Beruflich Dokumente
Kultur Dokumente
The Mechanical Earth Model Concept and Its Application to High-Risk Well
Construction Projects
Richard Plumb, Stephen Edwards, Gary Pidcock, Donald Lee, Brian Stacey, Schlumberger, SPE Members
during early field delineation has payback extending over the Figure 1 shows a 1-dimensional representation of a mechanical
life of the field. earth model and links to the stratigraphy and 3D-seismic cube.
This paper defines the mechanical earth model, explains From left to right the profiles include: Poisson’s ratio (ν),
why it is important, how it is developed and how it is applied to Young’s modulus (E), unconfined compressive strength
well construction and field development. Also discussed are (UCS), friction angle (φ), pore pressure (Pp), minimum
sources of information and the multi-disciplinary team horizontal stress (σh), maximum horizontal stress (σH), vertical
approach required to generate, revise and maintain an earth
stress (σv), and the direction of horizontal stress axes.
model. Finally, three examples are given where mechanical
Lithologic variations in MEM parameters are governed by
earth models have benefited high-technical and high-economic
the mechanical stratigraphy. Research has shown that rock
risk well construction projects.
strength and earth stresses profiles are modulated according to
the nature of the dominant load-bearing solid phase. The
• In the deep water Gulf of Mexico, mechanical earth mechanical stratigraphy is a bimodal textural model of a
models have been used to forecast and revise pore stratigraphic sequence1. The bimodal classification
pressure while drilling. differentiates rocks with clays as the dominant load-bearing
solid from rocks with quartz or carbonate minerals as the
• Offshore Canada, a mechanical earth model helped to dominant load-bearing solid. Petrophysical models are then
successfully drill the first 6.8 km long extended reach used to transform the mechanical stratigraphy into elastic,
well in the Hibernia field. elasto-plastic and rock strength profiles.
Lateral variations in mechanical properties, associated with
• In a fold and thrust belt setting, in S.America a geologic structure, are captured by linking the mechanical
mechanical earth model was built and revised while stratigraphy to a 3-dimensional (3D) framework model. The
drilling exploration wells. The evolving model helped 3D-framework model consists of surfaces, such as formation
to minimize drilling risks, accelerate drilling learning tops and faults. The surfaces are interpreted from seismic
and to generate valuable knowledge of the reservoir data, guided by log data and the geologist’s lithostratigraphic
early in the life of the field. model.
In its most complete form, the MEM consists of a full 3D
Theory, Definitions and Motivation description of pore pressure, stress and mechanical properties.
The mechanical earth model is a numerical representation In practice, the complexity of the model evolves in step with
of the state of stress and rock mechanical properties for a the acquisition of new information. From exploration to
specific stratigraphic section in a field or basin. The model is development, the model evolves from of a sparse set of 1-
linked to geologic structure through the local stratigraphy and a dimensional profiles to a full 3D description of rock properties
3D seismic cube. Table 1 illustrates the relationship between and earth stresses. The degree of detail captured by the model
components of the model and drilling planning and execution will vary from field to field depending on the perceived
decisions. operational risks.
Table 1 - Relationship between drilling decisions and the
mechanical earth model Why Build an Earth Model
Drilling Earth Rock Failure Rock Geologic The mechanical earth model concept is one of the practical
Decision Stresses, Mechanisms Mechanical Structure
pore pressure Parameters Stratigraphy
outcomes of the Cusiana study. However, the need for
Well location x x x x information embodied in the earth has been know for many
Rig selection x years. Bradley2 focused attention on the role of earth stress in
and BOP wellbore stability analysis. The work of Bell and Gough3,
rating
Trajectory
Zoback et al4, and Plumb and Hickman5 demonstrated
x x x x
analysis conclusively that wellbore instability was due to borehole
Casing design x x stress concentrations and that the location of the stress
Safe Mud x concentrations could be measured using geophysical logs.
weight
Throughout the 1980s the practical theory of wellbore stability
Wellbore x x x x
stability advanced slowly in step with the development of faster
Drilling fluids x x x x computers and better logging tools, such as sonic and imaging
Drilling x x x x logs. At the same time geologists and engineers were gaining
practices experience applying wellbore stability modeling techniques of
Cementing x
Strategy
various levels of complexity (Anderson et al6., Mclean and
Bit selection x x Addis7, Charlez8, Bradford et al9). A breakthrough occurred in
the early 1990s when BP encountered severe wellbore
In its basic form, the MEM consists of depth profiles: of the instability in the Cusiana field in Colombia. The severity of the
elastic and/or elasto-plastic parameters, rock strength and the wellbore instability problems demonstrated that conventional
earth stresses referenced to the local stratigraphic section. approaches to solving wellbore stability problems simply did
IADC/SPE 59128 THE MECHANICAL EARTH MODEL CONCEPT AND ITS APPLICATION TO HIGH-RISK WELL CONSTRUCTION PROJECTS 3
not work (Last et al10). It took a multi-company team of risks. Figure 2 shows a process that has been used successfully
geoscientists and engineers almost 1 year to compile enough to minimize drilling risks on high-risk projects. Consider the
geomechanics information about the field to affect an case of drilling the first high-angle extended-reach well in a
improvement in the drilling performance. During the time field.
when the model was being compiled wellbore stability was a
continuing problem This experience motivated the 1. The first step in the process is to build a mechanical
development of the mechanical earth model. earth model. The MEM represents all geological and
Fortunately few fields in the world today have suffered rock mechanics information that currently exists in the
wellbore stability problems as severe as those in Cusiana. field.
However because of market pressures, operators and service 2. The second step uses information from the MEM to
companies are expected to drill more complex wells in less forecast wellbore stability along the planned well path.
time and at lower cost. Under these constraints, even relatively The stability forecast summarizes the expected drilling
minor-wellbore stability problems in tectonically quiescent performance as a function of measured depth in the
settings can be extremely expensive (from $100,000 to well.
$250,000/day offshore). Under these circumstances, the 3. The third step is to monitor the data while drilling and
tendency to design wells based on “closeology” can lead to to test the model predictions for anomalies.
costly surprises. Important lessons from Cusiana that apply Anomalies in the forecast indicate flaws in the data and
equally well to lower-risk projects include: or the MEM.
• Use of all available data to develop geomechanics 4. The anomalies are analyzed to determine the source(s)
of error. If immediate action is required on the rig, it
knowledge of the field.
can be based on informed decisions.
• Balance the complexity of data analysis with available 5. The fifth step is to correct the MEM. Correcting the
time and information. MEM may be done before remedial action is taken (e.g.
• The value of three specific types of information: failure if the pore pressure was too low, see example 1 below)
mechanisms, state of stress and rock mechanical or it may be done off-line if the geology or stress
properties. changes drastically.
• The value of real-time information. This process systematically captures potentially valuable
• The value of data management and good communication. information about the field earlier than would otherwise be
possible. The wellbore stability forecast is revised, as required
What Can Go Wrong by revisions to the model, and the loop continues.
Planning and drilling high-risk wells without the aid a Implementing this process requires team work, excellent
MEM can lead to numerous surprises which, if handled communications (human and electronic) among staff on the
improperly, could result in operational delays or loss of the planning team, at the rig site and between the rigsite and the
well. Some recent examples of unexpected events on high-risk planning team. The benefits of implementing this process are
projects include: fewer unexpected drilling events and accelerated learning about
the field.
• Wellbore instability develops unexpectedly. What action is Recent experience in a number of high technical and high-
required? Raise mud weight? Drop mud weight? Change economic risk projects has demonstrated that an initial MEM
drilling fluid? A combination of the above? can be generated for most fields in about 1 month. Three
• High angle extended reach wells can be drilled without projects, described below, indicate that the MEM can be refined
problems northeasterly but not southwesterly. Why? continuously as new wells are drilled. Table 2 shows typical
sources of information used to construct a MEM.
• Pore pressure increases faster than expected. Can the well
reach TD with remaining casing strings?
• The well can be drilled without any noticeable drilling Profile Logs Other
problems but time dependent wellbore failure prevents Mechanical Gamma ray, density, Cuttings, cavings,
Stratigraphy resistivity, sonic (Vp) sequence stratigraphy
casing from reaching bottom. What to do? Pore Pressure Sonic (Vp), check shot Interval velocity from
Substantial cost overruns occur when the first time these (Pp) survey, resistivity , seismic, formation test,
questions are asked is as the problem develops. An earth model daily drilling reports
Overburden Bulk density Cuttings
coupled with a process for planning and monitoring the well
Stress (σv)
while drilling can minimize the economic impact.
Stress Direction Oriented multi-arm Structural maps
calipers, borehole images, 3D seismic
Constructing the MEM Oriented velocity
anisotropy
Constructing the mechanical earth model and using it to Minimum Sonic (Vp & Vs), wireline Pp , leak-off tests,
generate a wellbore stability forecast helps to reduce drilling Horizontal stress stress tool, extended leak-off test,
4 RICHARD PLUMB,SPE, STEPHEN EDWARDS,SPE, GARY PIDCOCK,SPE, DONALD LEE,SPE, SCHLUMBERGER IADC/SPE 59128
(σ h ) weights from distant (20 miles) offset wells. Notice the two
microfrac, step-rate
injection tests, database,
daily drilling reports, profiles differ by as much as 3 ppg. The drilling program was
modeling based on the more optimistic pressure profile.
Pore pressure, σh, rock
Maximum borehole images, wellbore Several thousand feet below the mud line, the well took a
Horizontal stress, stress model
strength, database, kick indicating an error in the pressure profile. A real-time
(σ H )
Elastic Sonic (Vp & Vs ) bulk data base, laboratory
pore pressure-monitoring project was then initiated to revise
Parameters density tests on cores, cavings the pressure forecast and to re-evaluate the drilling program. A
E, G, ν real-time pore pressure analysis was performed using sonic
Rock strength Sonic (Vp & Vs) bulk data base, laboratory while drilling (LWD) data as input to the pore pressure model
parameters density, mechanical tests on cores, cavings instead of the seismic interval velocities. The real-time
(UCS, φ) stratigraphy
Failure Borehole image, oriented Daily drilling reports,
pressure analysis was calibrated using drilling data (kicks,
mechanisms multi-arm caliper Cavings (digital images) losses, cavings type etc.) acquired prior to running the LWD
Table 2 - Sources of information used to build the MEM tools.
Back-analysis of the initial earth model showed that the
Entries under the heading labeled logs can either be seismic-based prediction could be calibrated using drilling data
measurements made by wireline tools or logging while drilling to provide a look-ahead pore pressure prediction (Figure 4).
tools. Three of the greatest challenges in building the MEM At half the planned well depth a decision had to be made
are: whether or not to continue drilling the well. The higher than
1. Compiling data from a wide range of disciplines, anticipated pore pressure resulted in setting several casing
(drilling engineers, exploration geologists, mud loggers, shoes shallower than planned and the question was: could the
reservoir engineers etc.). target be reached given the remaining casing strings? On the
2. Data management and organizing the data onto a basis of the revised model the decision was made to continue
computer system. drilling using real-time pressure monitoring and evaluation for
3. Timely processing (editing, splicing data) and the remainder of the well. The real-time monitoring included a
interpreting the data in terms of geomechanics parameters (rock daily evaluation of the current pore pressure forecast using
strength, Pp, earth stresses etc.). LWD measurements of p-wave velocity, resistivity, annular
pressure and gamma ray and a continuous evaluation of the
Case Studies look-ahead prediction. Using this methodology the well
Following are brief descriptions of three high-risk well reached the planned TD safely.
construction projects where we have implemented the Comparison of the look ahead prediction (shown in Figure
integrated process illustrated in Figure 2. Application of this 4) with the actual ECD from this well (Figure 5) shows that
technology has enabled these and other challenging wells to be the look-ahead was accurate over the 8000 ft between the depth
drilled successfully with relatively minor difficulties. at which it was made and TD. The above example illustrates the
Example 1: Deepwater Exploration. Drilling in deep process shown in Figure 2. This process:
water is economically and technically challenging (water depth • resulted in rapid learning about the pressure profile and
>1500 ft). With typical well costs between $30-to $50 million earth stress while drilling the exploration well
the financial risk is very high. An accurate pre-drill pore
pressure forecast can reduce this financial risk significantly. • established effective communication between the
But, pore pressure prediction in deep water remains a challenge Geomechanics planning team onshore and the execution
for earth scientists and operators because of the multiplicity of team offshore
processes that can cause overpressure and the lack of • created a more complete earth model at the end of well,
subsurface data, particularly with exploration wells. At best, an including a mechanical stratigraphy and the state of stress
inaccurate forecast leads to cost over-runs for extra casing (direction, principal stress magnitudes).
strings and lost time due to well control problems. Worse, it
might not even be possible to reach the exploration target if the Experience gained on this first exploration well was then
forecast is seriously in error. Although pore pressure prediction applied to successfully drill a second deepwater exploration
can be a challenge, there are proven prediction methods where well in a different mini-basin.
abnormal pressure results from compaction disequilibrium.
Figure 3 shows a simple pre-drill earth model constructed for Example 2: Offshore Continental Margin. Drilling on
a deepwater exploration well in the Gulf of Mexico by third continental margins in harsh environments such as the North
parties. The model consists of an overburden stress and two Sea and eastern Canada presents other challenges. The
pore pressure curves. The overburden stress was estimated Hibernia field is one such example where high day rates mean
using an empirical model11 for the Gulf of Mexico. The lower the smallest problem can turn into a large expense. The
estimate of pore pressure was calculated using a standard Hibernia field is located in the Jeanne d'Arc Basin, 315 km east
compaction disequilibrium model driven by interval transit of St. John's, Newfoundland, in 80m water depth. The field
times obtained from 3D seismic. The other was based on mud was discovered in 1979 and consists principally of two early
IADC/SPE 59128 THE MECHANICAL EARTH MODEL CONCEPT AND ITS APPLICATION TO HIGH-RISK WELL CONSTRUCTION PROJECTS 5
1980 and 1984 nine vertical appraisal and delineation wells B-16 2
B-16 3
No
Yes
1230
1180
No WBS Model
No WBS Model
were drilled into the highly faulted structure. Development B-16 4 No 1200 No WBS model, Liner wouldn't go to bottom
drilling began in mid 1997 from a large Gravity Base Structure B-16 4Z
B-16 7
No
No
1200
1245
No WBS used
No WBS used
located near the top of the structure. The first four wells had a B-16 6 Yes 1270 No WBS used
maximum deviation angle ranging from 30 to 55 degrees and B-16 5 Yes 1340 W B S u s e d
extended reach wells were planned for waterflooding and gas B-16 11 Yes 1320 W B S u s e d
In economically challenging fields, timely access to fit-for- Table 3 - The impact of wellbore stability modeling on
purpose information is vital for optimizing real-time decisions reservoir evaluation.
on the rig. The Hibernia MEM was built as a contingency for
coping with unexpected wellbore stability problems. All of the Example 3: Fold and Thrust Belt. In addition to providing
early vertical wells were characterized by elongated, overgauge short-term benefits for drillers the MEM contains information
wellbores indicating that hole instability was a risk for that can help to optimize field development:
directional development wells. A mechanical earth model was
first developed for the Hibernia reservoir section using data • optimize location of producer-injector wells
from the vertical delineation wells and the first three deviated- • define trajectories that jointly optimize wellbore stability
development wells. The regional N-S trend for the minimum and reservoir productivity
horizontal stress, previously interpreted from caliper logs, was
confirmed using images from LWD. Log derived rock strength • early identification of reservoir compartmentalization and
profiles were calibrated to rock strength measurements made directional permeability trends.
on a representative range of lithologies in the reservoir section.
The minimum stress profile was calibrated to closure pressure Our third example is from a fold and thrust belt in a
measurements from extended leak-off tests and the maximum tectonically active region of South America. The objective of
horizontal stress was constrained using forward modeling the earth modeling project was to help reduce the risk of
techniques 12. wellbore instability while drilling the second and third
The first application of the MEM was to perform a exploration wells in the field and to develop geomechanics
wellbore stability (WBS) analysis for a high-angle extended- information that would help optimize field development
reach gas injector well, B-16 5 (Table 3). B-16 5 was needed planning.
to re-inject gas for reservoir pressure management. At the time The target formations are located in different fault blocks at
B-16 5 had the longest measure depth and the longest depths ranging from around 15,000 to 17,000 ft. Neither of the
horizontal displacement of any well drilled in Canada. The well fault blocks had been drilled previously. To reach the reservoir,
has a measured length of 6.8 km and is drilled at an azimuth of wells must penetrate about 10,000 ft of over-pressured shale.
The first exploration well was vertical and had to be side
N 25o E at deviation angle of 70 o- to-80 o (Figure 6). The
tracked due to problems in the over-pressured section.
well was completed successfully with out any wellbore
The MEM evolved over a six-month period from starting
stability problems following the mud weight recommendations
one week before the second exploration well spudded. Figure
set out in the wellbore stability forecast (Figure 7). The
wellbore stability analysis would not have been possible in the 8 shows a time line indicating when log data were available to
update the model and MEM-based products delivered to
available time frame with out the MEM. The MEM has been
drilling operations.
used to forecast mud weights on subsequent development
The overburden stress and pore pressure profiles were the
wells, to provide a field wide model of rock strength to support
first parts of the model to be built. A pore pressure forecast
efforts to optimize drill bit selection and to improve ROP.
was completed before the 2X well reached the over-pressured
The reliability of the mud weight forecasts have been validated
section. Drilling data confirmed the accuracy of the pore
in several cases where instability developed when mud weights
pressure profile when splintered cavings appear near the
fell below the recommended range (Table 3). The later point
predicted top of overpressure. The initial earth model,
reinforces the value of monitoring the prediction carefully.
consisting of stress and strength profiles, was completed in
Careful monitoring of actual vs. predicted wellbore stability
time to deliver a wellbore stability forecast before the 2X well
enables errors in the model to be corrected promoting
reached the reservoir section. As each hole section was
accelerated learning and provides confidence in the model
completed the LWD and wireline logs were processed and used
when the well performs as expected. The better the model the
to refine the MEM.
fewer the unexpected wellbore instability incidents and lower
Note the well 3X spudded before 2X reached TD.
costs related to non-productive time.
Information gained on 2X enabled a pore pressure forecast and
a wellbore stability forecast to be delivered for the entire well
6 RICHARD PLUMB,SPE, STEPHEN EDWARDS,SPE, GARY PIDCOCK,SPE, DONALD LEE,SPE, SCHLUMBERGER IADC/SPE 59128
10. Last, N., Plumb, R.A., Harkness, R., Charlez, P., Alsen, J.,
McLean, M., “An Integrated Approach to Managing
Wellbore Instability in the Cusiana Field, Columbia, South
America” SPE 30464, Dallas 22-25 (Oct 1995)
Stratigraphy 10 Young’s
Modulus (E)
100 0 F.Angle (φ) 70
Stress Direction σh
0 Poisson’s 1 20 UCS 400 0 Stress 200 W N E
Ratio (ν) MPa
Structure and Stratigraphy
1.0
Grain
Support
Facies
fault ?
Clay
Support
Facies ν Ε UCS φ
Regional
σh σH σv
Pp Trend
X1000
X2000
X3000
X4000
X5000
X6000
X1000
X4500
X8000
X11500
X15000
X18500
X1000
X4500
X8000
X11500
X15000
X18000
Plasticity effect
Start of Project
0 2 0 4 0 6 0 8 0 1 0 0 1 2 0 1 4 0 1 6 0 1 8 0 2 0 0
Logs
Logs Logs
1 6 0 0 0
Logs Logs
1 8 0 0 0
Figure 8 - Time depth curves showing key input from the geomechanics
team. Note that the MEM was supporting two wells that were being
drilled simultaneously. Two-way communication between the field and the
planning team can not be adequately shown in this figure.
12 RICHARD PLUMB,SPE, STEPHEN EDWARDS,SPE, GARY PIDCOCK,SPE, DONALD LEE,SPE, SCHLUMBERGER IADC/SPE 59128
200
180
Days from Spud to TD
160
140
120
100
80
60
40
20
0
1 2 3 4 5
Sequence Drilled
Geomechanics Projects