Professional Documents
Culture Documents
Summary
In high-angle wells the force of gravity pulls the drillstring against the low side of the hole. This stabilizes the
string and allows drillpipe to carry axial compressive
loads without buckling. For this reason it is practical to
run drillpipe in compression in high-angle drilling where
the drillcollar weights, which are needed to avoid compression, would cause excessive torque and drag.
Introduction
In drilling vertical wells it is common industry practice
to avoid loading drill pipe so that it would be unstable
from a simple column-buckling point of view. Drillpipe
is kept stable by making sure that the buoyed weight of
the drill collars and heavyweight pipe exceeds the weight
on bit (WOB). This practice was recommended by
Lubinski in 1950. 1
In drilling directional wells it is common practice to
use about the same bottomhole assembly (BRA) weight
that would be used for a vertical well. Most operators do
not add collars as hole angle increases even though they
may have several thousand feet of drillpipe in compression in high-angle holes. This practice seems to work
fairly well; the high incidence of drillstring failure that
would be expected if compression really were harmful
does not occur. This paper is intended to justify and support this practice by showing that drillpipe can tolerate
significant levels of compression in small-diameter highangle holes because of the support provided by the low
side of the hole.
The benefit of using drillpipe in compression is that
the BRA weight can be kept low in high-angle drilling.
This, in tum, helps reduce torque and drag, which are
often operational constraints in very deep directional
drilling.
Earlier Work
The fact that a high hole angle can stabilize drillpipe
against buckling seems to have been recognized first by
Lubinski and Woods in 1953. 2 They stated that for
given values of drill collar weight and compressive load,
"there is a value of Ir below which helical buckling
may occur and above which it cannot occur." Lubinski
and Woods mentioned that they had done some experimental studies to quantify this phenomenon and that
the results of these tests had been used in preparing some
of the figures in their paper. Unfortunately these brief
comments were not noticed widely by the drilling
industry.
After Lubinski and Woods' work, the contribution of
hole angle to drillpipe stability was ignored in the drill01492136/84/01011167 $00.25
Copyright 1984 Society of Petroleum Engineers of AIME
1734
Fcrit =2 (
................. (1)
FCrit=EI( ; )
2 .........................
(2)
Size
(in.)
Nominal
Weight
(lbm/tI)
Material
5
5
4112
4
3112
5
4112
25.6
19.5
16.6
14.0
13.3
10.0
8.4
steel
steel
steel
steel
steel
aluminum
aluminum
Notes:
Body
Body
00
ID
~ ~
5.00
5.00
4.50
4.00
3.50
5.15
4.60
E/t
Moment
of Inertia""
(in.4)
(psi-in. 4
x 10 6 )
(Ibmlfl)
18.11
14.27
9.61
6.46
4.50
20.66
13.73
543
429
288
194
135
219
146
21.4
16.4
13.9
12.2
10.7
6.2
4.9
4.00
4.28
3.83
3.34
2.76
4.10
3.60
pAgt
" 1=0.04909 (0 4 _d 4 ).
0.436
18
16
BUCKLING OCCURS!N THIS REGION
10
10
20
30
40
50
60
70
80
90
0.511
.................................... (5)
Graphical Presentation
Eq. 1 lends itself to simple graphical presentation provided a separate graph is made for each size of drillpipe.
Several of these graphs are included in this paper; their
preparation and use is discussed briefly here.
'Lubinski,
A.:
1735
. = (23.4XI0 8 SinX)
F ent
60
HOLE ANGLE
70
80
90
IDEGREES~
'/2
.............. (6)
30,000 Ibf
60
70
80
90
Discussion of Approximations
A number of idealizations and approximations were
made in the analysis scheme presented in the preceding
section.
JOURNAL OF PETROLEUM TECHNOLOGY
20
20
18
18
16
16
14
14
12
12
~
0
,0'
",\
,0,00'
10
80
90
6~0-L~'-~~--~--~--~50o---~60'--'7~0--~80'-~90
Favorable Approximations. The following approximations tend to make the analysis err on the side of safety
and reduce the chances of unexpected pipe buckling.
1. The radial clearance used is based on the pipe-body
diameter. Presence of tool joints will reduce this
clearance and increase the critical buckling loads.
2. The mud weight used in the analysis is 15.5 Ibm/gal
[1546 g/m3]. If the actual mud weight is lower than that,
the higher buoyed weight of the pipe will increase the
critical buckling load.
Unfavorable Approximations. The following approximations may allow the pipe to buckle even though it appears stable.
1. The wellbore is assumed to be straight. A compressive load in an angle drop-off region will reduce the
effective weight of the pipe there and, thereafter, lower
the critical buckling load proportionally.
2. API Class I (new) drillpipe dimensions have been
used in the analysis. Worn pipe will be less stiff, less
heavy, and buckle more easily.
3. The effects of torque are neglected in the equations
Conclusions
Drillpipe in high-angle holes can carry compressive axial
loads without buckling. At hole angles above 50, use of
drill pipe in compression will permit a major reduction in
BHA weight and a significant reduction in torque and
drag on the drillstring.
The purpose of this paper is not to revolutionize
drill string design but to provide an engineering basis for
analyzing current practice and making improvements in
the future.
Nomenclature
A = cross-sectional area of pipe, sq. in. [cm 2 ]
E = Young's modulus, psi [kPa]
20,---,----r------.----,---..-,--,-----,----,------,
5000 Ibf
18
16
14
1;00 Ibf
12
10
30.000 Ibf
70
80
80
90
90
Acknowledgments
We thank Exxon Production Research Co. for permission to publish this paper. We also thank Arthur Lubinski for providing unpublished results and Trudy Atkins
for typing the manuscript.
References
APPENDIX
Derivation of Eq. 1
The derivation of Eq. 1 in this paper begins with Eq. 33
in Ref. 4, which is
(1-1')2
(1 +1')(1-21')
11"2 (
L4pAg )
Ef-- n 2 of .
L2
n 211"4EIr
a
an
-Ferit
which is
....... (A-4)
or
Ferit
2n-
2L 4 pAg sin
(J
n 3 11"4Efr
0, ................... (A-5)
which gives
n
2=
4
(L pAg sin
(J)
'h
................. (A-6)
1I"4EIr
Ferit
-2
(EfPAg sin
(J)
'h
This equation is the same as Eq. 1 given in the text. Ignoring the eigenvalue nature of the problem has
eliminated both nand L from the final formula. A few
sample calculations made with the original Eq. A-I have
shown that the approximate result is quite accurate for
pipe lengths more than about 200 ft [61 m].
Note on Compression. Throughout this paper the
"compressive load" being considered is the load that exists above and beyond any isotropic compression caused
by hydrostatic pressure.
E+OO
E+OO
E-01
E+oo
cm
N
kg
kPa
JPT