Pipeline Design
8.1 Introduction
Expansion may be induced by internal pressure and temperature. In recent years, more and
more High Pressure and High Temperature (HP/HT) fields are developed using pipelines and
steel catenary risers. In this design scenario, axial creeping, buckling in the form of upheaval
movements and lateral movements or a combination of both, may take place.
In this Chapter, the basis for expansion analysis will be outlined first with a focus on effective
force calculation. An analysis model is introduced for axial creeping of a high temperature
flowline in clay soil condition. Simplified mathematical equations are presented for use in the
upheaval buckling and lateral buckling designs. Finally HP/HT flowlines will be considered
on uneven seabed, and their buckling behavior will be characterized as a combination of
upheaval and lateral buckling modes: the flowline will uplift first and then move more
laterally. There are many papers published on these subjects. Hence, we shall focus on the
mathematical formulation that has actually been adopted in the current pipeline design
practice. Some further readings include Ellinas et al. (1990), Kershenbaum et al. (1996), Nes
et al. (1996) and etc.
8.2 Expansion
The maximum pipeline end expansion is calculated for the lower bound friction coefficient
and the highest pipeline axial stresses are derived for the higher bound friction coefficient. The
116 Part 11Pipeline Design
increased axial resistance from backfill along the route should be also accounted for. The input
data for expansion analysis includes:
0 Pipeline and Coatings Properties;
0 Geotechnical Data;
0 Depth of Burial.
8.3.1 General
In recent years, continuous monitoring of some hightemperature subsea flowlines operated in
the North Sea has shown that the flowlines are experiencing a gradual overall axial
displacement towards the colder outlet ends. This net axial shift of the pipelines has proved to
be critical for the integrity of the tiein spools. In this Section, we shall present an analysis
model for axial creeping of flowlines caused by soil ratcheting, based on Tsrnes et a1 (2000).
 Curve 1
Curve 1
I
I *.
._
__
Curve 2
T
Unloading
+ >
Axial Displacement
The first part of the bilinear Coulomb friction curve is linear. Initially the expanding pipeline
stretches the surrounding soil and the soil resistance behaves purely elastic. When the static
friction limit is reached, the pipe surface starts to slide relative to soil at the lower residual
frictional value (plastic behaviour). The distance the pipe needs to travel before plastic sliding
occur, is normally referred to as the frictional mobilisation length (y).
This is shown in Figure 8.2 b), where the effective compressive force increases linearly from
the end to a point where the pipeline becomes fully restrained. In the midsection of the
pipeline, often referred to as the anchor zone, the effective compressive force will remain
constant. Since a section of a long pipeline is fully restrained at full operating condition, it
cannot undergo global axial creep.
URSTHEATUP
__ __.I.
a) Definitkw Strelch
EmfcSva
Tension
f
y
No fnclion build up
Figure 8.2 Expansion of a long flowline with free ends and constant pressure and temperature
(Tornes et al., 2000).
Chapter 8 Expansion, Upheaval & Lateral Buckling 119
(Ref. Fig. 2
800 hl
 h3
600  h4
400 h l
e  h6
. 200
2  h7
{ o  h8
.z* 200  h9
 hlO
0' hll
w 400 hll
600 h13
h14
800 h15
1000 h16
Only at the moment friction has been mobilised along the entire length, will the maximum
compressive force occur at midline. As can be seen from Figure 8.3, this occurs approximately
when temperature profile number 9 is applied to the flowline. The corresponding axial
displacement history from the FE analysis is shown in Figure 8.4. A schematic of a
temperature profile and the corresponding effective force and axial displacement distributions
at one instant during the heating up process, i.e. prior to reaching the full operating conditions,
is shown in Figure 8.5.
Since the thermal strain is proportional to the temperature in the line, the tendency to expand
will decrease gradually from the hot end to the point where the pipeline is at ambient
temperature. As the pipeline expands towards the hot end, the accumulation of frictional force
is built up away from this end of the flowline. This means that the further away from the hot
end, the more resistance (higher effective force due to the accumulation of friction force)
against the thermal expansion.
120 Part 11Pipeline Design
1.o
 0.6
6
I 0.2
an
P
0.2
0.6
1.o
0 200 400 600 800 1000 1200 1400 1600 180
Distance Along Pipe From Well [m]
Figure 8.4 Axial pipe displacements along the flowline through the temperature cycle 1 (Ternes et al.,
2000).
The analysis results showed that this heating up process causes the entire flowline to shift
towards the cold end compared to a case where it is heated up uniformly to its operating
conditions (i.e. without heating it up gradually from the hot towards the cold end).
Furthermore, the entire shift occurs before the maximum effective force is located at midline,
Le. before the frictional resistance is mobilised along the entire line.
The reason for this is as follows: Figure 8.5 shows how the pipeline expands in both directions
from either side of the stationary point at which the maximum effective force occurs. Since
this point gradually moves from the hot end towards the midline, this means that all the points
on the hot half of the pipeline will first experience an expansion movement towards the cold
end followed by movement in the reversed direction towards the hot end. The latter occurs as
soon as the point of maximum effective force passes the point in question.
8.4.1 General
Global buckling is a beammode buckling that is characterized by an amplitude and
wavelength. An upheaval buckling may take place for a buried pipeline that uplifts due to
excessive expansion.
the Pedersen and Jensen model that has been derived from an elastic beam theory for the
imperfect pipeline uplifted in the xw coordinate system.
a) Definition stretch
b) Temperature profile
,
X

KP
A' 7 f I T f ,
AKP
X
Figure 8.5 Expansion of a flowline with a free end & a temperature gradient (Ternes et al., 2000).
122 Part 11Pipeline Design
P
N9 D2dp
t I
I t I I
Figure 8.6 Uplifted pipelines (top), and variation in compressive axial force (bottom) (Pedersen &
Jensen, 1988).
1) Calculate the k value associated with a given free span (2L) from Equation (8.5) and
then predict the axial force N using Equation (8.6), where q is the total pipe weight per unit,
qpis the weight due to pipe plastic deformation, and Lois half wave length, L is half uplift
length, respectively .
a Y P 1
+)sin kL  (  K) cos kL + (aL + P ) = 0 (8.5)
(2 k kZ k2
N = EIk2
where
qp * (3L  2L0)+ 3 . q . L
P= 3.E.I
9
K=.f (L,L)*L f o r ~ s ~ , ,
6E.I
or K =0 for L > L o .
Chapter 8 Expansion, Upheaval & Lateral Buckling 123
n
Without initid impcrfection
Figure 8.7 Minimum temperature increase AT,,,i. for equilibrium curves associated with uplifted
pipelines (Pedersen & Jensen, 1988).
n n
N , = ;EA, AT  v D~hp + D~M
2 4
where NOis effective axial compression away from the buckle, N is compression in the uplift
buckle, D is steel pipe outside diameter, AT is the temperature change, AP is the pressure
 the product and the pressure at the seabed, A, is the pipe steel cross
difference between
sectional area, a is thermal coefficient, v is Poisson's ratio, w'is the differential deflection,
and is the differential initial deflection, respectively.
124 Part 11Pipeline Design
w = 0.064(Peff,)? 5
EI
(8.10)
where
w : total required download;
Pen : effective axial force;
fa : axial force factor;
H : prop height;
EI : flowline bending stiffness.
An analysis may be carried out to calculate the total required download for a series of prop
heights, which accounts for the true axial force in the flowline by the build up soil friction
forces along the route. A safety factor of 1.2 is normally applied to the axial force. The
required extra download is then calculated as:
wnet =  &PL  wbf (8.1 1)
where:
aet: net required download (rock dump);
OPL : pipeline unit weight;
wbf : backfill unit weight.
The calculation is performed for the design and operational temperature and pressure. In the
calculation, the effect of backfill is considered, for both the axial force and download
components. The net required download finally can be converted into an equivalent soil or
rock dump cover based on the geotechnical properties entered, as follows:
(8.12)
where:
y : submerged weight of the soil and rocks dumped on the pipeline;
Do : pipeline overall diameter;
z : the cover from the top of the pipe to the surface of the soil above the pipe
centreline (backfill depth);
f : an uplift coefficient determined experimentally, generally about 0.7 for rock and
0.5 for sand, but occasionally much smaller in loose sand.
Chapter 8 Expansion, Upheaval & Lateral Buckling 125
8.5.1 General
Unburied pipe on the seabed will buckle laterally instead of vertically unless the lateral
friction coefficients are very high. The resistance to sideways movement is the submerged
weight multiplied by lateral friction coefficient. The driving force for lateral buckling is the
compressive force induced by operational temperature and pressure. A lateral buckling
analysis may be carried out based on Hobbs (1984), Hobbs and Liang (1989). The effective
axial force is calculated for the operating condition defined within a design envelope (shown
by the series of graphs) for any of the four possible modes presented.
(8.14)
(8.15)
Values for the buckling constants are given in the following table.
126 Part 11Pipeline Design
The maximum amount of pipe which can feed into a buckling occurring at the pipeline end is
equal to the unrestrained expansion of the line. However, the force in the buckied section (P)
would have to be zero. The increase in length of pipe A1 in the buckled section from the
unbuckled state can be determined as:
L, = ('buck (8.19)
P*U
Total expansion into the buckle is,
A finite element analysis may be applied to gain an understanding of the complex mechanism
between the vertical and horizontal mode of buckling. This included a number of 2D and 3D FE
simulations using a typical 3D seabed model (Nystrom et al., 1997).
3D buckling model
In the first examples, a flat seabed with one realistic vertical seabed imperfection was used
(1.8 m high and 50 m long, see Figure 8.7). A 10'' oil flowline with a residual lay tension of
50 kN and no internal pressure is assumed. The seabed is assumed to be flat in the lateral
plane, however a pipe imperfection of Wom/Lo=O.OO1was introduced in the pipe horizontal
plane. In order to determine the effect of the restraining force, two friction coefficients were
checked (0.3 and 1.O). Temperature loads were applied to the model, in addition to the residual
lay tension.
Chapter 8 Expansion, Upheaval & Lateral Buckling 127
Seabed Profile
309.5
310
310.5
1
+ \
....
31 1
Figure 8.8 Vertical pipeseabed profile (Half mode) (Nystrom et al., 1997).
Figure 8.8 shows that, for a friction coefficient of 0.3, the flowline starts to lift off the seabed
at approximately 13C. Little horizontal movement has yet occurred. At 16"C, the vertical
displacement reaches 0.25 m but, suddenly drops down again to zero. In the same instant, the
pipe 'snaps' laterally by 1.85 m to a new equilibrium configuration. As stated previously the
pipe is now fallen back onto the seabed again.
The same effect can be seen for the higher friction case (p=l.O), the differences being that the
vertical movement starts at 15"C, the pipe lifts slightly higher, the lateral buckling temperature
is approx. 2 1"C and the resulting lateral displacement is 1.4 m.
Even though the higher friction case causes the pipeline to buckle laterally at a higher
temperature, Figure 8.9 shows that the flowline buckles at approximately the same effective
force for both fiiction cases. This is an important finding for the understanding of the 3D
buckling behaviour. On an uneven seabed, therefore, it appears that the critical lateral
buckling force is little affected by the lateral friction resistance. This phenomenon can be
explained by the fact that for a pipeline resting on an uneven seabed, the vertical imperfections
can be significantly larger than the horizontal imperfections. As the flowline initially 'buckles'
vertically, it looses lateral restraint (less contact, less lateral restraint) where upon it interacts
with the horizontal mode and buckles laterally. Specifically, the pipe starts to lift vertically at
approximately 170 kN (Ref. Figure 8.9) and starts to loose lateral frictional restraint. As a
result of this reduction in lateral restraint, the lateral buckling force also decreases. The critical
lateral buckling force is reached at an effective axial force of approx. 200 kN. Because the
pipe now has reached the point of axial instability in the horizontal plane, the path of least
resistance is chosen, and a lateral 'snap' to a new equilibrium configuration occurs. It should
be noted that, in theory, not all pipes which lift off vertically will buckle laterally. A situation
could occur where the flowline operational loads only are capable of lifting a section of the
pipeline. In this case the vertical mode would still be governing.
128 Part 11 Pipeline Design
Lateral Deflection 
2wwo
f
B 1
w
t
Pd lw0W
0
ow 050 100 150 2 00 250
AMPLITUDE [m]
Figure 8.9 Static 3D analysis, buck force vs. deflection (Nystrom et al., 1997).
8.7 References
1. Ellinas, CP, Supple, WJ and Vastenholt, H (1990): Prevention of Upheaval Buckling of
Hot Submarine Pipelines by Means of Intermittent Rock Dumping, Proceedings 22nd
Offshore Technology Conference, pp 5 19528.
2. Hobbs, RE (1984): InService Buckling of Heated Pipelines, Journal Transport
Engineering, pp 178179.
3. Hobbs, RE and Liang, F (1989): Thermal Buckling of Pipelines Close to Restraints,
Proceedings 8th Int. Conf on Offshore Mechanics and Arctic Eng, pp 121127.
4. Kershenbaum, NY, Harrison, GE and Choi, HS (1996): Subsea Pipeline Lateral
Deviation Due to High Temperature Product, Proceedings 6th Int. Offshore and Polar
Eng Conf, Vol2, pp 7479.
5. Nes, H, Szvik, S, Levold, E. and Johannesen, A., (1996): Expansion Control Design of
Large Diameter Pipelines, Proceedings 15th Int. Conf on Offshore Mechanics and Arctic
Eng, Vol5, pp 279285.
6. Nystram P.R., Terrnes K., Bai Y. and Damsleth P.A., (1998): 3D Dynamic Buckling and
Cyclic Behavior of HP/HT Flowlines, ISOPE1997.
7. Pedersen, P. T. and Jensen, J.J, Upheaval Creep of Buried Heated Pipelines with Initial
Imperfections, Marine Structures, 1988.
8. Palmer, A., Ellinas, C., Richards & Guijt (1990): Design of Submarine Pipelines against
Upheaval Buckling, OTC 6335.
9. Terrnes, K., Ose, B.A., Jury, J. and Thomson, P. (2000): Axial Creeping of High
Temperature Flowlines Caused by Soil Ratcheting, OMAE2000.