WELLBORE STABILITY AND SAND CONTROL
Assessing the Impact of Trajectory on Wells
Drilled in an Overthrust Region
N.C. Last, SPE, and M.R. McLean, SPE, BP Exploration
Summary
Drilling in the Cusiana field, located in the tectonically active foot
hills of the Casanare region of Colombia, has proved to be extremely
difficult. One major contributor to the operational difficulties is
poor hole conditions, which often lead to stuck pipe. However, ex
perience has shown that performance improves when drilling updip
of the major faults and bedding; downdip and crossdip well trajecto
ries are the most problematic. Stress modeling of the geological
cross sections indicates that the principal stresses in the Cusiana
field may be rotated significantly from the vertical and horizontal.
After adjusting a conventional wellborestability analysis to allow
for stress rotation, a reasonable match is obtained between predic
tion and field experience on the issue of stability variation with hole
trajectory for wells drilled in the region.
Introduction
In partnership with Ecopetrol, Total, and Triton, British Petroleum
(BP) has been drilling in the foothills of the Casanare region of
Colombia (Fig. 1) since the late 1980's. Both the Cusiana and Cu
piagua fields have now been declared commercial. Development
well drilling in Cusiana began in late 1993. Appraisal wells are cur
rently being drilled in Cupiagua. An aggressive drilling schedule is
now being pursued, with 10 rigs active. From the earliest well, op
erational difficulties were encountered. Possibly the biggest prob
lem associated with drilling wells in this region has been well bore
instability. Ref. I discusses the considerable effort that has gone into
tackling the problems and the improvements that have been made.
Here, the particular aspect of instability variation with well trajecto
ry is discussed.
Background to Wellbore Stability
Wellbore instability (Fig. 2) can result in lost circulation where ten
sile failure has occurred and in spalling and/or hole closure in the
case of compressive failure of the rock. In severe cases, hole insta
bility can lead to stuck pipe and eventual loss of the open hole sec
tion. Owing to the large costs associated with well bore instability,
it has received considerable attention, as is apparent from the num
ber of publications on the subject. The extent of the problem is real
ized to a greater extent in the drilling experiences encountered in the
foothills of the Casanare region of Colombia than elsewhere. For ex
ample, Fig. 3 shows caliper traces from a Cusiana development
well. Fig. 3a shows a hole drilled with a l2V,in. bit that has enlarged
to nearly 45 in. in places. The ingauge section is a sand interval.
Also, as Fig. 3b shows, the large overgauge reading in one caliper
pair accompanied by a relatively ingauge reading of the other pair
suggests that the stresses may be highly anisotropic.
Selection of mud types and weights to eliminate hole instability of
wells drilled in Cusiana has been found to be impossible. Loss zones,
which necessitate a low mud weight, frequently occur adjacent to col
lapse zones, which require a high mud weight. Incorporating suffi
cient intermediate casing strings to isolate each individual problem
zone is unrealistic. Hence, mudweight selections are made in an at
tempt to minimize the problems without eliminating them, and it has
been necessary to live with a certain degree of hole instability.1
Copyright t 995 Society of Petroleum Engineers
Original SPE manuscript received for review March 15. 1996. Revised manuscript received
April 2. 1996. Paper peer approved April 19. 1996. Paper (SPE 30465) first presented at the
1995 SPE Annual Technical Conference and Exhibition. Dattas, Oct. 2225.
620
Well borestability theory2.3 shows that many factors other than
mud weight are likely to influence the competency of the well bore
wall. Examples are formation pressure, depth, formation strength,
stress state, mudcake efficiency, and time. Of the parameters that
can be controlled by the operator, theory shows that wellbore trajec
tory has a significant influence on stability.2,4 Field experience from
wells in the Cusiana field also indicates that trajectory has a signifi
cant influence on hole stability. In cases where optimizing mud
weights alone cannot reduce instability to acceptable levels, then
optimizing well trajectory to reduce further, or at least limit, the ex
tent of problems is necessary.
Drilling Experiences in Cusiana
Fig. 4 is a schematic of the geological structure. The figure shows the
two main faults, Yopal and Cusiana, their dip and dip direction, and the
dipping nature of the formations. It also shows example wells that de
fine the meaning of updip, downdip, and crossdip trajectories.
Drilling experience in Cusiana showed that wells drilled in the
general updip direction appear to progress more smoothly than ver
tical wells and are a significant improvement on wells drilled in the
downdip direction, which have proved to be very difficult. Evidence
also exists that crossdip wells are more difficult than nominally ver
tical wells. The incidence of stuck pipe and the need to sidetrack also
increase significantly for downdip and crossdip wells.
Hole condition is one aspect that updip drilling seems to improve.
Assuming cumulative reaming time is a reasonable indication of the
extent of hole instability, then Fig. 5 can be used to assess the varia
tion of stability with well trajectory. Fig. 5 plots a number of wells,
showing their stepout and azimuth. The bars against each well show
the number of reaming hours for drilling through the most difficult se
quence of formations, known as the Carbonera (Fig. 6). In general,
wells in the southeast sector denote updip trajectories (Fig. 4), wells
in the northwest sector downdip trajectories, and wells in the north
east and southwest sectors crossdip trajectories. Wells with the larger
stepouts in the southeast sector (typically updip wells) require fewer
reaming hours than near vertical wells and considerably fewer ream
ing hours than wells of similar stepouts in the other three sectors.
Hole breakout and enlargement are also reduced in updip wells, as
measured by the volume of cavings produced at surface and the ap
parent rugosity of the hole (reduced tripping problems).
Hole condition is not the only issue to consider. For example, the
natural tendency for the bit to attack the formations normal to their
bedding (unless deliberately steered) must be a factor. However, at
least part of the performance improvement in drilling updip is the
result of reduced hole instability. To project the behavior of future
wells planned in the region and to plan trajectories to minimize hole
instability problems require a consistent explanation of the experi
ences encountered to date. A credible wellborestability model
needs to explain the directionally dependent nature of the holesta
bility problems encountered.
Directional Dependence of Hole Stability
For stability to be directionally dependent requires anisotropy in
loading and/or relevant formation properties. The two factors most
likely to contribute to directional dependency of wellbore stability
are anisotropy of formation strength and insitu stresses.
Effect of Strength Anisotropy. Rock can be anisotropic in its
strength behavior because of sedimentary features, such as bedding
planes and beddingparallel laminations, or because of deforma
July 1996 . JlFf
Fig. 1Location map.
o 4 8
...............
Miles
tioninduced features, such as fractures, joints, or cleavage
associated with shear zones adjacent to a fault. Strong evidence ex
ists for localized occurrences of all of these in the Cusiana overbur
den, but bedding planes are the most extensive and obvious contrib
utor to strength anistropy. A number of researchers have studied the
effect of rockstrength anisotropy on wellbore stability.5.6 Similar
approaches to those published are currently under investigation as
a means of improving the backanalysis of past wells and forward
prediction for planned wells in the Cusiana field. However, the in
depth studies of the stress state alone and its influence on stability
provide valuable insight and form the central theme of this paper.
Effect of Stress State. Formations are subjected to stresses derived
from the weight of the overburden and any regional tectonic loading
that exists. The stresses can be resolved into three principal compo
nents, which in most cases are the vertical, minimum horizontal, and
maximum horizontal stresses (Fig. 7). In passive sedimentary ba
sins, such as the North Sea, the vertical stress is usually considered
to be the largest (Fig. 7a). In a geologically active setting, such as
Cusiana, the tectonic forces dominate and the maximum horizontal
stress is the largest of the three components (Fig. 7b). Knowledge
of the stress state is important when diagnosing hole problems and
planning development wells. In a passive basin, a vertical well is
parallel to the maximum stress component, which theory shows to
12,350
15n
Gamma Ray
45" 121/4" 121/4" 45"
Caliper 13 Caliper 24
:.:
Lost
{
Circulation
:::I!
1
LL
Hole
Closure
Fig. 2Types of stressinduced instability.
be one of the most stable trajectories.
4
.
5
.
7
In general, field evidence
from the North Sea also shows that vertical or lowangle wells are
the least problematic and that instability is more associated with
moderate to highangle wells.
8

10
In Cusiana, which is in an active overthrust region, the maximum
horizontal stress is greater than the overburden stress; hence, the
faulting seen, the active tectonism, and the evolution of the Andes
mountains. For the overthrust stress state depicted in Fig. 7b, stan
dard wellborestability theory predicts that vertical wells are the
most prone to wellbore instability, while highangle wells with azi
muths close to the maximum horizontal stress direction are the least
prone to instability. These theoretical results are reported in a num
ber of publications.
4
, 11 However, little or no field evidence has been
published to support or refute the theoretically derived conclusions
for wells drilled in this type of stress environment. In Cusiana (in a
tectonically active region), where the conventional opinion is that
the maximum stress is horizontal, a vertical well is predicted to be
the least stable and highangle wells drilled subparallel to the direc
tion of the maximum horizontal stress are predicted to be the most
stable. Field experience from Cusiana shows that trajectory can
Depth
=""".. (It mO)
150
Gamma Ray
r=
'
);.
Th.
jJ
45" 12114" 121/4" 45"
Caliper 13 Caliper 24
Fig. 3Caliper traces from 12
1
/
4
in. section of a Cusiana well.
JlYl' July 1996 621
Fig. 4Schematic of structural geology and inferred stress
directions in the Cusiana field.
have a strong influence. However, the trends seen are not consistent
with the standard theory. This apparent mismatch and its resolution
are discussed in the following sections.
Stress Distribution in the Cusiana field
The interpretation of wellbore breakout from caliper logs reveals a
very consistent northwestsoutheast direction for the maximum hor
izontal stress across the Cusiana field,
1
which is perpendicular to the
mountains and in the regional thrust direction (Fig. 4). The degree
of breakout (e.g., Fig. 3b) suggests a significant difference in the two
horizontal stresses. On the basis of leakoff tests, mud losses, and
steprate tests, the minimum horizontal stress appears to be surpris
ingly low (equivalent to 0.7 psi/ft) and therefore horizontal. Hence,
the vertical stress must be intermediate. However, there is evidence
of significant variability in the stress magnitudes with depth and
formation type, which is to be expected.
The expense involved in conducting sufficient field measurements
to define the stress variations reliably throughout the structure would
be prohibitive. Therefore, a predictive stress analysis model has been
used to get trends in the variation of stress with depth and with posi
tion on the structure. In the model, horizontal loading is applied to a
geometric representation of the structure, which includes lithology
and major faults. The model predicts the spatial variation of stresses,
including both magnitude and directions. We plan to publish details
of the stressmodeling process at a later time. * To date, the results
have been used to determine the stress variation along predicted well
paths for wellborestability calculations, leading to mudweight rec
ommendations.
1
However, the predictive modeling has also led to the
observation of another potentially important aspect of the stress field
in Cusiana: the rotation of the principal stresses away from the com
monly assumed vertical and horizontal axes.
Fig. 6 shows the results from a typical simulation. This figure is
a simplification of the stresstrajectory plot given in Ref. 12. In Fig.
6, each cross represents the magnitude and orientation of the princi
pal components of stress. The subhorizontal stress is the largest, re
flecting the effects of the regional tectonics. The minimum horizon
tal stress is normal to the plane of the diagram and is determined by
the model on the basis of planestrain conditions. The simulation re
sult of interest here is that the principal stresses are not uniformly
vertical and horizontal, as is generally assumed. The assumption of
vertical and horizontal principal stresses is reasonable in a normal
sedimentary basin. However, the combined effects of the loading
and structure (faults and dips) produce shear stresses in the overbur
den that result in a rotation of the principal stresses away from the
vertical and horizontal.
Evaluation of Relative Stability of Directional Wells
Model Description. Standard wellborestability analysis tools can
be used to evaluate the effect of well trajectory on hole stability. The
equations presented in the Appendix for determining the elastic
= 4
o
3
= 0
1
2
o
U)
2023
I
...
'  ,
'"
I .  
 fUJ '
,/
hours
"
I
5 4 3 2
West (1 000 It)
,!J
,
111
900
.
,
,
,

]if
112
...... 4} fr',  u
" '
h.?
' ':J1
245
IBoltor \ hoI.
location
I I
2 3
East (1000 It)
Fig. 5Reaming time as a function of trajectory for wells in the
Cusiana and Cupiagua fields.
stress state around the wellbore are similar to those presented by oth
er authors, 2,13 except that rotation of the principal stresses from the
commonly assumed vertical and horizontal axes is incorporated. To
assess well bore stability, the computed stress state is compared with
a rockstrength criterion. A number of criteria have been proposed
for elasticlbrittle failure analyses of well bore stability. Here a Mohr
Coulomb criterion is assumed, Refs. 14 and 15 discuss the effect of
the choice of criterion on computed results.
Output from wellborestability models is typically presented in
terms of minimum mud weights required to prevent collapse and
maximum mud weights required to avoid exceeding the fracture gra
dient. These allowable mud weights can be presented as functions of
well deviation and/or azimuth.
4
,16,17 Our experience is that mud
weight predictions from elasticlbrittle failure models can be overly
conservative
14
and may overstate the sensitivity of hole stability to
well trajectory. Published data from other sources support our experi
ences,1820 although the view that an elasticlbrittle failure model is
too conservative is not unanimously supported. Nevertheless, we be
lieve that presenting model predictions in terms of mudweight rec
ommendations without carefully "calibrating" the results could be
misleading and potentially detrimental to the credibility of well bore
stability studies. Therefore, to avoid these problems the "stability
charts" presented here give only a qualitative guide to the stability of
a well as a function of its trajectory.
To illustrate the model output, Fig. 8 shows the results from an
analysis of the Ula field (Block 7112 in the Norwegian sector of the
North Sea). Ula is in the more conventional passive basin with nom
inally horizontally bedded formations and therefore it is reasonable
o 5,000 It 10,000 II 15,000 II 20,000 It 25,000 It
sea 0
o sea
5,000 It
2,000 m
15,000 It
20,000 It
2,000 m 4.000 m
'Last, N.C. and Harkness, M.R.: "Stress Distribution in Cusiana OverburdenEvaluation by
Modeling and Field Calibration," unpublished work (1995). Fig. 6Computed stress vectors in Cusiana field,
622 July 1996 .JIyr
(a) (b)
St ress state typ ical of a passive basi n. Stress state typical of a compressive
tectonically active region.
Fig. 7(a) Passive and (b) active stress states.
to assume vertical and horizontal insitu principal stresses. Breakout
analysis
21
of wells in the VIa field show a fairly consistent north
westsoutheast trend for the maximum horizontal stress. The direc
tion of the maximum horizontal stress entered into the model is indi
cated by the arrows on the boundary of the polar diagram (Fig. 8).
The stress state in shale formations susceptible to instability is esti
mated to be defined by the ratio of 1.0: 0.9: 0.8, respectively, for ver
tical stress, 0v; maximum horizontal stress, 0Hmax; and minimum
horizontal stress, OHmin' The polar diagram in the figure shows the
relative stability of wells as a function of well trajectory at 8,000 ft
total vertical depth (TVD). The center of the diagram represents a
vertical well. The concentric circles represent increasing inclina
tion, with the outer boundary representing horizontal wells. The ra
dial lines represent different azimuths. The darker the shading in a
sector, the greater the risk or extent of instability of wells drilled
with that trajectory. Conversely, the lighter the shading the lower the
risk of instability. The "stability numbers" against the shading chart
are a measure of stability relative to a vertical well. Thus, a number
N
More Stable
1.05
1.03
0.68
Less Stable
cr
Hmax
Fig. 8Hole stability vs. well trajectory, Ula field, Norwegian sec
tor of the North Sea. Depth = 8,000 ft, Ov equivalent to 1.00 psilft,
0Hmax equivalent to 0.90 psi/ft, 0Hmin equivalent to 0.80 psi/ft,
pore pressure gradient = 11.5 Ibm/gal, mud weight = 12.5 Ibm/gal,
and calculated UCS=4,191 psi.
,JlYI' July 1996
less than unity is less stable than a vertical well, and a number great
er than unity is more stable than a vertical well.
The stability of a vertical well is set to unity by determining a uni
axial compressive strength (VCS) for the formation so that a vertical
well is predicted to be on the limit between being stable and suffer
ing hole collapse, assuming an elasticlbrittle failure model. A mud
weight must be selected to enable a VCS to be backed out of the
analysis. The VCS determined from this calculation is unlikely to
match the laboratory VCS, if measured; in general, the computed
VCS will be significantly greater than the laboratorydetermined
VCS. Typically, the selected mud weight will be the mud weight
considered to be optimum for the vertical hole section in question.
In extreme cases, such as Cusiana, the optimum mud weight may
not be sufficient to prevent all occurrences of hole collapse. There
fore, a stability factor of one or greater does not necessarily mean
that there will be no spalling of the borehole wall. Conversely, in rel
atively easy drilling environments, a stability factor of less than one
does not necessarily mean that there will be hole instability. The ob
ject of the plots is simply to allow a qualitative assessment of how
the risk of instability varies with well trajectory.
The VCS backed out from the chosen stress conditions and opti
mum mud weight for a vertical well are then used to determine the
stability factor for all other well trajectories, assuming the mud
weight is kept constant. The stability factor is based on a compres
sivefailure criterion and is defined as allowable maximum princi
pal effective stress at the wellbore wall divided by actual maximum
principal effective stress. We can show that for typical overbal
ances, the stability factor is, in general, given by
Co + (PH  pp)(l + sin rp)/(l  sin rp)
F, = 0 _ P ......... (1)
I P
From Fig. 8, we can see that for VIa, northeastsouthwest direc
tional wells should be less prone to instability than northwestsouth
east wells. These theoretical conclusions are observed in practice in
the VIa field. Wells with trajectories in the least favorable direction
are planned with higher mud weights as a consequence of instability
problems seen early in the developmentdrilling program.
Model Application to Cusiana. Compared with the VIa field, the Cu
siana stresses arc different in three important ways: (1) the vertical
stress is not the maximum stress, (2) the ratio of the stresses is much
greater, and (3) the principal stresses are rotated from the vertical and
N
More Stable
1.62
1.58
1.54
1.50
1.46
1.42
1.38
1.34
1.30
1.26
1.22
1.18
1.14
1.10
1.06
1.02
Less Stable
Fig. 9Hole stability vs. well trajectory, Cusiana field, vertical and
horizontal principal insitu stresses. Depth = 10,000 ft, Ov equiva
lent to 1.1 0 psi/ft, 0Hmax equivalent to 1.40 psi/ft, 0Hmin equivalent
to 0.70 psilft, pore pressure gradient=8.5 Ibm/gal, mud
weight=12.0 Ibm/gal, and calculated UCS=19,856 psi.
623
N
More Stable
1.50
1.47
1.43
1.40
1.37
1.33
1.30
1.26
1.22
1.17
1.14
1.10
1.07
1.04
1.00
0.97
Less Stable
()
Hmax
Fig. 1 ~ H o l e stability vs. well trajectory, Cusiana field, principal
insitu stresses rotated 30 from vertical and horizontal axes.
Depth = 10,000 ft, 0v equivalent to 1.18 psi/ft, 0Hmax equivalent to
1.33 psi/ft, 0Hmin equivalent to 0.70 psilft, TVHmax equivalent to 0.17
psilft, porepressure gradient =8.5 Ibm/gal, mud weight = 12.0 Ibm/
gal, and calculated UCS=18,355 psi.
horizontal axes (the minimum principal stress is still horizontal). The
vast majority of well bore instability is seen in the Carbonera sequence
(see Fig. 6). Experience to date suggests that the three insitu principal
stresses in the Carbonera are in a ratio of 1.4: 1.1 : 0.7, although this
varies with depth and position on structure.
The relative stability of wells drilled through the Carbonera is first
investigated assuming the principal stresses are vertical and horizon
tal. as per standard approaches to wellborestability analysis. Hence,
0Hmax: Ov :OHmin = 1.4: 1.1 :0.7. Fig. 9 shows the predicted effect of
well trajectory at a depth of 10,000 ftTVD [typically 10werCarbonera
above the Yopal fault (Fig. 6)). The predictions are quite different
from the Ula predictions presented in Fig. 8. For Cusiana, hole condi
tions in wells drilled parallel to the mountains are predicted to be no
worse than nominally vertical wells, and conditions should, in fact,
improve slightly at higher deviation angles. The most stable orienta
tions are predicted to be in the updip and downdip directions (see Fig.
4); indeed, highangle wells with these azimuths should be signifi
cantly more stable than a vertical well. These conclusions are a direct
consequence of the predicted stress ordering and the assumption that
principal stresses are aligned vertical and horizontal.
Experience has shown that while the relative performance gener
ally improves when drilling updip, drilling downdip results in sig
nificantly worse hole conditions. Drilling crossdip has also proved
to result in worse conditions than nearvertical wells. Hence, the
predicted effect of trajectory on conditions (Fig. 9) shows little con
sistency with the actual drilling experiences.
The effect of stress rotation is now incorporated into the analysis.
The previously defined stress state is rotated so that the vertical and
maximum horizontal stresses are rotated 30
0
clockwise as one looks
at the northwestsoutheast section in Fig. 6. From Fig. 6, we can see that
significant rotations of the principal stresses are indicated and that a 30
0
rotation is not unrealistic within the Carbonera above the Yopal fault.
(Below the Yopal fault the stresses in the "repeat" Carbonera sequence
do appear to be more closely aligned to the vertical and horizontal and
the conventional approach may be reasonable here.)
Rotation of the vertical and maximum horizontal stresses is
achieved by applying a shear stress, TVHmax, (see the Appendix).
The shear stress component required to induce a clockwise rotation
of lJ1 is given by
For the stress rotation example, the stresses 0Hmax, 0v, and 0Hmin
are entered as 1.33, 1.18, and 0.7 psi/ft, respectively, and the shear
stress, TVHmax, entered as 0.13 psi/ft. From Eq. 2 we can see that this
provides a 30
0
clockwise rotation in the principal stresses. The values
of 0Hmax and 0v are now different from those used in the nonrotated
stress case (Fig. 9). This is because 0Hmax and 0v are no longer princi
pal stresses. However, we can show that the three principal compo
nents of the insitu stress state are still in the ratio of 1.4: 1.1 :0.7.
Fig. 10 shows the relative stability of wells as predicted, with
stress rotation incorporated into the analysis. The variation of stabil
ity with trajectory is now significantly different from that shown in
Fig. 9. Now updip (southwest) wells are predicted to be consider
ably more stable than downdip (northeast) wells, as observed in the
Cusiana field. The most stable wells are bracketed by azimuths of
100 to 170
0
and deviations between 40 and 70
0
These predictions
are consistent with the field observations, as characterized by the
variation in reaming hours with well trajectory (Fig. 5).
Conclusions
I. Field evidence has shown that drilling performance is im
proved relative to vertical wells when drilling updip of the major
faults and bedding in the Cusiana field.
2. Drilling downdip and crossdip has resulted in greater problems
than drilling of vertical wells in Cusiana.
3. Conventional wellborestability analyses assume insitu prin
cipal stresses are horizontal and vertical. The application of a con
ventional analysis to the Cusiana field predicts that highangle wells
drilled toward or away from the mountains are the most stable and
vertical wells the least stable.
4. Conventional wellborestability analysis results are inconsis
tent with the field experiences from Cusiana.
5. Stress modeling of the structural setting for the Cusiana field
reveals that the normal assumption of the principal stresses being
aligned with the horizontal and vertical axes may not be valid. In the
vicinity of the thrust faults, significant rotation is predicted.
6. When stress rotation is incorporated into the wellborestability
model, updip drilling is the most favorable direction.
7. The predicted trends from the modified model are in general
agreement with experiences seen to date in Cusiana.
8. Other factors may contribute to hole condition for directional
wells drilled in thrust regions, such as the angle between the well
and any planes of weakness that may exist (e.g., bedding). This is
currently under investigation. JPT
Nomenclature
Co = formation UCS, mlLt
2
, psi
F" = stability factor
Pp = formation pore pressure, mlLt
2
, psi
PI\' = wellbore pressure from mud column, mILt
2
, psi
a = wellbore inclination, degrees
f3 = wellbore azimuth relative to maximum horizontal
insitu stress, degrees
() = angular location around wellbore wall, degrees
0Hmax = maximum horizontal insitu stress, mlLt2, psi
0Hmin = minimum horizontal insitu stress, mILt
2
, psi
Or ,Go'Oz = radial, circumferential, and axial stresses at well
bore wall, mlLt
2
, psi
0v = vertical/overburden stress, mlLt2, psi
Ox,Oy,Ozz = normal stresses from transpose of insitu stresses
to wellborecoordinate system, m/Lt2, psi
01 = maximum principal stress at the wellbore wall (see
Appendix), mlLt
2
, psi
TVHmax = insitu shear stress (resulting in rotation of insitu
principal stresses), mlLt2, psi
T
xy
, T
yZ
' T;:x = shear stresses from transpose of in situ stresses to
wellborecoordinate system, mlLt2, psi
To
z
= shear stress at wellbore wall, mlLt2
1> = MohrCoulomb friction angle, degrees
lJ1 = rotation of principal insitu stresses from vertical
and horizontal axes, degrees
T VHmax = V2(0 Hmax  v) tan 2lJ1. . ................... (2) 'Compression is assumed positive throughout.
624
July 1996 .JIYI'
Acknowledgments
We thank the management of BP Exploration and partners, Ecopetrol,
Total, and Triton, for permission and encouragement to publish this pa
per and the many colleagues who participated in the studies described,
in particular, Patrick Collins, consultant; Richard Harkness of South
ampton U.: and Herve De Naurois of Total.
References
1. Last, N.e. et al.: "An Integrated Approach to Evaluating and Managing
Well bore Instability in the Cusiana Field, Colombia, South America,"
paper SPE 30464 presented at the 1995 SPE Annual Technical Confer
ence and Exhibition, Dallas, Oct. 2225.
2. Bradley, W.B.: "Mathematical Stress Cloud Can Predict Borehole Fail
ure," Oil & Gas 1. (Feb. ) 77, No.8, 92.
3. McLean, M.R.: "Analysis of Well bore Stability," PhD dissertation, U. of
London (1988) 43119.
4. Haiso, C.: "A Study of HorizontalWellbore Failure," SPEPE (Nov.
1988) 489; Trans., AI ME, 285.
5. Aadnoy, B.S. and Chenevert, M.E.: "Stability of Highly Inclined Bore
holes," SPEDE (Dec. 1987) 364; Trans., AIME, 281.
6. Ong, S.H. and Roegicrs, Je.: "Horizontal Well Collapse in an Aniso
tropic Formation," paper SPE 25504 presented at the 1993 SPE Produc
tion Operations Symposium, Oklahoma City, March 2123.
7. Charlez, P. and Heugas, 0.: "Evaluation of Mud Weight in Soft Shale
Levels," Rock Mechanics as a Multidisciplinary Science, Je. Roegiers
(cd.), A.A. Balkema, Rotterdam (1991) 100514.
8. Wilson. R.e. and Willis, D.N.: "Successful HighAngle Drilling in the
Statfjord Field," paper SPE 15465 presented at the 1986 SPE Annual
Technical Conference and Exhibition, New Orleans, Oct. 58.
9. Wong, SW. et al.: "Optimizing Shale Drilling in the Northern North Sea:
Borehole Stability Considerations." paper SPE 26736 presented at the
1993 Offshore Europe Conference, Aberdeen, Sept. 69.
10. Ewy, R.T. et al.: "North Sea Case Histories of Wellbore Stability Predic
tions for Successful HighAngle Nelson Field Wells," paper SPE 27495
presented at the I 994lADC/SPE Drilling Conference, Dallas, Feb. 15I 8.
I I. da Fontoura, S.A.B. and Santos, H.M.R: "Stability Studies for the Exper
imental Horizontal Well at Fazenda BeIem (e.E.), Brazil." paper SPE
2 I 064 presented at the 1990 SPE Latin American Petroleum Engineering
Conference, Rio de Janeiro, Oct. 1014.
12. Last, N.e. and McLean, M.R.: "Assessing the Impact of Trajectory on
Wells Drilled in an Overthrust Region." paper SPE 30465 presented at
the 1995 SPE Annual Technical Conference and Exhibition. Dallas. Oct.
2225.
Hmax
H m i ~ \
V ;.,..:....:,.......,.,
~
(xaxis lies in the Z
horizontal plane)
Fig. A1Transposing insitu stresses.
JlYI' July 1996
13. Leeman, E.R.: "The Determination of the Complete State of Stress in
Rock in a Single BoreholeLaboratory and Underground Measure
ments," 11111. 1. Rock Mech. Min. Sci. Geol11ech. Absli: (1968) 5, 31.
14. McLean, M.R. and Addis. M.A.: "Well bore Stability Analysis: A Re
view of Current Methods of Analysis and Their Field Application." pa
per SPE 19941 presented at the 1990 IADC/SPE Drilling Conference.
Houston, Feb. 27March 2.
15. McLean, M.R. and Addis, M.A.: "Wellbore Stability: The Effect of
Strength Criteria on Mud Weight Recommendations," paper SPE 20405
presented at the 1990 SPE Annual Technical Conference and Exhibition
New Orleans, Sept. 2326. '
16. Guild. GJ., Jeffery, J.T., and Carter, J.A: "Drilling ExtendedReach!
HighAngle Wells Through Overpressured Shale Formation in the Cen
tral Graben Basin, Arbroath Field. Block 22/17, U.K. North Sea,"
SPEDC (Sept. 1994) 161; Trans . AIME. 297.
17. Ottensen, S. and Kwakwa, K.A.: "A Multidisciplinary Approach To In
situ Stress Determination and Its Application to Wellbore Stability Anal
ysis," paper SPE 21915 presented at the 1991 SPE/IADC Drilling Con
ference. Amsterdam, March 1114.
18. Veeken. e.A.M. et al.:. "Use of Plasticity Models for Predicting Bore
hole Stability," Proc .. ISRM/SPE IntI. Rock at Great Depth Symposium.
A.A. Balkema, Rotterdam/Brookfield (1989) 2, 835.
19. Santarelli, FJ.: "Theoretical and Experimental Investigation of the Stabil
ity of the Axisymmetric Wellbore," PhD dissertation, U. of London (\ 987).
20. Santarelli. FJ . Brown, E.T., and Maury. v.: "Analysis of Borehole
Stresses Using Pressure Dependent Linear Elasticity," Illti. 1. Rock
Mech. Mill Sci. Geomech. (1986) No. 23,445.
21. Plumb. R.A. and Hickman. S.H.: "StressInduced Borehole Elongation:
A Comparison Between the FourArm Dipmeter and the Borehole Tele
viewer in the Auburn Geothermal Well." 1. Geophysical Research
(1985) 90, No. B7, 5513.
AppendixModified Wellbore Stability Model
Before determining the stress state at the wellbore wall, it is neces
sary to transpose the insitu stress tensor relative to a coordinate sys
tem with one of its axes parallel to the weIIbore axis and another
which lies in the horizontal plane (see Fig. AI). The transposed
stress state is given as
 2rvHmaxsina cos a cos f3, .................. (A2)
+ 2r \'IImax sin a cos a cos f3, .. . .............. (A3)
rxv = cos a sin f3 cos f3(a Hmax  aHmin)
~
(x.axis lies in the Z
horizontal plane)
(1\. well pressure from mud column)
Fig. A2Stress state at wall of deviated wellbore.
625
 2r VHmax sin a sin f3 ' (A4)
+ 2rVHmax cos f3(sin
2
a  cos2f3), ............. . (A5)
and r ~ , = sin a sin f3 cos f3{a Hillin  a Hma,)
 2rVHmax cos a sin f3 . .................... . (A6)
Fig. AI defines the inclination angle, a, and wellbore azimuth rela
tive to stress state, f3.
In the standard analysis approach, where the principal stresses are
vertically and horizontally oriented, then the shear component,
rHmax, is zero. In the more general case, where stresses are rotated
from the vertical and horizontal axes, then rHmax is nonzero and av
and aHmax are not principal stresses. In the structural modeling of
Cusiana aHmin is determined from plane strain conditions and al
ways taken to be a principal stress.
Having transposed the stress state, then the total stresses at the
wall of the wellbore (Fig. A2) can now be written as
a
r
= PH" ....... . ........... ... ......... . . (A7)
0.
9
= ax + a
r
 2(0. ..  aJ cos 28  4rxy sin 28  PH"
(A8)
(A9)
r 9, = 2( r" cos 8  r :.1 sin 8). . .................. (AIO)
The maximum principal stress at the wellbore wall, as used in Eq.
I, is then determined from
(0.
0
; 0.:)2 + r ~ " ........... (AII)
626
The stress state must be checked around the complete circumference
of the wellbore (e.g., by incrementing 8 in I ' steps from 0 to 359') so
that the most critically stressed point can be found and selected.
SI Metric Conversion Factors
ft x 3.048*
gal x3.785 412
in. x 2.54*
Ibm X 4.535 924
mile x 1.609 344*
psi x 6.894 757
Conversion factor is exact.
EOI =m
E03=m3
E+OO=cm
EOl =kg
E+OO=km
E+OO=kPa
Nigel Last is a senior drilling engineer with BP Exploration Colom
bia Ltd .. where he specializes in reducing wellbore instability. He
joined BP in 1988 as head of the geomechanics team at BP' s
Technology Provision Center in the U.K. and moved to Colombia
in 1994. Last holds BS and PhD degrees in engineering from the
U. of London. He was a member of the 199394 Rock Mechanics
in Petroleum Engineering Program Committee. Michael McLean
is a geomechanics engineer at BP's Technology Provision Center
in SunburyonThames. He joined BP in 1986 and has worked on
a variety of geomechanical issues, including wellbore stability,
sand production, and cuttings reinjection. He holds BS, MS, and
PhD degrees in engineering from the U. of London.
Last McLean
July 1996. JlY('
Mult mai mult decât documente.
Descoperiți tot ce are Scribd de oferit, inclusiv cărți și cărți audio de la editori majori.
Anulați oricând.