Predicting Bottomhole

Assembly Performance
J.S. Williamson, SPE, Smith Drilling Systems, Div. of Smith IntI. Inc.
A. Lubinski, SPE, Consultant

Summary. A computer program for prediction of bottomhole assembly (BHA) performance has been
developed. Input parameters include formation dip angle, hole and collar size, and stabilizer spacing. The
predictions are in drilling terms: hole curvature, hole angle, weight on bit (WOB), etc. Results of extensive
parametric studies and field use are presented.

Introduction
Poorly designed BHA's restrict the WOB, thus increasing (4) force on bit depends on only the mechanics of the
cost. A BHA computer program, BHAP, has been devel- BHA, not the formation.
oped to predict performance and has been used extensively To account for updip drilling, drillability in the direc-
in conducting parametric studies and in making field de- tion perpendicular to bedding planes is considered greater
cisions for the best performance. The program takes into than in the direction parallel to bedding planes. 19-21 This
consideration formation characteristics, any number of has been confirmed by laboratory experiments. 22 The ra-
stabilizers, drill-collar (DC) sizes, square collars, shock tio of the two drillabilities is related to the so-called for-
absorbers, motors, directional tools, and measurement- mation class. Thus formation crookedness depends on
while-drilling (MWD) tools. Predictions are given in drill- both the formation dip angle and formation class, Fe (see
ing terms, such as rate of building or dropping angle in Eqs. A-I through A-3). Such a model is simple. A more
degrees per 100 ft [0.57xlO- 3 rad/m]. The simplicity complicated model is not warranted because of the lack
of the mathematical model-two-dimensional, constant of precision of input parameters, such as hole gauge and
hole curvature, static-results in manageable computer stabilizer clearance. For an isotropic formation, Fe =0.
time, which is vital for field use. For a very anisotropic formation, corresponding to For-
mation Class A in Ref. 23, Fe = 100. (See Bit Anisotropy
in the Appendix.)
Bit/Formation Interaction
It is well known that dipping formations generally result Input
in high hole angles. Several papers on the effect of bit/for- The following parameters are entered: hole size; hole an-
mation interaction on deviation have been published. 1-5 gle; dip angle; formation class; mud weight; WOB; DC
Refs. 1, 3, and 4 are based on the mechanism of a single information; and stabilizer(s) information, including clear-
wedge or tooth, and not the bit itself. Application of these ance(s), bent sub, or other directional-tool information
models to a three-cone bit would be extremely difficult. (if any), and hole curvature (assumed constant)-i.e.,
Models for polycrystalline-diamond-compact bits are degrees per 100 ft [0.57x 10- 3 rad/m]. All these param-
more feasible, but expensive because of computer time. 6 eters should be known, except one-the unknown.
In most of these papers, the bit/formation interaction Noncylindrical cross sections (square collars) or tools
is not coupled to the BHA. Similarly, most of the publi- with reduced stiffness-such as shock absorbers, motors,
cations dealing with BHA computer programs (Refs. 7 and MWD tools-may be modeled.
through 17) do not include any bit/formation interaction. As an example, the following is used to model square
Therefore, the output is the lateral force on the bit and collars: (1) OD is taken as the distance across the cor-
the direction in which the bit points, but not the parame- ners, (2) ID is chosen to model the moment of inertia,
ters of greatest interest, such as hole angle and rate of and (3) the density is chosen to model the weight per unit
change of hole angle. length.
To predict the performance of BHA's in field-usable
terms, the effect of the formations must be included. One Prediction of Performance
paper attempts this by postulating that dipping formations To be successful in the application of BHAP in the field,
impose a bending moment at the bit. 18 Another supposes we first conducted extensive parametric studies. These
the existence of geologic forces 16 that push the bit up- studies provided basic understanding of field phenome-
dip. We believe that (1) drilling in the direction of the na. The following hypothetical cases are presented to show
force on bit occurs only in non sedimentary formations, the need for dip information and how BHAP may be
such as granite; (2) drilling is not in the direction in which applied.
the bit points; (3) bending moment at the bit is zero; and
Case 1. Consider the following at 2,800 ft [853 m]:
Copyright 1987 Society of Petroleum Engineers 12.25-in. [31. 12-cm] hole, 49,000-lbf [218-kN] WOB,
SPE Drilling Engineering, March 1987 37
 HOLE ANGLE (DEGREES) (j) 24
:l
3r-__. -__,2____3~~~4=_-,5----6.---17--~8 o 22 / . 2 STABILIZER
z /
;Ji 20 , /'1 STABILIZER
f=' ::::l
----...........( '
------
!Jj 4 oI
u.
o PACKED ASSEMBLY
t:.
~ ~;;;
......
........... ...... _-
~ 5 W
..J 16 A SLICK --
Ul
:;, ::;;!!l o ------------
o :;,Ul
14
E6 ::;;!!1
-::;;
xa:
w
I-
I
(!)

:r ::;;a. ~ 12
I-
fu 7~--------------------------~~ 10 ~~~-L-:....". j .~~~~~L~~~-'-'
o 90 95 100 105 110 115 120 125 130
(3 DC) FEET (4 D.C )
8 LENGTH FROM BIT TO FIRST STABILIZER

Fig. 1-Parametric study, Case 1-20 0 dip. Fig. 3-Parametric study, Case 2-pendulum optimization.

49,000-lbf [218-kN] WOB. Therefore, the decision is to

HOLE ANGLE (DEGREES) use a packed assembly.
o 2345678 40 [0. 7-rad] Dip. The slick and the previously accept-
3r---~---'--~~-'----r---~---r--~
, able packed assembly are no longer sufficient here. Fig.
f=' 2 shows the predicted trajectories of the previously con-
!Jj 4 W
u.
..J
::;;!!l
sidered packed assembly and an assembly with stabiliz-
o :;,Ul ers and a square collar. The conclusion is that the latter
::;;!!1
~ 5 -::;; would be required to allow the full 49,000-lbf [218-kN]
({) xa:
:;, w WOB to run without exceeding 7 at 7,000 ft [0.12 rad
o ::;;a.
E6 PACKED w! SQUARE
ASSEMBLY
at 2134 m].
i!= These three examples show that, for the same drilling
fu 7r------------------i--------~ data, three very different actions are required, depend-
o
ing on the actual dip angle.
8
Case 2. Consider the following: 12.2S-in. [31.12-cm]
Fig. 2-Parametric study, Case 1-40 0 dip. hole, no stabilizers, 8-in. [20.3-cm] DC's, and holding
3 for lS,OOO-lbf [O.OS rad for 66.7-kN] WOB. We wish
no stabilizers, 8-in. [20.3-cm] DC's, 1.so [0.026-rad] hole to increase the WOB by using stabilizers in a pendulum
angle, and zero hole curvature. Suddenly a new forma- assembly. The stabilizer position and the WOB allowed
tion is encountered, and at 3,000 ft [914 m], the hole an- to hold 3 [O.OS rad] are plotted in Fig. 3.
gle is increased to 3 [O.OS rad]. This means that the angle From the curve for a single stabilizer, we see that the
builds up at the rate of O. TSO /lOO ft [0.43 x 10 -3 rad/m]. WOB could be increased to more than 19,000 Ibf [84.S
We are to drill an additional 4,000 ft [1219 m]-i.e., kN] if a single stabilizer were placed at 108 ft [32.9 m].
from 3,000 to 7,000 ft [914 to 2134 m] with a maximum This calls for three DC's plus a short collar. Use of a short
limit of 7 [0.12 rad]. What will be done depends on for- collar in a pendulum assembly is generally considered un-
mation dip. Three values of dip will be considered. desirable.
6 [O.l-rad] Dip. By using the program with Fe as the Points A and B on the single-stabilizer curve show that
unknown, we obtain Fe =46.3. Then by using the pro- only 16,000- or 17,000-lbf[71.2- or 7S.6-kN] WOB may
gram with the same input values (except that Fe =46.3)- be carried if no short collar is used. This is for three or
zero hole curvature, and hole angle as the unknown-we four 31- ft [9 .4S-m] collars and one-half the length of a
obtain a hole angle of 4.7So [0.083 rad]. Thus the hole stabilizer. Thus much of the advantage of the stabilizer
angle will build to 4.7So [0.083 rad] and then stop. Be- is lost. To avoid use of a short collar, while obtaining
cause this is acceptable, no further action is required. practically the optimum WOB, two stabilizers may be
20 [0.35-rad] Dip. By proceeding as above, we ob- used, the first at three collars above the bit and the sec-
tain the equilibrium angle of7.So [0.13 rad], which may ond at two collars above the first stabilizer. This is shown
not be satisfactory, depending on the inclination the hole by Point C (19,000 Ibf [84.S kN]) on the two-stabilizer
will reach at 7,000 ft [2134 m]. One must therefore predict curve.
the hole trajectory in a manner described in the Appen- Finally, to maintain 3 [O.OS-rad] hole angle as stabiliz-
dix. The trajectory is the curve denoted "slick assembly" ers wear, the WOB must be reduced. For this example,
in Fig. 1, which shows that a slick assembly is unsatis- BHAP showed that this reduction is only about 1 %.
factory, because the hole angle of 7 [0.12 rad] would Stabilizer clearance in pendulum assemblies has been
be reached at about 4,300 ft [1311 m]. Therefore, either found unimportant. Unfortunately, the contrary holds true
the WOB must be reduced or use of a packed assembly for packed assemblies.
must be considered. The predicted trajectory of a packed Incidentally, solution of Case 2 problems does not re-
assembly is shown in Fig. 1. At 7,000 ft [2134 m], the quire the knowledge of the dip. This is the case when there
hole angle will be only 6.So [0.11 rad]. with the full is no change in hole angle and formation.
38 SPE Drilling Engineering, March 1987
 ---ASSEMBLY 0.60r-----------------,
W~ O-)(-ll X
O::G:
:::>0 4=:L I+l14 89-.J
~o 0.20 li!f
>:!::
o::cn 0
58
:>!::
0.20
----
:::>W
U W o::cn
:::>w O~----------~~~~~--------~
wO:: -0.20
-1 0 UW
OW
Ie. ----ASSEMBLy ~ -0.20
-0.40 OW
4i?=~ Ie. -0.40 ----ASSEMBLy
-0.60
-0.80
99--1
-0.60 4-}E;:r-99 J
50 60 70 80 90 100 110 -0.80 L...-.J.-...L.-......L--'---'----'_L...-.L--'--'--'---L.--I
W.O.B. 25 27 29 31 33 35 37
(THOUSANDS OF LBS.) FORMATION CLASS

Fig. 4-Sensitivity to WOB. Fig. 5-Sensitivity to formation change.

Case 3. Consider now a directional case. We are in a 0.60r-----------------.

12.25-in. [31.12-cm] hole and want an assembly to hold
22 with 50,000- to 85,000-lbf [0.38 rad with 222- to li! E' 0.40
378-kN] WOB. The dip is flat (0.5 [0.009 rad]). A sys-
tematic approach is to use the program in batch mode. ~ 0 0.20
This allows consideration of all three-stabilizer assemblies ~~
~~
UW
O~~~~~~,~,~------------~
, , __
where the length between stabilizers is varied by steps
of 5 ft [1.52 m]. The output gave 12 assemblies that satis- W 0::
-I (!)
-0.20 -_ ' -_ _ ,
o ~ -040
fied the above conditions. This number was reduced to
I _ . t
----ASSEMBLy

4--8=~::J
four when only 75,000- to 85,000-lbf [334- to 378-kN]
WOB was considered. The final decision of which assem- -0.60
bly to choose comes from the BHAP output, which gives -0.80L.----:7;------;"':::-----=':-:--_::_"
sensitivities. N 1~ F~
ew Worn Worn Worn Worn
Figs. 4 through 6 explain the meaning of the sensitivi- STABILIZER WEAR
ties. They show plots of hole curvature vs. WOB, for-
mation class, and stabilizer diametrical clearance for two
of the four assemblies. Ideally, we want as little change 0.043 0.070 0.098 0.125 0.152
in hole curvature as possible for unavoidable changes of NEAR BIT STABILIZER CLEARANCE (Inches)
WOB, formation type, and stabilizer clearance.
These figures show that Assembly 4-14-89 is more sen-
0.094 0.125 0.156 0.188 0.219
sitive to changes of formation class and WOB, but less
sensitive to change in stabilizer wear than Assembly OTHER STABILIZER CLEARANCE (Inches)
4-49-99. The sensitivities are essentially the slopes of Fig. 6-Sensitivity to wear.
curves.
The program also gives forces on stabilizers, which in-
fluence decision making. Smaller forces minimize dig-
ging into the formation and stabilizer wear, which is very
important for packed assemblies. With all the above con- Case 4. Consider the following: 12.25-in. [31.12-cm]
siderations, Assembly 4-49-99 would be chosen. hole, 31.4 [0.548-rad] hole angle, 50 [0.87-rad] for-
There are three abscissa scales in Fig. 6. The upper mation dip, no stabilizers, 7-in. [17.8-cm] DC's,
is the stabilizer wear in a hole on gauge. The two lower 1O,560-lbf [47-kN] WOB, and zero hole curvature. The
ones are stabilizer clearances that may result from stabiliz- unknown is Fe. From the program, we obtain Fe =63.4.
er wear or an oversized hole, or both. A packed assembly Consider now the same input, except that F c=63.4
is very sensitive to clearance. For Assembly 4-49-99, the and there is a 7-in. [l7.8-cm] motor whose housing is bent
hole curvature varies from +0.12' to -0.341100 ft at 5 ft [1.52 m] above the bit. Additional input is the an-
[+0.069 x 10 -3 to -0.19 x 10 -3 rad/m] from one end gle of the bend given in Table 1. The output (i.e., the
of the scale to another. unknown) is hole curvature. An additional output, which
Generally, the hole angle increases with the WOB. actually is not needed, is the angle between the hole axis
When the hole angle is greater than the dip angle, how- and the direction of the bit axis.
ever, the contrary may happen for a packed assembly. Hole curvature is found to be of the same sign as the
Fig. 4 is an example of such a case. For Fig. 4, increasing angle of the bend. On the other hand, for the first two
the WOB to more than 85,000 lbf [378 kN] results in lines in Table 1, the angle in which the bit points is not
negative hole curvature (reducing hole angle). As we will of the same sign as the angle of the bend. For a better
show, this unexpected phenomenon actually occurs. understanding, see Bit Anisotropy in the Appendix.
SPE Drilling Engineering, March 1987 39
 WOB
TABLE 1-INPUT AND OUTPUT FOR CASE 4 HOLE ANGLE
(THOUSAND LB)
(DEGREES)

~~OO--r
15 16 17 18 19 20 21 22 23 24 25
Angle Between
Angle of Hole Bit Axis and J-A FIRST

"
rCK'~"
Bend' Curvature' Hole Axis' I ASSEMBLY

...
01
(degrees) (degrees/100 ft) (degrees) 14000
FT
B fT~'
-0.25 -0.14 +0.44
o o +0.64
,/ "wc5 =I
'I
",:,:
ANHY RITE
+0.25 +0.15 +0.83 '1;
+1.25 +0.84 +1.57 SECOND
ASSEMBLY

~ I
* Positive means upward .
.. Positive means building up the angle.
15000
FT
\ PREDICTED

HOLE ANGLE
(DEGREES)
WOB
(THOUSAND LB)
.
~ MEASURED

o 1 2 3 4 o 10 2030 40 ASSEMBLY FORMATION Fig. 8-Directional-hole field case.

I
t
PEND.
I 1
I
---r 1
m], the hole angle increased from 1 to 3 [0.02 to 0.05
SHA LE rad], despite a reduced WOB.
... From this information and the knowledge that the dip
r~ PACKED angle was about 25 [0.44 rad], the formation class of

~
b the new formation was determined. The combination of
I dip angle and formation class resulted in very crooked
PEND. I hole conditions. The decision was made to reduce the hole
r ~-t HA RO
SANDS TONE
angle by running a pendulum assembly optimized by the
program. This was done between 4,800 and 5,000 ft [1463
-
r, 1--1--

I i
PACKED
w/SQUARE
!
I
I
~
and 1524 m]. As can be seen, the predicted and meas-
ured results agreed very well.
Next, BHAP was used to analyze feasible options, which
\ PREDICTED included using several conventional packed assemblies,
".MEASURED drilling smaller hole followed by hole opening, and using
a square DC. A square collar was determined to be nec-
Fig. 7-Straight-hole field case.
essary.
The only 17.5-in. [44.5-cm] square collar available was
worn to 16.5 in. [41.9 cm]. Initially, this seemed unac-
A bent sub may not contact the hole at the bend. The ceptable. However, the program evaluation of stabilized
above is a no-contact case. assemblies containing the worn square led to an assembly
Decisions pertaining to the bent-sub angle to be used whose predicted rate of angle buildup was acceptable. Evi-
are made according to the desired rate of buildup or drop- dently, the stiffness and weight of the worn square more
off (hole curvature). than compensated for its lack of wall contact.
The assembly was put in at 5,000 ft [1524 m]. Fig. 7
Field Cases shows that the predicted performance matched the actual
We have chosen one straight-hole case and one directional results.
case.
Directional-Hole Case. This well was drilled in Mobile
Straight-Hole Case. This well was drilled in Grady Bay, AL. Results are shown in Fig. 8. The interval un-
County, OK, without benefit of a nearby offset well. The der consideration, for which there was no formation
drilling contract called for a 4 [0.07-rad] maximum hole change, is from 13,600 to 15,100 ft [4145 to 4602 m].
angle to the end ofthe 17 .5-in. [44.45-cm] hole at 7,000 The assembly run between 13,600 and 14,000 ft [4145
ft [2100 m]. From the surface to 4,000 ft [1219 m], a and 4267 m] was analyzed by the program, with known
pendulum assembly-optimized by the program-was run, drilling data and the knowledge that the dip was flat (near
allowing 35,000-lbf [156-kN] WOB while keeping the 0 [0 rad]). Despite the operator's hope that the assem-
hole angle to around 1 [0.02 rad]. Fig. 7 shows a plot bly would hold angle, BHAP correctly predicted that it
of hole angle and WOB vs. depth from 3,500 to 5,600 would drop, as seen between Points A and B.
ft [1067 to 1707 m]. The solid lines are program results; At 14,000 ft [4267 m], the assembly was changed to
the circles are survey data. the one designed by the program to hold angle with
At 4,000 ft[1219 m], the pendulum assembly was re- 40,000-lbf [178-kN] WOB. Between 14,000 and 14,200
placed by a packed assembly. This was done in anticipa- ft [4267 and 4328 m], a WOB of 50,000 lbf [222 kN]
tion of a formation change. Prediction of the performance was run instead of the design value of 40,000 lbf [178
of this assembly in the next formation was not possible kN]. The hole began dropping angle. To counteract this,
at this time, because no drilling data were available for the operator increased the WOB even more, to 60,000
that formation. The formation change occurred at 4,500 lbf [267 kN]. Unexpectedly, the drop became even faster
ft [1372 m]. Between 4,500 and 4,700 ft [1372 and 1433. (from 14,200 to 14,500 ft [4328 to 4420 m]).
40 SPE Drilling Engineering, March 1987
 When we were contacted, we investigated the situation D II = component of displacement in direction
by using BHAP. The results were that an increase ofWOB parallel to bedding planes
should, in fact, decrease the angle; the converse is also E = Young's modulus
true. These results are shown in Fig. 8 as Lines BC and fo = reduced value F 0 (Eq. A-5)
CD. Thus, our recommendation was to decrease the WOB F B = magnitude of force of bit on formation
to the design value of 40,000 lbf [178 kN] even though
(Axis OF)
such a procedure contradicted visualization.
Fe = formation class (Eq. A-3)
Conditions encountered here were the same as described
in parametric studies (Case 3). The WOB was decreased F 0 = Component OY of reaction of section of
to 40,000 lbf [178 kN] and, as Fig. 8 shows, the hole . assembly below origin, 0, on section
angle held at about 19 [0.33 rad].0
above
F T = component of force in direction
Data Reliability perpendicular to bedding planes
The main limitation of using any computer model of F II = component of force in direction parallel to
BHA's is the reliability of input data. In particular, three bedding planes
parameters-hole curvature, dip angle, and stabilizer I = moment of inertia with respect to diameter
clearance-are difficult to obtain accurately and have a I fani = formation anisotropy index (Eq. A-2);
strong effect on the results. denoted "h" in Ref. 19
Hole-angle errors are Ys [O.002 rad]. Hole curva-
0
K = reduced compression (Eq. 2C of Ref. 24)
ture is calculated from the differences of hole angle, which L = length of a section
results in large hole-curvature errors. o = origin of coordinates
Except in broken formations, dip logs are generally
q = reduced value of Q sin {3 (lateral load per
good enough for our purposes. Dip angles may also be
obtained from geologic or seismic data. In pendulum as- unit length), Eq. 4tc of Ref. 24
semblies, stabilizer clearance is unimportant. Unfortunate- Q = weight per unit length
ly, the contrary holds true for packed assemblies. In spite R = reaction of stabilizer on collar
of input errors, the output of the program provides useful W = weight on bit
information in almost all cases. X = abscissa
Y = ordinate
Conclusions a 0 = angle between hole axis and elastic line at
1. A computer program has been developed to predict the origin, 0 (Fig. A-5)
the performance of, and thus allow optimization of, any {3 = angle between hole axis and vertical at the
BHA. The effects offormation anisotropy and formation origin, 0 (Fig. A-5); hole angle at bit
dip angle are included. (Fig. A-I)
2. Knowledge of the dip angle and predictions of its l' = formation dip angle
effects on assembly performance are essential in making o = angle between bit force and vertical (Figs.
the correct field decision on BHA design and WOB.
A-I and A-5); denoted "ell" in Ref. 19
3. Sensitivities to changes of the WOB, formation
change, and stabilizer clearance are essential in choosing n = angular discontinuity
the best assembly.
4. Field cases in which BHAP predictions match actual
performance have been presented.
Subscripts
5. Satisfactory performance of a pendulum assembly
L = end of section
may require the use of two stabilizers.
6. In some cases, increasing the WOB may result in o = start of section
dropping hole angle. 1,2,3 = Sections 1, 2, 3
7. Stabilizer clearance is extremely important in the per-
formance of packed assemblies but not pendulum as-
semblies. Acknowledgments
Nomenclature * Eugene Usdin, Southwestern Computing Service, Tulsa,
contributed to a major extent in both mathematical and
a = radial hole-to-collar clearance
computational fields. James May, Leguin Williams,
Ch = hole curvature . George McKown, John McCormick, and Al Odell, all
Co = curvature of elastic line at origin with Sii Drilco, contributed in gathering and/or processing
Co = Co +q (Eq. 6c of Ref. 24) field data. J.P. Bernhard, Regie Autonome des Petroles,
C = compression, assumed constant in a section Paris, supplied Eq. A_2.I9
DB = magnitude of displacement of bit in Moreover, this work could not have been successful
direction of drilling (Axis OD) without the invaluable cooperation of the many operators
D T = component of displacement in direction and contractors to whom we express our great ap-
perpendicular to bedding planes preciation.
Finally, we thank the management of Smith Drilling
Symbols used here should not be confused with those that usually represent depth
or length in standard SPE nomenclature; these variables represent displacements
Systems, Div. of Smith IntI. Inc., for permission to pub-
and are not units of length. lish this paper.
SPE Drilling Engineering, March 1987 41
References Murphey, C.E. and Cheatham, 1.B. Jr.: "Hole Deviation and Drill String
Behavior," SPEJ (March 1965) 44-54; Trans., AIME, 237.
1. McLamore, R.T.: "The Role and Rock Strength Anisotropy in
Natural Hole Deviation," JPT (Nov. 1971) 1313-21; Trans.,
AIME,251. Appendix
2. Timofeev, N.S. et al.: "Predicting the Rate of Well Curvature,"
Burenie (1973) No. 10, 8-9. Bit/Formation Interaction. Formation Anisotropy. The
3. Bradley, W.B.: "Deviation Forces from Wedge Penetration Failure bit/formation interaction used is the same as in Ref. 19.
of Anisotropic Rock," J. Eng. Ind.; Trans., AS ME (1974) 95, A clearer presentation of the topic will be provided.
Series B, No.4, 1093. Laboratory experiments have shown that the hole be-
4. Smith, M.B. and Cheatham, 1.B. lr.: "Deviation Forces Arising
From Single Bit Tooth Indentation of an Anisotropic Porous Media,"
comes inclined by repeated fracturing of the layers of the
J. Pressure Vessel Tech.; Trans., ASME (May 1977) 362. formation in the direction perpendicular to the bedding
5. Millheim, K.K. and Warren, T.M.: "Side Cutting Characteristics planes. In other words, drilling is easier in the direction
of Rock Bits and Stabilizers While Drilling," paper SPE 7518 perpendicular rather than parallel to the bedding planes. 22
presented at the 1978 SPE Annual Technical Conference and In Fig. A-I, dashed lines represent bedding planes, Axis
Exhibition, Houston, Oct. 1-3.
6. Baird, 1.A. et al.: "GEODYN2: A Bottomhole Assembly/ OP L is parallel to bedding planes, and Axis OPR is per-
Geological Formation Dynamic Interaction Computer Program," pendicular to bedding planes. The force with which the
paper SPE 14328 presented at the 1985 SPE Annual Technical bit acts on the formation is shown as a vector from Points
Conference and Exhibition, Las Vegas, Sept. 22-25.
7. Amara, M.H.: "Use of Drillstring Models and Data Bases for the
o to F, whose magnitude is denoted F B. F t t is the
Scientific Control of Vertical and Directional Hole Paths," paper
projection of the force in Direction OP L. Similarly, F T
SPE 13495 presented at the 1985 IADC/SPE Drilling Conference, is the projection of the force in Direction OP R.
New Orleans, March 5-9. Displacement of the bit as it drills ahead may also be
8. Nicholson, R.: "Analysis of Constrained Directional Drilling taken as having components parallel and perpendicular
Assemblies," PhD dissertation, U. of Tulsa, Tulsa, OK (1972). to the bedding planes. D t I is the displacement compo-
9. Walker, B.H.: "Some Technical and Economical Aspects of
Stabilizer Placement," JPT (June 1973) 663-72. nent parallel to bedding planes, while DT is the displace-
10. Fischer, F.J.: "Analysis of Drill Strings in Curved Boreholes," ment component perpendicular to bedding planes.
paper SPE 5071 presented at the 1974 SPE Annual Technical The scales for forces and displacements used in Fig.
Conference and Exhibition, Houston, Oct. 6-9. A-I are chosen such that D II is the same magnitude as
11. Wolfson, L.: "Three-Dimensional Analysis of Constrained
F 11 . Because drilling in Direction OPR is easier than in
Directional Drilling Assemblies in a Curved Hole," MS thesis, U.
of Texas, Austin (1974). Direction OPL, DT is greater than FT' When the dis-
12. Walker, B.H. and Friedman, M.B.: "Three Dimensional Force placement components are summed vectorially, the re-
and Deflection Analysis of a Variable Cross Section Drill String, " sult is the vector shown from Points 0 to D whose
J. Pressure Vessel Tech. (May 1977) 367. magnitude is denoted DB' Thus the direction in which
13. Millheim, K.K., Jordan, S., and Ritter, C.l.: "Bottomhole
Assembly Analysis Using the Finite-Elements Method," JPT(Feb.
we drill is different from the direction of the force on bit.
1978) 265-74. Axis OV is a vertical. Angles (), {3, and l' are angles
14. Callas, N.P. and Callas, R.L.: "Boundary Value Problem is of force on bit, direction of drilling, and OP R, with
Solved," Oil & Gas J. (Dec. 1980) 62. respect to vertical. l' is, of course, the formation dip an-
15. Millheim, K.K. and Apostal, M.C.: "The Effect of Bottomhole
Assembly Dynamics on the Trajectory of a Bit," JPT(Dec. 1981)
gle, and {3 is the hole angle at the bit.
2323-38.
16. Millheim, K.K. and Apostal, M.: "Bottomhole Assembly Dynamics It is evident that
and Its Effect on the Bit's Trajectory, " paper presented at the 1981 DT = DB cos (-Y-{3),
IADC/Cdn. Assn. of Oilwell Drilling Contractors Drilling Tech- D II = DB sin (1' - (3),
nology Conference, Calgary; Trans., IADC, Houston, and Cdn.
Assn. of Oilwell Drilling Contractors, Calgary (1981) 89-108. F T = F B cos ('Y-(), and
17. Toutain, P.: "Analyzing Drill String Behavior," World Oil (July FII = FB sin ('Y-().
1981) 221.
18. Enen, J., Callas, N.P., and Sullivan, W.: "Rig Sit Computer
Optimizes Bit Weight," Oil & Gas J, (Feb. 1984) 68.
Therefore
19. Lubinski, A. and Woods, H.B.: "Factors Affecting the Angle of
Inclination and Dog-Legging in Rotary Bore Holes," Drill. and DIIIDT tan('Y-{3)
Prod. Prac., API (1953) 222. .......... (A-I)
20. Woods, H.B. and Lubinski, A.: "Practical Charts for Solving Prob- F 11 IF T tan('Y-()
lems on Hole Deviation," Drill. and Prod. Prac., API (1954) 56.
21. Lubinski, A. and Blenkarn, K.A.: "Usefulness of Dip Information
in Drilling Crooked Formations," The Drilling Contractor (April which is a fraction denoted by I - I fan;, where I fan; is
1956) 53-56. called the "anisotropy index" of the formation.
22. Rollins, H.M.: "Are 3 and 5 Straight Holes Worth Their Cost?, "
Oil & Gas J. (Nov. 1959). tan('Y-{3)
23. Rollins, M.: "Drilling Straight Holes in Crooked Hole Country," - - - = I - Ifan ; . ..................... (A-2)
Sii Drilco, Div. of Smith IntI. Inc., Houston (1956) (updated tan('Y-()
regularly) .
24. Lubinski, A.: "Maximum Permissible Dog-Legs in Rotary Bore-
holes," JPT (Feb. 1961) 175-94; Trans., AIME, 222. In the computer program, a so-called formation class,
F c, is used instead of I fan; .
General References
Bradley, W.B.: "Factors Affecting the Control of Borehole Angle in F C= I33.97Ifani 0.5. . . . . . . . . . . . . . . . . . . . . . (A-3)
Straight and Directional Wells," JPT (June 1975) 679-88.

Daring, D.W.: "Drilling Directional Holes Having Constant Curva- For Ifani=O, we have {3=()-Le., in Fig. A-I, DT=FT,
ture," paper SPE 3508 presented at the 1971 SPE Annual Meeting, which means that the formation is isotropic, such as
New Orleans, Oct. 3-6. granite. For I fani > I,(DII IF 11 )/(Dr/F T) is negative.
42 SPE Drilling Engineering, March 1987
 ..J

u
~
w
>

(j
~
i=
~--'a:
w
-_--.I>

v v
Fig. A-1-Bitlformation-interaction hole angle smaller than Fig. A-2-Bitlformation-interaction hole angle greater than
dip angle. dip angle.

Because the denominator is positive, the numerator garding anisotropy index as a function of only the forma-
would have to be negative, which means that the compo- tion, extensive field use also indicates that the three-cone
nent of bit displacement in the direction parallel to bed- bit anisotropy is small compared with formation
ding planes would have to be in the direction opposite to anisotropy.
the component of force in that direction, which is absurd. Here the program also assumes zero bit anisotropy and
Thus, 0:5, I fani < l. determines formation anisotropy, Ifani' from drilling data.
In Fig. A-I, /3 < ,,(, which is frequently the case for Use of I fani in solving drilling problems yields results that
straight-hole drilling in crooked-hole formations. In direc- are in agreement with field results. This holds true if I fani
tional drilling, frequently /3>"(. Fig. A-2 pertains to such is determined and then used in three-cone bits or in drag
a case. By following the same steps as for Fig. A-I, one bits. It does not hold true if Ifani is obtained from use of
may prove that Eqs. A-I and A-2 remain valid. a three-cone bit and then used in solving problems per-
As mentioned, experiments have shown that in both taining to drag bits.
figures, DT>FT-i.e., the right member in Eq. A-2 is Because of this, we conclude that in most, but not all,
in fact a fraction. On the other hand, that this fraction cases the assumption of zero bit anisotropy is satisfacto-
is a constant-which depends only on the formation-is ry. A new experimental investigation on bit anisotropy,
an assumption. Since 1953, however, this assumption has however, with superior experimental means now available
been used successfully in thousands of wells, indicating could be useful.
that it is satisfactory. 19,20,23
Bit Anisotropy. In 1952, the mathematics to handle both
formation and bit anisotropies was developed, but not in- Sections and Junctions. Hole curvature is assumed con-
cluded in Ref. 19. For an isotropic formation drilled with stant. The BHA is divided into sections. Points between
an anisotropic bit, the direction of drilling would have two successive sections are called junctions and are lo-
to be between the' direction of the force applied by the cated at each stabilizer, at each change of cross section
bit and the direction in which the bit is pointing. or stiffness, and at any additional point where a change
If bit anisotropy were small compared with formation in compression is wanted.
anisotropy, however, the effect of tilt angle would be In each section, the compression is constant. At each
negligible. junction, the compression (and therefore the bending mo-
Extensive, expensive experiments were conducted in ment) is changed in a discontinuous manner. The coor-
1952 to determine anisotropy indices of various three-cone dinate axes are changed from section to section. The origin
bits. The project did not produce constant, usable results. is at the lower end of a section and is on the axis of the
Consequently, the results were not included in Ref. 19 hole. The X axis is parallel to the hole axis (Figs. A-3
and bits were assumed to be isotropic. As mentioned, re- and A-4).
SPE Drilling Engineering, March 1987 43
 '\',
\ ' I ,......
!'; .....

\ ' I \ ........
Vc'(\"'~
\
<:.", "
i \ ~"' . . . . . . .
~v'(\~:
...
...

\ I Cn4 \ .....
\ -{'!>

Fig. A-4-Condltions at an active stabilizer and the point

Fig. A-3-Cross-sectional change. of tangency.

There are two types of junctions-A and B. At a B junc- Conditions at the Bit (Fig. A-5). At the bit, Y01 is
tion, the Y coordinate of the elastic line is not known. Such assumed to be zero and COl = W. The bending moment
is the case when there is a change of cross section at the is assumed to be zero; thus COl =0 and Cal =q. F 01 is
junction (Fig. A-3), or a change of stiffness, or a point the 0 I Y I component of the reaction ofthe bit on the as-
where a change in the compression or an inactive stabilizer sembly.
is wanted. "Inactive" means that because of clearance From Fig. A-S,
there is no contact on either the low or high side of the
hole. F 01 =- CI tan(~ -8), ................... (A-6)
At an A junction, the Y coordinate of the elastic line
is known, which is the case for an active stabilizer (Fig. in which the sign is "minus" because both ~ I and 8 are
A-4). negative. 8 is obtained from Eq. A-2. Thus all the pa-
rameters in Eq. A-4 are known except tan a 0 .
Differential Equation and Solution. The differential
equation (Eq. It of Ref. 24) remains valid, except - K2 Y Solving a Problem. For the sake of explicitness, consider
becomes + K2 Y because of compression instead of ten- that there are only two junctions and that Junction 1-2
sion and the shear, So, is replaced by F o. The follow- (i.e., the junction between Sections 1 and 2) is a B junc-
ing may easily be checked as a solution of the differential
tion (Fig. A-3) and that Junction 2-3 is an A junction (Fig.
equation. A-4). BHAP may handle any number of A and B junctions
in any order.
Y=Y o + ;2 [co(l-cos KX)+lo(KX-sin KX)
The.input parameters were listed under Input in the body
of the paper. Anyone of these parameters may be the
unknown of the problem. For the sake of explicitness,
consider the WOB, W, as the unknown. An arbitrary value
-f(KX)2+tan ao sin KX] ........ (A-4) is assigned: W = COl' Another arbitrary value is assigned
to tan ad. Therefore the problem is started with two ar-
and bitrarily assigned values. With Eq. A-4 and its deriva-
tives, the following are obtained: YL1' Yit, and YZ I
1 Fo From these-and taking into consideration rotation of
10= K EI' ........................... (A-S) hole angle and coordinate axes by ChLI (Fig. A-3), dis-
continuity of the compression, C, and therefore of the
bending moment, and equilibrium conditions-the follow-
where F 0 is Component OY of the reaction of the por- ing are obtained: Y 02 , tan a02, K2, C02, C02, F 02 ,
tion of assembly below origin 0 on the portion above. 102, q2, and ~2' Therefore all the parameters in Eq. A-4
Eq. A-4 is written in a forin that satisfies the boundary pertaining to Section 2 are known. No new unknown has
conditions at the origin-i.e., for X=O, been introduced. Such is always the case in a B junction.
At Junction 2-3 (Fig. A-4), one proceeds exactly as in
Y= Yo, Junction 1-2. This is an A Junction, however, for which
Y' = tan ao, and the actual value of Y 03 is known. Generally, the calcu-
Y" = Co -q=co (from Eq. 6c of Ref. 24). lated and actual value of Y 03 are not the same. Therefore
44 SPE Drilling Engineering. March 1987
 Fig. A-5-Conditions at the bit.

a new value of tan ex 01 is tried at the start of Section 1,

and the process is repeated until the correct value of
tan ex 01 is found for which the calculated and actual
values of tan ex 03 are equal.
At Junction 1-2, F 02 has been calculated. On the other
hand, at Junction 2-3 (Fig. A-4)-i.e., at an A junction-
F 03 cannot be calculated. An arbitrary value of F 03 (and
the corresponding f 03) is assigned. Thus, the initially
unknown value of tan ex 01 has been resolved, but a new Fig. A-6-Keeping BHA inside the hole.
value, f 03, became a new unknown.
As before, for tan ex OJ , the value of f 03 will be re-
solved at the next active stabilizer up the assembly or at Then one must restart, assuming contact on the high
the point of tangency (Point T in Fig. A-4), whichever side of the hole. If R2-3 is negative, the new solution is
comes first. valid. If R 2 - 3 is positive, the stabilizer does not contact
At each active stabilizer encountered, the preceding the hole on either side, and the problem must be restarted,
value of f 0 is resolved, while a new value of the next assuming a Junction B at R 2 - 3 . If there are five stabiliz-
fo appears. ers, there are 3 5 , or 243 possibilities, one of which is
correct.
Point of Tangency. In Fig. A-4, Section 3 ends at the
point of tangency, T. The length, L 3, is unknown. Thus False Stabilizers. Until now, we were not concerned with
we arrive at T with three unknowns-the original whether the assembly was inside the hole. The solution
unknown, in our case W = C 1; f 0 at the top active obtained so far is invalid if such is not the case, as in Fig.
stabilizer, which in our case is f 03 (or tan ex 01 if there A-6a. Point P is the point of maximum excursion. The
were no stabilizers); and L3 (Fig. A-4)-to which ar- next step is to place a stabilizer, called a false stabilizer,
bitrary values have been assigned. at Point P, with a hole-to-collar radial clearance, Q. The
At Point T, values of Y L3, Yb, and YZ 3 are calculat- result is shown in Fig. A-6b. Generally, the BHA is still
ed, then coordinate axes are rotated as if for the start of not entirely inside the hole. The process is repeated until
a nonexistent Section 4. The following three conditions either a satisfactory solution is obtained or no solution
must be satisfied at Point T: Y 04 = -Q, Y 04 =0, and is possible, which means that the portion of the assembly
Y 04 = C h. Because there are three unknowns and three above Point P is decoupled from the portion below. Point
equations, the problem is numerically solved. P is the point of tangency-i.e., the portion of the assem-
Next, the reaction, R 2 -3 , of the active stabilizer (Fig. bly above Point P has no influence on drilling.
A-4) on the assembly is calculated. As the stabilizer in A false stabilizer is handled by the computer in the same
Fig. A-4 contacts the low side of the hole, R 2-3 must be manner as a real stabilizer, except that coordinate axes
positive. If R 2-3 is negative, the solution of the problem are not rotated and there are no discontinuities of com-
is invalid. pression and bending moment.
SPE Drilling Engineering. March 1987 45
 Bent Sub. If in Fig. A-3, Section 0 1 B were a bent sub This procedure violates the assumption of constant hole
(or motor housing) whose angle at Point B is 0, then an curvature between the bit and the point of tangency. It
angular discontinuity, 0, is used in the equations. The is justified through satisfactory correlation with a predicted
same holds true if, at the bend, the sub contacts the hole hole trajectory for variable curvature, on which we are
wall. currently working.
Predicting Hole Trajectory. Consider Case 1 in the body
of the paper with a 20 [0.35-rad] formation dip. One SI Metric Conversion Factors
of the input values is 0.75/100 ft [0.43 x 10- 3 rad/m].
Consider that it will remain constant for about 50 ft [15 degree x 1.745329 E-02 = rad
. m]. Thus at 3,050 ft [930 m], the hole angles will increase ft X 3.048* E-Ol m
from 3 to 3.375 [0.052 to 0.059 rad]. Consider now the in. x 2.54* E+OO em
same input as the original, but with a 3.375 [0.059-rad] lbf x 4.448 222 E+OO N
hole angle and hole curvature as the unknown. By usin~
the program, we obtain 0.685/100 ft [0.392 x 10- Conversion factor is exact. SPEDE
rad/m]. Consider again that this value will remain con-
stant for another interval of 50 ft [15 m]. This is how the Original manuscript received in the Society of Petroleum Engineers office Feb. 9, 1986.
trajectory of the slick assembly in Fig. 1 has been ob- Paper accepted for publication Aug. 7, 1986. Revised manuscript received Dec. 9,
1986. Paper (SPE 14764) first presented at the 1986 IADC/SPE Drilling Conference
tained. held in Dallas, Feb. 10-1;1.

46 SPE Drilling Engineering, March 1987

Discussion of Predicting Bottomhole
Assembly Performance
Hwa-Shan Ho, NL Technology Systems/MWD R&D
-+
This paper discusses "Predicting Bottomhole Assembly the actual drilling direction to be D da' also lying in the
Performance" (March 1987 SPEDE, Pages 37-46). Over same Plane PI. (Therefore, no walk is predicted.)
the past 35 years, Lubinski has made major technical On the other hand, in a 3D analysis, one considers in-
contributions and has been instrumental in demonstrat- stead the true formation normal vector, jJN, which in this
ing the importance of understanding and quantifying the particular case points toward the front of Plan~ PI. The
deformation of the bottomhole assembly (BHA) to resultant drilling direction will then be Vector D d, which
characterize the deviation tendencies during drilling. This ,*0 points toward the front of ~. Even though Vector
has recently led to the development of several more DNa is the projection of ~ector D N onto Base Plane PI,
sophisticated BHA-analysis programs. 1-4 While I join in the projection of VeEor D d onto Plane P I does not coin-
congratulating Lubinski for his great achievements, I cide with Vector D da' assuming that the same rock-
regret that he and Williamson did not mention these new anisotropy characteristics are used in both 2D and 3D
results in the subject paper. analyses. -+
I concur with the authors about the importance of under- This is further illustrated in Fig. D-3. Vectors D F and
standing the effect of formation/bit interaction on the drill- jJNa define Base Plane PI. With a 2D analysis, one
ing direction. They included only the formation anisotropy would decompose the bit force into the normal and parallel
effect, however, and claimed that bits are isotropic. I components OBa and ABa. Anisotropy of t!!.e formation
respectfully disagree with this conclusion. Recently, I de- would cause the apparent drilling vector, D da' to pass
veloped a general three-dimensional (3D) rock/ bit inter- through Point Ca. The ratio CaBalABa describes the
action model and carried out limited field-data analysis. 5 degree of anisotropy of the formation, which is an
Soft-formation three-cone bits were used in these wells, anisotropy index.
and I found that they are highly anisotropic. In fact, the Likewise, in a 3D analysis, one would obtain the bit
bits are seen to be several times more effective in for- force components OB and AB, and the drilling direction,
ward drilling than in lateral drilling. Brett et al. 3 also im- jJd, would pass through Point C. The ratio CB/ AB is
ply a similar trend (see Model Verification in Ref. 3), again the anisotropy index, and is also the same as
though their model does not address this effect directly. CpBp/ABp (where p denotes the projection onto the base
One possible explanation for Lubinski's 1952 and Wil- plane) caused by parallel projections. We can ~en con-
liamson and Lubinski's 1987 conclusions-that bits are clude that Line C a C p is parallel to Vec..!?r DNa, and
isotropic-is that both studies use only a two-dimensional therefore cannot--+'be parallel to . Vector

D da. In other
--+
(2D) analysis model, which is somewhat limited in capa- words, Vector D d does not project mto Vector D da.
bility. As a result, formation dip is assumed to lie within Therefore, the out-of-plane dipping of an anisotropic
the same vertical plane as the well trajectory. Any effects formation will not only cause the BHA to walk (to the
caused by out-of-plane dipping may be masking the bit left in this particular case), but will also change the
effect. build/drop angle from that predicted by a 2D analysis
In a 2D model, where the entire wellbore and drillstring model with the apparent dip angle. The precise nature of
are assumed to lie in the same vertical plane (Plane PI build/drop depends on the borehole orientation relative
in Fig. D-l), the formation dip (Plane P 2 ) is seen as the to the other directions. In fact, it is conceivable that the
apparent dip and not as the true dip. The apparent dip true drilling direction might have a building tendency
angle, 8 a' is equal to the true dip angle (not shown in while the apparent drilling direction might show a drop-
Fig. D-l) only when the relative strike angle, an of the ping tendency, or vice versa. In anisotropic formations,
dipping angle is 90 [1.6 rad]. Otherwise, the apparent there are only two exceptions to the above conclusion:
dip angle is always smaller than the true dip angle. In the when the relative strike angle, a" is 90 or 0 [1.6 or
extreme case when the relative strike angle, an is zero, Orad].
the apparent dip angle is always zero, even when the true 1. If a r is 90 [1.6 rad], then the 2D and 3D analyses
dip angle is 90 [1.6 rad]! coincide. A subsidiary case occurs when the true dip angle
A quantitative treatment of the full 3D formation effect is zero. Then the strike direction of the bedding normal
is best handled by vector algebra. 5 Furthermore, a full is arbitrary, and can be set to 90 [1.6 rad].
3D BHA-analysis program is needed for such analysis, 2. If a r is zero, then formation anisotropy causes only
allowing the bit-force vector and the bit-pointing vector walk deviation but no build/drop deviation.
to deviate from the presumed common vertical plane, PI. It is perhaps this small but significant difference that
The purpose of this discussion, however, is to demonstrate is masking the bit anisotropy effect in the authors'
such a difference in a qualitative-.manner. Therefore, in analysis.
Fig. D-2, the bit-force vector, D F, is set to lie in the An additional factor could be the lack of precision of
same vertical plane, PI, as the well trajectory. field data, which, as the authors pointed out, could sig-
In a 2D formation-effect analysis, one would assume nificimtly alter the analysis result. I have also experienced
the formation-normal vector to be jJNa' lying in the same this effect in my analysis. It is imperative for the industry
vertical plane, PI. The result is that one would conclude to recognize the need for improved and more timely field-
SPE Drilling Engineering, September 1987 283
 HORIZONTAL
(STRIKE)

/HORIZONTAL
2-D WELLBORE (STRIKE)
~

FORMATION
DIPPING PLANE
P2

VERTICAL PLANE P1
WELLBORE PLANE

Fig. D-l-Apparent dip.

data measurement and recording. Only then will it be pos-

sible to exploit the full potential of more sophisticated and Fig. D-2-Apparent and true drilling directions.
rationally based drilling-analysis models to control the
drilling process better, to improve drilling economy, and
to predict and to help avoid potential drilling troubles.
In this regard, the recent development of various measure-
ment-while-drilling tools provides a major impetus toward
the development of improved drilling-analysis-model
programs.
It should be pointed out that when the effects of both
bit and rock anisotropy are considered simultaneously,
concepts of various drillabilities need to be redefined. In
addition, the process of deducing the rock and bit anisotro-
pies from the known bit-force, bit-tilt, and formation-
normal directions is a significant, nontrivial problem. 5

Nomenclature
->
D d = drilling direction vector
->
D F = bit-force direction vector
->
D N = bedding-normal direction vector
PI = vertical plane
P 2 = formation dipping plane
a r = relative strike angle o A
8 a = apparent dip angle
Fig, D-3-Projected and apparent vectors.
Subscripts
a = apparent 3. Brett, J. F. et al.: "A Method of Modeling .the Directional Behavior
p = projected of Bottomhole Assemblies Including Those With Bent Subs and
Downhole Motors." paper SPE 14767 presented at the 1986
References IADC/SPE Drilling Conference, Dallas, Feb. 9-12.
4. Ho, H.-S.: "General Formulation of Drillstring Under Large
I. Rafie, S., Ho, H.-S., and Chandra, U.: "Applications of a BHA Deformation and Its Use in BHA Analysis," paper SPE 15562
Analysis Program in Directional Drilling," paper SPE 14765 presented at the 1986 SPE Annual Technical Conference and
presented at the 1986 IADC/SPE Drilling Conference, Dallas, Feb. Exhibition, New Orleans, Oct. 5-8.
9-12. 5. Ho, H.-S.: "Prediction of Drilling Trajectory in Directional Wells
2. Jogi, P.N., Burgess, T.M., and Bowling, J.P.: "Three-Dimensional Via A New Rock-Bit Interaction Model," paper 16658 available
Bottomhole Assembly Model Improves Directional Drilling," paper at SPE Headquarters, Richardson, TX.
SPE 14768 presented at the 1986 IADC/SPE Drilling Conference,
Dallas, Feb. 9-12. (SPE 17015) SPEDE
284 SPE Drilling Engineering, September 1987
 Authors' Reply to Discussion of Predicting
Bottomhole Assembly Performance
J.S. Williamson, SPE, Smith Drilling Systems, Div. of Smith IntI. Inc.
Arthur Lubinski, SPE, Consultant

We appreciate Ho's discussion of our paper and reply as "effective" dip angle in a fashion similar to Ho's detailed
follows. discussion of "apparent" dip angle.
Not referencing papers presented at the same conference Ho argues that bits are not isotropic. As we stated, we
as our original paper was an oversight. We agree that these have seen evidence of this when comparing three-cone
should have been added for SPEDE publication. bits to drag bits. However, more data are required to allow
We are pleased that Ho agrees that including a bit/for- quantifying bit anisotropy. Our experience is that the as-
mation interaction is essential in predicting bottornhole sumption of zero bit anisotropy leads to satisfactory re-
assembly performance for both two- and three-dimen- sults in most, but not all, cases. Ho alludes to having data
sional analyses. on bit anisotropy but does not present them in his discus-
Ho also points out that the well trajectory may not be sion. We look forward to the publication of such data in
in the same plane as the strike of the dip. We certainly the future.
agree, especially in directional drilling. Though not stated
in our paper, we take this into account through use of the (SPE 17075) SPEDE

SPE Drilling Engineering, September 1987 285