SPE

SPE 16658

Prediction of Drilling Trajectory in Directional Wells Via a New

Rock-Bit Interaction Model
by H-S. Ho, NL Petroleum Services
SPE Member

Copyright 1987, Soclely of Petroleum Engineers

This paper was prepared for presentahon at the 62nd Annual Technical Conference and Exhlbmon of the Society of Petroleum Engineers held in
Dallas, TX September 27-30, 1987

This paper was selected for presentation by an SPE Program Committee following rewew of Information contained in an abstract submitted by the
author(s) Contents of the paper, as presented, have not been rewewed by the Society of Petroleum Engineers and are subject to correction by the
author(s) The material, as presented, does not necessarily reflect any posltton of the Society of Petroleum Engineers, Its officers, or members Papers
presented at SPE meetmgs are subject to publication rewew by Edltonal CommNtees of the Society of Petroleum Engineers, Permwslon to copy IS
restricted to an abstract of not more than 300 words. Illustrations may not be copied. The abstract should contain conspicuous acknowledgment of
where and by whom the paper IS presented, Write Pubhcatlons Manager, SPE, P.O Box 833836, Richardson, TX 75083-3836. Telex, 730989 SPEDAL

ABSTRACT
more emhasis will be placed on directional
drilling. At the same time, the increased cost of
This paper presents the basic concepts and metho- such rigs has also heightened the need to reduce
dologies of a new general rock-bit interaction drilling costs (including the tripping while while
model useful in predicting drilling trajectories drilling ) and avoid drilling troubles due to
in directional (and deep vertical) wells. This unwanted hole deviations.
model is a generalization of existing similar Recognizing these dual economic constraints,
models. It accounts for the anisotropic drilling the petroleum industry has recently devoted major
characteristics of both the formation and the bit. efforts to develop capabilities to better under-
The model is developed in a 3-D geometry. There- stand and control drilling trajectories. These
fore, it is capable of predicting the walk efforts fall into two categories: active downhole
tendency and the build/drop tendency of a given directional control tools, and analysis models and
BHA (bottomhole assembly) under any drilling computer programs.
condition. The model can be used as follows: in
the forward mode to predict the drilling This paper will address only the latter. Com-
direction; in the inverse mode to generate the prehensive directional drilling analysis program
rock and bit anisotropy indices; in the will eventually help predict and/or control the
log-generation mode to generate drilling logs, drilling trajectory. In addition, they can also
such as a drilling dip log. Examples of the first help to better understand the behavior of downhole
two uses are shown. directional control tools, and to improve their
accuracy and reliability.
The drilling directions predicted using various
practices and theories are compared to those using Since the pioneering work by Lubinski [1,2] in
this general theory in parametric form. 1953, and particularly in the last 10 years, there
Significant differences can be seen, not only in have been many published works on this subject [3-
the predicted walk tendency, which is often 22]. With the recent development of MWD tools,
missed, but also in the predicted build/drop these analysis programs have taken on greater
tendency of the BHA. significance, as they can be used at the rig site
to help the directional driller make timely deci-
This paper also shows why a 3-D analysis is essen- sions. No longer are they regarded merely as
tial to properly predict the drilling direction. A qualitative well-planning aids.
2-D analysis not only completely ignores the walk
tendency of the BHA, it also incorrectly predicts B. Elements of a Predictive Directional Drilling
the build/drop tendency. Program
INDRODUCTICION A comprehensive directional drilling analysis
program will contain the following elements: (1) A
A. Role of a Predictive Directional Drilling Pro- BHA (Bottomhole Assembly) analysis program; (2) A
gram predictive model which relates the drilling direc-
tion to the bit used, the operating conditions,
Due to diminishing world oil reserves, future the borehole geometry, and the formation drilled;
exploration for fossil fuels will gradually shift and (3) A drill ahead/post analysis feature. Ele-
to more difficult reservoirs, requiring deeper ments (1) and (3) will only be briefly discussed
and/or offshore drilling. In either case, rig in the following. The capabilities of existing
costs will be much higher than in conventional directional drilling programs are still limited,
land drilling of vertical wells. Thus more and particularly with respect to the predictive aspect
83
 2 PREDICTION OF DRILLING TRAJECTORY . . . . .
(the “rock-bit interaction” modeling), element
(2), which is the subject of this paper.
Many BHA analysis programs have been developed
[1-17]. A good BHA program is the heart of a com-
prehensive directional drilling analysis program,
and can seine the following functions (See [13,
14]):
a. Quantitatively describe the deformation of
the BHA, including the total bit force
(build/drop and walk) components, and the
bit tilt direction. Ihese data, alone and/
or in conjunction with a rock-bit interac-
tion model, can be used to infer the build/
drop and, for a 3-D program, the walk
trend(s).
b. Determine the locations and magnitudes of
contact forces between the BHA and the
borehole wall. These data are useful in
estimating the wear rates of tool joints,
stabilizers, casings, and boreholes. They
are also useful in torque and drag
computations (See (e) below).
c* Compute the stresses in the BHA, which can
be used to locate the critically stressed
sections. This is particularly valuable for
the expensive downhole tool subs.
d. Calculate the difference between the survey
sub axial direction and the borehole cen-
terline direction, leading to a correction
of MWD survey data.
e. Form a part of a torque-drag model program
to enable more accurate computation of the
torque and drag in a directional and deep
vertical well. Such models are useful in
optimum well planning; in the designs of
surface equipnent, drillstring, and casing;
and in the diagnosis and avoidance of
drilling troubles.
The existing BHA programs use different
approaches (semi-analytic method, finite-element
method, or finite-difference method), and contain
different features. The pros and cons of these
approaches have been discussed in [14], and will
not be repeated here. Some of them are 2-D analy-
sis programs.
The usefulness of a BHA analysis program
depends on its inherent features and capabilities.
Selection of a BHA analysis program should be made
by matching the user’s needs with program fea-
tures. Other considerations include the quality
and rigor in the methodology used in the program,
user-friendliness, and the speed of computation,
which becomes critical if the program is to be
used at the rig site for “real-time” operations.
A drill-ahead program allows repeated calcula-
tions at different projected bit locations, thus
leading to a predicted drilling trajectory. AS a
companion feature, post drilling analysis allows
for a more detailed comparison of actual vs.
predicted drilling trajectories, and can provide
much other useful information about the well in
the form of generated “drilling logs”.This will
be discussed later in this paper.
84
SPE 16658
HISTORICAL PERSPECTIVES
Many drillers have sometimes observe rather
severe deviations. Deviation angles of up to 60°
have sometimes been observed in supposedly
vertical wells . Such phenomena were semi-
qualitatively explained by several concepts,
including the ‘miniature whipstock theory”, which
attributed them to the effect of different
formation drillabilities.
A. Practices in the control of directional
drllling
Improvements in our understanding of the
deviation tendencies of various BHA’S have come
slowly. Fig. 1 shows the key features in a predic-
tive directional drilling problem. At the present,
there is a heavy reliance on trial and error,
though one can use any one of the following exist-
ing practices for directional control:
1. Prior experience and standard BHA types
(building, dropping, or holding): This is
the most common approach;
2. Bit side force as a qualitative measure of
deviation tendency ([8]);
3. Resultant bit force direction as the actual
drilling direction [18-20];
4. Borehole curvature that induces zero side
force as the actual drilling curvature
[3,9];
5. Rock-bit interaction modeling to define the
drilling direction [l,15,21,22].
Additionally, one can use the following:
6. Bit axis direction as the projected drill-
ing direction.
methods (2-6) require the use of a suitable BHA
analysis program.
In method (l), a suitable type of BHA is
selected for a depth region to match the planned
borehole curvature, e.g., a building BHA for a
building section of the borehole. Though simple,
such an approach poses two problems. First, though
BHA ’S do generally behave as expected in a
straight hole, their drilling tendencies are
strongly influenced by the borehole curvature and
inclination, and, to a lesser extent, by the WOB
(weight on bit). A “building” BHA will become a
dropping assembly in a hole that builds at a suf-
ficient curvature, and vice versa [13,14]. Second,
such a practice does not account for the effects
of formation, borehole geometry, and operating
conditions. AS a result, what worked in one well
or depth interval may not work in another. The
consequence is that frequent correction runs are
needed.
Method (2) is an improvement over method (1)
in that it provides a semi-quantiative means of
predicting the deviation tendency of a BHA.
Methods (3-6) provide a quantitative predic-
tion of the actual drilling direction. They differ
in how the actual drilling trajectory is defined
by the known parameters, i.e., by how the “rock-
bit interaction” is modeled. The degree of success
 SPE 16658 HWA-SHAN HO
of each such method lies in how well each model
accounts for the relevant parameters affecting the
drilling direction. Some of these methods are
clearly inadequate because important parameters
are neglected.
B. Research on Quantitative Prediction of Devia-
tion Tendency
Drilling deviation is the result of rock remo-
val under the complex action of the bit. Research
on the fundamental problems of rock removal and
deviation involve three approaches: (1) laboratory
studies, (2) stress calculations, and (3) simpli-
fied analytical (“rock-bit interaction”) modeling.
The first two approaches examine the actual, if
simplified, rock removal and drilling deviation
under given bit loads, which must include a devia-
tion side force. Results of the tests or analyses
hopefully will lead to useful (even if empirically
fitted) relations that describe the deviation
tendencies of bits in any particular situation.
In terms of the first approach, earlier exper-
imental works dealt primarily with the effects of
various drilling conditions on the drilling rate
of various bits [23-32]. McLamore [33] and Bradley
[34] were the first to address the side force
generated at a single bit tooth under axial
impact. Their results confirm, at least qualita-
tively, the common observation that both the bit
and the formation exhibit anisotropic drilling
characteristics. The deviation tendency was found
to depend on the bit geometry and dip angle.
Millheim and Warren [35] carried out valuable lab
drilling tests using a rock cradle that was sub-
jected to a side force, and measured the side and
axial penetration rates. Using isotropic rocks,
they concluded that bits indeed drill anisotropi-
tally. The effect of friction between the guide
rails and the cradle was not measured, and may
influence the data interpretation.
In terms of the second approach, plasticity
theory was employed [33,36-40] to study the limit
(failure) stress state under a single bit tooth,
which was idealized as a 2-D wedge or punch. Among
these, McLamore [33], Bradley [34,39] and Smith
and Cheatham [40] considered the side force gener-
ated on the bit tooth, using simplified 2-D (upper
bound) analysis in plasticity. Though useful in
providing some insights, these static analyses
clearly do not simulate actual drilling
conditions. The results are also not easily
interpreted in terms of quantitative deviation
trends. More recently, a large scale computer
program was developed to carry out numerical
analyses [41,42] to study the simulated dynamic
response of PDC bits. The modeling and solution
processes are extremely cumbersome and require
detailed apriori knowledge of all parameters
affecting the system. Most of these data are not
available at present (and perhaps for a long time
to come). This approach is clearly not yet practi-
cal .
The above discussion shows that rational and
useful ways of describing the deviation tendencies
of a drill bit will not be possible at such a
fundamental level, at least for quite some time to
come . This makes the last approach, “rock-bit
interaction” modeling, the most appealing
approach.
85
3
INTRODUCTION TO ROCK-BIT INTERACTION MODELING
Relevant parameters that affect the deviation
tendency of a given BHA may be grouped into the
following: (1) the BHA configuration (with or
without stabilizers); (2) the borehole trajectory
and geometry; (3) the operating conditions; (4)
the bit; and (5) the formation being drilled. Each
of these groups further contain many parameters.
The essential features are illustrated in Fig. 1.
Because of the large numbers of paramters
involved, a more fundamental understanding can be
achieved only by reducing the number of inmediate
parameters by rational synthesis and grouping of
the contributing effects. A rock-bit interaction
model relates the drilling direction to the fol-
lowing vectors, as shown in Fig. 1: the bit orien-
tation, the bit force, and the formation normal.
The parameters listed in groups (l-3) are
implicitly considered through their influences on
the bit orientation and the bit force vectors. Use
of a BHA analysis program is required.
The pioneering work in this respect was by
Lubinski and Woods [1,2] (and recently, Williamson
and Lubinski [15] with the same model). The
Lubinski model includes two elements: a 2-D BHA
analysis program using a semi-analytic method to
predict the side (build/drop) force on the bit in
slick assemblies, and a formation anisotropy
effect model to account for the commonly exper-
ienced up-dip tendency in directional drilling.
They defined a rock anisotropy index to account
for the different drillabilities parallel and
perpendicular to the formation bedding plane. This
model assumes bits to be isotropic. This was
recently disputed by Ho [43], and will be further
discussed in Appendix III.
Nevertheless, since its inception in 1953, the
Lubinski model has stood for a long time as the
only rationally derived rock-bit interaction
model.
Recently, Brett et al [21] developed a bit effect
model, using the test data obtained by Millheim
and Warren [35]. Their model accounts for the
anisotropic effects of the bit, but assumed the
formation to be isotropic. Ma & Juzar [19] and
Brakel & Azar [20] also developed a bit effect
model that is coupled with BHA analysis, though
their model in effect assumes the drilling direc-
tion to be coincident with the bit force.
A NEW ROCK-BIT INTERACTION MODEL
A more general 3-D rock-bit interaction model was
recently developed at NL Technology Systems. This
mdel accounts for the simultaneous effects of
rock and bit anisotropies on the drilling
direction in the following manner (See Figure
1). The drilling direction vector is thought of
as a linear function of the following three vec-
tors: the resultant bit force f, the bit axis
and the normal vector to the formation bedding~,
as follows:
 4 PREDICTION OF DRILLING TRAJECTORY . . . . .
Here, Ir and are the the rock and bit aniso-
tropy Indices which describe the anisotropic
drilling characteristics of the rock and bit; rN
is the “normalized” drilling efficiency under
general situations; and A d is the angle between
the drilling direction and the formation normal.
For definitions, see Appendix I.
Two degenerate cases of this model are described
in Appendix II. First, if the bit is isotropic
(Fig. 2), the model in effect reduces to the
Lubinski model [1,2,15] if the bit force, bit axis
and formation normal all lie in the same vertical
plane of the borehole (i.e., the 2-D case). Se-
condly, if the rock is isotropic (Fig. 3), the
model then reduces to the Brett model [21] for a
linearly dependent drilling efficiency on the bit
force.
Since this model accounts for both the bit and the
formation effect, it has the potential to provide
accurate predictions of drilling trajectories.
Other operating parameters are considered impli-
citly by carrying out the BHA analysis program (to
generate the bit force and the bit axis vectors).
In addition, effects of RPM and hydraulics are
deemed as unimportant. These affect both the
lateral and forward drilling and will be cancelled
out, since the anisotropy indices are ratios of
two drilling efficiencies. (See Appendix I).
The normalized drilling efficiency factor r N as
defined in this model is used to define the true
“base” rock penetration rate. It is dimensionless,
and independent of the units of measurements used.
This rN should not be confused with the normalized
drilling rate [44] used to define the D-exponent,
etc. In common practice, effects of deviation from
such a “base” condition are not accounted for. In
fact, rN is the additional normalization one needs
to carry out in order to filter out the effects of
formation dip and bit on the drilling rate.
Bradley [45] previously postulated such an r N to
be less than unity, and having different patterns
for roller cone bits and PDC bits(Fig. 4). Ac-
cording to the present model, rN is merely des-
cribed by the bit anisotropy index Ib (if Ir = 1),
and has the pattern shown in Fig. 5. The situation
when I > 1 is unlikely. Interestingly, Bradley’s
model or the PDC bits coincides with the present
model when Ib = O.
APPLICATIONS OF THE ROCK-BIT INTERACTION MODEL
The rock-b it interactlon model can bet used in the
following ways, when a true 3-D BHA analysis
program is used to define the bit force and bit
axis:
1. Inverse Modeling: With known formation dip and
instantaneous drilling direction, the model
computes the rock and bit anisotropy indices.
This process is required to generate the
anisotropy indices for the next application.
2. Forward Modeling: With known formation dip,
and rock and bit anisotropy indices, the model
predicts the instantaneous drilling direction.
3. Modeling to Generate Drilling Logs: With known
anisotropy indices and the instantaneous
drilling direction, we can, in principle,
generate a “drilling dip log”. This drilling
86
SPE 16658
dip log will provide both the true dip angle
and the true dip direction.
The computer program which runs the rock-bit
interaction model can be used in stand alone mode,
or linked, as a subroutine, to a main program as
any of NL’s DIDRIL-sm programs.
The first two of these applications will be demon-
strated in the following.
APPLICATION OF INVERSE MODELING: GENERATING ROCK
AND BIT ANISOTROPY INDICES
The first application of this rock-bit interaction
model has been that of inverse modeling by evalu-
ating some old well data. only limited application
has been made so far.
To this end, well data were first screened for
suitability. The following information are needed:
1. Detailed information about the BHA assembly;
2. Survey data;
3. Operating conditions: WOB (weight on bit), TOB
(torque on bit), and mud weight;
4. Bit type/size and bit trip (and/or daily)
report;
5. Formation dip.
In addition, a lithology log and caliper log are
useful.
Data are first screened to select suitable depth
points. For each depth point, NL’s 3-D BHA analy-
sis program DIDRIL-I [13,14] is used to define the
bit force and the bit axis. The actual drilling
direction is defined by the tangent vector to the
borehole centerline, which is obtained from inter-
polating the survey data (using the circular arc
method ) . Finally, the normal to the formation
bedding is provided by 3-D formation dip informa-
tion. The rock-bit interaction model is then used
to generate the rock and bit anisotropy indices.
Use of the dip information requires same care.
Dipmeter logs, which directly provide the dip
angle and dip direction, are available only for a
few-wells. Even then, many depth sections exhi-
bited erratic dip data. In this case, only
sections with reasonably smooth dip data were
used. In other wells, only regional dip informa-
tion was available. In the Gulf coast, such
regional dip data may be acceptable if no local-
ized structures, such as salt domes, are present
in the particular well (or depth region) being
analyzed. Otherwise, results may not be reliable.
Another important factor that can significantly
influence the data interpretation is the borehole
caliber (and similarly, the stabilizer wear). A
change in borehole diameter, be it overgage due to
washouts or instability, or undergage due to bore-
hole creep, can significantly influence the BHA
deformation which not not be accounted for in the
model, particularly if this occurs near the bit or
the first couple of stabilizers. In such situa-
tions, the bit axis and the bit force directions
obtained from the BHA analysis maybe inaccurate.
DIDRIL SM is a service mark of ML Petroleum
Services
 SPE 16658
In this case, unreasonable anisotroy indices
(such as negative numbers) may be obtained. This
problem points out the importance of knowing the
borehole conditions accurately. The use of MWD
surveys will alleviate this problem to some extent
due to more timely and more frequent data
collection.
Our limited results show the following average
values:
Ib = .194; Ir = .999.
The bits used are soft-formation roller cone bits,
and are shown to be very anisotropic. The forma-
tion is only slightly anisotropic. Table 1 sum-
marizes a portion of the data upon which the
averages are based. These data are obtained in the
depth internal using the same building BHA as
described in Table 1.
APPLICATION OF FORWARD MODELING: PREDICTION OF
DRILLING DIRECTIONS
The model can also be used to predict the instan-
taneous drilling direction with a single analysis,
or the drilling trajectory with repeated analyses.
Using the average Ir and 1b obtained from the
inverse modeling, the rock-bit Interaction program
recomputes the predicted survey data, using the
same BHA for the same depth interval as in the
example above.
Table 2 summarizes the result. In the table, the
"actual" borehole deviation and azimuth angles are
computed through survey interpolation using the
circular arc method. As can be seen, the model
predicts the drilling directions very well. The
average difference over a depth interval of about
300’ between the predicted and the actual survey
data are:
Deviation angle difference: .037”;
(Variance: .0200).
Azimuth angle difference: .031”;
(Variance: .0360).
IMPORTANCE OF BOTH THE ROCK AND BIT ANISOTROPIES
Although the rock is found to be much less aniso-
tropic than the bit, this does not mean we can
arbitrarily set it to be unity and use the degen–
crate model for isotropic rocks (Appendix II).
There are two reasons: (1) The angle between the
bit force and the bit axis is limited by the
borehole confinement and drillstring deformation,
and is therefore very small (on tie order of a few
degrees). on the other hand, the angle between the
bit force and the formation normal is quite arbi-
trary, and may be as large as 90°. (2) The devia-
tion (measured from the bit force) is much more
sensitive to changes in the rock anisotropy index
I t h a n i n Ib . F i g .
r 6 and Fig. 7 illustrate these
sensitivities.
Furthermore, because the angle between the bit
force and the bit axis is generally very small, it
is important to have a reliable BHA analysis
program. Small errors in the computed bit force
and bit axis vectors may cause large errors in the
generated anisotropy indices.
HWA-SHAN HO 5
COMPARISON OF PREDICTION METHODS
In this section, comparisons will be made between
the drilling directions predicted using several
different approaches.
The following parameters
are held constant:
WOB - 40K; TUB = 5’-K; MU DWT. = 10 ppg;
HOLE INCLINATION = 45°; HOLE AZIMUTH = 90° at bit.
along with the same “typical” building BHA.
Three different well trajectories are examined:
(Table 3): straight well;
(Table 4): 2-D well building at 20/100’;
(Table 5): 3-D well additionally walking at
2°/100’ to the right.
For each situation, five prediction methods are
presented:
1.
2.
3.
4.
5.
Tables
data groups: . .
a. Dip angles at O, 20, 40 and 60°;
For O dip angle, results are independent of
the azimuth angle, and are shown under the
table.
b. Formation normal Azimuths at 90 (hole nearly
Perpendiculaar to bedding), -90 (hole nearly
parallel to bedding), O (out-of plane dip) and
45°.
For isotropic rocks (Ir = 1), results are indepen-
dent of dip variation. Therefore only one case is
shown in each of the tables. In the tables, the
prediction method number is shown in parenthesis.
A deviation angle from hole axis of .3° will be
mild, while l.O will be strong. Since this devia-
tion angle is the instantaneous drilling deviation
angle, it is not directly translated into the more
comnon notion of change in hole curvature. To
compute that, one needs to carry out successive
calculations after each finite drilling distance,
and then take the average curvature. This incre-
mental approach is probably more realistic than
the common notion, as it more closely duplicates
the actual drilling process.
In Table 3, we see the bit force to be strongly
building, while the bit axis is actually slightly
dropping. AS a result, method (2) would predict a
very mild dropping trend, while all other methods
predict mild to strong building trends. AS
expected, methods 3 & 4 predict similar left-
walking, but differ very significantly in the
build trend prediction.
87
 6 PREDICTION OF DRILLING TRAJECTORY . . . . . SPE 16658

In Table 4, the inherent hole curvature causes

both the bit force and the bit axis to be drop-
ping. This is due to the stiffness of the BHA, as
pointed out previously [13,14]. Therefore, all Vector A, with magnitude A, and unit
methods predict a dropping trend. Methods 3 & 4 vector
also predict a left-walking trend. The severity of
the dropping trend varies according to the Components of vector A in (X,Y,Z)
methods. Note that, once drilling is allowed to directions;
proceed according to the predicted direction
(dropping), the hole curvature is reduced, and Unit base vectors along (X,Y,Z) di-
therefore the inherent dropping tendency of the rections;
BHA will also be reduced. This will then change
the future drilling direction to be either less Unit vector along bit axis direction
dropping, or even return to slightly building.
Such repetitive computations and case studies will
be presented in later papers. Unit vector normal to formation bed-
ding;
In Table 5, the right-walking hole curvature
further causes left-walking trends in both the bit Unit vector along the resultant bit
force and the bit axis. As a result, all methods force on formation;
now predict moderate to strong left-walking ten-
dencies. Unit vector along the drilling di-
In both 2- and 3-D holes, we see that using the rection;
bit force (method (2)) as the predictor of drill-
ing direction actually provides the greatest Resultant bit force on the forma-
scatter. Most current practices are in fact based tion;
on this method.
CONCLUSIONS Lubinski’s rock anisotropy index =
1- Ir;
In summary, the following conclusions may be
drawn: Bit anisotropy index;
1. A new rock-bit interaction model is now avail- Rock anisotropy index = 1 - h;
able. It is a generalization of existing
models. Drilling rate along direction ();
2. The rock-bit interaction model simultaneously Drilling efficiency along direction
accounts for both the rock and bit anisotropy
effects in 3-D space.
Subscripts ():
3. The model requires a reliable 3-D BHA analysis
program to generate the bit force and bit axis o: Base quantities, referring to situa-
directions. tion when both rock and bit are iso-
4. The model can be used in the forward mode to tide;
predict the drilling trajectory. a: Bit’s axial direction;
5. The model can be used in the inverse mode to d: Formation normal direction;
generate rock and bit anisotropy indices.
f: Bit force direction;
6. The model can also be used to generate other
drilling logs, such as a drilling dip log. 1: Bit’s lateral direction;
7. Complete and accurate field data is important n: Bedding’s normal direction;
for-generating the anisotropy indices. -
p: Bedding’s parallel direction;
ACKNOWLEDGEMENTS
N: “Normalized” quantity;.
The author wishes to express his appreciation to
the following persons: To Paul Rodney, John r: Drilling direction.
Fontenot and Vik Rao for their sustained support
of the directional drilling seRVice programs at NL * Note * When two subscripts appear, that
Industries; To William Bradley of Standard oil, pertains to bit direction comes
for discussion on the model while he was at NL; To first.
Carolyn McFarland and Mary Fouts for their typing
support and firmly, the author thanks the NL (X,Y,Z): Fixed global coordinate system,
management for the permission to publish this x —> East, Y —> North, Z —> Ver-
paper. tical up;
e: Inclination angle;
: Azimuth angle, measured C.W. from
north.
88
 SPE 16658 HWA-SHAN HO
REFERENCES
[1 ] Lubinski, A. and Woods, H.B.: “Factors
Affecting the Angle of Inclination and Dog-
legging in Rotary Bore Holes”, API Drilling &
Prod. Pratt., 1953, pp.222-250.
12 ] Woods, H. B. & Lubinski, A.: “Use of Stabi-
lizers in Controlling Hole Deviation”, API
Drill. & Prod. Pratt., 1955, pp.165-182.
[3 1 Murphy, C. E. & Cheatham, J. B. Jr.: “Hole
Deviation and Drillstring Behavior”, SPE J.,
Mar. 1966, p. 44.
[4 ] Fischer, F. J.: “Analysis of Drill Strings in
Curved Boreholes”, SPE Ann. Conf., 1974,
Houston. SPE Paper #5071.
[5 ] Bradley, W. B., Murphy, C. E., McLamore, R.
T. & Dickson, L. L.: “Advantages of Heavy
Metal Collars in Directional Drilling and
Deviation Control” SPE Annual Mtg., Dallas,
1975. SPE Paper #5545.
[6 ] Walker, B. H. & Friedman, M. B.: “Three-
Dimensional Force and Deflection Analysis of
a Variable Cross-Section Drillstring”, J.
Press. Vessel Tech., Trans. ASME, May 1977,
pp. 367-373.
[7 ] Millheim, K., Jordan, S. and Ritter, C. J.:
“Bottomhole Assembly Analysis Using the
Finite Element Method”, JPT, Feb. 1978, pp.
265-274.
[8 ] Millheim, K. K.: Eight part series on Direc-
tioml Drilling, OGJ., Nov. 1978 to Feb.
1979.
[9 ] Callas, N. P. & Callas, R. L.: “Boundary
Value Problem is Solved”, OGJ., wc. 1980,
pp. 62-66.
[10] Millheim, K. K. & Apostal, M.: “The Effect of
Bottomhole Assembly Dynamics on the Trajec-
tory of a Bit", JPT, Dec. 1981. pp. 2323-
2338.
[11] Toutain, P.: ‘Analyzing Drillstring Beha-
vior”, World Oil, 1981. Part I: June; Part
II: July; Part III: Sept.
[12] Dunayevsky, V. A. & Judzis, A.: “Conservative
and Nonconservative Buckling of Drill Pipe”,
58th Annual Mtg. SPE of AIME. San Francs.
1983.
[13] Rafie, S., Ho, H.-S. and Chandra, U.: “Appli-
cation of a BHA Analysis Program in Direc-
tional Drilling”, IADC/SPE Conference,
Dallas, 1986. SPE paper #14765.
[14] Ho, H.-S.: “General Formulation of Drill-
string Under Large Deformation and Its Use in
BHA Analysis”, SPE Ann. Tech. Conf., 1986,
New Orleans, SPE paper #15562.
[15] Williamson, J. S. and Lubinsk, A.: ‘Predict-
ing Bottomhole Assembly Performance”, SPE/
IADC conference, Feb. 1986, Dallas, SPE paper
#14764 . Also in SPE/Drilling Engng., Mar.
1987, pp. 37-46.
89
7
[16] Jogi, P. N., Burgess, T. M. & Bowling, J. P.:
“Three-Dimensional Bottomhole Assembly Model
Improves Directional Drilling”, SPE/IADC
Conf., Dallas, 1986. SPE Paper #14768.
[17] Birades, M.: “ORPHEE 3D: Static and Dynamic
Tridimensional BHA Computer Models”, SPE
Annual Technical Conf. and Exh., Oct. 1986,
New Orleans, SPE paper 15466.
[18] Sutko, A. A., Myers, G. M. and Gaston, J. D.:
“Directional Drilling - A Comparison of
Measured and Predicted Changes in Hole
Angle”, SPE Paper 8336, 54th Ann. SPE Conf.,
1979, 8p.
[19] Ma, D. & Azar, J. J.: “A Study of Rock-Bit
Interaction and Wellbore Deviation”, J.
Energy Rests. Techn., Trans. ASME, V. 108,
Sept. 1986, pp. 228-233.
[20] Brakel, J. D. & Azar, J. J.: “Prediction of
Wellbore Trajectory Considering Bottomhole
and Drill Bit Dynamics”, SPE/IADC Conf., New
Orleans, 1987, pp. 1077-1090. SPE Paper
#16172 .
[21] Brett, J. F.; Gray, J. A.; Bell, R. K. and
Dunbar, M. E.: “ A Method of Modeling the
Directional Behavior of Bottomhole Assemblies
Including Those with Bent Subs and Downhole
Motors”, SPE/IADC conference, Feb. 1986,
Dallas. SPE paper 14767.
[22] Codling, J.: “Heuristic Program - A Guide to
Directional Drilling”, ASNE Drilling & Pro-
duction Syq., Feb. 17-21, 1985, pp. 127-137.
[23] Somerton, W. H.: “A Laboratory Study of Rock
Breakage by Rotary Drilling”, Petroleum
Trans., AIME, V. 216, 1959, pp. 92-97.
[24] Gamier, A. J. & Van Lingen, N. H.: “Pheno-
mena Affecting Drilling Rates at Depth”,
Petrol. Trans., Arm, v. 216, 1959, pp. 232-
246.
[25] Cunningham, R. A.: “Laboratory Studies of the
Effect of Rotary Speed on Rock-Bit Perfor-
mance and Drilling Cost”, API Drill. Prod.
Pratt., 1960, pp. 7-14.
[26 Simon, R.: “Energy Bahnce in Rock Drilling”,
SPE J., Dec. 1963, pp. 298-306. Paper pre-
sented at SPE - univ. Texas Symp. on Drill.
Rock Mech., Jan. 1963.
[27] Garner, N. E. and Gatlin, C.: “Experimental
Study of Crater Formation in Plastically
Deforming Synthetic Rocks”, JPT, Sept. 1963,
pp. 1025-1030.
[28] Burdine, N. T.: “Rock Failure under Dynamic
Loading Conditions”, SPE J., Mar. 1963, pp.
1-8.
[29] Gnirk, P. F. & Geatham, J. B. Jr.: "Ah
Experimental Study of Single Bit-Tooth Pene-
tration Into Dry Rock at Confining Pressures
O to 5000 psi”, SPE J., June 1965, pp. 117-
130.
[30] Maurer, W. C.: “Bit-Tooth Penetration under
Simulated Borehole Conditions”, JPT, Dec.
1965, pp. 1433-1442.
 8 PREDICTION OF DRILLING TRAJECTORY . . . . . SPE 16658

[31] Yang, J. H. & Gray, K. E.: “Single-Blow Bit- APPENDIX

Tooth Impact Tests on Saturated Rocks under
Confining Pressure: II. Elevated Pore Pres- I. DEFINITIONS OF ANISOTROPY INDICES
sure”, SPE J., Dec. 1967, pp. 389-408.
A. Rock Anisotropy Index Ir
[32] Peterson, C. R.: “Roller Cutter Forces”, SPE
J.,Mar. 1979, pp. 37-65. The rock anisotropy index Ir is directly
definable if the bit is isotropic, or if the
[33] McLamore, R. T.: “The Role of Rock Strength resultant bit force is along the bit axis. Under
Anisotropy in Natural Hole Deviation”, JPT, these situations, we can define the normal and
NOV. 1971, pp.1313-1321. parallel drilling efficiencies, r n and rp, as:
[34] Bradley, W. B.: “Deviation Forces From the
Wedge Penetration Failure of Anisotropic (A-1)
Rock “ , J. Engineering for Industries, Trans.
ASME, Nov. 1973, pp. 1093-1100.
[35] Millheim, K. K. and Warren, T. M.: “Side (A-2)
Cutting Characteristics of Rock Bits and
Stabilizers While Drilling”, SPE paper 7518,
Fall Annual SPE Conf. 1978, 8p.
[36] Paul, B. & Sikarskie, D. L.: “A Preliminary (A-3)
Theory of Static Penetration by a Rigid Wedge It has the following ranges:
into a Brittle Material”, Trans. SME of AIME,
1965, p. 372. Ir = 0: drilling only perpendicular to bedding;
[37] Cheatham, J. B., Jr., Paslay, P. R. & < 1: faster drilling along normal to bedding
Fulcher, G. W. G.: “Analysis of the Plastic
Flow of Rock under a Lubricated Punch”, J. (up-dip tendency);
Appl. Mechanics, Trans. ASME, 1968, pp. 87- = 1: isotropic rock, no formation effect;
94.
> 1: slower drilling along normal to bedding
[38] Pariseau, W. G.: “Wedge Indentation of Aniso- (down-dip tendency );
tropic Geologic Media”, Dynamic Rock Mecha-
nics, Ed. G. B. Clark, Proc. 12th U. S. Symp. drilling only parallel to bedding.
Rock Mechanics, AIME, 1971, pp.529-546.
B. Bit Anisotropy Index Ib
[39] Bradley, W. B.: “Factors Affecting the Con-
trol of Borehole Angle in Straight and If an anisotropic bit is drilling into isotro-
Directional wells”, JPT, June 1973, pp. 679- pic rock, we can define the axial and lateral
688. drilling efficiencies, ra and rl, as:
[40] Smith, M. B. and Cheatham, J. B. Jr.: “Devia-
tion Forces Arising From Single Bit Tooth
Indentation of an Anisotropic Porous Media”,
J. Press. Vessels Tech., Trans. ASME, May (A-4)
1977, pp. 362-366.
[41] Baird, J. A., Tinianow, M. A., Caskey, B. C.
& Stone, C. M.: “GEODYN: A Geological Forma-
tion/Drillstring Dynamics Computer Program”, (A-5 )
59th Ann. Conf., SPE of AIME, Houston, 1984. The bit anisotropy index is then:
SPE Paper #13023.
Ib = rl/ra. (A-6 )
[42] Baird, et. al: “GEODYN2: A Bottomhole Assem-
bly/Geological Formation Dynamic Interaction It has the following ranges:
Computer Program”, SPE Ann. Conf., Las Vegas,
1985. SPE Paper #14328. Ib= O: drilling only along axial direction;
[43) Ho, H.-S.: ‘Discussion on: ‘Predicting Bot- < 1: faster drilling along bit’s axial di-
tomhole Assembly Performance’ by J. S. rection
Williamson & A. Lubinski, SPE/Drilling Engng.
J ? Mar. 1987, pp. 37-46”, to appear.
●
= 1: isotropic bit, no bit effect;
[44] Jordan, J. R. & Shirley, O. J.: “Application > 1: slower drilling along bit’s axial di-
of Drilling Performance Data to Overpressure rection;
Detection”, JPT,Nov. 1966, pp. 1387-1394.
-> : drilling only lateral to bit’s axis.
[45] Bradley, W. B.: “Formation Characteristics
Have a Key Effect on Hole Direction”, Oil/Gas II. SPECIAL CASES OF THE GENERAL MODEL
J ? Aug. 1975, pp.77-80.
●
A. Isotropic Bits
This case degenerates essentially into the
90
 SPE 16658
Lubinski model [1,2,151, though the latter was
derived specifically only for a 2-D situation,
namely the bit force, drilling direction, and the
formation normal vectors all lie in the same ver-
tical plane as the well trajectory. The Lubinski
model does not account for any walk tendencies,
while this isotropic bit model does. Note that the
rock anisotropy index h as defined by Lubinski is
related to the current definition Ir by the fol-
lowing relation:
h = l - Ir.
Eq.(1) can be reduced to the following simple
..
This relation is shown in Fig. 2 in the general
situation when and do not lie in the same
vertical plane, and thus requires a 3-D special
description.
Fig. 8 shows a series of curves describing the
deviation angle (measured from the bit force) as a
function of the rock anisotropy index Ir, and Afd,
the angle between the bit force and the format ion
normal. In all cases, the maximum deviation occurs
when Afd is 45°, while no deviations exist when
Afd is zero (normal drilling) or 90° (parallel
drilling).
B. Isotropic Rocks
In this case, equ. (1) reduces to the follow-
ing:
and is illustrated in Fig. 3. For “normally ani-
sotropic” bits, Ib is less than unity.
Curves similar to Fig. 8 can be used if one re-
plaCeS Ir and Ib and respectively.
III. DIFFERENCE BETWEEN 2-D AND 3-D ANALYSES
Some existing models utilize a 2-D analysis,
resulting in only a build/drop prediction. AS an
example, in assessing the formation effect, HO
[43] recently showed that, due to the difference
in the apparent dip angle (seen in the common
vertical plane) and the true dip angle (tilting
away from the vertical plane), the predicted
drilling direction (in the common vertical plane)
will change. This will affect the result of build/
drop prediction. It may also mask the bit aniso-
tropy effect.
Parallel arguments exist when one examines
only the bit effect, and will not be repeated
here.
In a 2-D model, where the entire wellbore and
drillstring are assumed to lie in the same verti-
cal plane. the formation dip is seen as the
apparent dip and not the true dip. These angles
are equal only when the relative strike angle of
the dipping plane is 90°. Otherwise, the apparent
HWA-SHAN HO 9
dip angle is always smaller than the true dip
angle. In the extreme case when the relative
strike angle is zero, the apparent dip angle is
always zero, even when the true dip angle is 90°!
In analyzing the formation effect, the above
difference is illustrated in the following (Fig.
9). In a 2-D analysis, all relevant vectors are
assumed to lie on the common vertical plane, which
is the base plane. The formation normal vector is
a; the bit force is decomposed into the normal
(A-7)
vector Era to pass through the point Ca. The ratio
CaBa/ABa describes the degree of anisotropy of the
formation, which is an anisotropy index. Vector
In a 3-D analysis, one uses the true formation
which in this particular case
points above the base plane. The similar bit force
components are OB and AB, and the drilling direc-
AB is again the anisotropy index, which is also
the same as CpBp/ABp (where the subscript p de-
notes the projection onto the base plane) due to
parallel projections. We can then conclude that
therefore cannot be parallel to the vector
other words, the vector Er does not project into
the vector Era Additionally, the 3-D analysis al-
so results in a walk component of Er pointing
above the base plane.
(A-9)
Using 3-D vector analysis, one can derive the in-
plane build/drop deviation angle Aa (from 2-D
analysis) and (from projected 3-D analysis),
relative to the bit force vector, as follows:
It is conceivable that the true drilling
direction might have a building tendency while the
apparent drilling direction might show a dropping
tendency, or vice versa. In anisotropic forma-
tions, there are only two exceptions to the above
conclusion: when the relative strike angle A r is
AOO or OO.
91