0.1 cm2/g
The dissolution of minerals is a thermally activated phe- Feldspar Few m2/g
nomenon; thus, the rates increase greatly as a function of Kaolinite 15–30 m2/g
temperature (approximately multiplied by 2 for quartz Illite 113 m2/g
for a 25°C increment), and the penetration depths of live
Smectite 82 m2/g
acid diminish accordingly. In the case of quartz, the acti-
vation energy is about 5.2 kcal/mol, and in the case of
the previous feldspar, it is about 8 kcal/mol (Fogler et Clays react much faster than feldspars, which react
al., 1976). much faster than quartz, especially in the presence of high
Figure 18-2 shows the variation of the reaction rate proton (H+) concentrations. Thus, most of the quartz
of mud acid with vitreous silica (more reactive than matrix (about 95%) can be considered inert with respect
quartz) as a function of both HF concentration and to the dissolution reaction, and the mineralogical nature
temperature (Smith and Hendrickson, 1965). Alumi- of the accessible rock components determines the overall
num and iron solubilities also increase slightly with reaction rate. This situation also emphasizes the necessity
a rise in temperature. of HCl preflushes and excess HCl in the HCl-HF mixture.
Calcite reacts at the highest rate of all the minerals that
7 can be present in a sandstone, leading to HF microchan-
neling, but the mechanism of attack is not comparable
Reaction rate (lbm ⁄ ft2/s × 10–6)
6
because protons coming from either HCl or HF can pro-
5°
F
5 17 voke the dissolution.
0°F
4 15
°F
125
3 18-4.5. Pressure
2 An increase in pressure speeds up the overall dissolution
1
Reaction time, 60 min reaction slightly, because dissolved silicon tetrafluoride
can be transformed partially into an acidic species
0 (H2SiF6) and can quickly initiate further reactions. For
0 1 2 3 4 5 6 7 8 9 10
quartz, a 24% rise in the reaction rate was noticed
HF (%)
between the two extreme conditions (Smith et al., 1965).
Figure 18-2. Reaction rate of HCl-HF on silicate glass In a radial injection situation, the mineral pore-
(Smith and Hendrickson, 1965). space texture that determines flow partition around the
wellbore (most live acid flows through the large
pores) is also a relevant parameter; clay clasts can
be bypassed by the acid flow (Williams, 1975).
18-4.4. Mineralogical composition and
accessible surface area
18-5. Hydrofluoric acid reaction
The relatively high total specific surface area of sand-
stone rocks is the primary parameter determining mud modeling
acid spending because of the heterogeneous nature of The parameters that affect the reaction rate of HF on
the dissolution reaction. However, if the contribution sandstone minerals are incorporated in a model that

predicts the evolution of formation parameters when the volume of clays (or fast-reacting dissolvable mate-
acid is injected. rial) and of the acid concentration.
In terms of surface reaction rates, sandstones are The acid concentration (or spending) front can be
typically considered a two-component system: modeled similarly. The thickness of the front depends
on the Damköhler number Da, which is a function of
• slow-reacting pseudocomponent, forming the crys-
the reaction rate and the acid velocity. These simula-
talline quartz fraction
tions show why HF does not penetrate deeply into the
• fast-reacting pseudocomponent, comprising all reservoir before spending unless unrealistically large
other species (e.g., clays, feldspars and poorly volumes are used. (These large volumes would almost
crystallized silica). dissolve everything around the wellbore and thus
For both pseudocomponents, the overall kinetics, leave the reacted formation totally unconsolidated.)
which includes the diffusion of HF-reactant species to
the surface, surface reactions and the diffusion of react-
ed products back to the bulk solution, is governed by 18-6. Other acidizing formulations
the surface reactions because they are the slowest step.
Problems related to the use of mud acid to remove
Therefore, matrix acidizing of sandstones with HF is
damage in sandstone formations include the following:
called surface reaction limited.
This is the major difference from the matrix acidiz- • Rapid spending provides only a short penetration,
ing of limestone, where the process is diffusion con- especially at high temperatures (maximum depth
trolled. In sandstones, the increase in permeability about 12 in.).
results from damage removal and is correlated with • Fines, composed of either mostly quartz or mostly
a small increase in rock porosity. Quartz reacts very clay minerals, can be generated during the acid
slowly with HF; reactions with most aluminosilicates reaction and can migrate with the fluid flow. The
provoke a rapid spending of the acid. A pseudo- destabilization of fines can lead to a quick produc-
stationary state reflects the much faster variation in tion decline after treatment. Gravel-packed gas
species concentration (chemical modifications) than wells can exhibit a 50% productivity reduction.
the one within the rock porosity (resulting in physical • The high dissolving power of mud acid destroys
modification). The HF progresses and homogeneously rock integrity at the formation face.
dissolves every pore and never forms conductive
channels or wormholes. The flow is stable, and sharp New sandstone acidizing systems are designed to
fronts are formed in response to the dissolution of dif- alleviate these shortcomings.
ferent mineral species as acid injection progresses
radially (McCune et al., 1975).
Several authors have tried to model this process. 18-6.1. Fluoboric acid
Taha et al. (1986) used the reaction model developed Fluoboric acid is recommended by Thomas and
by Fogler and various coworkers (see particularly Crowe (1981) as an alternative to mud acids. It does
Hekim et al., 1982). Such a simplified, two-pseudo- not contain large amounts of HF at any given time
component model and macroscopic description can be and thus has a lower reactivity. However, it generates
used because Fogler et al. (1976) showed that the order more HF, as HF is consumed, by its own hydrolysis.
of reaction of HF with each pseudocomponent is equal Therefore, its total dissolving power is comparable to
to unity relative to the concentrations of HF and of the a 2% mud acid solution. Fluoboric acid solutions are
pseudocomponent. The flow is considered stable. used as a preflush before treating formations sensitive
In this model the mineral dissolution fronts can be to mud acid; this avoids fines destabilization and sub-
computed and the concentration of remaining clays sequent pore clogging. They are also used as a sole
(or fast-reacting materials) can be calculated. Then, treatment to remove damage in a sandstone matrix
the permeability increase can be estimated from the with carbonate cement or in fissures that contain many
change in porosity (or amount of material dissolved). clay particles. Another use is as an overflush after a
The velocity of the mineral dissolution front depends mud acid treatment that has removed near-wellbore
on the acid capacity number Ac, which is a function of damage (up to 0.5 ft) to allow easier penetration of the

fluoboric acid solution (a few feet). Fluoboric acid is the following text, BF3OH– hydrolysis is neglected at
recommended when the sandstone contains potassic the usual acid concentrations.
minerals to avoid damaging precipitates and in the The hydrolysis reaction kinetics of fluoborate ions
case of fines migration owing to its fines stabilization is affected by
properties.
• concentration of the fluoborate ions
In the field, fluoboric acid is easily prepared by
mixing boric acid (H3BO3), ammonium bifluoride • medium acidity, which has a catalyzing effect (reac-
(NH4F ⋅ HF) and HCl. Ammonium bifluoride, an tion is proportional to the proton concentration)
acidic salt of HF, reacts first with HCl to generate HF: • temperature, through the usual activation energy
effect.
NH4F ⋅ HF + HCl → 2HF + NH4Cl
Thus, the reaction rate, assuming the reverse reac-
Tetrafluoboric acid is formed as a reaction product
tion is negligible, can be expressed after Kunze and
of boric acid with HF, according to
Shaughnessy (1983) as
H3BO3 + 3HF → HBF3OH + 2H2O (quick reaction)
d[BF4 − ]
HBF3OH + HF HBF4 + H2O (slow reaction) r= = K[H 3O + ][BF4 − ] , (18-4)
dt
Hydroxyfluoboric acid (HBF3OH) probably does not where
exist in aqueous solutions unless it is in equilibrium
K = 1.44 × 1017 exp −
with fluoboric acid (Wamser, 1948). The preceding 26,183 
slow reaction is of an order equal to unity with respect  1.987T 
to both HF and HBF3OH. For this reaction, equilibrium
is attained at room temperature after nearly 40 min for in (mol/L)–1min–1 and T is the temperature in kelvin.
a resulting lM HBF4 solution. Because the equilibrium Thus, the reaction rate is increased 300-fold when the
constant at 75°F is K = 2.3 × 10–3 (Wamser, 1948), mixture is heated from 75° to 150°F [25° to 65°C] and
about 6% (molar) HBF4 is converted into HBF3OH at is increased 12,000-fold when heated from 75° to 220°F
equilibrium for a lM HBF4 solution. These equilibrium [105°C]. Because the hydrolysis reaction kinetics is
considerations mean that at any given time and place not affected by clays, fluoboric acid can be considered
there is only between 0.1% and 0.2% (weight) of free a retarded acid in normal use (i.e., less than 200°F). In
HF at ambient temperature and 212°F [100°C], respec- the presence of excess bentonite, pure 0.1M fluoboric
tively. acid is spent within 30 min at 150°F (Kunze and
Fluoboric acid is a strong acid with strength compa- Shaughnessy, 1983). In a slurry test, which has an infi-
rable to that of HCl (Maya, 1977); thus, the following nite surface area (1 L of acid with 20 g of bentonite or
reaction occurs in solution: 1600 m2 of surface area, which is equivalent to several
football fields of exposed area), the reaction rate is a
HBF4 + H2O → H3O+ + BF4– function of the rate of hydrolysis. However, in the
In the following text, reactions are written using BF4– matrix, where there is a finite amount of clay surface,
instead of HBF4. Acid strength diminishes in the fol- the reaction rate is a function of the amount of HF pres-
lowing order: fluoboric, hydroxyfluoboric (the strength ent, which in the case of fluoboric acid is low.
of which can be compared to that of trichloroacetic The reaction of fluoboric acid in sandstones
acid; Maya, 1977) and boric acid (K H3BO 3 = 9.2 at involves at the same time the hydrolysis reaction of
75°F). fluoboric acid, standard reactions of the generated HF
The dissolving power of fluoboric acid results from with minerals and additional slow reactions related to
the generation of HF through its hydrolysis: the fluoborate ions in the liquid phase. As expected,
the dissolution reaction of clays with fluoboric acid
BF4– + H2O BF3OH– + HF is a first-order reaction with respect to the fluoborate
The BF3OH anions can be further hydrolyzed suc- concentration, similar to the relation of the reaction
cessively into BF2(OH)2–, BF(OH)3– and H3BO3 with of mud acid to the HF concentration.
correlated HF formation, but these reactions must be The spending rate of fluoboric acid on glass slides
taken into account only when the BF3OH– concentra- at 150°F is one-tenth that of a mud acid with the same
tion is lower than 3 × 10–3 at 75°F (Wamser, 1948). In total HF content (Thomas and Crowe, 1981). Amor-

phous silica reacts faster than quartz, which limits the forms after treatment with mud acid. This shows that
destruction of cores near the injection face during flow nondamaging by-products are formed by fluoboric acid,
tests with fluoboric acid. Significantly less destruction whereas formation plugging by alkali fluosilicates can
is noted than during mud acid flow testing. The occur with mud acid. This is another advantage of using
reduced destruction with fluoboric acid results in 30% fluoboric acid in some “acid-sensitive” formations.
to 50% higher compressive strengths than observed During the injection period, fluoboric acid behaves
for mud acid. like a weak HF solution, but one in which the HF is
The unique advantage of fluoboric acid is that it pro- constantly replenished. The small amount of fluoride
vides efficient stabilization of clays and fines through ions available at any time limits the danger of precipi-
reactions related to borate and fluoborate ions. Swelling tating aluminum species. Only the first acidity is used
clays are desensitized by fluoboric acid, and there is a during this step. Hydroxytrifluoboric acid (HBF3OH)
large decrease in the CEC (e.g., a 93% decrease after buffers the solution and prevents other undesirable
18 hr in fluoboric acid at 150°F for a Wyoming ben- precipitations.
tonite was observed by Thomas and Crowe). During shut-in, HBF4 and HBF3OH continue to
After a fluoboric acid treatment, migrating clays react, but at a slow pace because the hydrolysis is
and other fines stabilize as a result of the rock’s expo- minimal. The liberated HF reacts further with mineral
sure to acid. This is why a long shut-in time is recom- species. It also reacts by topochemical reactions, in
mended in fluoboric acid treatments. During injection, which the aluminum from the undissolved clay struc-
while the acid spends normally, cores treated only ture is put into solution by forming one of the fluoalu-
with fluoboric acid exhibit a normal increase in per- minate complex ions (depending on F–), and the
meability. However, no long-term stabilization occurs surface of the mineral is therefore enriched in silicon
after treatment because only a portion of the clay was and boron. An amorphous coating of silica and boro-
dissolved; the remainder did not have time to stabi- silicate glass is then formed over the remaining silicate
lize. Additional shut-in time allows this stabilization. and fine silica grains, welding them to the framework
When treated by fluoboric acid, montmorillonite pro- and thus preventing their migration.
gressively decreases in aluminum content and then pro- This effect is clearly seen in Figs. 18-3 and 18-4,
gressively incorporates boron atoms; silicon precipitates where the same pore, containing two different clays
from the solution. Cores originally containing 30% sili- (kaolinite and illite), is shown before and after reac-
coaluminates at 150°F attain maximum static solubili- tion with a fluoboric acid solution. The quartz is
ties after only 24 hr in the presence of lM HBF4 (4 hr barely etched, whereas the high-surface-area, fast-
for mud acid), whereas the maximum increase in per- reacting illite is completely dissolved. The kaolinite
meability is obtained after only 4 hr under dynamic platelets are about half-dissolved, and an amorphous
conditions (Thomas and Crowe, 1981). These results material is coating the undissolved kaolinite, welding
prove the dissociated effects of mineral dissolution by them together and to the underlying quartz grain.
the generated HF (essentially kinetically controlled) and
of particle stabilization resulting from the slow complex
dissolution/reprecipitation mechanisms (toward thermo- 18-6.2. Sequential mud acid
dynamic equilibrium) during the shut-in period. The sequential mud acid system involves the in-situ
Examination under a scanning electron microscope generation of HF, occurring from the alternate injec-
(SEM) shows that the original kaolinite clay platelets— tion of HF and ammonium fluoride (Hall et al., 1981).
pure aluminosilicates—that are not dissolved by fluo- The reactions of HF are thought by some to take place
boric acid appear welded together and to the quartz at the rock surface by adsorption followed by ion
grains. A type of chemical fusion of any fines seems to exchange, but the yield of this heterogeneous process
take place slowly onto the silica surface. The formation seems highly doubtful for several reasons:
of borosilicate “glass” has been assumed to account for
this reaction. • If HF were generated through such a process, it
Bertaux (1989) observed that in silicoaluminates con- would be a small quantity, hardly enough to etch
taining potassium, such as illite, potassium fluoborate the surface of the clay material.
forms after treatment with fluoboric acid as a nondam- • Because this process is based on the CEC of the
aging coating on the clay; potassium hexafluosilicate clays, migrating kaolinite would hardly be touched.

Figure 18-4. SEM photograph of the same pore after an 8%

Figure 18-3. SEM photograph of kaolinite (K) and illite (I) HBF4 treatment.
clays in a pore (Q = quartz), before attack.
• This process supposes the initial adsorption of the This procedure is tantamount to adding dissolution
hydronium (H3O+) ions on the clay surface, fol- reaction products to the mixture before the reactions
lowed by their exchange with NH4+, to generate HF occur (i.e., the injection of spent acid). In theory this
in situ. Exchange and replacement of H3O+ by NH4+ should slow the rates. However, the retardation of clay
depends on many parameters and cannot be ascer- dissolution has not been proved experimentally because
tained. Therefore, even the generation of HF is of the prime importance of the high surface area on
dubious. clay reactivity, which is much more important than a
slight depletion of acid at high temperatures. The risk
of early precipitation of damaging products, such as
18-6.3. Alcoholic mud acid AlF3 or fluoaluminates, is probably increased by the use
of an acid that already contains aluminum ions before
Alcoholic mud acid formulations are a mixture of reaction. Flow tests have shown a smaller effective live
mud acid and isopropanol or methanol (up to 50%). acid penetration than in the case of mud acid. In addi-
The main application is in low-permeability dry gas tion, field experience has shown that the addition of
zones. Dilution with alcohol lowers the acid-mineral aluminum to the system increases the precipitation of
reaction rate and provides a retarding effect. amorphous aluminosilicate scale. This white material
Cleanup is facilitated; acid surface tension is plugs near-wellbore perforations and gravel packs.
decreased by the alcohols while the vapor pressure
of the mixture is increased, which improves gas perme-
ability by reducing water saturation. 18-6.5. Organic mud acid
Because total acidity speeds mineral dissolution with
18-6.4. Mud acid plus aluminum chloride mud acid, organic mud acid involves replacement of
for retardation the 12% HCl component with organic acids (9%
formic acid, a weak acid that only partially dissoci-
An acidizing system to retard HF-mineral reactions has ates), mixed with 3% HF, to retard HF spending. This
been proposed in which aluminum chloride (AlCl3) is system is particularly suited for high-temperature
added to mud acid formulations to complex some of wells (200° to 300°F [90° to 150°C]), for which pipe
the fluoride ions in the injected mixture, according to corrosion rates are diminished accordingly. This sys-
the reactions (Gdanski, 1985) tem also reduces the tendency to form sludge.
AlCl3 + 4HF + H2O AlF4– + 3HCl + H3O+
AlF4– + 3H3O+ AlF2+ + 3HF + 3H2O

18-6.6. Self-generating mud acid systems depth of damage. The true degree of retardation depends
on the temperature and pumping time. These esters are
Self-generating acidizing systems were originally more expensive and more dangerous to handle because
developed by Templeton et al. (1975), and their appli- of their flammability than HCl or inorganic salts. More
cation was widened by Abrams et al. (1983). They precipitates are formed as a result of the poor solubility
involve the hydrolysis of organic esters into the corre- of the organic by-products. The only advantage over
sponding carboxylic acids, followed by the reaction of reduced-strength HF is lower corrosion rates.
these acids with ammonium fluoride to yield HF.
Because the hydrolysis reaction is activated by tem-
perature and the acidity obtained is not as strong as 18-6.7. Buffer-regulated hydrofluoric
with mud acid, a low corrosion rate of tubular goods acid systems
and delayed reaction of the progressively generated
HF are expected. The latter would allow deep penetra- Other high-pH acidizing systems proposed for use
tion of live HF. up to 360°F [180°C] involve the buffering effect of
Depending on the bottomhole temperature, different an organic acid and its ammonium salt, mixed with
organic esters are used: ammonium fluoride, as an HF precursor (Abrams et
al., 1983). To minimize corrosion, the use of the same
• methyl formate between 130° and 180°F [55° and uninhibited buffer without ammonium fluoride as a
80°C] with the reactions preflush has been recommended up to 350°F [175°C].
HCOOCH3 + H2O HCOOH + CH3OH The ammonium salt of the organic acid is generated
from the partial neutralization of the acid with ammo-
HCOOH + NH4F NH4+ + HCOO– + HF nium hydroxide. The proposed buffered systems are
(the latter is the slow, rate-controlling reaction). • formic acid/ammonium formate with pH = 3.5 to 4
• ammonium salt of monochloroacetic acid between • acetic acid/ammonium acetate and citric acid/
180°F and 215°F [102°C]: ammonium citrate with pH = 4.5 to 5.
NH4+ + ClCH2COO– + H2O To extend the application to higher temperatures
HOCH2COOH + NH4+ + Cl– (up to 550°F [290°C]), an excess of ammonium salt
• methyl acetate between 190° and 280°F [90° and is formed by using a higher ratio of ammonium
140°C]. hydroxide to organic acid. Because the kinetics of clay
dissolution increases with the fluoride ion concentra-
The reagent choice is intended to limit at 30% tion, more ammonium fluoride is added to compensate
(maximum) the generation of HF during pumping of for the pH increase (Scheuerman, 1988). Successful in-
the mixture in the tubing; thus, a minimum of 40 min depth stimulation has been observed with this system
of spending time seems necessary. However, field only for bottomhole temperatures lower than 129°F
tests of these systems have not been conclusive. Many [54°C]. In most cases using this system, many damag-
precipitates form in these low acidic systems, such as ing precipitates are noticed (e.g., fluosilicates, fluoalu-
ralstonite (NH4MgAlF6) and other fluoaluminates minate usually involving ammonium), the formation
(silicates) upon spending of these mixtures on clays; of which is related to the weak acidity in the near-
thus, the use of complexing agents or acids, such as wellbore area. These systems suffer from the same
citric acid, is suggested. Furthermore, formation sensi- drawbacks as the self-generated mud acid system.
tivity after treatment has not been tested, and handling
problems arise from the high flammability of methyl
formate.
Overall, these systems have many drawbacks. Based
18-7. Damage removal mechanisms
on the hydrolysis of various organic esters, they are Selection of a chemical as a treatment fluid for any
temperature activated. Unlike fluoboric acid, which application depends on the contaminants plugging the
generates new HF only upon spending, no equilibrium formation. HCl does not dissolve pipe dope, paraffin
is reached. This means that more HF is generated as the or asphaltenes. These solids or plugging agents are
temperature increases, and the ester can eventually be organic in nature, and their treatment requires an
completely hydrolyzed long before reaching the final effective organic solvent (usually an aromatic solvent

such as toluene, xylene or orthonitrotoluene). Because ing fluids to predict how the spent acid will react as it
different plugging solids require a variety of solvents penetrates the formation. Potential incompatibilities can
for their removal, there is no universal solvent for be prevented by proper log and core evaluation. Because
wellbore damage. The proper evaluation of damage the secondary reactions can be just as damaging, defining
and treatment design are illustrated by Clementz et al. the potential problems generated by long-term exposure
(1982) for the successful removal of bacterial damage should also be evaluated.
in water injection wells. Solvent or acid should never Two key formation characteristics for fluid selection
be pumped into a well until the probable causes of are mineralogy and permeability. Defining formation
damage and the best chemical to remove the damage mineralogy helps to confirm the types of acid systems
have been defined. and acid concentrations to use. Defining formation
Compatibility with formation fluids and mineralogy permeability provides the information required to esti-
is extremely important in sandstone acidizing. Thus, mate the matrix injection rate and the maximum
determination of the precipitation potential of mud bottomhole pressure allowed before hydraulically
acid mixtures requires close scrutiny of the mineral- fracturing the formation.
ogy and connate water present. Acid reaction products Pore pressure, temperature and the mechanical condi-
are not necessarily soluble in the spent solution or in tion of the formation are influential in the design. High-
certain ionic environments. pore-pressure formations fracture at much lower pressure
Formation damage is fully discussed in Chapter 14. differentials than depleted formations. Depleted forma-
tions have a lower fracture pressure than that originally
observed. Temperature significantly affects the selected
18-7.1. Formation response to acid fluid’s reaction rate with different mineral types. Acid
Incompatibilities may occur even if the damage is concentrations are usually lower for higher temperatures.
identified, an appropriate removal fluid system is The mechanical integrity of the formation biases the fluid
available and the probable response of the formation selection in that the acid concentrations are usually
fluids and minerals to the acid and spent acid solution reduced in less consolidated formations.
has been determined. These incompatibilities can
result in solid or gelatinous precipitates, which can
plug pores and offset the improvement the acid was
18-7.3. Formation brine compatibility
intended to create. Results can range from no harmful Compatibility with formation brines must be considered
effects and complete cleanup of the damage to less when treating with mud acid. Mud acid mixtures can
than optimum improvement or plugging of the forma- form CaF2 (a solid) when excess Ca2+ ions are present.
tion with acid-generated precipitates. Similar solid materials are also created with K+ and Na+
When detailed petrographic core analyses are avail- ions. The use of clear brines as completion and work-
able, geochemical simulators can be used to estimate over fluids has increased the necessity of checking the
potential problems. This type of simulation requires formation waters for compatibility. This brine usage has
detailed definition of the chemistry of the treating also increased the necessity of ensuring that sufficient
fluid, formation damage and matrix mineralogy. The compatible preflush is used to dilute and remove these
release of fines and undefined spent-acid precipitates ionic species prior to injection of the mud acid system.
still have the potential to damage the formation and Several available methods have been tested. The salinity
are not identified by core testing. of the connate brine is in equilibrium with the native
minerals and their CEC. When possible, the salinity of
the preflush and overflush fluids should closely approx-
18-7.2. Formation properties imate that of the connate brine. Historically, the use of
Damage prevention and dealing with formation response low-salinity brines has rarely presented catastrophic
before acidizing are the goals of proper design. Although problems when used in conjunction with acid treatments.
it may be easy to dissolve formation damage, success is Several additives have been demonstrated to posi-
dependent on dissolving this material without damaging tively affect the formation’s sensitivity to changes in
the formation. This is possible, yet it is paramount to salinity. Other species in connate water have equal,
define the chemistry of the formation minerals and treat- if not more, influence on the success of the treatments.

Each of these species has specialty chemical additives completed with oil-base muds, presoaks with an aro-
that address them individually. While there is documen- matic solvent and producing back before acidizing are
tation on the benefits of certain types of additives helpful. Solvent formulations and surfactant solutions
(Gidley, 1971; Hall, 1975), other authors have reported are available as a pretreatment to clean up oil-base
damage caused by a similar system in multiphase-flow mud filtrates and restore the formation to a water-wet
environments (Muecke, 1979; Davies et al., 1988). condition.
Shaughnessy and Kline (1983) showed the difficulties Gidley (1985) reported that the use of CO2 as a pre-
with high bicarbonate ion content in formation waters. flush to acid treatments has many benefits, including
The use of HCl was not sufficient to keep the well from reducing the volumes of acid required to generate suc-
redamaging itself quickly. They used an ingenious cessful production increases. This type of preflush has
treatment with a form of ethylenediamenetetraacetic worked well in core studies to enhance crude oil dis-
acid (EDTA) to both remove the calcium carbonate placement and improve mobility.
scale that had damaged productivity and prevent recur-
rence of the scale for long periods of time.
High sulfate ion contents (>1000 ppm) exist in some 18-7.5. Formation mineral compatibility
formation waters. Spending HCl on calcium carbonates with fluid systems
generates a high concentration of calcium ions that will An analysis of the formation minerals is important for
precipitate calcium sulfate when the spent acid mixes designing the HCl preflush, mud acid treatment and
with formation water. This can be prevented by pre- overflush in sandstone formations. Basic questions
flushing the formation water away from the wellbore. that must be answered are listed here.
In sandstone formations, water containing ammonium
chloride (NH4Cl) should be used as a preflush fluid. 1. How much of the formation will dissolve in HCl?
Where a high HCl solubility exists (20% or more),
mud acid should not be used. This statement is based
18-7.4. Crude oil compatibility on the assumption that HCl-soluble compounds are
Another serious problem with formation fluids is the carbonate-base minerals. These minerals are the
reaction of crude oil with acid. Removal of the residual common cementing material of sandstone forma-
hydrocarbon phase improves the effectiveness of aque- tions. Dissolution of this cementing material releases
ous acid systems. Some oils, particularly black heavy particles that can decrease the permeability. In addi-
oils (less than 30°API gravity), react with acid to form tion, precipitants exist as small discrete particles that
either damaging sludge (precipitated asphaltenes) or a cannot be produced back through the perforations
stable emulsion. Moore et al. (1965) reported this prob- and out of the well. The use of mud acid in sand-
lem and gave the treatment to prevent it. Sometimes stones with a high carbonate content produces
sludge preventers and emulsion breakers cannot prevent numerous precipitates.
the formation of stable emulsions. Houchin and Hudson Calcium carbonate, magnesium carbonate and
(1986) discussed similar problems with organic deposits. iron compounds are soluble in HCl. Even feldspars
Recent work shows how dissolved iron creates more and chlorite clay are slightly soluble (Gdanski and
stable sludges and emulsions with these crude oils. Some Peavy, 1986). Recent investigations of HCl involve-
“difficult” crude oils require a hydrocarbon solvent buffer ment in the HF reaction with clays show that the
between the crude oil and the acid that is mutually com- HCl is consumed on the clay surfaces, and this
patible with both the crude oil and the acid. The buffer should also be accounted for in the preflush volumes
reduces contact between the acid and the problem oil and in the HCl:HF ratio of the main fluid stage
and prevents or reduces the problems with sludge and (Gdanski and Peavy, 1986). Zeolite minerals can
emulsion. Using this technique in one Wyoming oil produce gelatinous precipitates when exposed to
field increased the success rate from 25% to 75%. HCl. This can be avoided by the use of organic acid
Asphaltene particles can precipitate during produc- mixtures, as discussed later in this chapter. Sufficient
tion as a result of a pressure drop. Solvents can be volumes of HCl must be injected ahead of the mud
used to loosen and partially or completely disperse acid to dissolve enough of the HCl-soluble materials
them. This action helps the acid do a better job of before the mud acid or spent mud acid reaches them.
dissolving acid-soluble solids. When a well has been

2. How much of the formation will dissolve in mud acid stage is in the near-wellbore area. If the pre-
acid? Will acid reaction by-products precipitate? cipitates are diluted and flushed, the likelihood of
The volume of mud acid used depends on the con- permanent damage is reduced. The by-products can
centration of the acid and the amount of damage. be flushed away and sometimes even stabilize for-
Optimizing this volume can be done only by detail- mation fines in the process. If the well is then
ing the damage in a valid numerical simulator returned to flow quickly, some of the precipitate
(Perthuis et al., 1989). The HCl:HF ratio and con- may be produced back. “Quickly” refers to non-
centration are selected to prevent or reduce the for- producing time, not the rate at which flowback is
mation of damaging precipitates (Table 18-5). accomplished. The quick return of fluids can help
Some minerals automatically precipitate fluoride improve cleanup of the formation after acid treat-
compounds when high concentrations of HF are ment, regardless of the flow rate. If an inadequate
used, particularly 6% HF. Even 3% HF will precip- amount of HCl preflush is used in formations with
itate potassium fluosilicate when mud acid reacts 5% to 15% carbonate, residual carbonate near the
with potassium feldspar. HF-dissolved sodium wellbore reacts with spent HF (fluosilicic acid or
feldspars do not usually precipitate sodium silicate AlF3), and voluminous precipitates form. The
with 3% or less HF. hydrated precipitates occupy a much larger volume
When HF is used in a formation containing clay, than that of the original clays and carbonate dis-
feldspar and micas, hydrous silica always precipi- solved.
tates. Hydrated silica has been reported in a sticky, 3. Will iron be a problem?
gelatinous form that if left stagnant can attach to Where a lot of iron-rich minerals are in the formation,
the mineral surfaces. However, Crowe’s (1986) dissolved iron can precipitate in the formation. It is
work on sandstone cores demonstrates that hydrated well known that ferric iron precipitates as acid spends
silica does not precipitate as a sticky, gelatinous to a pH of 2 to 4. The precipitation of iron hydroxide,
mass. The reaction between the spent mud acid and where concentrations as high as 10,000-ppm iron are
formation fines is a topochemical reaction, with present in solution, can be prevented by adequate
hydrated silica deposited on the surface of the fines. treatment with a sequestering agent such as nitrilotri-
It is important to design the overflush to dilute acetic acid (NTA), EDTA, citric acid or combinations
and displace the hydrous silica at least 3 to 5 ft of acetic and citric acid (Shaughnessy and Kunze,
away from the wellbore to reduce the effect of the 1981; McLeod et al., 1983; Paccaloni, 1979a, 1979b)
damage. To avoid silica-creating damage, it is (see Chapter 15). Crude oil with a high asphaltene
important to limit any static time while the mud content should be tested for sensitivity to different iron
concentrations. Sludge and ridged-film emulsions are
Table 18-5. Acid use guidelines for sandstone common problems for these crude oils.
acidizing (McLeod, 1984). Damage with iron hydroxides can be compounded
by the high iron concentration that comes off the sur-
Condition or Mineralogy Acid Strength (blend)
face of the tubing during acid injection (De Ghetto,
HCl solubility > 20% HCl only 1982). Injecting acid through new tubing can be
High permeability (>50 md) highly damaging in this respect (Fogler and Crain,
1980; Lybarger and Gates, 1978a, 1978b). Newly
High quartz (>80%), low clay (<5%) 12% HCl–3% HF†
manufactured tubing has a crust of mill scale, or
High feldspar (>20%) 13.5% HCl–1.5% HF†
magnetite, which is a form of ferric and/or ferrous
High clay (>10%) 10% HCl–1% HF‡ oxide. The mill scale is dissolved and loosened by
High iron/chlorite clay (>15%) 10% acetic acid–1% HF§ the acidic fluid, and in the early stages, partially
Low permeability (≤10 md) spent, iron-rich weak acid is injected. Particles of
Clay (<10%) 6% HCl–1% HF
mill scale can then be injected into the perforations
and trapped there. Injected acid will continue to dis-
Clay (>10%) 6% HCl–0.5% HF
solve the mill scale, creating ferric chloride that
†
‡
Preflush with 15% HCl enters the formation. If the ferric chloride combines
Preflush with 10% HCl
§
Preflush with 10% acetic acid with iron leached out of iron-rich chlorite clay or

other iron compounds, a large amount of iron 18-7.6. Acid type and concentration
hydroxide precipitates is possible, which can severely
damage the formation. This aggravated iron damage Permeability and mineralogy determine the compati-
can be prevented by pickling (cleaning) new tubing ble concentration of HCl or acetic acid in the preflush
to remove mill scale and then circulating the pick- stage and HF and HCl in the mud acid stage. Con-
ling acid back out of the well, as discussed later. centration recommendations are provided in Table
18-6 for preflush fluids and Table 18-7 for mud acid
4. Do the sidewall core samples contain drilling mud- fluids. The previously presented acid use guidelines in
cake? Table 18-5 were published in 1984 (McLeod, 1984).
Testing results from samples of sidewall cores with Lower mud acid concentrations were first recom-
excessive mudcake should be reviewed closely and mended in 1970 by Farley et al. (1970) to prevent
compared to the log response and other data unconsolidation in California sandstones. U.S. West
sources such as a produced water sample analysis. Coast sandstones are generally rich in potassium feld-
High concentrations of drilling mud solids (e.g., spars. Holcomb (1975) published work on the first
barite, smectite, mica, bentonite or illite minerals) successful acid stimulation of the Morrow formation
should not be present in clean, high-porosity sand- in West Texas–New Mexico with weak acid (6% HCl–
stone formations. The solubility of the samples in 1.0% HF and 3% HCl–0.5% HF). Lybarger and Gates
mud acid mixtures may be exaggerated. (1978b) subsequently developed the slow-rate, low-
Table 18-6. Fluid selection guidelines for preflush fluids.
Mineralogy Permeability
>100 md 20 to 100 md <20 md
<10% silt and <10% clay 15% HCl 10% HCl 7.5% HCl
>10% silt and >10% clay 10% HCl 7.5% HCl 5% HCl
>10% silt and <10% clay 10% HCl 7.5% HCl 5% HCl
<10% silt and >10% clay 10% HCl 7.5% HCl 5% HCl
Note: Selection guidelines for all temperatures

For 4% to 6% chlorite/glauconite, use <20-md guidelines with 5% acetic acid.
For >6% to 8% chlorite/glauconite, do not use HCl; use 10% acetic acid preflush to mud acid plus 5% acetic acid.
For >8% chlorite/glauconite, do not use HCl; use 10% acetic acid and organic mud acid.
For <2% zeolite, use 5% acetic acid in all fluids containing HCl and preflush with 10% acetic acid.
For >2% to 5% zeolite, do not use HCl preflush; use 10% acetic acid preflush and overflush to mud acid containing 10% acetic acid.
For >5% zeolite, do not use HCl in any system; use 10% acetic acid preflush and overflush to organic acid prepared from 10% citric acid/HF.
Table 18-7. Fluid selection guidelines for mud acid fluids.
Mineralogy Permeability
>100 md 20 to 100 md <20 md
<10% silt and <10% clay 12% HCl–3% HF 8% HCl–2% HF 6% HCl–1.5% HF

>10% silt and >10% clay 13.5% HCl–1.5% HF 9% HCl–1% HF 4.5% HCl–0.5% HF
>10% silt and <10% clay 12% HCl–2% HF 9% HCl–1.5% HF 6% HCl–1% HF
<10% silt and >10% clay 12% HCl–2% HF 9% HCl–1.5% HF 6% HCl–1% HF
Notes: Selection guidelines for all temperatures

For 4% to 6% chlorite/glauconite, use <20-md guidelines with 5% acetic acid.
For >6% to 8% chlorite/glauconite, use 10% acetic acid preflush to mud acid plus 5% acetic acid.
For >8% chlorite/glauconite, use 10% acetic acid and organic mud acid.
For <2% zeolite, use 5% acetic acid in all fluids containing HCl.
For >2% to 5% zeolite, use 10% acetic acid preflush and overflush to mud acid containing 10% acetic acid.
For >5% zeolite, use 10% acetic acid preflush and overflush to 10% citric acid/HF.

pressure injection technique, in which they used 7.5% HCl– The early work of Smith et al. (1965) in acidizing
1.5% HF for Gulf Coast sandstones. various cores with different permeabilities shows differ-
The guidelines are based on industry practices and ent responses to mud acid. C. F. Smith (pers. comm.,
the chemistry of sandstone acidizing from limited 1979) found it more difficult to stimulate wells produc-
research studies; however, many case histories have ing from sandstones with permeabilities of 10 to 60 md,
corroborated them with high levels of success. which are much lower than the usual Berea sandstone
From 1975 through 1980, poor success in acidizing permeability (100 to 300 md) in mud acid experiments.
several formations such as the Frio and Wilcox in Smith attributed much of the difficulty to the release of
Texas led to the concern that spent acid generated fines by the acid.
damaging precipitates. Quick, qualitative laboratory Long-core tests performed by R. D. Gdanski (pers.
bench tests confirmed that precipitates occur depend- comm., 1985) in low-permeability sandstone at high
ing on the solubility of the acid reaction products. temperatures demonstrate increased permeability with
These same observations were first pointed out by mud acid in the first two 4-in. cores in series and
Smith and Hendrickson (1965), in particular the prob- decreased permeability in the third 4-in. core in a total
lem with sodium fluosilicate. Labrid (1971) discussed core length of 16 in. Gdanski and Peavy (1986) also
the precipitation of hydrous silica, which caused some discussed the depletion of the preflush HCl in sand-
plugging in cores. This damage was later demonstrat- stone acidizing by ion exchange of H+ with K+ or Na+
ed by Shaughnessy and Kunze (1981) by leaving ions on the formation clay minerals. This gives new
spent acid in the core for several hours, a condition insights into potential problems with sandstones rich
that occurs in an actual acid job. This allows the slow in clay minerals with high CECs (smectite and illite).
reaction rate between the spent mud acid and the clay Simon et al. (1979) showed that HCl attacks chlo-
minerals and feldspars (aluminosilicates) to produce rite clay, extracting the iron and magnesium and leav-
hydrous silica that decreases the permeability. ing an amorphous aluminosilicate residue. J. M.
Crowe (1986) showed that there was little or no Kullman (pers. comm., 1988) observed plugging
plugging during the injection of spent acid (fluosilicic problems with these residues as well as with rim coat-
acid) in a Berea core. This reassuring result matches ings of chlorite liberated by HCl in core flow tests.
the behavior seen during acid injection; however, Chlorite is prevalent in the Morrow formation in the
plugging conditions are worse during the static condi- same areas where Holcomb (1975) worked and could
tions of shut-in examined by Shaughnessy and Kunze be the reason why weaker acids worked better in that
(1981). Crowe’s work does not address shut-in condi- environment (i.e., they were easier on the chlorite).
tions or conditions of inadequate preflush with HCl. Thus, weaker acids are recommended for use in sand-
Walsh et al. (1982) presented theoretical work on the stones with significant chlorite content and acetic acid
equilibrium of spent acid and showed that plugging is recommended to dissolve the carbonate and not
precipitates are possible with various acid concentra- attack chlorite ahead of the mud acid.
tions and mineral compositions in sandstones. A common misunderstanding about the recom-
Research by Bertaux (1986) addresses reprecipitation mended acid concentrations is that they are not
and plugging problems in acidizing sandstones contain- absolute. The guidelines are a conservative approach
ing potassium feldspars. The solubility of potassium to avoid problems with spent acid precipitates when
fluosilicate is less than one-half of the solubility of sodi- no previous experience exists in acidizing a particular
um fluosilicate, which is why lower mud acid concen- formation. Significant deviation from these guidelines
trations are recommended in the presence of potassic should not be necessary. Unless evaluated experience
feldspars such as orthoclase or microcline (KAlSi3O8). exists, the guidelines are the most reliable source of
The amount of potassium in the mineral orthoclase information. Also, acid flow tests with cores are reli-
(potassium feldspar) is enough that the solubility of able if long cores are used and if the spent acid is left
potassium fluosilicate is exceeded at normal reservoir in a portion of the unacidized core for the same period
temperatures (less than 200°F) by dissolving pure of time and at the same temperature that will occur in
orthoclase in regular mud acid (12% HCl–3% HF). the downhole treatment. These tests are expensive and
Bryant and Buller (1990) observed the generation of therefore seldom performed.
fines by the reaction of HCl with feldspars.

18-8. Methods of controlling • Concentration

precipitates Regular mud acid (12% HCl–3% HF) is the normal
concentration to use to remove damage in clean
The methods used to control the precipitates caused quartzose sands. Field experience has shown that
by acidizing are proper acid staging, lower acid con- weaker concentrations (0.5% to 1.5% HF) can be
centrations, correct usage of preflushes and sufficient effective for other sands. Mineral composition from
overflushing, as illustrated in the following guidelines. a laboratory analysis can also dictate when less
than 3% HF should be used. If the combined per-
centage of clay and feldspar is more than 30%,
18-8.1. Preflush 1.5% HF or less should be used. Field experience
Preflush with with some tight sandstones has shown that concen-
trations as low as 0.6% HF may be used (e.g., the
1. 5% to 15% HCl Morrow formation in Texas and New Mexico;
2. acetic acid (see Section 18-3). Holcomb, 1975). If the appropriate concentration
The preflush displaces formation brine away from is in doubt, an acid response test on a typical core
the wellbore to prevent it from mixing with reacted should be performed if a core sample is available.
mud acid and causing a damaging precipitate. If the
formation contains more than 1% to 2% carbonate,
an HCl preflush is necessary to dissolve the carbonate,
18-8.3. Postflush or overflush
prevent the waste of mud acid and prevent formation The overflush is an important part of a successful sand-
of the insoluble precipitate CaF2. stone acidizing treatment. An overflush has several
If completion brines such as seawater, potassium purposes:
chloride (KCl), calcium chloride (CaCl2) or calcium
• to displace nonreacted mud acid into the formation
bromide (CaBr) have been used in the well prior to
acidizing, the brines will mix with the mud acid in the • to displace mud acid reaction products away from
formation. Preflushing the mud acid with HCl or brine the wellbore
containing ammonium chloride to dilute the brines • to remove oil-wet relative permeability problems
and remove them away from the wellbore helps avoid caused by some corrosion inhibitors.
this problem. When overflushing the acid treatment, it is impor-
Preflushes can also be used to displace and isolate tant to remember that miscible fluids are required to
incompatible formation fluids (either brine or crude perform these listed functions. Aqueous-base liquids
oil), as previously discussed. should therefore be considered as the first displacing
and flushing fluid. Another fluid system can then be
used for addressing the other concerns as the condi-
18-8.2. Mud acid volume and concentration
tions dictate. This suggests that multiple fluid types
• Volume should be used as overflush stages for a given set of
Gidley (1985) reported that for the most successful circumstances.
mud acid treatment, more than 125 gal/ft of mud acid Typical overflushes for mud acid treatments are
is required. Less may be used where only shallow • water containing 3% to 8% ammonium chloride
damage exists around new perforations (e.g., 25 to
• weak acid (3% to 10% HCl)
75 gal/ft is used to remove mud damage or in a spear-
head treatment as an aid to perforation breakdown • diesel oil (oil wells only and only following a water
prior to hydraulic fracturing). or weak acid overflush)
When the damage is quantified, a simulator can • nitrogen (gas wells only and only following a water
be used to optimize the volumes of mud acid mix- or weak acid overflush).
tures to be used. Simulators can be used to aid the
Studies of displacement fronts indicate that the
modification of volumes if several job stages are
reactivity and fluid character of the overflush have a
used (see Chapter 14).
major influence on the volume required to displace the

spent mud acid. For most overflush fluids (weak HCl estimation and calculation of critical parameters is
and water containing ammonium chloride), volumes required. Pressures, rates and volumes must conform
less than twice the mud acid stage should be consid- to the constraints of the mechanical conditions of the
ered inappropriate. The volume of overflush should well equipment and the available space for surface
never provide less than 3 ft of radial penetration. This and pumping equipment, along with logistical time
means that for most situations, the overflush should be constraints. The following discussion includes the dif-
at least 200 gal/ft of perforations to push all the spent ferent types of acid sequences, how and why attempts
acid past the critical flow radius of 3 to 5 ft. A large are made to retard the acid reaction rate, potential con-
overflush is necessary to prevent the near-wellbore tamination from various sources and the resultant
precipitation of amorphous silica, which occurs after damaging precipitation. The basic quality assurance
spent HF contacts the clay in the formation. At forma- and quality control (QA/QC) checks and the design
tion temperatures of 200°F or higher, amorphous sil- of treatments from both a formation compatibility and
ica precipitation occurs while the mud acid is being operational standpoint are included.
pumped into the formation. The precipitate is some- An acid design technique based on the work of
what mobile at first but may set up as a gel after flow Williams et al. (1979) for mud acid injection is in the
stops. If it is kept moving by overflushing with water SPE Monograph Acidizing Fundamentals. Although
containing ammonium chloride or weak acid, it is the technique is based on studies of one sandstone, it
diluted and dispersed far enough away from the well- does show the important effects of temperature and
bore to where it has a less harmful influence. injection rate on live mud acid penetration. Well illus-
Recent experience indicates the advantage of includ- trated is the small depth of invasion of mud acid in
ing HCl or acetic acid in the first part of the overflush sandstone, particularly when formation temperatures
to maintain a low-pH environment for the displaced are greater than 200°F. Live mud acid usually pene-
spent mud acid stage. This supports the original recom- trates only about 6 to 12 in. into the sandstone before
mendations of Smith and Hendrickson (1965). As the spending. This work was extended by Hill et al. (1977),
hydrogen ions adsorb on nonreacted clay deeper in the who incorporated the effect of specific mineralogy and
formation, the pH rises unless it is replaced by fresh added the reaction kinetics of HF to the slower but
acid in the first part of the overflush. Although the most finite quartz reaction rate. They also discussed the dif-
economic overflush of a mud acid treatment is water ferent reactivities of clay minerals and the importance
containing 3% to 8% ammonium chloride with 10% of their morphology in the pore network. McElhiney
ethylene glycol monobutyl ether (EGMBE) and a et al. (1979) also reviewed the progress in methods of
polyquarternary amine clay stabilizer, it does not predicting live mud acid penetration and permeability
address the pH problem without acetic acid addition. increases in sandstone. These are worthwhile develop-
Also, certain chemicals can be added to acids to pre- ments, but a simple guideline of wellbore contact time
vent or reduce the precipitation of some compounds offers a practical solution to determining acid volumes
(e.g., iron complexing agents, sulfate scale inhibitors to remove near-wellbore damage.
and antisludge agents).
An example of the role of reservoir mineralogy was
presented by Boyer and Wu (1983) in evaluating acid 18-9.1. Selection of fluid sequence stages
treatments in the Kuparuk River formation in Alaska. The damage type dictates the sequence of acid sys-
Their results indicate that fluoboric acid significantly tems used for each treatment. The preflushes, main
reduces the amount of hydrated silica formed in com- stage and overflush should be matched to the type of
parison with conventional HCl-HF systems. damage. Diversion should be matched to formation
characteristics and the type of treating fluid. Diversion
guidelines are provided in Chapter 19. Each type of
18-9. Acid treatment design diversion technique is addressed as it pertains to sand-
considerations stone treatments in this section. The sequence of fluids
that compose an acid treatment can be the key to mak-
Once a well is determined to be a candidate for a ing a treatment successful.
matrix acid treatment, the design should account for
many different issues. A systematic approach to the

18-9.2. Typical sandstone acid job stages preparatory flush to help remove and dilute acid-incom-
patible species (e.g., potassium or calcium). An exam-
A preflush stage should be used ahead of the HCl ple of a preflush sequence is preceding the HCl portion
especially when high sulfate ion or high bicarbonate of the preflush with a large quantity of brine containing
ion concentrations exist in the formation connate ammonium chloride followed by a hydrocarbon-base
water or seawater or when CaCl2, KCl or CaBr com- surfactant mixture. The purpose of the brine preflush is
pletion fluids have been used and calcium carbonate to dilute the incompatible species to soluble levels. The
is a formation mineral. HCl dissolution of the calcite hydrocarbon mixture has the same purpose as men-
generates high calcium ion concentrations that mix tioned previously.
with the incompatible formation water and generate The next consideration for preflushes is compatibil-
scale (calcium sulfate or calcium carbonate). ity with formation fluids. Certain crude oils have a
high sensitivity to acidic mixtures. These situations
may require dilution with hydrocarbons or other iso-
18-9.3. Tubing pickle
lating or buffering fluid systems (e.g., foams). Further
One of the first items to be addressed when matrix compatibility consideration should be given to the
treatments are considered should always be a tubing iron content of the initial injection fluids that contact
pickle (cleaning). This one step can have a significant the crude or condensate, because even low iron con-
impact on the success of treatments. Tubulars, regard- centrations can cause sludge formation. Displacement
less of how new, have scale, rust and other debris that of the fluids away from the near-wellbore region
result from handling, installation and production and reduces the potential of problems that can reduce pro-
that can be loosened by the solvents and acid injected duction success and limit or halt the injection process.
into the well. The pickling process may be multiple HCl preflushes in sandstone acidizing are extremely
staged and may involve expensive solvent packages. important. Their function is to remove as much of the
Typically, a small treatment containing solvent and calcareous material as possible prior to injection of the
acid stages will greatly improve, if not completely mud acid. Strength and volume guidelines are based
eliminate, the problems associated with tubular debris. on the criteria set in work by Labrid (1971), Fogler et
The pickling process should be included in the proce- al. (1976), Kline (1980), Kline and Fogler (1981) and
dure and time allotted for job execution. The purpose Walsh et al. (1982). Their theoretical work was further
of the pickling process is to investigated and confirmed by field work by Gidley
• remove rust, iron oxides and scale (1985), McLeod (1984), Thomas and Crowe (1981)
and others. Table 18-6 provides selection guidelines
• dissolve oily films and pipe dope that could plug for the appropriate strength of the HCl preflush. The
the downhole equipment and perforations table is based on the solubility of the formation in HCl
• limit the amount of iron that gets into the formation and the requirement of minimizing the remaining car-
and contacts the crude oil. bonate or calcite prior to introducing the mud acid.
Figure 18-5 summarizes Walsh et al.’s (1982) work
on the selection of HCl-HF formulations based on the
18-9.4. Preflushes amount of calcite remaining after the preflush. The
The sequence of fluids in sandstone treatments is figure illustrates the importance of HCl preflushes.
dependent largely on the damage type or types. The use The HCl preflush step should never be neglected
of multiple-stage preflushes should functionally address when using mud acid mixtures. A few systems con-
the different types of damage and thereby prepare the taining HF can be injected without an HCl preflush,
surfaces for the main treatment fluids. Hydrocarbon sol- but these are systems with extremely low HF concen-
vents are used to remove oil films and paraffin deposits trations, such as fluoboric acid. These systems can be
so the aqueous acid systems can contact the mineral used without an HCl preflush because the HF concen-
surfaces. These types of preflushes affect treatment suc- tration in fluoboric acid is low enough not to present a
cess and should not be overlooked or demoted in precipitation potential.
importance. Acid-compatible brines (e.g., brine contain-
ing ammonium chloride) can be used as an excellent

16
14
Maximum weight % of HF 12
in acid formulation Ideal case
10
3% calcite
8
0% calcite
6
4
6% calcite
2
0
0 10 20 30 40
Weight % of HCI in acid formulation
Figure 18-5. HCl-HF treatment fluid selection based on AlF3 or CaF2 precipitation (Walsh et al., 1982).
18-9.5. Main fluid stage lapse. Table 18-7 is derived from this guideline on the
basis of further laboratory testing and extensive field
The HCl-HF mixture used in each treatment should experience.
conform to the guidelines in Table 18-7. Work by
Walsh et al. (1982) demonstrates that low HF concen-
trations should be used to avoid the precipitation of 18-9.6. Overflush stage
AlF3 or CaF2 if the remaining calcite cannot be quanti-
fied. Their work also suggests that 12% HCl–3% HF The purpose of the overflush is twofold. First, it should
can be used even in low-calcite environments without displace the main fluid stage more than 3 to 4 ft away
a precipitation problem. Some significant problems from the wellbore, which is the critical matrix area for
that may occur in high-clay-content formations radial flow. Second, the portion of the main stage that
include compromised formation integrity and exces- is not displaced should be diluted. Both of these factors
sive fines generation. These conditions can be the help to eliminate damage in the near-wellbore area
result of too high HF concentrations. The volumes caused by the precipitation potential of the spent main
should be determined using a field-validated simulator fluid stage. Overflush fluids must be chosen carefully to
to sensitize the severity of the damage. Gidley (1985) avoid creating damage during the treatment flowback.
reported that the percentage of acidizing successes Overflush systems should meet the following crite-
increases as the volume of mud acid increases for gas ria. The portion of the overflush immediately follow-
wells, whereas a maximum of 100 to 125 gal/ft of per- ing the main fluid stage should be aqueous based,
forations is required to maximize success for oil wells. have a low pH value and have dilution potential for
This study did not take into account the preflush used the spent mud acid. Smith et al. (1965) recommended
or the quantity of overflush. If diversion is maximized an HCl overflush to maintain a low-pH environment
and the damage is known or perceived to be shallow, and match the fluid density of the previous stages. The
then smaller quantities per foot can be used. The acid remainder of the overflush should be miscible and
strength is important, because precipitation potential compatible with the previous stages. The total mini-
and formation matrix collapse are problems that can mum overflush volume must completely displace the
be irreversible. Table 18-5 provides the original guide- main fluid stage at least 4 ft away from the wellbore.
lines for HCl-HF mixtures to obtain the appropriate Any anisotropy of the formation permeability can
HCl:HF ratio to avoid precipitation and formation col- warrant doubling or tripling the overflush volume

if the energy in the reservoir is sufficient to unload the Operational considerations should always be taken
injected fluid. Although not previously reported, one into account when designing diversion stage sequences.
of the authors of this chapter has achieved notable The use of oil-soluble resins (OSRs) dictates that the
improvement where larger overflushes were used. method should be slug application. The last stage of
This is especially true for wells where heavy bromide preflush can contain a solvent to help dissolve the OSR
brines are used during the completion phase. material, creating uniform injectivity of the last sequence
throughout the interval. A few exceptions apply to using
certain acid systems. For example, when using fluo-
18-9.7. Diversion techniques boric acid as the overflush to a mud acid treatment for
Common practice in sandstone acidizing is for the silt and clay control, the fluids should be staged as in
diverter stage to be applied as merely another stage. Table 18-8.
This is an excellent way to ensure that the main fluid Other sequences could include brine flushes sepa-
stages are properly isolated by the preflush and over- rating the hydrocarbon preflush from the HCl preflush
flush fluids. Some methods described in other chapters before the main fluid stage; brine or weaker acid
(e.g., ball sealers, rock salt) are not suitable for use in stages could be used to increase the volume of the
sandstone acidizing. The compatibility of the diverting overflush stage.
agent with the live and spent acid species requires
knowledge of the chemicals. Some forms of benzoic
acid solids should not be used because the sodium
18-9.8. Typical sandstone acid job stages
content in some environments causes precipitation. The key to successful staging is to address all damage
Rock salt should never, under any circumstances, be types present and maintain compatibility with forma-
used as a diverter with HF mixtures. Other materials tion fluids and formation mineralogy while minimizing
can be incompatible with the solvents and surfactants the quantities of fluids injected. Table 18-9 provides a
used in the acid systems. listing of typical stage sequences for a sandstone acidiz-
ing treatment.
Table 18-8. Acid treatment sequence and fluid options.
Stage Fluid System
1. Preflush Brine
Hydrocarbons
HCl
2. Main fluid HCl-HF formulation
3. Overflush HCl or NH4Cl
4. Diverter Foam or slug OSR
5. Repeat stages 1–4 as necessary with 1–3 as the last fluid sequence
6. Fluoboric acid With diverter solvent for OSR or foam-weakening agent

(mutual solvent)
7. Fluoboric acid diverter Fluoboric acid–based fluid system, either foamed

or slug OSR
8. Fluoboric acid Fluid left at the perforations

Table 18-9. Typical stage sequence for a sandstone acidizing treatment.
Stage Stage Reason for Stage Information Stage Stage Volume

Number Source Composition
1 Crude oil displacement To prevent oil sludge Acid–crude oil Aromatic solvent To achieve 3-ft radial
formation by the acid sludge test displacement
2 Formation water To prevent scale HCO3 and SO4 Ammonium chloride To achieve 3-ft radial
displacement deposition contents from (NH4Cl) at 3%–8% displacement
formation water depending on the
analysis salinity of the for-
mation water
3 Acetic acid Iron compounds in X-ray-diffraction 3%–10% acetic acid CaCO3 (%) Volume (gal/ft)
formation (pyrite, siderite, (XRD) analysis 0–5 25
hematite), chlorite, clay, 5–10 50
zeolites 10–15 75
15–20 100
4 Hydrochloric acid CaCO3 or other HCl- HCl solubility According to core Calculated on the basis of
soluble minerals test and/or XRD mineralogy: 3%–15% HCl solubility and porosity
analysis HCl (see Table 18-5) or this
schedule:
HCl Solubility Stage
of HF (%) Volume
(gal/ft)
<5 50
5–10 100
10–20 200
5 Hydrofluoric acid To remove clay, other XRD analysis, According to formation 75–100 gal/ft
(not used for carbon- formation fines and SEM analysis, mineralogy:
ates and sandstones mud damage HCl:HF 3%–13.5% HCl
where HCl solubility solubilities with 0.5%–3% HF
> 20%)
6 Overflush To spend acid and flush Always used 3%–8% NH4Cl or One to two volumes of
spent acid away from 3%–5% HCl in all the HCl:HF volume or to
the near-wellbore area wells followed by achieve 5-ft radial
nitrogen (gas wells), displacement
kerosene (oil wells)
or 5% HCl (water
injection wells)
7 Diversion To improve injection Used as OSR for oil or low gas/

throughout the interval required for oil ratio wells, foam for
heterogeneous either oil or gas wells
formation and water-soluble resins
permeability for water injector wells
18-10. Matrix acidizing design 1. Estimate safe injection pressures:

guidelines a. determine present fracturing gradient
b. determine present bottomhole fracturing pressure
Matrix acidizing is the process of injecting acid into
the formation in radial flow below fracturing pressure c. determine allowable safe injection pressure at
to remove damage and restore the permeability to the both the wellbore and at the surface.
original reservoir permeability or higher. More 2. Estimate safe injection rate into the damage-free
detailed procedures are available from McLeod et formation.
al. (1983), who recommended the following steps for 3. Estimate safe injection rate into the damaged formation.
treatment design: 4. Select stages required for fluid compatibility.

5. Calculate volume of each stage required: radius in ft, rw is the wellbore radius in ft, and s is
a. crude oil displacement the skin effect factor. B is the formation volume fac-
tor and has a value of 1 for noncompressible fluids.
b. formation brine displacement
Equation 18-5 is a simple way of estimating the
c. HCl stage or acetic acid stage injection rate. However, Eq. 18-5 does not account
d. mud acid stage for several factors, which are detailed in Chapter 20
e. overflush stage. for accurately modeling the injection rate. Equation
6. Select acid concentrations according to formation 18-5 with zero skin effect and with the estimated
mineralogy. value of the skin effect provides respective values
for the minimum and a maximum pump rate during
the job. These values enable allocating appropriate
equipment for the treatment. True transient injection
18-10.1. Calculations
monitoring can be done in real time on location to
• Fracturing pressure monitor the progress of the job.
Matrix treatments are defined as fluid injection • Friction pressure estimation
occurring below fracturing pressure. If the fluid Accurate fluid friction pressure is a difficult parame-
is injected above fracturing pressure, the acid may ter to obtain. Because the tubular arrangement can
bypass the damage. It is important to perform some be different in each case, a fairly accurate number is
basic calculations to ensure that this pressure is not important. The following limited-range equation has
exceeded, and the exercise also provides the pres- been used with relatively good accuracy for estimat-
sure and rates that may occur. Thorough discussions ing friction pressures for Newtonian fluids at rates
of fracturing pressure and bottomhole injection less than 9 bbl/min:
pressure and how these aspects are derived are
provided in Chapters 3 and 20, respectively. 518ρ0.79 q1.79 µ 0.207
p pipe friction = , (18-6)
An important item to keep in mind with matrix D 4.79
treatments is that fracturing pressure is related to
the pore pressure but is not directly proportional. where ppipe friction is the friction pressure in psi/1000 ft,
As the pore pressure declines, so does the fractur- ρ is the density of the fluid (specific gravity) in g/cm3,
ing pressure. Although this is not a one-to-one q is the pump rate in bbl/min, and D is the diameter of
relationship, it can be important when treating the pipe in in. Coiled tubing friction pressures can also
low-bottomhole-pressure wells. The hydrostatic be calculated using Eq. 18-6.
pressure exerted by the column of fluid in the tubu- • Fluid volumes
lars can be sufficient to fracture the formation. If it is assumed that acid flows through porous
• Injection rates media with a front that is uniform and stable, then
The injection rate can be as significant as the injec- the injection is piston-like and the first fluid in
tion pressure. The maximum injection rate that should be the last fluid out. To calculate fluid vol-
does not fracture the formation can be estimated by ume, the following equation should be sufficient:
[( )
4.917 × 10 −6 kh g f × H − ∆psafe − p ] , (18-5) [
Vp = 7.48 φ(rs2 − rw2 )π , ] (18-7)
qi,max =
 r  where Vp is the pore volume for the distance s in
µB ln e + s
 rw  gal/ft, φ is the fractional porosity, and rs is the dis-
tance it is necessary to penetrate the damaged or
where qi,max is the injection rate in bbl/min, k is the displaced section in ft.
effective permeability of the undamaged formation Mud acid treatments do not dissolve much of the
in md, h is the net thickness in ft, gf is the fracture formation minerals but rather dissolve the materials
gradient in psi/ft, H is the depth in ft, ∆psafe is the clogging the pore throats. This means that significant
safety margin for the pressure in psi (usually 200 changes in the flow distribution of the injected fluids
to 500 psi), p is the reservoir pressure in psi, µ is the occur during the treatment as the pore-plugging
viscosity of the injected fluid in cp, re is the drainage materials are dissolved (see Chapter 19 on diver-

sion). Because the acid does not follow the ideal π(1 − φ) X HCl [rs2 − rw2 ] (18-8)
mode, adjustments to the injection volumes must be VHCl = 7.48 ,
β where
made. Significant changes in the fluid can also occur
in the tubulars, before the fluid reaches the forma- VHCl is the volume of HCl required in gal/ft, XHCl is
tion. The dilution of stage composition and spending the fraction of the bulk rock dissolved by HCl, and
are just some of the complications that must be β is the dissolving coefficient expressed as the
addressed by the designer. The use of smaller tubu- amount of rock dissolved per gallon of acid and is
lars, such as coiled tubing, during acid treatments can related to the acid strength.
contribute to a better acid job by facilitating the main-
tenance of stage integrity and reducing displacement
volumes. Mechanical limitations associated with arti- 18-10.2. Flowback and cleanup techniques
ficial lift (e.g., gas lift) are more easily overcome by Selection of the correct flowback procedure is critical.
the use of coiled tubing. The risks of leaking valves, The flowback during multiphase transition periods can
undiluted acid remaining in the mandrels and acid cause irreversible damage. The fines loosened during
leaking into the tubing/casing annulus are avoided. the acid job are invariably produced back into the
The limited injection rate coincidentally controls the near-wellbore area. These fines can be removed in
contact time. The pump rate and extraction out of diluted concentrations that pass through the comple-
the tanks holding the acid can create a bottleneck tion if small, gradual pressure drops are created. This
during execution. A complete understanding of the was demonstrated by Krueger (1986).
operational aspects is necessary for proper execution. The following are key factors to consider for flow-
One of the considerations in selecting the stage back in sandstone formations:
volumes is the tubulars. The volumes of diverter and
their location in the tubulars while injecting must be • The fluids flowing back are more viscous than
considered, especially for the use of foam diverters. those injected. They are capable of carrying natural
When using foam diversion techniques, brief shut- formation fines and other partially dissolved solids
downs or momentum changes are called for to maxi- at lower velocities, which can cause plugging
mize diversion. If the foam is in the tubulars when before the well cleans up completely.
the shutdown occurs, phase separation of the foam • The spent acid usually has a higher density than the
can occur, affecting the foam diverter performance. formation water. The tubing pressure should be
Another consideration is the preflush activity. lower than when connate water is produced, owing
If formations do not have much solubility in HCl, to the higher hydrostatic pressure of the spent acid.
operators have tended to lower the volume of acid • Spent acid has an equilibrium established of poten-
preflush and use brine. However, Gdanski and tial precipitants, held in place by dissolved gases and
Peavy (1986) reported that this is not a good idea dissolved salts. Should these gases (e.g., CO2) be
because the HCl preflush performs the vital func- removed from the spent fluid as a result of creating
tion of cation exchange, which prepares the mineral an excessive pressure drop, precipitation will occur.
surfaces for the HF mixture. The cation exchange • A minimum velocity is necessary for liquid to be
must otherwise be done by the HCl portion of the voided from the tubing without slippage occurring.
HF mixture, which raises the pH of the acid system The minimum velocity to the unload tubing can be
and induces the precipitation of silicate complexes. calculated. The flow rate and tubing pressure in this
As a minimum, the preflush should penetrate the calculation should include the heavier liquid den-
same distance as the HF mixture (e.g., if the HF sity. The flow rate should be achieved gradually but
blend volume penetrates 2 ft, then the preflush sufficiently soon to avoid precipitation in the for-
should penetrate a minimum of 2 ft). mation. The rate should then be maintained until all
Where the HCl solubility is moderate to high, injected fluids are returned and both the tubing
more HCl is necessary. The following equation is pressure and production rate are steady. Plotting the
used to calculate this volume and address the HCl- gradual incremental choke changes as pressure and
soluble materials: rate stabilize provides insight to the affect of the
acid treatments on the formation and completion.

• HF systems should be flowed back immediately • Did any of the treatment fluids or stages create prob-
after injection of the overflush. The potential dam- lems during the execution? During the treatment
aging precipitates that are generated form when the were there mixing or handling problems associated
pH increases as the HCl is spent. If the acid is with any of the additives or the fluid system?
returned quickly, then the pH change may not reach • What are the properties of the formation fluids (i.e.,
the range for precipitation. Many iron precipitates hydrocarbons and brine) and are these compatible
also drop out when the pH increases. The exception with the treating fluids? Post-treatment flowback
is fluoboric acid treatments. The shut-in time inspection and analysis of fluids identifies emul-
required for complete HF generation and fines sta- sions (treating fluid additive formulation), solid
bilization varies on the basis of temperature. debris (proper acid strength) and other telltale signs
• The majority of the additives that are injected are of precipitation caused by incompatibilities.
produced back. Because the acidizing additives are • What is the type of completion injected into and
by design water soluble, they are partitioned into the was this a consideration during the treatment?
water phase. This can cause separation and floatation When injecting into gravel packs or frac and pack
equipment problems. The return fluids are also acidic, completions, injection rates should be limited if
which creates problems for chemical-electric detec- the injected height is limited. Too high an injection
tion devices in the separation equipment. Local envi- rate through the perforations can evacuate them of
ronmental regulations may dictate water quality gravel and create an unstable and unsatisfactory
standards that are difficult to achieve, and disposal environment where the potential exists for forma-
of the returned fluids can be cost prohibitive. tion sand production. This is especially true for
Alternatives to disposal of the returns as hazardous high-solubility zones where a small percentage
waste have been developed, including filtration of the perforations is taking fluid.
through inexpensive media (Hebert et al., 1996).
• Was the proper diversion technique or sequence
chosen and applied? Most acid treatments require
diversion. The application of proper diversion tech-
18-11. Acid treatment evaluation niques with the selected acid system is vital to the
Matrix treatment evaluation is the subject of Chapter ultimate success of the treatment.
20. The following is a partial list of the basic ques- • Was the well appropriately prepared before acidiz-
tions that should be answered during the evaluation of ing? When key steps of preparing the wellbore for
an acid treatment to help determine the success or fail- the acidizing process are left out (e.g., not pickling
ure of the treatment. the tubing, not removing the gas lift valves, not
removing the rods or an electrical submersible
• Was the well damaged? Was there an improvement pump), the prospects for ultimate success are
in the injectivity or transient skin during the treat- reduced. Wellbore preparation is especially critical
ment? Is there evidence that the well was dam- for acid treatments. Injection of tubular debris into
aged? the formation can be disastrous, and acid in the
• Which fluid system or stage accomplished the most annulus of a gas lift completion string is corrosive.
damage removal? Injectivity values or transient • Were the injected fluids checked using quality con-
skin values for each of the fluid stages must be trol steps? Acid strength and certain additives must
evaluated to help identify what damage was caus- meet at least threshold ranges for activity and com-
ing the most significant production impairment. patibility limits.
• Were emulsions observed during cleanup? During • Was the tubing acid cleaned (pickled)? Rust and
the cleanup of the treatment is when the effective- mill scale must be removed, even with new pipe.
ness of additives and treatment fluid packages
demonstrates value. Cleanup time, emulsion prob- • Were the pumped fluids sampled and checked for
lems and facility upset have an economic impact cleanliness and concentration? Although samples
and can be cause for considering different methods are routinely taken and checked before the job
of handling the problems. starts, samples should be taken during the pumping
of each stage. Many changes in injection behavior

can be explained when these samples are analyzed. brought back on line, the production does not
Fluctuations in injectivity may be due to a process improve because of limitations associated with the
problem that was innocently incorporated for oper- wellbore construction or production facilities.
ational expedience or safety compliance. • Was a pressure buildup test performed and inter-
• Was an injection test with the appropriate fluid preted? A pressure test analysis is the definitive
made before pumping acid to establish a base injec- method to answer whether the treatment is a suc-
tivity before acidizing? Injection testing conducted cess or failure.
with platform or field equipment not intended for
this purpose can produce misleading results. Con-
taminated fluids or poorly equipped monitoring can 18-12. Conclusions
result in bad data.
• Did the acid response during the injection validate Acidizing sandstone formations is not an impossible
the damaging substance identified? HCl-soluble task, but it is not simple either. Success requires a
scale may be revealed as the obvious contributor methodical, systematic approach. It can be accom-
to the injectivity problem when large pressure drops plished without any detrimental effects by analyzing
occur when the HCl reaches the formation. If the vital information. The flow chart in Fig. 18-6 shows
injection rate is increased, similar pressure drops the steps the process encompasses.
could also be noted when the HCl-HF mixture The following conclusions can be made about sand-
reaches the formation. stone acidizing:
• Did the pressure response indicate good diversion? • Damage identification determines the types of acids
Pressure increases may be interpreted as diverter and other solvents to use in a sandstone acidizing
action, but this is not always the case. Diverter treatment.
response should coincide with the use of a diverter; if • A knowledge of the chemical reactions involved
not, other parts of the process should be investigated. among acids and formation minerals and connate
• How long was the spent acid left in the well before fluids provides some guidelines for acid types, con-
flowback? Some secondary and tertiary reactions centrations and the sequence to prevent or reduce
require time to produce precipitates. Quick turn- the precipitation of insoluble reaction products.
around for flowback, not high production rates, • The selection of appropriate types and volumes of
lessens the potential for these reactions to create preflushes and overflushes also helps prevent incom-
damage. Most of these reactions result in damage patibilities between formation fluids and acid systems.
that is detected only after production is initiated. • A numerical simulator should be used to quantify
• Were spent acid samples recovered and analyzed? acid volumes, although simple guidelines are pro-
Flowback fluid samples should be acquired regard- vided to assist in the selection of treatment vol-
less of the volume of the treatment. These samples umes. The most important factor in successful acid
should be marked, with the date and time, total vol- stimulation is to provide clean, filtered acids at the
ume recovered to that point and other pertinent perforations by filtering all fluids and cleaning
data, such as choke size, flowing tubing pressure, (pickling) the tubing before the acid treatment
water cut and produced quantities. is injected into the formation.
• Were the production results consistent with the acid • Evaluating the executed acid treatment provides
injection pressure response? If the injection pres- information to improve subsequent acid treatments
sure declines too quickly, the acid treatment causes in the same or similar formations.
the well to develop a vacuum. Once the well is

Well Core Pressure

schematic Logs data data
Evaluate as an acidizing candidate

using
NODAL analysis
Mechanical Liquid
No Acidizing Yes samples
problems
required
Yes Laboratory
analysis
Design process
Fluid design using
expert system or How do the fluids
simulators affect cleanup?
Is the
chemistry No Equipment
good? constraints:
Treatment requirements
Yes 1. Fluid 1. Stage Location requirements Calculate
2. Fluid 2. Stage Fluid system cleanup
flow rate
3. Fluid 3. Stage
Prepare
equipment and
Pump
materials and
the
instruct personnel
Record job
Monitor the
job
Yes Finish
Cleanup
Treatment job
process
parameters execution
Alter job
execution No
Review and reference

for future work
Figure 18-6. Sandstone acidizing treatment design process.

◆
Fluid Placement and
Pumping Strategy
J. A. Robert, Schlumberger Dowell
W. R. Rossen, University of Texas at Austin
19-1. Introduction The last method is specific to carbonate formations.

It involves pumping gelled acid, with the viscosity
Fluid placement is critical to the success of a matrix varying with the degree of spending. Injectivity is
stimulation treatment (Pye et al., 1970; Cooper and reduced in zones containing large volumes of spent
Bolland, 1984). Damage, depending often on fluid- acid, and fresh acid is directed to lower injectivity
rock interactions, may be unevenly distributed along zones. This method has been increasingly used recently
the net thickness. Also, the natural reservoir perme- and has produced good results in terms of placement. It
ability may vary considerably, with substantial con- is described in more detail in Chapter 17.
trasts. In this environment, matrix stimulation tends The choice of the pumping strategy must take into
to remove or bypass the damage that is easier to account several parameters, such as producing-interval
reach (i.e., a lower degree of damage or higher per- thickness, lithology and permeability profile, damage
meability) and becomes self-defeating. Each addi- distribution and the economics of the treatment.
tional volume of stimulation fluid follows the path
of least resistance, and more of it invades the layer
where it is least required. 19-2. Choice of pumping strategy
Several methods have been developed to improve
fluid placement during matrix acidizing. Nonmech- 19-2.1. Importance of proper placement
anical methods include the use of particles accumu-
lating at the sandface to form a low-permeability Matrix stimulation is almost always performed in
cake (i.e., particulate diverters). The size of the parti- multilayer reservoirs containing zones with, at times,
cles varies from a few tens of micrometers to a few wide injectivity contrasts caused either by different
millimeters. Other methods involve the use of vis- permeabilities or by uneven severity of damage.
cous fluid slugs or the use of foam either as staged Other phenomena causing vertical heterogeneities
slugs or combined with the acid stages (i.e., foamed within a completion interval include permeability
acid). Each of these methods is based on the tempo- gradients in thick reservoirs and selective damage in
rary impairment of the high-injectivity zones accept- some perforations. The natural trend of stimulation
ing most of the diverting material, which results in an fluids is to follow the path of least resistance (i.e.,
increase of the proportion of fluid going into the low- to invade the most permeable or least damaged
injectivity zones. It is essential, of course, to avoid zones). To optimize treatment results, most of the
permanently damaging the higher injectivity zones. open interval must be treated, and thus, treating fluids
Mechanical techniques consist of ball sealers, must also be injected in significant volumes into the
which are rubber-lined balls added to treating fluids least permeable and most damaged zones.
to plug fluid-taking perforations, and packers, which In many cases, the use of diverting techniques has
enable isolating a given zone during a treatment. proved successful in improving fluid placement and
Mechanical techniques differ from the other diversion treatment results, especially in cases with large vari-
methods in that they completely shut off a part of the ations of initial injectivity throughout the open inter-
reservoir to direct all the stimulation fluid to a sub- val (Brannon et al., 1987).
section of the open zone. Another technique is coiled As an example, consider the three-layer reservoir
tubing placement in which acid is spotted or injected described in Table 19-1. Middle layer 2 is a high-
across the zone of interest. This is especially impor- injectivity zone of limited thickness (10% of the total
tant in horizontal wells or vertical wells with long pay zone) with high permeability and a lower skin
producing zones (Thomas and Milne, 1995). effect. A simple calculation, using a steady-state

Table 19-1. Example three-layer reservoir. for the first layer and 10 md and 10 ft for the second
layer. Simulation results show that the MAPDIR tech-
Layer Permeability Net Thickness Skin nique allows the fastest total skin effect reduction in
(md) (ft) Effect
terms of pumping time but at the expense of large acid
1 50 50 10 volumes injected into the layer with lower damage.
2 500 10 5 Furthermore, it does not allow treating the low-perme-
3 100 40 10 ability layer fully.
In terms of the total pumped volume, the continu-
injection relationship, suggests that the high-perme- ous injection of particulate diverter appears to be the
ability layer would initially take 51% of the total flow. most efficient technique (Fig. 19-1). The viscous fluid
Assuming that 150 gal/ft of perforated interval is method appears to be far less efficient than the other
required in each zone to remove the damage, the total diversion methods. Moreover, because of the implied
volume required to place at least this amount in each reduction in the injection rate, this last technique
zone depends on the placement strategy: requires the largest pumping time for treatment com-
pletion. However, the study assumed Newtonian flu-
• Total volume is equal to 150 gal/ft × 100 ft = ids. Power law gels may be more effective. Another
15,000 gal in the case of “perfect” placement simulation with the same permeabilities in each layer,
(i.e., the same injectivity in the three layers). but different initial values of skin effect, led to similar
• With no diversion, a volume of about 39,500 gal conclusions.
is necessary to treat the entire interval, including The choice of the best diversion technique depends
the layer with the lowest injectivity. In this calcu- also on a variety of other parameters. Particulate diver-
lation it is assumed that the ratio of the flow rate sion can be used in most cases, with some restrictions
into each layer remains constant throughout the in the case of gravel-pack completions. Materials have
treatment, but because damage is removed more been developed for a wide range of reservoir properties,
rapidly in the highest injectivity zone, an even the most important of which are the reservoir perme-
larger volume would be actually required (see also ability and the pore-throat size and distribution. Proper
Chapter 13 for numerical simulations). design of particulate diverters requires consideration of
the compatibility of the diverting agent with the well-
bore and reservoir conditions (e.g., temperature, reser-
19-2.2. Comparison of diversion methods voir fluid) and with other treating fluids (e.g., solubility
and flocculation in the carrying fluid or adjacent stages
A good early review of the development and use can impair particulate diverter efficiency).
of diverting techniques, including mechanical tech-
niques (packers and plugs) and various diverting
10
agents, was presented by Harrison (1972). MAPDIR
Hill and Rossen (1994) presented a comparative Viscous diversion
Continuous particulates
study of the efficiency of the existing nonmechanical 8 Batch foam
Batch particulates
placement techniques, including the maximum pressure Continuous foam
Skin effect factor
differential and injection rates (MAPDIR) technique 6

introduced by Paccaloni and coworkers (Paccaloni and
Tambini, 1990; Paccaloni, 1992). The MAPDIR tech- 4
nique involves pumping at the highest possible rate
without fracturing. Hill and Rossen also considered par-
2
ticulate diverters, foams and viscous fluids. In their
paper, they compared the evolution of skin effect for
a hypothetical two-layer case. For an assumed initial 0
0 10 20 30 40 50 60
skin effect of 10, the skin effect decrease is proportional
V (gal/ft)
to the acid volume injected per layer, and total damage
removal is achieved with 50 gal/ft of acid. The perme- Figure 19-1. Total skin effect evolution versus volume for
ability and thickness are, respectively, 100 md and 1 ft different placement techniques (Hill and Rossen, 1994).
19-2 Fluid Placement and Pumping Strategy

In the case of gravel-pack completions, large-parti- (1984) and Schechter (1992), however, argued that
cle-size chemical diverters are not used. Oil-soluble there exists an optimum rate for sandstone acidizing,
resin (OSR) may be used, but it must be mixed well based on the reaction rate with damage and acid
and injected continuously with the acid to avoid the and/or damage contact time. For carbonate acidizing,
risk of plugging the screen or gravel pack (McLeod, it has been shown that the injection rate must be high
1984). Foam diversion is also common for gravel- enough to allow wormholes to form and propagate.
pack completions and where particulate diverters are Pumping below this rate can cause a considerable
not appropriate. reduction in treatment effectiveness.
In the case of oil wells with a well-defined water
zone, foam has been found to plug the water zone
successfully and direct acid to the oil-bearing interval. 19-2.4. MAPDIR method
This is due to the detrimental effect of oil on foam The MAPDIR method suggests pumping treating fluids
formed by certain surfactants. Acidizing with foam as fast as possible below the fracturing limit without
diversion has been reported to improve overall pro- using any diversion technique. It allows a decrease of
duction in some cases without increasing the water pumping time and minimizes the risk of treatment fail-
cut, but instead reducing it (Zerhboub et al., 1994). ure caused by low pumping rates. However, MAPDIR
For cemented and perforated completions, buoyant is not a true diversion technique, as it does not modify
ball sealers have been shown to give good results the natural flow profile, nor does it necessarily distrib-
(Erbstoesser, 1980; Bale, 1983). Conventional ball ute stimulation fluids or remove all damage.
sealers are generally not recommended for long Paccaloni and coworkers analyzed a large series
intervals or high-shot-density completions, as they of matrix-acidizing treatments to identify the factors
require a minimum injection rate per perforation to leading to success or failure, including pump rate and
prevent settling in the rathole (McLeod, 1984). the use of diversion (Paccaloni et al., 1988; Paccaloni
Packers are the most effective means of selective and Tambini, 1990). Criteria for success or failure were
fluid placement. However, they require the use of specific to the authors. With development wells, job
coiled tubing or rig operations. They can be used after success involved treating the entire interval. On the
completion or workover, if a rig is already on site. basis of the analysis of more than 170 jobs, the authors
Coiled tubing has been increasingly used for claimed that the MAPDIR technique is sufficient to
matrix acidizing over the past few years. Treatment obtain full coverage of the pay zone if the permeability
analysis shows that coiled tubing is especially help- contrast is less than 300 md and the pay zone does not
ful for acidizing long intervals by allowing spotting exceed 200 ft. They concluded that the method actually
successive diverter and acid stages throughout the gives better results than diverting agents.
open interval, thereby ensuring good coverage of the In the case of carbonate formations, they reported
entire producing zone (Economides et al., 1991; that relatively small volumes of concentrated acid (5
Thomas and Milne, 1995). to 10 gal/ft) led to a substantial skin effect reduction,
allowing for reasonable injectivity of the zone for fur-
ther reduction. This could explain why the MAPDIR
19-2.3. Fluid placement versus method can eventually provide full zone coverage: if
injection rate the treatment is maintained long enough to inject the
Selection of the treatment strategy must be based on few gallons required to improve injectivity in the highly
the trade-off between pump rate and fluid placement. damaged zones, then injectivity contrasts are reduced
Except for the MAPDIR technique, placement meth- and the entire interval can be treated. However, this
ods result in a reduction in the injection rate, even in method can be applied only for intervals with limited
low-injectivity zones, with possible loss of treatment contrasts of permeability, and it does not optimize fluid
efficiency and creation of precipitation damage. placement.
Williams et al. (1979) mentioned that sandstone For better efficiency, the authors recommended the
acidizing is more efficient at higher rates, both avoid- use of coiled tubing for preacidizing operations such
ing the creation of precipitates near the wellbore and as wellbore cleanup, acid circulation for completion
extending the radius of live-acid penetration. McLeod string cleaning and acid spotting in front of the pay

zone. Then a T connection can be used to pump acid 19-3.2. Diverting agent properties
through both the coiled tubing and tubing string to
maximize the pump rate during matrix injection. An effective diverting agent must meet both physical
and chemical requirements.
• Physical requirements
19-3. Chemical diverter techniques Cake permeability—Cakes created on the reser-
Chemical diverters, which are materials insoluble voir walls by agents must be as impermeable as
in acid but highly soluble in water or hydrocarbons, possible to the acid for the maximum diverting
have been used either to form a low-permeability filter effect. If the permeability of the diverter cake is
cake at the sandface or to reduce the injectivity of greater than or equal to the permeability of the
high-permeability zones with the injection of a vis- tightest zone, little or no diversion occurs.
cous polymer slug. The first technique has been found Invasion—Deep invasion of the reservoir rock
to be more effective and can provide faster cleanup. by the diverter particles must be prevented, regard-
It has prevailed over the viscous slug technique. less of the nature of the rock, for maximum diverter
effectiveness and minimum problems in cleanup.
This and the previous requirement suggest that an
19-3.1. Historical background optimum particle size must be determined.
Dispersion—Diverting agent particles must be
The first attempts at fluid placement made use of properly dispersed in the carrying fluid. If floccu-
chemical additives. Harrison (1972) reported the lation occurs, the cake will be thicker but with
injection, as early as 1936, of soap solutions that high porosity and permeability.
could react with calcium chloride (CaCl2) to form
water-insoluble, but oil-soluble, calcium soaps. The • Chemical requirements
precipitate acted as a diverting agent for hydrochlo- Compatibility—Diverting agents must be compati-
ric acid (HCl). The generation of solid precipitates in ble with the bulk treatment fluid and with other
the formation was not, of course, desirable, because additives, such as corrosion inhibitors, surfactants
they could cause permanent damage. Thus, in the and antifoaming agents. They must be inert toward
late 1930s, more sophisticated systems were used, the carrying fluid at the well treating conditions.
such as heavy CaCl2 solutions. Diversion effects Cleanup—Diverting agents must be soluble in
were possible by impairing injectivity in the zones either the production or injection fluids. Having
accepting most of the high-viscosity solution. Other acted as diverters, they should undergo a rapid
systems utilized cellophane flakes suspended in a and complete cleanup.
water gel with a bacterial breaker. Later, gels were
replaced by oil-external emulsions.
In 1954, naphthalenes were first used as a blocking 19-3.3. Classification of diverting agents
material. Oil-soluble naphthalenes were thought to be Diverting agents can be classified, according to their
ideal diverters because they sublime above 175°F particle size, as bridging agents or particulate diverters.
[80°C]. Also, crushed limestone, sodium tetraborate,
oyster shells, gilsonite, perilite, paraformaldehyde and • Bridging and plugging agents
“chicken feed” were used as diverters with mixed suc- These diverting agents consist of large-size parti-
cess. These compounds were replaced progressively cles, from 10/20 to 100 mesh. They are used as
by rock salt, which is partly soluble in the acid but diverters in carbonate formations, where natural
inexpensive and easy to handle. fractures are common. However, their efficiency
A major improvement in diversion techniques was is limited by the high permeability of the cakes
brought about by completely soluble materials, they create. As an example, a cake formed with
including wax-polymer blends and hydrocarbon 10-mesh particles has a permeability between
resins in production wells and rock salt and benzoic 20,000 and 40,000 md, whereas a cake created
acid in water-injection wells. with 100-mesh particles has a permeability
between 1,000 and 3,000 md.

When effective diversion is required in fractured – Oil-soluble, for production wells—Oil-soluble

zones, a slug of bridging agent is injected first, fol- agents are blends of hydrocarbon resins. They
lowed by the treating fluid containing a diverting are totally inert in an acidic medium, yet quickly
agent. The bridging agent fills the fractures, creates and completely dissolve in the produced oil after
a coating in front of high-permeability formations treatment. It is often difficult to disperse these
or both. If the plugging agent consists of particles resins in the acid. They are usually injected as
small enough (e.g., 100 mesh), the diverting agent dispersions in aqueous solutions.
does not penetrate the coating and allows effective
diversion through the development of low-perme-
ability cakes. 19-3.4. Potential problems during
Bridging agents with various chemical compo- diversion treatment
sitions are available:
During a diversion treatment several major problems
– inert materials, such as silica sand or a mixture
may occur.
of silica, nonswelling clay and starch (Soluble
materials reduce the risk of permanently plug- • Decantation
ging the formation.) Bridging agents are subject to gravity effects dur-
– water-soluble bridging agents, including rock ing injection. They tend to settle in the fluid that
salt and benzoic acid fills the well, and the best results are therefore
– oil-soluble bridging agents, including graded obtained when the solid additives are placed in a
oil-soluble resins, naphthalenes flakes and beads water-base gel. Decantation problems rarely occur
made of wax-polymer blends that are soluble in with diverting agents because of their small parti-
oil, deformable and temperature degradable. cle size and the low density of their constituent
(Their composition can be selected according materials (1.1 g/cm3 for resin, 1.3 g/cm3 for ben-
to the bottomhole temperature.) zoic acid).
• Particulate diverters • Solubility
Particulate diverters are characterized by very Water-soluble agents are also slightly soluble in
small particle sizes, well below 0.004 in. in diam- acid. Therefore, a portion of these additives dis-
eter. Both water- and oil-soluble particulate divert- solves during mixing with the acid. Another por-
ers are available. tion dissolves during injection because of inter-
– Water-soluble, for injection wells—A fine grade mixing with wellbore fluids and heating. If the
of benzoic acid is typically used as a water-sol- plugging agent is injected in slugs, after it is
uble diverting agent. Because this product deposited in the fractures or on the formation
agglomerates during storage, it is difficult to walls it undergoes additional dissolution. Thus,
achieve a constant particle-size distribution the effectiveness of water-soluble products is
before injection. For this reason, salts (i.e., always low, and large initial concentrations are
ammonium or sodium benzoate) can be used required. The dissolution rate during the different
instead. In HCl, the salts are converted to ben- stages of injection must be considered in the job
zoic acid. For instance, sodium benzoate reacts design.
according to • Particle-size distribution
C6H5COONa + HCl C6H5COOH + Na+ + Cl– The particle size of the diverting agent must corre-
spond to the petrophysical properties of the treated
Benzoic acid is only slightly soluble in HCl but zones, such as permeability and pore-size distribu-
highly soluble in water or alkaline solutions. tion. If bridging agents are used in a reservoir with
After acting as a diverter, this compound dis- permeability from 100 to 1000 md, the resistance
solves in the injection water. Because benzoic of the diverter cake may not be high enough to
acid particles are hydrophobic, surfactants are avoid fluid penetration in the high-permeability
required to properly disperse the agent in the zones. On the other hand, if too fine a diverting
treating fluid. agent is used, the solid particles migrate through

the porous medium with the treating fluid and neous reservoirs. Pressure drops through cakes of
diversion will not occur. Where the rock pores have diverting agents were measured under various well-
a slightly larger diameter than the diverter particles, bore and fluid conditions, such as temperature, flow
an internal cake can be created, and although the rate, concentration of diverting agent and nature
diverter will be efficient, cleanup can be difficult. of the carrying fluid. Filtration theory was used to
• Flocculation express the experimental results in a more general
form:
Incompatibility of diverting agents with additives
used in the same stage or in adjacent stages can ∆p
Rcake = , (19-1)
cause diverter flocculation. Carrier fluid composi- µu
tion should follow technical memoranda recom-
where Rcake is the cake resistance in m–1, ∆p is the
mendations or be tested in the laboratory.
pressure drop across the diverter cake, µ is the carry-
• Compatibility ing fluid viscosity, and u is the superficial velocity
Rock salt should never be used as a bridging agent across the cake. By analogy with the pressure drop
in hydrofluoric acid (HF) treatments or before HF in a porous medium, the cake resistance can be
treatments because it may increase the risk of sodi- defined as the inverse of the cake permeability divid-
um fluosilicate precipitation. Such a problem is not ed by the cake thickness. Laboratory testing allows
anticipated when using sodium benzoate, because correlation between the cake resistance and the vol-
this compound is readily converted into benzoic ume of diverter deposited at the sandface per unit
acid in HCl, and the released sodium ions do not area, which can then be used to simulate the diverter
contact the subsequent HF flush. effect at reservoir conditions.
Taha et al. (1989) and Economides et al. (1994)
extended the model introduced by Hill and Galloway.
19-3.5. Laboratory characterization They defined a specific cake resistance α in m/kg:
Crowe (1971) designed equipment that injects slugs 1
α= , (19-2)
of diverting agent under constant pressure into parallel ρdiv (1 − φ cake )kcake
cores of different permeabilities. By measuring the
time required to equalize the flow rates entering each which then leads to the pressure drop ∆p across the
core, he compared the efficiency of different diverters. cake:
The best results were obtained using OSRs. αµuCdivρdiv V
∆p = . (19-3)
Other investigators (Hill and Galloway, 1984; A
Houchin et al., 1986) attempted to use a slightly dif-
In Eqs. 19-2 and 19-3, ρdiv is the density of diverter
ferent experimental setup consisting of a well model
particles in kg/m3, φcake is the cake porosity, kcake is
with three cores of various permeabilities to predict
the cake permeability, Cdiv is the concentration of
flow distribution at presumed reservoir conditions.
diverter particles in m3 of particles per m3 of solution,
This approach, because of differing pressure-drop
V is the total volume of diverter solution injected,
ratios and slug-volume ratios between laboratory and
and A is the surface of cake deposition.
reservoir conditions, cannot always be extrapolated
Equation 19-3 is obtained by writing Darcy’s law
to field conditions (Doerler and Prouvost, 1987).
across the cake:
Hill and Galloway’s original attempt to measure
the pressure drop across a diverter cake included the µlu
∆p = . (19-4)
experimental data in a simple numerical reservoir kcake
model. However, they assumed the pressure drop
The cake thickness l varies with the diverter vol-
to be a linear function of the cumulative volume
ume as
of diverter and neglected the effect of flow rate and
Cdiv V
other important parameters, such as diverter concen- l= . (19-5)
tration and temperature. (1 − φ cake ) A
Doerler and Prouvost investigated separately the
Combining Eqs. 19-4 and 19-5 and using α as defined
intrinsic properties of the cakes created by diverters
in Eq. 19-2 yields the expression of the pressure drop
and their effects on the flow distribution in heteroge-
across the cake as in Eq. 19-3.

The cake resistance Rcake and the specific cake

resistance α are related by With diverter cake
Without diverter cake
αCdivρdiv V
Rcake = . (19-6)
Pressure, p
A
pw
Either characterization of the diverter cake can ∆p

be used. However, Doerler and Prouvost took into
account the cake compressibility in their correlations
pe
for Rcake, whereas the other authors neglected the
cake compressibility and assumed α to be constant. rw re
Radial distance, r
19-3.6. Modeling diverter effects Figure 19-2. Radial pressure profile in the reservoir with
and without a diverter cake.
Presented in this section is a methodology using cake
resistance data obtained in the laboratory to predict
the efficiency of the treatment in terms of flow distri- 6
With diverting agent
bution between the different layers. The concept of Without diverting agent
diverter cake pseudoskin effect is introduced.
Acid penetration (in.)

Drilling damage zone
The diverter cake created at the sandface of a layer 4 CHF = 0.1
induces a pressure drop across a zone of negligible CHF = 0.5
thickness. Figure 19-2 depicts the radial pressure
profiles in a formation layer during injection for two 2
CHF = 0.9
different situations:
Perforation
• fluid injection without a diverter (dashed line) Crushed zone
0
• injection after a diverter cake has been deposited 0 2 4 6 8 10
at the sandface (solid line). Acid penetration (in.)
The injection and reservoir pressures are identical in Figure 19-3. Hydrofluoric acid concentration CHF contours
both cases, but the injection rates are different. The around a perforation (Lea et al., 1993).
cake acts as a temporary skin effect. Consequently,
a pseudoskin effect factor can be defined by the stan- drilling mud damage (4 in. depth). The figure shows
dard (van Everdingen and Hurst, 1949) with an that the diverting agents do not significantly modify
injection rate qi: the flow patterns around the perforation, and therefore
they have little impact on the rate of skin effect reduc-
2πkh
scake = ∆p , (19-7) tion for a single perforation (Fig. 19-4). These results
µqi indicate that particulate diverters do not provide diver-
where k is the permeability and h is the thickness. sion within the perforations. They provide diversion
Use of particulate diverters in a perforated comple- from one perforated zone to another. The model pre-
tion requires special attention because they can mod- sented in this section is thus applicable also to a perfo-
ify the flow pattern in the perforation tunnel itself. Lea rated completion.
et al. (1993) showed that a diverting agent tends to Using the definition of cake resistance (Eq. 19-1)
equalize the flux entering the formation at the perfora- and expressing the fluid superficial velocity u through
tion walls. However, the flow profile in the formation the cake as qi /A in Eq. 19-7, then
is determined mostly by the damage distribution, as 2πkh
acid tends to migrate toward the higher permeability scake = Rcake , (19-8)
A
zone. Figure 19-3 shows the acid concentration con-
tours obtained by numerical simulation of a perfora- where A is the sandface area exposed to flow and
tion affected by both a crushed zone and shallow available for cake deposition.

q
q i = 0.007 bbl/min
15 q i = 0.0035 bbl/min
Perforation skin effect
10
Depth
5 Permeability
(md)
5000
scake,1
0
q1 k1, s1 h1
–5 5050
0 2 4 6 8 10 q2 scake,2 k2, s2 h2
Injection volume (gal)
q3 k3, s3 h3
scake,3
Figure 19-4. Skin effect evolution for different injection 5100
rates and diverting agents (Lea et al., 1993).
Figure 19-5. Flow distribution in a multilayered reservoir:

• For an openhole: A = 2πrwh, where rw is the well- on the left, without a diverter; on the right, with a diverter.
bore radius.
• For a cased hole: A = n perf2πrperf l p, where nperf is equations, the wellbore radius is rw and the reservoir
the number of perforations in the layer, rperf is the radius is re. Each layer is characterized by its net
radius of a perforation, and lp is the length of the thickness hj, undamaged permeability kj and skin
perforation tunnel. effect sj. The subscript j denotes the layer number.
For the assumptions that the cake thickness is neg- For simplicity, a single-phase, radial and horizontal
ligible and the fluid velocity across the cake does not steady-state flow is assumed to prevail. Whenever
vary, the cake resistance obtained under linear flow more complicated geometries exist, such as commin-
conditions in the laboratory can be used. gled layers and crossflow, a more sophisticated inflow
The growth rate of the diverter cake can be related performance relationship may be used. This would
to the injection rate. If ρa is denoted as the mass of require the results of layered reservoir testing, such
cake per unit area of sandface available for deposi- as described by Ehlig-Economides and Joseph (1985).
tion, its derivative with respect to time can be With the simplified assumptions, and assuming
expressed for any layer by steady state, the injection rate entering layer j is
dρa Cdiv
′ qi 2πk j h j ∆p
= , (19-9) qj = , (19-10)
dt A  r  
µ ln e  + s j + scake, j 
where t is the time and Cdiv′ is the net diverter con-   rw  
centration in kg/m3 after dissolution effects are con- where s is the skin effect from damage and ∆p is the
sidered. In Eq. 19-9, it is assumed that the injected pressure differential of the wellbore pressure pw,j
fluid is evenly distributed on the sandface. If this minus the outer reservoir pressure pe. The skin effect
were not the case, as along a perforation tunnel, a scake,j is time dependent because of cake buildup, and
finer modeling involving infinitesimal elements of sj varies with time if the effect of damage removal by
the surface dA would be required. acid is considered.
To calculate the evolution of flow and pressure dur- Equations 19-8, 19-9 and 19-10 are coupled
ing diverter injection requires expressing the inflow because the cake resistance and pseudoskin effect
performance relationship (IPR) for a multilayered depend on ρa , which is itself dependent on the injec-
reservoir and linking the diverter pseudoskin effect to tion rate. Eliminating qj between Eqs. 19-9 and 19-10
the flow rate entering each layer. Figure 19-5 depicts and using Eq. 19-8 to replace scake,j , an equation gov-
a cylindrical reservoir made of n horizontal layers, at erning the rate of growth of the cake in layer j is
the center of which is a vertical well. In the following obtained:

dρa , j 2 πk j h j ∆p ′
Cdiv
= , 60
dt   re   Aj Layer 1
µ ln  + s j + 2 πk j h j Rcake, j Aj  Layer 2
  rw   Layer 3
(19-11)
50
where ρa,j is the mass of diverter per unit area in
front of layer j. Equation 19-11 is readily solved for
two cases:
Fractional flow rate (%)

40
• If the injection is performed at constant excess
pressure ∆p, each differential equation can be
solved independently. If the cake resistance varies
linearly with the mass of diverter deposited, the
30
system of equations can be solved analytically
(Economides et al., 1994; Hill and Rossen, 1994).
• If the injection is performed at a constant total rate
qtot, the differential equations are coupled through 20
the expression of total rate for n layers:
qtot = ∑ q j . (19-12)
j
10
The term ∆p derived from Eqs. 19-10 and 19-12 is 0 4 8 12 16
substituted into Eq. 19-11, which is then solved for Time (min)
each layer. A numerical method is required in this
case (Doerler and Prouvost, 1987). Hill and Rossen Figure 19-6. Flow redistribution caused by a diverter slug.
(1994) presented a simplified numerical method that
applies in some cases.
Once Eq. 19-11 is solved for ρa,j (t), the other para- Table 19-2. Reservoir data for
meters of interest, such as q j (t) or ∆p(t) can be the example in Fig. 19-6.
derived. The solution shows that the system tends
toward equalization of the layer injectivities. Figure Wellbore radius 0.35 ft
19-6 illustrates the evolution of the fractional flow rate Reservoir radius 2000 ft
under diverter injection in a three-layer case. At the Carrying fluid viscosity 1 cp
end of the diverter stage, the rate per layer is roughly Diverter concentration 150 lbm/1000 gal
proportional to the layer thickness. Table 19-2 details Temperature 150°F [65°C]
the reservoir geometry corresponding to this case.
Cake resistance function at 150°F Rcake = 5 × 1012 ρa
For a more accurate calculation of the diverter effect,
Total injection rate 2 bbl/min
this model can be included in a finite-difference matrix
acidizing simulator. This enables due accounting of Layer 1 Layer 2 Layer 3
other factors such as the effect of damage removal on Layer permeability 50 md 100 md 500 md
injectivity and transient effects and the use of more Layer thickness 15 m 3m 12 m
realistic and complicated relations for the cake resis-
Initial skin effect 10 5 10
tance versus deposited diverter mass.
• The chemical nature of the diverter must be selected

19-3.7. Field design considering the type of well (injector or producer),
To ensure proper coverage of the zones to be stimu- formation type (sandstone or carbonate), bottomhole
lated, the diverter placement design must include the temperature, treating fluid and cleanup procedure.
following several steps.

• The particle-size distribution must be adapted to can be foamed with the addition of gas and surfac-
the pore size of the formation upon which the cake tant, or, more commonly, foam can be injected in
will be deposited. This should prevent any inva- alternating slugs with acid. Until recently, there was
sion of the porous medium by diverter particles. little published information on acid diversion with
In general, the particle size must be adapted to the foam, in contrast to the extensive literature on foam
formation layer with the highest permeability. for diverting gas flow in improved oil recovery
• A dispersant must be used to ensure that particles (IOR) processes (Hirasaki, 1989a, 1989b; Schramm,
will not flocculate in the carrying fluid. Although 1994; Rossen, 1996). The findings of IOR foam
not recommended, bridging agents must be mixed research have exerted a strong influence on the
in gel pads to prevent sedimentation in some cases. understanding of foam acid diversion. More recent
research suggests that although the basic mecha-
• Once the proper material has been selected, the
nisms of the two processes are similar, differences
volume, concentration and injection rate of the
in the flow regime between near-well and reservoir
diverter must be determined. Generally, concen-
applications cause significant differences in behavior.
trated slugs have been found more effective than
This section summarizes the current consensus on
continuous injection at a lower concentration. The
foam behavior and process design. This is an active
pumping parameters, namely bottomhole pressure
area of research, and this consensus may well shift
and injection rate, must then be estimated to
within the next few years.
ensure their compliance with field operational con-
straints, such as fracturing pressure. The overall
injectivity of the well should not be lowered to the 19-4.2. Foam mechanisms
point that only small injection rates are possible.
Foams act fundamentally differently as they penetrate
The method described in this section can be used and plug the pore space of rock compared with the
to optimize diverter slugs if the well and reservoir behavior of foams used for drilling, cementing, frac-
data are known. Unfortunately, in many cases, critical turing or well cleanout. Therefore, analogies to these
parameters such as the exact perforation geometry, other foams can be misleading. For example, foams in
pore size near the sandface and distributions of perme- pipes are created and maintained by shear forces, and
ability and damage are poorly characterized. In the the resulting bubbles are much smaller than the pipe
past, this has limited the use of mathematical models diameter. Within the tiny pores of an oil or gas reser-
for diverter design. Many practical designs are based voir, however, foam bubbles larger than the individual
on local experience (i.e., trial and error). With the use pores are squeezed down into elongated shapes, span-
of coiled tubing, which greatly decreases the volume ning several pores, with a liquid film or lamella
of the injection string, design can be modified on the between each pair of bubbles (Ettinger and Radke,
fly and the pressure response of the well to acid and 1992; Falls et al., 1989). The behavior of foam is
diverter slugs can help guide diversion optimization. dominated by the capillary forces on the lamellae;
A flat pressure response to acid indicates that a divert- viscous forces in the conventional sense are relatively
er is required, and the efficiency of cake buildup can unimportant.
be assessed from the reservoir response (pressure Numerous IOR foam studies agree that foam does
increase) to diverter injection. This method of design not directly alter liquid mobility in porous media
optimization based on real-time well response is more (Bernard et al., 1965; Friedmann et al., 1991; Huh and
reliable than a priori estimates and is most likely to be Handy, 1989; de Vries and Wit, 1990). In other words,
effective if facilities are available for on-site stimula- the aqueous-phase relative permeability krw is the same
tion monitoring. function of its saturation Sw as in the absence of foam.
(For the remainder of this section, for simplicity the
aqueous phase is referred to as water.) Evidently, upon
19-4. Foam diversion entering rock most of the water leaves the foam and
flows through the same network of narrow pores and
19-4.1. Historical background pore crevices through which it would flow at that same
Foams have been used for acid diversion since at Sw in the absence of foam. As a result, overall mobility
least the 1960s (Smith et al., 1969). The acid itself with foam can be inferred from Darcy’s law if the rela-

tive permeability function krw (Sw) without foam and the These forces completely trap gas bubbles in from 70%
water saturation in the presence of foam are known: to 99% of the gas-filled pore space even as injected
foam flows through the remaining pores (Radke and
∆p =
(u g + uw )= uw
(19-13) Gillis, 1990; Friedmann et al., 1991). The fraction of
kλ rt (k k rw uw ) the pore space completely blocked by foam declines as
the pressure gradient increases. Because of these same
krw ( Sw ) capillary forces, the effective viscosity of the gas that
λ rt = (19-14) flows is much higher than that of gas alone.
uw fw
Both gas trapping and effective viscosity are even
100 − Γ more sensitive to bubble size than to flow rates, how-
fw = , (19-15) ever. There are many processes that can spontane-
100
ously alter bubble size as foam flows through rock
where ug and uw are the volumetric fluxes of gas and (Hirasaki, 1989a; Kovscek and Radke, 1994; Rossen,
water, respectively, λrt is the total mobility as defined 1996). For example, if unfoamed gas and water with
by Eq. 19-14, µw is the water viscosity, fw is the water surfactant are injected into rock under appropriate
fractional flow (i.e., water volume fraction of the conditions, foam may be created within the first inch
injected foam), and Γ is the foam quality (i.e., volume or two of the rock face (Ettinger and Radke, 1992).
percentage of gas in the injected foam). Therefore, the Although it may be convenient to assume that foam is
only way to divert acid (i.e., to reduce krw) is to reduce created instantaneously at the rock face, this entrance
the water saturation Sw by increasing the gas satura- region of low ∆p in which foam is created can be sig-
tion. During steady foam injection, Sw is governed by nificant in a well treatment focused on the near-well
the equation for water fractional flow fw: region. Similarly, if a foam of very small bubbles is
uw 1 injected into rock, there may be an entrance region
fw = = , (19-16) with ∆p higher than that downstream. In addition, if
uw + ug  krg µ w 
1 + µ k ( S ) 
flow rates are suddenly changed, it may take some
 g rw w  time for foam to adjust to the changes in conditions.
where krg and µg are the relative permeability and the The various processes that create and destroy lamel-
viscosity of gas, respectively. lae, together with the non-Newtonian mobility of gas
Capillary forces tend to lower the gas mobility at a fixed bubble size, enormously complicate the pre-
(krg/µg) and drive down Sw (Eq. 19-16). For example, diction of foam behavior in rock. Fortunately, two
the IOR foam of Persoff et al. (1990, 1991) reduced regimes observed experimentally greatly simplify the
gas mobility by a factor of almost 20,000, which in description of foam mechanisms.
turn drove down the liquid saturation Sw and relative First, under some conditions bubble size is domi-
permeability krw to about 0.37 and 0.001, respectively. nated by processes that destroy lamellae, and these
During the injection of surfactant solution without gas processes appear to be sensitive to capillary pressure
following foam, krw remained at 0.001 during several pc in the rock (Khatib et al., 1988): foam properties
pore volumes of injection. As discussed subsequently, change greatly upon a small change in pc. Because pc
foams used for acid diversion are neither so strong nor depends on the water saturation Sw, the large changes
so durable, probably as a result of the higher flow in foam properties upon small changes in Sw, com-
rates used in acid diversion. bined with Eq. 19-13, suggest that ∆p is a simple
The key to the success of foam as a diverter is low function of the water saturation Sw* at which this
gas (and consequently liquid) mobility during foam change occurs (Zhou and Rossen, 1994; Rossen and
injection and gas trapping during the subsequent injec- Zhou, 1995). Under some conditions common in IOR
tion of liquid. Foam reduces gas mobility proportion- foam processes, Sw* appears to be independent of flow
ally to the number of liquid films, or lamellae, blocking rates in a given rock (Persoff et al., 1990). This
the flow of gas or, viewed another way, in inverse pro- implies that foam in this regime can be treated simply
portion to the bubble size (Falls et al., 1989; Ettinger as a Newtonian fluid of low mobility: the effect of pc
and Radke, 1992; Kovscek and Radke, 1994). This on bubble size just balances the non-Newtonian
reduction is due primarily to the capillary forces resist- mobility of foam at fixed bubble size. The fixed-pc*
ing movement of the lamellae separating gas bubbles. model for foam, discussed subsequently, is based on

this mechanism. In conditions where this regime 800

applies, foams are stronger in higher permeability
rock (higher pc; i.e., lower Sw*), diverting flow into
Pressure gradient (psi/ft)

lower permeability zones (Lee et al., 1991; Zhou and 600
Rossen, 1995).
The second regime in which foam behavior appears 400
to be simplified is that in which capillary trapping of
the foam bubbles controls overall mobility. In some
cases discussed in the following sections, ∆p during 200
foam flow or subsequent liquid injection is nearly inde-

pendent of the foam or liquid flow rate (Parlar et al., 0
1995). As mentioned, the trapping of foam bubbles in 0 2 4 6 8 10
much of the pore space of the rock depends on a bal- Foam flow rate (cm3/min)
ance of capillary forces trapping the bubbles and

Figure 19-7. Example of ∆p versus foam flow rate (Robert
applied ∆p (Rossen, 1990). Evidently, in some cases and Mack, 1995).
this balance is highly sensitive to ∆p: as the flow rate
increases, causing a slight rise in ∆p, previously trapped
foam bubbles are liberated and allowed to flow, accom- mobility is controlled by mobilization and libera-
modating the increase in flow rate without a further rise tion of trapped gas in this high-flow-rate regime.
in ∆p. Whether this conjecture is correct, and its full Parlar et al. reported that the threshold ∆p at
implications for the design of foam diversion processes, which flow becomes independent of flow rate
remains to be confirmed. scales roughly as the –1⁄2 power of permeability
values greater than 800 md.
– During foam injection, foams are stronger (i.e.,
19-4.3. Foam behavior in porous media higher apparent viscosity) in high-permeability
• Mobility of injected foam rock, implying the diversion of flow from higher
There is a wide body of experimental data on foam permeability to lower permeability or more dam-
behavior in rock for IOR foams and a growing aged intervals (see, e.g., Kibodeaux et al., 1994).
body of data on foams for acid diversion. It appears Whether similar trends apply to diversion
that there is a substantial difference between foam between layers differing in the extent of damage
behavior in the two applications, probably because rather than permeability per se is not clear.
of the lower quality Γ and much higher flow rates Behenna (1995) used a thin wafer of 0.3-md
of foams for acid diversion. Observed trends in sandstone to simulate a damaged zone at the face
behavior include the following. of a 1-ft sandstone core. Thompson and Gdanski
(1993) used the same approach with carbonates.
– Foam mobility is at a minimum (i.e., foam is However, most foam studies use relatively high-
strongest) at qualities between about 70% and permeability rock (hundreds of millidarcies).
90% (Thompson and Gdanski, 1993). Foams More studies are required to characterize foam
are weaker for qualities greater than 90%. This behavior in lower permeabilities.
is probably due to the collapse of foams at high
capillary pressures; the reason for weak foams – Oil weakens or destroys most foams (Jensen
at low quality is not clear. and Friedmann, 1987; Schramm and Novosad,
1992; Rossen, 1996; Manlowe and Radke, 1990).
– Foams are extremely non-Newtonian at high flow Oil is extremely detrimental to foam in oil-wet
rates (Parlar et al., 1995; Zerhboub et al., 1994), rock. It is possible to tailor a foam formulation
as shown in Fig. 19-7. In this case, the pressure to partially withstand the adverse effects of oil, but
gradient responds nearly linearly to the flow rate this weakening can be an advantage, causing foam
at low values. At high flow rates ∆p becomes to collapse in oil-saturated layers and diverting
independent of flow rate. This suggests that foam acid to stimulate those productive layers.

• Mobility of acid injected after foam of the foam injection rate preceding liquid injec-
Most laboratory studies have used an aqueous tion, the liquid flow rate after foam and whether
slug without acid to simulate acid injection fol- there is a shut-in period preceding liquid injec-
lowing foam. This is a reasonable simplification tion (Zeilinger et al., 1995; Parlar et al., 1995).
if the effects of acid removal of damage and foam In other words, there is virtually unlimited liquid
diversion are separable. This is clearly not the flow above a threshold ∆p. Evidently, ∆p during
case for carbonates, as discussed subsequently. this period is controlled by the mobilization of
The mobility of the liquid injected following foam trapped gas bubbles, as discussed previously
is the key to processes of alternating slugs of acid (Rossen, 1990). Parlar et al. reported that the
and foam; diversion of the foam itself is useless threshold ∆p value scales with the –1⁄2 power of
unless the acid following the foam is diverted as permeability, which suggests that at any given
well. Observations from laboratory corefloods ∆p, high-permeability zones would take large
include the following: amounts of fluid while low-permeability zones
would receive little. This finding considerably
– Trapping of gas by liquid injected following
affects strategy for foam diversion.
foam is incomplete, and the pressure gradient
declines nearly simultaneously throughout the – In the field, however, the entire diversion process
core to a lower, uniform value (Fig. 19-8) may occur within the period of decline in ∆p to its
(Parlar et al., 1995; Kibodeaux et al., 1994). lower, steady value. The rate of decline depends
There is a later decline in ∆p that starts near the on the rate of liquid injection. In Fig. 19-8, it takes
inlet and proceeds through the core. The second about 21⁄2 hr for ∆p to decline to the plateau, with a
decline is due to the dissolution of trapped gas liquid velocity of about 1 m/d. Parlar and cowork-
in injected liquid, which is undersaturated with ers found that for a velocity of 6 m/d (equivalent
gas at the high pressure of the core (Robert and to a pump rate of 0.05 bbl/min/10 ft at the well-
Mack, 1995; Zeilinger et al., 1995). The second bore), the plateau value is reached within 10 min.
decline may be unimportant in the field, owing In field application, injection of an entire acid slug
to the large number of pore volumes of liquid may last only 30 min. Thus the rate of decline in
required to dissolve the gas. In any event, it can ∆p may be as important as the lower, steady value.
be avoided by including a small amount of gas What controls this rate of decline is not yet clear.
with the acid. • Foam propagation rate
– The lower, steady pressure gradient observed Foam cannot propagate faster than the surfactant
after liquid follows foam is nearly independent advances into the porous medium, so surfactant
adsorption losses to the rock play an important
part in foam propagation. The latter can lag
400 behind surfactant propagation, however. Whether
350 Foam injection ends controlled by surfactant adsorption or by other
factors, foam propagation is faster in high-perme-
300
ability rock than lower permeability rock and

250 faster with a surfactant preflush than without
(Friedmann and Jensen, 1986; Kibodeaux et al.,
200
Liquid injection
1994; Parlar et al., 1995). Both effects help the
150 diversion process. Because most preflush enters
100
the higher permeability or less damaged intervals
Section 1 2 3 4 that are to be blocked with foam, a preflush helps
50 to place more foam in the layer to be blocked
0
(Zerhboub et al., 1994; Zhou and Rossen, 1994).
340 345 350 355 360 365 370 375 • Interactions between foam and acid
Pore volumes injected
As noted, most studies of foam diversion do not
Figure 19-8. Pressure evolution during liquid injection fol- include acid, on the assumption that the effects of
lowing foam (Zeilinger et al., 1995). acid and foam are separable. For carbonate rocks,

however, acid dissolves the rock matrix in which in cores with permeabilities between 100 and 1200 md.
foam resides, forming channels or wormholes Behenna (1995) also reported diversion with this tech-
through the rock (see Chapters 16 and 17). Two nique in two cores with permeabilities of 18 and 126 md.
effects of foam on the formation of wormholes are However, when strong foam was generated in both
crucial to the acid stimulation of carbonates. First, cores by the injection of larger foam slugs, no diversion
foam causes the formation of longer wormholes or even reverse effects were obtained.
with fewer branches, leading to more efficient Behenna (1995) also tested the effect of acid on
stimulation of a formation with less acid required foam diversion. A 5-mm thick wafer of Ohio sand-
(Bernardiner et al., 1992). Second, foam evidently stone was placed at the face of the high-permeability
fills and plugs wormholes, diverting acid to layers core (322-md Berea sandstone) to simulate a thin
with fewer or shorter wormholes. This leads to a layer of near-wellbore damage. Injection of a small
more even distribution of acid between intervals foam slug slightly increased the fraction of flow enter-
(Thompson and Gdanski, 1993). ing the “damaged” core. The subsequent acid stage,
however, directed most of the flow to the damaged
core by increasing substantially the permeability of
19-4.4. Foam diversion experiments the wafer while the permeability of the foamed
The first published foam diversion tests in connection undamaged core remained fairly constant. In this case
with acidizing used high-permeability sandpacks (20/40 diversion resulted from the combined effect of both
and 40/60 mesh) with a permeability ratio of 20 and foam and acid.
alternate injection of foaming solution and gas. Smith Laboratory experiments show that the keys to foam
et al. (1969) reported successful plugging of the high- diversion are placing more foam and (if possible)
permeability pack, even during the liquid stages fol- stronger foam in the higher permeability or less dam-
lowing the gas stages. Later, Burman and Hall (1986) aged intervals and keeping the gas trapped during sub-
used lower permeability cores with permeability ratios sequent liquid injection. As discussed previously, foam
between 1.6 and 3.8. They observed durable diversion itself is stronger in higher permeability rock, but ineffi-
with core permeabilities of 100 md and higher: equal cient trapping of gas during the stages following foam
flow in both cores was eventually obtained after several negates some of this effect. Other factors such as slower
foam stages, and diversion was maintained during the foam propagation caused by surfactant adsorption in
liquid stage. For experiments conducted with lower lower permeability zones can help achieve foam diver-
permeability cores (10 to 66 md), little or no improve- sion during the liquid stages following foam. Further-
ment was achieved in the flow profile during liquid more, if partial temporary diversion is achieved, the
stages. However, some diversion was obtained during effect of acid on damaged zones can combine with the
the foam stages, with better efficiency obtained with foam effect to complete the reversal in flow distribution
lower quality foam. between the damaged and undamaged zones after the
Zerhboub et al. (1994) and Parlar et al. (1995) used first exposure to a foam slug–acid sequence.
small preflush volumes of surfactant solution ahead of Zeilinger et al. (1995) showed, however, that a
the foam stage and limited the size of the foam slug to process that relies primarily on faster foam propaga-
a fraction of the pore volume. They obtained successful tion in higher permeability or undamaged rock may
diversion, which was maintained during the liquid work better in the linear flow geometry of laboratory
stages following the foam slugs. Zerhboub et al. tested corefloods than in radial flow in reservoirs. In essence,
sandpacks from 300 md to 20 d and reported better effi- in radial flow, faster propagation helps only if the
ciency with higher permeability cores. Applying a shut- slugs are small, because otherwise foam fills the cru-
in after the foam stage and increasing the foam-stage cial near-well region in all layers.
volume resulted in better diversion. A test performed In carbonate cores, foam diversion is coupled to the
with two cores of the same permeability, one saturated formation of wormholes. Thompson and Gdanski (1993)
with brine and the other containing residual oil, showed showed that foam can help equalize flow between two
that foam was able to impair temporarily the brine-satu- cores with wormholes of different lengths; evidently bulk
rated core and redirect flow into the oil-bearing core. foam fills and blocks the wormhole, diverting acid to the
Parlar et al. obtained diversion with the slug technique core with shorter or no wormholes.

19-4.5. Modeling and predicting effect sfoam,j replaces scake,j in Eq. 19-10. The foam
foam diversion skin effect is
IOR foam researchers use a variety of mathematical  k  r 
s foam , j =  j − 1 ln f , j  , (19-17)
models to describe foams (Kular et al., 1989; Fisher  k foam , j   rw 
et al., 1990). In principle, the most complete, but also
highly complex, population balance model (Friedmann where kj/kfoam,j is the mobility reduction resulting
et al., 1991; Falls et al., 1989; Kovscek and Radke, from foam (assumed for simplicity to affect only
1994) attempts to represent all mechanisms of non- permeability) in layer j and rf,j is the radial extent
Newtonian mobility and processes altering bubble size. of the foam bank in that layer. Foam advances in
There is no published application of this model to each layer at a constant volumetric rate that reflects
foams for diverting acid. permeability and the volume of preflush in that layer.
The much simpler approach of the fixed-pc* model Upon switching from foam to acid injection in a slug
(Rossen and Zhou, 1995; Zhou and Rossen, 1994, process, the mobility reduction factor kj/kfoam,j sud-
1995), focuses on the coalescence of foam as a func- denly takes a new, usually higher, constant value,
tion of capillary pressure pc or water saturation Sw. and rf,j stays constant during acid injection. An ana-
Combined with fractional-flow modeling, this foam lytical solution is obtained for constant-pressure
model predicts some important aspects of foam injection, and other cases, such as injection at a fixed
behavior: the benefits of a surfactant preflush, the overall injection rate, can be derived easily.
importance of compatibility of acid and foam slugs, This approach uses many simplifying assumptions,
foam diversion between layers differing in perme- of which two now appear to distort predictions of
ability and differences between diversion processes foam performance in the field: assuming mobility is
in linear and radial flow. In this approach, laboratory independent of flow rates and assuming instantaneous
corefloods are used to determine the rates of advance attainment of steady-state mobilities upon a change in
and mobility of foam during the injection of foam injection conditions. As noted previously, foams are
and subsequent injection of liquid (Kibodeaux et al., strongly non-Newtonian in the flow regime relevant
1994). Mobilities are assumed independent of flow to acid diversion, and the period of transition between
rates. These parameters are then used with equations high ∆p with foam and lower ∆p during liquid injec-
for radial flow such as Eq. 19-10 to predict diversion tion (Fig. 19-8) may be crucial to the diversion proc-
in a field application. The model identifies the keys ess. Work to extend this model to account for these
to effective foam diversion as the rate of foam prop- effects is ongoing.
agation, foam mobility during injection and foam The simulator described by Robert and Mack
mobility during liquid injection in each layer. (1995) and Bartko et al. (1996) does not make these
Hill and Rossen (1994) presented an even simpler assumptions and fits their laboratory corefloods well
version that gives essentially equivalent results for (Fig. 19-9). This approach is closer to full reservoir
many applications. In this model, the foam skin simulation: the computer solves a partial differential
equation for pressure and flow rates as functions of
Foam Surfactant Foam Surfactant
800 800
0–1 in. 0–1 in.

1–3 in. 1–3 in.
600 3–5 in. 600 3–5 in.
5–7 in. 5–7 in.
7–11 in. 7–11 in.
400 9–12 in. 400 9–11 in.
200 200
0 0
0 100 200 300 0 100 200 300
Figure 19-9. Comparison of simulation (right) with laboratory coreflood (actual data, left) (Robert and Mack, 1995).

position and time, interpolating mobilities from tables formation in porous media is a dynamic process in
of laboratory data. Local gas and liquid saturations are which bubbles (or lamellae) are constantly created
computed from mobilities and local foam flow condi- and broken. Surfactant adsorption on the rock sur-
tions using Eqs. 19-14 and 19-15. A mass balance is face, elasticity of lamellae and kinetics of foam
then used to compute the local flow rates for each formation are important properties that cannot be
phase from the variations of saturation. Both regimes measured by static half-life tests. Therefore, only
shown in Fig. 19-7—the low-flow-rate regime, in core flow tests can be used to select the most
which mobilities are roughly Newtonian at constant appropriate surfactant formulation for diversion.
quality and the constant-pressure-gradient regime at Laboratory experiments indicate that diversion
high flow rates—are included in the simulator foam- requires rapid formation and propagation of strong
mobility model. For liquid injection after a foam stage foam in higher permeability cores, with delayed
(i.e., the reimbibition stage), the simulator assumes a foam formation in the lower permeability zones.
sharp transition in gas mobility from complete plug- This suggests that surfactant adsorption is a criti-
ging to high mobility at the threshold pressure gradi- cal parameter for diversion. Oil present in the for-
ent. This tends to maintain the actual pressure gradient mation is another parameter to take into account
at the threshold value during liquid injection, as for surfactant selection: the presence of oil can
observed experimentally. An empirical correlation is impair significantly the foam strength or delay the
used to account for the delay of gas removal (and the onset of effective foam banks, especially in oil-
corresponding delay of pressure decrease) observed at wet formations.
low liquid rates. This is the most difficult part of the • Pumping strategy
corefloods to model accurately, and in some cases the
On the basis of laboratory experiments, Zerhboub
simulator tends to overestimate this delay slightly,
et al. (1994) proposed a set of guidelines for foam
especially in the downstream section of the cores.
diversion. The principle of the method consists of
Effects of surfactant adsorption, gas compressibility
pretreating the matrix to facilitate the formation of
and gas solubility are also included. In principle, there
strong foam with a minimum foam volume and
is no limit to the complexity of mobility behavior that
adding surfactant to acid stages to preserve the
can be incorporated in this approach.
foam bank stability. Experiments on sandstones and
carbonates show that effective and durable diver-
19-4.6. Application to field design sion can be obtained with repetitive slugs of foam
and acid. Indeed, once a damaged zone has been
Recent laboratory results have led to the development acidized, it must be plugged by foam to redirect
of guidelines for foam diversion (Gdanski, 1993; acid to zones that have not been treated. Further-
Zerhboub et al., 1994). Foam diversion has been used more, foam in place tends to undergo slow decay
increasingly in recent years, particularly in cases where (presumably because of gas dissolution), and peri-
other diversion techniques are impractical or less effi- odic regeneration of the foam bank is required.
cient such as the application of particulate diverters and Use of a mutual solvent is recommended ahead
ball sealers over large intervals. Foams also help in of the treatment to reduce interfacial tension and
cleanup and back-production of treating fluids. to help sweep oil from the near-wellbore region.
• Choice of surfactant For a well-defined water zone, treatment can be
started with a foam slug aimed at selectively plug-
Traditionally, half-life measurements have been
ging the water zone, while the oil prevents strong
used to rank the foaming performance of surfac-
foam formation in the hydrocarbon-bearing zone.
tants. In these static tests, foam is generated and
In that case, a mutual-solvent preflush is not bene-
allowed to decay slowly in a beaker under the
ficial. A gel-based water- or oil-specific diverter
action of liquid drainage and gas migration. The
can be also used in addition to foam to help plug
time for one-half of the liquid to drain out of the
water zones (Kennedy et al., 1992).
foam is called the foam half-life. These tests are
The size of foam slugs can be extrapolated from
useful for determining chemical compatibility and
laboratory studies. As with particulate diverters, the
identifying ineffective foams but not for optimiz-
design consists of pumping enough foam to plug
ing surfactant formulation (Rossen, 1996). Foam

higher injectivity zones while keeping the overall dient in the foam bank (Parlar et al., 1995). How-
injectivity high enough to perform the treatment in ever, little variation of foam behavior was observed
a reasonable time and ensure effective acid action. during the following liquid stage; therefore, diver-
Generally, the volume of the foam stages is gradu- sion efficiency is not expected to vary.
ally increased throughout the treatment to provide For long intervals, especially in horizontal
diversion over increasing interval lengths and wells, coiled tubing is often used to help place
replenish the degenerating foam pumped during treating fluid throughout the interval. Treatment is
prior stages. For carbonate acidizing, increasing the started at the bottom of the interval, and the coiled
volume of the foam stages accommodates the void tubing is pulled out while pumping. Foam stages
space created by rock dissolution during the previ- are pumped at regular intervals, usually of 30 to
ous acid stages. Gdanski (1993) recommended also 50 ft. This method has been shown to yield better
that the volume of the foam slugs be sufficiently results than simple bullheading (i.e., injection at
large to allow transport in the tubing without exces- the top of the interval). Analysis of stimulation
sive mixing with other stages. results for water injection wells with long open-
Zerhboub et al. (1994) found that better diver- hole intervals (e.g., 300 ft) in a carbonate forma-
sion is obtained on parallel sandpack cores if mod- tion showed that replacing the conventional bull-
erate acid rates are used. This phenomenon can be heading technique (with or without a particulate
explained by the behavior of foam during subse- diverter) with the use of coiled tubing and foam
quent liquid injection: mobility in the foam bank diversion doubled the average post-treatment
appears to be governed by a threshold ∆p for foam injectivity index while reducing the required acid
mobilization and therefore increases strongly with volume from 100 to 50 gal/ft (Ginest et al., 1993).
increasing pump rate (Parlar et al., 1995). At high Figure 19-11 is an example of a post-treatment
flow rates, foam directs a greater fraction of flow flow profile after acidizing with coiled tubing and
into foam-filled high-permeability intervals. Sim- foam diversion. It shows that acidizing opened the
ulation of field application shows that under certain entire interval to flow, which indicates successful
conditions diversion can depend strongly on the damage removal over the entire zone.
pumping rate (Fig. 19-10). Coiled tubing can also help in designing diversion
Foam quality in diverter slugs varies generally stages on the fly. It allows using downhole sensors to
from 60% to 80%. Experiments on low-permeabil- measure pressure or to get an estimate of the bottom-
ity carbonate samples showed better diversion effi- hole pressure from the pressure at the wellhead in the
ciency if higher quality foams (70% to 80%) were annulus between the production tubing and the coiled
used (Thompson and Gdanski, 1993). Single-core tubing. An increase of bottomhole pressure when
tests on Berea sandstones showed that increasing foam is at bottomhole indicates that foam is entering
foam quality from 65% to 80% at a constant liquid the formation, but this is not a sufficient indication to
rate resulted in a slight increase of the pressure gra- prove that diversion is taking place. A significant
Acid Foam Acid Foam Acid Acid Foam Acid Foam Acid
3.5 3.5
Layer 1 = 300 md Layer 1 = 300 md
3.0 3.0
Rate per layer (bbl/min)
Rate per layer (bbl/min)
Layer 2 = 100 md Layer 2 = 100 md

2.5 2.5
2.0 2.0
1.5 1.5
1.0 1.0
0.5 0.5
0 0
0 50 100 150 0 50 100 150
Volume (bbl) Volume (bbl)
Figure 19-10. Effect of acid injection rate on diversion efficiency (Robert and Mack, 1995). Left: high injection rate (no
diversion). Right: low injection rate (diversion).

tions to keep the balls in place. Their effectiveness is

1500
Prejob also limited by the roundness and smoothness of the
Postjob
perforation holes. Moreover, ball sealers are not effec-
tive in wells with a large number of perforations.
1000 Better efficiency can be gained by using buoyant
Rate (RB/D/ft)
ball sealers instead of conventional nonbuoyant ones

(Erbstoesser, 1980; Gabriel and Erbstoesser, 1984). As
500 shown in Fig. 19-12, when a buoyant ball sealer is
transported to the perforations, it either seals an upper
perforation or is carried to the lowest one that is
0
accepting fluid. However, because of its buoyancy, it
7000 7050 7100 7150 7200 7250 7300 does not remain in the quiescent fluid in the rathole.
Depth (ft) Contrary to chemical diverters, ball sealers seated
on the perforations facing a high-permeability inter-
Figure 19-11. Flow profile after acidizing with foam diver-
sion (Ginest et al., 1993). Prejob = 102,000 BWPD at
val effectively stop fluid entry and direct the entire
1500 psig; postjob = 608,000 BWPD at 1180 psig. flow to areas with lower injectivity. Figure 19-13
shows the injected flow distribution resulting from
reduction of pressure during the following acid stage this diversion technique.
can indicate that a new zone is acidized and that The density of buoyant ball sealers must be com-
diversion was obtained. However, laboratory experi- patible with the density of the carrying fluid to allow
ments show that a drop in injection pressure during transport downhole while prohibiting settling in the
the acid stage may also reflect foam displacement and rathole. Ball sealers are transported downhole as
decay (Parlar et al., 1995; Kibodeaux et al., 1994). long as the rising velocity does not exceed the fluid
Therefore, pressure evolution is more difficult to inter- velocity in the wellbore. The rising velocity ur of
pret for foam than for particulate diverters. spherical particles in a fluid is
Foam may be mixed with acid throughout the
4 ∆ρgd
treatment. Generally, lower qualities are used for ur = , (19-18)
foamed acid than for foam slugs. This method was 3 ρ f fd
first tried for acidizing gravel-pack formations
(Burman and Hall, 1986). Laboratory data indicate
that foamed acid provides better coverage than Flow
straight acid, even if inversion of the flow profile
(i.e., more flow into the damaged or low-permeabil-
ity zones) cannot be attained. However, this tech-
nique requires higher gas volumes and pumping
times than the slug method. Thompson and Gdanski
(1993) reported that foamed acid does not provide Conventional Buoyant ball sealer
sufficient diversion in high-porosity carbonates. density
ball sealer
Flow
19-5. Ball sealers

Originally introduced in 1956, ball sealers are small
spheres intended to seal perforations on the inside of
the casing. Added to treating fluids during stimulation,
they are carried downhole and seal the perforations
accepting the largest quantities of fluid. Although they
are widely used, conventional ball sealers often fail to Figure 19-12. Two ball sealer seating processes
(Erbstoesser, 1980). The dashed trajectory shows the con-
provide diversion. A sufficient rate must be available ventional density (nonbuoyant) ball sealer, and the solid
to maintain a differential pressure across the perfora- trajectory shows the 100%-efficient buoyant ball sealer.

100 To recorder
To pump truck
90 To test tank To pump truck

Flow rate distribution (%)
Oil
Casing
Tubing
Electrical cable
Openhole
10 Layer 1
Layer 2
Time Oil
Figure 19-13. Flow distribution using ball sealers in a two-

layer reservoir ( h1/ h2 = 1, k1/ k2 = 9). Perimeter nipple
where ∆ρ is the difference between the fluid and ball Acid

densities, d is the ball diameter, g is the acceleration of
gravity, rf is the fluid density, and fd is the drag coeffi-
cient. The drag coefficient depends on the Reynold’s
number NRe and can be obtained from the well-known
Moody diagram or from empirical expressions:
ur dρ f
N Re = , (19-19) Figure 19-14. Diversion in openhole using an “interface
µ locator” tool (Harrison, 1972).
where µ is the dynamic viscosity. A trial and error

technique is required to solve Eq. 19-18 rigorously
(Brown et al., 1963), because NRe depends on the without moving the completion. Although this is an
calculated velocity. However, in high-flow regimes, effective means of obtaining excellent control on cov-
and thus at large values of NRe (between 2,000 and erage, it is expensive and time consuming.
100,000), fd varies slightly about an average value of Frac baffles were invented in 1965 to provide eco-
0.44. If fd is assumed constant over this interval, then nomically attractive completions in multilayered
the rising velocity can be estimated readily (Gabriel reservoirs. In this method, shown in Fig. 19-15, one or
and Erbstoesser, 1984). more concentric baffle rings of different diameters are
designed to accept a “bomb.” Bombs of different
diameters are dropped to sit on the correct-diameter
19-6. Mechanical tools baffle and thus isolate the zone of treatment. However,
this technique limits the treatment to downcasing and
During the mid-1940s, downhole tools were used to requires accurate placement of the baffles when run-
locate the contact between a conductive fluid (acid) ning the casing.
and a nonconductive fluid (oil). Both fluids were Retrievable packers have also been developed that
pumped simultaneously to maintain the interface at use various techniques for setting and retrieving.
a constant level. This allowed treatment of a selected Tension-set packers are particularly suitable for matrix
portion of the reservoir. A schematic of this procedure acidizing operations. They are initially set by pulling
is given in Fig. 19-14. In 1950, the development of tension on the tubing and held in place by the pressure
hydraulically operated inflatable packers led to differential between the tubing and the annulus above
improved diverting techniques. When these packers the packer.
are used as a straddle tool, it is possible to treat three Conventional mechanical techniques are efficient
intervals (below, between and above the packers) at controlling the placement of acid. However, they

Fluid
flow
Frac baffle
Cast-iron Frac baffle

bomb
.....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
.....
Figure 19-15. Frac baffle diverting technique (Harrison, ..
.. .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... ..
..
1972). .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
.....
.. .....
.. ..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... ..
..
.....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
.....
..
.. .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... ..
..
.....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
.....
are cumbersome and generally require a rig on site. ..
.. .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... ..
..
They are also expensive and time consuming. .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
.....
New techniques have been developed to convey
packers with coiled tubing. Inflatable straddle pack-
ers can provide zonal isolation during matrix treat-
ment. Before packer setting, the treating fluid can be
circulated down the coiled tubing above the packers.
Setting is obtained by picking up on the coiled tub-
ing and applying pump pressure to inflate the packer
elements. Treating fluid can then be directed to the
interval between the packers by slacking off weight
on the tool string (Fig. 19-16). Picking up over the
string weight and then slacking off closes the injec-
tion port and reopens the circulating port above the Figure 19-16. Treating between the packers (Milne, 1991).
packers. Finally, sustained overpull allows the packer
elements to deflate, and the packer can be moved to
another zone (Milne, 1991).
fluid over the entire open interval is more difficult to
achieve. Furthermore, the limits of corrosion-inhibitor
efficiency, as well as other practical considerations,
19-7. Horizontal wells restrict the treatment duration and therefore require
pumping smaller volumes per length of open interval,
19-7.1. Optimal treatment which makes the issue of placement control even
Producing sections in horizontal wells are typically more critical. Better results are generally reported
between 1 and 2 orders of magnitude longer than in when a diversion technique is used than when acid is
vertical wells. Therefore, even distribution of treating simply bullheaded via the production tubing. In sever-

al cases, production logs performed after acidizing treatment configurations. Simulations were performed
without the use of any diversion technique show that with a 2000-ft well with 12-in. damage penetration
only a small portion of the well was open to flow, and a permeability reduction of 90% in the damaged
generally the area nearest to the vertical section (Frick zone. The original vertical and horizontal permeabili-
and Economides, 1991; Thomas and Milne, 1995). ties were 90 and 10 md, respectively. A total volume
Efforts have been made to determine the optimum of 100,000 gal was pumped, corresponding to 50 gal/ft
treatment design, taking into account the limitation for full coverage and 125 gal/ft for 40% coverage. For
of the total volume that can be pumped. Economides partial coverage, the treated area was divided into five
and Frick (1992) recommended adjusting the injec- equally spaced intervals. In the particular case shown
tion volumes according to the severity of damage. in Fig. 19-18, optimum treatment would be achieved
For instance, in the case of completion-induced dam- with partial coverage. However, these simulation
age, cone-shaped damage invasion is expected as results must be interpreted with caution, as they
damage severity increases with the time of exposure depend on the total volume required for total damage
to drilling mud, which decreases with increasing removal. For example, if only 50 gal/ft were required
measured depth (Fig. 19-17). In that case, a to achieve zero skin effect, then full coverage would
“tapered” injection would be optimal, with more acid obviously become the optimal configuration.
pumped near the vertical section of the well. If com- The total volume used in horizontal well treatments
plete damage removal requires impractical acid vol- is usually limited by practical constraints not directly
umes, partial damage removal throughout the open linked to the acidizing itself, such as pumping time
interval should be targeted, leaving a uniform dam- (especially if coiled tubing is used). However, because
age collar around the wellbore. of the relatively low cost of matrix acidizing com-
Da Motta et al. (1994) suggested that in some cases pared with the return in terms of a production
treatment efficiency can be improved by selective increase, the highest net present value (NPV, or the
acidizing, where some intervals in the horizontal sec- difference between revenue from additional produc-
tion of the well are intentionally left unperforated or tion and the treatment cost) is usually obtained when
untreated. For example, Fig. 19-18 shows the normal- the entire interval is treated, even if this entails large
ized productivity index (ratio of the productivity index acid volumes (Economides and Frick, 1992).
to undamaged productivity index) for three different
rs
Vertical section
dV
dx
asV, max asV, min
x
asH, min
asH, max
Figure 19-17. Cone-shaped damage (Frick and Economides, 1991). rs = radius of the damaged section, L = length, as =
axis of the damaged ellipse with the subscripts V and H denoting vertical and horizontal dimensions and min and max
denoting minimum and maximum values, respectively.

Normalized productivity index 1.0 3.5
0.9
3.0
0.8
2.5
0.7
Coiled tubing withdrawal (ft/min)

0.6 2.0
20% treated
40% treated
Fully treated
0.5
0 50,000 100,000 150,000 200,000
1.5
Acid volume (gal)
2
Figure 19-18. Normalized productivity index for different bb
l/m
1.0 in
fractions of the treated interval (da Motta et al., 1994).
1.5
bb
l/m
in
1b
0.5 bl/m
in
19-7.2. Placement techniques

Different placement techniques can be used to treat 0
horizontal wells. As most completions are openhole 0 25 50 75 100 125 150
Volume injected (gal/ft)
or with slotted liner, ball sealers are generally not
applicable. Furthermore, conventional ball sealers Figure 19-19. Coiled tubing withdrawal rate (Economides
are not recommended in horizontal wells, as seating et al., 1991).
requires a minimum velocity per perforation.
Most recommended techniques include the use of
coiled tubing. A standard practice consists of starting
6.0
the treatment with coiled tubing at bottomhole, then
slowly pulling the coiled tubing out while acid is
4.8
injected, with the rate of coiled tubing retrieval set
according to the desired volumetric coverage (Fig. 19-19). 100 gal/ft
50 gal/ft
3.6
This technique provides contact of the entire interval
Skin effect
with the treating fluid. A diverter stage in the form of 100 gal/ft
2.4
foam or particulate agents is usually pumped at regular
intervals to avoid the formation of thief zones.
1.2
Figure 19-20 shows the rate of skin effect reduction Without coiled tubing
with and without the use of coiled tubing as determined With coiled tubing
0
by simulation, assuming a 2000-ft horizontal section 0 12 24 36 48 60
with an initial skin effect of 6, damage penetration of Injection time (hr)
1.5 ft and damage consisting solely of clays. Clay dis-
solution with 12% HCl–3% HF was simulated, and the Figure 19-20. Skin effect evolution versus time with and
without coiled tubing (Economides et al., 1991).
evolution of skin effect with time calculated from the
permeability profile around the wellbore. The injection
rate was equal to 1.5 bbl/min, corresponding to a total Economides et al. (1991) recommended pumping
volume of 100 gal/ft over 52 hr. Without coiled tubing, a nonreactive fluid through the annulus between the
it was found that only the first half of the interval could production tubing and coiled tubing to provide back-
be stimulated, whereas coiled tubing provided good pressure and force the treating fluid to enter the res-
coverage accompanied with better skin effect reduction. ervoir next to or below the coiled tubing end. For a

tapered treatment, the rate of coiled tubing retrieval Coiled tubing can also be used to convey inflatable
should be progressively decreased to follow the pat- straddle packers in cased completions or completions
tern of damage severity and gradually increase the including external casing packers. The drawback of
volumetric coverage. this technique is that the interval between packers is
Tambini (1992) showed that the method developed limited to 30 ft, which entails many setting and
by Paccaloni for vertical wells can be adapted to hori- unsetting operations. When zonal segmentation is
zontal completions, mainly by adding a friction reduc- provided by slotted liner with external packers, each
er to allow pumping at the maximum rate under segment (500 ft long on average) must be treated
fracturing pressure. The technique consists of spotting with retrievable packers set opposite to the casing
acid (one-third of the total volume) throughout the packers to prevent treating fluid migration along the
entire interval by pumping through coiled tubing only wellbore behind the slotted liner (Milne, 1991).
(still starting at total depth and pumping while the
coiled tubing is retrieved). Then acid is pumped
through both the annulus and coiled tubing at the 19-8. Conclusions
maximum rate allowed by the surface equipment and
reservoir. Good results are reported with this tech- Pumping strategy is one of the major issues for the
nique for production interval lengths up to 1500 ft. success of matrix acidizing. Improving the place-
Pressure analysis during the treatment and production ment of treating fluids also addresses increasing con-
logs run on cemented completions indicate that treat- cerns for environmental protection and cost control.
ment efficiency depends heavily on the injection rate Four placement techniques are discussed in this
during the bullheading phase, with the best results chapter: particulate diverter, foam, mechanical tools
reported for rates as high as 5 to 6 bbl/min/100 ft in and ball sealers. The design of particulate diverter or
15-md formations (Fig. 19-21). However, this tech- bridging agents requires checking compatibility with
nique requires additional pumping equipment and the formation and carrying fluids. Particle size must
therefore can be more costly than other methods. be adapted to the pore-throat distribution of the for-
Moreover, in some cases the required rates cannot be mation, and the diverter must be soluble in the reser-
achieved because of well and surface equipment limi- voir fluid to help flowback. Compatibility with the
tations. Finally, as mentioned earlier, the method does carrying fluid is required to ensure proper dispersion
not provide fluid placement optimization, which is of the particles and prevent dissolution or reaction
critical for long intervals. with other additives.
Foam diversion has been used with success in long
intervals. Qualities from 65% to 70% are commonly
100 used. Current practice involves pumping surfactant-
laden preflush to saturate high-permeability zones and
ensure foam stability in these zones. Better foam dura-
80
bility is obtained if surfactant is added to all stages,
Completion efficiency (%)
even at moderate concentrations. Laboratory tests and

60 computer simulations indicate that better diversion is
obtained if moderate pump rates are applied.
40 Contrary to the other techniques, mechanical tools
and ball sealers allow complete shut-off of a portion
of the wellbore. Mechanical tools provide the best
20
placement control but require additional equipment
and costly operations. Ball sealers are normally lim-
0 ited to a single stage in cased completions and can-
0 1 2 3 4 5 6 not be used for long intervals.
Injection rate (bbl/min/100 ft) One other technique has emerged for carbonate
acidizing: self-diverting acid. This technique is pre-
Figure 19-21. Completion efficiency versus injection rate sented in Chapter 17. Field and laboratory data indi-
(Tambini, 1992).
cate that it is highly effective for plugging high-
injectivity streaks created by acid.

The application of coiled tubing in matrix stimula- Acknowledgments

tion has improved results, especially for the stimula-
tion of long intervals. Combining most of the place- Some of the material in this chapter was written by
ment techniques (i.e., particulate diverter, foam, L. P. Prouvost and N. Doerler and previously pub-
packers) with the use of coiled tubing increases their lished in second edition of this volume. The authors
efficiency. In long intervals, as in horizontal wells, also thank O. Liétard and R. Thomas for reviewing
coiled tubing is run to the well bottom and then this chapter.
withdrawn while pumping acid. Diverter stages are
pumped between acid stages. This technique allows
spotting acid across the entire interval and minimizes
the effect of thief zones.

◆
Matrix Stimulation
Treatment Evaluation
Carl T. Montgomery, ARCO Exploration and Production Technology
20-1. Introduction 20-2. Derivation of bottomhole

As for any stimulation operation, it is important to parameters from wellhead
evaluate the effectiveness of a matrix treatment. In measurements
Chapter 12, the methodology of evaluating hydraulic
fracturing through pretreatment and post-treatment In many matrix stimulation treatments, wellhead
well tests is outlined. For matrix stimulation, such a pressures and injection rates are recorded. Ideally,
comprehensive approach is rarely justified economi- these variables should be measured bottomhole.
cally. Usually, the effectiveness is gauged by apparent However, the hostile nature of the stimulation fluids
increases in the productivity index, without the bene- prevents the use of downhole pressure gauges and
fits of a post-treatment test. Clearly, increases in the flowmeters. Thus, bottomhole parameters are often
productivity (or injectivity) index, although desirable, extrapolated from measurements made at the surface
may not provide an adequate picture of optimum well by simple mechanical energy balance applications
conditions. and estimates of the hydrostatic and friction pressure
Estimation of the well skin effect and identification losses. When stimulation fluids are injected through
of its individual components are covered in Chapter 1. a conduit such as coiled tubing or rigid tubing, mea-
The total skin effect can be determined from a well surement of the bottomhole injection pressure can be
test, and test analysis is outlined in Chapter 2. Mech- done in the “backside” (i.e., at the annulus formed
anical skin effects (such as from partial completion and between the injection tubing and other well tubulars).
inadequate perforations) may often overwhelm the Echometers have also been employed to measure the
damage skin component. An appropriate production level of backside fluids if the wellhead pressure is
engineering analysis should enable the identification below hydrostatic.
and separation of these skin effects. Only the portion
of the skin effect resulting from damage can be re-
moved by a matrix treatment, which should reduce 20-3. Monitoring skin effect
the total skin effect by that amount. Furthermore, to evolution during treatment
be cost effective, the injected volume and the pumping
Different attempts have been made to interpret
time should be minimized.
recorded wellhead pressures (or derived bottomhole
Several attempts have been made to evaluate the
pressures) and injection rates in terms of progress of
effectiveness of a remedial matrix treatment by moni-
the remedial treatment. Several techniques have been
toring evolution of the skin effect in real time. Mon-
used to derive the evolution of the skin effect from
itoring to evaluate whether an adequate fluid volume
these records.
has been injected indicates whether the treatment
requires modification and helps to improve future
designs in similar situations. 20-3.1. McLeod and Coulter technique
Techniques for real-time monitoring have been
introduced by McLeod and Coulter (1969) and Pac- Each stage of injection or shut-in during the treatment
caloni (1979a, 1979b) and further augmented by is considered a short, individual well test (McLeod
Prouvost and Economides (1987, 1989), Behenna and Coulter, 1969). The transient reservoir pressure
(1994) and Hill and Zhu (1994). They are described response to the injection of fluids is analyzed and
in this chapter. interpreted to determine the skin effect and the for-

mation transmissibility. This is accomplished using Paccaloni (1979a) used an arbitrary constant value for
conventional pressure transient analysis methods, an acid bank radius rb and the concept of the effective
namely the solution to the diffusivity equation and the wellbore radius rw′ = rwe–s to transform Eq. 20-1 into
superposition principle (see Chapters 1 and 2.)
141.2 qi Bµ  rb 
Although the results are presented not in terms of piw − pe =  ln  . (20-2)
skin effect but in terms of the permeability and the kh  rw′ 
radius of the damaged zone (i.e., the short-term nature Assuming that the permeability-thickness product
of the test has a limited radius of investigation), the kh of the reservoir is known, the bottomhole pres-
objective is essentially the same: the skin effect can sures, corresponding to the injection rate, can be pre-
be related to the properties of the damaged zone. For dicted for any value of the skin effect. The wellhead
example, if the well is initially injected at a constant injection pressures are then derived from the bottom-
rate, a graph of the pressure response plotted against the hole pressures, and at any time, a match of the mea-
logarithm of elapsed time since injection started yields sured wellhead pressures with the ones corresponding
the permeability in both the damaged and undamaged to different values of the skin effect yields the actual
zones and the radial extent of the damaged zone. With skin effect at that time.
a little modification, the same information can be Paccaloni used the concept of a damage ratio DR,
obtained from pressure falloffs when injection stops. which is simply the rewriting of Eq. 20-1 under ideal
The technique has several drawbacks: (s = 0) and real (s ≠ 0) conditions, and obtained the
ratio of the productivity indices:
• The analysis of pressure transients is valid only if
the skin effect factor does not change while a set of Jideal ln(re rw ) + s
DR = = . (20-3)
pressure data for one particular interpretation is col- Jreal ln(re rw )
lected. This is not the case when reactive fluids are
injected into the formation to remove damage. • Example application of the Paccaloni method
Hence, to be correct, this method requires the injec- Figure 20-1 depicts a series of damage ratio curves
tion of a slug of inert fluid each time the damage for a range of injection rates and calculated well-
removal is assessed. This is not usually practical. head pressures. Table 20-1 lists the well and reser-
• No continuous assessment of the evolution of the voir variables used for this construction.
treatment is possible. Skin effect measurements are The Paccaloni method (1979a, 1979b) implies
allowed only at discrete times during the course of that as acid is injected, superimposition of the mea-
the treatment, usually before and after the treat- sured wellhead pressure on the graph indicates the
ment. Thus, real-time analysis is not feasible. progress of the stimulation. This is shown by the
dashed line on the figure.
Among the assumptions listed in the Paccaloni
20-3.2. Paccaloni technique method, the “steady-state” assumption is one that
may cause problems and, potentially, errors.
A second method, presented by Paccaloni (1979a, Although a steady-state relationship is assumed,
1979b), uses instantaneous pressure and rate values transient behavior is in effect for a time duration
to compute the skin effect at any given time during far exceeding the injection test. For a radial reser-
the treatment. This method is based on the equation voir, for example, the time to a pseudosteady-state
for steady-state, single-phase, radial and horizontal condition is
flow in the reservoir: 1000φµct re2
t pss ≈ , (20-4)
141.2 qi Bµ  re  k
piw − pe =  ln + s , (20-1)
kh  rw  where ct is the total compressibility. For a typical
oil reservoir (e.g., φ = 0.25, µ = 1 cp, ct = 5 psi–1,
where piw is the bottomhole injection pressure, pe is the
k = 50 md and 80-acre spacing or re ≈ 1000 ft), the
reservoir pressure, qi is the injection rate, B is the for-
time to pseudosteady state (which may be perceived
mation volume factor, µ is the viscosity, k is the perme-
as a succession of steady states) is more than 50 hr,
ability, h is the reservoir thickness, re is the reservoir
which is significantly longer than the normal pump-
radius, rw is the wellbore radius, and s is the skin effect.
ing times for a typical matrix acidizing job.
20-2 Matrix Stimulation Treatment Evaluation

5000
DR = 10.0
3.3
4500
5.0
ring
fractu
avoid
=
DR =
ssure to
DR
ad pre
4000 m wellhe
Maximu
Wellhead injection pressure, pit (psia)
3500
rs
ed
3000
mete
predict
.5
=2
.0
7
.
4
para
1.
2500 =1
=2
=
DR
DR 25 b
Initial
l jo
DR
DR
1.
ping
= sfu
es
2000 DR .1 cc
pum
=1 su
ully
DR af
.0 or
1500 =1 sf
ter es
DR e loss
ram ion
pa Frict
1000 ping
m
pu
al
Fin
500
0
1 2 3 4 5 6 7 8 9
Injection rate, qi (bbl ⁄min)
–989
Fluid level (ft)
Figure 20-1. Job control chart with the damage ratio DR plotted as a parameter. The measured injection rates and injection
pressures are recorded on the chart (Paccaloni, 1979b).
Table 20-1. Well and fluid data used for Fig. 20-1 1625
(Paccaloni, 1979a).
Depth, H 13,540 ft
Density, ρ 8.9 lbm/gal [66.6 lbm/ft3]
1500
Viscosity, µ 0.7 cp
Bottomhole pressure (psia)
Reservoir pressure, pe 5805 psi

Hydrostatic pressure drop, ∆ph 6262 psi
Reservoir thickness, h 30 ft 1375
Permeability, k 35 md
Wellbore radius, rw 0.25 ft
Reservoir radius, re 800 ft
1250
Tubing 1 ID 2.992 in.
Tubing 1 length 12,750 ft
Tubing 2 ID 2.441 in.
Tubing 2 length 750 ft 1125
Acid bank radius, rb 4 ft 0 12.5 25.0 37.5 50.0 62.5
Time (min)
Figure 20-2 is a graph comparing bottomhole Figure 20-2. Comparison of calculated steady-state and
pressure evolution using steady-state and transient transient bottomhole pressure responses.
models. The simulations were generated for qi =
0.5 bbl/min, kh = 1000 md-ft, φ = 0.2, ct = 1.5 × 10–5 was used. It is obvious that the bottomhole transient
psi–1 and rw = 0.35 ft. The initial skin effect is equal pressure evolution could lead to errors. The pressure
to 10. For the Paccaloni (1979b) method, rb = 3 ft departure between the two curves in Fig. 20-2 could

be identified as the result of an additional skin effect where s(t) is the skin effect at time t, and q(t), B(t) and
where such a skin effect does not exist. For the µ(t) are the injection rate, formation volume factor
example in Fig. 20-2, after about 1 hr of injection and viscosity of each slug at time t, respectively.
and a pressure departure ∆pdeparture equal to 400 psi, The method can be executed in real time. The
this additional false skin effect is equal to “measured” bottomhole pressure can be calculated
kh∆pdeparture from measured wellhead values as shown earlier.
∆s = , (20-5) Hence, a skin effect evolution with time can be
141.2 qi Bµ obtained using Eq. 20-6.
a value equal to approximately 4. The practical
implication is that more acid may be employed to
eradicate this seemingly present skin effect when, 20-4.2. Determining reservoir characteristics
in fact, the pressure difference is due to easily before treatment
explainable transient phenomena. Simulation of the pressure response, described in the
previous section, requires a good knowledge of the
reservoir and the initial value of the skin effect. Even
20-4. Prouvost and Economides if a well test was performed on the well earlier, the
method following procedures are recommended to alleviate
any changes brought about by time. They apply mainly
A technique presented by Prouvost and Economides to oil producers and water injectors and are based on
(1987, 1989) enables continuous calculation of the an injection/falloff test, in which native reservoir
skin effect factor during the course of the treatment fluids in the wellbore are displaced and injected into
and accounts for transient response. The technique is the formation. There are three steps:
based on a continuous comparison of measured and
simulated pressures. The diminishing difference is 1. The reservoir fluid contained in the tubular string
attributed to the diminishing skin effect. A good reser- before treatment is partly or totally displaced into
voir description is presumed, including the type of the reservoir at a matrix rate (i.e., at a rate that no
model and well and reservoir variables. Section 20-4.2 fracture is initiated in the reservoir). To save pump-
presents an associated methodology to obtain ing time, this fluid can be displaced by the first
unknown variables just prior to the treatment. treating fluid.
2. The injection is stopped before any foreign fluid is
injected into the formation. The evolution of pres-
20-4.1. Deriving skin effect during treatment sure with time during this period is recorded typi-
The Prouvost and Economides (1987, 1989) method cally for a duration of 1 hr, depending on the
simulates the reservoir pressure response to the injec- reservoir parameters, until valuable information on
tion of fluids using the sequence of flow rates mea- the reservoir can be obtained using the appropriate
sured during job execution. Each simulation is pressure testing methods extensively described in
performed with a constant skin effect so. the literature and Chapter 2.
At any time, the difference between the simulated 3. Analysis of the pressure falloff data yields a model
pressure response psim(t, so) and the measured value of reservoir behavior and the associated well and
pmeas(t) is interpreted as the result of the difference reservoir variables, such as the permeability-thick-
between the actual skin effect value and the value used ness product and especially the initial value of the
for the simulation (so). The latter can be zero or some skin effect.
other finite value for a known nondamage skin effect. In the case of layered reservoirs and the absence of
Because all other effects that influence the pressure more detailed downhole measurements, this technique
response are accounted for in the simulation, the dif- results only in an averaged skin effect factor. Never-
ference between the simulated and actual pressure theless, this remains a valuable piece of information,
responses is attributed to the changing skin effect: because variation in the average skin effect during treat-
s(t ) = so +
kh
141.2 q(t ) B(t )µ(t )
[ ]
pmeas (t ) − psim (t, so ) , ment can be interpreted as the progress of damage
removal by the stimulation fluid. For example, flat-
(20-6)

tening of the s versus t curve at a nonzero level could ods of rapidly changing skin effect and has the effect
indicate poor fluid placement (or, of course, complete of smoothing the effects of noisy data. The slope of
damage removal in a single zone). Knowledge of the the derivative is useful in quantifying the rate at which
type of behavior and parameters enables computing the formation damage is removed. The steeper the slope,
pressure response at the wellbore to any injection his- the faster the rate of change.
tory, as described in Section 20-4.1.
The technique can be generalized by varying the
injection rate instead of just stopping the injection, as 20-6. Inverse injectivity
described in the second step of the procedure. Although
it can introduce some error from computation of the
diagnostic plot
friction pressure drop, this modified procedure is of Hill and Zhu (1994) proposed a technique that is as sim-
interest for depleted reservoirs (i.e., when reservoir ple to use as the Paccaloni (1979a, 1979b) method and
pressure is not sufficient to balance the head of a col- also takes into account the effects of transient flow as
umn of fluid up to the surface). suggested by Prouvost and Economides (1987, 1989).
An example calculation of the Prouvost and The method utilizes an expression derived from transient
Economides method is presented in Sidebar 20A. analysis methods extensively described in the literature
(Earlougher, 1977):
pi − pwf N  q −q
( ) 
20-5. Behenna method
qN
= m′∑ 
j =1 
j
qN
j −1
( )
log t − t j −1  + b ′, (20-7)

A technique that extends the Prouvost and Economides 
(1987, 1989) approach was provided by Behenna (1994). where pi is the initial reservoir pressure, pwf is the bot-
He accounted for the effects of diverting agents by sub- tomhole flowing pressure, and
tracting the pressure drop created by the diverting agent 162.6 Bµ
from the bottomhole pressure used in the skin effect cal- m′ = (20-8)
kh
culation. The pseudoskin effect caused by the placement
of diverting agent slugs corrects the observed pressure   k  
response by an amount equal to the pressure increase b ′ = m ′ log 2
− 3.23 + 0.87s  . (20-9)
  φµct rw  
caused by the diverter placement. Figure 20-3 from
Behenna’s paper shows an example application. An From Eq. 20-7, a plot is made of inverse injectivity
assumption made with this technique is that all subse- versus the superposition time function ∆tsup, which is
quent response is due to removal of damage and not defined as
inadvertent removal of diverter.
(q − q j −1 )log t − t .
( )
N
∆tsup = ∑
j
A second extension that the Behenna method offers (20-10)
j −1
is use of the derivative of the skin effect with time as j =1 qN
a diagnostic aid. The derivative plot accentuates peri- With all parameters remaining constant, a straight line
with a slope m′ and intercept b′ can be produced. To
Pseudoskin effect utilize the technique, a series of straight lines with dif-
Skin effect (minus diverter effect)
1 2
ferent skin effects at varying rates and times is pro-
Increasing skin effect
3 duced before treatment.

1
2 3 An example application of the Hill and Zhu method
1 2 is presented in Sidebar 20B.
3
Fluid at perforation
1. 15% HCI 20-7. Limitations of matrix treat-
2. 7.5% HCl–1.5% HF
3. 15% HCI + OSR diverting agent ment evaluation techniques
Elapsed time Coupling the calculated bottomhole pressure (based
on measured variables) and the anticipated reservoir
Figure 20-3. Skin effect evolution with diverters (Behenna,
1994). OSR = oil-soluble resin. response leads to calculation of the reservoir response.

20A. Example calculation of the Prouvost and Economides method
In the injection profile in Fig. 20A-1 the rate was maintained

at approximately 1 bbl/min and then dropped to 0.25 bbl/min 2125
after approximately 1.5 hr of injection. The “measured” bot- = Simulated pressure
tomhole pressure that appears as the top curve in Fig. 20A-2 = Measured pressure
is based on the recorded wellhead pressures and the calcu-
lated bottomhole pressures using the procedure outlined in
Section 20-2. 2000
Underneath this curve is the simulated expected pressure
response, using the injection profile of Fig. 20A-1 as an input
Bottomhole pressure (psia)

and a skin effect equal to zero. For this well, the permeability-
thickness product kh = ~5000 md-ft, the wellbore storage con- 1875
stant C = 2.6 × 10–3 bbl/psi, and the porosity φ = 0.22. As
usual, both the formation volume factor B and the viscosity
µ are taken as equal to unity (water). As outlined earlier, the
Prouvost and Economides (1987, 1989) method attributes the
departure between the two curves in Fig. 20A-2 to the skin 1750
effect. As can be seen, the departure diminishes as injection
progresses, and in spite of the reduced rate, after 1.5 hr the
two curves coincide and show the effectiveness of the job.
The associated skin effect evolution is shown in Fig. 20A-3.
The bottom curve is the predicted skin effect using the 1625
Prouvost and Economides technique. The initial value is pro-
vided by the injection/falloff test described in Section 20-4.2.
The skin effect reduces smoothly during the acid injection and
approaches a zero value after approximately 2 hr of opera-
1500
tion. Above this curve is the skin effect curve calculated by
0 0.4 0.8 1.2 1.6 2.0
the Paccaloni (1979a, 1979b) technique. A continuous over-
estimation is shown, as explained earlier. This is due to the Time (hr)
misidentification of a certain amount of pressure transients
as caused by an additional skin effect. Figure 20A-2. Calculated bottomhole pressure (based
An interesting observation is what happens when the flow on measured well pressure) and simulated pressure
rate is reduced. The steady-state relationship of the Paccaloni
response with zero skin effect.
method requires an analogous reduction in the pressure drop.
Because this effect is not forthcoming, a substantial increase
(usually unrealistic) in the calculated skin effect is obtained.
Such an increase is easy to surmise from Eq. 20-5 and is
shown graphically in Fig. 20A-3.
30
2000
1600 20
Surface injection rate (B/D)
1200
Skin effect
Paccaloni
10
800
0
400 Prouvost and Economides
0 –10
0 0.4 0.8 1.2 1.6 2.0 0 0.4 0.8 1.2 1.6 2.0
Time (hr) Time (hr)
Figure 20A-1. Example injection rate profile. Figure 20A-3. Example skin effect evolution calculation.

20B. Example application of the Hill and Zhu method
An example of the Hill and Zhu (1994) method is shown in Fig. 20B-1 from the data in Tables 20B-1 and 20B-2 (Prouvost and
Economides, 1989). As the skin effect changes during an acid treatment, ∆p/qi shifts from one line to another. The changing skin
effect is monitored by comparing ∆p/qi versus the ∆tsup response. As with the Paccaloni (1979a, 1979b) technique, this job control
graph is used in the field to monitor job performance. The advantage of this technique in comparison with the Paccaloni technique
is that it includes transient effects caused by injection rate changes.
900
800
700
600
s = 40
Inverse injectivity
500
s = 30
400
s = 20
300
200 s = 10
s=5
100 s=0
0
–1.50 –0.10 –0.50 0 0.50 1.00
tsup
Figure 20B-1. Inverse injectivity plot using the Hill and Zhu (1994) technique and data from Tables 20B-1 and 20B-2.
Table 20B-1. Well parameters used in example Table 20B-2. Injection data for example
application of the Hill and Zhu (1994) method application of the Hill and Zhu (1994) method
(Prouvost and Economides, 1989). (Prouvost and Economides, 1989).
Parameter Value Stage Fluid Time Time qi pit

(hr) (min) (bbl/min) (psi)
k 80 md
1 Water 0.15 9 2 4500
h 253 ft
2 Water 0.3 18 2 4480
µ 1 cp 3 Water 0.45 27 2 4400
rw 0.51 ft 4 Xylene 0.6 36 2 4390
5 Xylene 0.75 45 2.5 4350
φ 0.2
6 Xylene 0.9 54 2.4 4270
B 1.2 RB/STB 7 Xylene 1.05 63 2.7 4250
ct 10–5 psi–1 8 15% HCl 1.2 72 2.5 4200
9 15% HCl 1.35 81 3.2 3970
pi 2750 psi 10 3% HCl–12% HF 1.5 90 3.7 3800
qi 2 bbl/min 11 3% HCl–12% HF 1.65 99 4.1 3750
12 3% HCl–12% HF 1.8 108 4.4 3670
so 40
13 3% HCl–12% HF 1.9 114 4.2 3600
14 3% HCl–12% HF 2.1 126 0.9 2900
15 3% HCl–12% HF 2.25 135 5.6 3730
16 3% HCl–12% HF 2.4 144 5.6 3750
17 3% HCl–12% HF 2.55 153 5.6 3650
18 3% HCl–12% HF 2.7 162 5.7 3600
19 3% HCl–12% HF 2.85 171 5.8 3400

However, removal of damage by the stimulation fluid of the pressure response that should be isolated for
is not the only reason why the bottomhole pressure a proper evaluation of the stimulation treatment.
evolves during the course of a matrix stimulation Compared with this physical analysis, some of the
treatment. It is important that the engineer is aware of techniques presented in this chapter pose additional
the different components that can affect the reservoir limitations:
pressure response.
• The McLeod and Coulter (1969) method properly
• Reservoir response to changing injection rate acknowledges the importance of transient reservoir
When fluid is injected at a constant rate into the pressure response, but it fails to provide a continu-
reservoir, a pressure transient response develops ous, real-time measurement of the skin effect. It
that at early injection times is dominated by well- implicitly assumes that while one measurement
bore storage and skin effects. Later, the infinite- is taken the evolution of the skin effect is frozen.
acting radial flow regime prevails, and usually for • Although the Paccaloni (1979a, 1979b) method
long injection times the effects of the outer reser- enables continuous, real-time measurement of the
voir boundaries are not felt. The same series of skin effect for the duration of the treatment, it is
flow patterns develops when the rate is varied, and limited by not properly accounting for transient
the overall response is obtained through the super- phenomena. As a result, all transient phenomena
position principle. This is described in detail by occurring in the reservoir are interpreted as changes
Earlougher (1977). in the skin effect.
• Different viscosities of the various treating fluids
During a typical stimulation treatment, several dif-
ferent fluids are pumped in sequence. Because 20-8. Treatment response diagnosis
these fluids have different rheologies and viscosi-
ties at reservoir conditions, the pressure response Montgomery et al. (1995) described techniques that
is affected. can be used during a treatment to provide several lev-
els of diagnostics and interpretation. Figure 20-4 pro-
• Temporary skin effects created by diverters or ball vides an overall schematic of how these diagnostics
sealers can be used. An example determination of a well
When a diversion technique is used, with either stimulation candidate is provided in Sidebar 20C.
chemical diverters or ball sealers, a temporary skin
effect is created that reduces the well injectivity • Well diagnostics
and then affects the pressure evolution. Compu- To determine if a well is a candidate for stimula-
tation of the pseudoskin effect created by diverter tion, a step rate or injectivity test is performed. A
cakes is detailed in Section 20-5. nonreactive fluid that is compatible with the forma-
• Other effects tion is injected at various subfracturing rates, with
the rate and pressure response carefully monitored.
Other effects can obscure the pressure response: At a constant rate and pressure, a skin effect value
– Two-phase flow can occur when water-base fluid can be calculated. When a new constant rate and
is injected into an oil or gas reservoir. The pressure are achieved, a new skin effect is calcu-
reduced mobility of the injected foreign fluid lated. At the various rates and pressures the calcu-
can complicate the pressure response. lated skin effect value should be constant. If a shift
– The reservoir can be multilayered with commin- is noted in the skin effect value, a new assumption
gled flow, or it can be fissured. In the latter case, for kh should be used until a constant value of skin
the fissures may be opened by injection. effect is achieved. Although the calculated skin
– In carbonate reservoirs, the pressure response effect is an apparent skin effect, the test gives an
may be related to reaction kinetics. indication of kh and the skin effect. If the value of
the skin effect is zero or less, the need for stimula-
• Removal of damage by reactive fluids
tion should be questioned. An analytical technique
As a result of the stimulation and damage removal, for the analysis of this type of step rate test is pro-
the skin effect from damage should decrease during vided in Earlougher (1977).
the course of the treatment. This is the component

Is there
Perform Determine a positive No Reevaluate
step rate initial skin effect? well
test skin effect concerns
Yes
Perfom
matrix
stimulation
treatment
Determine Is Stage to
apparent skin effect No
next fluid
skin effect decreasing? or stop job
in real time
Yes
Yes
Does skin
effect response No Continue
level off? stage
Yes
Evaluate the efficiency of

the treatment for this stage
Yes
Does evaluation Drop Was

indicate untreated Yes diverter diverter
zone or height? (if no increase effective?
in concentration)
No
Stop No
job
Figure 20-4. Job treatment diagnosis.

20C. Production indications for matrix stimulation requirements
For this example, the well is in a two-phase reservoir with k = 50 md, h = 70 ft, B = 1.1 RB/STB, µ = 0.8 cp, rw = 0.328 ft and
—
p = 4700 psi. The drainage radius re is unknown but presumed to be at least 750 ft. A well test 6 months ago suggested zero skin
effect. The original inflow performance relationship (IPR) curve (using Eq. 1-17) and the vertical lift performance (VLP) curve are
drawn with solid lines on Fig. 20C-1. The original production rate was more than 6600 STB/D, but 6 months later dropped to less
than 5150 STB/D. The question is whether this well is a stimulation candidate.
5000
4500
4000
3500
IPR (p = 4700 psi, s = 0)
3000
pwf (psia)
2500
VLP
2000
1500
IPR (p = 4700 psi, s = 2.4)
1000
IPR (p = 4100 psi, s = 0)
500
0
0 2000 4000 6000 8000 10,000 12,000
q (STB ⁄ D)
Figure 20C-1. Deliverability of the producing system.
Solution
From the VLP curve, the flowing bottomhole pressure should be about 2550 psi (for q = 5150 STB/D). Assuming that the average
reservoir pressure remains approximately equal to 4700 psi, then from Eq. 1-17 a skin effect equal to 2.4, represented by one of the
dashed IPR curves on Fig. 20C-1, would result in this production performance. However, the same performance can be observed
with zero skin effect but with the reservoir pressure declining to about 4100 psi. This is shown by the second dashed IPR curve in
Fig. 20C-1. Clearly, this example shows the necessity of well management and testing for the appropriate production engineering
decision (i.e., deciding whether the well is damaged and requires stimulation or whether the well drainage area is limited).
• Treatment optimization – Treatment volume—Most of the damage was

The shape and period of the treatment response removed by the first one-third of the HCl pumped.
curve provide a number of clues about the treat- This is in the time period from 27 to 60 min. The
ment. Figure 20-5 is an acidizing response curve volume used to remove this damage is 33 min ×
from a water injection well (Montgomery et al., 1.2 bbl/min ≈ 40 bbl. Using the 33-ft height
1995). The relevant parameters used to interpret determined from a step srate test, the damage was
this curve are φ = 22%, h = 33 ft (from a step rate removed by a treatment of approximately 50 gal/ft.
interpretation), average flow rate qave = 1.2 bbl/min – Damage radius—Using a porosity equal to 0.22,
and rw = 0.178 ft. the radius of formation damage rs is 3.26 ft. This
From this plot the following conclusions can be ignores residual water saturation and assumes
drawn: that 100% of the pore volume is displaced.
– Damage mechanism—The damage is clearly – Treatment improvement—On the basis of post-
removable by hydrochloric acid (HCl). Mud acid treatment spinner surveys, it was determined that
does not have an effect. Carbonate scale or iron only about one-third of the 100-ft perforated
oxide damage are the most plausible. interval was treated. To improve effectiveness,

80
70
12% HCI–3% HF on formation

15% HCI on formation
60
Transient skin effect
50
NH4CI on formation
40
30
20
10
0
0 20 40 60 80 100 120 140 160 180 200 220 240
Time (min)
Figure 20-5. Example matrix acid treatment response curve (Montgomery et al., 1995).
the use of a diverter is warranted. In addition, the that can be extracted from this type of analysis. The
use of hydrofluoric acid (HF) in this well only table shows the concentration of aluminum and silicon
increases the cost of the treatment, without any contained in the returned fluids from two wells in the
measurable stimulation benefit. same formation. Comparison of the actual Si and Al
concentrations with the theoretical Si and Al concentra-
tions expected from the volume percent of HF indicates
20-9. Post-treatment evaluation that severe secondary clay reaction precipitation
occurred. Much less Si and Al were observed in the
Several other techniques are available to evaluate the return samples than expected from the mass-balance
effectiveness of matrix stimulation treatments. In addi- equations. On the basis of the tertiary reaction kinetic
tion to the techniques described in this section, pre- work of Gdanski (1996) as described in Chapter 15,
and postbuildup tests and production logging tech- future jobs should be conducted with higher HCl to HF
niques as described in Chapters 1 and 2 can be used to ratios than the 6.5–0.5 used for these treatments to min-
evaluate acid treatments. It should always be remem- imize secondary precipitation problems.
bered that the primary criterion for a treatment is the
economic rather than the technical success.
20-9.2. Tracer surveys
20-9.1. Return fluid analysis The use of various radioactive tracers placed strategi-
cally in the fluid or with the diverter can provide insight
Almond et al. (1990) showed that monitoring the ionic on the effectiveness of matrix treatments. Figure 20-6
content of returned fluids following acidizing treat- shows a log for two treatments placed in the same well.
ments can provide insight into the actual chemical acid- The treatments were into a limestone formation perfo-
spending processes downhole. The mass balance of the rated at 4 shots per foot (spf) at a depth of 6902 to
injected and returned fluids and the relative percent of 6982 ft. The first treatment was a 6000-gal, three-stage
silicate reprecipitation allow the determination of rela- treatment (2000 gal/stage) diverted with rock salt and
tive spending levels of various acids on the formation pumped at a rate of 7 bbl/min. The acid was traced with
and well scale. Table 20-2 is an example of the results liquid 46scandium. The first 1000-lbm rock salt diverter

Table 20-2. Example acid flowback analysis showing aluminum and silicon concentrations in returned fluids
(from Almond et al., 1990).
Flowback Theoretical
Concentration (mg/L) Concentration (mg/L)
Well HF (%) Si Al Si Al Si:Al
1 16.5 216 2560 1311 15,543 0.08:1

1 12.1 190 1700 1573 14,070 0.11:1
1 7.5 142 990 1903 13,270 0.14:1
1 3.7 112 312 3060 8,525 0.35:1
1 2.5 102 117 4064 4,661 0.84:1
1 0.4 39 6 9512 1,415 6.5:1
2 7.1 162 675 2270 9,450 0.23:1
2 18.5 60 3810 325 20,630 0.015:1
stage was tagged with 192iridium, and the second

1000 lbm of rock salt was tagged with 124antimony.
Scandium
Iridium The log indicates that the diverter stages were ineffec-
Antimony
tive and that the thief zone at the top took the majority
of the treatment.
A second three-stage treatment was attempted in
which 4500 gal of acid was staged (1000, 1000 and
1500 gal/stage) by placing rock salt between the
stages. In the second treatment the upper thief zone
was isolated mechanically by placing tubing and a
packer at 6886 ft. The first 1000 gal of acid was
traced with 192iridium, the second 1000 gal with
46scandium and the third 1500 gal with 124antimony.
In this case, much better placement of the acid was

achieved, indicating that mechanical isolation was
much more effective than chemical diversion. The
scan also indicates that the rock salt dropped at the
end of the second stage was effective in diverting into
a lower zone. It would also appear that there was
some transfer of acid around the packer into the upper
zone. Explanations for this are a poor cement job,
poor packer efficiency or a fractured formation. The
postproduction curves for the second treatment are
shown in Fig. 20-7.
20-10. Conclusions
Methods to assess the effectiveness of matrix stimula-
Treatment 1 Treatment 2 tion treatments have been sought over the years. The
low overall cost and “low technology” of these treat-
Figure 20-6. Tracer scan comparison of two matrix acid
treatments on the same well treatment (courtesy of
ments usually preclude the justification of comprehen-
ProTechnic International, Inc.). sive pretreatment and post-treatment well tests.

700 Before restimulation After restimulation

Oil (BOPD)
Water (BWPD)
600
Gas (Mcf ⁄ D)
500
Producing rate
400
300
200
100
0
Nov-95 Dec-95 Jan-96 Feb-96 Mar-96 Apr-96 May-96
Monthly production
Figure 20-7. Results of the second treatment based on use of tracers to evaluate the effectiveness of the primary acid treat-
ment (courtesy of ProTechnic International, Inc.).
Several real-time skin effect evolution techniques Several techniques for evaluating the effectiveness
are presented in this chapter. The recent ones enable of matrix stimulation treatments are also presented.
definitive calculation of the initial value of the skin These include the use of pre- and postbuildup tests,
effect and its evolution during injection, taking into production logging, tracer scans, fluid flowback analy-
account pressure transients and appropriate reservoir sis and comparison of inflow performance relationship
models. The efficiency of the stimulation treatment (IPR) curves.
may then be assessed in real time. Changes in the
design and optimum termination of the injection are
therefore possible.


◆
References
Chapters 1–12 Acharya, A.R.: “Particle Transport in Viscous and

Viscoelastic Fracturing Fluids,” paper SPE 13179,
Abass, H., Meadows, D., Brumley, J. and Venditto, SPE Production Engineering (1986), 104.
J.: “Oriented Perforations—A Rock Mechanics
View,” paper SPE 28555, presented at the SPE Acharya, A.R.: “Viscoelasticity of Crosslinked
Annual Technical Conference and Exhibition, New Fracturing Fluids and Proppant Transport,” paper
Orleans, Louisiana, USA (September 25–28, 1994). SPE 16221, presented at the SPE Production
Operations Symposium, Oklahoma City, Oklahoma,
Abass, H.H, Saeed, H. and Meadows, D.L.: “Non- USA (March 8–10, 1987); also paper SPE 15937,
planar Fracture Propagation from a Horizontal SPE Production Engineering (November 1988), 483.
Wellbore: Experimental Study,” paper SPE 24823,
presented at the SPE Annual Technical Conference Acharya, A.R. and Deysarkar, A.K.: “Rheology of
and Exhibition, Washington, D.C., USA (October Fracturing Fluids at Low-Shear Conditions,” paper
4–7, 1992). SPE 16917, presented at the SPE Annual Technical
Conference and Exhibition, Dallas, Texas, USA
Abdulraheem, A., Roegiers, J.-C. and Zaman, M.: (September 27–30, 1987).
“Mechanics of Pore Collapse and Compaction in
Weak Porous Rocks,” Proc., 33rd U.S. Symposium Advani, S.H., Lee, T.S. and Dean, R.H.: “Variational
on Rock Mechanics, Santa Fe, New Mexico, USA Principles for Hydraulic Fracturing,” Journal of
(1992), 233–242. Applied Mechanics (December 1992) 59, No. 4,
819–826.
Abou-Sayed, A.S. and Dougherty, R.L.: “Evaluation
of the Influence of In Situ Reservoir Conditions on Advani, S.H., Lee, T.S., Dean, R.H., Pak, C.K. and
the Geometry of Hydraulic Fractures Using a 3D Avasthi, J.M.: “Consequences of Fracturing Fluid
Simulator,” paper SPE 13275, presented at the SPE Lag in Three-Dimensional Hydraulic Fractures,”
Annual Technical Conference and Exhibition, paper SPE 25888, presented at the SPE Rocky
Houston, Texas, USA (September 16–19, 1984). Mountain Regional/Low Permeability Reservoirs
Symposium, Denver, Colorado, USA (April 12–14,
Abou-Sayed, I.S., Schueler, S., Ehrl, E. and Hen- 1993).
dricks, W.: “Multiple Hydraulic Fracture Stimulation
in a Deep Horizontal Tight Gas Well,” paper SPE Agarwal, R.G., Al-Hussainy, R. and Ramey, H.J. Jr.:
30532, presented at the SPE Annual Technical “An Investigation of Wellbore Storage and Skin
Conference and Exhibition, Dallas, Texas, USA Effect in Unsteady Liquid Flow: I. Analytical
(October 22–25, 1995). Treatment,” paper SPE 2466, SPE Journal
(September 1970) 10, No. 3, 279–290.
Abousleiman, Y., Cheng, A.H.-D. and Gu, H.:
“Formation Permeability Determination by Micro Agarwal, R.G., Carter, R.D. and Pollock, C.B.:
or Mini-Hydraulic Fracturing,” J. Ener. Res. Tech. “Evaluation and Performance Prediction of Low
(June 1994) 116, No. 2, 104–114. Permeability Gas Wells Stimulated by Massive
Hydraulic Fracturing,” paper SPE 6838, Journal
Abramowitz, M. and Stegun, I.A. (eds.): Handbook of Petroleum Technology (March 1979) 31, No.
of Mathematical Functions, ninth edition, New York, 3, 362–372; also in Trans., AIME (1979) 267.
New York, USA, Dover (1989).
Reservoir Stimulation R-1

Ahmed, G., Horne, R.N. and Brigham, W.E.: Residue and Flow Impairment in 20/40 Mesh Sand,”
“Theoretical Development of Flow into a Well paper CIM 84-35-30, presented at the 35th Annual
Through Perforations,” report DE90000251, U.S. Technical Meeting of Petroleum Society of CIM and
DOE (August 1990). Canadian Association of Drilling Engineers, Calgary,
Alberta, Canada (June 10–13, 1984).
Ahmed, U., Newberry, B.M. and Cannon, D.E.:
“Hydraulic Fracture Treatment Design of Wells with Almond, S.W. and Garvin, T.R.: “High Efficiency
Multiple Zones,” paper SPE 13857, presented at the Fracturing Fluids for Low Temperature Reservoirs,”
SPE Low Permeability Gas Reservoir Symposium, Proc., 31st Annual Southwestern Petroleum Short
Denver, Colorado, USA (May 19–22, 1985). Course, Lubbock, Texas, USA (1984), 76–88.
Ainley, B.R.: “Development of Foam Fracturing and Amadei, B., Swolfs, H.S. and Savage, W.S.: “Gravity
Cementing Fluids for Use in Tight Gas Sands,” pre- Induced Stresses in Stratified Rock Masses,” Rock
sented at the 1983 AIChE National Meeting, Denver, Mechanics and Rock Engineering (January 1988) 21,
Colorado, USA (August 28–31, 1983). No. 1, 1–20.
Akram, A.H., Fitzpatrick, A.J. and Halford, F.R.: “A Amyx, B., et al.: Petroleum Reservoir Engineering
Model to Predict Wireline Formation Tester Sample Physical Properties, McGraw Hill (1960).
Contamination,” paper SPE 48959, presented at the
Anderson, A.J., Ashton, P.J.N., Lang, J. and Samuel-
SPE Annual Technical Conference and Exhibition,
son, M.L.: “Production Enhancement Through
New Orleans, Louisiana, USA (September 27–30,
Aggressive Flowback Procedures in the Codell
1998).
Formation,” paper SPE 36468, presented at the SPE
Albright, J.N. and Pearson, C.F.: “Acoustic Emissions Annual Technical Conference and Exhibition,
as a Tool for Hydraulic Fracture Location: Experience Denver, Colorado, USA (October 6–9, 1996).
at the Fenton Hill Hot Dry Rock Site,” paper SPE
Anderson, B., Barber, T., Druskin, V., Lee, P., Dussan,
9509, SPE Journal (August 1982) 22, No. 4, 523–
V.E., Knizhnerman, L. and Davydycheva, S.: “The
530.
Response of Multiarray Induction Tools in Highly
Aleksandrowksi, P., Inderhaug, O.H. and Knapstad, Dipping Formations with Invasion and in Arbitrary
B.: “Tectonic Structures and Wellbore Breakout 3D Geometries,” Transactions of the SPWLA 37th
Orientation,” Proc., 33rd U.S. Symposium on Rock Annual Logging Symposium (June 16–19, 1996),
Mechanics, Santa Fe, New Mexico, USA (June 3–5, paper A.
1992).
Anderson, E.M.: The Dynamics of Faulting and Dyke
Al-Hussainy, R. and Ramey, H.J. Jr.: “Application of Formation with Application to Britain, second edition,
Real Gas Theory to Well Testing and Deliverability Edinburgh, Scotland, Oliver and Boyd (1951).
Forecasting,” paper SPE 1243-B, Journal of Petrol-
Archie, G.E.: “The Electrical Resistivity Log as an
eum Technology (May 1966) 18, 637–642.
Aid in Determining Some Reservoir Characteristics,”
Al-Hussainy, R., Ramey, H.J. Jr. and Crawford, P.B.: Journal of Petroleum Technology (January 1942) 5.
“The Flow of Real Gases Through Porous Media,”
Asadi, M. and Preston, F.W.: “Characterization of
paper SPE 1243-A, Journal of Petroleum Technology
the Crushed-Zone Formed During Jet Perforation
(May 1966) 18, 624–636.
by Qualitative Scanning Electron Microscopy and
Allen, L.S., Tittle, C.W., Mills, W.R. and Caldwell, Quantitative Image Analysis,” paper SPE 22812, pre-
R.L.: “Dual Spaced Neutron Logging for Porosity,” sented at the SPE Annual Technical Conference and
Geophysics (1967), 32, No. 1. Exhibition, Dallas, Texas, USA (October 6–9, 1991).
Allen, T.O. and Worzel, H.C.: “Productivity Method Aseltine, C.: “Flash X-Ray Analysis of the
of Evaluating Gun Perforating,” Drilling and Pro- Interaction of Perforators with Different Target
duction Practice 1956, New York, New York, USA, Materials,” paper SPE 14322, presented at the SPE
American Petroleum Institute (1957), 112–125. Annual Technical Conference and Exhibition, Las
Vegas, Nevada, USA (September 22–25, 1985).
Almond, S.W., Bland, W.E. and Ripley, H.E.: “The
Effect of Break Mechanisms on Gelling Agent
R-2 References

Asgian, M.I., Cundall, P.A. and Brady, B.H.: “Mech- Baker, H.R., Bolster, R.N., Leach, P.B. and Little,
anical Stability of Propped Hydraulic Fractures: A R.C.: “Association Colloids in Nonaqueous Fluids,”
Numerical Study,” paper SPE 28510, Journal of Ind. Eng. Chem., Prod. Res. Develop. (1970) 9, No.
Petroleum Technology (March 1995) 47, No. 3, 4, 541–547.
203–208.
Bale, A., Owren, K. and Smith, M.B.: “Propped
Atkinson, B.K. (ed.): Fracture Mechanics of Rock, Fracturing as a Tool for Sand Control and Reservoir
London, UK, Academic Press (1987). Management,” paper SPE 24992, presented at the
SPE European Petroleum Conference, Cannes,
Atkinson, B.K. and Meredith, P.G.: “Experimental
France (November 16–18, 1992); also in SPE
Fracture Mechanics Data for Rocks and Minerals,”
Production & Facilities (February 1994a) 9, No.
B.K. Atkinson (ed.): Fracture Mechanics of Rock,
1, 19–28.
London, UK, Academic Press (1987), 477–525.
Bale, A., Smith, M.B. and Settari, A.: “Post-Frac
Aud, W.W., Sullivan, R.B., Coalson, E.B., Poulson,
Productivity Calculation for Complex Reservoir/
T.D. and Warembourg, P.A.: “Acid Refracturing Pro-
Fracture Geometry,” paper SPE 28919, presented at
gram Increases Reserves, Cottonwood Creek Unit,
the European Petroleum Conference, London, UK
Washakie County, Wyoming,” paper SPE 21821,
(October 24-27, 1994b).
Journal of Petroleum Technology (January 1992),
11; also presented at the Rocky Mountain Regional Baranet, S.E. and Ainley, B.R.: “Delayed Crosslinking
Meeting and Low-Permeability Reservoirs Symposium, Improves Performance of Fracturing Fluids in Deep
Denver, Colorado, USA (April 15–17, 1992). Formations,” paper CIM 85-36-39, presented at the
36th Annual Technical Meeting of the Petroleum
Aud, W.W., Wright, T.B., Cipolla, C.L., Harkrider,
Society of CIM, Edmonton, Alberta, Canada (June
J.D. and Hansen, J.T.: “The Effect of Viscosity on
2–5, 1985).
Near-Wellbore Tortuosity and Premature Screenouts,”
paper SPE 28492, presented at the SPE Annual Barenblatt, G.I.: “Mathematical Theory of Equilibrium
Technical Conference and Exhibition, New Orleans, Cracks,” Advances in Applied Mech. (1962) 7, 56–
Louisiana, USA (September 25–28, 1994). 131.
Ayoub, J.A., Bourdet, D.P. and Chauvel, Y.L.: Barree, R.D.: “A Practical Numerical Simulator for
“Impulse Testing,” paper SPE 15911, SPE Formation Three-Dimensional Fracture Propagation in Hetero-
Evaluation (September 1988), 534. geneous Media,” paper SPE 12273, presented at the
SPE Reservoir Simulation Symposium, San Francisco,
Ayoub, J.A., Brown, J.E., Barree, R.D. and Elphick,
California, USA (November 15–18, 1983).
J.J.: “Diagnosis and Evaluation of Fracturing Treat-
ments,” SPE Production Engineering (February Barree, R.D.: “A New Look at Fracture-Tip Screen-
1992a), 39. out Behavior,” paper SPE 18955, Journal of Petro-
leum Technology (February 1991) 43, No. 2, 138–
Ayoub, J.A., Kirksey, J.M., Malone, B.P. and
143.
Norman, W.D.: “Hydraulic Fracturing of Soft
Formations in the Gulf Coast,” paper SPE 23805, Barree, R.D. and Conway, M.W.: “Experimental and
presented at the SPE Formation Damage Control Numerical Modeling of Convective Proppant
Symposium, Lafayette, Louisiana, USA (February Transport,” paper SPE 28564, presented at the SPE
26–27, 1992b). Annual Technical Conference and Exhibition, New
Orleans, Louisiana, USA (September 25–28, 1994);
Babcock, R.E., Prokop, C.L. and Kehle, R.O.:
also in Journal of Petroleum Technology (March
“Distribution of Propping Agents in Vertical Frac-
1995) 47, No. 3, 216–222.
tures,” Production Monthly (November 1967), 11.
Barree, R.D. and Mukherjee, H.: “Determination of
Badon Ghyben, W.: “Nota in Verband Met de Voor-
Pressure Dependent Leakoff and Its Effect on Fracture
genomen Putboring Nabij Amsterdam,” Tijdschr.
Geometry,” paper SPE 36424, presented at the SPE
Kon. Inst. Ing. (1888–1889), 1988, 8–22.
Annual Technical Conference and Exhibition, Denver,
Colorado, USA (October 6–9, 1996).

Bartko, K.M., Robertson, B. and Wann, D.: “Imple- Behrmann, L. and Elbel, J.: “Effect of Perforations
menting Fracturing Technology to the UKCS on Fracture Initiation,” paper SPE 20661, Journal of
Carboniferous Formation,” paper SPE 38609, pre- Petroleum Technology (May 1991) 43, No. 5, 608–
sented at the SPE Annual Technical Conference and 615; also presented at the SPE Annual Technical
Exhibition, San Antonio, Texas, USA (October 5–8, Conference and Exhibition, New Orleans, Louisiana,
1997). USA (September 23–26, 1992).
Barton, C. and Zoback, M.: “Stress Pertubations Behrmann, L., et al.: “Quo Vadis, Extreme Over-
Associated with Active Faults Penetrated by Bore- balance?” Oilfield Review (Autumn 1996) 8, No.
holes: Evidence for Near Complete Stress Drop and a 3, 18–33.
New Technique for Stress Magnitude Measurement,”
Behrmann, L.A.: “Underbalance Criteria for Mini-
Journal of Geophysical Research (1994) 99, 9373–
mum Perforation Damage,” paper SPE 30081, pre-
9390.
sented at the SPE Formation Damage Symposium,
Bartusiak, R., Behrmann, L.A. and Halleck, P.M.: The Hague, Netherlands (May 15–16, 1995).
“Experimental Investigation of Surge Flow Velocity
Behrmann, L.A. and Halleck, P.M.: “Effects of Well-
and Volume Needed to Obtain Perforation Cleanup,”
bore Pressure on Perforator Penetration Depth,” paper
Journal of Petroleum Science & Engineering
SPE 18243, presented at the SPE Annual Technical
(February 1997) 13, Nos. 1–2, 19–28.
Conference and Exhibition, Houston, Texas, USA
Batchelor, A.S., Baria, R. and Hearn, K.: “Monitoring (October 2–5, 1988).
the Effects of Hydraulic Stimulation by Microseismic
Behrmann, L.A. and McDonald, B.: “Underbalance
Event Location: A Case Study,” paper SPE 12109,
or Extreme Overbalance,” paper SPE 31083, pre-
presented at the SPE Annual Technical Conference
sented at the SPE International Symposium on
and Exhibition, San Francisco, California, USA
Formation Damage Control, Lafayette, Louisiana,
(October 5–8, 1983).
USA (February 14–15, 1996).
Baumgartner, W.E., Shlyapobersky, J., Abou-Sayed,
Behrmann, L.A., Pucknell, J.K. and Bishop, S.R.:
I.S. and Jacquier, R.C.: “Fracture Stimulation of a
“Effects of Underbalance and Effective Stress on
Horizontal Well in a Deep, Tight Gas Reservoir: A
Perforation Damage in Weak Sandstone: Initial
Case History from Offshore The Netherlands,” paper
Results,” paper SPE 24770, presented at the SPE
SPE 26795, presented at the Offshore European
Annual Technical Conference and Exhibition,
Conference, Aberdeen, Scotland (September 7–10,
Washington, D.C., USA (October 4–7, 1992).
1993).
Bell, W.T.: “Perforating Underbalanced—Evolving
Beg, M.S., Kunak, A., Oguz, G., Ming, Z.D. and
Techniques,” paper SPE 13413, Journal of Petroleum
Hill, A.D.: “A Systematic Experimental Study of
Technology (October 1984) 36, No. 11, 1653–1662.
Acid Fracture Conductivity,” paper SPE 31098, pre-
sented at the SPE International Symposium on Bell, W.T., Brieger, E.F. and Harrigan, J.W. Jr.:
Formation Damage Control, Lafayette, Louisiana, “Laboratory Flow Characteristics of Gun Perfora-
USA (February 14–15, 1996). tions,” paper SPE 3444, Journal of Petroleum Tech-
nology (September 1972) 36, No. 11, 1095–1103.
Beggs, H.D. and Brill, J.P.: “A Study of Two-Phase
Flow in Inclined Pipes,” paper SPE 4007, Journal of Bell, W.T., Sukup, R.A. and Tariq, S.M.: Perforating,
Petroleum Technology (May 1973) 25, No. 5, 607– Henry L. Doherty Monograph Series, Richardson,
617. Texas, USA, Society of Petroleum Engineers (1995)
16.
Behie, A. and Settari, A.: “Perforation Design
Models for Heterogeneous, Multiphase Flow,” paper Ben Naceur, K. and Economides, M.J.: “Production
SPE 25901, presented at the SPE Rocky Mountain from Naturally Fissured Reservoirs Intercepted by a
Regional/Low Permeability Reservoirs Symposium, Vertical Hydraulic Fracture,” paper SPE 17425, pre-
Denver, Colorado, USA (April 12–14, 1993). sented at the SPE California Regional Meeting, Long
Beach, California, USA (March 23–25, 1988); also in
R-4 References

SPE Formation Evaluation (December 1989) 4, No. Biot, M.A.: “Theory of Propagation of Elastic Waves
4, 550–558. in a Fluid-Saturated Porous Solid. II. High Frequency,”
J. Acoust. Soc. Am. (1956c) 28, 179–191.
Ben Naceur, K. and Roegiers, J.-C.: “Design of
Fracturing Treatments in Multilayered Formations,” Bird, K. and Dunmore, S.: “Optimising Perforation
paper SPE 17712, presented at the SPE Gas Tech- Performance for the Armada Gas Condensate
nology Symposium, Dallas, Texas, USA (June 13–15, Development,” paper SPE 30083, presented at the
1988); also in SPE Production Engineering (February 1995 European Formation Damage Conference, The
1990) 5, No. 1, 21–26. Hague, Netherlands (May 15–16, 1995).
Bennett, C.O., Reynolds, A.C. and Raghavan, R.: Boone, T.J., Kry, P.R., Bharatha, S. and Gronseth,
“Analysis of Finite-Conductivity Fractures Intercepting J.M.: “Poroelastic Effects Related to Stress
Multilayer Commingled Reservoirs,” paper SPE Determination by Micro-Frac Tests in Permeable
11030, SPE Formation Evaluation (June 1986) 1, Rock,” Proc., 32nd U.S. Symposium on Rock
259–274. Mechanics, Norman, Oklahoma, USA (1991), 25–34.
Bennett, C.O., Rodolpho, C.-V., Reynolds, A.C. and Boone, T.J., Wawryznek, P.A. and Ingraffea, A.R.:
Raghavan, R.: “Approximate Solutions for Fractured “Simulation of the Fracture Process in Rock with
Wells Producing Layered Reservoirs,” paper SPE Application to Hydrofracturing,” International
11599, SPE Journal (October 1985) 25, 729–742. Journal of Rock Mechanics and Mining Sciences
& Geomechanics Abstracts (June 1986) 23, No. 3,
Bennett, C.O., Rosato, N.D., Reynolds, A.C. and
255–265.
Raghavan, R.: “Influence of Fracture Heterogeneity
and Wing Length on the Response of Vertically Bourdet, D.: “Pressure Behavior of Layered
Fractured Wells,” paper SPE 9886, SPE Journal Reservoirs with Crossflow,” paper SPE 13628, pre-
(April 1983) 23, 219–230. sented at the SPE California Regional Meeting,
Bakersfield, California, USA (March 27–29, 1985).
Besson, J.: “Performance of Slanted and Horizontal
Wells on an Anisotropic Medium,” paper SPE 20965, Bourdet, D., Alagoa, A., Ayoub, J.A. and Pirard,
presented at Europec 90, The Hague, Netherlands Y.M.: “New Type Curves Aid Analysis of Fissured
(October 22–24, 1990). Zone Well Tests,” World Oil (April 1984) 198, No.
5, 111–124.
Bhalla, K. and Brady, B.H.: “Formation and Fracture
Characterization by Inversion of Fracture Treatment Bourdet, D., Ayoub, J.A. and Pirard, Y.M.: “Use
Records,” paper SPE 26526, presented at the SPE of Pressure Derivative in Well-Test-Interpretation,”
Annual Technical Conference and Exhibition, paper SPE 12777, SPE Facilities & Engineering
Houston, Texas, USA (October 3–6, 1993). (June 1989), 293–302; also in Trans., AIME 287.
Bilhartz, H.L. and Ramey, H.J. Jr.: “The Combined Bourdet, D., Whittle, T.M., Douglas, A.A. and Pirard,
Effects of Storage, Skin, and Partial Penetration on Y.M.: “A New Set of Type Curves Simplifies Well
Well Test Analysis,” paper SPE 6753, presented at Test Analysis,” World Oil (May 1983) 196, No. 6,
the SPE Annual Technical Conference and Exhi- 95–106.
bition, Denver, Colorado, USA (October 9–12, 1977).
Boutéca, M.J., Bary, D., Piau, J.M., Kessler, N.,
Biot, M.A.: “General Theory of Three-Dimensional Boisson, M. and Fourmaintraux, D.: “Contribution
Consolidation,” Journal of Applied Physics (February of Poroelasticity to Reservoir Engineering: Lab
1941) 12, 155–164. Experiments, Application to Core Decompression
and Implication in HP-HT Reservoirs Depletion,”
Biot, M.A.: “General Solutions of the Equations of
Proc., Eurock ’94 Symposium, Delft, Netherlands
Elasticity and Consolidation for a Porous Material,”
(1994), 525–533.
Journal of Applied Mechanics (March 1956a) 23, No.
1, 91–96. Bradley, W.B.: “Failure of Inclined Boreholes,”
Journal of Energy Research Technology (December
Biot, M.A.: “Theory of Propagation of Elastic Waves
1979) 101, 232–239.
in a Fluid-Saturated Porous Solid. I. Low Frequency
Range,” J. Acoust. Soc. Am. (1956b) 28, 168–178.

Brady, B.H., Ingraffea, A.R. and Morales, R.H.: Britt, L.K. and Bennett, C.O.: “Determination of
“Three-Dimensional Analysis and Visualization of Fracture Conductivity in Moderate-Permeability
the Wellbore and the Fracturing Process in Inclined Reservoirs Using Bilinear Flow Concepts,” paper
Wells,” paper SPE 25889, presented at the SPE SPE 14165, presented at the SPE Annual Technical
Rocky Mountain Regional/Low Permeability Conference and Exhibition, Las Vegas, Nevada, USA
Reservoirs Symposium, Denver, Colorado, USA (September 22–25, 1985).
(April 12–14, 1993).
Britt, L.K., Hager, C.J. and Thompson, J.W.:
Branagan, P.T., Warpinski, N.R., Engler, B.P. and “Hydraulic Fracturing in a Naturally Fractured Res-
Wilmer, R.: “Measuring the Hydraulic Fracture- ervoir,” paper SPE 28717, presented at the SPE
Induced Deformation of Reservoirs and Adjacent International Petroleum Conference and Exhibition
Rocks Employing a Deeply Buried Inclinometer of Mexico, Veracruz, Mexico (October 10–13, 1994).
Array: GRI/DOE Multi-Site Project,” paper SPE
Brown, E.T. (ed.): Rock Characterization Testing
36451, presented at the SPE Annual Technical
and Monitoring, International Society for Rock
Conference and Exhibition, Denver, Colorado, USA
Mechanics, Commission on Testing Methods,
(October 6–9, 1996).
Oxford, UK, Pergamon Press (1981).
Brannon, H.D. and Pulsinelli, R.J.: “Breaker Concen-
Brown, J.E. and Economides, M.J.: “An Analysis of
trations Required to Improve the Permeability of
Horizontally Fractured Horizontal Wells,” paper SPE
Proppant Packs Damaged by Concentrated Linear
24322, presented at the SPE Rocky Mountain
and Borate-Crosslinked Fracturing Fluids,” paper
Regional Meeting, Casper, Wyoming, USA (May
SPE 21583, SPE Production Engineering (November
18–22, 1992).
1992) 7, No. 4, 338–342.
Brown, J.E., King, L.R., Nelson, E.B. and Ali, S.A.:
Brannon, H.D. and Tjon-Jo-Pin, R.M.: “Biotechno-
“Use of a Viscoelastic Carrier Fluid in Frac-Pack
logical Breakthrough Improves Performance of
Applications,” paper SPE 31114, presented at the
Moderate to High-Temperature Fracturing Appli-
SPE International Symposium on Formation Damage
cations,” paper SPE 28513, presented at the SPE
Control, Lafayette, Louisiana, USA (February 14–15,
Annual Technical Conference and Exhibition, New
1996).
Orleans, Louisiana, USA (September 25–28, 1994).
Brown, K.E., et al.: The Technology of Artificial Lift
Bratli, R.K. and Risnes, R.: “Stability and Failure of
Methods, Tulsa, Oklahoma, USA, PennWell
Sand Arches,” paper SPE 8427, SPE Journal (April
Publishing Co. (1984).
1981) 21, No. 2, 236–248.
Bui, H.D.: “An Integral Equations Method for
Bratton, T., Bornemann, T., Li, Q., Plumb, R. and
Solving the Problem of a Plane Crack of Arbitrary
Rasmus, J.: “Logging-While-Drilling Images for
Shape,” J. Mech. Phys. Solids (February 1977) 25,
Geomechanical, Geological and Petrophysical
No. 1, 29–39.
Interpretations,” Transactions of the SPWLA 40th
Annual Logging Symposium, Olso, Norway (June Burnham, J.W., Harris, L.E. and McDaniel, B.W.:
8–10, 1999). “Developments in Hydrocarbon Fluids for High-
Temperature Fracturing,” paper SPE 7564, presented
Breckels, I.M. and van Eekelen, H.A.M.: “Rela-
at the SPE Annual Technical Conference and
tionship Between Horizontal Stress and Depth in
Exhibition, Houston, Texas, USA (October 1–3,
Sedimentary Basins,” paper SPE 10336, Journal of
1978); also in Journal of Petroleum Technology
Petroleum Technology (September 1982) 34, No. 9,
(February 1980) 32, No. 2, 217–220.
2191–2199.
Camacho-V., R.G., Raghavan, R. and Reynolds,
Brie, A., Endo, T., Johnson, D.L. and Pampuri, F.:
A.C.: “Response of Wells Producing Layered
“Quantitative Formation Permeability Evaluation
Reservoirs: Unequal Fracture Length,” paper SPE
from Stoneley Waves,” presented at the SPE Annual
12844, SPE Formation Evaluation (March 1987)
Technical Conference and Exhibition, New Orleans,
2, 9–28.
Louisiana, USA (September 27–30, 1998).
R-6 References

Cameron, J.R.: “Rheological Modeling and Scale-Up Carter, R.D.: “Derivation of the General Equation
of a Delayed-Crosslinked Gel in Nonhomogeneous for Estimating the Extent of the Fractured Area,”
Flow,” paper SPE 21858, presented at the SPE Rocky Appendix I of “Optimum Fluid Characteristics for
Mountain Regional Meeting and Low-Permeability Fracture Extension,” Drilling and Production
Reservoirs Symposium, Denver, Colorado, USA Practice, G.C. Howard and C.R. Fast, New York,
(April 15–17, 1991); also in SPE Production & New York, USA, American Petroleum Institute
Facilities (February 1993), 23. (1957), 261–269.
Cameron, J.R., Joha, R.J. and Veatch, R.W. Jr.: “The Castillo, J.L.: “Modified Fracture Pressure Decline
Nonhomogeneous Flow Character of a Delayed Analysis Including Pressure-Dependent Leakoff,”
Crosslinked Gel,” paper SPE 20638, presented at the paper SPE 16417, presented at the SPE/DOE Low
SPE Annual Technical Conference and Exhibition, Permeability Reservoirs Symposium, Denver,
New Orleans, Louisiana, USA (September 23–26, Colorado, USA (May 18–19, 1987).
1990).
Cawiezel, K.E. and Elbel, J.L.: “A New System for
Cameron, J.R. and Prud’homme, R.K.: “Fracturing- Controlling the Crosslinking Rate of Borate Frac-
Fluid Flow Behavior,” Recent Advances in Hydraulic turing Fluids,” paper SPE 20077, presented at the
Fracturing, J.L. Gidley, S.A. Holditch, D.E. Nierode SPE California Regional Meeting, Ventura, Cali-
and R.W. Veatch Jr. (eds.), Monograph Series, fornia, USA (April 4–6, 1990).
Richardson, Texas, USA, Society of Petroleum
Cawiezel, K.E. and Niles, T.D.: “Rheological Prop-
Engineers (1989), 177–209.
erties of Foam Fracturing Fluids Under Downhole
Cantu, L.A. and Boyd, P.A.: “Laboratory and Field Conditions,” paper SPE 16191, presented at the SPE
Evaluation of a Combined Fluid-Loss Control Hydrocarbon Economics and Evaluation Symposium,
Additive and Gel Breaker for Fracturing Fluids,” Dallas, Texas, USA (March 2–3, 1987).
paper SPE 18211, presented at the SPE Annual
Cesaro, M., Cheung, P., Etchecopar, A. and Gonfalini,
Technical Conference and Exhibition, Houston,
M.: “Shaping Up to Stress in the Appenines,” Italy
Texas, USA (October 2–5, 1988).
2000: Value Added Reservoir Characterization (Well
Card, R.J., Howard, P.R. and Féraud, J.-P.: “Control Evaluation Conference, Italy), Schlumberger (1997).
of Particulate Flowback in Subterranean Wells,” U.S.
Chambers, M.R., Mueller, M.M. and Grossmann, A.:
Patent No. 5,330,005 (July 19, 1994).
“Well Completion Design and Operations for a Deep
Card, R.J., Howard, P.R. and Féraud, J.-P.: “A Novel Horizontal Well with Multiple Fractures,” paper SPE
Technology to Control Proppant Back Production,” 30417, presented at the Offshore Europe Conference,
paper SPE 31007, SPE Production & Facilities Aberdeen, Scotland (September 5–8, 1995).
(November 1995), 271.
Chaveteau, G.: “Fundamental Criteria in Polymer
Carico, R.D. and Bagshaw, F.R.: “Description and Flow Through Porous Media and Their Relative
Use of Polymers Used in Drilling, Workovers, and Importance in the Performance Differences of
Completions,” paper SPE 7747, presented at the SPE Mobility Control Buffers,” Water Soluble Polymers,
Production Technology Symposium, Hobbs, New J.E. Glass (ed.), Advances in Chemistry Series
Mexico, USA (October 30–31, 1978). (1986), 213.
Carmen, P.C.: “Fundamental Principles of Industrial Cheatham, J.B. Jr. and McEver, J.W.: “Behavior of
Filtration,” Trans. Inst. Chem. Eng. (1938) 16, Casing Subjected to Salt Loading,” paper SPE 828,
168–188. Journal of Petroleum Technology (September 1964)
16, 1069–1075.
Carnahan, B., Luther, H.A. and Wilkes, J.O.: Applied
Numerical Methods, New York, New York, USA, Chen, W.F. and Han, D.J.: Plasticity for Structural
John Wiley (1969). Engineers, New York, New York, USA, Springer-
Verlag (1988).
Carslaw, H.S. and Jaeger, J.C.: Conduction of Heat in
Solids, second edition, Oxford, UK, Clarendon Press Chen, Z. and Economides, M.J.: “The Effect of Near-
(1959). Wellbore Fracture Geometry on Fracture Execution

and Post-Treatment Production of Deviated and Cinco-Ley, H., Samaniego-V., F. and Dominguez, N.:
Horizontal Wells,” paper SPE 39425, presented at “Transient Pressure Behavior for a Well with a Finite-
the SPE International Formation Damage Control Conductivity Vertical Fracture,” paper SPE 6014,
Conference, Lafayette, Louisiana, USA (February presented at the SPE Annual Technical Conference
18–19, 1998). and Exhibition, New Orleans, Louisiana, USA
(October 3–6, 1976); also in SPE Journal (August
Cheng, C.H. and Johnston, D.H.: “Dynamic and
1978) 18, 253–264.
Static Moduli,” Geophys. Res. Letters (January 1981)
8, No. 1, 39–42. Clark, J.B.: “A Hydraulic Process for Increasing the
Productivity of Wells,” Petroleum Trans., AIME
Cherepanov, G.P.: Mechanics of Brittle Fracture,
(January 1949) 186, 1–8.
New York, New York, USA, McGraw-Hill (1979).
Clark, K.K.: “Transient Pressure Testing of Fractured
Chrisp, J.: U.S. Patent No. 3,301,723 (1967).
Water Injection Wells,” paper SPE 1821, Journal of
Cinco-Ley, H. and Meng, H.Z.: “Pressure Transient Petroleum Technology (June 1968) 20, 639–643.
Analysis of Wells with Finite Conductivity Vertical
Clark, P.E.: “Stimulation Fluid Rheology—A New
Fractures in Double Porosity Reservoirs,” paper SPE
Approach,” paper SPE 8300, presented at the SPE
18172, presented at the SPE Annual Technical Con-
Annual Technical Conference and Exhibition, Las
ference and Exhibition, Houston, Texas, USA
(October 2–5, 1988).
Clark, P.E., Balakrishnan, M. and Sundram, L.:
Cinco-Ley, H., Ramey, H.J. Jr. and Miller, F.G.:
“Crosslinking of Hydroxypropyl Guar with Metal
“Pseudoskin Factors for Partially Penetrating Direc-
Ions,” paper SPE 25208, presented at the SPE Inter-
tionally Drilled Wells,” paper SPE 5589, presented
national Symposium on Oilfield Chemistry, New
at the SPE Annual Technical Conference and Exhi-
Orleans, Louisiana, USA (March 2–5, 1993).
bition, Dallas, Texas, USA (September 28–October
1, 1975a). Clark, P.E. and Barkat, O.: “Hydraulic Fracturing
Fluids: The Crosslinking of Hydroxypropyl Guar
with Titanium Chelates,” paper SPE 19331, presented
“Unsteady-State Pressure Distribution Created by
at the SPE Eastern Regional Meeting, Morgantown,
a Well with an Inclined Fracture,” paper SPE 5591,
West Virginia, USA (October 24–27, 1989).
and Exhibition, Dallas, Texas, USA (September Clark, P.E. and Courington, T.J.: “Visualization of
28–October 1, 1975b). Flow Into a Vertical Fracture,” paper SPE 27692,
presented at the SPE Permian Basin Oil and Gas
Cinco-Ley, H. and Samaniego-V., F.: “Effect of
Recovery Conference, Midland, Texas, USA (March
Wellbore Storage and Damage on the Transient
16–18, 1994).
Pressure Behavior of Vertically Fractured Wells,”
paper SPE 6752, presented at the SPE Annual Clark, P.E. and Guler, N.: “Prop Transport in Vertical
Technical Conference and Exhibition, Denver, Fractures: Settling Velocity Correlations,” paper SPE
Colorado, USA (October 9–12, 1977). 11636, presented at the SPE/DOE Symposium on
Low Permeability, Denver, Colorado, USA (March
Cinco-Ley, H. and Samaniego-V., F.: “Transient
14–16, 1983).
Pressure Analysis: Finite Conductivity Fracture Case
Versus Damaged Fracture Case,” paper SPE 10179, Clark, P.E., Halvaci, M., Ghaeli, H. and Parks, C.F.:
presented at the SPE Annual Technical Conference “Proppant Transport by Xanthan and Xanthan-
and Exhibition, San Antonio, Texas, USA (October Hydroxypropylguar Solutions: Alternatives to Cross-
5–7, 1981a). linked Fluids,” paper SPE/DOE 13907, presented at
the SPE/DOE Low Permeability Gas Reservoirs
Cinco-Ley, H. and Samaniego-V., F.: “Transient
Conference, Denver, Colorado, USA, (May 19–22,
Pressure Analysis for Fractured Wells,” paper SPE
1985).
7490, Journal of Petroleum Technology (September
1981b) 33, 1749–1766.
R-8 References

Clark, P.E., Harkin, M.W., Wahl, H.A. and Sievert, Conference and Exhibition, New Orleans, Louisiana,
J.A.: “Design of a Large Vertical Prop Transport USA (September 25–28, 1994).
Model,” paper SPE 6814, presented at the SPE
Cleary, M.P. and Fonseca, A.: “Proppant Convection
and Encapsulation in Hydraulic Fracturing: Practical
Denver, Colorado, USA (October 9–12, 1977).
Implications of Computer Laboratory Simulations,”
Clark, P.E., Manning, F.S., Quadir, J.A. and Guler, paper SPE 24825, presented at the SPE Annual
N.: “Prop Transport in Vertical Fractures,” paper SPE Technical Conference and Exhibition, Washington,
10261, presented at the SPE Annual Technical Con- D.C., USA (October 4–7, 1992).
ference and Exhibition, San Antonio, Texas, USA
Cleary, M.P., Keck, R.G. and Mear, M.E.: “Micro-
(October 5–7, 1981).
computer Models for the Design of Hydraulic
Clark, P.E. and Zhu, Q.: “Flow Visualization of Flow Fractures,” paper SPE 11628, presented at the SPE/
Into 8 Point-Source Fractures II: The Effect of DOE Symposium on Low Permeability, Denver,
Fracture Non-Uniformity and Fluid Crosslinking,” Colorado, USA (March 14–16, 1983).
paper SPE 29503, presented at the SPE Production
Cleary, M.P., Wright, C.A. and Wright, T.B.:
Operations Symposium, Oklahoma City, OK, USA
“Experimental and Modeling Evidence for Major
(April 2–4, 1995a).
Changes in Hydraulic Fracturing Design and Field
Clark, P.E. and Zhu, Q.: “Fluid Flow in Vertical Procedures,” paper SPE 21494, Proc., SPE Gas
Fractures from a Point Source,” paper SPE 28509, Technology Symposium, Houston, Texas, USA
presented at the SPE Annual Technical Conference (1991), 131–146.
and Exhibition, New Orleans, Louisiana, USA
Clifton, R.J. and Abou-Sayed, A.S.: “On the Com-
(September 25–28, 1994); also in Journal of
putation of the Three-Dimensional Geometry of
Petroleum Technology (March 1995b) 47, No.
Hydraulic Fractures,” paper SPE 7943, presented
3, 209–215.
at the SPE/DOE Low Permeability Gas Reservoirs
Cleary, M., Johnson, D., Kogsboll, H., Owens, K., Symposium, Denver, Colorado, USA (May 20–22,
Perry, K., de Pater, C., Stachel, A., Schmidt, H. and 1979).
Tambini, M.: “Field Implementation of Proppant
Clifton, R.J. and Abou-Sayed, A.S.: “A Variational
Slugs to Avoid Premature Screen-Out of Hydraulic
Approach to the Prediction of the Three-Dimensional
Fractures with Adequate Proppant Concentration,”
Geometry of Hydraulic Fractures,” paper SPE 9879,
paper SPE 25892, presented at the SPE Rocky
presented at the SPE Low Permeability Symposium,
Mountain Regional/Low Permeability Reservoirs
Denver, Colorado, USA (May 27–29, 1981).
1993). Clifton, R.J. and Wang, J-J.: “Multiple Fluids,
Proppant Transport, and Thermal Effects in Three-
Cleary, M.P.: “Analysis of Mechanisms and Proced-
Dimensional Simulation of Hydraulic Fracturing,”
ures for Producing Favourable Shapes of Hydraulic
Fractures,” paper SPE 9260, presented at the SPE
Annual Technical Conference and Exhibition, Dallas,
Texas, USA (September 21–24, 1980a).
Colle, E.A., Halleck, P.M., Saucier, R.J. and Wesson,
Cleary, M.P.: “Comprehensive Design Formulae for
D.S.: “Discussion of Simple Method Predicts
Hydraulic Fracturing,” paper SPE 9259, presented
Downhole Shaped-Charge Gun Performance,” SPE
Production & Facilities (November 1994) 9, No. 4,
bition, Dallas, Texas, USA (September 21–24, 1980b).
292–293.
Cleary, M.P., Doyle, R.S., Teng, E.Y., Cipolla, C.L.,
Collins, R.E.: Flow of Fluids Through Porous
Meehan, D.N., Massaras, L.V. and Wright, T.B.:
Materials, New York, New York, USA, Van Nostrand
“Major New Developments in Hydraulic Fracturing,
Reinhold (1961).
with Documented Reductions in Job Costs and
Increases in Normalized Production,” paper SPE Constien, V.G., Brannon, H.D. and Bannister, C.E.:
28565, presented at the SPE Annual Technical “Continuous Mix Technology Adds New Flexibility

to Frac Jobs,” Oil & Gas Journal (June 6, 1988) 86, Shear History Simulation,” paper SPE/DOE 11635,
No. 23, 49–54. presented at the SPE/DOE Symposium on Low
Permeability, Denver, Colorado, USA (March 14–16,
Constien, V.G., Fellin, E.L. and King, M.T.: “Auto-
1983).
mated Rheology Laboratory—Part 1,” ACS Symp.
Ser. (1986) 313, 105. Cramer, D.: “Limited Entry Extended to Massive
Hydraulic Fracturing,” Oil & Gas Journal (December
Conway, M.W., Almond, S.W., Briscoe, J.E. and
14, 1987) 85, No. 50, 40–50.
Harris, L.E.: “Chemical Model for the Rheological
Behavior of Crosslinked Fluid Systems,” paper SPE Cramer, D.D., Dawson, J. and Ouabdesselam, M.:
9334, presented at the SPE Annual Technical “An Improved Gelled Oil System for High-
Conference and Exhibition, Dallas, Texas, USA Temperature Fracturing Applications,” paper SPE
(September 21–24, 1980). 21859, presented at the Rocky Mountain Regional
Meeting and Low Permeability Reservoirs
Cook, J.M., Bradford, I.D.R. and Plumb, R.A.: “A
Symposium, Denver, Colorado, USA.
Study of the Physical Mechanisms of Sanding and
Application to Sand Production Prediction,” paper Crawford, D.L., Earl, R.B. and Monroe, R.F.:
SPE 28852, Proc., European Petroleum Conference, “Friction Reducing and Gelling Agent for Organic
London, UK (1994), 473–480. Liquids,” U.S. Patent No. 3,757,864 (September 11,
1973).
Cook, N.G.W.: “The Design of Underground
Excavations,” Proc., 8th Symposium on Rock Crockett, A.R., Willis, R.M. and Cleary, M.P.:
Mechanics, Minneapolis, Minnesota, USA (1967), “Improvement of Hydraulic Fracture Predictions by
167–193. Real-Time History Matching on Observed Pressures,”
SPE Production Engineering (November 1989) 4,
Cooke, C.E. Jr.: “Conductivity of Fracture Proppants
No. 4, 408–416.
in Multiple Layers,” paper SPE 4117, Journal of
Petroleum Technology (September 1973) 25, Crouch, S.L. and Starfield, A.M.: Boundary Element
1101–1107; also in Trans., AIME (1973) 255. Methods in Solid Mechanics, London, UK, Allen and
Unwin (1983).
Cooke, C.E. Jr.: “Effect of Fracturing Fluids on
Fracture Conductivity,” paper SPE 5114, Journal Crowe, C.W.: “Clay Stabilization Ups Productivity,”
of Petroleum Technology (October 1975) 27, Oil and Gas Suppliers’ Digest (1979) 1, No. 6.
1273–1282; also in Trans., AIME (1975) 259.
Crowe, C.W., Hutchinson, B.H. and Trittipo, B.L.:
Coon, R.: “Correlation of Engineering Behaviour “Fluid-Loss Control: The Key to Successful Acid
with the Classification of In Situ Rock,” PhD disser- Fracturing,” SPE Production Engineering (August
tation, University of Illinois, Illinois, USA (1968). 1989) 4, No. 3, 215–220.
Cotton, F.A. and Wilkinson, G.: Advanced Inorganic Crowe, C.W., Martin, R.C. and Michaelis, A.M.:
Chemistry, New York, New York, USA, Interscience “Evaluation of Acid-Gelling Agents for Use in Well
Publishers (1972). Stimulation,” paper SPE 9384, SPE Journal (August
1981), 415.
Coulter, A.W., Crowe, C.W., Barrett, N.D. and
Miller, B.D.: “Alternate Stages of Pad Fluid and Acid Crowe, C.W., McGowan, G.R. and Baranet, S.E.:
Provide Improved Leakoff Control for Fracture “Investigation of Retarded Acids Provides Better
Acidizing,” paper SPE 6124, presented at the SPE Understanding of Their Effectiveness and Potential
Annual Technical Conference and Exhibition, New Benefits,” paper SPE 1822, SPE Production
Orleans, Louisiana, USA (October 3–6, 1976). Engineering (May 1990), 166.
Craft, B.C., Kemp, C.E. and Caulde, B.H.: Well Crump, J. and Conway, M.: “Effects of Perforation-
Design: Drilling and Production, Englewood Cliffs, Entry Friction on Bottomhole Treating Analysis,”
New Jersey, USA, Prentice Hall (1962), 494. paper SPE 15474, Journal of Petroleum Technology
(August 1988) 40, No. 8, 1041–1048.
Craigie, L.J.: “A New Method for Determining the
Rheology of Crosslinked Fracturing Fluids Using
R-10 References

Daccord, G., Lamanczyk, R. and Vercaemer, C.: Sand Control,” paper SPE 26545, presented at the
“Method for Obtaining Gelled Hydrocarbon Compo- SPE Annual Technical Conference and Exhibition,
sitions According to Said Method and Their Appli- Houston, Texas, USA (October 3–6, 1993).
cation the Hydraulic Fracturing of Underground
de Kruijf, A., Roodhart, L.P. and Davies, D.R.: “The
Formations,” U.S. Patent No. 4,507,213 (March 26,
Relation Between Chemistry and Flow Mechanics of
1985).
Borate Crosslinked Fracturing Fluids,” paper SPE
Dake, L.: Fundamentals of Reservoir Engineering, 25206, presented at the SPE International Sympo-
Developments in Petroleum Science, Amsterdam, sium on Oilfield Chemistry in New Orleans,
Netherlands, Elsevier Scientific Publishing Company Louisiana, USA (March 2–5, 1993); also in SPE
(1982) 8. Production & Facilities (August 1993), 165.
Daneshy, A.A.: “Experimental Investigation of de Pater, C.J., Desroches, J., Gronenbroom, J. and
Hydraulic Fracturing Through Perforations,” paper Weijers, L.: “Physical and Numerical Modelling of
SPE 4333, Journal of Petroleum Technology Hydraulic Fracture Closure,” paper SPE 28561,
(October 1973) 25, No. 10, 1201–1206. presented at the SPE Annual Technical Conference
Darcy, H.: Les Fontaines Publiques de la Ville
(September 25–28, 1994); also in SPE Production
de Dijon, Paris, France, Victor Dalmont (1856).
& Facilities (May 1996), 122.
Davidson, B.M., Saunders, B.F. and Holditch, S.A.:
de Pater, C.J., Weijers, L., Savic, M., Wolf, K.H.A.A.,
“Field Measurement of Hydraulic Fracture Fluids:
van den Hoek, P.J. and Barr, D.T.: “Experimental
Case Study,” paper SPE 27694, presented at the SPE
Study of Nonlinear Effects in Hydraulic Fracture
Permian Basin Oil and Gas Recovery Conference,
Propagation,” paper SPE 25893, presented at the
Midland, Texas, USA (March 16–18, 1994).
SPE Joint Rocky Mountain Regional Meeting/Low
Davies, D.R., Bosma, M.G.R. and de Vries, W.: Permeability Reservoirs Symposium, Denver,
“Development of Field Design Rules for Viscous Colorado, USA (April 12–14, 1993); also in SPE
Fingering in Acid Fracturing Treatments: A Large- Production Engineering (November 1994) 9, No.
Scale Model Study,” paper SPE 15772, presented at 4, 239–246.
the SPE Middle East Oil Show, Manama, Bahrain
Deo, M., Tariq, S.M. and Halleck, P.M.: “Linear and
(March 7–10, 1987).
Radial Flow Targets for Characterizing Downhole
Davies, D.R., Lau, H.C., Kamphuis, H. and Rood- Flow in Perforations,” paper SPE 16896, SPE Pro-
hart, L.P.: “Development and Field Testing of a duction Engineering (August 1989) 4, No. 3, 295–
Novel Fracturing Fluid,” paper SPE 23109, presented 300.
at the SPE Offshore Petroleum Conference, Aber-
de Rozières, J.: “Measuring Diffusion Coefficients
deen, Scotland (September 3–6, 1991).
in Acid Fracturing Fluids and Their Application to
Davis, P.M.: “Surface Deformation Associated with Gelled and Emulsified Acids,” paper SPE 28552, pre-
Dipping Hydrofracture,” Journal of Geophysical sented at the SPE Annual Technical Conference and
Research (July 10, 1983) 88, No. B7, 5826–5834. Exhibition, New Orleans, Louisiana, USA (Sep-
Dawson, J.: “Method and Composition for Delaying tember 25–28, 1994).
the Gellation of Borated Galactomannons,” U.S. Desbrandes, R. and Gualdron, J.: “In Situ Rock
Patent No. 5,082,579 (1992). Wettability Determination with Wireline Formation
Dawson, J.C.: “A Thermodynamic Study of Borate Tester Data,” The Log Analyst (July–August 1988)
Complexation with Guar and Guar Derivatives,” 29, No. 4, 244–252.
paper SPE 22837, presented at the SPE Annual Desroches, J. and Kurkjian, A: “Applications of
Technical Conference and Exhibition, Dallas, Texas, Wireline Stress Measurements,” paper SPE 58086,
USA (October 6–9, 1991). SPE Reservoir Evaluation (October 1999).
Dees, J.M. and Handren, P.J.: “A New Method of Desroches, J., Lenoach, B., Papanastasiou, P. and
Overbalance Perforating and Surging of Resin for Thiercelin, M.: “On the Modelling of Near Tip

Processes in Hydraulic Fracturing,” International Exhibition, Houston, Texas, USA (October 3–6,
Journal of Rock Mechanics and Mining Sciences 1993).
& Geomechanics Abstracts (1993) 30, No. 7,
Dietz, D.N.: “Determination of Average Reservoir
1127–1134.
Pressure from Build-Up Surveys,” paper SPE 1156,
Desroches, J. and Thiercelin, M.: “Modelling the Journal of Petroleum Technology (August 1965) 17,
Propagation and Closure of Micro-Hydraulic Frac- No. 8, 955–959.
tures,” International Journal of Rock Mechanics and
Dobecki, T.L.: “Hydraulic Fracture Orientation Using
Mining Sciences & Geomechanics Abstracts (1993)
Passive Borehole Seismics,” paper SPE 12110, pre-
30, No. 7, 1231–1234.
sented at the SPE Annual Technical Conference and
Desroches, J. and Woods, T.E.: “Stress Measurements Exhibition, San Francisco, California, USA (October
for Sand Control,” paper SPE 47247, presented at the 5–8, 1983).
SPE/ISRM Eurock ‘98, Trondheim, Norway (July
Dobkins, T.A.: “Improved Methods to Determine
8–10, 1998).
Hydraulic Fracture Height,” paper SPE 8403, Journal
Detournay, E. and Carbonell, R.: “Fracture Mechanics of Petroleum Technology (April 1981) 33, No. 4,
Analysis of the Breakdown Process in Minifrac or 719–726.
Leak-Off Tests,” paper SPE 28076, presented at the
Dovan, H.T. and Hutchins, R.D.: “Crosslinked
SPE/ISRM Rock Mechanics in Petroleum Engi-
Hydroxyethylcellulose and Its Uses,” U.S. Patent
neering Conference, Delft, Netherlands (August
No. 5,207,934 (May 4, 1993).
29–31, 1994).
Dunand, A., Guillot, D. and Soucemarianadin, A.:
Detournay, E. and Cheng, A.: “Poroelastic Response
“Viscous Properties of Glass-Bead Suspensions and
of a Borehole in a Non-Hydrostatic Stress Field,”
Falling Sphere Experiments in Hydroxypropyl Guar
International Journal of Rock Mechanics and Mining
Solutions,” Proc., Ninth International Congress on
Sciences & Geomechanics Abstracts (June 1988) 25,
Rheology, Mexico City, Mexico (1984).
No. 3, 171–182.
Dunning, J.D.: “Surfactants,” Chemtech (June 1980)
Detournay, E. and Cheng, A.: “Plane Strain Analysis
10, No. 6, 356–359.
of a Stationary Hydraulic Fracture in a Poroelastic
Medium,” Int. J. Solids & Structures (March 1991) Dyes, A.B., Kemp, C.E. and Caudle, B.H.: “Effect
27, No. 13, 1645–1662. of Fractures on Sweep-Out Pattern,” Trans., AIME
(1958) 213, 245–249.
Detournay, E. and Cheng, A.: “Influence of Pressuri-
zation Rate on the Magnitude of the Breakdown Earlougher, R.C. Jr.: Advances in Well Test Analysis,
Pressure,” Proc., 33rd Symposium on Rock Dallas, Texas, USA, Society of Petroleum Engineers
Mechanics, Santa Fe, New Mexico, USA (1992), (1977).
325–333. Economides, M.J.: “Implications of Cementing on
Detournay, E. and Cheng, A.: “Fundamentals of Well Performance,” Well Cementing, E.B. Nelson
Poroelasticity,” Comprehensive Rock Engineering, (ed.), Sugar Land, Texas, USA, Schlumberger (1990),
J.A. Hudson (ed.), Oxford, UK, Pergamon Press 1-1–1-6.
(1993) 2, 113–171. Economides, M.J., Brand, C.W. and Frick, T.P.: “Well
Deuel, H. and Neukorn, H.: “Uber die Reacktion Configurations in Anisotropic Reservoirs,” SPE
von Borsaure und Borax mit Polysacchariden und Formation Evaluation (December 1996) 11, No. 4,
Anderen Hochmolekularen Polyoxy-Verbindungen,” 257–262.
Makromol. Chem. (1949) 13. Economides, M.J., Deimbacher, F.X., Brand, C.W.
Dewprashad, B., Abass, H.H., Meadows, D.L., and Heinemann, Z.E.: “Comprehensive Simulation
Weaver, J.D. and Bennett, B.J.: “A Method to Select of Horizontal Well Performance,” SPE Formation
Resin-Coated Proppants,” paper SPE 26523, pre- Evaluation (December 1991) 6, No. 4, 418–426.
R-12 References

Economides, M.J., Hill, A.D. and Ehlig-Economides, Elbel, J.L. and Ayoub, J.A.: “Evaluation of Apparent
C.A.: Petroleum Production Systems, Englewood Fracture Lens Tip Indicated from Transient Tests,”
Cliffs, New Jersey, USA, Prentice-Hall (1994). paper 91-44, presented at the CIM-AOSTRA Tech-
nical Conference, Banff, Alberta, Canada (April
Economides, M.J., McLennan, J.D., Brown, E. and
21–24, 1991); also in Journal of Canadian Petroleum
Roegiers, J.-C.: “Performance and Stimulation of
Technology (December 1992) 31, No. 10, 41–46.
Horizontal Wells,” World Oil (June–July 1989) 208,
No. 6, 41–45, and 209, No. 1, 69–72 and 76–77. Elbel, J.L., Howard, R.L., Talley, G.R. and
McLaughlin, B.H.: “Stimulation Study of Cottage
Edimann, K., Somerville, J.M., Smart, B.G.D.,
Grove Formation,” paper SPE 11564, Journal of
Hamilton, S.A. and Crawford, B.R.: “Predicting Rock
Petroleum Technology (July 1984), 1199–1205.
Mechanical Properties from Wireline Porosities,”
paper SPE 47344, presented at the SPE/ISRM Eurock Elbel, J.L., Piggott, A.R. and Mack, M.G.:
‘98, Trondheim, Norway (July 8–10, 1998). “Numerical Modeling of Multilayer Fracture
Treatments,” paper SPE 23982, presented at the SPE
Ehlig-Economides, C.A.: “Implications of the Test
Area of Investigation in Nonradial Geometries,” SPE
Memorial Series, Richardson, Texas, USA, Society of
Petroleum Engineers (1995) 1, 103–112. Elbel, J.L. and Sookprasong, P.A.: “Use of
Cumulative-Production Type Curves in Fracture
Ehlig-Economides, C.A., Chan, K.S. and Spath, J.B.:
Design,” SPE Production Engineering (August 1987)
“Production Enhancement Strategies for Strong
2, No. 3, 191–198.
Bottomwater Drive Reservoirs,” paper SPE 36613,
presented at the SPE Annual Technical Conference El Rabaa, W.: “Experimental Study of Hydraulic
and Exhibition, Denver, Colorado, USA (October Fracture Geometry Initiated from Horizontal Wells,”
6–9, 1996). paper SPE 19720, presented at the SPE Annual
Technical Conference and Exhibition, San Antonio,
Ehlig-Economides, C.A. and Economides, M.J.:
“Pressure Transient Analysis in an Elongated Linear
Flow System,” SPE Journal (December 1985) 25, Ely, J.W.: Stimulation Treatment Handbook, Tulsa,
No. 6, 347–360. Oklahoma, USA, PennWell Publishing Co. (1985).
Ehlig-Economides, C.A., Joseph, J., Erba, M. and Ely, J.W.: “GRI’s Fracture Fluid Rheology Unit—
Vik, S.A.: “Evaluation of Single-Layer Transients in Portable Equipment for Measuring Fluid Behavior
a Multilayered System,” paper SPE 15860, presented Under Simulated Downhole Conditions,” In Focus—
at the European Petroleum Conference, London, UK Tight Gas Sands, Gas Research Institute (September
(October 20–22, 1986). 1987) 4, No. 2.
Elbel, J.L.: “Designing Hydraulic Fractures for Endo, T., Brie, A. and Badri, M.: “Fracture Evaluation
Efficient Reserve Recovery,” paper SPE 15231, pre- from Dipole Shear Anisotropy and Borehole Stoneley
sented at the SPE Unconventional Gas Technology Waves,” Transactions of the SPWLA 2nd Annual
Symposium, Louisville, Kentucky, USA (May 18–21, Logging Symposium of Japan (1996), paper U.
1986).
Engelder, T.: Stress Regimes in the Lithosphere,
Elbel, J.L.: “Considerations for Optimum Fracture Princeton, New Jersey, USA, Princeton University
Geometry Design,” SPE Production Engineering Press (1993).
(August 1988) 3, No. 3, 323–327.
Enzendorfer, C., Harris, R.A., Valkó, P., Economides,
Elbel, J.L.: “A Method To Estimate Multizone M.J., Fokker, P.A. and Davies, D.D.: “Pipe Vis-
Injection Profiles During Hydraulic Fracturing,” cometry of Foams,” J. Rheol. (March–April 1995)
paper SPE 21869, presented at the Rocky Mountain 39, No. 2, 345–358.
Regional Meeting and Low-Permeability Reservoirs
Esmersoy, C., Koster, K., Williams, M., Boyd, A. and
Kane, M.: “Dipole Shear Anisotropy Logging,”
1993).

Proc., 64th International Meeting of the Society Spherical Particles in Newtonian Liquids,” J. Phys.
of Exploration Geophysicists (1994), 1139–1142. Chem. (1960), 1168–1174.
Evans, K.F., Engelder, T. and Plumb, R.A.: Ford, W.G.F.: “Foamed Acid—An Effective Stim-
“Appalachian Stress Study: 1. A Detailed Description ulation Fluid,” paper SPE 9385, Journal of Petroleum
of In-Situ Stress Variations in Devonian Shales of the Technology (July 1981) 33, 1203–1210.
Appalachian Plateau,” Journal of Geophysical
Fragachan, F.E., Mack, M.G., Nolte, K.G. and Teggin,
Research (1989) 94, No. B6, 7129–7154.
D.E.: “Fracture Characterization from Measured and
Fair, W.B. Jr.: “Pressure Buildup Analysis with Simulated Bottomhole Pressures,” paper SPE 25848,
Wellbore Phase Distribution Effects,” SPE Journal presented at the SPE Rocky Mountain Regional/Low
(April 1981) 21, 259–270. Permeability Reservoirs Symposium, Denver, Col-
orado, USA (April 12–14, 1993).
Fan, Y.: “Measurements of Viscous and Degradation
Properties of Fracturing Fluids at Insitu Conditions Franquet, J.A. and Economides, M.J.: “Effect of
of Shear Rate and Temperature Using FANN 50C Stress and Stress Path on Young’s Modulus and
Viscometer,” PhD dissertation, Texas A&M Poisson Ratio of Unconsolidated Rocks: A New Idea
University, College Station, Texas, USA (1992). for Hydraulic Fracturing,” paper SPE 54012, pre-
sented at the SPE Latin America and Caribbean
Fan, Y. and Holditch, S.A.: “Testing Crosslinked
Petroleum Engineering Conference, Caracas,
Fluids on the Fann 50 Viscometer Using the Vol-
Venezuela (April 21–23, 1999).
umetric Average Shear Rate,” paper SPE 26895, pre-
sented at the SPE Eastern Regional Conference and Fraser, C.D. and Petitt, B.E.: “Results of a Field Test
Exhibition, Pittsburgh, Pennsylvania, USA to Determine the Type and Orientation of a
(November 2–4, 1993). Hydraulically Induced Formation Fracture,” paper
SPE 180, Journal of Petroleum Technology (May
Farris, R.F.: “Fracturing Formations in Wells,”
1962) 14, 463–466.
reissue No. 23,733 of U.S. Patent No. 2,596,843
(November 1953). Frick, T.P. and Economides, M.J.: “Horizontal Well
Damage Characterization and Removal,” SPE Pro-
Fast, C.R., Holman, G.B. and Colvin, R.J.: “The
duction & Facilities (February 1993) 8, No. 1,
Application of Massive Hydraulic Fracturing to the
15–22.
Tight Muddy ‘J’ Formation, Wattenberg Field,
Colorado,” paper SPE 5624, Journal of Petroleum Friedman, G.M. and Sanders, J.E.: Principles of
Technology (January 1977) 29, 10–17. Sedimentology, John Wiley & Sons, New York, New
York, USA (1978).
Felsenthal, M.: “Step Rate Tests to Determine Safe
Injection Pressures in Floods,” Oil & Gas Journal Friedman, M., Handin, J. and Alani, G.: “Fracture-
(October 1994), 49–54. Surface Energy of Rocks,” International Journal of
Rock Mechanics and Mining Sciences & Geomech-
Fenner, R.: “Untersuchen zur Erkentnis des
anics Abstracts (November 1972) 9, No. 6, 757–766.
Gebirgsdrucks,” Glukauf (1938) 74.
Frimann-Dahl, C., Irvine-Fortescue, J., Rokke, E.,
Fletcher, P.A., Montgomery, C.T., Ramos, G.G.,
Vik, S. and Wahl, O.: “Formation Testers Vs. DST—
Miller, M.E. and Rich, D.A.: “Using Fracturing as a
The Cost Effective Use of Transient Analysis to Get
Technique for Controlling Formation Failure,” paper
Reservoir Parameters,” paper SPE 48962, presented
SPE 27899, presented at the Western Regional
Meeting, Long Beach, California, USA (March
Exhibition, New Orleans, Louisiana, USA (Sep-
23–25, 1984).
tember 27–30, 1998).
Forchheimer, P.: “Wasserbewegung Durch Boden,”
Fung, R.L., Vilajakumar, S. and Cormack, D.E.:
ZVDI (1901) 45, 1781.
“Calculation of Vertical Fracture Containment in
Ford, T.F.: “Viscosity Concentration and Fluidity- Layered Formations,” paper SPE 14707, SPE
Concentration Relationships for Suspensions of Formation Evaluation (December 1987) 2, No.
4, 518–523.
R-14 References

Gadala-Maria, F. and Acrivos, A.: “Shear-Induced Fractures,” paper SPE 2458, Journal of Petroleum
Structure in a Concentrated Suspension of Solid Technology (December 1969) 21, 1571–1581.
Spheres,” J. Rheol. (1980) 24, No. 6, 799–814.
Geertsma, J. and Haafkens, R.: “A Comparison of the
Gall, B.L. and Raible, C.J.: “A Method to Study Theories to Predict Width and Extent of Vertical,
Fracturing Fluid Polymer Degradation Using Size Hydraulically Induced Fractures,” Trans., AIME
Exclusion Chromatography,” DOE/BETC/RI-83/10 (March 1979) 101, 8.
(1984).
Geilikman, M.B, Dusseault, M.B. and Dullien, F.A.:
Gall, B.L. and Raible, C.J.: “Molecular Size Studies “Sand Production as a Viscoplastic Granular Flow,”
of Degraded Fracturing Fluid Polymers,” paper SPE paper SPE 27343, Proc., SPE International Sympo-
13566, presented at the SPE Oilfield and Geothermal sium on Formation Damage Control, Lafayette,
Chemistry Symposium, Phoenix, Arizona, USA Louisiana, USA (1994), 41–50.
(April 9–11, 1985).
Germanovich, L.N., Salganik, R.L., Dyskin, A.V. and
Gardner, D.C.: “High Fracturing Pressures for Shales Lee, K.K.: “Mechanisms of Brittle Fracture of Rock
and Which Tip Effects May Be Responsible,” paper with Pre-Existing Cracks in Compression,” Pure and
SPE 24852, presented at the SPE Annual Technical Applied Geophysics (1994) 143, Nos. 1–3, 117–149.
Conference and Exhibition, Washington, D.C., USA
Gidley, J.L.: “A Method for Correcting Dimensionless
(October 4–7, 1992).
Fracture Conductivity for Non-Darcy Flow Effects,”
Gardner, D.C. and Eikerts, J.V.: “The Effects of Shear paper SPE 20710, presented at the SPE Annual
and Proppant on the Viscosity of Cross-Linked Technical Conference and Exhibition, New Orleans,
Fracturing Fluids,” paper SPE 11066, presented at the Louisiana, USA (September 23–26, 1990); also in
SPE Annual Technical Conference and Exhibition, SPE Production Engineering (November 1991) 6,
New Orleans, Louisiana, USA (September 26–29, No. 4, 391–394.
1982).
Gidley, J.L., Holditch, S.A., Nierode, D.E. and
Gatens, J.M., Lee, W.J., Hopkins, C.W. and Lan- Veatch, R.W. Jr.: Recent Advances in Hydraulic
caster, D.E.: “The Effect of Permeability Anisotropy Fracturing, Monograph Series, Richardson, Texas,
on the Evaluation and Design of Hydraulic Fracture USA, Society of Petroleum Engineers (1989).
Treatments and Well Performance,” paper SPE
Gnirk, P.F.: “The Mechanical Behavior of Uncased
21501, presented at the SPE Gas Technology
Wellbores Situated in Elastic/Plastic Media Under
Symposium, Houston, Texas, USA (January 23–25,
Hydrostatic Stress,” paper SPE 3224, SPE Journal
1991).
(February 1972) 12, No. 1, 49–59.
Gauthier, F., Goldsmith, H.L. and Mason, S.G.:
Godbey, J.K. and Hodges, H.D.: “Pressure
“Particle Motions in Non-Newtonian Media,” Rheol.
Measurements During Formation Fracturing
Acta (September 1971) 10, No. 3, 344–364.
Operations,” Trans., AIME (1958) 213, 65–69.
Gdanski, R.D. and Norman, L.R.: “Using the
Goel, N., Willingham, J.D., Shah, S.N. and Lord,
Hollow-Core Test to Determine Acid Reaction
D.L.: “A Comparative Study of Borate-Crosslinked
Rates,” paper SPE 12151, SPE Production
Gel Rheology Using Laboratory and Field Scale
Engineering (March 1986), 111.
Fracturing Simulations,” paper SPE 38618, presented
Geertsma, J.: “Estimating the Coefficient of Inertial at the SPE Annual Technical Conference and Exhi-
Resistance in Fluid Flow Through Porous Media,” bition, San Antonio, Texas, USA (October 5–8,
SPE Journal (October 1974) 14, No. 5, 445–450. 1997).
Geertsma, J.: “Some Rock-Mechanical Aspects of Oil Goode, P.A., Pop, J.J. and Murphy, W.F. III:
and Gas Well Completions,” paper SPE 8073, SPE “Multiple-Probe Formation Testing and Vertical
Journal (December 1985) 25, No. 6, 848–856. Reservoir Continuity,” paper SPE 22738, presented
Geertsma, J. and de Klerk, F.: “A Rapid Method of
Exhibition, Dallas, Texas, USA (October 6–9, 1991).
Predicting Width and Extent of Hydraulic Induced

Goodman, R.E.: Methods of Geological Engineering SPE 25425, presented at the Production Operations
in Discontinuous Rocks, St. Paul, Minnesota, USA, Symposium, Oklahoma City, Oklahoma, USA
West Publishing (1976). (March 21–23, 1993).
Graham, J.W., Kiel, O.M., Terry, W.M. and Sinclair, Guar and Derivatives, Henkel Corporation (1986).
A.R.: “Hydraulic Fracturing Using Petroleum Coke,”
Guenot, A.: “Instability Problems at Great Depth
U.S. Patent No. 3,664,420 (May 23, 1972).
Drilling Boreholes and Wells,” Proc., Rock at Great
Grebe, J.J. and Stoesser, M.: “Increasing Crude Depth, International Symposium of Rock Mechanics,
Production 20,000,000 Bbl. from Established Fields,” Pau, France (1989), 1199–1208.
World Petroleum (August 1935), 473–482.
Guillot, D. and Dunand, A.: “Rheological Charac-
Griffith, A.A.: “The Phenomena of Rupture and Flow terization of Fracturing Fluids by Using Laser Ane-
in Solids,” Phil. Trans. Roy. Soc. (1921) 221, 163– mometry,” paper SPE 12030, SPE Journal (February
198. 1985), 39.
Griffith, A.A.: “The Theory of Rupture,” Proc., 1st Gulbis, J.: “Dynamic Fluid Loss Study of Fracturing
International Congress for Applied Mechanics, Delft, Fluid,” paper CIM 82-33-18, presented at the 33rd
Netherlands (1924), 55–63. Annual Technical Meeting of Petroleum Society of
CIM (1982).
Gringarten, A.C., Bourdet, D.P., Landel, P.A. and
Kniazeff, V.: “A Comparison Between Different Skin Gulbis, J.: “Dynamic Fluid Loss of Fracturing
and Wellbore Storage Type-Curves for Early-Time Fluids,” paper SPE 12154, presented at the SPE
Transient Analysis,” paper SPE 8205, presented at Annual Technical Conference and Exhibition, San
the SPE Annual Technical Conference and Exhi- Francisco, California, USA (October 5–8, 1983).
bition, Las Vegas, Nevada, USA (September 23–
Gulbis, J., King, M.T., Hawkins, G.W. and Brannon,
26, 1979).
H.D.: “Encapsulated Breaker for Aqueous Polymeric
Gringarten, A.C. and Ramey, H.J. Jr.: “The Use of Fluids,” SPE Production Engineering (February
Source and Green’s Functions in Solving Unsteady 1992) 7, No. 1, 9–14.
Flow Problems in Reservoirs,” SPE Journal (October
Gulrajani, S.N., Mack, M.G. and Elbel, J.L.:
1973) 13, No. 5, 285–296.
“Pressure History Inversion for Interpretation of
Gringarten, A.C. and Ramey, H.J. Jr.: “Unsteady- Fracture Treatments,” paper SPE 36439, presented
State Pressure Distributions Created by a Well with at the SPE Annual Technical Conference and
a Single Horizontal Fracture, Partial Penetration, or Exhibition, Denver, Colorado, USA (October 6–9,
Restricted Entry,” paper SPE 3819, SPE Journal 1996); also in SPE Production & Facilities
(August 1974) 14, No. 4, 413–426. (November 1998) 13, No. 4, 258–266.
Gringarten, A.C., Ramey, H.J. Jr. and Raghavan, R.: Gulrajani, S.N., Nolte, K.G. and Romero, J.:
“Unsteady State Pressure Distributions Created by a “Evaluation of the M-Site B-Sand Experiments: The
Well with a Single Infinite-Conductivity Vertical Evolution of a Pressure Analysis Methodology,”
Fracture,” paper SPE 4051, SPE Journal (August paper SPE 38575, presented at the SPE Annual
1974) 14, No. 4, 347–360. Technical Conference and Exhibition, San Antonio,
Texas, USA (October 5–8, 1997a).
Gross, J.M.: “Gelling Organic Liquids,” U.S. Patent
No. 5,190,675 (March 2, 1993). Gulrajani, S.N. and Romero, J.: “Evaluation and
Modification of Fracture Treatments Showing Near-
Gruesbeck, C. and Collins, R.E.: “Particle Transport
Wellbore Effects,” paper SPE 36901, presented at the
Through Perforations,” paper SPE 7006, presented at
SPE European Petroleum Conference, Milan, Italy
the SPE Symposium on Formation Damage Control,
(October 22–24, 1996).
Lafayette, Louisiana, USA (February 15–16, 1978).
Gulrajani, S.N., Tibbles, R.J. and Nolte, K.G.:
Gu, H., Elbel, J.L., Nolte, K.G., Cheng, A.H.-D. and
“Evaluation of Calibration Treatments, for Frac-Pack
Abousleiman, Y.: “Formation Permeability Deter-
Completions, in Offshore West Africa,” paper SPE
mination Using Impulse Fracture Injection,” paper
R-16 References

38192, presented at the SPE European Formation Fracturing Breakdown Pressure,” Proc., 32nd U.S.
Damage Symposium, The Hague, Netherlands (June Symposium on Rock Mechanics, Norman, Okla-
2–3, 1997b). homa, USA (1991), 191–199.
Guo, F., Morgenstern, N.R. and Scott, J.D.: Hainey, B.W., Weng, X. and Stoisits, R.F.: “Mit-
“Interpretation of Hydraulic Fracturing Breakdown igation of Multiple Fractures from Deviated Well-
Pressure,” International Journal of Rock Mechanics bores,” paper SPE 30482, presented at the SPE
and Mining Sciences & Geomechanics Abstracts Annual Technical Conference and Exhibition, Dallas,
(1993) 30, No. 6, 617–626. Texas, USA (October 22–25, 1995).
Guppy, K.H., Cinco-Ley, H. and Ramey, H.J. Jr.: Hall, B.E. and Houk, S.G.: “Fluid-Loss Control in the
“Pressure Buildup Analysis of Fractured Wells Naturally Fractured Buda Formation,” paper SPE
Producing at High Flow Rates,” paper SPE 10178, 12152, presented at the SPE Annual Technical Con-
Journal of Petroleum Technology (November 1982a) ference and Exhibition, San Francisco, California,
34, 2656–2666. USA (October 5–8, 1983).
Guppy, K.H., Cinco-Ley, H., Ramey, H.J. Jr. and Hall, C.D. Jr. and Dollarhide, F.E.: “Effects of
Samaniego-V., F.: “Non-Darcy Flow in Wells with Fracturing Fluid Velocity on Fluid Loss Agent Per-
Finite-Conductivity Vertical Fractures,” paper SPE formance,” paper SPE 736, Journal of Petroleum
8281, SPE Journal (October 1982b) 22, No. 5, Technology (May 1964), 555–560; also in Trans.,
681–698. AIME (1964) 231.
Gupta, D.V.S. and Cooney, A.P.: “Encapsulations for Hall, C.D. Jr. and Dollarhide, F.E.: “Performance
Treating Subterranean Formations and Methods for of Fracturing Fluid Loss Agents Under Dynamic
the Use Thereof,” U.S. Patent No. 5,164,099 Conditions,” paper SPE 1904, Journal of Petroleum
(November 17, 1992). Technology (July 1968), 763–769.
Hagedorn, A.R. and Brown, K.E.: “Experimental Halleck, P.M.: “The Effects of Stress and Pore
Study of Pressure Gradients Occurring During Pressure on Penetration of Jet Perforators in Berea
Continuous Two-Phase Flow in Small-Diameter Sandstone,” final report, API Project 86-36, Amer-
Vertical Conduits,” paper SPE 940, Journal of ican Petroleum Institute (1987).
Petroleum Technology (April 1965) 17, 475–484.
Halleck, P.M.: “Further Effects of Stress on Pene-
Hagoort, J.: “Waterflood Induced Hydraulic Frac- tration Flow Performance of Jet Perforations,” final
turing,” PhD thesis, Delft University of Technology, report, API Project 87-36, American Petroleum
Delft, Netherlands (1980). Institute (1988).
Haimson, B. and Fairhurst, C.: “Hydraulic Fracturing Halleck, P.M.: “Minimum Size and Stress Require-
in Porous-Permeable Materials,” paper SPE 2354, ments for a Possible API Standard Test Target for
Journal of Petroleum Technology (July 1969) 21, Performance Evaluation of Shaped Charge Oil Well
811–817. Perforators in Stressed Rock,” final report, API
Project 88-36, American Petroleum Institute (1989).
Haimson, B.C.: “The Hydraulic Fracturing Method
of Stress Measurement: Theory and Practice,” Halleck, P.M.: “The Application of Prototype
Comprehensive Rock Engineering, J. Hudson (ed.), Penetration and Flow Tests in Stressed Targets to
Oxford, UK, Pergamon Press (1993), 3, 297–328. Several Perforator Charges and Rocks,” final report,
API Project 89-36, American Petroleum Institute
Haimson, B.C. and Huang, X.: “Hydraulic Fracturing
(1990).
Breakdown Pressure and In-Situ Stress at Great
Depth,” Rock at Great Depth, V. Fourmaintraux and Halleck, P.M.: “Minimizing Skin in Fractured
D. Maury (eds.), Rotterdam, Netherlands, Balkema Reservoirs,” final report GRI-95/0472, Contract
(1989), 939–946. 5094-210-2901, Gas Research Institute (March
1996a).
Haimson, B.C. and Zhongliang, Z.: “Effect of
Borehole Size and Pressurization Rate on Hydraulic

Halleck, P.M.: “Studies Reveal Fractured Reservoir SPE 22808, presented at the SPE Annual Technical
Perforating Damage,” Petroleum Engineering Conference and Exhibition, Dallas, Texas, USA
International (April 1996b) 69, No. 4, 35–44. (October 6–9, 1991).
Halleck, P.M., Atwood, D.C. and Black, A.D.: “X-Ray Handren, P.J., Jupp, T.B. and Dees, J.M.: “Overbalance
CT Observations of Flow Distribution in a Shaped- Perforating and Stimulation Method for Wells,” paper
Charge Perforation,” paper SPE 24771, presented at SPE 26515, presented at the SPE Annual Technical
the SPE Annual Technical Conference and Exhibi- Conference and Exhibition, Houston, Texas, USA
tion, Washington, D.C., USA (October 4–7, 1992). (October 3–6, 1993).
Halleck, P.M. and Behrmann, L.A.: “Penetration of Hannah, R.R., Harrington, L.J. and Anderson, R.W.:
Shaped Charges in Stressed Rock,” Rock Mechanics “Measurement of Dynamic Proppant Fall Rates in
Contributions and Challenges, Rotterdam, Nether- Fracturing Gels Using a Concentric Cylinder Tester,”
lands, Balkema (1990), 629–636; also presented at paper SPE 7571, presented at the SPE Annual
the 31st U.S. Symposium on Rock Mechanics (1990). Technical Conference and Exhibition, Houston,
Halleck, P.M. and Damasena, E.: “Sand Production
Tests Under Simulated Down Hole Pressure Hannah, R.R., Harrington, L.J. and Lance, L.C.: “The
Conditions Using Shaped-Charge Perforated Well Real Time Calculation of Accurate Bottomhole
Cores,” paper SPE 19748, presented at the SPE Fracturing Pressure from Surface Measurements with
Annual Technical Conference and Exhibition, San Measured Pressure as a Base,” paper SPE 12062, pre-
Antonio, Texas, USA (October 8–11, 1989). sented at the SPE Annual Technical Conference and
Exhibition, San Francisco, California, USA (October
Halleck, P.M. and Deo, M.: “The Effect of
5–8, 1983).
Underbalance on Perforation Flow,” paper SPE
16895, presented at the SPE Annual Technical Hannah, R.R., Park, E.I., Porter, D.A. and Black,
Conference and Exhibition, Dallas, Texas, USA J.W.: “Combination Fracturing/Gravel Packing
(September 27–30, 1987); also in Journal of Completion Technique on the Amberjack, Mississippi
Production Engineering (May 1989) 4, No. 2, Canyon 109 Field,” paper SPE 26562, SPE Produc-
113–116. tion & Facilities (November 1994) 9, No. 4,
262–267.
Halleck, P.M., Poyol E. and Santarelli, F.J.:
“Estimating Perforation Flow Performance from Hannah, R.R. and Walker, E.J.: “Fracturing a High-
Variation in Indentation Hardness,” paper SPE 24769, Permeability Oil Well at Prudhoe Bay, Alaska,” paper
presented at the SPE Annual Technical Conference SPE 14372, presented at the SPE Annual Technical
and Exhibition, Washington, D.C., USA (October Conference and Exhibition, Las Vegas, Nevada, USA
4–7, 1992); also in SPE Drilling and Completion (September 22–25, 1985).
(December 1995) 10, No. 4, 271–275.
Harrington, L. and Hannah, R.R.: “Fracturing Design
Halleck, P.M., Safargar, M.S. and Dogulu, Y.S.: Using Perfect Support Fluids for Selected Fracture
“Perforating Naturally Fractured Formations: Proppant Concentrations in Vertical Fractures,” paper
Fracture Plugging by Jet Metal and Rock Defor- SPE 5642, presented at the SPE Annual Meeting,
mation,” Turkish Journal of Oil and Gas (1996) Dallas, Texas, USA (September 28–October 1, 1975).
2, 20–38.
Harrington, L.J., Hannah, R.R. and Williams, D.:
Halleck, P.M., Saucier, R.J., Behrmann, L.A. and “Dynamic Experiments on Proppant Settling in
Ahrens, T.J.: “Reduction of Jet Perforator Penetration Crosslinked Fracturing Fluids,” paper SPE 8342, pre-
in Rock Under Stress,” paper SPE 18245, presented sented at the SPE Annual Technical Conference and
at the SPE Annual Technical Conference and Exhi- Exhibition, Las Vegas, Nevada, USA (September
bition, Houston, Texas, USA (October 2–5, 1988). 23–26, 1979).
Halleck, P.M., Wesson, D.S., Snider, P.M. and Harrington, L.J., Whitsitt, N.F. and Hannah, R.R.:
Navarrete, M.: “Prediction of In-Situ Shaped Charge “Prediction of the Location and Movement of Fluid
Penetration Using Acoustic and Density Logs,” paper Interfaces in the Fracture,” Southwestern Petroleum
R-18 References

Short Course, Texas Tech University, Lubbock, Journal of Petroleum Technology (August 1963) 15,
Texas, USA (April 1973). 896–900.
Harris, L.E., Holtmyer, M.D. and Pauls, R.W.: Hawkins, G.W.: “Molecular Weight Reduction and
“Method for Fracturing Subterranean Formations,” Physical Consequences of Chemical Degradation of
U.S. Patent No. 4,622,155 (November 11, 1986). Hydroxypropylguar in Aqueous Brine Solutions,”
Proc., ACS Division of Polymeric Materials: Science
Harris, M.H.: “The Effect of Perforating on Well
and Engineering (1986) 55.
Productivity,” paper SPE 1236, Journal of Petroleum
Technology (April 1966) 18, 518–528. Hawkins, G.W.: “Laboratory Study of Proppant-Pack
Permeability Reduction Caused by Fracturing Fluids
Harris, P.C.: “Dynamic Fluid-Loss Characteristics
Concentrated During Closure,” paper SPE 18261,
of Nitrogen Foam Fracturing Fluids,” Journal of
Petroleum Technology (October 1985) 18, No. 10,
and Exhibition, Houston, Texas, USA (October 2–5,
1847–1852.
1988).
Harris, P.C.: “Chemistry and Rheology of Borate
Hawkins, M.F.: “A Note on the Skin Effect,” Journal
Crosslinked Fluids at Temperatures Up to 300°F,”
of Petroleum Technology (December 1956) 8, 356–
paper SPE 24339, presented at the SPE Rocky
357.
Mountain Regional Meeting, Casper, Wyoming, USA
(May 18–21, 1993). Hejjo, H., Fagan, J.E. and Shah, S.N.: “Flow Visual-
ization Technique for Investigation of Hydraulic
Harris, P.C. and Heath, S.J.: “Rheology of Cross-
Fracturing,” paper SPE 38579, presented at the SPE
linked Foams,” paper SPE 28512-P, SPE Production
Annual Technical Conference and Exhibition, San
& Facilities (May 1996) 11, No. 2, 113–116.
Antonio, Texas, USA (October 5–8, 1997).
Harris, P.C. and Heath, S.J.: “Rheological Properties
Herron, M.M., Johnson, D.L. and Schwartz, L.M.: “A
of Low-Gel-Loading Borate Fracture Gels,” paper
Robust Permeability Estimator for Siliclastics,” pre-
SPE 52399, SPE Production & Facilities (1998).
Harrison, E., Kieschnick, W.F. Jr. and McGuire, W.J.: Exhibition, New Orleans, Louisiana, USA (Sep-
“The Mechanics of Fracture Induction and Exten- tember 27–30, 1998).
sion,” Trans., AIME (1954) 201, 252–263.
Herzberg, A.: “Die Wasserversorgung einiger
Hart, C.M., Engi, D., Fleming, R.P. and Morris, H.E.: Nordseebaden,” Z. Gasbeleucht. Wasserverzorg.
“Fracture Diagnostics Results for the Multiwell (1901) 44, 815–819, 824–844.
Experiment’s Paludal Zone Stimulation,” paper SPE
Hilbert, L.B., Hwong, T.K., Cook, N.G.W., Nihei,
12852, presented at the SPE/DOW/GRI Uncon-
K.T. and Myer, L.R.: “Effects of Strain Amplitude
ventional Gas Recovery Symposium, Pittsburgh,
on the Static and Dynamic Nonlinear Deformation
Pennsylvania, USA (May 13–15, 1984).
of Berea Sandstone,” Proc., 1st North American
Hashem, M.N., Thomas, E.C., McNeil, R.I. and Symposium on Rock Mechanics, Austin, Texas, USA
Mullins, O.C.: “Determination of Producible (1994), 497–504.
Hydrocarbon Type and Oil Quality in Wells Drilled
Hill, R.: The Mathematical Theory of Plasticity,
with Synthetic Oil-Based Muds,” paper SPE 39093,
Oxford, UK, Clarendon Press (1951).
and Exhibition, San Antonio, Texas, USA (October Himes, R.E. and Vinson, E.F.: “Environmentally Safe
5–8, 1997). Salt Replacement for Fracturing Fluids,” paper SPE
23438, presented at the SPE Eastern Regional
Haskin, C.A.: “A Review of Hydroxy Aluminum
Meeting, Lexington, Kentucky, USA (October 2–5,
Treatments,” paper SPE 5692, presented at the SPE
1991).
Symposium on Formation Damage Control, Houston,
Texas, USA (January 29–30, 1976). Hinkel, J.J.: “Breaker System for High Viscosity
Fluids,” U.S. Patent No. 4,250,044 (February 10,
Havlena, D. and Odeh, A.S.: “The Material Balance
1981).
as an Equation of a Straight Line,” paper SPE 559,

Hirth, J.P. and Lothe, J.: Theory of Dislocations, New Holditch, S.A. and Morse, R.A.: “The Effects of
York, New York, USA, McGraw-Hill (1968). Non-Darcy Flow on the Behavior of Hydraulically
Fractured Gas Wells,” paper SPE 5586, Journal of
Hodge, R.M.: “Hydraulic Fracturing Method Using
Petroleum Technology (October 1976) 28, 1169–
Delayed Crosslinker Composition,” U.S. Patent No.
1178.
4,749,041 (June 7, 1988a).
Holditch, S.A. and Rahim, Z.: “Developing Data Sets
Hodge, R.M.: “Method of Fracturing a Subterranean
for 3D Fracture Propagation Models,” paper SPE
Formation Using Delayed Crosslinker Compositions
26155, SPE Production & Facilities (November
Containing Organic Titanium Complexes,” U.S.
1994) 9, No. 4, 257–261.
Patent No. 4,749,040 (June 7, 1988b).
Hornby, B.E.: “Tomographic Reconstruction of Near-
Hodge, R.M. and Baranet, S.E.: “Evaluation of Field
Borehole Slowness Using Refracted Borehole Sonic
Methods to Determine Crosslink Times of Fracturing
Arrivals,” Geophysics (December 1993) 58, No. 12,
Fluids,” paper SPE 16249, presented at the SPE
1726–1738.
International Symposium on Oilfield Chemistry, San
Antonio, Texas, USA (February 4–6, 1987). Horne, R.N.: Modern Well Test Analysis: Computer-
Aided Approach, Palo Alto, California, USA,
Hodgson, K. and Joughin, N.C.: “The Relationship
Petroway (1990).
Between Energy Release Rate, Damage and Seis-
micity in Deep Mines,” Failure and Breakage of Horner, D.R.: “Pressure Buildup in Wells,” Proc.,
Rock, C. Fairhurst (ed.), Minneapolis, Minnesota, Third World Petroleum Congress, The Hague,
USA, University of Minnesota Press (1967), 194– Netherlands (1951) Sec. II, 503–523; also in Pressure
203. Analysis Methods, SPE Reprint Series, Richardson,
Texas, USA, Society of Petroleum Engineers (1967)
Holder, J., Morita, N., Kenrick, A.J., Thallak, S. and
9, 25–43.
Gray, K.E.: “Measurements of Effective Fracture
Toughness Values for Hydraulic Fracture: Depen- Houzé, O.P., Horne, R.N. and Ramey, H.J. Jr.:
dence on Pressure and Fluid Rheology,” paper SPE “Pressure-Transient Response of an Infinite-
25491, presented at the SPE Production Operations Conductivity Vertical Fracture in a Reservoir with
Symposium, Oklahoma City, Oklahoma, USA Double-Porosity Behavior,” paper SPE 16778, SPE
(March 21–23, 1993). Formation Evaluation (September 1988) 3, No. 3,
510–518.
Holditch, S.A.: “Criteria of Propping Agent Selection,”
prepared for the Norton Company (1979a). Howard, G.C. and Fast, C.R.: “Optimum Fluid
Characteristics for Fracture Extension,” Drilling and
Holditch, S.A.: “Factors Affecting Water Blocking
Production Practice, New York, New York, USA,
and Gas Flow from Hydraulically Fractured Gas
American Petroleum Institute (1957) 24, 261–270.
Wells,” Journal of Petroleum Technology (December
(Appendix by E.D. Carter)
1979b) 31, No. 12, 1515–1524.
Howard, G.C. and Fast, C.R.: Hydraulic Fracturing,
Holditch, S.A. and Blakeley, D.M.: “Flow Char-
Monograph Series, Richardson, Texas, USA, Society
acteristics of Hydraulic Fracture Proppants Subjected
of Petroleum Engineers (1970) 2.
to Repeated Production Cycles,” paper SPE 19091,
SPE Production Engineering (February 1992) 7, No. Howard, P.R., James, S.G. and Milton-Tayler, D.:
1, 15–20. “High Permeability Channels in Proppant Packs
Containing Random Fibers,” SPE Production &
Holditch, S.A., Jennings, J.W., Neuse, S.H. and
Facilities (August 1999) 14, No. 3, 197.
Wyman, R.E.: “The Optimization of Well Spacing
and Fracture Length in Low Permeability Gas Howard, P.R., King, M.T., Morris, M., Féraud, J.-P.,
Reservoirs,” paper SPE 7496, presented at the SPE Slusher, G. and Lipari, S.: “Fiber/Proppant Mixtures
Annual Technical Conference and Exhibition, Control Proppant Flowback in South Texas,” paper
Houston, Texas, USA (October 1–3, 1978). SPE 30495, presented at the SPE Annual Technical
(October 22–25, 1995).
R-20 References

Hsia, T.-Y. and Behrmann, L.A.: “Perforating Skins James, S.G., Samuelson, M.L., Reed, G.W. and
as a Function of Rock Permeability and Under- Sullivan, S.C.: “Proppant Flowback Control in High
balance,” paper SPE 22810, presented at the SPE Temperature Wells,” paper SPE 39960, presented at
Annual Technical Conference and Exhibition, Dallas, the SPE Rocky Mountain Regional/Low Permeability
Texas, USA (October 6–9, 1991). Reservoirs Symposium and Exhibition, Denver,
Colorado, USA (April 5–8, 1998).
Hubbert, M.K. and Willis, D.G.: “Mechanics of
Hydraulic Fracturing,” paper 686-G, presented at the Jasinski, R., Redwine, D. and Rose, G.: “Boron
SPE Annual Meeting, Los Angeles, California, USA Equilibria with High Molecular Weight Guar: An
(October 14–17, 1956); also in Journal of Petroleum NMR Study,” J. Polymer Science: Part B: Polymer
Technology (September 1957) 9, No. 6, 153–168 and Physics (1996) 34, 1477–1488.
Trans., AIME (1957) 210.
Jeffrey, R.G.: “The Combined Effect of Fluid Lag
Huddleston, D.A.: “Liquid Aluminum Phosphate Salt and Fracture Toughness on Hydraulic Fracture
Gelling Agent,” U.S. Patent No. 5,110,485 (May 5, Propagation,” paper SPE 18957, presented at the SPE
1992). Joint Rocky Mountain Regional/Low Permeability
Reservoirs Symposium, Denver, Colorado, USA
Hunt, J.L.: “Production Systems Analysis for Frac-
(March 6–8, 1989).
tured Wells,” paper SPE 15931, presented at the SPE
Eastern Regional Meeting, Columbus, Ohio, USA Jennings, A.R.: “Fracturing Fluids—Then and Now,”
(November 12–14, 1986). paper SPE 36166-P, Journal of Petroleum Technology
(July 1996) 48, No. 7, 604–611.
Hunter, J. and Walker, R.N. Jr.: “Clean Frac Fluids
Improve Load Recovery, Tight-Gas Production,” Jizba, D. and Nur, A.: “Static and Dynamic Moduli
World Oil (October 1991) 212, No. 10, 76–78. of Tight Gas Sandstones and Their Relation to
Formation Properties,” Transactions of the SPWLA
International Society for Rock Mechanics Commission
31st Annual Logging Symposium, Lafayette,
on Testing Methods: Suggested Methods for Deter-
Louisiana, USA (June 24–27, 1990), paper BB.
mining the Fracture Toughness of Rocks, Lisbon,
Portugal, ISRM (1988). Johnson, D.E., Wright, T.B., Tambini, M., Moroli,
R. and Cleary, M.P.: “Real-Data On-Site Analysis of
Irwin, G.R.: “Analysis of Stresses and Strains Near
Hydraulic Fracturing Generates Optimum Procedures
the End of a Crack Traversing a Plate,” Journal of
for Job Design and Execution,” paper SPE 25920,
Applied Mechanics (September 1957) 24, 361–364.
presented at the SPE Rocky Mountain Regional
Ispas, I.N., Britt, L.K., Tiab, D., Valkó, P. and Meeting, Denver, Colorado, USA (April 12–14,
Economides, M.J.: “Methodology of Fluid Leakoff 1993).
Analysis in High Permeability Fracturing,” paper
Johnson, E. and Cleary, M.P.: “Implications of Recent
SPE 39476, presented at the SPE Formation Damage
Laboratory Experimental Results for Hydraulic
Conference, Lafayette, Louisiana, USA (February
Fractures,” paper SPE 21846, Proc., Rocky Mountain
18–19, 1998).
Regional/Low-Permeability Reservoirs Symposium,
ISRM Commission on Testing Methods: “Suggested Denver, Colorado, USA (1991), 413–428.
Methods for Determining Mode I Fracture Toughness
Jones, J. and Soler, D.: “Fracture Stimulation of
Using Cracked Chevron Notched Brazilian Disc
Shallow, Unconsolidated Kern River Sands,” paper
(CCNBD) Specimens,” International Journal of Rock
SPE 54102 (in press).
Mechanics and Mining Sciences & Geomechanics
Abstracts (1995) 32, No. 1, 59–64. Jones, L.G., Tibbles, R.J., Myers, L., Bryant, D.,
Hardin, J. and Hurst, G.: “Gravel Packing Horizontal
Ito, T. and Hayashi, K.: “Physical Background to the
Wellbores with Leak-Off Using Shunts,” paper SPE
Breakdown Pressure in Hydraulic Fracturing Tectonic
Stress Measurements,” International Journal of Rock
Conference and Exhibition, San Antonio, Texas,
USA (October 5–8, 1997).
Abstracts (1991) 28, No. 4, 285–293.

Joshi, S.D.: “Augmentation of Well Productivity with Kim, C.M. and Willingham, J.R.: “Flow Response
Slant and Horizontal Wells,” paper SPE 15375, of Propped Fracture to Repeated Production Cycles,”
Journal of Petroleum Technology (June 1988) 40, No. paper SPE 16912, presented at the SPE Annual
6, 729–739. Technical Conference and Exhibition, Dallas, Texas,
USA (September 27–30, 1987).
Kamphuis, H., Davies, D.R. and Roodhart, L.P.:
“A New Simulator for the Calculation of the In-Situ King, G.E.: “Foam Stimulation Fluids: What They
Temperature Profile During Well Stimulation Are, Where They Work,” Pet. Eng. Int. (July 1982)
Fracturing Treatments,” paper CIM/SPE 90-46, 54, No. 8, 52–60.
Journal of Canadian Petroleum Technology (May
King, G.E., Bingham, M.D. and Kidder, R.W.:
1993) 30, No. 5, 38–47.
“Factors Affecting Perforating Charge Performance
Kanninen, M.F. and Popelar, C.H.: Advanced and Relationship to Port Plug Condition,” paper SPE
Fracture Mechanics, New York, New York, USA, 13128, SPE Production Engineering (September
Oxford University Press (1985). 1986) 1, No. 5, 379–387.
Karakas, M. and Tariq, S.: “Semi-Analytical King, M.S.: “Static and Dynamic Elastic Properties
Productivity Models for Perforated Completions,” of Rock from the Canadian Shield,” International
paper SPE 18247, presented at the SPE Annual Journal of Rock Mechanics and Mining Sciences &
Technical Conference and Exhibition, Houston, Geomechanics Abstracts (October 1983) 20, No. 5,
Texas, USA (October 2–5, 1988). 237–241.
Karakas, M., Yokoyama, Y. and Arima, E.: “Well Kirkby, L.L. and Rockefeller, H.A.: “Proppant
Test Analysis of a Well with Multiple Horizontal Settling Velocities in Nonflowing Slurries,” paper
Drainholes,” paper SPE 21424, presented at the SPE/DOE 13906, presented at the SPE/DOE Low
Middle East Oil Show, Bahrain (November 16–19, Permeability Gas Reservoirs Symposium, Denver,
1991). Colorado, USA (May 19–22, 1985).
Kaufman, H. and Falcetta, J.: Introduction to Polymer Knoll, S.K.: “Wall Slip Evaluation in Steady Shear
Science and Technology, John Wiley and Sons, New Viscosity Measurements of Hydraulic Fracturing
York, New York, USA (1977), 153. Fluids,” paper SPE/DOE 13904, presented at the
SPE/DOE Low Permeability Gas Reservoirs Sym-
Kern, L.R., Perkins, T.K. and Wyant, R.E.: “The
posium, Denver, Colorado, USA (May 19–22, 1985).
Mechanics of Sand Movement in Fracturing,”
Journal of Petroleum Technology (July 1959), 55; Kooijman, A.P., Halleck, P.M., de Bree, P. and
also in Trans., AIME (1959) 216, 403–405. Kenter, C.J: “Large Scale Laboratory Sand Pro-
duction Test,” paper SPE 24798, presented at the
Kesavan, S., Prud’homme, R.K. and Parris, M.D.:
“Crosslinked Borate HPG Equilibria and Rheological
Washington, D.C., USA (October 4–7, 1992).
Characterization,” paper SPE 25205, presented at the
SPE International Symposium on Oilfield Chemistry, Kowalski, J.: “Formation Strength Parameter from
New Orleans, Louisiana, USA (March 2–5, 1993). Well Logs,” Transactions of the SPWLA 16th Annual
Logging Symposium (June 4–7, 1975), paper N.
Khristianovich, S.A. and Zheltov, Y.P.: “Formation
of Vertical Fractures by Means of Highly Viscous Kuchuk, F.J., Goode, P.A., Brice, B.W., Sherrard,
Liquid,” Proc., Fourth World Pet. Congress, Rome D.W. and Thambynayagam, R.K.M.: “Pressure
(1955) 2, 579–586. Transient Analysis and Inflow Performance for
Horizontal Wells,” paper SPE 18300, presented at the
Khristianovich, S.A., Zheltov, Y.P., Barenblatt, G.I.
and Maximovich, G.K.: “Theoretical Principles of
Houston, Texas, USA (October 2–5, 1988).
Hydraulic Fracturing of Oil Strata,” Proc., Fifth
World Petroleum Congress, New York (1959). Kulhman, R.D., Heemstra, T.R., Ray, T.G., Lin, P.
and Charlez, P.A.: “Field Tests of Downhole Exten-
Kiel, O.M.: “Process of Hydraulic Fracturing with
someter Used to Obtain Formation In-Situ Stress
Viscous Oil-in-Water Emulsion,” U.S. Patent No.
Data,” paper SPE 25905, Proc., Rocky Mountain/
3,552,494 (January 5, 1971).
R-22 References

Low Permeability Reservoirs Symposium, Denver, Lee, M.H. and Roberts, L.D.: “Effect of Heat of
Colorado, USA (1993), 625–634. Reaction on Temperature Distribution in a Fracture,”
paper SPE 7893, SPE Journal (December 1980) 20,
Labrid, J.C.: “Thermodynamic and Kinetic Aspects
No. 12, 501–507.
of Argillaceous Sandstone Acidizing,” paper SPE
5165, SPE Journal (April 1975), 117–128. Lee, M.Y. and Haimson, B.C.: “Statistical Evaluation
of Hydraulic Fracturing Stress Measurement
Labuz, J.F., Shah, S.P. and Dowding, C.H.:
Parameters,” International Journal of Rock
“Experimental Analysis of Crack Propagation in
Granite,” International Journal of Rock Mechanics
Abstracts (December 1989) 26, No. 6, 447–456.
and Mining Sciences & Geomechanics Abstracts
(April 1985) 22, No. 2, 85–98. Lee, W.J. and Holditch, S.A.: “Application of
Pseudotime to Buildup Test Analysis of Low-
Lagrone, K.W. and Rasmussen, J.W.: “A New
Permeability Gas Wells with Long-Duration Wellbore
Development in Completion Methods—The Limited
Storage Distortion,” paper SPE 9888, Journal of
Entry Technique,” paper SPE 530, Journal of
Petroleum Technology (December 1982) 34, No.
Petroleum Technology (July 1963) 15, 695–702.
12, 2877–2887.
Lamb, H.: Hydrodynamics, sixth edition, New York,
Leighton, D. and Acrivos, A.: “The Shear-Induced
New York, USA, Dover Publications (1932),
Migration of Particles in Concentrated Suspensions,”
581–587.
J. Fluid Mech. (1987) 181, 415–439.
Lambert, M.E.: “A Statistical Study of Reservoir
Lenoach, B.: “Hydraulic Fracture Model Based on
Heterogeneity,” MS thesis, The University of Texas
Analytical Near-Tip Solutions,” Computer Methods
at Austin, Austin, Texas, USA (1981).
and Advances in Geomechanics (1994), 1597–1602.
Lane, H.S.: “Numerical Simulation of Mud Filtrate
Lenoach, B.: “The Crack Tip Solution for Hydraulic
Invasion and Dissipation,” Transactions of the
Fracturing in a Permeable Solid,” J. Mech. Phys.
SPWLA 34th Annual Logging Symposium (June
Solids (1995) 43, No. 7, 1025.
13–16, 1993), paper D.
Levine, D.C., Thomas, E.W., Bezner, H.P. and Talle,
Larsen, L. and Hegre, T.M.: “Pressure-Transient
G.C.: “Annular Gas Flow After Cementing: A Look
Behavior of Horizontal Wells with Finite-
at Practical Solutions,” paper SPE 8255, presented at
Conductivity Vertical Fractures,” paper SPE 22076,
the SPE Annual Technical Conference and Exhi-
presented at the International Arctic Technology
bition, Las Vegas, Nevada, USA (September 23–
Conference, Anchorage, Alaska, USA (May 29–31,
26, 1979).
1991).
Li, Y., Sullivan, R.B., de Rozières, J., Gaz, G.L. and
Last, N., Plumb, R., Harkness, R., Charlez, P., Alsen,
Hinkel, J.J.: “An Overview of Current Acid Frac-
J. and McLean, M.: “An Integrated Approach to
turing Technology with Recent Implications for
Evaluating and Managing Wellbore Instability in the
Emulsified Acids,” paper SPE 26581, presented at the
Cusiana Field, Colombia, South America,” paper
(October 22–25, 1995). Lipton, D. and Burnett, D.B.: “Comparisons of
Polymers Used in Workover and Completion Fluids,”
Lea, F.M.: The Chemistry of Cement and Concrete,
paper SPE 5872, presented at the Improved Oil
third edition, New York, New York, USA, Chemical
Recovery Symposium, Tulsa, Oklahoma, USA
Publishing Co. (1971).
(March 22–24, 1976).
Lear, T.A.: “Near-Wall Velocity Measurements Using
Locke, S.: “An Advanced Method for Predicting the
Laser Doppler Velocimetry in the FFCF Simulator,”
Productivity Ratio of a Perforated Well,” paper SPE
MS thesis, University of Oklahoma, Oklahoma, USA
8804, Journal of Petroleum Technology (December
(1996).
1981) 33, 2481–2488.

Lord, D.L., Shah, S.N., Rein, R.G. and Lawson, J.T. Multiphase Pressure Buildup Analyses,” paper SPE
III: “Study of Perforation Friction Pressure 1235-G, Journal of Petroleum Technology (October
Employing a Large-Scale Fracturing Flow Sim- 1959) 11, 321–323; also in Trans., AIME (1959) 216.
ulator,” paper SPE 28508, presented at the SPE
Martinez, A.D., Pardo, C.W., Ellis, P.D. and Pittman,
R.W.: “Determination of Fracture Geometry Using
Field Data Coupled with a Fracture Simulator,” paper
Love, A.E.H.: A Treatise on the Mathematical Theory SPE 24826, presented at the SPE Annual Technical
of Elasticity, fourth edition, New York, New York, Conference and Exhibition, Washington, D.C., USA
USA, Dover Publications (1927). (October 4–7, 1992).
Lubinski, A.: “The Theory of Elasticity for Porous Martins, J.P., Abel, J.C., Dyke, C.G., Michel, C.M.
Bodies Displaying a Strong Pore Structure,” Proc., and Stewart, G.: “Deviated Well Fracturing and
2nd U.S. National Congress of Applied Mechanics Proppant Production Control in the Prudhoe Bay
(1954), 247. Field,” paper SPE 24858, presented at the SPE
Lund, K. and Fogler, H.S.: “Acidization V. The
Washington, D.C., USA (October 4–7, 1992a).
Prediction of the Movement of Acid and Permeability
Fronts in Sandstone,” Chemical Engineering Science Martins, J.P., Collins, P.J., Rylance, M., Ibe, O.E.,
(1976) 31, No. 5, 381–392. Kelly, R.T. and Bartel, P.A.: “Small, Highly Con-
ductive Fractures Near Reservoir Fluid Contacts:
Mack, M.G. and Elbel, J.L.: “A Simulator for
Application to Prudhoe Bay,” paper SPE 24856,
Modeling Acid Fracturing Treatments,” Proc., 8th
International Conference on Computer Methods and
and Exhibition, Washington, D.C., USA (October 4–
Advances in Geomechanics, Morgantown, West
7, 1992b).
Virginia, USA (1994).
Martins, J.P. and Harper, T.R.: “Mini-Frac Pressure
Mack, M.G., Piggott, A.R. and Elbel, J.L.: “Numer-
Decline Analysis for Fractures Evolving from Long
ical Representation of Multilayer Hydraulic Frac-
Perforated Intervals and Unaffected by Confining
turing,” Proc., 33rd U.S. Symposium on Rock
Strata,” paper SPE 13869, presented at the SPE/DOE
Mechanics, Santa Fe, New Mexico, USA (1992),
Low Permeability Gas Reservoirs Symposium,
335–344.
Denver, Colorado, USA (May 19–22, 1985).
Mahoney, J.V., Stubbs, P.B., Schwerer, F.C. III and
Martins, J.P., Leung, K.H., Jackson, M.R., Stewart,
Dobscha, F.X.: “Effects of a No-Proppant Foam
D.R. and Carr, A.H.: “Tip Screen-Out Fracturing
Stimulation Treatment on a Coal-Seam Degasifi-
Applied to the Ravenspurn South Gas Field Devel-
cation Borehole,” Journal of Petroleum Technology
opment,” paper SPE 19766, presented at the SPE
(November 1981), 2227.
Malone, B.P., Moran, J.H., Nasution, R.D., Putra, Antonio, Texas, USA (October 8–11, 1989; also in
P.H. and Thurston, F.K.: “Start-Up of a TSO Frac- SPE Production Engineering (August 1992c) 7, No.
turing Campaign in Shallow, Heavy Oil Steamflood,” 3, 252–258.
paper SPE 38096, presented at the Asia Pacific Oil
Martins, J.P., Milton-Tayler, D. and Leung, H.K.:
and Gas Conference, Kuala Lumpur, Malaysia (April
“The Effects of Non-Darcy Flow in Propped
14–16, 1997).
Hydraulic Fractures,” paper SPE 20709, presented
Maloney, D.R., Gall, B.L. and Raible, C.J.: “Non- at the SPE Annual Technical Conference and Exhi-
Darcy Flow Through Propped Fractures: Effect of bition, New Orleans, Louisiana, USA (September
Partial Saturation, Gel Damage, and Stress,” paper 23–26, 1990).
Mason, J.N., Dees, J.M. and Kessler, N.: “Block
Tests Model the Near-Wellbore in a Perforated
(September 27–30, 1987).
Sandstone,” paper SPE 28554, presented at the SPE
Martin, J.C.: “Simplified Equations of Flow in Gas Annual Technical Conference and Exhibition, New
Drive Reservoirs and the Theoretical Foundation of Orleans, Louisiana, USA (September 25–28, 1994).
R-24 References

Mathur, A.K., Ning, X., Marcinew, R.B., Ehlig- McGuire, W.J. and Sikora, V.T.: “The Effect of Ver-
Economides, C.A. and Economides, M.J.: “Hydraulic tical Fractures on Well Productivity,” paper SPE
Fracture Stimulation of Highly Permeable Forma- 1618-G, Journal of Petroleum Technology (October
tions: The Effect of Critical Fracture Parameters on 1960) 12, 72–74; also in Trans., AIME (1960) 219,
Oilwell Production and Pressure,” paper SPE 30652, 401–403.
McKenzie, L.F. and Hughes, B.J.: “Hydrocarbon
and Exhibition, Dallas, Texas, USA (October 22–
Gels of Alumino Alkyl Acid Orthophosphates,”
25, 1995).
paper SPE 9007, presented at the 5th International
Maury, V.: “Rock Failure Mechanisms Identification: Symposium on Oilfield and Geothermal Chemistry,
A Key for Wellbore Stability and Reservoir Behav- Stanford, California, USA (May 28–30, 1980).
iour Problems,” Proc., Eurock ’94, Rock Mechanics
Meehan, D.N., Horne, R.N. and Aziz, K.: “Effects
in Petroleum Engineering, Delft, Netherlands (1994),
of Reservoir Heterogeneity and Fracture Azimuth on
175–182.
Optimization of Fracture Length and Well Spacing,”
Maury, V., Etchecopar, A. and Pezard, P.A.: “New paper SPE 17606, presented at the SPE International
Borehole Imagery Techniques: An Aid for Failure Meeting on Petroleum Engineering, Tianjin, China
Modes and In Situ Stress Analysis and for Mini- (November 1–4, 1988).
mizing Drilling Incidents,” Transactions of the
Meese, C.A., Mullen, M.E. and Barree, R.D.: “Off-
SPWLA 40th Annual Logging Symposium, Olso,
shore Hydraulic Fracturing Technique,” paper SPE
Norway (June 8–10, 1999).
28159, Journal of Petroleum Technology (March
Mayerhofer, M.J., Economides, M.J. and Ehlig- 1994) 46, No. 3, 226–229.
Economides, C.A.: “Pressure Transient Analysis of
Meng, H.Z. and Brown, K.E.: “Coupling of Pro-
Fracture Calibration Tests,” paper SPE 26527, pre-
duction Forecasting, Fracture Geometry Require-
sented at the SPE Annual Technical Meeting and
ments and Treatment Scheduling in the Optimum
Exhibition, Houston, Texas, USA (October 3–
Hydraulic Fracture Design,” paper SPE 16435, pre-
6, 1993).
sented at the SPE/DOE Low Permeability Reservoirs
Mayerhofer, M.J., Economides, M.J. and Nolte, K.G.: Symposium, Denver, Colorado, USA (May 18–
“An Experimental and Fundamental Interpretation of 19, 1987).
Fracturing Filter-Cake Fluid Loss,” paper SPE 22873,
Meng, H.Z., Proano, E.A., Buhidma, I.M. and Mach,
J.M.: “Production Systems Analysis of Vertically
Fractured Wells,” paper SPE/DOE 10842, presented
9, 1991).
at the SPE/DOE Unconventional Gas Recovery
McCabe, M.A., Terracina, J.M. and Kunzl, R.A.: Symposium, Pittsburgh, Pennsylvania, USA (May
“Continuously Gelled Diesel Systems for Fracturing 16–18, 1982).
Applications,” paper CIM/SPE 90-93, presented at
Menjivar, J.: “On the Use of Gelation Theory to
the Petroleum Society of CIM/SPE International
Characterize Metal Crosslinked Polymer Gels,”
Technical Meeting, Calgary, Alberta, Canada (June
Proc., ACS Division of Polymer Materials: Science
10–13, 1990).
and Engineering, Washington, D.C., USA, American
McClain, C.: Fluid Flow in Pipes, New York, New Chemical Society (1984), 88–95.
York, USA, The Industrial Press (1963), 117–128.
Meyer, B.R.: “Design Formulae for 2-D and 3-D
McConnell, B.: “Fracturing Fluid High-Temperature Vertical Hydraulic Fractures: Model Comparison and
Breaker for Improving Well Performance,” paper Parametric Studies,” paper SPE 15240, presented at
SPE 28219, Journal of Petroleum Technology (May the SPE Unconventional Gas Technology Sympo-
1994) 46, No. 5, 382. sium, Louisville, Kentucky, USA (May 18–21, 1986a).
McDaniel, R.R., Deysarkar, A.K., Callanan, M.J. and Meyer, B.R.: “Generalized Drag Coefficient Appli-
Kohlhaas, C.A.: “An Improved Method for Mea- cable for All Flow Regimes,” Oil & Gas Journal
suring Fluid Loss at Simulated Fracture Conditions,” (May 1986b), 71–77.
paper SPE 10259, SPE Journal (August 1985), 482.

Meyer, B.R.: “Heat Transfer in Hydraulic Frac- Mondshine, T.C.: “Process for Decomposing Poly-
turing,” paper SPE 17041, presented at the SPE saccharides in Alkaline Aqueous Systems,” U.S.
Eastern Regional Meeting, Pittsburgh, Pennsylvania, Patent No. 5,253,711 (October 19, 1993).
USA (October 21–23, 1987).
Montgomery, K.T. and Berthelot, J.M.: “Effects of
Meyer, B.R. and Hagel, M.W.: “Simulated Mini-Frac Fracture Fluid Invasion on Cleanup Behavior and
Analysis,” Journal of Canadian Petroleum Engi- Pressure Buildup Analysis,” paper SPE 20643, pre-
neering (September–October 1989) 28, No. 5, 63–73. sented at the SPE Annual Technical Conference and
Exhibition, New Orleans, Louisiana, USA (Sep-
Michels, P.C. and Clark, D.S.: “Pressure Dependence
tember 23–26, 1990).
of Enzyme Catalysis,” Biocatalysis at Extreme Tem-
peratures, M.W.W. Adams and R.M. Kelly (eds.), Monus, F.L., Broussard, F.W., Ayoub, J.A. and
Symposium Series, Washington, D.C., USA, Amer- Norman, W.D.: “Fracturing Unconsolidated Sand
ican Chemical Society (1992), 498, 108–121. Formations Offshore Gulf of Mexico,” paper SPE
Miller, B.D. and Warembourg, P.A.: “Prepack Tech-
Conference and Exhibition, Washington, D.C., USA
nique Using Fine Sand Improves Results of Frac-
(October 4–7, 1992).
turing and Fracture Acidizing Treatments,” paper
SPE 5643, presented at the SPE Annual Technical Moore, J.E.: “Clay Mineralogy Problems in Oil
Conference and Exhibition, Dallas, Texas, USA Recovery,” Petroleum Engineering (1960) B40.
(September 28–October 1, 1975).
Morales, R.H., Brady, B.H. and Ingraffea, A.R.:
Miller, W.K. II and Smith, M.B.: “Reanalysis of the “Three-Dimensional Analysis and Visualization of
MWX Fracture Stimulation Data from the Paludal the Wellbore and the Fracturing Process in Inclined
Zone of the Mesaverde Formation,” paper SPE Wells,” paper SPE 25889, presented at the SPE
19772, SPE Annual Technical Conference and Exhi- Rocky Mountain Regional/Low Permeability Reser-
bition, San Antonio, Texas, USA (October 8–11, 1989). voirs Symposium, Denver, Colorado, USA (April
12–14, 1993).
Millheim, K.K. and Cichowicz, L.: “Testing and
Analyzing Low-Permeability Fractured Gas Wells,” Morales, R.H., Brown, J.E., Norman, W.D., DeBonis,
paper SPE 1768, Journal of Petroleum Technology V., Mathews, M., Park, E.I. and Brown, D.: “Mech-
(February 1968) 20, 193–198; also in Trans., AIME anical Skin Damage on Wells,” paper SPE 30459,
(1968) 243. presented at the SPE Annual Technical Conference
Minh, C.C., Freedman, R., Crary, S. and Cannon, D.:
25, 1995).
“Integration of NMR with Other Openhole Logs for
Improved Formation Evaluation,” paper SPE 49012, Morales, R.H. and Marcinew, R.P.: “Fracturing of
presented at the SPE Annual Technical Conference High-Permeability Formations: Mechanical Prop-
and Exhibition, New Orleans, Louisiana, USA erties Correlations,” paper SPE 26561, presented
(September 27–30, 1998). at the SPE Annual Technical Conference and Exhi-
bition, Houston, Texas, USA (October 3–6, 1993).
Misak, M.D.: “Temperature-Stable Aqueous Gels,”
U.S. Patent No. 3,922,173 (November 25, 1975). Moricca, G., Ripa, G., Sanfilippo, F. and Santarelli,
F.J.: “Basin Scale Rock Mechanics: Field Obser-
Mitchell, B.J.: “Viscosity of Foam,” PhD dissertation,
vation of Sand Production,” Proc., Eurock ’94, Rock
Oklahoma University, Norman, Oklahoma, USA
Mechanics in Petroleum Engineering, Delft, Nether-
(1969).
lands (1994), 317–328.
Mitchell, B.J.: “Test Data Fill Theory Gap on Using
Morita, N., Burton, R.C. and Davis, E.: “Fracturing,
Foam as a Drilling Fluid,” Oil & Gas Journal
Frac Packing and Formation Control: Can Screenless
(September 6, 1971), 96–100.
Completions Prevent Sand Production,” SPE Drilling
Mondshine, T.C.: “Crosslinked Fracturing Fluids,” and Completion (September 1998), 157–162.
U.S. Patent No. 4,619,776 (October 28, 1986).
R-26 References

Morita, N., Whitfill, D.L., Fedde, O.P. and Levik, also paper SPE 28529-P, SPE Production & Facilities
T.H.: “Parametric Study of Sand Production Pre- (August 1996), 138.
diction: Analytical Approach,” paper SPE 16990,
Navarrete, R.C. and Mitchell, J.P.: “Fluid-Loss
SPE Production Engineering (February 1989) 4, No.
Control for High-Permeability Rocks in Hydraulic
1, 25–33.
Fracturing Under Realistic Shear Conditions,” paper
Morita, N., Whitfill, D.L. and Wahl, H.A.: “Stress- SPE 29504, presented at the SPE Production Oper-
Intensity Factor and Fracture Cross-Sectional Shape ations Symposium, Oklahoma City, Oklahoma, USA
Predictions from a Three-Dimensional Model for (April 2–4, 1995).
Hydraulically Induced Fractures,” paper SPE 14262,
Nelson, E.B.: “Introduction,” Well Cementing, E.B.
Journal of Petroleum Technology (October 1988) 40,
Nelson (ed.), Houston, Texas, USA, Schlumberger
No. 10, 1329–1342.
(1990), I-1–I-3.
Morris, C.W. and Sinclair, A.R.: “Evaluation of
Nelson, E.B., Baret, J.-F. and Michaux, M.: “Cement
Bottomhole Treatment Pressure for Geothermal Well
Additives and Mechanisms of Action,” Well
Hydraulic Fracture Stimulation,” paper SPE 11157,
Cementing, E.B. Nelson (ed.), Houston, Texas, USA,
Journal of Petroleum Technology (May 1984), 829.
Schlumberger (1990), 3-1–3-37.
Morse, R.A. and Von Gonten, W.D.: “Productivity of
Nelson, E.B., Constien, V.G. and Cawiezel, K.E.:
Vertically Fractured Wells Prior to Stabilized Flow,”
“Delayed Borate Crosslinked Fracturing Fluid
paper SPE 3631, Journal of Petroleum Technology
Having Increased Temperature Range,” U.S. Patent
(July 1972) 24, 807–812.
No. 5,658,861 (1997).
Mukherjee, H., Paoli, B.F., McDonald, T., Cartaya,
Nester, J.H., Jenkins, D.R. and Simon, R.: “Resis-
H. and Anderson, J.A.: “Successful Control of Frac-
tances to Failure of Oil-Well Casing Subjected to
ture Height Growth by Placement of Artificial Bar-
Non-Uniform Transverse Loading,” Drilling and
rier,” paper SPE 25917, presented at the SPE Rocky
Production Practice, New York, New York, USA,
Mountain Regional/Low Permeability Reservoirs
American Petroleum Institute (1956), 374–378.
1994); also in SPE Production & Facilities (May Nguyen, H.X. and Larson, D.B.: “Fracture Height
1995), 89. Containment by Creating an Artificial Barrier with
a New Additive,” paper SPE 12061, presented at the
Mullen, M.E., Norman, W.D. and Granger, J.C.:
“Productivity Comparison of Sand Control Tech-
San Francisco, California, USA (October 5–8, 1983).
niques Used for Completions in the Vermilion 331
Field,” paper SPE 27361, presented at the SPE Inter- Nierode, D.E. and Kruk, K.F.: “An Evaluation of
national Symposium on Formation Damage Control, Acid Fluid Loss Additives, Retarded Acids, and
Lafayette, Louisiana, USA (February 7–10, 1994). Acidized Fracture Conductivity,” paper SPE 4549,
presented at the SPE Annual Meeting, Las Vegas,
Muskat, M.: The Flow of Homogeneous Fluids
Nevada, USA (September 30–October 3, 1973).
Through Porous Media, New York, New York, USA,
McGraw-Hill Book Co. (1937). Nikolaevskiy, V.N. and Economides, M.J.: “The
Near-Well State of Stress and Induced Rock Dam-
Muskhelishvili, N.I.: Some Basic Problems of the
age,” paper SPE 58716, presented at the SPE Inter-
Theory of Elasticity, fourth edition, J.R. Radok (trans.),
national Symposium on Formation Damage, Lafay-
Groningen, Netherlands, P. Noordhoff (1953).
ette, Louisiana, USA (February 23–24, 2000).
Navarrete, R.C., Cawiezel, K.E. and Constien, V.G.:
Nimerick, K.H., McConnell, S.B. and Samuelson,
“Dynamic Fluid Loss in Hydraulic Fracturing Under
M.L.: “Compatibility of Resin-Coated Proppants with
Realistic Shear Conditions in High-Permeability
Crosslinked Fracturing Fluids,” paper SPE 20639,
Rocks,” paper SPE 28529, presented at the SPE
(September 23–26, 1990); also in SPE Production
Engineering (February 1992), 29.

Nolte, K.G.: “Determination of Fracture Parameters New Orleans, Louisiana, USA (September 23–26,
from Fracturing Pressure Decline,” paper SPE 8341, 1990); also in Journal of Petroleum Technology
presented at the SPE Annual Technical Conference (February 1991) 43, No. 2, 210–218.
and Exhibition, Las Vegas, Nevada, USA (September
Nolte, K.G.: “Background for After-Closure Analysis
23–26, 1979).
of Calibration Treatments,” paper SPE 39407 (1998).
Nolte, K.G.: “Fracture Design Considerations Based
Nolte, K.G. and Economides, M.J.: “Fracture Design
on Pressure Analysis,” paper SPE 10911, presented at
and Validation with Uncertainty and Model Limi-
the SPE Cotton Valley Symposium, Tyler, Texas,
tations,” paper SPE 18979, Journal of Petroleum
USA (May 20, 1982).
Technology (September 1991) 43, No. 9, 1147–1155.
Nolte, K.G.: “Fracturing Fluid Breaker System Which
Nolte, K.G., Mack, M.G. and Lie, W.L.: “A System-
Is Activated by Fracture Closure,” U.S. Patent No.
atic Method for Applying Fracturing Pressure
4,506,734 (March 26, 1985).
Decline: Part 1,” paper SPE 25845, presented at the
Nolte, K.G.: “A General Analysis of Fracturing SPE Rocky Mountain Regional Meeting and Low
Pressure Decline Analysis with Application to Three Permeability Reservoirs Symposium, Denver,
Models,” paper SPE 12941, SPE Formation Colorado, USA (April 12–14, 1993).
Evaluation (December 1986a) 1, No. 6, 571–583.
Nolte, K.G., Maniere, J.L. and Owens, K.A.: “After-
Nolte, K.G.: “Determination of Proppant and Fluid Closure Analysis of Fracture Calibration Tests,”
Schedules from Fracturing Pressure Decline,” paper paper SPE 38676, presented at the SPE Annual
SPE 13278, SPE Production Engineering (July 1986b) Technical Conference and Exhibition, San Antonio,
1, No. 4, 255–265. Texas, USA (October 5–8, 1997).
Nolte, K.G.: “Discussion of Influence of Geologic Nolte, K.G. and Smith, M.B.: “Interpretation of
Discontinuities on Hydraulic Fracture Propagation,” Fracturing Pressures,” paper SPE 8297, presented
paper SPE 17011, Journal of Petroleum Technology at the SPE Annual Technical Conference and Exhi-
(August 1987) 39, No. 8, 998. bition, Las Vegas, Nevada, USA (September 23–26,
1979); also in Journal of Petroleum Technology
Nolte, K.G.: “Application of Fracture Design Based
(September 1981) 33, No. 9, 1767–1775.
on Pressure Analysis,” paper SPE 13393, SPE
Production Engineering (February 1988a) 3, No. Nordgren, R.P.: “Propagation of a Vertical Hydraulic
1, 31–42. Fracture,” paper SPE 7834, SPE Journal (August
1972) 12, No. 8, 306–314.
Nolte, K.G.: “Fluid Flow Considerations in Hydraulic
Fracturing,” paper SPE 18537, Proc., SPE Eastern Norman, L.R.: “Properties and Early Field Results
Regional Meeting, Charleston, West Virginia, USA of a Liquid Gelling Agent for Acid,” paper SPE 7834,
(1988b), 145–156. presented at the SPE Production Technology Sym-
posium, Hobbs, New Mexico, USA (October 30–
Nolte, K.G.: “Principles for Fracture Design Based
31, 1978).
on Pressure Analysis,” paper SPE 10911, SPE
Production Engineering (February 1988c) 3, No. Norman, L.R., Hollenbeak, K.H. and Harris, P.C.:
1, 22–30. “Fracture Conductivity Impairment Removal,” paper
Nolte, K.G.: “Derivations and Considerations for
Pressure Decline Analysis,” Appendix J of Recent
Advances in Hydraulic Fracturing, J.L. Gidley, S.A.
Holditch, D.E. Nierode and R.W. Veatch Jr. (eds.), Norman, L.R., Terracina, J.M., McCabe, M.A. and
Monograph Series, Richardson, Texas, USA, Society Nguyen, P.D.: “Application of Curable Resin-Coated
of Petroleum Engineers (1989), 410–415. Proppants,” paper SPE 20640, presented at the SPE
Nolte, K.G.: “Fracturing Pressure Analysis for Non-
Ideal Behavior,” paper SPE 20704, presented at the
also in SPE Production Engineering (November
1992), 343.
R-28 References

Norman, R.O.C.: Principles of Organic Synthesis, Palmer, I.D.: “Induced Stresses due to Propped
London, UK, Methuen & Co., Ltd. (1968). Hydraulic Fracture in Coalbed Methane Wells,” paper
SPE 25861, Proc., Rocky Mountain Regional/Low
Norris, M.R., Berntsen, B.A., Myhre, P. and Winters,
Permeability Reservoirs Symposium, Denver,
W.J.: “Multiple Proppant Fracturing of a Horizontal
Colorado, USA (1993), 221–232.
Wellbore: An Integration of Two Technologies,”
paper SPE 36899, presented at the SPE European Palmer, I.D. and Veatch, R.W. Jr.: “Abnormally High
Petroleum Conference, Milan, Italy (October 22– Fracturing Pressures in Step-Rate Tests,” paper SPE
24, 1996). 16902, presented at the SPE Annual Technical
Novotny, E.J.: “Proppant Transport,” paper SPE
(September 27–30, 1987).
ference and Exhibition, Denver, Colorado, USA Papamichos, E., Vardoulakis, I. and Ouadfel, H.:
(October 9–12, 1977). “Permeability Reduction due to Grain Crushing
Around a Perforation,” International Journal of Rock
Olejniczak, S.J., Swaren, J.A., Gulrajani, S.N. and
Olmstead, C.C.: “Fracturing Bypassed Pay in Tubing-
Abstracts (December 1993), 30, No. 7, 1223–1229.
less Completions,” paper SPE 56467, presented at
the SPE Annual Technical Conference and Exhi- Papanastasiou, P.: “A Coupled Elasto-Plastic
bition, Houston, Texas, USA (October 3–6, 1999). Hydraulic Fracturing Model,” International Journal
of Rock Mechanics and Mining Sciences & Geome-
Ong, S.H. and Roegiers, J.-C.: “Influence of Aniso-
chanics Abstracts (1997) 34, 240.
tropies in Borehole Stability,” International Journal
of Rock Mechanics and Mining Sciences & Geome- Papanastasiou, P. and Thiercelin, M.: “Influence of
chanics Abstracts (June 1993) 30, No. 7, 1069–1075. Inelastic Rock Behavior in Hydraulic Fracturing,”
International Journal of Rock Mechanics and Mining
Ott, R.E., Bell, W.T., Golian, T.G. and Harrigan, J.W.
Sciences & Geomechanics Abstracts (June 1993) 30,
Jr.: “Simple Method Predicts Downhole Shaped
No. 7, 1241–1247.
Charge Gun Performance,” paper SPE 27424, SPE
Production & Facilities (August 1994a) 9, No. 3, Papanastasiou, P., Thiercelin, M., Cook, J. and
171–178. Durban, D.: “The Influence of Plastic Yielding on
Breakdown Pressure in Hydraulic Fracturing,” Proc.,
Ott, R.E., Bell, W.T., Golian, T.G. and Harrigan, J.W.
35th U.S. Symposium on Rock Mechanics, Reno,
Jr.: “Authors’ Reply to Discussion of Simple Method
Nevada, USA (1995), 281–286.
Predicts Downhole Shaped Charge Gun Perfor-
mance,” paper SPE 30069, SPE Production & Papinczak, A. and Miller, W.K. II: “Fracture Treat-
Facilities (November 1994b) 9, No. 4, 294–295. ment Design to Overcome Severe Near-Wellbore
Damage in a Moderate Permeability Reservoir,
Ouchterlony, F.: “Review of Fracture Toughness in
Mereenie Field, Australia,” paper SPE 25379, pre-
Rock,” Solid Mechanics Archives (1982) 7, No. 2,
sented at the Asia Pacific Oil and Gas Conference
131–211.
and Exhibition, Singapore (February 8–10, 1993).
Ozkan, E. and Raghavan, R.: “New Solutions for
Parcevaux, P., Rae, P. and Drecq, P.: “Prevention
Well Test Analysis Problems: Part I—Analytical
of Annular Gas Migration,” Well Cementing, E.B.
Considerations,” paper SPE 18615, SPE Formation
Nelson (ed.), Houston, Texas, USA, Schlumberger
Evaluation (September 1991) 6, No. 3, 359–368.
(1990), 8-1–8-22.
Pacelli Zitha: “Effects of Bridging-Adsorption of
Parlar, M., Nelson, E.B., Walton, I.C., Park, E. and
Macromolecules on the Flow of Linear Flexible
Debonis, V.: “An Experimental Study on Fluid-Loss
Polymer Solutions in Porous Media,” C.R. Acad. Sci.
Behavior of Fracturing Fluids and Formation Damage
Paris, Serie IIb (1995), 320, 447–453.
in High-Permeability Porous Media,” paper SPE
Paillet, F.L. and Chang, C.H.: Acoustic Waves in 30458, presented at the SPE Annual Technical Con-
Boreholes: CRC Press, Inc., Boca Raton, Florida, ference and Exhibition, Dallas, Texas, USA (October
USA (1991). 22–25, 1995).

Pattillo, P.D. and Smith, M.B.: “The Effect of Perkins, T.K. and Kern, L.R.: “Widths of Hydraulic
Formation Flow on the Integrity of Perforated Fractures,” paper SPE 89, Journal of Petroleum
Casing,” paper SPE 11123, SPE Journal (October Technology (September 1961) 13, No. 9, 937–949.
1985) 25, No. 5, 637–646.
Perkins, T.K. and Krech, W.W.: “The Energy Balance
Payne, K.L. and Harms, S.D.: “Chemical Processes Concept of Hydraulic Fracturing,” paper SPE 1901,
for Controlling Crosslinking Reaction Rates,” pre- SPE Journal (March 1968) 8, No. 1, 1–12.
sented at the AIChE National Spring Meeting, Ana-
Perkins, T.K. and Weingarten, J.S.: “Stability and
heim, California, USA (May 20, 1984).
Failure of Spherical Cavities in Unconsolidated Sand
Pearson, C.M., Bond, A.J., Eck, M.E. and Schmidt, and Weakly Consolidated Rock,” paper SPE 18244,
J.H.: “Results of Stress Oriented and Aligned Perfo- presented at the SPE Annual Technical Conference
rating in Fracturing Deviated Wells,” paper SPE and Exhibition, Houston, Texas, USA (October 2–
22836, Journal of Petroleum Technology (January 5, 1988).
1992) 44, 10–18.
Perrine, R.L.: “Analysis of Pressure-Buildup
Penny, G.S.: “An Evaluation of the Effects of Envi- Curves,” Drilling and Production Practice 1956,
ronmental Conditions and Fracturing Fluids upon the New York, New York, USA, American Petroleum
Long-Term Conductivity of Proppants,” paper SPE Institute (1957), 482–509.
Peterson, R.E., Wolhart, S.L., Frohne, K.-H., Warp-
inski, N.R., Branagan, P.T. and Wright, T.B.: “Frac-
(September 27–30, 1987).
ture Diagnostics Research at the GRI/DOE Multi-Site
Penny, G.S., Conway, M.W. and Lee, W.: “Control Project: Overview of the Concept and Results,” paper
and Modeling of Fluid Leakoff During Hydraulic SPE 36449, presented at the SPE Annual Technical
Fracturing,” paper SPE 12486, Journal of Petroleum Conference and Exhibition, Denver, Colorado, USA
Technology (June 1985) 37, No. 6, 1071–1081. (October 6–9, 1996).
Penny, G.S. and Jin, L.: “The Development of Lab- Petitjean, L., Couet, B., Abel, J.C., Schmidt, J.H. and
oratory Correlations Showing the Impact of Multi- Ferguson, K.R.: “Well Productivity Improvement
phase Flow, Fluid, and Proppant Selection upon Gas Using Extreme Overbalanced Perforating and
Well Productivity,” paper SPE 30494, presented at Surging—Case History,” paper SPE 30527, presented
the SPE Annual Technical Conference and Exhi- at the SPE Annual Technical Conference and Exhi-
bition, Dallas, Texas, USA (October 22–25, 1995). bition, Dallas, Texas, USA (October 22–25, 1995).
Penny, G.S., Soliman, M.Y., Conway, M.W. and Pezron, E., Leibler, L., Ricard, A. and Audebert, R.:
Briscoe, J.E.: “Enhanced Load Water-Recovery Tech- “Glactomannan-Borate Systems: A Complexation
nique Improves Stimulation Results,” paper SPE Study,” Giol. Synt. Polym. Network, O. Kramer (ed.)
12149, presented at the SPE Annual Technical Con- (1988a), 113–126.
ference and Exhibition, San Francisco, California,
Pezron, E., Ricard, A., Lafuma, F. and Audebert, R.:
“Reversible Gel Formations Induced by Ion Com-
Perkins, T.K.: “Discussion of ‘On the Design of plexation,” Macromolecules (1988b) 21, No. 4,
Vertical Hydraulic Fractures,’” Journal of Petroleum 1121–1125.
Technology (January 1973) 25, No. 1, 80–81; also
Phelps, G.D., Stewart, G., Peden, J.M. and Heriot-
in Trans., AIME (1973) 255, 93–95.
Watt, U.: “The Analysis of the Invaded Zone
Perkins, T.K. and Gonzalez, J.A.: “Changes in Earth Characteristics and Their Influence on Wireline Log
Stresses Around a Wellbore Caused by Radially and Well-Test Interpretation,” paper SPE 13287,
Symmetrical Pressure and Temperature Gradients,” presented at the SPE Annual Technical Conference
paper SPE 10080, SPE Journal (April 1984), 129– and Exhibition, Houston, Texas, USA (September
140. 16–19, 1984).
R-30 References

Piggott, A.R., Brady, B.H. and Gu, H.: “Reservoir Poe, B.D. Jr., Shah, P.C. and Elbel, J.L.: “Pressure
Formation Characterization from Hydraulic Frac- Transient Behavior of a Finite-Conductivity Fractured
turing Records,” Proc., Eurock ’92 Symposium, Well with Spatially Varying Fracture Properties,”
London, UK (1992), 360–365. paper SPE 24707, presented at the SPE Annual Tech-
nical Conference and Exhibition, Washington, D.C.,
Plahn, S.V., Nolte, K.G., Thompson, L.G. and Miska,
S.: “A Quantitative Investigation of the Fracture
Pump-In/Flowback Test,” paper SPE 30504, pre- Pop, J.J., Badry, R.A., Morris, C.W., Wilkinson, D.J.,
sented at the SPE Annual Technical Conference and Tottrup, P. and Jonas, J.K.: “Vertical Interference
Exhibition, Dallas, Texas, USA (October 22–25, Testing with a Wireline-Conveyed Straddle-Packer
1995); also in SPE Production & Facilities (February Tool,” paper SPE 26481, presented at the SPE
1997) 12, No. 1, 20–27. Annual Technical Conference and Exhibition,
Plona, T.J. and Cook, J.M.: “Effects of Stress Cycles
on Static and Dynamic Young’s Moduli in Castlegate Pope, D., Britt, L., Constien, V., Anderson, A. and
Sandstones,” Proc., 35th U.S. Symposium on Rock Leung, L.: “Field Study of Guar Removal from
Mechanics, Reno, Nevada, USA (1995), 155–160. Hydraulic Fractures,” paper SPE 31094, presented
at the SPE International Symposium on Formation
Plumb, R.A.: “Influence of Composition and Texture
Damage Control, Lafayette, Louisiana, USA
on the Failure Properties of Clastic Rocks,” paper
(February 14–15, 1996).
SPE 28022, presented at the SPE/ISRM Rock Mec-
hanics in Petroleum Engineering Conference, Delft, Pope, D.S., Leung, L.K.-W., Gulbis, J. and Constien,
Netherlands (August 29–31, 1994a). V.G.: “Effects of Viscous Fingering on Fracture Con-
ductivity,” paper SPE 28511, presented at the SPE
Plumb, R.A.: “Variations of the Least Horizontal
Stress Magnitude in Sedimentary Rocks,” Proc., 1st
North American Symposium on Rock Mechanics,
Austin, Texas, USA (1994b), 71–78. Pospisil, G., Carpenter, C.C. and Pearson, C.M.:
“Impacts of Oriented Perforating on Fracture Stim-
Plumb, R.A. and Cox, J.W.: “Stress Directions in
ulation Treatments: Kuparuk River Field, Alaska,”
Eastern North American Determined to 4.5 Km from
paper SPE 29645, presented at the SPE Western
Borehole Elongation Measurements,” Journal of
Regional Meeting, Bakersfield, California, USA
Geophysical Research (1987) 92, No. B6, 4805–
(March 8–10, 1995).
4816.
Poupon, A., Clavier, C., Dumanoir, J., Gaynard,
Plumb, R.A., Evans, K.F. and Engelder, T.: “Geo-
R. and Misk, A.: “Log Analysis of Sand-Shale
physical Log Responses and Their Correlation with
Sequences: A Systematic Approach,” Journal
Bed-to-Bed Stress Contrasts in Paleozoic Rocks,
of Petroleum Technology (July 1970), 867–881.
Appalachian Plateau, New York,” Journal of Geo-
physical Research (August 1991) 96, No. B9, Power, D.J., Paterson, L. and Boger, D.V.: “The
14509–14528. Influence of Fluid Properties on the Success of
Hydraulic Fracturing Operations,” paper SPE 28804,
Plumb, R.A. and Hickman, S.H.: “Stress Induced
presented at the SPE Asia Pacific Oil & Gas Con-
Borehole Elongation: A Comparison Between Four-
ference, Melbourne, Australia (November 7–10,
Arm Dipmeter and the Borehole Televiewer in the
1994).
Auburn Geothermal Well,” Journal of Geophysical
Research (1985), 90, No. B7, 5513–5521. Prats, M.: “Effect of Vertical Fractures on Reservoir
Behavior—Incompressible Fluid Case,” paper SPE
Poe, B.D. Jr., Elbel, J.L., Spath, J.B. and Wiggins,
1575-G, SPE Journal (June 1961) 1, No. 1, 105–118;
M.L.: “Prediction of Future Well Performance,
also in Trans., AIME (1961) 222.
Including Reservoir Depletion Effects,” paper SPE
29465, presented at the SPE Operations Symposium, Prats, M.: “Effect of Burial History on the Subsurface
Denver, Colorado, USA (April 2–4, 1995). Horizontal Stresses of Formations Having Different

Material Properties,” paper SPE 9017, SPE Journal Pucknell, J.K. and Behrmann, L.A.: “An Investi-
(December 1981) 21, No. 6, 658–662. gation of the Damaged Zone Created by Perforating,”
Prats, M., Hazebroek, P. and Strickler, W.R.: “Effect
Technical Conference and Exhibition, Dallas, Texas,
of Vertical Fractures on Reservoir Behavior—Com-
pressible Fluid Case,” paper SPE 98, SPE Journal
(June 1962) 2, No. 2, 87–94. Quadir, J.A.: “Particle Transport by Non-Newtonian
Fluids in Vertical Slots,” MS thesis, The University
Press, W.H., Flannery, B.P., Teukolsky, S.A. and
of Tulsa, Tulsa, Oklahoma, USA (1981).
Vetterling, W.T.: Numerical Recipes—The Art of
Scientific Computing, New York, New York, USA, Quirein, J., Kimminau, S., LaVigne, J., Singer, J. and
Cambridge University Press (1986). Wendel, F.: “A Coherent Framework for Developing
and Applying Multiple Formation Evaluation
Proano, E.A. and Lilley, I.J.: “Derivative of Pressure:
Models,” Transactions of the SPWLA 27th Annual
Application to Bounded Reservoir Interpretation,”
Logging Symposium (June 9–13, 1986), paper DD.
paper SPE 15861, presented at the SPE European
Petroleum Conference, London, UK (October 20– Raghavan, R.: “Some Practical Considerations in the
22, 1986). Analysis of Pressure Data,” paper SPE 5876, Journal
of Petroleum Technology (October 1976) 28, 1256–
Prud’homme, R.K.: “Rheological Characterization
1268.
of Fracturing Fluids,” final report, API Practices
Projects 84-45, 85-45 and 86-45, American Petrol- Raghavan, R., Chen, C. and Agarwal, B.: “An
eum Institute (1984–1986). Analysis of Horizontal Wells Intercepted by Multiple
Fractures,” paper SPE 27652, presented at the SPE
Prud’homme, R.K.: “Laboratory Preparation and
Rheological Characterization of Crosslinked Frac-
turing Fluids,” paper SPE 20637 (1990).
Raghavan, R. and Joshi, S.D.: “Productivity of
Prud’homme, R.K.: “Rheology of Fracturing Fluid
Multiple Drainholes or Fractured Horizontal Wells,”
Gels and Slurries,” GRI/API Quarterly Progress
paper SPE 21263, SPE Formation Evaluation (March
Report (April 1991).
1993) 8, No. 1, 11–16.
Prud’homme, R.K., Constien, V.G. and Knoll, S.:
Rahim, Z. and Holditch, S.A.: “The Effects of
“The Effects of Shear History on the Rheology of
Mechanical Properties and Selection of Completion
Hydroxypropyl Guar Gels,” Polymers in Aqueous
Interval upon the Created and Propped Fracture
Media, J.E. Glass (ed.), Advances in Chemistry
Dimensions in Layered Reservoirs,” paper SPE
Series, Washington, D.C., USA, American Chemical
24349, presented at the SPE Rocky Mountain
Society (1989), 223.
Regional Meeting, Casper, Wyoming, USA (May
Prud’homme, R.K., Ellis, S., Constien, V.G. and 18–21, 1992).
Knoll, S.: “Reproducible Rheological Measurements
Ramakrishnan, T.S., Dave, Y. and Cappiello, A.:
on Crosslinked Fracturing Fluids,” paper SPE 18210,
“A Laboratory Investigation of Permeability in
Hemispherical Flow with Application to Formation
and Exhibition, Houston, Texas, USA (October 2–
Testers,” paper SPE 22689, presented at the SPE
5, 1988).
Prud’homme, R.K. and Khan, S.A. (eds.): Foams: Texas, USA (October 6–9, 1991).
Theory, Measurements, and Applications, Marcel
Ramlakhan, T.: “Characterization of the Altered Zone
Dekker, Inc. (1996).
Around a Shaped-Charge Perforation Using X-Ray
Prud’homme, R.K. and Wang, J.K.: “Filter-Cake Computerized Tomography,” thesis, Pennsylvania
Formation of Fracturing Fluids,” paper SPE 25207, State University, Pennsylvania, USA (1994).
presented at the SPE International Symposium on
Ramos, G.G., Katahara, K.W., Gray, J.D. and Knox,
Oilfield Chemistry, New Orleans, Louisiana, USA
D.J.W.: “Sand Production in Vertical and Horizontal
(March 2–5, 1993).
R-32 References

Wells in a Friable Sandstone Formation, North Sea,” Roberts, B.E., van Engen, H. and van Kruysdijk,
Proc., Eurock ’94, Rock Mechanics in Petroleum C.P.J.W.: “Productivity of Multiply Fractured Hori-
Engineering, Delft, Netherlands (1994), 309–315 zontal Wells in Tight Gas Reservoirs,” paper SPE
23113, presented at the SPE Offshore Europe Con-
Raymond, L.R. and Binder, G.G.: “Productivity of
ference, Aberdeen, Scotland (September 3–6, 1991).
Wells in Vertically Fractured, Damaged Formations,”
paper SPE 1454, Journal of Petroleum Technology Roberts, C.N.: “Fracture Optimization in a Tight Gas
(January 1967) 19, 120–130. Play: Muddy-J Formation, Wattenberg Field, Colo-
rado,” paper SPE 9851, presented at the SPE Low
Recommended Practice for Standard Procedures for
Permeability Symposium, Denver, Colorado, USA
Evaluation of Hydraulic Fracturing Fluids (RP 39),
(May 27–29, 1981).
second edition, Dallas, Texas, USA, American
Petroleum Institute (1983). Robinson, B.M., Holditch, S.A. and Peterson, R.E.:
“The Gas Research Institute’s Second Staged Field
Recommended Practices for Evaluation of Well
Experiment: A Study of Hydraulic Fracturing,” paper
Perforators (RP 43), fourth edition, New York, New
SPE 21495, presented at the SPE Gas Technology
York, USA, American Petroleum Institute (August
Symposium, Houston, Texas, USA (January 23–25,
1985).
1991).
Recommended Practices for Testing Sand Used in
Robinson, G., Ross-Murphy, S.B. and Morris, E.R.:
Hydraulic Fracturing Operations (RP 56), first edi-
“Viscosity-Molecular Weight Relationships, Intrinsic
tion, Dallas, Texas, USA, American Petroleum
Chain Flexibility, and Dynamic Solution Properties
Institute (1983).
of Guar Galactomannan,” Carbohydrate Research
Regalbuto, J.A. and Riggs, R.S.: “Underbalanced (1982) 107, 17–32.
Perforation Characteristics as Affected by Differential
Rochon, J., Creusot, M.R., Feugas, D., Thibeau, S.
Pressure,” paper SPE 15816, SPE Production Engi-
and Bergerot, J-L.: “Viscous Fluids Characterize the
neering (February 1988) 3, No. 1, 83–88.
Crushed Zone,” paper SPE 27383, SPE Drilling and
Reidenbach, V.G., Harris, P.C., Lee, Y.N. and Lord, Completion (September 1995) 10, No. 3, 198–203.
D.L.: “Rheological Study of Foam Fracturing Fluids
Rodriguez, F., Horne, R.N. and Cinco-Ley, H.: “Par-
Using Nitrogen and Carbon Dioxide,” paper SPE
tially Penetrating Vertical Fractures: Pressure Trans-
12026, SPE Production Engineering (January 1986)
ient Behavior of a Finite-Conductivity Fracure,”
1, No. 1, 31–41.
paper SPE 13057, presented at the SPE Annual Tech-
Reimers, D.R. and Clausen, R.A.: “High-Permeability nical Conference and Exhibition, Houston, Texas,
Fracturing at Prudhoe Bay, Alaska,” paper SPE USA (September 16–19, 1984).
Romero, J. and Féraud, J.P.: “Stability of Proppant
ference and Exhibition, Dallas, Texas, USA (October
Packs Reinforced with Fiber for Proppant Flowback
22–25, 1991).
Control,” paper SPE 31093, presented at the SPE
Rice, J.R.: “Mathematical Analysis in the Mechanics International Symposium on Formation Damage
of Fracture,” Fracture, An Advanced Treatise, New Control, Lafayette, Louisiana, USA (February 14–15,
York, New York, USA, Academic Press (1968), 2. 1996).
Rice, J.R.: “Pore Pressure Effects in Inelastic Con- Romero, J., Mack, M.G. and Elbel, J.L.: “Theoretical
stitutive Formulations for Fissured Rock Masses,” Model and Numerical Investigation of Near-Wellbore
Advances in Civil Engineering Through Engineering Effects in Hydraulic Fracturing,” paper SPE 30506,
Mechanics, New York, New York, USA, American presented at the SPE Annual Technical Conference
Society of Civil Engineers (1977), 295–297. and Exhibition, Dallas, Texas, USA (October 22–
Risnes, R., Bratli, R.K. and Horsrud, P.: “Sand 25, 1995).
Stresses Around a Wellbore,” paper SPE 9650, SPE Roodhart, L.P.: “Fracturing Fluids: Fluid-Loss Mea-
Journal (December 1982) 22, No. 6, 893–898. surements Under Dynamic Conditions,” paper SPE
11900, SPE Journal (October 1985a), 629–636.

Roodhart, L.P.: “Proppant Settling in Non-Newtonian Russell, D.G. and Truitt, N.E.: “Transient Pressure
Fracturing Fluids,” paper SPE/DOE 13905, presented Behavior in Vertically Fractured Reservoirs,” paper
at the SPE/DOE 1985 Low Permeability Gas Reser- SPE 967, Journal of Petroleum Technology (October
voirs Symposium, Denver, Colorado, USA (May 19– 1964) 16, 1159–1170.
22, 1985b).
Rutqvist, J. and Stephansson, O.: “A Cyclic
Roodhart, L.P., Kamphuis, H. and Davies, D.R.: Hydraulic Jacking Test to Determine the In Situ
“Improved Acid-Fracturing Treatment Designs Based Stress Normal to a Fracture,” International Journal
on In-Situ Temperature Calculations,” paper SPE of Rock Mechanics and Mining Sciences & Geome-
26185, presented at the SPE Gas Technology Sym- chanics Abstracts (1996) 33, No. 7, 695.
posium, Calgary, Alberta, Canada (June 28–30,
Sack, R.A.: “Extension of Griffith’s Theory of
1993).
Rupture to Three Dimensions,” Proc. Phys. Soc.
Roodhart, L., Kulper, T.O.H. and Davies, D.R.: of London (1946) 58, 729–736.
“Proppant Pack and Formation Impairment During
Salz, L.B.: “Relationship Between Fracture Propa-
Gas-Well Hydraulic Fracturing,” paper SPE 15629,
gation Pressure and Pore Pressure,” paper SPE 6870,
and Exhibition, Denver, Colorado, USA (October
(October 5–8, 1986).
9–12, 1977).
Rosen, M.J.: “Relation of Structure to Properties in
Samuel, M., Card, R.J., Nelson, E.B., Brown, J.E.,
Surfactants,” J. Amer. Oil Chem. Soc. (1972) 49, No.
Vinod, P.S., Temple, H.L., Qu, Q. and Fu, D.K.:
5, 293–297.
“Polymer-Free Fluid for Hydraulic Fracturing,” paper
Rosepiler, M.J.: “Determination of Principal Stresses SPE 38622, presented at the SPE Annual Technical
and Confinement of Hydraulic Fractures in Cotton Conference and Exhibition, San Antonio, Texas,
Valley,” paper SPE 8405, presented at the SPE USA (October 5–8, 1997).
Santarelli, F.J., Dahen, D., Baroudi, H. and Sliman,
K.B.: “Mechanisms of Borehole Instability in
Rowan, G.: “Theory of Acid Treatment of Limestone Heavily Fractured Rock Media,” International
Formations,” International Institute Petroleum (1957) Journal of Rock Mechanics and Mining Sciences
45, No. 431. & Geomechanics Abstracts (October 1992) 29, No.
5, 457–467.
Royce, T.N., Beck, L.M. and Rickards, A.R.: “Rhe-
ological Characteristics of Adjustable Crosslinked Santarelli, F.J., Ouadfel, H. and Zundel, J.P.: “Opti-
Fracturing Fluids,” paper SPE 13178, presented at mizing the Completion Procedure to Minimize Sand
the SPE Annual Technical Conference and Exhi- Production Risk,” paper SPE 22797, presented at the
bition, Houston, Texas, USA (September 16–19, SPE Annual Technical Conference and Exhibition,
1984). Dallas, Texas, USA (October 6–9, 1991).
Rudnicki, J.W.: “Effect of Pore Fluid Diffusion on Sattler, A.R., Raible, C.J. and Gall, B.R.: “Integration
Deformation and Failure of Rock,” Mechanics of of Laboratory and Field Data for Insight on the
Geomaterials, Proc., IUTAM William Prager Sym- Multiwell Experiment Paludal Stimulation,” paper
posium on Mechanics of Geomaterials: Rocks, Con- SPE 13891, presented at the SPE/DOE 1985 Low
crete, Soils, Chichester, UK, Wiley (1985), 315–347. Permeability Gas Reservoirs Symposium, Denver,
Colorado, USA (May 19–22, 1985).
Rummo, G.J.: “Crosslinking Delay Achieved with
Gel Additive,” Oil & Gas Journal (September 13, Saucier, R.J. and Lands, J.F.: “A Laboratory Study
1982) 80, No. 37, 84–89. of Perforations in Stressed Formations,” paper SPE
Ruseska, I., Robbins, J. and Costerton, J.W.: “Biocide
1978) 30, 1447–1353.
Testing Against Corrosion-Causing Oil-Field Bacteria
Helps Control Plugging,” Oil & Gas Journal (March Savage, W.Z., Swolfs, H.S. and Powers, P.S.: “Gravi-
8, 1982) 80, No. 10, 253–264. tational Stresses in Long Symmetric Ridges and
R-34 References

Valleys,” International Journal of Rock Mechanics 4834, presented at the SPE European Spring Meeting,
and Mining Sciences & Geomechanics Abstracts Amsterdam, Netherlands (May 29–30, 1974).
(October 1985) 22, No. 5, 291–302.
Schriefer, F.E. and Shaw, M.S.: “Use of Fine Salt as
Savins, J.G.: “Generalized Newtonian (Pseudoplastic) a Fluid Loss Material in Acid Fracturing Stimulation
Flow and Stationary Pipes and Annuli,” Trans., Treatments,” paper SPE 7570, presented at the SPE
AIME (1958) 213, 325. Annual Technical Conference and Exhibition,
Schatz, J.F., Olszewski, A.J. and Schraufnagel, R.A.:
“Scale Dependence of Mechanical Properties: SCR Geomechanics Group: “On the Modelling of
Application to the Oil and Gas Industry,” paper SPE Near Tip Processes in Hydraulic Fractures,” Inter-
25904, Proc., Rocky Mountain Regional/Low national Journal of Rock Mechanics and Mining
Permeability Reservoirs Symposium and Exhibition, Sciences & Geomechanics Abstracts (1993) 30, No.
Denver, Colorado, USA (1993), 615–624. 7, 1127–1134.
Schechter, R.S. and Gidley, J.L.: “The Change in “Section 2,” Recommended Practices for Evaluation
Pore Size Distribution from Surface Reactions in of Well Perforators (RP 43), fifth edition, Dallas,
Porous Media,” AIChE J. (May 1969) 15, 339–350. Texas, USA, American Petroleum Institute (January
1991).
Scherubel, G.A. and Crowe, C.W.: “Foamed Acid:
A New Concept in Fracture Acidizing,” paper SPE “Section 4,” Recommended Practices for Evaluation
7568, presented at the SPE Annual Technical of Well Perforators (RP 43), fifth edition, Dallas,
Conference and Exhibition, Houston, Texas, USA Texas, USA, American Petroleum Institute (January
(October 1–3, 1978). 1991).
Schiuma, M., Claverie, M. and Brie, A.: “Hydro- Senseny, T.W. and Pfiefle, T.W.: “Fracture Toughness
carbon Identification from Shear Sonic Logs in the of Sandstones and Shales,” Proc., 25th U.S. Sym-
Newquen Basin, Argentina,” Transactions of the posium on Rock Mechanics (1984), 390–397.
SPWLA 38th Annual Logging Symposium (1997),
Serra, O. and Andreani, M.: Thin Beds, Schlumberger
paper EEE.
(1991).
Schlottman, B.W., Miller, W.K. II and Lueders, R.K.:
Settari, A.: “A New General Model of Fluid Loss
“Massive Hydraulic Fracture Design for the East
in Hydraulic Fracturing,” paper SPE 11625, SPE
Texas Cotton Valley Sands,” paper SPE 10133, pre-
Journal (August 1985) 25, No. 4, 491–501.
Exhibition, San Antonio, Texas, USA (October 5– Settari, A.: “Modeling of Acid-Fracturing Treat-
7, 1981). ments,” paper SPE 21870, presented at the SPE
Rocky Mountain Regional/Low Permeability Res-
Schmidt, R.A.: “Fracture Toughness Testing of Lime-
ervoirs Symposium, Denver, Colorado, USA (April
stone,” Exp. Mech. (May 1976) 16, No. 5, 161–167.
15–17, 1991); also in SPE Production & Facilities
Schmidt, R.A. and Huddle, C.W.: “Effect of Con- (February 1993) 8, No. 1, 30–38.
fining Pressure on Fracture Toughness of Indiana
Settari, A. and Cleary, M.P.: “Three-Dimensional
Limestone,” International Journal of Rock Mechanics
Simulation of Hydraulic Fracturing,” paper SPE
10504, presented at the SPE Symposium on Reser-
(November 1977) 14, Nos. 5–6, 289–293.
voir Simulation, New Orleans, Louisiana, USA
Schmidt, R.A. and Lutz, T.J.: “Klc and Jlc of Westerly (January 31–February 3, 1982); also in Journal
Granite—Effects of Thickness and In-Plane Dimen- of Petroleum Technology (July 1984) 36, No. 7,
sions,” Fracture Mechanics Applied to Brittle 1177–1190.
Materials, S.W. Freidman (ed.), Philadelphia,
Shah, S.N.: “Proppant-Settling Correlations for Non-
Pennsylvania, USA, ASTM STP (1979), 166–182.
Newtonian Fluids,” paper SPE 13835, SPE Pro-
Schols, R.S. and Visser, W.: “Proppant Bank Buildup duction Engineering (November 1986), 446.
in a Vertical Fracture Without Fluid Loss,” paper SPE

Shah, S.N. and Asadi, M.: “Convection/Encapsulation Shlyapobersky, J., Walhaug, W.W., Sheffield, R.E.
in Hydraulic Fracture,” paper SPE 39961, presented and Huckabee, P.T.: “Field Determination of Frac-
at the SPE Rocky Mountain Regional/Low Permea- turing Parameters for Overpressure Calibrated Design
bility Reservoirs Symposium, Denver, Colorado, of Hydraulic Fracturing,” paper SPE 18195, pre-
USA (April 5–8, 1998). sented at the SPE Annual Technical Conference
and Exhibition, Houston, Texas, USA (October 2–
Shah, S.N., Asadi, M. and Lord, D.L.: “Proppant
5, 1988a).
Transport Characterization of Hydraulic Fracturing
Fluids Using a High Pressure Simulator Integrated Shlyapobersky, J., Wong, G.K. and Walhaug, W.W.:
with a Fiber Optic/LED Vision System,” paper SPE “Overpressure Calibrated Design of Hydraulic
49040, presented at the SPE Annual Technical Fracture Simulations,” paper SPE 18194, presented
Conference and Exhibition, New Orleans, Louisiana, at the SPE Annual Technical Conference and Exhi-
USA (September 27–30, 1998). bition, Houston, Texas, USA (October 2–5, 1988b).
Shah, S.N. and Lee, Y.N.: “Friction Pressures of Siegfried, R. and Simmons, G.: “Characterization of
Proppant-Laden Hydraulic Fracturing Fluids,” paper Oriented Cracks with Differential Strain Analysis,”
SPE 13836, SPE Production Engineering (November Journal of Geophysical Research (February 1978) 83,
1986), 437–445. No. B3, 1269–1278.
Shah, S.N. and Lord, D.L.: “Hydraulic Fracturing Sievert, J.A., Wahl, H.A., Clark, P.E. and Harkin,
Slurry Transport in Horizontal Pipes,” paper SPE M.W.: “Propped Transport in a Large Vertical
18994, SPE Drilling Engineering (September 1990) Model,” paper SPE 9865, presented at the SPE/DOE
5, No. 3, 225–232. Low Permeability Symposium, Denver, Colorado,
USA (May 27–29, 1981).
Shah, S.N. and Subramanian, R.: “Effects of Coiled
Tubing Shear History on the Rheological and Simmons, G. and Brace, W.F.: “Comparison of Static
Hydraulic Properties of Fracturing Fluids,” paper and Dynamic Measurements of Compressibility of
SPE 38421, presented at the SPE 2nd North Amer- Rocks,” Journal of Geophysical Research (November
ican Coiled Tubing Roundtable, Montgomery, Texas, 1965) 70, No. 22, 5649–5656.
USA (April 1–3, 1997).
Simonson, E.R., Abou-Sayed, A.S. and Clifton, R.J.:
Shah, S.N., et al.: New Correlations for Perforation “Containment of Massive Hydraulic Fractures,”
Pressure Loss, GRI Technical Summary (1996) paper SPE 6089, SPE Journal (February 1978)
GRI-96/0208. 18, No. 1, 27–32.
Shaly Sands, Reprint Volume, SPWLA (July 1982). Sinclair A.R.: “Heat Transfer Effects in Deep Well
Fracturing,” paper SPE 3011, Journal of Petroleum
Sharman, W.R., Richards, E.L. and Malcolm, G.N.:
Technology (November 1971) 23, No. 11, 1484–
“Hydrodynamic Properties of Aqueous Solutions of
1492.
Galactomannans,” Biopolymers (1978) 17, 2817–
2833. Singh, P.K., Agarwal, R.G. and Krase, L.D.: “System-
atic Design and Analysis of Step-Rate Tests to
Sheppard, C.S. and Kamath, V.R.: “Azo and Peroxide
Determine Formation Parting Pressure,” paper SPE
Free Radical Initiators: What They Are, How They
Differ, and How They Work,” Proc., 33rd Annual
Technical Conference of Reinforced Plastics/
(September 27–30, 1985).
Composites Institute, The Society of the Plastics
Industry, Inc. (February 7–10, 1978), section 5-D. Sinha, B.K.: “Fluid Leak-Off Under Dynamic and
Static Conditions Utilizing the Same Equipment,”
Shlyapobersky, J.: “Energy Analysis of Hydraulic
Fracturing,” Proc., 26th U.S. Symposium on Rock
Technical Conference and Exhibition, New Orleans,
Mechanics, Rapid City, South Dakota, USA (1985),
Louisiana, USA (October 3–6, 1976).
539–546.
R-36 References

Sizer, J.P., Moullem, A.S. and Abou-Sayed, I.S.: Technology Symposium, Dallas, Texas, USA (June
“Evaluation of Closed Fracture Acidizing Performed 7–9, 1989).
in a Tight Limestone Formation,” paper SPE 21440,
Smith, M.B., Rosenberg, R.J. and Bowen, J.F.:
presented at the SPE Middle East Oil Show, Bahrain
“Fracture Width: Design vs. Measurement,” paper
(November 16–19, 1991).
SPE 10965, presented at the SPE Technical Confer-
Skempton, A.W.: “Effective Stress in Soils, Concrete ence and Exhibition, New Orleans, Louisiana, USA
and Rocks,” Proc., Pore Pressure and Suction in (September 26–29, 1982).
Soils, London, UK, Butterworth (1960), 4–16.
Smith, R.C.: “Successful Primary Cementing Can Be
Slattery, J.C. and Bird, R.B.: “Non-Newtonian Flow a Reality,” paper SPE 13498, Journal of Petroleum
Past a Sphere,” Chem. Eng. Sci. (1961) 16, 231–241. Technology (November 1984) 36, No. 11, 1851–
1858.
Smith, M.B.: “Effect of Fracture Azimuth on
Production with Application to the Wattenberg Gas Smith, V.T., Colt, J.R., Johnson, R.L. II and Sinclair,
Field,” paper SPE 8298, presented at the SPE Annual A.R.: “Proppant Effects on Zirconium Crosslinked
Technical Conference and Exhibition, Las Vegas, Frac Fluids,” paper SPE 27720, presented at the SPE
Nevada, USA (September 23–26, 1979). Permian Basin Oil and Gas Recovery Conference,
Smith, M.B.: “Stimulation Design for Short, Precise
Hydraulic Fractures,” paper SPE 10313, SPE Journal Smits, R.M.M., de Waal, J.A. and van Kooten, J.F.C.:
(June 1985), 371. “Prediction of Abrupt Reservoir Compaction and
Subsurface Subsidence Caused by Pore Collapse
Smith, M.B., Bale, A., Britt, L.K., Hainey, B.E. and
in Carbonates,” paper SPE 15642, SPE Formation
Klein, H.K.: “Enhanced 2D Proppant Transport
Evaluation (June 1988) 3, No. 2, 340–346.
Simulation: The Key to Understanding Proppant
Flowback and Post-Frac Productivity,” paper SPE Sneddon, I.N.: “The Distribution of Stress in the
38610, presented at the SPE Annual Technical Con- Neighbourhood of a Crack in an Elastic Solid,” Proc.,
ference and Exhibition, San Antonio, Texas, USA Royal Soc. London (1946) 187, Ser. A., 229–260.
(October 5–8, 1997).
Sneddon, I.N. and Elliot, A.A.: “The Opening of a
Smith, M.B. and Hannah, R.R.: “High Permeability Griffith Crack Under Internal Pressure,” Quarterly
Fracturing: The Evolution of a Technology,” paper of Appl. Math. (1946) 4, 262–267.
SPE 27984-P, Journal of Petroleum Technology (July
Snider, P.M., Hall, F.R. and Whisonant, R.J.: “Exper-
1996), 628.
iences with High Energy Stimulations for Enhancing
Smith, M.B. and Klein, H.H.: “Practical Applications Near-Wellbore Conductivity,” paper SPE 35321,
of Coupling Fully Numerical 2-D Transport Calcu- presented at the SPE International Petroleum Con-
lation with a PC-Based Fracture Geometry Simu- ference and Exhibition, Villahermosa, Mexico
lator,” paper SPE 30505, presented at the SPE (March 5–7, 1996).
Soliman, M.Y.: “Design and Analysis of a Fracture
with Changing Conductivity,” J. Cdn. Pet. Tech.
Smith, M.B., Miller, W.K. and Haga, J.: “Tip Screen- (September–October 1986a) 25, No. 5, 62–67.
Out Fracturing: A Technique for Soft, Unstable
Soliman, M.Y.: “Fracture Conductivity Distribution
Formations,” paper SPE 13273, presented at the
Studied,” Oil & Gas Journal (February 10, 1986b)
84, No. 6, 89–93.
Houston, Texas, USA (September 16–19, 1984); also
in SPE Production Engineering (May 1987) 2, No. Soliman, M.Y.: “Technique for Considering Fluid
2, 95–103. Compressibility and Temperature Changes in Mini-
Frac Analysis,” paper SPE 15370, presented at the
Smith, M.B., Reeves, T.L. and Miller, W.K. II:
“Multiple Fracture Height Measurements: A Case
New Orleans, Louisiana, USA (October 5–8, 1986c).
History,” paper SPE 19092, presented at the SPE Gas

Soliman, M.Y., Hunt, J.L. and El Rabaa, A.M.: View,” paper SPE 10773, Journal of Petroleum
“Fracturing Aspects of Horizontal Wells,” paper SPE Technology (March 1985) 37, No. 3, 521–526.
18542, Journal of Petroleum Technology (August
Strickland, F.G. and Ren, N.-K.: “Predicting the In-
1990) 42, No. 8, 966–973.
Situ Stress for Deep Wells Using Differential Strain
Soliman, M.Y., Kuhlman, R.D. and Poulsen, D.K.: Curve Analysis,” paper SPE 8954, Proc., SPE/DOE
“Minifrac Analysis for Heterogeneous Reservoirs,” Symposium on Unconventional Gas Recovery,
paper SPE 21554, presented at the CIM/SPE Inter- Pittsburgh, Pennsylvania, USA (1980), 251–258.
national Technical Meeting, Calgary, Alberta, Canada
Sukup, R.A., Ott, R.E., Robson, M.K. and Bell, W.T.:
(June 10–13, 1990).
“Simple Method Tracks Charge Performance,” paper
Souhaite, P., Misk, A. and Poupon, A.: “Rt Determi- SPE 17172, Journal of Petroleum Technology
nation in the Eastern Hemisphere,” Transactions of (October 1989) 41, No. 10, 1026–1032.
the SPWLA 16th Annual Logging Symposium (June
Swanson, P.L. and Spetzler, H.: “Ultrasonic Probing
4–7, 1975), paper LL.
of the Fracture Process Zone in Rock Using Surface
Spath, J.B., Ozkan, E. and Raghavan, R.: “An Effi- Waves,” Proc., 25th U.S. Symposium on Rock
cient Algorithm of Well Responses in Commingled Mechanics, Evanston, Illinois, USA (1984), 67–76.
Reservoirs,” paper SPE 21550, SPE Formation
Talley, G.R., Swindell, T.M., Waters, G.A. and Nolte,
Evaluation (June 1994) 9, No. 2, 115–121.
K.G.: “Field Application of After-Closure Analysis
Stadulis, J.: “Development of a Completion Design of Fracture Calibration Tests,” paper SPE 52220,
to Control Screenouts Caused by Multiple Near- presented at the SPE Mid-Continent Operations
Wellbore Fractures,” paper SPE 29549, presented at Symposium, Oklahoma City, Oklahoma, USA
the SPE Rocky Mountain Regional/Low Permeability (March 28–31, 1999).
Talobre, J.: La Mécanique des Roches, Paris, France,
(March 20–22, 1995).
Dunod (1957).
Starfield, A.M., Smith, K.A. and Bleloch, A.L.: How
Talobre, J.: “Dix Ans de Mesures de Compression
to Model It—Problem Solving for the Computer Age,
Interne des Roches: Progrès et Résultats Pratiques,”
New York, New York, USA, McGraw-Hill (1990).
Geologie Bauwes. (1958) 25, 148–165.
Stewart, B.R., Mullen, M.E., Howard, W.J. and
Tan, H.C., Wesselowski, K.S. and Willingham, J.D.:
Norman, W.D.: “Use of a Solids-Free Viscous
“Delayed Borate Crosslinked Fluids Minimize Pipe
Carrying Fluid in Fracturing Applications: An
Friction Pressure,” paper SPE 24342, presented at the
Economic and Productivity Comparison in Shallow
SPE Rocky Mountain Regional Meeting, Casper,
Completions,” paper SPE 30114, presented at the
Wyoming, USA (May 18–21, 1992).
SPE European Formation Damage Conference, The
Hague, Netherlands (May 15–16, 1995). Tariq, S.M.: “Evaluation of Flow Characteristics of
Perforations Including Nonlinear Effects Using Finite-
Stewart, G., Heriot-Watt, U. and Ayestaran, L.: “The
Element Method,” paper SPE 12781, presented at the
Interpretation of Vertical Pressure Gradients Mea-
SPE California Regional Meeting, Long Beach,
sured at Observation Wells in Developed Reservoirs,”
California, USA (April 11–13, 1984).
Technical Conference and Exhibition, New Orleans, Tariq, S.M.: “New, Generalized Criteria for Deter-
Louisiana, USA (September 26–29, 1982). mining the Level of Underbalance for Obtaining
Clean Perforations,” paper SPE 20636, presented
Stiles, D.A.: “Compatibility of Curable Resin-Coated
Proppants with Breakers in Crosslinked Fracturing
bition, New Orleans, Louisiana, USA (September 23–
Fluids,” presented at the 1991 Southwestern Petro-
26, 1990).
leum Short Course, Lubbock, Texas, USA (April 17–
18, 1991). Tehrani, M.A.: “An Experimental Study of Particle
Migration in Pipe Flow of Viscoelastic Fluids,” J.
Strickland, F.G.: “Reasons for Production Decline in
Rheol. (1996) 40, 1057–1077.
the Diatomite, Belridge Oil Field: A Rock Mechanics
R-38 References

Terzaghi, K. van: “Die Berechnung der Durchlassig- Colorado, USA (April 15–17, 1991); also in SPE
keitsziffer des Tones aus dem Verlauf der Hydro- Formation Evaluation (December 1994) 4, No. 4,
dynamischen Spannungserscheinungen,” Sitzungsber. 251–258.
Akad. Wiss., Wien Math. Naturwiss. Kl. (1923) 132,
Thiercelin, M., Plumb, R.A., Desroches, J.,
105–124.
Bixenman, B., Jonas, J.K. and Davie, W.: “A New
Teufel, L.W.: “Determination of the In-Situ Stress Wireline Tool for In-Situ Stress Measurements,”
from Anelastic Strain Recovery Measurements of paper SPE 25906, Proc., Rocky Mountain Regional/
Oriented Cores,” paper SPE 11649, Proc., SPE/DOE Low Permeability Reservoirs Symposium and Exhi-
Symposium on Low-Permeability Reservoirs, Denver, bition, Denver, Colorado, USA (1993), 635–646.
Colorado, USA (1983), 421–427.
Thiercelin, M. and Roegiers, J.-C.: “Toughness
Teufel, L.W. and Farrell, H.E.: “In-Situ Stress and Determination with the Modified Ring Test,” Proc.,
Natural Fracture Distribution in the Ekofisk Field, 27th U.S. Symposium on Rock Mechanics, Tusca-
North Sea,” presented at the 3rd North Sea Chalk loosa, Alabama, USA (1986), 615–622.
Symposium, Copenhagen, Denmark (June 11–12,
Thomas, R.L., Crowe, C.W. and Simpson, B.E.:
1990).
“Effect of Chemical Treatment upon Formation Clays
Teufel, L.W. and Rhett, D.W.: “Geomechanical Evi- Is Revealed by Improved SEM Technique,” paper
dence for Shear Failure of Chalk During Production SPE 6007, presented at the SPE Annual Technical
of the Ekofisk Field,” paper SPE 22755, presented Conference and Exhibition, New Orleans, Louisiana,
at the SPE Annual Technical Conference and Exhi- USA (October 3–6, 1976).
bition, Dallas, Texas, USA (October 6–9, 1991).
Thomas, R.L. and Elbel, J.L.: “The Use of Viscosity
The Log Analyst, Special Issue on NMR Logging Stabilizers in High-Temperature Fracturing,” paper
(November–December 1996) 37, No. 6. SPE 8344, presented at the SPE Annual Technical
Conference and Exhibition, Las Vegas, Nevada, USA
The Log Analyst, Special Issue on NMR Logging
(September 23–26, 1979).
(March–April 1997) 38, No. 2.
Thompson, G.D.: Effects of Formation Compressive
Thiercelin, M.: “Fracture Toughness Under Confining
Strength on Perforator Performance,” Drilling and
Pressure Using the Modified Ring Test,” Proc., 20th
Production Practice, 1962, New York, New York,
U.S. Symposium on Rock Mechanics, Tucson,
USA, American Petroleum Institute (1963) 191–197.
Arizona, USA (June 29–July 1, 1987), 149–156.
Thompson, M. and Willis, J.R.: “A Reformulation
Thiercelin, M.: “Fracture Toughness and Hydraulic
of the Equations of Anisotropic Poro-Elasticity,”
Fracturing,” International Journal of Rock Mechanics
Journal of Applied Mechanics (September 1991) 58,
No. 3, 612–616.
(July 1989) 26, Nos. 3–4, 177–183.
Thorne, B.J. and Morris, H.E.: “Advances in Bore-
Thiercelin, M., Desroches, J. and Kurkjian, A.: “Open
hole Seismic Fracture Diagnostics,” paper SPE/DOE
Hole Stress Test in Shales,” Proc., Eurock ’94 Sym-
16405, presented at the SPE/DOE Low Permeability
posium, Delft, Netherlands (1994), 921–928.
Thiercelin, M., Jeffrey, R.G. and Ben Naceur, K.: (May 18–19, 1987).
“Influence of Fracture Toughness on the Geometry
Tiab, D. and Kumar, A.: “Application of the pD´
of Hydraulic Fractures,” paper SPE 16431, SPE
Function of Interference Analysis,” paper SPE 6053,
Production Engineering (November 1989) 4, No.
Journal of Petroleum Technology (August 1980) 32,
4, 435–442.
1465–1470.
Thiercelin, M. and Plumb, R.A.: “A Core-Based
Tiab, D. and Puthigai, S.K.: “Pressure-Derivative
Prediction of Lithologic Stress Contrasts in East
Type Curves for Vertically Fractured Wells,” paper
Texas Formations,” paper SPE 21847, presented at
SPE 11028, SPE Formation Evaluation (March
the SPE Rocky Mountain Regional Meeting/Low
1988) 3, No. 1, 156–158.
Permeability Reservoirs Symposium, Denver,

Timoshenko, S.P. and Goodier, J.N.: Theory of Unwin, A.T. and Hammond, P.S.: “Computer Simula-
Elasticity, third edition, New York, New York, USA, tions of Proppant Transport in a Hydraulic Fracture,”
McGraw-Hill Book Company (1970). paper SPE 29649, presented at the SPE Western
Regional Meeting, Bakersfield, California, USA
Timur, A.: “An Investigation of Permeability, Por-
(March 8–10, 1995).
osity and Water Saturation Relationships,” Trans-
actions of the SPWLA 9th Annual Logging Sym- Valkó, P. and Economides, M.J.: “A Continuum-
posium, New Orleans, Louisiana, USA (June 23–26, Damage-Mechanics Model of Hydraulic Fracturing,”
1968), paper J. paper SPE 25304, Journal of Petroleum Technology
(March 1993a) 45, No. 3, 198–205.
Tinker, S.J.: “Equilibrium Acid Fracturing: A New
Fracture Acidizing Technique for Carbonate For- Valkó, P. and Economides, M.J.: “Fracture Height
mations,” paper SPE 18883, SPE Production Engi- Containment with Continuum Damage Mechanics,”
neering (February 1991) 6, No. 1, 25–32. paper SPE 26598, presented at the SPE Annual
Tinsley, J.M., Williams, J.R., Tiner, R.L. and Malone,
Texas, USA (October 3–6, 1993b).
W.T.: Vertical Fracture Height—Its Effect on Steady
State Production Increase,” paper SPE 1900, Journal Valkó, P. and Economides, M.J.: “Propagation of
of Petroleum Technology (May 1969) 21, 633–638. Hydraulically Induced Fractures—A Continuum
Damage Mechanics Approach,” International
Tittman, J. and Wahl, J.S.: “The Physical Foundations
Journal of Rock Mechanics and Mining Sciences
of Formation Density Logging (Gamma/Gamma),”
& Geomechanics Abstracts (June 1994) 31, No. 3,
Geophysics (1965) 30, No. 2.
221–229.
Torrest, R.S.: “Aspects of Slurry and Particle Settling
Valkó, P. and Economides, M.J.: Hydraulic Fracture
and Placement for Viscous Gravel Packing (AQUA-
Mechanics, Chichester, UK, Wiley (1995).
PAC),” paper SPE 11009, presented at the SPE
Annual Technical Conference and Exhibition, New Valkó, P. and Economides, M.J.: “Performance of
Orleans, Louisiana, USA (September 26–29, 1982). a Longitudinally Fractured Horizontal Well,” paper
SPE 31050-P, SPE Journal (March 1996a) 1, No. 1,
Trouiller, J.-C., Delhomme, J.-P., Carlin, S. and
11–19.
Anxionnaz, H.: “Thin-Bed Reservoir Analysis from
Borehole Electrical Images,” paper SPE 19578, pre- Valkó, P. and Economides, M.J.: “Performance of
sented at the SPE Annual Technical Conference and Fractured Horizontal Wells in High-Permeability
Exhibition, San Antonio, Texas, USA (October 8– Reservoirs,” paper SPE 31149, presented at the SPE
11, 1989). International Symposium on Formation Damage
Control, Lafayette, Louisiana, USA (February 14–
Truby, L.S., Keck, R.G. and Withers, R.J.: “Data
15, 1996b).
Gathering for a Comprehensive Hydraulic Fracture
Diagnostic Project: A Case Study,” paper SPE 27506, Valkó, P. and Economides, M.J.: “Fluid Leakoff
presented at the IADC/SPE Drilling Conference, Delineation in High-Permeability Fracturing,” paper
Dallas, Texas, USA (February 15–18, 1994). SPE 37403, presented at the SPE Production Oper-
ations Symposium, Oklahoma City, Oklahoma, USA
Turian, R.M.: “An Experimental Investigation of the
(March 9–11, 1997).
Flow of Aqueous Non-Newtonian High Polymer
Solutions Past a Sphere,” AiChE J. (1967) 13, No. van Dam, D.B., de Pater, C.J. and Romijn, R.:
5, 999–1006. “Analysis of Fracture Closure in Laboratory Exper-
iments,” paper SPE 47380, presented at the SPE/
Underdown, D.R., Das, K. and Nguyen, H.: “Gravel
ISRM Eurock ’98, Trondheim, Norway (July 8–
Packing Highly Deviated Wells with a Crosslinked
10, 1998).
Polymer System,” paper SPE 12481, presented at the
SPE Formation Damage Symposium, Bakersfield, Vandamme, L., Jeffrey, R.G. and Curran, J.H.: “Pres-
California, USA (February 13–14, 1984). sure Distribution in Three-Dimensional Hydraulic
Fractures,” paper SPE 15265, SPE Production
Engineering (May 1988), 181.
R-40 References

van de Ketterij, R.G.: “Raw Data of Hydraulic Conference and Exhibition, New Orleans, Louisiana,
Fracturing Model Tests—Series 5,” test report, Delft USA (October 5–8, 1986).
University of Technology, Delft, Netherlands (April
Veatch, R.W. Jr. and Crowell, R.F.: “Joint Research
1996).
Operations Programs Accelerate Massive Hydraulic
van der Vlis, A.C., Haafkens, R., Schipper, B.A. and Fracturing Technology,” paper SPE 9337, Journal
Visser, W.: “Criteria for Proppant Placement and of Petroleum Technology (December 1982) 34, No.
Fracture Conductivity,” paper SPE 5637, presented 2, 2763–2775.
Veeken, C.A.M., Davies, D.R., Kenter, C.J. and
bition, Dallas, Texas, USA (September 28–October
Kooijman, A.P.: “Sand Production Prediction
1, 1975).
Review: Developing an Integrated Approach,” paper
Van Domelen, M.S.: “Optimizing Fracture Acidizing SPE 22792, presented at the SPE Annual Technical
Treatment Design by Integrating Core Testing, Field Conference and Exhibition, Dallas, Texas, USA
Testing, and Computer Simulation,” paper SPE (October 6–9, 1991).
22393, presented at the SPE International Meeting
Veley, C.D.: “How Hydrolyzable Metal Ions React
on Petroleum Engineering, Beijing, China (March
with Clays to Control Formation Water Sensitivity,”
24–27, 1992).
Van Everdingen, A.F. and Hurst, W.: “The Application (September 1969) 21, No. 9, 1111–1118.
of the Laplace Transformation to Flow Problems in
Venghiattis, A.A.: “Prediction of the Efficiency of
Reservoirs,” Trans., AIME (1949) 186, 305–324.
a Perforator Down-Hole Based on Acoustic Logging
van Heerden, W.L.: “General Relations Between Information,” paper SPE 422, Journal of Petroleum
Static and Dynamic Moduli of Rocks,” International Technology (July 1963) 15, No. 6, 761–768.
Journal of Rock Mechanics and Mining Sciences
Vincent, M.C. and Pearson, C.M.: “The Relationship
& Geomechanics Abstracts (1987) 24, No. 6, 381–
Between Fractured Well Performance and Hole
385.
Deviation,” paper SPE 29569, presented at the SPE
van Kruysdijk, C.P.J.W.: “Semianalytical Modeling Rocky Mountain Regional Low-Permeability Res-
of Pressure Transients in Fractured Reservoirs,” paper ervoirs Symposium, Denver, Colorado, USA (March
SPE 18169, presented at the SPE Annual Technical 20–22, 1995).
Conference and Exhibition, Houston, Texas, USA
Vinegar, H.J., Wills, P.B., DeMartini, D.C., Shlya-
(October 2–5, 1988).
pobersky, J., Deeg, W.F.J., Adair, R.G., Woerpel, J.C.,
van Kruysdijk, C.P.J.W. and Dullaert, G.M.: “A Fix, J.E. and Sorrells, G.G.: “Active and Passive
Boundary Element Solution to the Transient Pressure Seismic Imaging of a Hydraulic Fracture in Diato-
Response of Multiple Fractured Horizontal Wells,” mite,” paper SPE 22756, Journal of Petroleum
presented at the 2nd European Conference on the Technology (January 1992) 44, 28–34, 88–90.
Mathematics of Oil Recovery, Cambridge, UK (July
Vinod, P.L., Flindt, M.L., Card, R.J. and Mitchell,
25–27, 1989).
J.P.: “Dynamic Fluid-Loss Studies in Low-
van Poollen, H.K., Tinsley, J.M. and Saunders, C.D.: Permeability Formations with Natural Fractures,”
“Hydraulic Fracturing: Fracture Flow Capacity vs. paper SPE 37486, presented at the SPE Production
Well Productivity,” Trans., AIME (1958) 213, 91–95. Operations Symposium, Oklahoma City, Oklahoma,
USA (March 9–11, 1997).
Veatch, R.W. Jr.: “Overview of Current Hydraulic
Fracturing Design and Treatment Technology—Part Vliet, M.R.A. v.: “Size Effect of Concrete and Rock
1,” paper SPE 10039, Journal of Petroleum Tech- Under Uniaxial Tension,” Delft University of Tech-
nology (April 1983) 35, No. 4, 677–687. ology, Delft, Netherlands (1999).
Veatch, R.W. Jr.: “Economics of Fracturing: Some Vogel, J.V.: “Inflow Performance Relationships for
Methods, Examples, and Case Studies,” paper SPE Solution-Gas Drive Wells,” paper SPE 1476, Journal
15509, presented at the SPE Annual Technical of Petroleum Technology (January 1968) 20, 83–92.

Voight, B.: “Determination of the Virgin State of Warpinski, N.R.: “Interpretation of Hydraulic
Stress in the Vicinity of a Borehole from Measure- Fracture Mapping Experiments,” paper SPE 27985,
ments of a Partial Anelastic Strain Tensor in Drill presented at the University of Tulsa Centennial
Cores,” Felsmech. Ingenieur Geol. (1968) 6, No. 4, Petroleum Engineering Symposium, Tulsa,
201–215. Oklahoma, USA (August 29–31, 1994).
Wagoner, J.C. van, Mitchum, R.M., Campion, K.M. Warpinski, N.R. and Branagan, P.T.: “Altered Stress
and Rahmanian, V.D.: Siliclastic Sequence Stratig- Fracturing,” paper SPE 17533, Proc., SPE Rocky
raphy in Well Logs, Core and Outcrops, Tulsa, Mountain Regional Meeting, Casper, Wyoming, USA
Oklahoma, USA, American Association of Petroleum (1988), 465–476.
Geologists (1990).
Warpinski, N.R., Branagan, P.T., Peterson, R.E. and
Walker, M.L., Shuchart, C.E., Yaritz, J.G. and Wolhart, S.L.: “An Interpretation of M-Site Hydraulic
Norman, L.R.: “Effects of Oxygen on Fracturing Fracture Diagnostic Results,” paper SPE 39950, pre-
Fluids,” paper SPE 28978, presented at the SPE sented at the SPE Rocky Mountain Regional/Low
International Symposium on Oilfield Chemistry, San Permeability Reservoirs Symposium, Denver, Colo-
Antonio, Texas, USA (February 14–17, 1995). rado, USA (April 5–8, 1998a).
Walles, W.E., Williamson, T.D. and Tomkinson, D.L.: Warpinski, N.R., Branagan, P. and Wilmer, R.: “In-
“Method for Testing Subterranean Formations,” U.S. Situ Stress Measurements at the DOE’s Multiwell
Patent No. 4,741,401 (May 3, 1988). Experiment Site, Mesaverde Group, Rifle, Colorado,”
Walsh, J.B.: “Effect of Pore Pressure and Confining
(March 1985) 37, 527–536.
Pressure on Fracture Permeability,” International
Journal of Rock Mechanics and Mining Sciences Warpinski, N.R., Engler, B.P., Young, C.J., Peterson,
& Geomechanics Abstracts (1981) 18, 429–435. R.E., Branagan, P.T. and Fix, J.E.: “Microseismic
Mapping of Hydraulic Fractures Using Multi-Level
Ward, V.L.: “Nitrogen and Carbon Dioxide in the Oil
Wireline Receivers,” paper SPE 30507, presented at
Field: Stimulation and Completion Applications,”
paper SPE 12594, presented at the SPE Permian Basin
Oil & Gas Recovery Conference, Midland, Texas,
USA (March 8–9, 1984). Warpinski, N.R., Lorenz, J.C., Branagan, P.T., Myal,
F.R. and Gall, B.L.: “Examination of a Cored
Warpinski, N.R.: “Investigation on the Accuracy and
Hydraulic Fracture in a Deep Gas Well,” paper SPE
Reliability of In Situ Stress Measurements Using
22876, SPE Formation Evaluation (September 1993)
Hydraulic Fracturing in Perforated Cased Holes,”
8, No. 3, 150–158.
Rock Mechanics: Theory-Experiment-Practice,
Proceedings of the 24th U.S. Symposium on Rock Warpinski, N.R., Peterson, R.E., Branagan, P.T.,
Mechanics (June 1983), 773–776. Engler, B.P. and Wolhart, S.L.: “In Situ Stress and
Moduli: Comparison of Values Derived from Mul-
Warpinski, N.R.: “Measurement of Width and Pres-
tiple Techniques,” paper SPE 49190, presented at
sure in a Propagating Hydraulic Fracture,” paper SPE
11648, SPE Journal (February 1985) 25, No. 1, 46–
bition, New Orleans, Louisiana, USA (September 27–
54.
30, 1998b).
Warpinski, N.R.: “Dual Leakoff Behavior in
Warpinski, N.R. and Teufel, L.W.: “In-Situ Stresses
Hydraulic Fracturing of Tight, Lenticular Gas Sands,”
in Low-Permeability, Nonmarine Rocks,” paper SPE
paper SPE 18259, SPE Production Engineering
16402, Journal of Petroleum Technology (April 1989)
(August 1990), 243.
41, No. 4, 405–414.
Warpinski, N.R.: “Hydraulic Fracturing in Tight,
Warpinski, N.R., Wright, T.B., Uhl, J.E., Engler, B.P.,
Fissured Media,” paper SPE 21054, Journal of
Drozda, P.M., Peterson, R.E. and Branagan, P.T.:
Petroleum Technology (February 1991) 43, No.
“Microseismic Monitoring of the B-Sand Hydraulic
2, 146–152, 208–209.
Fracture Experiment at the DOE/GRI Multi-Site Proj-
R-42 References

ect,” paper SPE 36450, presented at the SPE Annual Field,” paper SPE 16403, presented at the SPE/DOE
Technical Conference and Exhibition, Denver, Colo- Low Permeability Reservoirs Symposium, Denver,
rado, USA (October 6–9, 1996). Colorado, USA (May 18–19, 1987).
Watkins, E.K., Wendorff, C.L. and Ainley, B.R.: “A Whitsitt, N.F. and Dysart, G.R.: “The Effect of Tem-
New Crosslinked Foamed Fracturing Fluid,” paper perature on Stimulation Design,” paper SPE 2497,
SPE 12027, presented at the SPE Annual Technical presented at the SPE Annual Meeting, Denver,
Conference and Exhibition, San Francisco, Cali- Colorado, USA (September 28–October 1, 1969);
fornia, USA (October 5–8, 1983). also in Journal of Petroleum Technology (April
1970), 493–502; also in Trans., AIME (1970) 249.
Wattenbarger, R.A. and Ramey, H.J. Jr.: “Gas Well
Testing with Turbulence, Damage, and Wellbore Wilkinson, D.J. and Hammond, P.A.: “A Pertubation
Storage,” paper SPE 1835, Journal of Petroleum Theorem for Mixed Boundary Value Problems in
Technology (August 1968) 20, No. 8, 877–887; also Pressure Transient Testing,” Transport in Porous
in Trans., AIME (1968) 243. Media (1992) 5, 609–636.
Wattenbarger, R.A. and Ramey, H.J. Jr.: “Well Test Willberg, D.M., Steinsberger, N., Hoover, R., Card
Interpretation in Vertically Fractured Gas Wells,” R.J. and Queen, J.: “Optimization of Fracture Clean-
paper SPE 2155, Journal of Petroleum Technology up Using Flowback Analysis,” paper SPE 39920,
(May 1969) 21, No. 5, 625–632; also in Trans., presented at the SPE Rocky Mountain Regional/Low
AIME (1969) 246. Permeability Reservoirs Symposium and Exhibition,
Denver, Colorado, USA (April 5–8, 1998).
Webster K.R., Goins, W.C. and Berry S.C. “A Con-
tinuous Multistage Fracturing Technique,” paper SPE Williams, B.B.: “Fluid Loss from Hydraulically
977, Journal of Petroleum Technology (June 1965) Induced Fractures,” paper SPE 2769, Journal of
17, No. 6, 619–625. Petroleum Technology (July 1970) 22, No. 6, 882–
888.
Weijers, L.: The Near-Wellbore Geometry of Hydrau-
lic Fractures Initiated from Horizontal and Deviated Williams, B.B., Gidley J.L. and Schechter, R.S.:
Wells, Delft, Netherlands, Delft University Press Acidizing Fundamentals, Richardson, Texas, USA,
(1995). Society of Petroleum Engineers (1979).
Wendorff, C.L. and Ainley, B.R.: “Massive Hydraulic Williamson, C.D. and Allenson, S.J: “A New Non-
Fracturing of High-Temperature Wells with Stable damaging Particulate Fluid-Loss Additive,” paper
Frac Foams,” paper SPE 10257, presented at the SPE SPE 18474, presented at the SPE International Sym-
Annual Technical Conference and Exhibition, San posium on Oilfield Chemistry, Houston, Texas, USA
Antonio, Texas, USA (October 5–7, 1981). (February 8–10, 1989).
Weng, X.: “Incorporation of 2D Fluid Flow into a Willson, S.M.: “Maximizing Completions Efficiency
Pseudo-3D Hydraulic Fracturing Simulator,” paper Through High Overbalanced Perforating,” TerraTek
SPE 21849, presented at the SPE Rocky Mountain (February 1995) CEA 61 Phase I.
Regional Meeting and Low-Permeability Reservoirs
Winkler, K.W. and Murphy, W.F. III: “Acoustic
Velocity and Attenuation in Porous Rocks,” Rock
1991).
Physics and Phase Relations. A Handbook of Phys-
Whistler, R.L. (ed.): Industrial Gums, New York, ical Constants, AGU Reference Shelf, Washington,
New York, USA, Academic Press (1959). D.C., USA, American Geophysical Union (1995) 3,
20–34.
White, J.L. and Daniel, E.F.: “Key Factors in MHF
Design,” paper SPE 9065, Journal of Petroleum Winkler, W., Valkó, P. and Economides, M.J:
Technology (August 1981) 33, No. 8, 1501–1512. “Laminar and Drag-Reduced Polymeric Foam Flow,”
J. Rheol. (January–February 1994) 38, No. 1, 111–
Whitehead, W.S., Hunt, E.R. and Holditch, S.A.:
127.
“The Effects of Lithology and Reservoir Pressure on
the In-Situ Stresses in the Waskom (Travis Peak)

Wiprut, D.J. and Zoback, M.D.: “High Horizontal Yale, D.P., Nieto, J.A. and Austin, S.P.: “The Effect
Stress in the Visund Field, Norwegian North Sea: of Cementation on the Static and Dynamic Mech-
Consequences for Borehole Stability and Sand anical Properties of the Rotliegendes Sandstone,”
Production,” paper SPE 47244, presented at the Proc., 35th U.S. Symposium on Rock Mechanics,
SPE/ISRM Eurock ’98, Trondheim, Norway (July Reno, Nevada, USA (1995), 169–175.
8–10, 1998).
Yeager, R.R. and Bailey, D.E.: “Diesel-Based Gel
Withjack, E.M.: “Computed Tomography for Rock- Concentrate Improves Rocky Mountain Region
Property Determination and Fluid-Flow Visualization,” Fracture Treatments,” paper SPE 17535, presented
paper SPE 16951, presented at the SPE Annual at the SPE Rocky Mountain Regional Meeting,
Technical Conference and Exhibition, Dallas, Texas, Casper, Wyoming, USA (May 11–13, 1988).
Yew, C.H. and Li, Y.: “Fracturing of a Deviated Well,”
Wong, D.W., Harrington, A.G. and Cinco-Ley, H.: paper SPE 16930, SPE Production Engineering
“Application of the Pressure Derivative Function in (November 1988) 3, No. 4, 429–437.
the Pressure Transient Testing of Fractured Wells,”
Yew, C.H. and Liu, G.F.: “Fracture Tip and Critical
paper SPE 13056, presented at the SPE Annual Tech-
Stress Intensity Factor of a Hydraulically Induced
nical Conference and Exhibition, Houston, Texas,
Fracture,” paper SPE 22875, SPE Production &
Facilities (August 1993) 8, No. 3, 171–177.
Woo, G.T. and Cramer, P.D.: “Laboratory and Field
Yew, C.H., Mear, M.E., Chang, C.C. and Zhang,
Evaluation of Fluid-Loss Additive Systems Used in
X.C.: “On Perforating and Fracturing of Deviated
the Williston Basin,” paper SPE 12899, presented at
Cased Wellbores,” paper SPE 26514, presented at
the SPE Rocky Mountain Regional Meeting, Casper,
Wyoming, USA (May 21–23, 1984).
bition, Houston, Texas, USA (October 3–6, 1993).
Wood, M.D., Smith, M.B., Abbot, R.L., Cox, D. and
Yew, C.H., Schmidt, J.H. and Li, Y.: “On Fracture
O’Shea, P.: “Fracture Proppant Mapping by Use of
Design of Deviated Wells,” paper SPE 19722, pre-
Surface Superconducting Magnetometers,” paper
SPE 11612 (1983).
Exhibition, San Antonio, Texas, USA (October 8–
Worlow, D.W.: “Development and Testing of a Stan- 11, 1989).
dard Procedure for Determining the Viscous Prop-
Yew, C.H. and Zhang, X.: “A Study of the Damage
erties of Crosslinked Fracturing Fluids,” MS thesis,
Zone Created by Shaped-Charge Perforating,” paper
Texas A&M University, College Station, Texas, USA
SPE 25902, presented at the Rocky Mountain
(1987).
Regional/Low Permeability Reservoirs Symposium,
Wright, C.A., Weijers, L., Minner, W.A. and Snow, Denver, Colorado, USA (April 12–14, 1993).
D.M.: “Robust Technique for Real-Time Closure
Zasadzinski, J.A.N., Kramer, J., Chu, A. and
Stress Estimation,” paper SPE 30503, presented at
Prud’homme, R.A.: “Transmission Electron Micros-
copy of Aqueous Polymer Gel Network Morph-
ology,” Chem. Eng. Comm. (1987) 52, 283–289.
Wroth, C.: “In-Situ Measurement of Initial Stresses
Zemanek, J., Caldwell, R.L., Glenn, E.F., Holcomb,
and Deformation Characteristics,” Proc., ACE Spec-
S.V, Norton, L.J. and Straus, A.J.D: “The Borehole
ialty Conference in In-Situ Measurements of Soil
Televiewer—A New Logging Concept for Fracture
Properties, Raleigh, North Carolina, USA (1975),
Location and Other Types of Borehole Inspection,”
181–230.
Wylie, D.J. and Rose, W.D.: “Some Theoretical Con- (June 1969) 21, 762–774; also in Trans., AIME
siderations Related to the Quantitative Evaluation of (1969) 246.
Physical Characteristics of Reservoir Rocks from
Zemlak, W., Lemp, S. and McCollum, R.: “Selective
Electrical Log Data,” Trans., AIME (1950) 189, 105–
Hydraulic Fracturing of Multiple Perforated Intervals
108.
with a Coiled Tubing Conduit: A Case History of the
R-44 References

Unique Process, Economic Impact and Related Adair, P. and Smith, P.S.: “The Miller Field:
Production Improvements,” paper SPE 54474, pre- Appraisal and Development Formation Damage
sented at the SPE/ICOTA Coiled Tubing Roundtable, Experiences,” paper SPE 27406, presented at the
Houston, Texas, USA (May 25–26, 1999). SPE International Symposium on Formation Damage
Control, Lafayette, Louisiana, USA (February 7–10,
Zhao, X.L., Roegiers, J.-C. and Guo, M.: “The
1994).
Determination of Fracture Toughness of Rocks by
Chevron-Notched Brazilian Disk Specimens,” 4th Addison, G.E.: “Identification and Treating of
Annual SCA Technical Conference, Dallas, Texas, Downhole Organic Deposits,” paper SPE 18894,
USA (August 15–16, 1990), paper 9014. presented at the SPE Production Operations
Symposium, Oklahoma City, Oklahoma, USA
Zhao, Z.-Y., Cui, M., Conway, M.W. and Chang, F.F.:
(March 13–14, 1989).
“The Rheology of Chinese XD and ZW Borate
Crosslinked Fracturing Fluids and Their Proppant Ali, S.A., Durham, D.K. and Elphingstone, E.A.:
Transport Capability,” paper SPE 29991, presented “Testing Identifies Acidizing Fluid/Crude
at the SPE International Meeting on Petroleum Engi- Compatibility Problems,” Oil & Gas Journal
neering, Beijing, People’s Republic of China (No- (March 1994), 47–51.
vember 14–17, 1995).
Ali, S.A., Hill, D.G., McConnell, S.B. and Johnson,
Zimmerman, R.W., Somerton, W.H. and King, M.S.: M.R.: “Process, Optimized Acidizing Reduce
“Compressibility of Porous Rock,” Journal of Geo- Production Facility Upsets,” Oil & Gas Journal
physical Research (November 1986) 91, No. B12, (February 1997), 44–48.
12765–12777.
Allen, T.O.: “Creative Task Force Attack on Profit
Zimmerman, T., MacInnis, J., Hoppe, J., Pop, J. and Loss Due to Formation Damage,” paper SPE 4658,
Long, T.: “Application of Emerging Wireline For- presented at the SPE Annual Meeting, Las Vegas,
mation Testing Technologies,” paper OSEA 90105, Nevada, USA (September 30–October 3, 1973).
presented at the Offshore South East Asia Confer-
Almond, S.W., Brady, J.L. and Underdown, D.R.:
ence, Singapore (December 4–7, 1990).
“Prudhoe Bay Field Study—Return Fluid Analysis,”
Zoback, M.D., Moos, D. and Mastin, L.: “Well Bore paper SPE 18223, Journal of Petroleum Technology
Breakouts and In Situ Stress,” Journal of Geophysical (April 1990) 42, No. 4, 466.
Research (1985) 90, No. B7, 5523–5530.
Amaefule, J.O., et al.: “Advances in Formation
Damage Assessment and Control Strategies,” paper
CIM 88-39-65 (1988).
Chapters 13–20 Amaefule, J.O., Padilla, P.C., McCaffery, F.G. and
Abdel-Mota’al, A.A.: “Detection and Remedy Teal, S.L.: “Steam Condensate: Formation Damage
of Behind-Casing Communication During Well and Chemical Treatments for Injectivity Improve-
Completion,” paper SPE 11498, presented at the ment,” paper SPE 12499, presented at the SPE
SPE Middle East Oil Technical Conference, Formation Damage Control Symposium, Bakersfield,
Manama, Bahrain (March 14–17, 1983). California, USA (February 13–14, 1984).
Abrams, A.: “Mud Design to Minimize Rock Antheunis, D., Luque, F.R., van der Vlis, A.C. and
Impairment Due to Particle Invasion,” paper SPE Vriezen, P.B.: “The Onset of Sand Influx from Gas-
5713, Journal of Petroleum Technology (March Producing Friable Sandstone Formations—Labor-
1977), 586–592. atory Investigations,” paper SPE 8031 (1979).
Abrams, A., Scheuerman, R.F., Templeton, C.C. and Antheunis, D., Vriezen, P.B., Schipper, B.A. and
Richardson, E.A.: “Higher-pH Acid Stimulation van der Vlis, A.C.: “Perforation Collapse: Failure
Systems,” paper SPE 7892, Journal of Petroleum of Perforated Friable Sandstones,” paper SPE 5750,
Technology (December 1983), 2175–2184. presented at the SPE European Spring Meeting,
Amsterdam, Netherlands (April 8–9, 1976).

Ayorinde, A., Granger, C. and Thomas, R.L.: “The Bartko, K.M., Montgomery, C.T., Boney, C.L. and
Application of Fluoboric Acid in Sandstone Matrix Ward, V.L.: “Development of a Stimulation Treat-
Acidizing: A Case Study,” paper IPA92-2.3.23, ment Integrated Model,” paper SPE 35991, presented
presented at the 21st Annual Indonesian Petroleum at the SPE Petroleum Computer Conference, Dallas,
Association Convention, Jakarta, Indonesia (October Texas, USA (June 2–5, 1996).
1992).
Bazin, B., Roque, C. and Bouteca, M.: “A Labor-
Azari, M. and Leimkuhler, J.: “Completion Fluid atory Evaluation of Acid Propagation in Relation to
Invasion Simulation and Permeability Restoration Acid Fracturing: Results and Interpretation,” paper
by Sodium- and Potassium-Based Brines,” paper SPE 30085, presented at the SPE European
SPE 19431, presented at the SPE Formation Damage Formation Damage Conference, The Hague,
Control Symposium, Lafayette, Louisiana, USA Netherlands (May 15–16, 1995).
(February 22–23, 1990a).
Beadie, G.: “Well Productivity Awareness School
Azari, M. and Leimkuhler, J.: “Formation (WPAS),” paper SPE 30131, presented at the SPE
Permeability Damage Induced by Completion European Formation Damage Conference, The
Brines,” paper SPE 17149, Journal of Petroleum Hague, Netherlands (May 15–16, 1995).
Technology (April 1990b), 486.
Behenna, F.R.: “Interpretation of Matrix Acidizing
Bale, G.E.: “Matrix Acidizing in Saudi Arabia Using Treatments Using a Continuously Monitored Skin
Buoyant Ball Sealers,” paper SPE 11500, presented Factor,” paper SPE 27401, presented at the SPE
at the SPE Middle East Oil Technical Conference, International Symposium on Formation Damage
Manama, Bahrain (March 14–17, 1983). Control, Lafayette, Louisiana, USA (February 7–10,
1994).
Ballard, T.J. and Dawe, R.A.: “Wettability Alteration
Induced by Oil-Based Drilling Fluid,” paper SPE Behenna, F.R.: “Acid Diversion from an Undamaged
17160, presented at the SPE Formation Damage to a Damaged Core Using Multiple Foam Slugs,”
Control Symposium, Bakersfield, California, USA paper SPE 30121, presented at the SPE European
(February 8–9, 1988). Formation Damage Conference, The Hague,
Netherlands (May 15–16, 1995).
Bandbach, P.L.: “The How and Why of Emulsions,”
Oil & Gas Journal (September 7, 1970). Behrmann, L.A.: “Underbalance Criteria for Min-
imum Perforation Damage,” paper SPE 30081, pre-
Bansal, K.M.: “Effect of Nonproduced Fluids on
sented at the SPE European Formation Damage
Produced Water Treatment Equipment Efficiency,”
Conference, The Hague, Netherlands (May 15–16,
paper SPE 25199, presented at the SPE International
1995).
Symposium on Oilfield Chemistry, New Orleans,
Louisiana, USA (March 2–5, 1993). Ben Marek, F.: “Permeability Loss in Depletion of
Reservoirs,” paper SPE 8433, presented at the SPE
Barna, B.A. and Patton, J.T.: “Permeability Damage
from Drilling Fluid Additives,” paper SPE 3830, pre-
sented at the SPE Rocky Mountain Regional Meeting,
Denver, Colorado, USA (April 10–12, 1972). Bergman, I.: “Silica Powders of Respirable Sizes IV.
The Long-Term Dissolution of Silica Powders in
Barron, A.N., Hendrickson, A.R. and Wieland, D.R.:
Dilute Hydrofluoric Acid: An Anisotropic Mech-
“The Effect of Flow on Acid Reactivity in a Carbon-
anism of Dissolution for the Coarser Quartz
ate Fracture,” Journal of Petroleum Technology
Powders,” J. Appl. Chem. (August 1963) 3, 356–361.
(April 1962), 409–415.
Bernadiner, M.G., Thompson, K.E. and Fogler, H.S.:
Bartko, K.M., Acock, A.M., Robert, J.A. and
“The Effect of Foams Used During Carbonate
Thomas, R.L.: “A Field Validated Matrix Acidizing
Acidizing,” paper SPE 21035, SPE Production
Simulator for Production Enhancement in Sandstone
Engineering (November 1992), 350–356.
and Carbonates,” paper SPE 38170, presented at the
SPE European Formation Damage Conference, The Bernard, G.G., Holm, L.W. and Jacobs, W.L.:
Hague, Netherlands (June 2–3, 1997). “Effect of Foam on Trapped Gas Saturation and
R-46 References

on Permeability of Porous Media to Water,” SPE Boyer, R.C. and Wu, C.-H.: “The Role of Reservoir
Journal (December 1965), 295. Lithology in Design of an Acidization Program:
Kuparuk River Formation, North Slope, Alaska,”
Bertaux, J.: “Treatment Fluid Selection for Sand-
paper SPE 11722, presented at the SPE California
stone Acidizing: Permeability Impairment in
Regional Meeting, Ventura, California, USA (March
Potassic Mineral Sandstones,” paper SPE 15884,
23–25, 1983).
SPE Production Engineering (February 1989), 41.
Brannon, D.H., Netters, C.K. and Grimmer, P.J.:
Bikerman, J.J.: “Foams and Emulsions—Formation
“Matrix Acidizing Design and Quality-Control
Properties and Breakdowns,” presented at the
Techniques Prove Successful in Main Pass Area
Chemistry and Physics of Interfaces—Symposium
Sandstone,” paper SPE 14827, Journal of Petroleum
on Interfaces (June 15–16, 1964).
Technology (August 1987), 931–942.
Bird, K. and Dunmore, S.: “Optimising Perforation
Brannon, H.D. and Pulsinelli, R.J.: “Evaluation of
Performance for the Armada Gas Condensate
the Breaker Concentrations Required to Improve the
Development,” paper SPE 30083, presented at the
Permeability of Proppant Packs Damaged by
SPE European Formation Damage Conference, The
Hydraulic Fracturing Fluids,” paper SPE 19402,
Hague, Netherlands (May 15–16, 1995).
presented at the SPE Formation Damage Control
Black, A.D., Dearing, H.L. and DiBona, B.G.: Symposium, Lafayette, Louisiana, USA (February
“Effects of Pore Pressure and Mud Filtration on 22–23, 1990).
Drilling Rates in a Permeable Sandstone,” paper SPE
Brown, R.W., Neill, G.H. and Loper, R.G.: “Factors
Influencing Optimum Ball Sealer Performance,”
1985), 1671.
Journal of Petroleum Technology (April 1963),
Black, S.J. and Rike, J.L.: “The Role of the Consult- 450–454.
ant in Meeting the Formation Damage Challenge,”
Browne, S.V., Ryan, D.F., Chambers, B.D., Oilchrist,
paper SPE 5700, presented at the SPE Symposium
J.M. and Bamforth, S.A: “Simple Approach to the
on Formation Damage Control, Houston, Texas,
Clean-Up of Horizontal Wells with Pre-Packed
USA (January 29–30, 1976).
Screen Completions,” paper SPE 30116, presented
Blacker, L.K.: “An Analysis of Rate-Sensitive Skin at the SPE European Formation Damage Conference,
in Oil Wells,” paper SPE 11187, presented at the The Hague, Netherlands (May 15–16, 1995).
Brownson, G.W., Peyden, J.M. and Heriot-Watt, U.:
New Orleans, Louisiana, USA (September 26–29,
“A Systematic Approach to Assessing Permeability
1982).
Impairment Due to Drilling Fluids,” paper SPE 9727
Blann, J.R. and Williams J.D.: “Determining the (November 18, 1980).
Most Profitable Gas Injection Pressure for a Gas Lift
Bruist, E.H.: “Better Performance of Gulf Coast
Installation,” paper SPE 12202, Journal of Petro-
Wells,” paper SPE 4777, presented at the SPE
leum Technology (August 1994), 1305.
Symposium on Formation Damage Control, New
Blanton, R.J.: “Formation Damage Control During Orleans, Louisiana, USA (February 7–8, 1974).
Underreaming and Gravel Packing in an Over-
Bryant, S.L.: “An Improved Model of Mud Acid/
pressured Reservoir,” paper SPE 23804, presented
Sandstone Chemistry,” paper SPE 22855, presented
Exhibition, Dallas, Texas, USA (October 6–9, 1991).
(February 26–27, 1992).
Bryant, S.L. and Buller, D.C.: “Formation Damage
Blanton, T.L. III: “Deformation of Chalk Under
from Acid Treatments,” paper SPE 17597, SPE
Confining Pressure and Pore Pressure,” paper SPE
Production Engineering (November 1990), 455.
8076, presented at the SPE European Offshore
Petroleum Conference and Exhibition, London, Bryant, S.L., Mellor, D.W. and Cade, C.A.: “Physi-
UK (October 24–27, 1978). cally Representative Network Models of Transport

in Porous Media,” AIChE J. (March 1993) 39, No. Christofferson, J. and Christofferson, M.R.: “The
3, 387. Kinetics of Dissolution of Calcium Sulphate Dihy-
drate in Water,” J. Crystal Growth (1976) 35, 79–88.
Buijse, M.A.: “Understanding Wormholing Mech-
anisms Can Improve Acid Treatments in Carbonate Chuah, B.-S., Hasumi, A.R., Samsudin, N. and
Formations,” paper SPE 38166, presented at the SPE Matzain, A.: “Formation Damage in Gravel Packed
European Formation Damage Conference, The and Non-Gravel Packed Completions: A Compre-
Hague, Netherlands (June 2–3, 1997). hensive Case Study,” paper SPE 27360, presented
Bunger, J.W.: “Chemistry of Asphaltenes—Summary
of Symposium,” Proc., Salt Lake City, Utah, USA
(February 7–10, 1994).
(1979), 1028–1031.
Burger, E.D., Perkins, T.K. and Striegler, J.H.:
“Unsteady-State Pressure Distribution Created by
“Studies of Wax Deposition in the Trans Alaska
a Well with an Inclined Fracture,” paper SPE 5591,
Pipeline,” paper SPE 8788, Journal of Petroleum
Technology (June 1981), 1075–1086.
and Exhibition, Dallas, Texas, USA (September
Burman, J.W. and Hall, B.E.: “Foam as a Diverting 28–October 1, 1975).
Technique for Matrix Sandstone Stimulation,” paper
Clementz, D.M.: “Alteration of Rock Properties by
Adsorption of Petroleum Heavy Ends: Implication
Conference and Exhibition, New Orleans, Louisiana,
for Enhanced Oil Recovery,” paper SPE 10683, pre-
sented at the SPE/DOE Third Joint Symposium on
Burton, B.: “Estimate Formation Damage Effects on Enhanced Oil Recovery, Tulsa, Oklahoma, USA
Horizontal Wells,” Petroleum Engineer International (April 4–7, 1982).
(August 1995), 29–34.
Clementz, D.M., Patterson, D.E., Aseltine, R.J. and
Carlson, V., Bennett, E.O. and Rowe, J.A. Jr.: Young, R.E.: “Stimulation of Water Injection Wells
“Microbial Flora in a Number of Oilfield Water- in the Los Angeles Basin by Using Sodium Hypo-
Injection Systems,” paper SPE 1553-G, SPE Journal chlorite and Mineral Acids,” paper SPE 10624,
(June 1961), 71. Journal of Petroleum Technology (September 1982),
Chatelain, J.C., Silberberg, I.H. and Schechter, R.S.: 2087–2096.
“Thermodynamic Limitations in Organic-Acid/ Clemmit, A.F., Ballance, D.C. and Hunton, A.G.:
Carbonate Systems,” paper SPE 5647, SPE Journal “The Dissolution of Scales in Oilfield Systems,”
(August 1976), 189–195. paper SPE 14010, presented at the SPE Offshore
Chenevert, M.E. and Thompson, T.W.: “Perforation Europe 85 Conference, Aberdeen, Scotland, UK
Stability in Low-Permeability Gas Reservoirs,” (September 10–13, 1985).
paper SPE 13902, presented at the SPE/DOE Low Cole, R.J. and Jessen, F.W.: “Paraffin Deposition,”
Permeability Gas Reservoirs Symposium, Denver, Oil & Gas Journal (September 19, 1960), 87–91.
Colorado, USA (May 19–22, 1985).
Collins, S.H. and Melrose, J.C.: “Adsorption of
Cheung, S.K.: “Effect of Acids on Gravels and Asphaltenes and Water on Reservoir Rock
Proppants,” paper SPE 13842, SPE Production Minerals,” paper SPE 11800, presented at the SPE
Engineering (May 1988), 201. International Symposium on Oilfield and Geothermal
Christian, W.W. and Ayres, H.J.: “Formation Chemistry, Denver, Colorado, USA (June 1–3,
Damage Control in Sand Control and Stimulation 1983).
Work,” paper SPE 4775, presented at the SPE Cooper, R.E. and Bolland, J.A.: “Effective Diversion
Symposium on Formation Damage Control, New During Matrix Acidization of Water and Injection
Orleans, Louisiana, USA (February 7–8, 1974). Wells,” paper OTC 4795, presented at the Offshore
Technology Conference, Houston, Texas, USA
(1984).
R-48 References

Coppel, C.P.: “Factors Causing Emulsion Upsets Cunningham, W.C. and Smith, D.K.: “Effect of Salt
in Surface Facilities Following Acid Stimulation,” Cement Filtrate on Subsurface Formations,” paper
paper SPE 5154, Journal of Petroleum Technology SPE 1920, Journal of Petroleum Technology (March
(September 1975), 1060–1066. 1968), 259–264.
Cowen, J.C. and Weintritt, D.J.: Water Formed Scale Cussler, E.L.: Diffusion: Mass Transfer in Fluid
Deposits, Houston, Texas, USA, Gulf Publishing Co. Systems, New York, New York, USA, Cambridge
(1976). University Press (1984).
Crowe, C.W.: “Evaluation of Oil Soluble Resin Daccord, G.: “Acidizing Physics,” Reservoir
Mixtures as Diverting Agents for Matrix Acidizing,” Stimulation, M.J. Economides and K.G. Nolte (eds.),
paper SPE 3505, presented at the SPE Annual second edition, Englewood Cliffs, New Jersey, USA,
Meeting, New Orleans, Louisiana, USA (October Prentice Hall (1989), 13-1–13-13.
3–6, 1971).
Daccord, G. and Lenormand, R.: “Fractal Patterns
Crowe, C.W.: “Evaluation of Agents for Preventing from Chemical Dissolution,” Nature (1987) 325,
Precipitation of Ferric Hydroxide from Spent 41–43.
Treating Acid,” paper SPE 12497, Journal of
Daccord, G., Lenormand, R. and Liétard, O.:
Petroleum Technology (April 1985), 691–695.
“Chemical Dissolution of a Porous Medium by
Crowe, C.W.: “Precipitation of Hydrated Silica from a Reactive Fluid,” Chemical Engineering Science
Spent Hydrofluoric Acid—How Much of a Problem (1993) 48, No. 1, 169–186.
Is It?,” paper SPE 13083, Journal of Petroleum
Daccord, G., Touboul, E. and Lenormand, R.:
Technology (1986), 1234.
“Carbonate Acidizing: Toward a Quantitative Model
Crowe, C.W. and Cryar, H.B. Jr.: “Development of of the Wormholing Phenomenon,” paper SPE 16887,
Oil Soluble Resin Mixture for Control of Fluid Loss SPE Production Engineering (February 1989),
in Water Base Workover and Completion Fluids,” 63–68.
Dahlgaard, C.C.: “Compressed Air and Foam,
Technical Conference and Exhibition, Dallas, Texas,
Cleanout Technique for Old Open Hole
USA (September 28–October 1, 1975).
Completions,” paper SPE 11854, presented at the
Crowe, C.W., Martin, R.C. and Michaelis, A.M.: SPE Rocky Mountain Regional Meeting, Salt Lake
“Evaluation of Acid Gelling Agents for Use in Well City, Utah, USA (May 23–25, 1983).
Stimulation,” paper SPE 9384, SPE Journal (August
da Motta, E.P., Hill, A.D. and Sepehrnoori, K.:
1981), 415.
“Selective Matrix Acidizing of Horizontal Wells,”
Crowe, C.W., McGowan, G.R. and Baranet, S.E.: paper SPE 27399, presented at the SPE International
“Investigation of Retarded Acids Provides Better Symposium on Formation Damage Control, Lafay-
Understanding of Their Effectiveness and Potential ette, Louisiana, USA (February 7–10, 1994).
Benefits,” paper SPE 18222, SPE Production
da Motta, E.P., Plavnik, B. and Schechter, R.S.:
Engineering (May 1990), 166.
“Optimizing Sandstone Acidization,” SPE Reservoir
Crowe, C.W. and Minor, S.S.: “Acid Corrosion Engineering (February 1992a), 149–153.
Inhibitor Adsorption and Its Effect on Matrix Stim-
da Motta, E.P., Plavnik, B., Schechter, R.S. and Hill,
ulation Results,” paper SPE 10650, presented at the
A.D.: “The Relationship Between Reservoir Min-
SPE Formation Damage Control Symposium,
eralogy and Optimum Sandstone Acid Treatment,”
Lafayette, Louisiana, USA (March 24–25, 1982).
Crowe, C.W. and Minor, S.S.: “Effect of Acid Cor- Symposium on Formation Damage Control, Lafay-
rosion Inhibitors on Matrix Stimulation Results,” ette, Louisiana, USA (February 26–27, 1992b).
Danesh, A., et al.: “Asphaltene Deposition in
(1985), 1853.
Miscible Gas Flooding of Oil Reservoirs,” Chem.
Eng. Res. Des. (July 1988) 66, 339–344.

Davies, D.K.: “Clay Technology and Well Stim- SPE Formation Damage Control Symposium,
ulation,” presented at Southwestern Petroleum Short Bakersfield, California, USA (February 8–9, 1988).
Course, Lubbock, Texas, USA (April 20–21, 1978).
Dullien, F.A.L.: Porous Media: Fluid Transport
Davies, D.R., Faber, R., Nitters, G. and Ruessink, and Pore Structure, New York, New York, USA,
B.H.: “A Novel Procedure to Increase Well Academic Press (1979).
Response to Matrix Acidising Treatments,” paper
Dunlap, D.D. and Houchin, L.R.: “Evaluation of
SPE 23621, presented at the Second SPE Latin
Acid System Design and Formation Damage Using
American Petroleum Engineering Conference,
Polarized Microscopy,” paper SPE 19425, presented
Caracas, Venezuela (March 8–11, 1992).
at the SPE Formation Damage Control Symposium,
Davies, D.R., Lievaart, L. and Nitters, G.: “The Lafayette, Louisiana, USA (February 22–23, 1990).
Effects of Corrosion Inhibitors and Mutual Solvents
Dunn, L.A. and Stokes, R.H.: “The Diffusion of
on Matrix Acidizing Treatments,” paper SPE 17153,
Monocarboxylic Acids in Aqueous Solution at 25°,”
Aust. J. Chem. (1965) 18, 285–296.
Symposium, Bakersfield, California, USA (February
8–9, 1988). Durham, D.K., Ali, S.A. and Stone, P.J.: “Causes
and Solutions to Surface Facilities Upsets Following
De Ghetto, G.: “Chart Eases Field Use of AGIP
Acid Stimulation in the Gulf of Mexico,” paper SPE
Matrix-Stimulation Method If Computer Program
29528, presented at the SPE Production Operations
Not Available,” Oil & Gas Journal (November 29,
Symposium, Oklahoma City, Oklahoma, USA (April
1982), 76–77.
2–4, 1995).
Denson, A.H., Smith, J.T. and Cobb, W.M.: “Deter-
Dusseault, M.B. and Gray, K.E.: “Mechanisms
mining Well Drainage Pore Volume and Porosity
of Stress-Induced Wellbore Damage,” paper SPE
from Pressure Buildup Tests,” SPE Journal (August
23825, presented at the SPE International Sympo-
1976), 209–216.
sium on Formation Damage Control, Lafayette,
de Rozières, J., Chang, F.F. and Sullivan, R.B.: Louisiana, USA (February 26–27, 1992).
“Measuring Diffusion Coefficients in Acid Frac-
Earlougher, R.C. Jr.: Advances in Well Test Analysis,
turing Fluids and Their Application to Gelled and
Dallas, Texas, USA, Society of Petroleum Engineers
Emulsified Acids,” paper SPE 28552, presented at
(1977).
bition, New Orleans, Louisiana, USA (September Eaton, B.A. and Smithey, M.: “Formation Damage
25–28, 1994). from Workover and Completion Fluids,” paper SPE
3707, presented at the SPE Annual California
de Vries, A.S. and Wit, K.: “Rheology of Gas/Water
Regional Meeting, Los Angeles, California, USA,
Foam in the Quality Range Relevant to Steam
(November 4–5, 1971).
Foam,” SPE Reservoir Engineering (May 1990),
185–192. Economides, M.J., Ben Naceur, K. and Klem, R.C.:
“Matrix Stimulation Method for Horizontal Wells,”
Doerler, N. and Prouvost, L.P.: “Diverting Agents:
Laboratory Study and Modeling of Resultant Zone
(July 1991), 854–861.
Injectivities,” paper SPE 16250, presented at the SPE
International Symposium on Oilfield Chemistry, San Economides, M.J., Cikes, M., Pforter, H., Udick,
Antonio, Texas, USA (February 4–6, 1987). T.H. and Uroda, P.: “The Stimulation of a Tight,
Very-High-Temperature Gas-Condensate Well,”
Dria, M.A., Schechter, R.S. and Lake, L.W.: “An
paper SPE 15239, SPE Formation Evaluation
Analysis of Reservoir Chemical Treatments,” paper
(March 1989), 63.
SPE 13551, SPE Production Engineering (February
1988), 52. Economides, M.J. and Frick, T.P.: “Optimization
of Horizontal Well Matrix Stimulation Treatments,”
Droddy, M.J., Jones, T.A. and Shaw, D.B.: “Drill
Fluid and Return Permeability Tests of South Texas
Meeting on Petroleum Engineering, Beijing, China
Wilcox Cores,” paper SPE 17159, presented at the
(March 24–27, 1992).
R-50 References

Economides, M.J., Hill, A.D. and Ehlig-Economides, Hydrofluoric Acid,” paper SPE 2621, Journal of
C.: Petroleum Production Systems, Englewood Petroleum Technology (April 1970), 433–440.
Cliffs, New Jersey, USA, Prentice-Hall (1994).
Fetkovitch, M.J. and Vienot, M.E.: “Shape Factor,
Efthim, F.P., Garner, J.J., Bilden, D.M., Kovacevich, CA, Expressed as Skin, sCA,” paper SPE 13304,
S.T. and Pence, T.C.: “Evaluation and Treatment of Journal of Petroleum Technology (February 1985),
Organic and Inorganic Damage in an Unconsolidated 321.
Asphaltic Crude Reservoir,” paper SPE 19412,
Fischer, P.W., Gallus, J.P., Krueger, R.F., Pye, D.S.,
Simmons, F.J. and Talley, B.E.: “An Organic ‘Clay
Symposium, Lafayette, Louisiana, USA (February
Substitute’ for Nondamaging Water Base Drilling
22–23, 1990).
and Completion Fluids,” paper SPE 4651, presented
Ehlig-Economides, C.A. and Joseph, J.A.: “A New at the SPE Annual Meeting, Las Vegas, Nevada,
Test for Determination of Individual Layer USA (September 30–October 3, 1973).
Properties in a Multilayered Reservoir,” paper SPE
Fisher, A.W., Foulser, R.W.S. and Goodyear, S.G.:
“Mathematical Modeling of Foam Flooding,” paper
Conference and Exhibition, Las Vegas, Nevada,
SPE 20195, presented at the SPE/DOE Symposium
on Enhanced Oil Recovery, Tulsa, Oklahoma, USA
Elbel, J.L., Navarrete, R.C. and Poe, B.D. Jr.: (April 22–25, 1990).
“Production Effects of Fluid Loss in Fracturing
Flanigan, M.J.: “Smaller Gravel and Coated Screens
High-Permeability Formations,” paper SPE 30098,
Enhance 50-Year-Old Field,” paper SPE 7973,
presented at the SPE European Formation Damage
Journal of Petroleum Technology (May 1980),
Conference, The Hague, Netherlands (May 15–16,
757–763.
1995).
Fletcher, P.A., Montgomery, C.T. and Ramos, G.G.:
Erbstoesser, S.R.: “Improved Ball Sealer Diversion,”
“Optimizing Hydraulic Fracture Length to Prevent
Formation Failure in Oil and Gas Reservoirs,” Proc.,
(November 1980), 1903.
35th U.S. Rock Mechanics Symposium, Reno,
Ettinger, R.A. and Radke, C.J.: “Influence of Texture Nevada, USA (June 1995), 293–298.
on Steady Foam Flow in Berea Sandstone,” paper
Fogler, H.S. and Crain, E.R.: “Stimulation of Gas
SPE 19688, SPE Reservoir Engineering (February
Storage Fields to Restore Deliverability,” paper SPE
1992), 83.
Ezeukwu, T., Thomas, R.L. and Gunneroed, T.: 1980), 1612–1620.
“Fines Migration Control in High-Water-Cut
Fogler, H.S., Lund, K. and McCune, C.C.: “Acid-
Nigerian Oil Wells: Problems and Solutions,” paper
ization. Part 3. The Kinetics of the Dissolution of
SPE 39482, presented at the SPE International
Sodium and Potassium Feldspar in HF/HCl Acid
Symposium on Formation Damage Control,
Mixtures,” Chem. Eng. Sci. (1975) 30, No. 11,
Lafayette, Louisiana, USA (February 18–19, 1998).
1325–1332.
Faber, M.J., Davies, D.R., Nitters, G. and Ruessink,
Fogler, H.S., Lund, K. and McCune, C.C.:
B.H.: “A Novel Procedure to Increase Well
“Predicting the Flow and Reaction of HCl/HF
Response to Matrix Acidising Treatments,” paper
Mixtures in Porous Sandstone Cores,” paper SPE
SPE 23621, SPE Advanced Technology Series (1994)
5646, SPE Journal (October 1976), 248–260.
2, No. 1, 5–14.
Fogler, H.S., Lund, K., McCune, C.C. and Ault,
Falls, A.H., Musters, I.J. and Ratulowski, J.: “The
J.W.: “Dissolution of Selected Minerals in Mud
Apparent Viscosity of Foams in Homogeneous Bead
Acid,” paper 52C, presented at the American
Packs,” paper SPE 16048, SPE Reservoir Engi-
Institute of Chemical Engineers 74th National
neering (May 1989), 1263.
Meeting, New Orleans, Louisiana, USA (March
Farley, J.T., Miller, B.M. and Schoettle, V.: “Design 1973).
Criteria for Matrix Stimulation with Hydrochloric-

Frasch, H.: “Increasing the Flow of Oil Wells,” Frick, T.P., Mostofizadeh, B. and Economides, M.J.:
U.S. Patent No. 556,669 (March 17, 1896). “Analysis of Radial Core Experiments for Hydro-
chloric Acid Interaction with Limestones,” paper
Fredd, C.N.: “The Influence of Transport and
SPE 27402, presented at the SPE International
Reaction on Wormhole Formation in Carbonate
Symposium on Formation Damage Control, Lafay-
Porous Media: A Study of Alternative Stimulation
ette, Louisiana, USA (February 7–10, 1994b).
Fluids,” PhD thesis, University of Michigan, Ann
Arbor, Michigan, USA (1998). Friedmann, F., Chen, W.H. and Gauglitz, P.A.:
“Experimental and Simulation Study of High-
Fredd, C.N. and Fogler, H.S.: “Alternative
Temperature Foam Displacement in Porous Media,”
Stimulation Fluids and Their Impact on Carbonate
paper SPE 17357, SPE Reservoir Engineering
Acidizing,” paper SPE 31074, presented at the SPE
(February 1991), 37–75.
International Symposium on Formation Damage
Control, Lafayette, Louisiana, USA (February Friedmann, F. and Jensen, J.A.: “Some Parameters
14–15, 1996). Influencing the Formation and Propagation of Foams
in Porous Media,” paper SPE 15087, presented at the
Fredd, C.N. and Fogler, H.S.: “Alternative Stimula-
SPE California Regional Meeting, Oakland,
tion Fluids and Their Impact on Carbonate Acid-
California, USA (April 2–4, 1986).
izing,” SPE Journal (March 1998a) 13, No. 1, 34.
Fulford, R.S.: “Effects of Brine Concentration and
Fredd, C.N. and Fogler, H.S.: “Influence of Trans-
Pressure Drop on Gypsum Scaling in Oil Wells,”
port and Reaction on Wormhole Formation in Porous
Media,” AIChE J. (September 1998b), 1933–1949.
(June 1968), 559–564.
Fredd, C.N. and Fogler, H.S.: “The Influence of
Gabriel, G.A. and Erbstoesser, S.R.: “The Design of
Chelating Agents on the Kinetics of Calcite Dissol-
Buoyant Ball Sealer Treatments,” paper SPE 13085,
ution,” J. Colloid and Interface Science (August
1998c) 204, No. 1, 187–197.
and Exhibition, Houston, Texas, USA (September
Fredd, C.N. and Fogler, H.S.: “The Kinetics of 16–19, 1984).
Calcite Dissolution in Acetic Acid Solutions,” Chem.
Gadiyar, B.R. and Civan, F.: “Acidization-Induced
Eng. Sci. (October 1998d) 53, No. 22, 3863–3894.
Formation Damage: Experimental and Modeling
Fredd, C.N., Tija, R. and Fogler, H.S.: “The Studies,” paper SPE 27400, presented at the SPE
Existence of an Optimum Damköhler Number for International Symposium on Formation Damage
Matrix Stimulation of Carbonate Formations,” paper Control, Lafayette, Louisiana, USA (February 7–10,
SPE 38167, presented at the SPE European Forma- 1994).
tion Damage Control Conference, The Hague,
Garrels, R.M. and Christ, C.L.: Solutions, Minerals
Netherlands (June 2–3, 1997).
and Equilibria, New York, New York, USA, Harper
Frick, T.P. and Economides, M.J.: “Horizontal Well & Row (1965).
Damage Characterization and Removal,” paper
Gdanski, R.: “Kinetics of Tertiary Reaction of HF on
SPE 21795, presented at the SPE Western Regional
Alumino-Silicates,” paper SPE 31076, presented at
Meeting, Long Beach, California, USA (March
the SPE International Symposium on Formation
20–22, 1991).
Frick, T.P., Kürmayr, M. and Economides, M.J.: (February 14–15, 1996).
“Modeling of Fractal Patterns in Matrix Acidizing
Gdanski, R.D.: “AlCl3 Retards HF Acid for More
and Their Impact on Well Performance,” paper SPE
Effective Stimulations,” Oil & Gas Journal (October
23789, presented at the SPE International Sympo-
28, 1985), 111–115.
sium on Formation Damage, Lafayette, Louisiana,
USA (February 26–27, 1992); also in SPE Gdanski, R.D.: “Experience and Research Show Best
Production & Facilities (February 1994a), 61–68. Designs for Foam-Diverted Acidizing,” Oil & Gas
Journal (September 1993), 85–89.
R-52 References

Gdanski, R.D. and Peavy, M.A.: “Well Return Gougler, P.D. Jr., Hendrick, J.E. and Coulter, A.W.:
Analysis Causes Re-Evaluation of HCl Theories,” “Field Investigation Identifies Source and Magnitude
paper SPE 14825, presented at the SPE Symposium of Iron Problems,” paper SPE 13812, presented at
on Formation Damage Control, Lafayette, Louisiana, the SPE Production Operations Symposium, Okla-
USA (February 26–27, 1986). homa City, Oklahoma, USA (March 10–12, 1985).
Gidley, J.L.: “Stimulation of Sandstone Formations Gudmundson, J.S. and Ortiz-R., J.: “Two-Phase
with the Acid-Mutual Solvent Method,” paper SPE Flow and Calcite Deposition,” paper SPE 12741,
3007, Journal of Petroleum Technology (May 1971), presented at the SPE California Regional Meeting,
551–558. Long Beach, California, USA (April 11–13, 1984).
Gidley, J.L.: “Acidizing Sandstone Formations: A Guidry, G.S., Ruiz, G.A. and Saxon, A.: “SXE/N2
Detailed Examination of Recent Experience,” paper Matrix Acidizing,” paper SPE 17951, presented at
SPE 14164, presented at the SPE Annual Technical the SPE Middle East Oil Technical Conference and
Conference and Exhibition, Las Vegas, Nevada, Exhibition, Manama, Bahrain (March 11–14, 1989).
Guin, J.A. and Schechter, R.S.: “Matrix Acidization
Gidley, J.L., Brezovec, E.J. and King, G.E.: “An with Highly Reactive Acids,” SPE Journal
Improved Method for Acidizing Oil Wells in Sand- (December 1971), 390–398.
stone Formations,” paper SPE 26580, SPE
Guin, J.A., Schechter, R.S. and Silberberg, I.H.:
Production & Facilities (February 1996).
“Chemically Induced Changes in Porous Media,”
Gidley, J.L. and Hanson, H.R.: “Prevention of Ind. & Eng. Chem. Fund. (February 1971) 10, No.
Central Terminal Upsets Related to Stimulation and 1, 50–54.
Consolidation Treatments,” paper SPE 4551, pre-
Gulati, M.S. and Maly, G.P.: “Thin-Section and
sented at the SPE Annual Meeting, Las Vegas,
Permeability Studies Call for Smaller Gravels in
Nevada, USA (September 30–October 3, 1973).
Gravel Packing,” paper SPE 4773, Journal of
Gidley, J.L., Penny, G.S. and McDaniel, R.R.: Petroleum Technology (January 1975), 107–112.
“Effect of Proppant Failure and Fines Migration
Gunter, W.D., Zhou, Z. and Perkins, E.H.:
on Conductivity of Propped Fractures,” paper SPE
“Modeling Formation Damage Caused by Kaolinite
24008, presented at the SPE Permian Basin Oil and
from 25 to 300 Degrees Centigrade in the Oil Sand
Gas Recovery Conference, Midland, Texas, USA
Reservoirs of Alberta,” paper SPE 23786, presented
(March 18–20, 1992); also in SPE Production &
Facilities (February 1995), 20.
Ginest, N.H., Phillips, J.E., Al-Gamber, A.W.A. and (February 26–27, 1992).
Wright, D.W.: “Field Evaluation of Acid Stimulation
Hall, B.E.: “The Effect of Mutual Solvents on
Diverter Materials and Placement Methods in Arab-
Adsorption in Sandstone Acidizing,” Journal of
D Injection Wells with Openhole Completions,”
Petroleum Technology (December 1975), 1439.
paper SPE 25412, presented at the SPE Middle East
Oil Technical Conference and Exhibition, Bahrain Hall, B.E.: “Workover Fluids: Parts 1–5,” World Oil
(April 3–6, 1993). (May–July, October and December 1986), 111–114,
64–67, 65–68, 61–63 and 49–50.
Givens, J.C.: “Old Idea (with New Application)
Increases Oil and Gas Production,” paper SPE 6499 Hall, B.E. and Dill, W.R.: “Iron Control Additives
(November 1976). for Limestone and Sandstone Acidizing of Sweet
and Sour Wells,” paper SPE 17159, presented at the
Goode, D.L, Berry, S.D. and Stacy, A.L.: “Aqueous-
SPE Formation Damage Control Symposium,
Based Remedial Treatments for Reservoirs Damaged
Bakersfield, California, USA (February 8–9, 1988).
by Oil-Phase Drilling Muds,” paper SPE 12501, pre-
sented at presented at the SPE Formation Damage Hall, B.E., Tinnemeyer, A.C. and Underwood, P.J.:
Control Symposium, Bakersfield, California, USA “Stimulation of the North Coles Levee Field with a
(February 13–14, 1984). Retarded HF-Acid,” paper SPE 9934, presented at

the SPE California Regional Meeting, Bakersfield, Hill, A.D. and Zhu, D.: “Real-Time Monitoring of
California, USA (March 25–26, 1981). Matrix Acidizing Including the Effects of Diverting
Agents,” paper SPE 28548, presented at the SPE
Harrison, N.W.: “Diverting Agents—History and
Application,” paper SPE 3653, Journal of Petroleum
Technology (May 1972), 593.
Himmatramka, A.K.: “Analysis of Productivity
Hassen, B.R.: “New Technique Estimates Drilling
Reduction Due to Non-Darcy Flow and True Skin in
Filtrate Invasion,” paper SPE 8791, presented at the
Gravel-Packed Wells,” paper SPE 10084, presented
SPE Symposium on Formation Damage Control,
Bakersfield, California, USA (January 28–29, 1980).
bition, San Antonio, Texas, USA (October 5–7,
Hausler, R.H.: “Guidelines Help Avoid, Treat Oil 1981).
Field Emulsions,” Oil & Gas Journal (September
Hinchman, S.P. and Barree, R.D.: “Productivity Loss
4, 1978).
in Gas Condensate Reservoirs,” paper SPE 14203,
Hebert, P.B., Khatib, Z.I., Norman, W.D., Acock, A. presented at the SPE Annual Technical Conference
and Johnson, M: “Novel Filtration Process Elimin- and Exhibition, Las Vegas, Nevada, USA
ates System Upset Following Acid Stimulation (September 22–25, 1985).
Treatment,” paper SPE 36601, presented at the SPE
Hirasaki, G.J.: “A Review of Steam-Foam Process
Mechanisms,” paper SPE 19518 (1989a).
Denver, Colorado, USA (October 6–9, 1996).
Hirasaki, G.J.: “The Steam-Foam Process,” paper
Hekim, Y., Fogler, H.S. and McCune, C.C.: “The
SPE 19505, Journal of Petroleum Technology (May
Radial Movement of Permeability Fronts and Mul-
1989b), 449–456.
tiple Reaction Zones in Porous Media,” SPE Journal
(February 1982), 99–107. Hirschberg, A., deJong, L.N.J., Schipper, B.A. and
Meijer, J.G.: “Influence of Temperature and Pressure
Hendrickson, A.R.: “Stimulation of Carbonate
on Asphaltenes Flocculation,” paper SPE 11202,
Reservoirs,” Oil and Gas Production from Carbon-
SPE Journal (1984), 283.
ate Rocks, G.V. Chilingar, R.W. Manson and H.H.
Ricke (eds.), New York, New York, USA, Elsevier Hodge, R.M., MacKinlay, W.M. and Landrum, W.R.:
(1972), 309–339. “The Selection and Application of Loss Control
Materials to Minimize Formation Damage in Gravel
Hendrickson, A.R., Hurst, R.E. and Wieland, D.R.:
Packed Completions for a North Sea Field,” paper
“Engineered Guide for Planning Acidizing Treat-
SPE 30119, presented at the SPE European Forma-
ments Based on Specific Reservoir Characteristics,”
tion Damage Conference, The Hague, Netherlands
Trans., AIME (January 1960) 219, 16–23.
(May 15–16, 1995).
Hill, A.D. and Galloway, P.J.: “Laboratory and
Hoefner, M.L. and Fogler, H.S.: “Effective Matrix
Theoretical Modeling of Diverting Agent Behavior,”
Acidizing in Carbonates Using Microemulsions,”
Chem. Eng. Prog. (May 1985), 40–44.
(July 1984), 1157.
Hoefner, M.L. and Fogler, H.S.: “Pore Evolution and
Hill, A.D., Lindsay, D.M., Schechter, R.S. and
Channel Formation During Flow and Reaction in
Silberberg, I.H.: “Sandstone Acidizing: The Devel-
Porous Media,” AIChE J. (January 1988) 34, No.
opment of Design Method,” paper SPE 6607, pre-
1, 45–54.
sented at the SPE-AIME International Symposium
on Oilfield and Geothermal Chemistry, La Jolla, Hoefner, M.L., Fogler, H.S., Stenius, P. and Sjöblom,
California, USA (June 27–28, 1977). J.: “Role of Acid Diffusion in Matrix Acidizing of
Carbonates,” Journal of Petroleum Technology
Hill, A.D. and Rossen, W.R.: “Fluid Placement and
(February 1987), 203–208.
Diversion in Matrix Acidizing,” paper SPE 27982,
presented at the University of Tulsa Centennial Holcomb, D.L.: “Low Surface Tension Hydro-
Petroleum Engineering Symposium, Tulsa, Okla- chloric-Hydrofluoric Acid Mixtures in Low Porosity,
homa, USA (August 29–31, 1994). Low Permeability Sandstones,” paper SPE 5411,
R-54 References

presented at the SPE Oklahoma City Regional Annual Technical Conference and Exhibition, Dallas,
Meeting, Oklahoma City, Oklahoma, USA (March Texas, USA (October 6–9, 1991).
24–25, 1975).
Huang, T., Hill, A.D. and Schechter, R.S.: “Reaction
Holub, R.W., Maly, G.P., Noel, R.P. and Weinbrandt, Rate and Fluid Loss: The Keys to Wormhole Initia-
R.M.: “Scanning Electron Microscope Pictures of tion and Propagation in Carbonate Acidizing,” paper
Reservoir Rocks Reveal Ways to Increase Oil SPE 37312, presented at the SPE International
Production,” paper SPE 4787, presented at the SPE Symposium on Oilfield Chemistry, Houston, Texas,
Symposium on Formation Damage Control, New USA (February 18–21, 1997).
Orleans, Louisiana, USA (February 7–8, 1974).
Huh, D.G. and Handy, L.L: “Comparison of Steady-
Hong, K.C: “Productivity of Perforated Completions and Unsteady-State Flow of Gas and Foaming
in Formations With or Without Damage,” paper SPE Solution in Porous Media,” paper SPE 15078, SPE
4653, Journal of Petroleum Technology (August Reservoir Engineering (February 1989), 77–84.
1975), 1027–1038.
Hung, K.M.: “Modeling of Wormhole Behavior in
Hoover, J.E. (ed.): Remington’s Pharmaceutical Carbonate Acidizing,” PhD dissertation, The Uni-
Sciences, 14th edition, Easton, Pennsylvania, USA, versity of Texas at Austin, Austin, Texas, USA
Mack Publishing Co. (1970). (1987).
Horton, H.L., Hendrickson, A.R. and Crowe, C.W.: Hung, K.M., Hill, A.D. and Sepehrnoori, K.:
“Matrix Acidizing of Limestone Reservoirs,” paper “A Mechanistic Model of Wormhole Growth in
API N906-10-F, presented at the Spring Meeting of Carbonate Matrix Acidizing and Acid Fracturing,”
the Southwestern District, Dallas, Texas, USA paper SPE 16886, Journal of Petroleum Technology
(March 10–12, 1965). (January 1989) 41, No. 1, 59–66.
Houchin, L.R., Dunlap, D.D., Arnold, B.D. and Hurst, R.E.: “Our Petroleum Reservoirs Deserve
Domke, K.M.: “The Occurrence and Control of Better Care,” paper SPE 4659, presented at the SPE
Acid-Induced Asphaltene Sludge,” paper SPE Annual Meeting, Las Vegas, Nevada, USA
19410, presented at the SPE Formation Damage (September 30–October 3, 1973).
Control Symposium, Lafayette, Louisiana, USA
Iler, R.K.: The Chemistry of Silica: Solubility, Poly-
(February 22–23, 1990).
merization, Colloid and Surface Properties, and
Houchin, L.R., Dunlap, D.D., Hudson, L.M. and Biochemistry, New York, New York, USA, J. Wiley
Begnaud, P.C.: “Evaluation of Oil-Soluble Resin as and Sons, Inc. (1979).
an Acid-Diverting Agent,” paper SPE 15574, pre-
Jacobs, I.C. and Thorne, M.A.: “Asphaltene Precip-
itation During Acid Stimulation Treatments,” paper
Exhibition, New Orleans, Louisiana, USA (October
SPE 14823, presented at the SPE Symposium on
5–8, 1986).
Houchin, L.R. and Hudson, L.M.: “The Prediction, USA (February 26–27, 1986).
Evaluation and Treatment of Formation Damage
Jensen, J.A. and Friedmann, F.: “Physical and
Caused by Organic Deposition,” paper SPE 14818,
Chemical Effects of an Oil Phase on the Propagation
presented at the SPE Symposium on Formation
of Foam in Porous Media,” paper SPE 16375, pre-
sented at the SPE California Regional Meeting,
(February 26–27, 1986).
Ventura, California, USA (April 8–10, 1987).
Hower, W.F.: “Influence of Clays on the Production
Jiang, R. and Anson, F.C.: “Association of Electro-
of Hydrocarbons,” paper SPE 4785, presented at the
active Counterions with Polyelectrolytes. 4.
SPE Symposium on Formation Damage Control,
Coordinative Binding of Ru (EDTA) to Poly (4-
New Orleans, Louisiana, USA (February 7–8, 1974).
Vinylpyridine),” J. Phys. Chem. (1992) 96, 452.
Hsia, T-Y. and Behrmann, L.A.: “Perforating Skins
Jiao, D. and Sharma, M.M.: “Formation Damage
as a Function of Rock Permeability and Under-
Due to Static and Dynamic Filtration of Water-Based
balance,” paper SPE 22810, presented at the SPE
Muds,” paper SPE 23823, presented at the SPE

International Symposium on Formation Damage Thermodynamic Predictive Technique,” paper SPE

Control, Lafayette, Louisiana, USA (February 17376, SPE Reservoir Engineering (May 1991), 185.
26–27, 1992).
Keelan, D.K. and Koepf, E.H.: “The Role of Cores
Jones, A.T. and Davies, D.R.: “Qualifying Acid and Core Analysis in Evaluation of Formation
Placement: The Key to Understanding Damage Damage,” paper SPE 5696, Journal of Petroleum
Removal in Horizontal Wells,” paper SPE 31146, Technology (May 1977), 482–490.
presented at the SPE International Formation Control
Keeney, B.R. and Frost, J.G.: “Guidelines Regarding
Symposium, Lafayette, Louisiana, USA (February
the Use of Alcohols in Acidic Stimulation Fluids,”
14–15, 1996).
Jones, D.L. and Brown, K.: “Application of Systems (May 1975), 552–554.
Analysis Can Increase Production by 200 Percent in
Keese, J.A. and Oden, A.L.: “A Comparison of Jet
High Volume Gas Lift Wells,” paper SPE 3547, pre-
Perforating Services: Kern River Field,” paper SPE
sented at the SPE Annual Meeting, New Orleans,
5690, presented at the SPE Symposium on Forma-
Louisiana, USA (October 3–6, 1971).
tion Damage Control, Houston, Texas, USA (January
Jones, F.O. Jr.: “Influence of Chemical Composition 29–30, 1976).
of Water on Clay Blocking of Permeability,” paper
Kennedy, D.K., Kitziger, F.W. and Hall, O.E.: “Case
SPE 631, Journal of Petroleum Technology (April
Study on the Effectiveness of Nitrogen Foams and
1964), 441.
Water-Zone Diverting Agents in Multistage Matrix
Jones, L.G., Blount, E.M. and Glaze, O.H.: “Use of Acid Treatments,” paper SPE 20621, SPE
Short Term Multiple Rate Flow Tests to Predict Production Engineering (May 1992), 203.
Performance of Wells Having Turbulence,” paper
Khatib, Z.I., Hirasaki, G.J. and Falls, A.H.: “Effects
of Capillary Pressure on Coalescence and Phase
Conference and Exhibition, New Orleans, Louisiana,
Mobilities in Foams Flowing Through Porous
Media,” SPE Reservoir Engineering (August 1988),
Jones, L.G. and Watts, J.W.: “Estimating Skin Effect 919.
in a Partially Completed Damaged Well,” paper SPE
Khilar, K.C. and Fogler, S.H.: “Water Sensitivity of
2616, Journal of Petroleum Technology (February
Sandstones,” paper SPE 10103, presented at the SPE
1971), 249–252.
Jones, L.G., Yeh, C.S., Yates, T.J., Bryant, D.W., Antonio, Texas, USA (October 5–7, 1981); also in
Doolittle, M.W. and Healy, J.C.: “Alternate Path Journal of Petroleum Technology (February 1983),
Gravel Packing,” paper SPE 22796, presented at the 55–64.
Kibodeaux, K.R., Zeilinger, S.C. and Rossen, W.R.:
Dallas, Texas, USA (October 6–9, 1991a).
“Sensitivity Study of Foam Diversion Processes for
Jones, R.R., Carpenter, R.B. and Conway, M.W.: Matrix Acidization,” paper SPE 28550, presented at
“A Study of Formation Damage Potential During the SPE Annual Technical Conference and Exhi-
Cementing Operations,” paper SPE 22777, presented bition, New Orleans, Louisiana, USA (September
at the SPE Annual Technical Conference and Exhi- 25–28, 1994).
bition, Dallas, Texas, USA (October 6–9, 1991b).
King, G.E., Anderson, A. and Bingham, M.: “A
Kalfayan, L.J. and Watkins, D.R.: “A New Method Field Study of Underbalance Pressures Necessary to
for Stabilizing Fines and Controlling Dissolution Obtain Clean Perforations Using Tubing-Conveyed
During Sandstone Acidizing,” paper SPE 20076, Perforating,” paper SPE 14321, presented at the SPE
presented at the SPE California Regional Meeting, Annual Technical Conference and Exhibition, Las
Ventura, California, USA (April 4–6, 1990). Vegas, Nevada, USA (September 22–25, 1985).
Kawanaka, S., Park, S.J. and Mansoori, G.A.: King, G.E. and Hollingsworth, H.F.: “Evaluation of
“Organic Deposition from Reservoir Fluids: A Diverting Agent Effectiveness and Cleanup Charac-
R-56 References

teristics Using a Dynamic Laboratory Model—High Krueger, R.F.: “An Overview of Formation Damage
Permeability Case,” paper SPE 8400, presented at and Well Productivity in Oilfield Operations: An
the SPE Annual Technical Conference and Exhi- Update,” paper SPE 17459, presented at the SPE
bition, Las Vegas, Nevada, USA (September 23–26, California Regional Meeting, Long Beach,
1979). California, USA (March 23–25, 1988).
King, G.E. and Lee, R.M.: “Adsorption and Chlor- Krueger, R.G., Vogel, L.C. and Fischer, P.W.: “Effect
ination of Mutual Solvents Used in Acidizing,” of Pressure Drawdown on Clean-Up of Clay- or Silt-
paper SPE 14432, SPE Production Engineering Blocked Sandstone,” paper SPE 1605, Journal of
(May 1988), 205–209. Petroleum Technology (March 1967), 397–403.
Kline, W.E.: “The Catalyzed Dissolution of Silicate Kular, G.S., Lowe, K. and Coombe, D.: “Foam
Materials by Hydrofluoric Acid,” PhD dissertation, Application in an Oil Sands Steamflood Process,”
University of Michigan, Michigan, USA (1980). paper SPE 19690, presented at the Annual Technical
Kline, W.E. and Fogler, H.S.: “Dissolution Kinetics:
The Nature of the Particle Attack of Layered Sili-
cates in HF,” Chem. Eng. Sci. (1981a) 36, 871–884. Kunze, K.R. and Shaughnessy, C.M.: “Acidizing
Sandstone Formations with Fluoboric Acid,” paper
Kline, W.E. and Fogler, H.S.: “Dissolution of Sili-
SPE 9387, SPE Journal (February 1983), 65–72.
cate Minerals by Hydrofluoric Acid,” Ind. Eng.
Chem. Fundam. (1981b) 20, 155–161. Kuo, C.H.: “On the Production of Hydrogen Sulfide-
Sulfur Mixtures from Deep Formations,” Journal of
Klotz, J.A., Krueger, R.F. and Pye, D.S.: “Maximum
Petroleum Technology (September 1972),
Well Productivity in Damaged Formations Requires
1142–1146.
Deep, Clean Perforations,” paper SPE 4792, present-
ed at the SPE Symposium on Formation Damage Kutasov, I.M. and Bizanti, M.S.: “Fluid Losses
Control, New Orleans, Louisiana, USA (February While Drilling,” paper SPE 13963 (1985).
7–8, 1974).
Labrid, J.C.: “Stimulation Chimique: Etude Theor-
Knobloch, T.S., Farouq Ali, S.M. and Diaz, M.J.T.: ique et Experimentale des Equilibres Chimiques
“The Role of Acid-Additive Mixtures on Asphaltene Decrivant l’Attaque Fluorhydrique d’un Gres
Precipitation,” paper SPE 7627, presented at the SPE Argileux,” Revue d’Institut Français du Pétrole
Eastern Regional Meeting, Washington, D.C., USA (October 1971) 26, No. 10, 855–876.
(November 1–3, 1978).
Labrid, J.C.: “Thermodynamic and Kinetic Aspects
Kovscek, A.R. and Radke, C.J.: “Fundamentals of of Argillaceous Sandstone Acidizing,” paper SPE
Foam Transport in Porous Media,” Foams: Funda- 5165, SPE Journal (April 1975), 117–128.
mentals and Applications in the Petroleum Industry,
Lambert, M.E.: “A Statistical Study of Reservoir
L.L. Schramm (ed.), ACS Advances in Chemistry
Heterogeneity,” MS thesis, The University of Texas
Series, American Chemical Society, Washington,
at Austin, Austin, Texas, USA (1981).
D.C., USA (1994) 3.
Lauritzen, S.E., Odling, N. and Petersen, J.: “Model-
Krause, D.C.: “Solids-Free Brines: Efficient Dis-
ling the Evolution of Channel Networks in Carbon-
placement Techniques Can Save You Money,” paper
ate Rocks,” presented at the Eurock Conference
SPE 14830, presented at the SPE Symposium on
(September 1992).
USA (February 26–27, 1986). Lea, C.-M., Hill, A.D. and Sepehrnoori, K.: “Effect
of Fluid Diversion on the Acid Stimulation of a
Krueger, R.F.: “An Overview of Formation Damage
Perforation,” paper SPE 22853, SPE Production
and Well Productivity in Oil Field Operations,”
& Facilities (May 1993), 131.
(February 1986), 131–152. Lee, H.O., Heller, J.P. and Hoofer, A.M.W.: “Change
in Apparent Viscosity of CO2 Foam with Rock

Permeability,” paper SPE 20194, SPE Reservoir North Sea,” paper SPE 30411, presented at the
Engineering (November 1991), 421–428. Offshore Europe Conference, Aberdeen, Scotland
(September 5–8, 1995).
Lejon, K., Thingvoll, J.T., Vollen, E.A. and
Hammonds, P.: “Formation Damage Due to Losses Liétard, O. and Daccord, G.: “Acid Wormholing in
of Ca-Based Brine and How It Was Revealed Carbonate Reservoirs: Validation of Experimental
Through Post Evaluation of Scale Dissolver and Growth Laws Through Field Data Interpretation,”
Scale Inhibitor Squeeze Treatments,” paper SPE presented at the 195th National Meeting of the
30086, presented at the SPE European Formation American Chemical Society, Toronto, Ontario,
Damage Conference, The Hague, Netherlands (May Canada (June 5–11, 1988).
15–16, 1995).
Lieu, V.T., Miller, S.G. and Miller S.: “Chemical
Leon, L.: “The Role of the Service Company in Reactions with Reservoir Sand in the Recovery of
Minimizing and Reducing Formation Contamin- Petroleum by Alkaline Flooding,” paper SPE 12561
ation,” paper SPE 4660, presented at the SPE (1983).
Annual Meeting, Las Vegas, Nevada, USA
Lissant, K.J.: Emulsions and Emulsion Technology,
(September 30–October 3, 1973).
part 1, New York, New York, USA, Dekker, Inc.
Leontaritis, K.J.: “Asphaltene Deposition: A (1974).
Comprehensive Description of Problem
Lissant, K.J. and Mayhan, K.G.: “A Study of
Manifestations and Modeling Approaches,” SPE
Medium and High Internal Phase Ratio Water/Poly-
18892, presented at the SPE Production Operations
mer Emulsions,” J. Colloid and Interface Science
Symposium, Oklahoma City, Oklahoma, USA
(January 1973), 201–208.
(March 13–14, 1989).
Locke, S.: “An Advanced Method for Predicting the
Leontaritis, K.J., Amaefule, J.O. and Charles, R.E.:
Productivity Ratio of a Perforated Well,” paper SPE
“A Systematic Approach for the Prevention and
8804, Journal of Petroleum Technology (Decmeber
Treatment of Formation Damage Caused by Asphal-
1981), 2481.
tene Deposition,” paper SPE 23810, presented at the
SPE International Symposium on Formation Damage Lund, K. and Fogler, H.S.: “Acidization V. The
Control, Lafayette, Louisiana, USA (February Prediction of the Movement of Acid and Perme-
26–27, 1992). ability Fronts in Sandstone,” Chemical Engineering
Science (1976) 31, No. 5, 381–392.
Leontaritis, K.J. and Mansoori, G.A.: “Asphaltene
Flocculation During Oil Production and Processing: Lund, K., Fogler, H.S. and McCune, C.C.:
A Thermodynamic Colloidal Model,” paper SPE “Acidization I: The Dissolution of Dolomite in
16258, presented at the SPE International Hydrochloric Acid,” Chemical Engineering Science
Symposium on Oilfield Chemistry, San Antonio, (1973) 28, 691–700.
Texas, USA (February 4–6, 1987). Lund, K., Fogler, H.S., McCune, C.C. and Ault,
Levich, V.G.: Physicochemical Hydrodynamics, J.W.: “Acidization II—The Dissolution of Calcite in
Englewood Cliffs, New Jersey, USA, Prentice Hall Hydrochloric Acid,” Chemical Engineering Science
(1962). (1975) 30, 825–835.
Li, Y., Sullivan, R.B., de Rozières, J., Gaz, G.L. and Lybarger, J.H. and Gates, H.R.: “SGMA Perfor-
Hinkel, J.J.: “An Overview of Current Acid Frac- mance, Design—Part 1: New Acidizing System
turing Technology with Recent Implications for Yields Sustained Production Increase,” Oil & Gas
Emulsified Acids,” paper SPE 26581, presented at Journal (October 16, 1978a), 59–64.
the SPE Annual Technical Conference and Exhi- Lybarger, J.H. and Gates, H.R.: “SGMA
bition, Houston, Texas, USA (October 3–6, 1993). Performance, Design—Part 2: Proper Reservoir
Liétard, O., Bellarby, J. and Holcomb, D.: “Design, Choice, Fluid Design Keys to SGMA Success,” Oil
Execution, and Evaluation of Acid Treatments of & Gas Journal (October 23, 1978b), 137–142.
Naturally Fractured Carbonate, Oil Reservoirs of the
R-58 References

Mahajan, N.C. and Barron, B.M.: “Bridging Particle sented at the SPE Formation Damage Control
Size Distribution: A Key Factor in the Designing of Symposium, Bakersfield, California, USA (February
Non-Damaging Completion Fluids,” paper SPE 8–9, 1988).
8792, presented at the SPE Symposium on
McLeod, H.O. Jr.: “The Effect of Perforating
Formation Damage Control, Bakersfield, California,
Conditions on Well Performance,” paper SPE 10649,
USA (January 28–29, 1980).
Journal of Petroleum Technology (January 1983), 21.
Maly, G.P.: “Close Attention to the Smallest Job
McLeod, H.O. Jr.: “Matrix Acidizing,” paper SPE
Details Vital for Minimizing Formation Damage,”
13752, Journal of Petroleum Technology (December
paper SPE 5702, presented at the SPE Symposium
1984) 36, 2055–2069.
on Formation Damage Control, Houston, Texas,
USA (January 29–30, 1976). McLeod, H.O. and Coulter, A.W.: “The Use of
Alcohol in Gas Well Stimulation,” paper SPE 1663,
Manlowe, D.J. and Radke, C.J.: “A Pore-Level
presented at the SPE Eastern Regional Meeting,
Investigation of Foam/Oil Interactions in Porous
Columbus, Ohio, USA (November 10–11, 1966).
Media,” paper SPE 18069, SPE Reservoir
Engineering (November 1990), 495. McLeod, H.O. Jr. and Coulter, A.W. Jr.: “The
Stimulation Treatment Pressure Record—An
Maya, L.: “Fluoroboric Acid and Its Hydroxy Deriv-
Overlooked Formation Evaluation Tool,” paper SPE
atives—Solubility and Spectroscopy,” J. Inorg. Nucl.
2287, Journal of Petroleum Technology (August
Chem. (1977) 39, 225–231.
1969) 71, No. 8, 951–960.
McClaflin, G.G. and Whitfill, D.L.: “Control of
McLeod, H.O. Jr., Ledlow, L.B. and Till, M.V.: “The
Paraffin Deposition in Production Operations,” paper
Planning, Execution and Evaluation of Acid Treat-
ments in Sandstone Formations,” paper SPE 11931,
Conference and Exhibition, San Francisco, Cali-
presen

Reservoir Stimulation

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Reservoir Stimulation

Hochgeladen von

Copyright:

Verfügbare Formate

elgoajiroblanco@hotmail.

© 2010 COPYRIGHT MERCADO NEGRO, LAS PLAYITAS. MARACAIBO-EDO. ZULIA, VENEZUELA.

MICHAEL J. ECONOMIDES is Professor of Chemical Engineering at

KENNETH G. NOLTE has held various senior technical and marketing

© 2010 COPYRIGHT MERCADO NEGRO, LAS PLAYITAS. MARACAIBO-EDO. ZULIA, VENEZUELA.

Technical editor: Elsa Kapitan-White

Published by John Wiley & Sons Ltd,

National 01243 779777

e-mail (for orders and customer service enquires): es-books@wiley.co.uk

Visit our Home Page on http://www.wiley.co.uk

All right reserved. No part of this publication may be reproduced, stored

Other Wiley Editorial Offices

John Wiley & Sons, Inc., 605 Third Avenue,

Wiley-VCH Verlag GmbH, Pappelalee 3,

Jacaranda Wiley Ldt, 33 Park Road, Milton,

John Wiley & Sons (Canada) Ltd, 22 Worcester Road,

British Library Cataloguing in Publication Data

ISBN 0 471 49192 6

Produced from camera-ready copy supplied by Schlumberger Dowell

© 2010 COPYRIGHT MERCADO NEGRO, LAS PLAYITAS. MARACAIBO-EDO. ZULIA, VENEZUELA.

Preface: Hydraulic Fracturing, A Technology for All Time

Chapter 1 Reservoir Stimulation in Petroleum Production

Chapter 2 Formation Characterization: Well and Reservoir Testing

© 2010 COPYRIGHT MERCADO NEGRO, LAS PLAYITAS. MARACAIBO-EDO. ZULIA, VENEZUELA.

2-3.3. Spherical flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-10

Chapter 3 Formation Characterization: Rock Mechanics

© 2010 COPYRIGHT MERCADO NEGRO, LAS PLAYITAS. MARACAIBO-EDO. ZULIA, VENEZUELA.

3-6.3. Fracture calibration techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-34

Chapter 4 Formation Characterization: Well Logs

Chapter 5 Basics of Hydraulic Fracturing