3L6+

4!34-/
Directional Survey by the Circular Arc Method
STANDARD OIL CO. OF CALIFORNIA, WI
W. A. ZAREMBA* LA HABRA, CAL IF.
I

ABSTRACT angle and the depth data are also subject to

measurement inaccuracies due to small instrumental
A method for computing directional surveys is errors; thus, other things being equal, the method
presented that eliminates the abrupt changes in minimizing the effect of those errors is preferable
direction of the tangent to wellbore. For any well provided that we can be reasonably sure it makes
course not located in a single plane the twist o~ no unrealistic assumptions and, as a practical tool,
tb e representing curve will still experience is not too costly to use.
,?ie~,,-+;-.,;+;o. &~ :,he o{!oe+ m“., ho OY~OCfOd t=
L4.ecu,’..u.u’. LC.J,
-//-’-. J,. -, “.- -zr -------- The coinmrirdy ...--2USCU
-.. A =llIlpLC-.
cUIU
-: -..1---- .-- 1,..:/.- is. ~~
=V.LI.IVLL

be small in practical cases. Ease o{ interpolation take the well course between consecutive stations
between the computed location of station points is as a series of straight segments —each one inclined
retained’, since the coora’inates 0/’ any intermea’iaie and directed as measured at its deeper end. Tiiis
point are explicitly expressible in terms of the so-called ‘ ‘tangent projection” suffers from an
measured depth increment to such point, while the obvious defect that all changes of direction are
in-bole depth needed to reach any given horizon is assumed to be concentrated at the survey points,
..-:,...,.1., ,?”,,..-”, J,J /.-fi.,; JnA +Lnt ~~e ~~:~~r ~~ ~:
ur~, yucty c4c Lcr IILLrtt=u, yrvv. wcu .r. w. whereas we can expe~[ . IIO1n L–-— plly
-L.. SICiil
-:1-1 LULl>1UWaL1U112
__--:J -. -.:-.. -

an intermediate depth to that of some two such as drill pipe rigidity, etc. , that those changes
consecutive station points. The formulas are are actually distributed over a considerable portion
somewhat tedious for any extended routine band of each segment.
calculation, but a digital computer program can be It has been proposed by Wilsonl to interpolate
constructed to carry it out, and a copy of one written between consecutive stations by segments of a
in ALGOL 60 is included here. spat e curve formed from a circular arc wrapped
around on a vertical right circular cylinder, the two
INTRODUCTION radii (arc and cylinder) being chosen so as to match
Accurate directional surveys are required for the the measured direction of the hole at two respective
correlation of bottom-hole location with the survey points, while the length of curve is equal to
.L– alllercnce
me
J:rz ______ ul -L mcta>uLcu
—A-- ... . ..l UCPLU=I.
J-... L.- l-a.:.-
LUI=
q~.”A:..e
~aulu=
-$
WL
geological data, for avoiding interference with
existing wells while drilling, and for possible curvature” method removes abrupt tangent vector
assessment of property rights in jurisdictional changes, but the intrinsic shape of the interpolating
curve is, in general, a rather complex space helix
disputes. Though the well course is ultimately
with the total curvature varying from point to point.
determined on the basis of simultaneous measure-
In this article, we want to point out that it is
ments of hole direction and depth at a number of
possible to retain the tangent continuity across
survey points, the true shape between those
survey points more simply by fitting true circular
“stations” is unknown; hence, some reasonable
arcs between them. Each arc is, in general, located
assumptions must be made as to its nature. Thus,
on a skew plane whose orientation can be easily
various interpretations of the same data may lead
to different computed locations of the same point, determined from the known inclination and direction
and the question of which one to choose becomes angles at each end, while the radius of arc follows
important. Unfortunately, detailed comparisons of from the requirement that the developed length of
the effects of various possible assumptions on the curve be the same as the measured separation of
predicted positions vs those known from closely relevant station points. For any well located in a
vertical plane, both Wilson’s approach and the
spaced and accurate surveys seem to be lacking;
... . . C.G ~~f~~-~~~ ~fISWC~ can be given.
hence
... current method give identical results; but if there
Ouite apart from shape uncertainty, both the is an out-of-plane component the computed locations
wi 11 diverge.
Ori-inel msafi,,scdnt
-.. e --------------- . -r . .ree.=ived
. _________ in —__
So.-ietv
--., Qf PetrQ!ewn En-ineers
~-. .--. — It must be stressed here that neither method
office Tune 25. 1971. Revised manuscript of SPE 3664 received
Aug. 3;, 1972.’ @ Copyright 1973 Amer;can Institute of Mining,
removes disco ntinuities in the twist of the
Metallurgical, and Petroleum Engineers, Inc. constructed space curve, while both retain ease of
preferences given at end of PaPer.
interpolating for the locations between station
*Now with Bechtel, Inc., San Francisco.
This paper will be printed in Transactions volume 255, which
points, since the measured and the vertical depths
will cover 1973. are explicitly solvable functions of each other.

FEBRUARY, 1973 SPE7 5

This advantage would be lost if some more EXAMPLES
complicated approach were taken, such as the use
of segments of a spiral helix or the minimum Consider Fig. I showing a well defined by five
------L&J &tAulc
cllc - .L; Sp.:nes.
1“ survey points PI through P5 . The example is not
Yet another interesting approach to the problem intended to be reaiistic, but was constructed sc
has been taken by Walstrom et al? Its distinguishing that the locations of points can be verified easily
feature is that tine coofdifl~te ifiCi&tl~tltS kMtl Or%? by elementary geometry. The input triples:
station to the next are obtained by applying some measured depth, inclination, and azimuth for aii
postulated averaging law of variation to the change points except P3 can be read off from Fig. 1, while
of inclination and of azimuth over a length of for P3 we get, after a simple calculation,
segment between these stations. Five different
methods of averaging are considered, some applying
to the angles themselves and some to their
functions, each method defining, in effect, the 93 = arctiue (C) = 59.73 “
shape of interpolating curve which in one case
turns out to be identical to that introduced by the Locations are required for all stations P, for all
“radius of curvature” method. 1 In general, the intermediate points at 300-ft measured depth
authors demonstrate that the results yielded by intervals, and for all horizons spaced 200 ft apart.
~b:~~ ~ifferenr rn.~dels are in close agreement. The program, whose major part is the algorithm
On the other hand, in this paper it is the shape iisted in the Appendix B, tabUiatt?S ~CSUiCS fQ!

of interpolating segment that is postulated input stations P and, depending on the option, for
(circular arc), and the variations of inclination and either set of intermediate points as shown in the
Of azimuth over its length are constrained to be two sections of Table 1. (To save space entries
consistent with this shape. beyond swim F’4 have h.=em
-+b.. omitted.
..-------- )
NumericaI examples showing application of the For comparison, the bottom-hole locations
method and a comparison of bottom-hole locations computed by this and by two other alternative
derived by this and by other procedures in use for methods are given below.
computing directional surveys are presented, Survey Method Ft-North Ft-East Ft-Vertical
followed by a short discussion of assumptions, Circular arc 2,353.33 745.79 2,880.65
resu. 1cs and experiences arising out of the current
Rad of curv* 2,432.36 699.92 2,821.47
implementation. Derivation of formulas is given in
Tangential* 2,267.78 707.11 2,767.78
the Appendix A, and a listing of a computer
program realizing the described procedure is given
*Radius of curvature method according to Ref. 1. Tangential
in Appendix B.
method modified to assign sum of half-distances to the adjacent
stations on either side of the considered one as the length of
tangent. Direct tangential method returns identical bottom-hole
coordinates when the kickoff point is assumed at a depth of .500 ft.

TABLE 1 — DIRECTIONAL SURVEY RESULTS FOR THE

EXAMPLE IN FIG. 1
Me~~g~=~ Latitude Departure Vertical
Point Depth Inclination Azimuth North East Depth
Number (ft) (ft)
—— (0) ——(0) (f’) _ (ft)
P, 0.00 0.00 0.00 0.00 0.00 0.00
300.00 13.50 0.00 Xi. i S CLoc 297.23
600.00 27.00 0.00 138.78 0.00 578.04
900.00 40.50 0.00 305.06 0.00 826.90
P2 1,000.00 45.00 0.00 372.92 0.00 900.32
1,200.00 45.70 12.62 513.76 15.67 1,041.16
1,500.00 49.21 30.36 717.47 96.91 1,244.85
1,800.00 55.11 45.77 902.13 243.15 1,429.50
P3 2,000.00 60.00 54.73 1,009.57 372.89 J,536.93
2,100.00 57.47 50.37 1,061.49 440.75 1,588.84
2,400.00 50.95 35.77 1,237.48 607.02 1,764.81
2,700.00 46.56 18.75 1,436.05 710.61 1,963.37
P4 3,000.00 45.00 0.00 1,646.22 745.79 2,173.54
P, 0.00 0.00 0.00 0.00 0.00 0.00
200.83 9.04 0.00 15.81 0.00 200.00
406.89 18.31 0.00 64.46 0.00 400.00
624.77 28.11 0.00 150.24 0.00 600.00
865.03 38.93 0.00 282.72 0.00 800.00
?2 1,000.00 45.00 0.00 372.92 0.00 900.32
1,141.26 45.35 8.95 472.61 7.83 1,000.00
1,432.07 48.18 26.53 672.61 7Z60 1,200.00
1,749.15 53.97 43.33 872.63 214.09 1,400.00
Pa 2,Gotl.oo ~~.go <4 7-4
<., ” 1., nf)Q.
.. ---- s7 372,89 1:s36.93
FIG. 1 — EXAMPLE OF CIRCULAR ARC INTERPOLA- 2,120.61 56.97 49.45 1,072.65 454.01 1,600.00
TION OF A DIRECTED HOLE. S- VALUES ARE THE 2,455.26 49.96 32.81 1,272.67 631.03 1,800.00
MEASURED Damws. ...wm~.rvnl
. .. .. . . . .. _AR.E
,,. .Glsl”,u”.m . . uCIR~Z~N~
ATU
2,753.03 46.07 15.52 J,472.68 721.91 2,000.00
SPACED ~V APART. P4 3,000.00 45.00 0.00 i,646.22 745.79 2, J73.54

6 sOCIETY OF PETROLEUM ENGINEERS JOURNAL

 .,
One need not be alarmed by the discrepancies to an existing well of tRe total depth coimpa~ab!e
revealed by the tabulation since they reflect the to that given in the earlier example. Here, the
the nature of the example designed specifically to station coordinates relative to the surface location,
amplify the differences between the methods. In as computed by the arc and by the tangential
practical cases, the number of survey stations is methods, differ very little, the vector distance at
much larger, distances between them much ~tm~ll~i the 60th station —, being in this particular example
and the variation of tangent direction much less, just 14.5 ft for a measured depth of more than 3,000
so that the final results are quite close. This is ft, which corresponds to only about 5 ft divergence
exemplified by data shown in Table 2, which refers per 1,000 ft of measured depth increment.

TABLE 2 - COMPARISON OF DIRECTIONAL SURVEYS BY ARC AND BY TANGENTIAL

METHODS (WELL AT SURFACE LOCATION: 300,515.0 FT N, 2,000,742.8 FT E: CAL IF.
LAMBERT ZONE 6)
Survey Data Arc Method Tangential Method
Latitude Departure Vertical Latitude Departure Vertical
Station Depth Inclination Direction North East Depth North East Depth
Number (ft) (0 rein) (ft) (ft) (ft) (ft) (ft) (ft)
—— —. (0) —. ——
1 0 0 00 N 0.0 E 0.0 0.0 0.0 0.0 0.0 0.0
2 108 000 NO.OE 0.0 0.0 108.0 0.0 0.0 108.0
3 139 1 00 s 59.3 w -0.2 -0.2 139.0 -0.3 -0.4 139.0
4 170 3 00 s 59.3 w -0.7 -1.1 170.0 -1.2 -1.8 170.0
s 202 400 s 60.3 w -1.7 -2.8 201.9 -2.3 -3.7 201.9
233 5 15 S 78.3 W -2.5 -5. 1 232. i? -~eg -6,5 232.7
6
? ~(j~ 6 00 S 83.3 W -3.0 -8.2 2647 -3.2 -9.8 264,6
-)&J ~p~e~
8 296 8 00 N 88.3 W -3.2 -12,0 295.4 -3. ;
9 327 9 00 S 85.3 W -3.3 -16.6 326.1 -3.5 -19.0 325.9
10 358 11 00 S 85.3 W -3.7 -21.9 356.6 -4.0 -24.9 356.3
11 390 11 45 N 89.8 W -4.0 -28.2 388.0 -4.0 -31.4 387.6
12 421 13 45 N 86.8 W -3.8 -35.1 418,2 -3.5 -38.8 417.8
13 452 15 00 S 85.3 W -3.9 -42.7 448.3 -4.2 -46.8 447.7
14 483 J7 00 N 88.8 W -4.1 -51.3 478.1 -4.0 -55.8 477.3
15 515 18 30 S 88.3 W -4.2 -61.0 508.5 -4.3 -66.0 507,7
16 547 20 00 S 85.3 W -4.8 -71.5 538.7 -5.2 -76.9 537.8
17 578 21 45 S 89.8 W -5.2 -82.6 567.7 -5.3 -88.4 566.6
18 609 24 00 S 89.8 W -5.3 -94.6 S96. 3 -5.3 -101.0 594.9
19 640 24 15 S 85.3 W -5.9 -107.3 624.6 -6.4 -113.7 623.1
20 67 i 25 45 cec.-lw
Q (2J..l.! ~&g -120.3 652.7 -7.5 -127.1 651.1
21 703 25 15 s 79.3 w -8.8 -134.0 681.5 -10.1 -140.5 680.0
J log ?;Q*4 -13,5 -154.1 708.8
22 735 26 00 s 75.g ‘w - ! 47.5
23 766 27 00 S 77.8 W -15.0 -160.9 738.1 -16.5 -167.8 736.4
---- - ;g.~ . In’).
24 797 28 00 S 80.3 W -17.7 - I (aou 765.6 ,“--- ? 7~3, a
25 g~~ 30 00 S 80.3 W -20.2 -189.8 792.8 -21.6 -197.5 790.6
26 860 33 00 S 80.3 W -23.1 -206.3 820.0 -24.5 -214.6 zi 7.4
27 892 36 00 S 80.3 W -26.1 -224.1 846.4 -27.7 -233.2 843.3
28 923 38 00 S 79.3 W -29.5 -242.5 871.2 -31.3 -2S1.9 867.8
29 954 40 00 S 77.8 W -33.3 -261.6 895.3 -35.5 -271.4 891.5
30 985 42 4S S 76.3 W -38.0 -281.6 918.5 -40.5 -291.8 914.3
31 1,016 45 00 s 75.3 w -43.3 -302.4 940.9 -46.1 -313.0 936.2
32 1,046 47 4s s 75.3 w -48.8 -323.4 961.6 -51.8 -3345 956.4
33 1,077 50 00 s 74.3 w -54,9 -345.9 981.9 -58.2 -3s7.4 976.3
34 1,108 52 15 S 74.3 W -61.5 -369.1 1,001,4 -64.9 -381.0 99s. 3
35 1,139 54 45 S 74.8 W -68.1 -393.1 1,019.8 -71.5 -405.4 1,013.2
a. 1 ,7, ,
1,t, 5$ so ~ ?~~ w -7S.2 -418.6 1,037.9 -78.8 -431.1 1,030.8
$ 1,203 58 30 S 73.8 W -82.6 -444.5 1,055.1 -86.4 -457.3 i,047.5
38 1,234 60 00 S 74.3 W -90.0 -470.1 1,070.9 -93.7 -483.1 1,063.0
39 1,265 62 00 S 748 W -97.2 -496.3 1,086.0 -100.9 -509.5 1,077.6
40 1,296 64 00 S 74.3 W -104.6 -522.9 1,100.0 -108.4 -536.3 1,091.2
41 1,327 66 00 s 74.3 w -112.2 -549.9 1,113.1 -116.1 -563.6 1,103.8
42 1,359 67 IS S 73.8 W -120.3 -578.2 1,125.8 -124.4 -591.9 1,116.2
43 1,391 68 00 S 73.8 W -J28.6 -606.6 1,138.0 -132.7 -620.4 1,128.2
44 1,422 68 30 S 73.8 W -136.7 -634.2 1,149.5 -140.8 -648.1 1,139.5
45 1,453 68 4s s 73,3 w -144.9 -661.9 1,160.8 -149.1 -675.7 1,150.7
46 1,483 69 00 S 72.8 W -153.3 -689.5 1,172.0 -157.7 -703.4 1,161.9
. ... 69 so
47 1,513 s 72, f3 Y -!~log -7j7*~ 1,1133.0 -166.3 -731.1 1,172.7
48 1,547 70 00 S 72.3 W -171.0 -745.9 1,194.0 -175.5 -759.8 1,183.7
49 1,689 70 Is S 72.3 W -211.7 -873.0 1,242.3 -216.2 -887.0 1,231.6
50 1,814 68 45 S 72.3 W -247.4 -984.5 1,286.1 -251.7 -998.0 1,276.9
1,924 67 15 S 77.8 W -273.8 -1.083.0 1,327.3 -273.2 -1.097.1 1.319.5
:; 2,007 68 00 s 79.3 w -289.1 -1;158.2 1,358.9 -287.6 -1;172.7 1;350.6
53 2,102 68 45 s 79.3 w -305.6 -1,245.0 1,393.9 -304.1 -1,259.7 1,385.0
54 2,221 68 45 S 78.8 W -326.7 - 1;353.8 1;437. 1 -325.8 - );368.5 1;428.1
55 2,344 68 00 S 78.3 W -349.5 -1,465.9 1,482.4 -349.0 -1,480,1 1,474.2
56 2,498 67 45 S 78.3 W -378.6 -1,605.6 1,540.4 -378.0 -1,619.7 1, S32.5
57 2,655 67 45 S 77.3 W -409.4 -J,747.6 1,599.8 -410.1 -1,761.4 1,592.0
58 2,810 67 30 s 75.3 w -443.5 -1,886.8 1,658.8 -446.6 -1,899.9 1,651.3
59 2,924 66 30 s 79.3 w -466.6 -1,989.1 1,703.4 -466.1 -2,002.6 1,696.7
60 3,035 64 4S S 82.3 W -482.9 -2,088.9 1,749.2 -479.6 -2,102.1 1,744.1

FEBRUARY, 1973 7
 CONCLUSIONS REFERENCES

The circular-arc method of directional survey 1. Wilson, G. J.:

({& Impm “ed Method for computing
preserves the continuity of tangent vectors at all Directional Survey s,” j. Pet. Tech (Aug., 1968)
871-876.
station points without introducing artificial
2. Walstrom, J. E., Brown, A. A. and Harvey, R P.:
assumptions about the shape of well center line. CtAn Analysis of Uncertainty in Directional %rveyit)g?”
The bottom-hole location computed by this procedure J. Pet. Tech. (April, 1969) 515-523.
is in fair agreement with those yielded by other
3. Walatrom, J. E., Harvey, R, P. and Eddy, H. D.: “A
methods. When checked on actual surveys, such Comparison of Various Directional Survey Models
differences as do exist are inevitable and reflect and an Approach to Model Error Analysis, ” J. Pet.
the effect of different assumptions about the Tech (Aug., 1972) 935-943.
unknown shape of wellbore. Agreement will, in
general, be closest between the arc and the
APPENDIX A
radius-of-curvature surveys, but if the stations are
closely spaced and the inclination buildup not too
.,. ,..
.+ti, -.,-- .he “ cq?et?cia! m.e&rc! tnriy be
= ‘=”” ‘“Q =’mP!~
expected to yield good agreement with the circular
arc or the radius of curvature methods. Consider Fig. 2 showing a circular arc extending
Currently, two operational programs at the in space between survey points P1 and P2 and
&-----
%anrl~r~ Q~] ~0, Of California, Western Operation directed there along the known vectors tl and t2.
Inc. utilize the arc interpolatio~ One is on a We require the orientation of the plane of arc and
time-sharing terminal written in GEC Dartmouth the magnitude of its radius.
ALGOL 60, containing as its key part the procedure Let N,E,V = orthogonal Cartesian frame of reference
listed in the Appendix B; the other is a PL/1 (north, east, downward vertical)
System 360 program with a user’s option to select
+1,+2 = meas~ed inclination angles from the
either the arc or the tangential method. Both have downward vertical vector V at
been in use for some time now and, in general, the Points PI and P2, respectively
users’ reaction is favorable.
61, 62 = measured direction angles from the
Regarding the assumptions underlying the current
,. . . . vector N for the horizontal projection
approacn, lr wouici be diffic-uit to CiKiiiKl ifi Vi~W Of
of well at Points PI and P2,
all uncertainties that they manage to represent the
respectively
true shape of the wellbore. But the same can, for
the lack of data substantiating them, be said of rl, r2, rc = radius vectors from the origin of the
other ways of computing surveys. The calculations N,E,V system to Points PI, P2 and
involved in arc survey are rather too tedious for a C12, respectively
routine hand work, but present no problem for .s12 = measured depth difference between
digital computers once the necessary code has Points PI and P2
been written.
Standard definitions of a vector product, of its
Finally, it is probable that the elimination of
absolute value, and of the length of a vector are
discontinuities in tangent direction resuits in a
listed for subsequent reference:
better approximation to bottom-hole location than
in the routinely used tangent projection method.
This last conclusion appears to result from the
application of reasoning in Walstrom et al.2
Briefly, if the measured survey angles (but not
depths) are assumed to be subject to small errors,
then the effect on the far-end displacement of
each section will be magnified in proportion to the
slant distance between the end points; and this
must be less for an arc than for a straight segment
of the same Iength. However, for the arc method
the angular error must also reflect in changes
of radius of curvature so that the net effect is
rather unpredictable, and the last conclusion must
remain tentative.

ACKNOWLEDGMENTS

The author wishes to extend his thanks to the

management of Standard Oil Co. of California,
Western Operations Inc., for permission to publish
this paper and to George D. Boozer, Roy p. Harvey
and John E. Walstrom for their comments and help FIG. 2 — THE GEOMETRY OF CIRCULAR ARC
with its presentation. INTERPOLATION.

8 SOCIETY OF PETROLEUM ENGINEERs JO URXAL

 system is immediately

x, X2 X3 r
Imc
=
J-h%- . . . . ..(A43
I Y, ~, Y’3 Now the vector from the center Cl z to an arbitrary
point P, defined by a parameter angle a on the arc,
relative to the triad at C12 is:

-II~,.m(d)
La-
-P L
rcos

o
(d)l

-1
. . . . . . .. (A-7)

Hence, its components in the reference frame

(N, E,V) are:

Now the tangent unit vector, t 1 expressed in the b JN t,N n,2N-

frame (N, E,Vj at point PI , is
b= b,i t,~ n,z~
-P
b Iv t,v n,zv
.
. . . . . . . . . . . . . . . (A.8)

where blN, blE, etc., are the already computed

components of bl in the reference frame (N, E,V)
A similar expression can be written for t2 by and the same holds for fl and r112. With bp known,
replacing the subscript. Assuming next that fl and the displacement vector from PI to P is easily
t2 are not parallel (to within some predefine shown to be
tolerance), the normaI vector to the plane of arc,
;___
1.’=.> n~2 iS

{Z, A z.) (A-2) In particular, for a= a12 we obtain the displacement

of P2 from PI; hence, by cumulating components of
4’2=L($, AZ.) “ ““”””
D for successive arcs, we get eventually the
The angle a12 subtended by the arc at its center coordinates of all the survey stations and of the
C 12 follows from the standard definition: bottom hole. If the vectors tl and t2 are parallel,
then this will be indicated by the vanishing of the
d,= = urc tan [ L (’-j-,
A~= ) absolute value of their vector product in Eq. A-2;
in such case, we can bypass Eqs. A-3 through A-9
and compute the components of D directly by
multiplying either tl or t2 by s12, and the same, if
///-L 2(’’,/+ )1 * 0. (A-3) necessary, will hold for an extrapolation to the
total depth. In practice, of course, due to inevitable
where we must use the principal value of a12 small roundoff errors, the value .L(tl A t2j wiii
between O and v. The required radius of curvature seldom drop to zero; hence, we must decide on
now follows immediately from the condition that the some suitable tolerance at which we shall begin to
developed length of arc be equal to the measured regard the vectors tl and t2 as essentially
parallel.
depth difference; hence, In the appended program this tolerance was
assumed to be 0.001r, which is equivalent to an
angle of divergence of about 0.057°. With the
f’., = S,2 /4,=. . . . . . .. (A-4)
coordinates of station points computed and avail-
able, the determination of coordinates of any other
The orthogonal triad at the center of curvature,
intermediate point on the well axis is easily
which will be needed later, is completed by a unit
accomplished.
vector Al normai to the arc at PI (note that fl at
~..”A+ In practice, two distinct cases arise: either the
12 ~..= , ~ ~~ .Pi ~~e ~a~al!el by definition). Thus,
~ea~uie~ de~tb, IS s.pec~f~e~ ~fi~ we seek the
(N, E,V) coordinates of a point at that depth, or
A -t
ml-w)
A~,=, . . . . . . . .. (A-5) else the horizon is given, i.e., the value of V, and
we want the measured depth and/or the remaining
from which the iocation of Point C12 in Eke (P-J,E,v) unknown ~oord~nate~ .N and E. It is clear that the

FEBRUARY, 1973 9
first case is already solved by Eqs. A-8 and A-9, ,8<1<:?.>
1”< >,< 1,1 H[,.Ul>.N,d X ,1X)1.4’3LC.V
VLY(Pti 1.1
<CaL adrlAY P!41,1UT. NOP, *CAL UXI lNICGEd ‘Vi

since for any measured depth we will know the C@ MM?VF

ANO UE%TICAL
C@ MPIJIES
OIUECrl
015PLACE.LNr
ONS .Y F1lll
S PXCI
NG
,31
CIRCULAR
1. THE
CINCS
NO NIM.
W
E&ST
TANGENC
I J> vECFOti S GIVEN aT N ST&ll@N PO INIS OY rHE AVGLES PM1 ANO
interval in which it occurs and, hence, the distance I 40 rnr. l*rc.f.OLarEs rw DtSPLACCMENTS AND .4NGLZs EITMEN ON
14> rME VGNTIC&L COMPONENT PX[31, OH aN rME MEASUREO
from the initial point of arc to the point in question. I >~ #lEPr H *OP lCC’GROING W THE CUO ICC 8F 1300 LEaN P6RAMCTER 1X
15> aN!) US,1F46 6 Cr?NSr tNrE*V4L OX. T.? Mfil NT&l N IWUND VALUES
If that distance is say As, then the angle a to be 160 ., P.. fiMEt, R THE ,N7EI?P0!_AT1@N sib, tiTS RT NEaI?ES1 MULTIPLE
16> & U. S$14LLEN rHAN MW( I 1:
used in Eqs. A-8 and A-9 is As/p12 and the required I 70 R6GIV dE6L .ww T1. T2. Tx. Nlz,61, Bx, pb, Pxc 1133:
Cuo
17, UCAL ALF2, &LFX, AX, NX, S. SL. V. V2, VL, RUO. PHIXBTHTXB A1. A2.43,

coordinates follow. 1 no
1%5
1NTCGC8 l,K,

In the second case the parameter a is unknown ! 90

195
aEnL
COMMENT
PeOCCDURE
C!3MPUTES CaMPONENTS
VPROIICaOq, C> I
OF VECT9N
RC6L
FROWCT
aRRAY A.O. ct
C OF VECTONS

and must be found first, but here we have to 200

eos
h ANU
13EG1N
R ..LSO
cci]i=Qt21
HETuMNS ITS
.nc3]-4c31*8t21a
ABWLUIE VRLUEI

stipulate that the well course has no horizontal 210

215
C[21t=A[
C[31:=AC
31*
11*
R[ll-A[ll*e
B[21-A[%l*BC!l
C31:
J
220
sections in the plane of the required horizon, since 225 ENO
VP ROD,
VPtiOO,
=S9RT<C[ 11*2+ CC21.2.C[31t~)

P 30
then we know from physical considerations that no
. 23>
240
HEAL
CWI!$ENT
PR9CEDU)IE
RETW?NS
AZ1l
WE
N,E)I
AZIMUTH
REAL
FROM
N,CI
THE LATITUDE N@RTH
unique point exists where the hole would cross this 245 AN? DCPART4RE EtisT cOORDINfiTCS N A*u EI
250 &zl, =lF E=O TMCN (1- SIGN CN>>*P1/2
horizon; mathematically this appears through a 25S ELSE IF ARS(E>b. ABSCN1 THEN P1.12-SI GNIE>)/2-ARC ThN<N/E>
260 ELSE P1*(E-SIGN<E >*ll*SIGNIN)l >/2+4 RCrfiN1E/Nll
becoming indeterminate. Assuming that this 265
210 PR0CCOU3E %U7PI

situation does not arise, we know this from the 215

280
c0w4ENT
fiND T!4E
OuTpdTs
91 SPL4CEMENTS
MEAsunED
AT
WpTH,
6 PI ON
AN6Lt
THE
S
WELL
w T4NGENT
AXIS I
vEcTa Rs.

already computed values of V at the station points, 285

290
PNINT<s, PM Ix/cu, THTX/CR, PX[ll, PXC21, PX[31>1

in which measured depth interval rhe required 295

300
PROCEDURE
ceMMENT C’SMPUTES
lNTERP#
THE VECTOR rROM CENTC?+ OF ARC w THE

horizon occurs, so let the vertical distance between 305

310
lNTERPOLAIED
AT TM&T PO INTJ
PO INT.4NLl THE DISPLACEMENT ANO T&NGEN1 VECTaHS

3!5 BEGIN FOR !(, =1.2,3 DO

it and the horizon through PI be AV. Then, from 320 wE(31N mxc M1, =91[K1. CO SC AWX>+Tl [K18S1NCALP’X>J
325 PXIKl,.Pl [N1. <Q XC K)-nl C 1(1). NHO
Eq. A-8 we get 330 END I
33s NX, =VPR@O<NIE. RX. TX>l
340 PHI X,. aZ11TXC31, SQRT1TX[ 11.2. TxC21t2))’
345 THTx,.AZ1[TX( ll. TX[21>I

(bpv - b,v)r,z = (blv 35Q n,,,,

b V= [os(~~ 355
360
365 cOWWCNT lNITl&LIZE M& RKERS V ANCI S 4ND DISPL.4CEMENPS rMEN
310 LOOP OVER THE NO 9F STAT19NSI

+ t, ~ S.n (d) - ‘$fY) f’/2 “ “ “ ‘A-lO)

375
380
CR,.
F9R
Pi/200,
1,=1,2 D@ PI[ll, =PX[ll, =OI
38s P[C31t. PXC3J, =U,. S,.”DP[ 11,
39Q VL, =SL, =lF DX=O THEN MDPII 1 ELSE [MDPII ,\ox>*Llx,
39 s Fe? 1,.1 STEP 1 uNTIL N 00
By standard substitutions, .400 13EG1N PMIX,. PM IC1l.CU’
40> TH1’Xt=THT[ll*CRl
410 AX, =S!NIPMIX)I
415 T!2[ll,. Ax. C0S<T14TX>,

ff = taa pj2j ~ .,n~d~. ~U 42Q

425
Tz[?),
T2c31,
.ax.,,.
=cas<P!-llx)’
ciiirx>i

Ijuz 430
43s
,F ,>,
azGiN”cOXIcNT
NX,.
,“, N

VPWO<TI,
SKIP
T2. N12),
CALC Feu TnE INITIAL pTI
440
44s IF NX. O. flO1 TMEq
450 BEGIN C9!4WCPJT N@ RMALIZE N12, THEN CALC CEN [E,? &NGLE fiLF2,
.55 R.40 2F CURV NMO, AND VECTO~ RI :
460 PnR K,.l ,2,3 0!3 Nlz[l(l, =N12[!41/Nx’
46> *Lr*,. Azl(sQtr( l-vx*2>, Nx>1
470 NM~:=[WDP1 ll-MDPK1-11>/ALF21
41S VX, =VP~OO(Tl, N12Sql>1

this reduces to a quadratic: 480

48s
IF
-EGIN
1X TMEV
C’dMMENT lNTEUPOLflTE @VEN W A*C USING
.,0 VERTICAL lNTENV&LI
49!l v2:=(R1c 31*(cascaLF2>-1 )+ Tlc31. sINIALr2)I. RHO I
50rl AX:= P1C31+V2,
50s S’?8 V:=V+VX WILE V. AX. AX 00
>lQ q%GIN al,. <u- Pl[31>/MHOl
>1> A2, =AI. <A I.2*RI[31), TIC 31121
52Q &3, =.5.419 <t* A2/4+A2.2/ql/T 1C311
52S ALFX, =2. AZ1[I. ARS(A3)>1
,30 IF V=&X rHEN ALFxz=C!LF2 ELSE VL, =V1
>35 S,. MDP[l - 1 l+ ALFX. HHOl
>40 lNT<tlP
545
‘~/v . . . . . . . . . ..(A-ll) 5>0 ENll
END

>55 ELSE Fad 5%= S+ll X !4H1LC S=WPCI lsMDPCI1 DO

560 BEGIN CL3MMENT [NTERPOL&TC OVE.1 hN ANC uSING
56 S ME& SU?CO WPTM lNTCWliLl

Solving for u and substituting the result into the 510

515
aLFX,
INTERP*
=< S- M9?[1-11)/*H@l

5w IF S=/ MDPC1l THEN SL,. S

inverse function for a, we ger: 3*, ENO
590 EW CLSC
. .
595 Ir ix ,.,.
600 iEGIN COt4MENT INTE?pOLfiTE OwR A srRAIGHT sccT1O*
60> dSING UEMTIC4L tNTERVALI
610 V2, .P1C31+TI [31*< MVP[11-MJP[ 1-1111
615 F@u V,= V+DX WHILE V.V2, W 03
620 L3EG1N &V,. (U- PlC311/Tl [311
6t5 FL73 M,=1>2,3 UO PxCN18=P1C U1. TIC Nl*AXt
630 S,= MDP[!-11.AXI
635 BdTP1
64Q IF V=/ U2 THEN VLZ=V
. . . . . . . . . . . . . . . . . (A-I2) 645 ENO
65Q EN,>
655 ELSE F9R SX=S+DX WILE S. MDP[l l, MOPC1 1 UO

where 669
665
REGIN
USING
ca.wrir
ME4SLIRCV
INTCR.OLfiTE
DEPTM lNTERV&Ll
awn A ST HAIGMT sccrlav

670 F03 M:=l .2.3 00

675 PXCK1,. PICK1. rl[K1. (S-MOP[ 1-11)1
680 OUTP#
68, If S=/ MDPL1 1 THCN SL1=S
690 EN,>
5?> ENI! r~sc OuTP,
100 COMWEYT U’JUE [N TEI?.0L4T10+ M’?RNEHS V AND S TD LAST PI

and 705
110
RE&CMW
FOR NLXT
2TH<2
L08P
rHQN
PcISS,
&
lM<N
STnll
S18RE
ON PT.
DIS.L
M8VE
AT
VLCTO?+
LASr
T2
PT
T9 r!
tlEACHEDl
715 V,=VL,
720 S:=SL,
AV 72s F8R K,. !,2,3 08
uz=— -Zb,v . . . . . .(A-13) 139 qzGIN
PIIK1,.
r,[K,
PXIK1
,. T*cl(, ,

F’,z 735
74g END
145 END
75’3 ENLI SURVEY,
Only the negative sign of the radical in Eq. A-12 FIG. 3 — SOURCE TEXT OF PROCEDURE FOR
is relevant since al is poqitive by definition, and COMPUTING DIRECTIONAL SURVEYS BY THE ARC
METHOD.
10 sOCIETY OF PETROLEUM EXGINEERS JO URIVAL
if a2 < 0, then t~v - a1a2 > t~v; thus the radical Inclination and direction then follow by simple
must be subtracted to make the argument of arctan transformation of tp to polar coordinates:
positive. Measured depth increment (and hence the
measurea , at?p~ll
~ - I- lL-=
:.--1/) ,, ~~ ‘be
. .- horizon follows from
& E a.P12, and the remaining two coordinates from
Eq. A-8.
. . . . . . . . . . . . . . . . (A-17)
Though the expressions given above solve the
problem, it would be fatal to use Eq. A-12 in
practice, since AV/P12 is normally a very small
9P = -cfQ~ (’ tpE / t=N) . . (A-W
number, and a2 may also be small, whence a heavy
cancellation occurs in Eq. A-12 and significance Note that, since tpv can never be zero (as we do
is .“U.
Inct . To overcome this, we can expand the not allow an arc in a horizontal plane coinciding
“ ---- T-+:mn ..”..
with the iilCclpo~-..”,. hn,i>fin)
_”..,, no nrobIetns
r-- .-– occur
argument of arctan in a binomial series and take
the first few terms as approximation. This results in the evaluation of Eq. A-17. On the other hand,
in tpN = O is quite possible and a numeric procedure
a must check for this condition. If it occurs, then the
(A-14) signs of the two components of Eq. A-18 are used
u=
L
i=l
Up’””””””””””” to choose the correct azimuth.

where che terms are defined recursively by APPENDIX B

a,
u,=~—
2 t,v COMPUTER PROGRAM FOR THE ARC METHOD

The algorithm listed in Fig. 3 is given in ALGOL

60, the currently accepted standard language for
communicating formalized procedures. The comments
The series converges very fast, and three or four in the text serve oniy to indicate -dtla.
k “ .=‘“ L ‘me
“e...e
terms are sufficient to give excellent accuracy. In done at various stages and have no other
the actual program implemented three terms were operational significance. The algorithm must be
used. embedded in a driving program providing for input
If the hole direction at the interpolated Point P and output of data and may serve as a part of larger
is also required, we get immediately the expression well data manipulation system. Pi at Line 375 is a
for the tangent vector tp there from: global constant declared in this outside program
and equals 3.14159265 for a centesimal division of
a right angle, and 3.49065850 for the standard 90°
division,
***

FE BBUARY, 1973 11