Petrotech
Petrotech
Petrotech
P09- 584
ABSTRACT
A comprehensive wellbore stability study was conducted for southern Cambay Basin, which includes
determination of petrophysical and mechanical properties of formations from core analysis and well logs,
pore pressure from well logs and seismic data, in-situ stresses from mini-fracture test data, and pore
pressure change due to osmosis and capillary effects in shale by chemical potential and mud pressure
penetration tests. An innovative poro-elastic wellbore stability model was used for wellbore stability analysis.
Introduction
Drilling practices in southern Cambay Basin face challenges from wellbore instability in shale sections.
Problems such as excessive cavings leading to well washouts, frequent held up and reaming during drilling
and round trips, stuck pipes, high torque and drag, and hole fill up in round trips are very common. These
problems led to poor hole condition, difficult logging operations, bad cementation job, time overrun and more
importantly formation damage arising out of long exposure of drilling mud.
An innovative wellbore stability model was used to analysis wellbore instability problem in southern Cambay
Basin. The model consists of a poroelastic wellbore deformation module, a chemical potential module for
predicting pressure buildup in shale and a reservoir module which takes into account of change in pore
pressure due to invasion of mud filtrate. Data required for the model were derived from well completion
reports, cores/cuttings, well logs and laboratory tests on core samples. The model simulates wellbore
stability under down hole conditions. The output includes time dependent safe mud weight window and mud
weight polar contour diagram which reflects the effects of well trajectory on wellbore stability.
The wellbore stability model was used to understand the behavior of shales in southern Cambay Basin.
Results of this study show that the borehole instability problem can be significantly reduced if mud system
can be tailored towards the mitigation of the shale related problems.
Vertical stress: Vertical stress or overburden stress originates from the combined weight of formation matrix
and the fluid in the pore space overlaying the formation of interest. It is expressed as
DS
: σ ob = ∫ ρ B × g dDs (1)
0
where, σob is the overburden stress at given depth, Ds, ρB is average bulk density and g is the acceleration
due to gravity. Since in the Dahej field the density log in the upper formations is not available, VSP data and
density log trend line were obtained from regression analysis and used to estimate bulk density in formations
where density log is absent. Figure 2 shows the density profile for Well A.
Horizontal stress orientation: The azimuth of horizontal stresses was determined from wellbore breakout
analysis with four arm caliper logs. A breakout rose diagram of Well D is shown in Figure 3 from which the
Page 1 of 7
PETROTECH – 2009
11 - 15 JANUARY 2009, NEW DELHI, INDIA
azimuth of minimum horizontal stress is determined to be N16.8W, which matches well with the regional
stress orientation.
Minimum horizontal stress: The minimum horizontal stress is usually determined from extended leak-off
test. Due to the lack of data in Dahej field, mini-fracture as well as hydraulic fracture data from the
neighboring Gandhar field were used to estimate minimum horizontal stress. The post-shut-in bottom hole
pressure (BHP) vs. shut-in time curves of four wells were analyzed to determine shut-in pressure and closure
pressure. Shut-in pressure was determined for each well using following methods: inflection point method; pw
vs. log(tp+∆t)/ ∆t method; pw vs. log∆t method; log(pw-pa) vs. ∆t method; log pw vs. log tp method; dpw/dt vs.
pw method; pw vs. ∆t method and maximum curvature method. Note that pw is wellbore pressure, pa is trial
pressure (usually taken to be zero), tp is pumping time and ∆t is shut-in time. Closure pressure was
evaluated by step rate test (SRT) analysis; pw vs ∆t method and derivative of G-function analysis. The
gradient of minimum horizontal stress calculated from the average of the initial shut-in pressure and closure
pressure for all the four wells is 0.0213 MPa/m.
Maximum horizontal stress and stress polygon: Stress bounding method - stress polygon was used to
determine the horizontal stress boundaries. This method essentially investigates the range of permissible
values of horizontal stresses by systematically narrowing stress bounds by imposing known constraints. The
stress polygon constructed for Well D at a depth of 2991.8 m is shown in Figure 4. The lower and upper
bounds of horizontal stress gradients obtained from stress polygon analysis range from 0.0167 to 0.0214
MPa/m for minimum horizontal stress and from 0.0234 to 0.0291 for maximum horizontal stress.
Horizontal stress calibration with wellbore stability data from Well D: Since hydraulic fracture data used
in the determination of horizontal stresses are from a neighboring field and are available only in Hazad
member, stress profile in Dahej field need to be calibrated using additional local data. One approach is to
conduct wellbore stability analysis and compare the safe mud weight with actual mud weight used in drilling
operations. Well completion report shows that Well D is a vertical well and was drilled down to a target depth
of 3232 m without wellbore collapse or mud loss complications. Hence this well was selected for horizontal
stress calibration. By comparing simulated mud weight with mud weight used in drilling, it is found that a
minimum horizontal stress gradient of 0.0185 MPa/m (a mean value of minimum horizontal stress gradients
determined from stress polygon) is mostly applicable to Telwa, Ardol, Kanwa and Hazad members and a
maximum horizontal stress gradient of 0.0234 MPa/m (the lower limit of maximum horizontal stress
gradients determined from stress polygon) is mostly applicable to Kanwa and Hazad members, while a
maximum horizontal stress gradient of 0.0220 MPa/m is mostly applicable to Telwa and Ardol members.
The simulated safe mud weight window as well as the mud density used in drilling of Well D are shown in
Figure 5 for comparison. It can be seen that the lower bound of the simulated safe mud weight matches well
to the mud density used in drilling.
Rock properties
Rock mechanical properties are determined from rock mechanics tests and log interpretations. Multivariate
regression technique was used to model relationships among the shear wave slowness and a series of other
log responses, such as compressional wave slowness, caliper, bulk density and neutron porosity.
Geomechanical parameters, such as Poisson’s ratio, shear modulus, bulk modulus, Young’s modulus,
unconfined compressive strength (UCS) and tensile strength, cohesive strength, internal friction angle were
estimated from well logs. All the log-derived geomechanical properties were calibrated with data obtained
from tri-axial rock mechanical tests. Figure 6 shows the rock mechanical property profile for Well D.
Page 2 of 7
PETROTECH – 2009
11 - 15 JANUARY 2009, NEW DELHI, INDIA
that the pore pressure predicted by the model is much higher than the experimental data. Possible
explanation for this discrepancy is that Darcy’s law may not be applicable to extremely low permeability
shale. However, by introducing a calibration factor 1/3 to the absolute permeability (k0), it was possible to
obtain a reasonable match.
Chemical potential test. Shale exhibits a non-ideal semi-permeable (“leaky”) membrane in response to
water-based solutions because shale has a range of pore sizes including wide pore throats, which result in
significant permeability to solutes. The efficiency of the process of osmosis of a leaky membrane is normally
termed as reflection coefficient, Θ, which is generally defined by the following equation when there is no net
fluid flux through the membrane:
Θ=
∆P with ∆ π = RT ln ⎛⎜ a 1 ⎞⎟ (2)
∆π ⎜a ⎟
V w ⎝ 2 ⎠
where ∆P is the actually observed, osmotically induced hydrostatic pressure difference across the
membrane, ∆π is the theoretical chemical potential difference of the two solutions separated by the
membrane, R is gas constant, T is temperature in °K, Vw is the mean partial molar volume of the water on
either side of the membrane, and a1 and a2 are the activities of water in the two solutions. The constants
used in this study are as follows: R = 0.083 L-atm/mol-K, Vw = 0.018 L/mol. In this study, the osmotically
induced hydrostatic pressure was experimentally measured from chemical potential test, and hence the
reflection coefficient can be estimated from Eq. 2. Three tests were conducted in a tri-axial chemical potential
cell at a temperature of 55°C. Test results are shown in Table 1.
Wellbore stability analysis for the determination of safe mud weight windows
Wellbore stability analysis primarily involves determination of the principal stresses around a borehole and
applying failure criterion to assess likelihood of borehole collapse or fracture for the given the rock and mud
properties. This analysis ultimately suggests a range of permissible mud weights required to maintain a
stable borehole. Based on the wellbore stability theories, a wellbore stability analysis program “Boremap”
developed at the School of Petroleum Engineering, UNSW was used in this study for wellbore stability
analysis.
Wellbore stability analysis was conducted for Wells A, B and C using the calibrated stress profile. The safe
mud weight windows as well as actual mud weight used in drilling these wells are presented in Figure 7. The
results of Well A shows that the lower bound of the simulated safe mud weight matches well to the mud
density used in drilling in Ardol, Kanwa and Hazad intervals and in most of Telwa interval, except in the
interval of 2726 – 2772 m (Telwa) where the simulated mud weight is slightly higher than the mud density
used in drilling. Reviewing the logs in this interval, it is found that the compressional wave slowness is
exceptionally higher in this interval than in other intervals. Higher compressional wave slowness results in
lower estimated UCS and lower cohesive strength, eventually the simulated safe mud weight will be higher.
In case of Well B, it can be seen that the lower bound of the simulated safe mud weight matches well to the
mud density used in drilling in Telwa, Ardol and Kanwa intervals. In Hazad interval (below 2471 m), however,
the mud density used in drilling is higher than the simulated lower bound of safe mud weight. As for Well C,
the mud density used in drilling is quite high comparing to the simulated safe mud weight in the interval
between 2900.78 m and 3229.81 m.
Effect of well trajectory on wellbore stability – Contour plot of safe mud density
Contour plot of safe mud weight is an effective way to present graphically safe mud weight windows for
different wellbore trajectories. The plot is made by calculating mud weights for any possible wellbore
trajectory (combination of inclination and azimuth) at a given depth and then plotting the mud weight
contours in polar coordinates, where the r and θ axes represent wellbore inclination (α) and wellbore azimuth
(β), respectively.
The effect of wellbore trajectory on wellbore stability or safe mud weight depends on the in-situ stress
regime. As determined from stress polygon, the Hazad member in Dahej field is most likely to be in a strike-
slip stress regime. To illustrate the effect of wellbore trajectory on the safe mud weight window, wellbore
stability analyses for different wellbore trajectories (inclination and azimuth) are conducted at a depth of
2966.01 m from Well D. From log interpretation and core/cuttings description presented in the well
completion report, the rock type at 2966.01 m is shale. The lower and upper bounds of safe mud weight
windows are presented in Figure 8 in the form of polar contour plot. It can be seen that drilling a vertical well
requires higher mud density (1.244 g/cm3) to prevent wellbore collapse compare to horizontal wells along the
directions of minimum horizontal stress (1.194 g/cm3) and maximum horizontal stress (1.207 g/cm3). While
Page 3 of 7
PETROTECH – 2009
11 - 15 JANUARY 2009, NEW DELHI, INDIA
lower mud density (2.276 g/cm3) is required to prevent wellbore tensile failure (fracture) for vertical well
compare to horizontal wells along the directions of minimum horizontal stress (3.328 g/cm3) and maximum
horizontal stress (2.362 g/cm3). The optimum wellbore direction for a horizontal well at this depth (2966.01
m) is along 45 degree referring to the direction of minimum horizontal stress, i.e along the direction of
N28.2E since the direction of minimum horizontal stress is N16.8W. The safe mud weight window is 1.176
g/cm3 for the lower bound and 2.791 g/cm3 for the upper bound.
The following three cases are studied for the time dependant wellbore stability analysis: Case 1. WBM with
mud pressure penetration without osmosis effect; Case 2 WBM with both mud pressure penetration and
osmosis effect and Case 3 Fresh water with both mud pressure penetration and osmosis effect. The upper
bound and lower bound of safe mud weight window at different time for the three cases are presented in
Figure 9. It can be seen from the figure that mud pressure penetration induced pore pressure increase
requires higher mud density to prevent wellbore collapse and lower mud density to prevent wellbore fracture.
Hence, the safe mud weight window becomes narrower with time. WBM with lower water activity than pore
fluid can compensate for the effect of mud pressure penetration on safe mud weight window. Since fresh
water has a higher water activity than pore fluid, the safe mud weight window becomes further narrower with
time than in the case of WBM. Considering mud pressure penetration when drilling with WBM, a mud
density of 1.35 g/cm3 will be required to keep the wellbore stable.
Conclusions
The wellbore stability study conducted in Dahej field provides an improved understanding on behavior of
shales in southern Cambay Basin. Results of this study show that the borehole instability problem can be
significantly reduced if mud system can be tailored by incorporating both physical and chemical effects.
Acknowledgements
The authors wish to acknowledge the support provided by Institute of Drilling Technology of Oil and Natural
Gas Corporation (ONGC), India, in particular the ONGC officers who worked on the project: Dr. O.P.
Sharma, Dr. Ajay Kumar, Dr. H.C. Tewari and Mr. S.K. Gauba. Also we wish to acknowledge Dr. Liuqi Wang
for work on analysis and interpretation of well logs.
References:
1. Aadnoy, B.S., and Chenevert, M.E., 1987, "Stability of highly inclined boreholes," SPE Drilling Engg.
Dec., pp.364-374.
2. Chang, Chandong, 2004,“Empirical Rock Strength Logging in Boreholes Penetrating Sedimentary
Formations,” Geophysical Exploration, Vol 7, No 3, pp. 174~183.
3. Gardner, G. H. P., Gardner, L. W., Gregory, A. R., 1974, Formation Velocity and Density – The
Page 4 of 7
PETROTECH – 2009
11 - 15 JANUARY 2009, NEW DELHI, INDIA
Figure 1 Pore pressure profile of Well B. Figure 2 Bulk density derived from VSP and density log
trend line for Well A.
Pressure (MPa)
0 10 20 30
0
1000
1500
Pore
2000 Pressure
Hydrostatic
2500
Pressure
Figure 3 Rose diagram of breakout (Well D). Figure 4 Stress polygon (Well D @ 2991.8 m).
250
σ h=σ v fracture constrain
200
RF σ H=σ h
150
SS
σ H =70 - 87 MPa
100 σH =σv
50
breakout constrain
NF σ h=50 - 64 MPa
0
0 100 200 300
Minimum horizontal stress (MPa)
Page 5 of 7
PETROTECH – 2009
11 - 15 JANUARY 2009, NEW DELHI, INDIA
Depth, m
2650
2750
2850
2950
3050
3150
Maximum horizontal stress
Minimum horizontal stress Poisson's ratio
Vertical stress
Pore pressure
Cohesive strength
Tensile strength Biot coefficient
Internal friction angle
Figure 5 Comparison of safe mud weight Figure 6 Rock property profile for Well D
with mud weight used in drilling of Well D.
Page 6 of 7
PETROTECH – 2009
11 - 15 JANUARY 2009, NEW DELHI, INDIA
Figure 8 Lower and upper bounds of safe mud weight window for Well D @ 2966.01 m.
Figure 9 Lower and upper bounds of safe mud weight window for Well D @ 2966.01 m. Case 1 – WBM
with mud pressure penetration only; Case 2 – WBM with both mud pressure penetration and osmosis
effects and Case 3 – Fresh water with both mud pressure penetration and osmosis effects.
Page 7 of 7