jpt forum

Lateral Buckling of Axially Constrained Pipelines

Introduction where L is the length of the pipe; that is, when

Imagine a length of straight uniform pipeline, rigidly p = 272EDT/(1 – 2V)L2.
attached to a rigid anchor block at each end, and
initially unstressed. If an internal pressure is applied, The above argument leads to a correct value for
can this pressure cause the pipe to buckle sideways? the buckling pressure, but leaves a certain uneasiness
Although a correct answer to this question follows in one’s mind, because an instinctive feeling that the
from a simple argument, it is nevertheless the subject axial compressive force must “follow” the buckling
of dispute among pipeline engineers. A simple but pipe seems to conflict with the fact that the ends are
misleading argument goes like this: pressure induces fixed. An alternative argument avoids this difficulty,
a hoop tensile stress around the pipe, and if it were and makes clearer what the force pushing the pipe
unconstrained it would contract along its length (be- sideways is.
T-.
Ill +1.,.“l.-
L1lC cIUUt
,-G ..”,, --” w~ll itci=lf and the.
~ I+Y=

cause Poisson’s ratio is positive). Since the anchor (.Z1~U,lLL.lL t thw y.y- ,,-----------------

blocks resist this contraction, a longitudinal tensile contained fluid within it were considered together.
stress is induced. Since the pipe wall is in longitudinal Instead, consider them separately, Fig. 1a shows the
tension, the pipe cannot buckle laterally. forces and moments acting on an element ds of the
It is the final step in this argument that is incorrect. pipe wall, and Fig. 1b shows the forces acting on the
This article sets down reasoning that leads to the cor- fluid contained within the element. The resultant force
rect conclusion—that the pipe can buckle—and com- exerted by the contained fluid pressure on the pipe
pares the predicted buckling pressure with experiment. wall is Rds, and ~ denotes the inclination of the de-
flected pipe to its original line. Since the fluid element
The Critical Pressure must be in equilibrium, resolving forces perpendicular
An elementary argument shows that the longitudinal to the pipe axis in Fig. lb,
stress set up in a straight thin-walled axially con- Rds = (7/4)Dzp d~
strained pipe by an internal pressure, p, is vpD/2T,
where v is Poisson’s ratio, D is diameter, and T is wall so that
thickness, The resultant force over a complete cross- R = (7/4)D’p d#/ds.
section perpendicular to the tube axis is (7r/4)D2p
(1 – 2.), compressive since v < 1%, the difference In words, the contained fluid exerts a lateral force on
between the compressive force (7/4)D2p carried by a deflected pipe; the magnitude of the force per unit
the tluid within the pipe and the tensile force (vpD/ length is the pressure multiplied by the cross-section
2T) (nDT) Carned by the waii of ti’Ie pipe. Tine pipe multiplied by the curvature, and it acts towards the
C~I th~~ be tho~~kf nf 9Cn enlllpm with ~i~rn~d ~ds
outside of the curve. If one analyzes buckling by ex-
~... “L s.” . . . . . . . . .

carrying this resultant compressive force, and buck- amining the pipe deflected from its inittal posittmi,
ling can be expected to occur when the force reaches one must not forget the existence of this force (which
the Euler buckling load. If Young’s modulus is E, the is of course well known in other contexts).
flexural rigidity is ZED3T/8; by this argument, buck- Analysis of the stability of the pipe can from here
ling will occur when on follow the conventional methods of beam-column
analysis: the differential equation governing the be-
(7r/4)D2P(l – 2v) = 4r2(rED3T/8)/L2, havior of a deflected pipe is derived, and one then

JPT Forum articles are limited to 1,500 words including 250 words for each table and figure, or a maximum of two
pages in JPT, A Forum article may present preliminary results or conclusions of an investigation that the author
wishes to publish before completing a full study it may impart general technical information that does not wa~
rant publication as a full-length paper. All Forum articlea are subject to approval by an editorial committee.

Letters to the editor are published under Dialogue, and may cover technical or nontechnical topics. SPE -AlME
resewes the right to edit letters for style and content.

k,AX7C!,fD13D 107” u-n-

1283
looks for conditions under which the pipe can remain 4 ~xperimental
in equilibrium in a deflected form, so that the equi- N/mm2
librium of a straight pipe is neutral or unstable. If N t~
is the axial force transmitted by the pipe wall and y
is the lateral deflection, the difIerential equation turns
out to be

(TED3T/8)$ + [(iT/4)D2P - N# = O

for small deflections. This is identical with the equa-

tion that governs a beam-column carrying an axial
~mm
load (7/4)D2p – N, and it follows that buckling will lateral deflection
occur when this force reaches the Euler buckling load. Fig. 3-Observed relation between pressure and
lateral deflection. (1 mm = 0.0394 in.;
The corresponding critical pressure is the same as 1 IV/mm’ = 145.0 lb/in.~
before.
walled stainless steel tube is held between tied steel
Experimental Results
blocks into which the ends of the tube are cemented.
The conclusions of the previous section were tested The tube is supported on thin steel strips, which are
experimentally in the apparatus shown in Fig. 2a, free tn
---- mnve
.- ...- . - cirlew~y~
“--- .. Qn s~.d! m!!ers. Inte.rna! Qii
which is a foreshortened end view, A straight thin- pressure is applied by a hand-operated pump, and
measured by the Bourdon gauge. A traveling micro-
scope (not shown) measures the central deflection.
Fig. 2b shows the tube under an internal pressure
of 4.9 N/mmz (710 lb/in2), and it is clear that it does
indeed buckle sideways. Fig. 3 shows how its central
lateral deflection depends on the pressure applied. For
the tube used in the experiment,
— .— -— —— -—
axis before deflection D= 7.80 mm (0.313 in.),
T = 0.151 mm (0.0593 in.),
L = i,fM5 mm (73.5 in.),
E = 204 kN/mm2 (2.96 X 107 lb/in,2),
v= 0.268,
and the critical pressure calculated from the expres-
sion derived earlier is
—- — .— —— p = 2.93 N/mm’ (425 lb/in. a).
Fig. l—(a) Forces on an element of pipe; (b) forces This agrees well with observation. Small lateral de-
on the fluid within the pipe element. flections do occur at rather lower pressures, presum-
ably because the tube is not perfectly straight initially,
but the deflection becomes large only at pressures
slightly above the calculated value.
If an ordinary beam has clamped ends that cannot
move inward as the center of the beam deflects, any
lateral deflection will induce an axial tensile strain in
the beam, though normally the effects of this tensile
strain are negligible unless the deflection is unusually
large. In the same way, lateral deflection of the pres-
surized pipe induces an axial tensile strain, which
alters the axial force in the pipe wall. It is this that
causes the gentle rise in the observed pressure-
deflection curve as the deflection becomes large. A
straightfonvard analysis gives the complete relation
between pressure and lateral deflection it is omitted
here for reasons of space.

Acknowledgment
We thank J. W. Ells for suggesting this problem.
A. C. Palmer
Cambridge U.
a b Cambridge, England
Fig. 2~An experiment on lateral buckling. J. A. S. Baldry
(a) Constrained tube initially free of pressure; Iranian Oil Supply Co.
(b) tube urider p~essurc greate[
L..... uu~~,,,,6
-------- -.--* -.. ●LIIa,, I.,,,.l,li”” p,==
mracc,,,a
-=,.-.. Tehran, Iran

1284 JOURNAL OF PETROLEUM TECHNOLOGY