Professional Documents
Culture Documents
This paper was prepared for presentation at Asheville 2009, the 43rd US Rock Mechanics Symposium and 4th U.S.-Canada Rock Mechanics Symposium, held in Asheville, NC June 28th – July 1,
2009.
This This paper was selected for presentation at the symposium by an ARMA Technical Program Committee based on a technical and critical review of the paper by a minimum of two technical
reviewers. The material, as presented, does not necessarily reflect any position of ARMA, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of ARMA 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 acknowledgement of where and by whom the paper was presented.
ABSTRACT:
Most borehole stability problems occur during drilling in the overburden shale and mudstone. Typically borehole instabilities are
associated with high pore pressures just above the hydrocarbon reservoir. Worldwide there is a tendency going towards deeper
HPHT reservoirs with increased pressure and temperature. Consequently the drilling window margins are reduced and more
emphasis is put on the borehole stability predictions. In order to prevent water influx the mud weight is kept above pore pressure
and the collapse pressure. Further to avoid fracturing and tensile cracks the mud weight is kept below the fracture gradient. In
order to predict the upper and lower limits for mud weight densities a proper description of the formation is needed. It is well
known that the shale exhibits anisotropic behaviour due to sedimentation processes. The anisotropy is apparent both with respect
to deformation and strength.
This paper presents the borehole stability simulations carried out for an HPHT field offshore Norway. An anisotropic shale model
is implemented and calibrated against laboratory test results. Both drained and partially undrained stress paths are compared. In
total three triaxial tests with different sample orientations are required to fully describe the anisotropic properties. The borehole
stability calculations are carried out in the finite element code Abaqus, where the anisotropic model is implemented as a user
routine. The simulation results are presented for different borehole inclinations, where emphasis is put on predicting the allowable
underpressure.
group, where the shale just above the reservoir has high
1. INTRODUCTION organic content.
1.1. General A narrow borehole stability window is predicted just
A borehole stability study is carried out for a planned above the reservoir. This is because of a high
HPHT (High Pressure High Temperature) field overpressure at top reservoir. In addition the collapse
development offshore Norway. The field is planned with pressure is limited by the predicted fracture gradient.
several production and injection wells. Some wells are The mud weight should be designed to lie between these
drilled prior to production, while others are planned to lower and upper boundaries. However, there are some
be drilled after production startup. Due to the potential uncertainties regarding the overpressure in the shale
problems related to drilling in a depleted reservoir, the section. To be conservative the maximum pressure at top
plan is to start with the deepest hydrocarbon layers. reservoir in the sandstone section is used.
Top reservoir is situated at about 4000m TVD MSL and In order to predict the allowable underpressure in the
consists of sandstone from Late Jurassic age. Typical shale section both analytical and numerical borehole
downhole reservoir pressure and temperature is 785 bar stability analyses were carried out. A one meter shale
and 137 degrees Celsius. The overburden consists of core was recovered and a comprehensive test program
shale formations from the Shetland group and Viking performed. From the test results a proper material design
basis was then established.
1.2. Borehole stability
The definition of borehole stability is often related to
“stuck pipe” or “tight hole” problems during drilling
operations [2]. Most often these incidents occur in shale
or shale-like materials as mudstones. Borehole
instability often occurs due to a mechanical collapse
situation during drillout and subsequent pore pressure
development in the surrounding formations. The
instability can occur both due to shear failure or tensile Figure 1:Dry and wet specimen. Parallel beddings are
failure in the surrounding formations. observed. The spacing grid in left hand picture is one
centimeter. A 2mm scale is provided on the right hand picture.
During drilling a typical undrained response is observed
at the borehole wall. This is due to the stress relaxation
from drilling in the shale. With time these excess pore Figure 1 shows the tested plugs from one dry and one
pressures will dissipate out and equilibrate with the wet specimen. A clear bedding lamination is shown on
surrounding pressures. A redistribution of stresses might both. Results from petrophysical testing are provided in
give time-delayed failure in the formation. Table 1.
The borehole collapse is strongly dependant on wellbore
orientation, especially in areas where the formation is Table 1:Petrophysical properties. Porosity φ is the volumetric
strongly bedded and the stress field is non-isotropic. percentage of lost water and TOC (Total organic carbon) is
given as weight percentage.
Sample Quartz Clay ρdry φ TOC
(%)1 content (g/cm3) (%) (%)
2. LABORATORY TESTING
(%)
A comprehensive test program was initialized for the 1 19.6 47.4 2.48 5.1 7.16
shale cores. The shale formation is a source rock with 2 18.7 64.2 2.51 6.3 3.57
high organic content. It was difficult to get intact
samples due to splitting along the bedding planes
As seen from the petrophysical testing the organic
probably from the coring process. After removal from
content is relatively high, in the range 3.5–7%. The clay
the liner the cores were stored in laboratory grade oil.
content comes mainly from Mica/Illite and Kaolonite.
Undrained compression tests (CIU) were performed with
The geomechanical testing gave an overview of the
angles 0, 60 and 90 degrees between sample axis and
material anisotropy both with respect to stiffness and
bedding normal. Also one drained compression test
strength. For the triaxial testing radial strains were
(CID) was performed to obtain effective stress
measured using a circumferential extensometer.
parameters. In addition one uniaxial compression test
However, it is recommended to use two independent
was carried out at downhole temperature (140 °C), at
radial measurements to fully describe the anisotropic
drained conditions. Four transient permeability tests on
properties.
thin shale discs were run with flow both normal to and
parallel with the bedding plane. The petrographical Elastic stiffness anisotropy was confirmed by acoustic
index testing included porosity, bulk and solid density logging of cores with different bedding orientations. In
measurements. general this shale material is strongly anisotropic,
especially for the elastic range. See Table 2 for a list of
For triaxial testing all samples were consolidated up to
interpreted elastic properties.
30 MPa pore pressure. The confining pressure was
ramped to 35, 40 and 50 MPa respectively to simulate a
range of downhole effective stress conditions. The
samples were left to consolidate for 18 hours overnight.
1
The mineral weight percentages are from whole rock.
νtp: Out of plane Poisson’s ratio. Found from testing
Table 2: Effective stress based Young’s modulus (E’) and normal to bedding = (Δε3/Δε1). Can be found
Poisson’s ratio (ν) from testing. Isotropic behavior is assumed. “exactly” from triaxial test.
Orientation is the angle between normal to bedding and Gtp: Found from testing on inclined sample
sample axis.
Orientation Confining stress E’ ν
[MPa] [GPa] During undrained testing of shale the confining stress is
90 10 16.1 0.17 usually kept constant on a total stress basis. However the
90 20 14.1 0.14 effective confining stress is affected by the pore pressure
0 5 8.1 0.2 response. Consequently the anisotropic effective stress
0 10 6.1 0.15 parameters are difficult to obtain as the radial effective
0 102 5.2 0.2 stress changes. In [8] the relationships between drained
0 20 7.6 0.23 and undrained moduli in poroelasticity are discussed.
60 5 12.3 0.55
Testing showed that the shale behavior was strongly
anisotropic both with respect to strength and stiffness. In
The elastic properties given in Table 2 are based on the addition the temperature affected these parameters. This
assumption of isotropic behavior. There is no special was confirmed by comparing tests run at room
consideration with respect to bedding orientation relative temperature (20 °C) and downhole temperature (140 °C).
to loading direction. Values are found at 50% of peak In Table 3 below the interpreted UCS values are shown.
stress.
2 1 3. MATERIAL MODELING
3 3 3.1. Constitutive formulation
In order to predict allowable underpressure during
Figure 2:Triaxial testing of samples with different loading to drilling it was decided to use numerical simulations. A
bedding plane orientations. Loading normal to bedding is constitutive model [1] was implemented in the Finite
shown left and parallel bedding right. Element Program Abaqus [5].The model is able to
capture the anisotropy both in strength and stiffness. The
To fully describe anisotropic behavior the material well defined concept of CamClay is used as a
parameters need to be redefined. By adding a reference framework. By incorporating an anisotropic tensor the
bedding plane we get a transversal isotropic behavior. basic CamClay equations are modified. In the elastic
This implies that two more Poisson’s ratios and one domain a transversal isotropic (TI) formulation is
additional Young’s modulus are required. Figure 2 implemented requiring five independent parameters to
shows loading directions normal to and parallel to be defined.
bedding plane. In [4] is a procedure for determining the
anisotropic properties from triaxial tests is given.
The yield function is given as:
Et: Transversal Young’s modulus. Found from
⎛ g (θ )q orth ⎞
2
1
testing with loading normal to bedding (θ=0) ⎟ + 2 (p orth σ − p t + a ) − a 2 = 0 (1)
2
F (σ , ε p ) = ⎜
Ep: Parallel Young’s modulus. Found from testing ⎝ 2 M ⎠ b
with loading parallel to bedding (θ=90)
νpp: Bedding plane Poisson’s ratio. Found from where
testing parallel to bedding. =(-Δε3/Δε2).
1 T
2
CID test at downhole temperature qorth = σ Pσ (2)
2
dε =
(λ − κ ) dpc'
p
v '
v pc (3)
v ⎛ ∂F ⎞
qorth : Transformed deviatoric stress i +1 i dλtr ⎜
(λ −κ ) ⎝ ∂σ ⎠
⎟
p c' = p c' e
P : Anisotropy matrix
M : Critical state line Here α and κ are non-dimensional soil constants
pt : Tensile strength describing the virgin loading and unloading-reloading
curves.
p orth σ : Transformed mean effective stress For numerical solution the model is written in Fortran
code and implemented in Abaqus as a user model. At
b : Shape parameter
material point level the implicit Backward Euler return
a : Preconsolidation parameter mapping algorithm is used [6]. The Backward Euler
method is favorable for large loading steps and in case of
curved yield surfaces.
P is an anisotropy matrix, see the Appendix for
more information. The other model parameters are
shown in Figure 3.
p’
Figure 4: Illustration of the Backward Euler for an elastic trial
stress σe.
t + Δt
∂F
Figure 3: Yield surface for anisotropic model [1]. Y1 = σ t + Δt − σ e + ΔλD
∂σ
2
⎛ g (θ )qorth t + Δt ⎞
In the deviatoric plane the model incorporates the Y2 = ⎜⎜
2M
1
( ) 2
⎟ + 2 p orth t + Δt σ − pt + a − a 2
⎟ b
(4)
120
110 θ =0°
100 (Figure 8) gives increased stiffness compared with the
90
80 normal to bedding simulation. However the simulated
70
60 volumetric strain is underpredicted, due to lower radial
θ =60°
50
40
strains compared to true behavior.
30
20 The summary plot of all three directions (θ=0º,60º and
10
0 90º) clearly show the transversal isotropic behavior both
0 -10 -20 -30 -40 -50 -60 -70 -80 -90 -10 -11 -12 -13 -14 -15
0 0 0 0 0 0 in stiffness and strength. All three tests give a softening
Effective mean stress p' [MPa] response post peak. This is due to the associated flow
Isotropic Camclay line 0 degrees fit rule and that the peak stress is reached at effective mean
90 degrees fit 60 degrees fit
0 degrees 90 degrees stress levels well below pc/2. The consolidation stress
60 degrees
level is lower for θ=60°.
Figure 5: Calibrated failure surfaces. Curves are calibrated
against lab test and UCS values. The UCS values are corrected
for downhole temperatures.
-120
-100
Model
-90
-100
-80
-60
-60
-50
Lab test
-40
-40
Lab test
-30
-20 -20
-10
Radial Axial
Radial Axial 0
0
3% 2% 1% 0% -1 % -2 % -3 %
2% 1% 0% -1 % -2 %
Strains
Strains
-0.1 % -0.2 %
Model
Lab test
-0.2 %
Volumetric strain
Volumetric strain
-0.4 % Model
-0.3 %
-0.6 %
-0.4 %
-0.8 %
-0.5 %
Lab test
-1.0 %
-0.6 %
-0.7 % -1.2 %
Figure 6: Undrained triaxial test loading perpendicular to Figure 7: Drained triaxial test perpendicular to bedding
bedding (θ=0˚). Room temperature conditions (20°C). (θ=0°). Downhole temperature conditions (140°C).
Consolidation stress 20 MPa. Axial stress versus axial/radial Consolidation stress 10 MPa. Axial stress vs. axial/radial
strain (top) and deviatoric strain versus volumetric strain strain (top) and volumetric strain vs. deviatoric strain
(bottom). (bottom).
-140 -60
-80
-30
Lab test
-60
Lab test
-20
-40
-10
-20
Volumetric strain
-0.05 %
Model
-0.2 %
-0.07 %
-0.3 %
Lab test
-0.4 % -0.09 % Model
-0.5 %
-0.11 %
-0.6 %
-0.13 %
-0.7 %
-0.8 % -0.15 %
Figure 8: Undrained triaxial test parallel to bedding (θ=90°). Figure 9: Undrained triaxial test inclined to bedding. Angle
Room temperature conditions (20°C). Consolidation stress 20 between loading axis and normal to bedding plane is θ=60º.
MPa. Axial stress vs. axial/radial strain (top) and volumetric Room temperature conditions (20°C). Consolidation stress 5
strain vs. deviatoric strain (bottom). MPa. Axial stress vs. axial/radial strain (top) and volumetric
strain vs. deviatoric strain (bottom).
-140
q
-120
θ=90° D=0
Axial effective stress [MPa]
-100
θ=0°
-80
-60
D<0 D>0
-40
θ=60°
-20
p’
0
2% 1% 0% -1 % -2 % Figure 11: Change in mean effective stress (p’) from
Strains
undrained shear loading (q).
Figure 10: Plot of undrained triaxial tests with bedding angles
(θ) 0º, 60º and 90º. For TI materials it is observed an implicit dilatancy
behavior through the Poisson’s ratios and Young’s
modulus. In other words the effective mean stress
3.3. Undrained response of transverse isotropic change during undrained shear loading depends on the
relative Poisson’s ratios and Young’s modulus.
material
The laboratory tests showed that during assumed
undrained conditions the effective mean stress was In order to explain the undrained Poisson’s effect, the
changing also in the elastic domain. This could be an loading perpendicular to bedding (θ=0) was studied in
indication of dilatancy, i.e. a pore pressure development detail. For undrained loading we have the requirement of
due to undrained shear loading. However for this lab zero volume change.
testing program the reason is probably partially drained
conditions during testing. This could be due to low shear
rates or that the valves were not fully closed. Δε v = 0 (6)
6.0E-04
paths will be different from isotropic materials.
5.0E-04 Although the numbers from the analytical and numerical
4.0E-04 approach are slightly different, the study gives
3.0E-04 confidence in the conclusion that balanced drilling
2.0E-04 (drilling fluid pressure = pore pressure) is not limited by
1.0E-04 well inclination.
0.0E+00
0.0 2.0 4.0 6.0 8.0 10.0 12.0 Should underbalanced drilling of the shale be required, it
Underbalance pressure[MPa] is recommended to do a separate and more in-depth
study of the zones where this is planned.
A detailed data collection program is recommended for
Figure 14: Deformation at borehole wall as a function of the first production well.
underpressure (horizontal axis). Critical underpressure is
defined at the intersection point from linear to non-linear
curve.
WELLBORE STABILITY
-200
-180
-160
Critical underpressure [Bar]
-140
-120
-100
-80
-60
-40
-20
0
0 10 20 30 40 50 60 70 80 90
Well inclination [deg]
5. CONCLUSION
An extensive study is carried out to investigate the
borehole stability for this HPHT field. The effect of
anisotropy was tested by using analytical and numerical
modeling. The effect is most pronounced when drilling
at high inclination angles near parallel to the bedding. A
mixture of failure modes may then appear.
The numerical model is able to capture the effect of
anisotropy both in stiffness and strength. However the
softening is limited to certain stress path, and the local
strain effects at large deformation is yet to be
ACKNOWLEDGEMENTS
The authors want to thank StatoilHydro for supporting
and giving permission to write this paper. Also the
extensive laboratory work carried out at SINTEF
Petroleum Research is appreciated.
REFERENCES
1. Crook A.J.L., Jian Guo Yu, S.M. Wilson m
2002. Development of an Orthotropic 3D
Elastoplastic Material Model for Shale,
SPE/ISRM 78238
2. Fjær, E., P. Horsrud , R.M. Holt, A.M. Raaen,
R. Risnes, 2008. Petroleum Related Rock
Mechanics 2nd edition. Elsevier B.V.
3. Schellekens J.C.J., R. De Borst, 1990. The use
of the Hoffmann yield criterion in finite element
analysis of anisotropic composites. Computers
and Structures vol 37 No 6.
4. Oka F., S. Kimoto, H. Kobayashi, T. Adachi
2003. An elasto-plastic consitutive model for
soft sedimentary rock. 16th ASCE Engineering
Mechanics Conference July 16–18.
5. SIMULIA, ABAQUS v.6.7.EF1. 2007
6. Sluys L.J. Computational Methods in Non-linear
Solid Mechanics. 1999. Koiter Institute Delft.
Delft University of Technology.
7. Janbu N. Grunnlag i geoteknikk, 2nd edition
1989. Tapir forlag. Norwegian Institute of
Technology.
8. Loret B., R. Egidio, Z. Zohra, 2001. Relations
between drained and undrained moduli in
anisotropic poroelasticity. Journal of the
Mechanics and Physics of Solids 40, 2593–
2619.
9. Wood D.M. 1990. Soil Behaviour and critical
state soil mechanics. Cambridge University
Press.
APPENDIX
ANISOTROPY MATRIX
The anisotropy matrix is controlling the deviatoric
strength anisotropy at different inclinations.
⎛Ω 0 ⎞
P = ⎜⎜ ⎟⎟
⎝ 0 Γ⎠
⎡2(α 4 + α 6 ) − 2α 4 − 2α 6 ⎤
⎢
Ω = ⎢ − 2α 4 2(α 4 + α 5 ) − 2α 5 ⎥⎥
⎢⎣ − 2α 6 − 2α 5 2(α 5 + α 6 )⎥⎦
⎡6α 7 0 0 ⎤
Γ = ⎢⎢ 0 6α 8 0 ⎥⎥
⎢⎣ 0 0 6α 9 ⎥⎦