SPE-174976-MS

How to Effectively Strengthen Wellbores in Narrow Drilling Margin

Wells: An Experimental Investigation
Omid Razavi, Ali Karimi Vajargah, and Eric van Oort, The University of Texas at Austin; Munir Aldin,
and Robert Patterson, Metarock Laboratories

Copyright 2015, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Annual Technical Conference and Exhibition held in Houston, Texas, USA, 28 30 September 2015.
This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
Wellbore strengthening (WBS) offers enabling technology for wells that are drilled in geological
environments with a narrow drilling margin. Through its deployment, costly lost circulation events may
be avoided, casing setting depths may be extended, and, in optimum cases, deeper targets may be reached
with a reduced or slimmed-down casing program.
The elevation of the fracture gradient offered by WBS is a complex issue that involves the growth of
fractures in permeable or impermeable rocks using non-Newtonian drilling fluids that are laden with
solids of varying types and sizes. Several plausible (and sometimes contradictory) models have been
proposed historically to explain the WBS phenomenon, and the only way to assess the correct explanation
is through dedicated experimentation. In this paper, an experimental technique to study WBS under
realistic conditions is introduced, and the results of a series of larger-scale fracturing experiments using
this technique are presented.
The experimental set-up described here consists of a dual flow-loop/ pressure-intensifying system to
carry out high-pressure borehole fracturing tests on cylindrical rock samples while maintaining continuous
circulation of the drilling fluid within the borehole. The system offers full control over pore pressure,
radial confining pressure and, if desired, independent axial pressure. Several injection cycles are performed to characterize the values of the fracture initiation pressure (FIP) and fracture propagation pressure
(FPP) and thereby characterize WBS effects. Typical experimental variables included: the type of base
fluid (water-based, oil- or synthetic-based), the concentration, type, and particle size distribution (PSD) of
lost circulation materials (LCMs) used to achieve WBS effects, and the type of rock tested (sandstone and
shale, i.e. permeable and impermeable rock media). Additionally, post-fracturing techniques such as
thin-section analysis were employed to study the fracture geometry and deposition structure of plugging
solids on the fracture surfaces.
The experiments clearly show that for any rock with a given set of rock strength and failure parameters,
there is an optimum PSD for maximizing WBS effects. Optimum PSD appears to be of primary
importance, almost irrespective of LCM type. The results furthermore show that although a minimum
concentration of LCM bridging agents is required for effective WBS, FPP does not increase significantly
for concentrations above a certain upper threshold value. Moreover, increasing the injection volume

SPE-174976-MS

during WBS squeeze treatments above a threshold value may actually lead to lower FPP values. All of
these findings have important implications for field application of WBS treatments. In addition, petrographic imaging of the fracture after testing show that fracture plugging occurs in the proximity of the
fracture tip and not close to wellbore face, in direct support of the Fracture Propagation Resistance (FPR)
model of WBS, and in disagreement with Wellbore Stress Augmentation/ Stress Cage models.
The results not only confirm information from previous investigations, but also provide new insight
into effective ways to strengthen wellbores in various formations. The experimental results are directly
applicable to improve well construction and to minimize non-productive time on narrow drilling-margin
wells such as (ultra-) deep-water wells by selecting the appropriate mud formulations, LCM materials and
their concentrations, as well as application treatments.

Introduction and Background

General
The drilling margin (or mud window) is defined as the difference between the fracture pressure and either
the pore pressure or the mud pressure required for mechanical wellbore stability, whichever of the two is
higher (Karimi Vajargah and van Oort, 2015). The margin shrinks in geo-pressured formations or in
open-hole zones with production depleted formations present. Note that both situations often occur at the
same time in more mature deepwater drilling environments. A narrow drilling margin complicates well
delivery by posing severe lost circulation, well control or well instability risks that (in extreme cases) may
lead to losing the well and having to drill costly sidetracks (see e.g., Gradishar et al., 2013).
WBS, which offers extension of the drilling margin through enhancement of the FPP, is an effective
method to drill wells in geological environments with a narrow drilling margin. The potential of WBS to
deliver such wells with reduced trouble time and cost was shown to compare very favorably with
mechanical methods to negotiate the drilling margin based on managed pressure drilling (MPD) and
dual gradient drilling (DGD) techniques (van Oort and Razavi, 2014). Implementation effort and costs are
significantly lower for WBS compared to MPD/DGD. Moreover, since WBS works on favorably
extending the drilling margin while MPD/DGD creates a more advantageous annular pressure profile that
better fits the available margin, the two technologies can be used synergistically to significant cumulative
benefit (Jacobs, 2014).
Although WBS has now been part of field drilling practice for several decades, the industry is still far
from exploiting its full potential. Confusion persists on the true mechanism underlying WBS and the best
way to apply WBS treatments in the field, as indicated by the contradictory theories to explain the
phenomenon. A critical review of all the major WBS theories was presented elsewhere (van Oort and
Razavi, 2014), and is not repeated here. The previous work, however, highlighted the evident need to
extend the literature on WBS experimentation to help clarify the true nature of the WBS phenomenon, and
to identify practical ways to optimally exploit it for well construction optimization purposes, in particular
for narrow margin wells.
This paper reviews the development and testing of a state-of-the-art experimental set-up designed and
manufactured to investigate the WBS phenomenon under realistic conditions. Large-scale fracturing
experiments were conducted to study various factors affecting the WBS phenomenon. A comprehensive
framework to maximize the strengthening effect in permeable formations is presented here. In addition,
the primary test results on shale samples are presented, and the challenges of performing WBS experiments in shale are discussed. Before discussing the new equipment, however, a succinct overview of
relevant prior WBS experimental studies and some of their main conclusions is given first.
DEA 13 Investigation
The history of WBS arguably starts with the groundbreaking DEA 13 investigation (19851988, see Black
et al., 1985, DEA 13 Final Report (Phase II) 1988, Onyia, 1994) into the effects of drilling fluid

SPE-174976-MS

formulation and additives on rock fracturing behavior. High-pressure fracturing experiments were
performed on large-scale samples under poly-axial stress conditions. The experiments showed FIP to be
independent of mud type and formulation. Furthermore, the FPP was found to be strongly dependent on
the mud formulation and the use of certain additives that could plug fractures. Significant FPP increase
(60%) was observed in sandstone samples by introducing 40 pounds per barrel (ppb) of calcium
carbonate particles in certain size ranges. However, the scope of the experiments was limited in terms of
the concentration, particle size distribution, and type of the WBS materials tested, with some of the
popular present-day WBS materials (e.g. graphite, fiber) not yet available at the time of the investigation.
In addition, it is probably fair to state that with todays inflation in cost, a repeat / extension of the DEA
13 investigation using such large-scale equipment and samples as were used in the 1980s would probably
be cost-prohibitive.
GPRI Project
The GPRI project on Minimizing Lost Circulation in Synthetic Based Mud (Dudley et al., 2000) was
performed with the aim to improve the lost circulation control capabilities of invert oil emulsion muds,
notably synthetic-based muds (SBM). WBS experiments were performed on cylindrical samples under
conditions of isostatic confining stress. The GPRI experiments confirmed the fundamental findings of
DEA 13 investigations on FIP and FPP. Moreover, fracture sealing capabilities of various types of lost
circulation materials such as graphite, calcium carbonate, and cellulosic fibers were also investigated. The
GPRI project postulated two distinct mechanisms for WBS generated by the use of LCMs: (1) fracture
bridging performed by coarser particles, (2) impairment of the fracture hydraulic conductivity using finer
particles. However, rigorous guidelines on how to engineer LCM types, loading and particle size
distributions were not provided. The results merely suggested that the ideal LCM blend should contain
coarse bridging agents to form a seal in the fracture and finer particles to reduce the permeability of the
formed seal. No post-testing analysis was performed on the possible relationship between the geometry
of the formed fracture and the optimum LCM particle size distribution. Due to lack of pressure control
to maintain constant confining pressure during the propagation injection, the fracture propagation
pressures (FPPs) were measured based on single fracture propagation cycles. Relying on single data points
for FPP measurement can be significantly misleading, as will be shown in this paper: a sufficient number
of fracture bridging and re-opening cycles is required to characterize FPP accurately. Additionally, in the
GPRI experiments, the majority of the fracture reopening and propagation cycles were performed under
the same in-situ stress conditions and the effect of confining pressure variation on FPP was not
investigated.
Other Relevant Experimental Studies
The Fracture Studies Joint Industry Project is another notable attempt to better understand the WBS
phenomenon and to find practical ways to exploit its benefits. Several papers have been published on the
findings of this Joint Industry Project (JIP) since its start in the mid-2000s (see e.g., Guo et al. (2009 and
2014), Kageson-Loe et al. (2008), and Sanders et al. (2008)). Several experimental set-ups have been used
in the experimental investigations, including a fracture sealing tester and block testing equipment. The
fracture sealing tester contains two parallel aluminum platens with adjustable opening to simulate fracture
faces. Extensive experimental investigations were conducted to study the effect of LCM PSD, type, and
concentration on fracture bridging and reopening. Although the results provide valuable general insights
into the fracture sealing capabilities of drilling fluids and LCM additives, the relatively simplistic
experimental set-up complicates the extension of the results to a real-life rock fracturing scenario. In
addition, fracturing experiments were performed using a block tester to characterize remedial WBS
treatments on drilling induced- and natural fractures (Guo et al., 2014). Compared to the base-mud tests,
measured FPPs for LCM-laden fluids were increased by more than 300 and 400 percent for induced and
natural fractures respectively. These high FPP values for drilling induced fractures even exceeded the

SPE-174976-MS

rocks initial FIP. This seems unrealistically high and generally not in agreement with the previous
investigations. In this paper, we present an independent study on the realistic increase of the magnitude
of the FPP (and thereby of the fracture gradient) that can be expected in field application.
Contreras et al. (2014 a-b) recently investigated applications of nanoparticle-based drilling fluids for
WBS in both permeable and impermeable formations. Fracturing injection tests were performed under
conditions of axial and radial confinement, but without any explanation of the mechanism(s) employed to
control theses stresses during the injection cycles. It was observed that the mud formulation affects the
borehole breakdown pressure (i.e., FIP), and that applying nanoparticles can increase the FIP significantly,
at least in permeable rocks. These observations, however, contradict the findings of almost all major WBS
investigations, including DEA 13 (Onyia, 1994) and the GPRI (Dudley, 2000) studies. No validation tests
were conducted to support the findings. Furthermore, no pressure time-history was presented for axial and
confining stress during the injections, which complicates the identification of artifacts caused by pressure
communication between the borehole and confining pressure prior to fracture initiation. In addition, the
FIP values for permeable rocks (see Contreras et al. (2014 a)) in some experiments were unrealistically
low, possibly due to fluid leakage from the borehole. Fracturing experiments on shale samples showed
that, despite a slightly positive effect, FIP and FPP values are not significantly increased (certainly
insufficiently enhanced to rely on it in an actual field application setting) by the nanoparticles suspended
in the test fluids when considering the experimental error of the performed tests. Nanoparticles at present,
therefore, do not appear to present a clear path to effective WBS in impermeable rocks such as shales.

Experimental
Equipment Set-Up: The UT MudFrac System
A state-of-the-art experimental set-up was designed and manufactured for in-depth WBS investigations
(Figure 1a-b). The UT MudFrac set up is a dual flow-loop/pressure-intensifying system. Cylindrical rock
samples of 4 inch in diameter and 6 inch in length are used. A 9/16 inch borehole is drilled and flow lines
are then inserted 2.5 inch into each end of the sample, leaving a 1 inch section of the rock exposed for
fracture initiation. The flow lines are epoxied to the rock sample to prevent pressure communication
between the borehole and the vessel. The sample is isolated by using two steel end-caps in the axial
direction and a rubber sleeve in the radial direction (Figure 2 a-d).

SPE-174976-MS

Figure 1The UT MudFrac system: (a) photograph with overlay, indicating essential equipment components; (b) schematic of the dual
flow loop set-up.

SPE-174976-MS

Figure 2Sample preparation and loading into the vessel: (a) rock sample with end caps; (b) sample mounted in the radial confining
sleeve; (c) sample loaded into the test vessel; (d) test vessel and connection flow lines.

A progressive cavity pump is used to circulate the drilling fluid through the flow-loop. In addition, a
rotary vane water pump is used to saturate the rock sample, to examine the connections for leakage before
running the test, and to flush the flow-lines after each test. Since the pressure required for fracture
initiation and propagation cycles is beyond the working pressure range of the mud pump, a pressure
intensifying system consisting of a fluid accumulator and a positive displacement pump is employed. The
accumulator is essentially a fluid container, which separates the pressurizing fluid (water) and drilling
fluid. The fluid accumulator connects the injection pump to the flow loop. Before each injection cycle, the
accumulator is filled with the drilling fluid while the fluid is circulating through the flow loop. The mud
is always sheared, either before or during the fracturing test, to prevent it from gelling-up (note that
allowing the mud to gel up during the fracturing cycles would create test artifacts). Once the accumulator
is filled with enough fluid to perform the fracture initiation or propagation cycles, pressure is intensified
by shutting in the flow loop and extending the injection pump piston to apply hydraulic pressure to the
borehole.
The UT MudFrac System applies isostatic stress to the sample by compressing the confining fluid
(water). Permeable rock samples were tested under isostatic stress condition. Fracturing experiments on
shale samples were typically performed under 1,500 psi axial stress (in addition to the confining pressure,
which could be varied at will) to avoid rock tensile failure along the lamination planes. All experiments
were conducted at room temperature (a planned upgrade to the equipment will allow the use of elevated
temperature and associated control in the experiments).

SPE-174976-MS

Positive displacement pumps are used to control borehole injection pressure, pore pressure, and
confining pressure. Utilizing Proportional Integral Derivative (PID) control algorithms and high performance motion control hardware, the system can accurately control several independent process variables
(e.g., the pressure and flow-rate of injection-, pore-, and confining pumps). In addition, safety shut downs
are hard-coded in the process control system. The data acquisition system records mud density, temperature, pressure and flow rate in real-time.
Rock and Fluid Samples
Two types of rocks were selected to represent permeable and impermeable formations respectively: Berea
Upper Gray Sandstone and Mancos Shale. The material properties of the samples are presented at Table
1. Permeability, porosity, and unconfined compressive strength (UCS) were provided by the supplier of
these rock samples. Berea sandstones elastic moduli (Youngs modulus and Poissons ratio) and fracture
toughness were measured separately. Inhibitive water based-mud (WBM) and SBM were used as the base
fluid systems to which LCM materials were added. Commercial grades of graphites, fibers and gilsonites
that are routinely used in field practice for lost circulation control purposes were used in the investigations.
Rheological properties of the drilling fluids were measured before and after each experiment. Typical
values for plastic viscosity (PV), yield point (YP) and 10 seconds gel strength for the two base mud
systems are presented in Table 2.
Table 1Typical rock properties of the samples in this study
Measured Property

Berea Sandstone

Mancos Shale

Brine Permeability
Gas Permeability
Porosity
UCS
Youngs Modulus
Poissons Ratio
Fracture Toughness

105 md
280-350 md
18 percent
6000-8000 psi
1.5-2 106 psi
0.23-0.3
230 psi

1 nD
1 nD
3.7-7.9 percent
6,300-9,800 psi

Table 2Rheological properties of the drilling fluid systems

used for WBS experiments
Drilling Fluid System

SBM

WBM

Plastic Viscosity
Yield Point
Gel Strength (10 sec)

22-26 cP
20-23 lb/100ft2
9 lb/100ft2

39-45 cP
14-20 lb/100ft2
13 lb/100ft2

Experimental Procedure: Fracture Initiation and Propagation

The result of a complete fracture initiation and propagation cycle on a Berea sandstone sample is
presented in Figure 3, in which injection, pore, and confining pressures are plotted versus time. The
confining pressure was set to 100 psi for the fracture initiation cycle. Since the pore line was open to
atmosphere, pore pressure readings remain very close to the atmospheric pressure during the injection
(i.e., pore pressure build-up is dissipated through the pore line). Fracture initiation was started by shutting
in the flow loop and intensifying pressure in the wellbore by injecting additional drilling fluid into the
closed loop at the rate of 1 cc/sec. The fracture initiation pressure (FIP) was marked by a distinct peak
in the injection pressure (approximately 1,900 psi in Figure 3 for Berea), followed by a sudden drop in
the injection pressure and a simultaneous jump in confining pressure (due to pressure communication
between the borehole and the outside of the sample). In order to prevent any expansion of the confining
sleeve, the injection was stopped immediately after fracture initiation. Subsequently, the created fracture

SPE-174976-MS

was closed by reducing the borehole pressure and increasing the confining pressure to 500 psi (cf. Figure
3). Then, fluid injection was resumed at a rate of 0.1 cc/sec to re-open and propagate the fracture.

Figure 3Fracture initiation and fracture propagation injection: the fracture initiation cycle was performed under 100 psi confining
pressure and atmospheric pore pressure. Note the peak in borehole injection pressure that characterizes fracture initiation, and the
subsequent communication of injection and confining pressure thereafter. The fracture propagation injection cycle was performed
under 500 psi confining pressure. The local injection pressure maxima, marked by the red dotted circles, indicate events where the
fracture re-opens and propagates. Note that the maxima have corresponding peaks in confining pressure due to pressure communication to the outside of the sample during the fracture re-opening/propagation events. FPP is best characterized by taking the overall
average of the propagation pressure values, as indicated in the graph.

As shown in Figure 3, each cycle of fracture re-opening/propagation is distinguished by an injection

pressure ramp up to a local maximum, (i.e., fracture propagation pressure, FPP) followed by an immediate
pressure reduction associated with fracture growth. Our experiments show that a single re-opening /
propagation cycle does not provide a reliable estimate of the fracture sealing capability and wellbore
strengthening effect. Instead, FPP is best characterized by taking the average of injection pressure local
maxima of several fracture re-openings / propagations (as shown in Figure 3). Moreover, performing
consecutive fracture propagations is an effective method to identify any experimental artifacts caused by
flow-line clogging, which results in unrealistically high FPPs. In order to perform the fracture re-opening/
propagation cycles, it is crucial for the pressure control system to re-adjust to the original confining
pressure after each cycle. The UT MudFrac System has the capability to automatically control the
confining pump and maintain a constant confining pressure throughout the propagation injections. A
relatively slow injection rate is also desirable to provide enough time for the confining pump to release
the residual confining fluid. For this experimental set-up, it has been found that 0.1 cc/sec is the best
suitable injection rate to perform the propagation injection cycle.
To study the effect of confining pressure on the formation of any seals along the fracture, fracture
propagation injections were performed at several confining pressures (i.e., in the range of 100-500 psi,
with 100 psi incremental steps), which also provides multiple data points to precisely evaluate the effect
of mud formulation on the FPP. To compare the fracture initiation and propagation injection cycles, the
results are shown in one graph in Figure 3. In the remainder of the paper, however, the initiation and
propagation injection cycles are presented separately for various confining pressures.
Two different stages of fracture propagation were identified during the fracturing experiments.
Initially, a fracture is created perpendicular to the borehole and grows quickly to the boundaries of the
sample (Figure 4a). Since no horizontal stress anisotropy was applied to the sample, the fracture direction

SPE-174976-MS

in horizontal plane is determined predominantly by the orientation of pre-existing micro-fractures and

other strength variation in the intact samples. Subsequently, the created fracture grows in height (i.e.,
along the sample length) during the fracture reopening and propagation cycles (Figure 4b). Post-testing
analysis of the sample shows that the fracture width is generally not a function of fracture height and hence
remains relatively constant along the sample length. In order to study the geometry of the initially created
fracture, in one of the experiments, fluid injection was stopped immediately after the fracture initiation (no
propagation injection was conducted on the sample) and the sample was unloaded and photographed
(Figure 4c). The created fracture propagates in the radial direction and reaches the boundaries of the
sample due to finite sample size and high level of stored energy. However, the created fracture has a
limited height as shown Figure 4c. In addition, it was observed that when the fluid injection was continued
after fracture initiation (to perform fracture re-opening and propagation cycles) the fracture propagates
along the length of the sample (Figure 4d).

Figure 4 Fracture initiation and propagation: (a) fracture initiation schematic: the created fracture initially propagates in the radial
direction; (b) fracture propagation schematic: fracture propagates along the sample length, the width of the fracture remains relatively
constant along the fracture height; (c) fractured sample unloaded after the fracture initiation injection: note the limited fracture height
at sample mid-length; (d) fractured sample after the fracture propagation injection: the fracture has propagated along the entire length
of the sample.

Validation Experiments
Validation experiments were performed to critically examine sample preparation and test procedure.
Fracture initiation and propagation injection cycles were conducted as described above for various fluid
systems. SBM and WBM fluids without any LCM and a mud density of 12 pounds per gallon (ppg),
achieved with the addition of barite, were used for all baseline tests. Subsequently, WBS experiments
were conducted by adding 20 ppb of a graphite-based LCM blend (developed in a previous study, see van
Oort et al., 2011) to the base muds. Figure 5a shows the fracture initiation curves for the validations tests,

10

SPE-174976-MS

including results for both mud types with and without LCM. FIP was observed to be relatively constant
for these tests taking into account the expected experimental error and the natural variation in strength
among Berea sandstone samples. FIP was clearly independent of drilling fluid type or formulation. This
is in full agreement with the results of the DEA 13 investigations (Onyia, 1994). In addition, these
experiments indicate that adding LCM materials to the base muds does not appear to change the near
wellbore tangential stress (hoop stress). Variation in the injection volume pumped until the FIP value is
reached indicates that there are variations in fluid compressibility between the tests, possibly caused by
such factors as entrainment of some air in the mud formulations upon mixing.

Figure 5Results of validation experiments: (a) fracture initiation curves at 100 psi confining pressure; (b-c) fracture reopening and
propagation curves for confining pressures of 100 psi and 500 psi respectively.

The results of fracture propagation injection cycles at 100 psi and 500 psi confining pressures are
compared in Figure 5b and c. Also, the average FPP values for all validation experiments are presented
in Table 3. Unlike FIP, the mud type and formulation can have a significant impact on FPP. The results
show that a significant increase (50% in our experiments) in FPP was achieved by adding the proper
LCM blends to the base fluid. In addition, the base WBM fluid had a higher FPP than the base SBM,
which is again in agreement with the DEA 13 (Onyia 1994) and the GPRI experiments (Dudley et al.,
2000). As explained by van Oort et al. (2011) WBM shields the pressure communication between the
borehole and fracture tip more effectively than SBM. This is due to the capability of WBM to form an
external filter cake in the vicinity of the fracture tip, which shields the tip better from the full hydraulic
force of the mud than an internal filter cake formed by OBM/SBM.

SPE-174976-MS

11

Table 3Average FPP measured in validation experiments

Drilling Fluid System
Confining
Pressure (psi)

100
500

SBM, no
LCM

WBM, no
LCM

SBM
LCM

WBM
LCM

486
1127

1424

661
1869

1150
2097

Results and Discussion

Permeable Rocks: Optimum WBS LCM Blend
The PSD of LCM blends is one of the most important parameters in WBS, affecting the sealing
capabilities of the drilling fluid. To investigate the effect of PSD on sealing properties of LCM blends in
SBM, various graphite-based LCM grades were tested and the results were compared with the base tests
(i.e. SBM without LCM). Figure 6a-b presents the PSD of various LCM grades: a fine and a coarse grade
were tested individually, and they were also combined to form a medium bi-modal grade which was also
tested. The median size (d50) of each grade is presented in Table 4.

Figure 6 Particle size analysis of various graphite-based LCM blends: (a) LCM blend PSD curves; (b) LCM blend cumulative PSD
curves.
Table 4 Median size (d50) of fine, medium, and coarse LCM
blend
LCM Blend

Fine Blend

Medium Blend

Coarse Blend

Median Size (microns)

90

217

393

The results of fracture propagation cycles in the confining pressure range of 100-500 psi are plotted
and compared with the base SBM result in Figure 7. It was observed that the medium grade LCM blend
maximizes the FPP values compared to the fine and coarse grades, which gave comparable results. This
observation clearly indicates that there is an optimum PSD which maximizes the strengthening effect.
Moreover, the result implies that any deviation from this optimum PSD will lead to a sub-optimum FPP
values and therefore a sub-optimum WBS effect. It is also clear that the optimum LCM blend in this case
has a distinct bi-modal (tri-modal when accounting for barite weighting material) distribution. This
important observation and its implications are explored further in the following sections.

12

SPE-174976-MS

Figure 7Effect of LCM PSD on FIP and FPP: (a) fracture initiation injection curve for 100 psi confining pressure; (b-f) fracture
propagation injection cycles for confining pressures of 100 psi, 200 psi, 300 psi, 400 psi and 500 psi respectively; (g) average FPP for
various LCM blends.

SPE-174976-MS

13

The dependence of the magnitude of the WBS effect on particle size distribution has profound
implications for field application. Not only do the optimum LCM PSD and concentration (see Section
Permeable Rocks: Effect of LCM Concentration below) need to be determined for the mud formulations
and mud weights employed as well as the rock formations drilled, they need to be rigorously managed /
characterized and maintained at optimum levels to achieve optimum WBS benefits. However, most LCM
materials, typically in medium to coarser size ranges, undergo significant size degradation with applied
shear, e.g. by shearing these materials through the nozzles of drill bits. It is important that this degradation
is well-understood in real-time and managed accordingly. This subject will be discussed in more detail in
a future paper (van Oort et al., 2016).
Thin-section analysis was conducted to analyze the generated fractures in detail and determine the
location of seals that were created along the fracture surfaces. Fractured samples were first epoxied to
preserve the geometry of the induced fractures after testing. Subsequently, two sections (section A and B)
were extracted along the length of fracture surfaces (i.e., in the radial direction of the cylindrical samples).
Petrographic images were then taken from the sections using the plane-polarized light to delineate the rock
mineralogy and the presence of LCM in the fractures. Plane-polarized images of sections A and B are
shown in Figure 8a-d. Figure 8a-c show section A with different magnifying factors (5X, 10X, 20X). Note
that the borehole is located on the right-hand side of each image, and the fracture tip is on the left-hand
side. Figure 8d shows section B at a 5X magnification.

Figure 8 Plane-polarized petrographic imaging of fracture surface thin section: (a) section A: 5X scale factor; (b) section A: 10X scale
factor; (c) section A: 20X scale factor; (d) section B: 5X scale factor

14

SPE-174976-MS

Void spaces in the images are shown in blue. In Figure 8a and 8d, the clear blue regions within the rock
matrix reflect the induced fractures. Note that the thin sections shown in Figure 8 only represent the small
fractures formed in the vicinity of main planar fracture, which had an average width of approximately 600
microns; it was not possible to take thin sections from this main fracture. The sections were extracted from
a Berea sandstone sample fractured with SBM loaded with a graphite-based LCM blend. Berea sandstone
is mostly composed of quartz and feldspar minerals, which easily can be distinguished in the images.
Organic (carbon-rich) LCMs such as the graphite used in the tests do not reflect plane-polarized light.
Therefore, the dark spots in the images indicate the presence of LCM at places where a seal was formed
along the fracture. In order to investigate the structure of the formed seal, magnified images (Figure 8b-c)
were taken from the zones with the highest concentration of carbon-rich materials. These magnified
images reveal the true nature of the seal structure: in all cases, a bridge is formed by graphite-based LCM
that acts as a dam, while any remaining, smaller flow channels were filled in with smaller-size gray
inorganic particles that were identified to be mainly composed of barium sulfate (barite). Previously
broken seals were identified (see Figure 8b) that ultimately led to new seals to be formed.
Thin section analysis provides several crucial insights into the underlying mechanism of WBS LCM
plugging in fractures:
1. Formed seals are always located in close vicinity of the fracture-tip and never at the borehole face.
In fact, in all of our petrographic image analysis the highest concentration of LCM was observed
close to the fracture tip and NOT at the fracture aperture close to the borehole. Such observations
are in full agreement with the FPR model, which is based on the formation of pressure barriers in
the vicinity of the fracture tip, and in direct contradiction with the stress cage / WSA model, which
attempts to explain WBS by assuming that near-wellbore hoop stress elevation occurs through
particles bridging close to the wellbore (see van Oort and Razavi, 2014, for further discussion).
2. The appearance of the seals confirms a well-known hypothesis on fracture sealing (e.g., Dudley
et al., 2000, Kageson-Loe et al., 2008). According to this hypothesis, two simultaneous mechanisms occur while sealing an induced fracture: (a) fracture bridging happens by deposition of
coarser particles along the fracture width and in relatively close proximity of the fracture-tip, (b)
pressure isolation of the fracture-tip happens by accumulation of finer particles behind the formed
bridge/seal.
3. Several zones with a high concentration of LCM and weighting material were identified (Figure
8a-b) at multiple locations along the length of the fracture. It shows that during fracture
propagation, pressure barriers will temporarily form and subsequently be compromised during
fracture re-opening and propagation. This is due to repeated cycles of fracture bridging -
pressure build-up behind the bridge/seal - breaching of the bridge followed by fracture tip
propagation - formation of a new bridge/seal, etc., as described e.g. by Morita et al. (1990 and
1996).
Permeable Rocks: LCM Type & PSD
To extend the findings on optimum LCM PSD, further tests were conducted with LCM blends that were
not primarily based on graphite. Figure 9a-b show the PSD of two gilsonite-based blends, which were
designed to mimic the PSD of the graphite-based blend. The primary gilsonite blend has a very similar
median size to the graphite-based blend (Table 5). However, unlike the graphite-based blend, the primary
gilsonite blend has unimodal PSD. The modified gilsonite-based blend consisted of both coarse (d50
300 microns) and fine gilsonite particle (d50 120 microns), to provide a broad and bi-modal PSD curve
with a d50 similar to the optimum graphite-based blend.

SPE-174976-MS

15

Figure 9 Particle size distribution of optimum graphite- and gilsonite-based blends: (a) particle size volume concentration, (b) particle
size cumulative volume concentration.
Table 5Median size of optimum graphite-based LCM, primaryand modified-gilsonite based blend
LCM Blend Type

Median Size (microns)

Optimum graphite-based blend

Primary gilsonite-based blend
Modified gilsonite-based blend

217
242
170

In a series of tests, 20 ppb of the primary and modified gilsonite-based materials were added to the base
SBM and the results of fracture propagation tests are compared with the graphite-based blend (Figure 10
b-f). It was observed that both gilsonite blends increase FPP significantly in comparison with the base
SBM with no LCM. However, the modified gilsonite blend shows better sealing properties at all confining
pressures (100-500 psi) in comparison with the primary gilsonite blend. It was observed that a combination of coarse and fine particles provides a superior strengthening effect: coarse particles form a bridge
and fine particles impair the hydraulic conductivity along the fracture. In fact, the modified gilsonite blend
provides strengthening benefits that are almost identical to the optimum graphite-based blend. These
experiment results show that the magnitude of the WBS effect is primarily determined by PSD and to a
much lesser extent by the type of LCM material. This may be the reason why different parties have
reported good WBS results with a variety of different LCM materials: the type of material may be
irrelevant (or at least less important) as long as PSD is optimized.

16

SPE-174976-MS

Figure 10 Fracture initiation and propagation injection cycles for graphite- and gilsonite-based LCM blends: (a) fracture initiation
injection curves; (b-f) fracture propagation injection curves for confining pressures 100 psi, 200 psi, 300 psi, 400 psi and 500 psi
respectively; (g) average FPP for tested LCM types note that the modified gilsonite blend closely mimics the optimum graphite blend
in performance.

SPE-174976-MS

17

Permeable Rocks: Effect of LCM Concentration

Considering practical application and plain economics, it is crucial to identify the optimum LCM
concentration that achieves the maximum WBS effect. In our experiments on Berea Sandstone, the effect
of LCM concentration on FPP was investigated by varying the addition of a graphite based blend that was
previously shown to provide excellent WBS effects (van Oort et al., 2011). As discussed earlier, FIP is
not affected by the LCM concentration (Figure 11a); in fact, the base fluid provided higher FIP than
LCM-laden fluids, probably due to variations in tensile strength/fracture toughness between the samples.
However, no strengthening effect is observed in FIP by adding LCM to the drilling fluids. Fracture
propagation cycles were performed under 100, 200, 300, 400, and 500 psi confining pressures (Figure
11b-f). The average measured FPP values as a function of confining pressure are compiled in Figure 11g.
As expected, increasing the LCM concentration resulted in a significantly higher FPP. However, beyond
a certain concentration (i.e. 20 ppb), no significant further increase in FPP was observed. Therefore, we
believe that 20 ppb is the optimum LCM concentration for WBS purposes, for the tested mud density (12
ppg) and the tested rock material (Berea). It should be noted that the optimum LCM concentration will
change with a change in mud density and a change in rock material, and that only experimental testing
as described here can clarify what the optimum concentration is for any particular case. Also, note that
FPP increase is more significant at higher confining pressures, supposedly because higher confining
pressure provides the possibility of forming more stable seals in narrower fractures.

18

SPE-174976-MS

Figure 11Effect of LCM concentration on FIP and FPP in Berea sandstone tested with 12 ppb SBM: (a) fracture initiation curve; (b-f)
fracture propagation curves for confining pressures 100 psi, 200 psi, 300 psi, 400 psi and 500 psi respectively; (g) average FPP for
tested confining pressures.

SPE-174976-MS

19

Permeable Rocks: Effect of Injection Volume

There are contradictory views provided by different models / techniques on the effect of injection volume
on WBS. The Fracture Closure Stress model (Dupriest, 2005), for instance, prescribes injecting a
relatively large volume of LCM-laden fluids (pills) during a hesitation squeeze to obtain the desired
strengthening effect. It is discussed elsewhere (van Oort and Razavi, 2014) that from a theoretical
standpoint, a large injection volume during squeezing may in fact result in a lower FPP and is therefore
not desirable. Instead, it is more effective to either perform small-volume squeezes or alternatively
practice continuous WBS to immediately arrest the propagation of any induced or re-opened natural
fractures using appropriate levels of optimum-sized LCM carried in the mud at all time.
An attempt was made to clarify the effect of injection volume on WBS using the MudFrac tester.
Consecutive propagation injections were performed at 100 psi confining pressures using SBM with 20 ppb
of graphite-based LCM on Berea sandstone samples, with results shown in Figure 12. It shows the result
of a first fracture re-opening and propagation cycle (blue curve) in a Berea sample with 100 psi
confinement pressure, which is the lowest confining pressure used in the tests and generally results in a
relatively wide fracture. It was observed that after reaching a maximum pressure of around 650 700 psi
at approximately 30 ccs of injection volume, a distinctly downward trend in pressure occurred. Clearly,
while the fracture is growing, the resistance to propagation is reduced. Such behavior is also observed in
field practice during hesitation squeezes when excess volume is squeezed (Dupriest, private communication). From an optimum strengthening perspective, it would therefore have been better to cut the
squeeze short at 30 ccs injection volume in this case. The elevated fracture propagation pressure was
fully recovered in a second cycle (red curve), which indicates that there may be value in repeat squeezing
and layering of LCM material in the fracture if the first squeeze has not delivered an optimum result.
Note also that the repeat squeeze appears to follow a slightly downward trend after 50 ccs of mud volume
was injected, again indicating that squeezing excess volume is not beneficial.

Figure 12Effect of injection volume on FPP: fracture re-opening/propagation cycles with 100 psi confining pressure. Note the
downward trend in the first squeeze cycle (blue curve) after 30 ccs of injection volume. High propagation pressure is recovered with
a second squeeze (red curve), but its propagation values appear to be declining as well after 50 ccs if injection volume.

In summary, these results indicate that: (1) squeezing excess volume beyond an optimum value lowers
the FPP, in agreement with earlier theoretical work (van Oort and Razavi, 2014); (2) if a squeeze
technique is used to achieve WBS, the squeeze volume should be kept at a minimum; (3) if the fractures
are wide (e.g. in a field setting when a natural fracture is encountered), it may be beneficial to carry out
repeat squeezes and layer material in the fracture, particularly if the first squeeze has not delivered an
optimum result.

20

SPE-174976-MS

Impermeable Rocks: Shale Fracturing

Preliminary experiments were performed on Mancos shale samples to investigate the fracturing behavior
and effectiveness of the conventional WBS techniques on impermeable formations. Four samples were
tested using two SBMs (with and without LCM) and two WBMs (Glycol-based and KCl-based). Figure
13a shows fracture initiation injection curves for the four Mancos samples. A significant discrepancy in
FIP values was observed for two shale samples (red and green lines in Figure 13a), which were fractured
with the same SBM (no LCM). Note that Mancos shale is highly heterogonous, which causes discrepancies in both FIP and FPP values. In Figure 13b-d, the fracture propagation curves are shown for 100 psi,
300 psi, and 500 psi confining pressure. It was observed that addition of LCM to the base SBMs might
provide some sealing capability at low confining pressure (100 psi). However, no major strengthening
effects were observed at other injection/confining pressures.

Figure 13Fracture initiation and propagation injection for Mancos shale samples: (a) fracture initiation curves; (b,c,d) fracture
propagation injection curves for confining pressures at 100 psi, 300 psi, and 500 psi respectively, as indicated.

Figure 14a shows the average FPPs obtained from the preliminary shale experiments. The results show
that in shale samples, WBMs do not appear to provide higher FPPs than SBMs. The results from the
various tests are within a relatively tight band, with variations attributed to sample variability. Also, Figure
14b compares the FPPs in sandstone and shale samples for base tests on SBM (no LCM) and for
strengthening test (SBM with 20 ppb graphite-based LCM). Adding LCM to the drilling fluid had only
minor effects on FPP enhancement in shale samples, whereas the effects on sandstone was very
significant. As explained by Morita et al. (1990) and van Oort and Razavi (2014), induced fractures in
shale are highly unstable due to pore pressure build-up and lack of fluid leak-off through the rock matrix.
The lack of fluid leak off also prevents formation of effective pressure barriers along the fracture surfaces.
More research work is currently ongoing to find viable WBS techniques for impermeable formations such

SPE-174976-MS

21

as shales. This work revolves around approaches that do not (solely) rely on fracture plugging by particles
to achieve effective WBS.

Figure 14 (a) Average Mancos shale FPP for various mud systems, with variations in FPP dominated by rock property variation and
experimental error with little discernable fluid effect; (b) average FPP in sandstone and shale samples for SBM, showing clear
strengthening effects achieved with LCM in sandstone (blue curves), but no discernable effect in shale (orange curves).

Conclusions and Recommendations

A state-of-the-art experimental set-up (UT-MudFrac) was developed for the cost effective evaluation of WBS under realistic test conditions. The equipment offers full control over borehole mud
flow and pressure, pore pressure, axial and radial confining pressures for fracturing cylindrical
rock samples. Validation results were in very good agreement with those of earlier investigations
(notably the DEA 13 and GPRI studies), showing FIP to be independent of mud type, and
indicating that FPP can be greatly enhanced using appropriate types and levels of LCM material.
Optimum PSD appears to be of overriding importance in maximizing WBS effects and achieving
elevated FPP values in permeable rock. The type of LCM seems to be of lesser importance. This
observation may explain why different parties historically have favored different WBS materials:

22

SPE-174976-MS

whether one favors calcium carbonate, graphite or gilsonite may simply be a matter of personal
preference (usually inspired by prior success with a particular material or treatment) as long as the
selected PSD is close to optimum. The optimum PSD is dependent on the rock properties that
govern fracture dimensions, and will therefore vary for different rock types. It needs to be carefully
assessed for each particular case.
For blends with similar median size (d50) values, those with a bi-modal distribution (tri-modal
when accounting for weighting material, which also plays an active part in WBS) have a clear
strengthening advantage in permeable rock over unimodal blends. The reason seems to be related
to the underlying mechanism of seal formation in the fractures: the main structure of these seals
is generated by coarser bridging agents, with finer particles deposited behind these coarser solids,
apparently to reduce the permeability of the formed seal and thereby pressure-isolating the fracture
tip. This is consistent with thin section observations.
Thin section analyses of fractured samples also show that seals that have formed in fractures were
located in the close vicinity to the fracture-tip and not at the borehole face. These observations are
in contradiction with the stress cage wellbore stress augmentation approach to wellbore strengthening, which relies on hoop stress elevation caused by near-wellbore fracture widening and
plugging. The results, however, are fully supportive of the fracture propagation resistance
approach to wellbore strengthening, showing that WBS happens deeper within induced or natural
fractures.
As expected, a higher concentration of a blend of LCM materials that is optimum for WBS
purposes will lead to higher FPP in permeable rocks. However, it appears that FPP does not
significantly increase above a certain threshold value. This more is not necessarily better
observation has important consequences for optimum economic field application of WBS treatments. For any WBS treatment, there will be an optimum in LCM concentration that balances
performance and cost. It should be noted that LCM materials, particularly those in medium to
coarse size range tend to degrade in the field under the influence of shear. This degradation,
which affects PSD and optimum concentrations, should be carefully quantified and managed in
order to maintain optimum WBS results.
Increasing the injection volume of an LCM-laden fluid does not result in higher FPP, on the
contrary. The FPP value typically reaches an optimum value after a limited volume has been
injected. Injecting excess fluid beyond this volume, however, may actually reduce the FPP. The
important field implication for WBS treatments that are delivered by squeeze techniques is that
lower-volume squeezes are likely going to be more successful in delivering meaningful WBS
effects than high volume squeezes. Moreover, the test results indicate that repeat squeezing may
be effective in layering LCM material in wider fractures.
Conventional WBS techniques based on solids plugging are not effective in impermeable formations (e.g., shale), which require a radically different approach. New experimental work using the
UT MudFrac equipment is currently underway to achieve WBS in shales using approaches that do
not rely or do not rely exclusively - on particle plugging.

Acknowledgements
The UT MudFrac equipment and the experiments done with it would not have been possible without an
enabling donation by Schlumberger and the active involvement of Susan Rosenbaum and Jim Friedheim,
to whom we owe a great deal of gratitude. The authors would like to sincerely thank ConocoPhillips and
Schlumberger as the lead sponsors of this R&D project. Our very special thanks for their guidance and
support go out to Dave Beardmore, Greg Mullen, Bob Pantermuehl, Gary Collins, Bret Borland, Ernie
Onyia, Son Pham and Kyle Fontenot, either formerly or currently with ConocoPhillips; Jim Friedheim,
Steve Young, and Quan Guo with Schlumberger; and to Greg Vardilos, Robert Joha, Oswaldo Nunez, and

SPE-174976-MS

23

the rest of the support staff at Metarock Laboratories. Additional financial support was provided by
American Gilsonite Company, with special thanks to Dwight Strickland, Ted Stevens and Dario Montes.
Nomenclature

d50
DEA

FIP

FPP

FPR

FRP

LCM

OBM

PSD

PV

SBM

UCS

WBM

WBS

WSA

YP

Median Particle Size in a Particle Size Distribution, micron

Drilling Engineering Association
Fracture Initiation Pressure
Fracture Propagation Pressure
Fracture Propagation Resistance
Fracture Re-opening Pressure
Lost Circulation Material
Oil Based Mud
Particle Size Distribution
Plastic Viscosity
Synthetic Based Mud
Unconfined Compressive Strength
Water Based Mud
Wellbore Strengthening
Wellbore Stress Augmentation
Yield Point

References
Abrams, A. (1977, May 1). Mud Design To Minimize Rock Impairment Due To Particle Invasion.
Society of Petroleum Engineers. doi:10.2118/5713-PA
Alberty, M. W., & McLean, M. R. (2004, January 1). A Physical Model for Stress Cages. Society of
Petroleum Engineers. doi:10.2118/90493-MS
Andreasen, A. H. M. and Andersen, J., Ueber die Beziehung zwischen Kornabstufung und Zwischenraum in Produkten aus losen Krnern (mit einigen Experimenten), Kolloid-Zeitschrift 50 (1930)
217228.
Barkman, J. H., & Davidson, D. H. (1972, July 1). Measuring Water Quality and Predicting Well
Impairment. Society of Petroleum Engineers. doi: 10.2118/3543-PA
Black, A, et al DEA 13 (Phase I) Final Report. Investigation of Lost Circulation Problems and
Apparent Fracture Gradient Reduction Encountered in the Field with Oil-Based Drilling Fluids During
Large-Scale Laboratory Fracturing Experiments. December 1985. Prepared by Drilling Research Laboratory, Inc.
Chellappah, K., & Aston, M. S. (2012, January 1). A New Outlook on the Ideal Packing Theory for
Bridging Solids. Society of Petroleum Engineers. doi: 10.2118/151636-MS.
Contreras, O., Hareland, G., Husein, M., Nygaard, R., & Alsaba, M. (2014a, September 10). Wellbore
Strengthening In Sandstones by Means of Nanoparticle-Based Drilling Fluids. Society of Petroleum
Engineers. doi: 10.2118/170263-MS
Contreras, O., Hareland, G., Husein, M., Nygaard, R., & Alsaba Mortadha T. (2014b, October 27).
Experimental Investigation on Wellbore Strengthening In Shales by Means of Nanoparticle-Based
Drilling Fluids. Society of Petroleum Engineers. doi: 10.2118/170589-MS
DEA 13 (Phase II) Final Report (Volume 1). Investigation of Lost Circulation Problems with Oil-Base
Drilling Fluids. March 1988. Prepared by Drilling Research Laboratory, Inc.
Dupriest, F. E. (2005, January 1). Fracture Closure Stress (FCS) and Lost Returns Practices. Society
of Petroleum Engineers. doi: 10.2118/92192-MS

24

SPE-174976-MS

Dudley, J. W., Fehler, D. F., and Zeilinger, S. Final Project Report of GPRI 2000 Project DC3:
Minimizing Lost Circulation in Synthetic Mud. 2000. Shell International Exploration and Production,
Inc. Bellaire Technology Center. Houston, TX
Gradishar, J., Ugueto, G., & van Oort, E. (2013, March 5). Setting Free The Bear: The Challenges And
Lessons Of The Ursa A-10 Deepwater ERD Well. Society of Petroleum Engineers. doi:10.2118/
163525-MS
Guo, Q., Roy, S., Simpkins, D., and Zamora, M. 2009. A Practical Design Tool for Wellbore
Strengthening. AADE 2009-NTCE-12-05, AADE Fluids Conference, New Orleans, 8-9 April.
Guo, Q., Cook, J., Way, P., Ji, L., & Friedheim, J. E. (2014, March 4). A Comprehensive Experimental
Study on Wellbore Strengthening. Society of Petroleum Engineers. doi:10.2118/167957-MS
Jacobs, T. (2014). Pushing the Frontier through Wellbore Strengthening. Journal of Petroleum
Technology, 66(11), 64 73.
Kageson-Loe, N. M., Sanders, M. W., Growcock, F., Taugbol, K., Horsrud, P., Singelstad, A. V., &
Omland, T. H. (2008, January 1). Particulate Based Loss-Prevention MaterialThe Secrets of Fracture
Sealing Revealed! Society of Petroleum Engineers. doi:10.2118/112595-MS.
Karimi Vajargah, A., & van Oort, E. (2015, March 17). Automated Drilling Fluid Rheology
Characterization with Downhole Pressure Sensor Data. Society of Petroleum Engineers. doi:10.2118/
173085-MS
Morita, N., Black, A. D., & Guh, G.-F. (1990, January 1). Theory of Lost Circulation Pressure. Society
of Petroleum Engineers. doi:10.2118/20409-MS
Morita, N., Black, A. D., Fuh, G.-F. Borehole breakdown pressure with drilling fluidsI. Empirical
results, International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts,
Volume 33, Issue 1, January 1996, Pages 39 51, ISSN 0148-9062
Onyia, E. C., Experimental Data Analysis of Lost Circulation Problems During Drilling With
Oil-Based Mud. SPE Drilling & Completion. Volume 9, Number 1. Pages 2531. March 1994. DOI:
0.2118/22581-PA.
Sanders, M., Young, S. and Friedheim, J. E. 2008. Development and Testing of Novel Additives for
Improved Wellbore Stability and Reduced Losses. AADE-08-DF-HO-19, AADE Fluids Conference,
Houston, 8-9 April.
van Oort, E., Friedheim, J. E., Pierce, T., & Lee, J. (2011, December 1). Avoiding Losses in Depleted
and Weak Zones by Constantly Strengthening Wellbores. Society of Petroleum Engineers. doi:10.2118/
125093-PA
van Oort, E., & Razavi, O. S. (2014, March 4). Wellbore Strengthening and Casing Smear: The
Common Underlying Mechanism. Society of Petroleum Engineers. doi: 10.2118/168041-MS
van Oort, E., Hoxha, B. B., Yang, L., & Hale, A. H.: Automated Drilling Fluid Analysis using
Advanced Particle Size Analyzers, paper in preparation, 2016