Beruflich Dokumente
Kultur Dokumente
www.elsevier.com/locate/oceaneng
Abstract
A new study investigates an unburied offshore “snaked” pipeline behavior under various
types of seismic faults. The snaking of the pipeline is caused by the thermal/pressure expansion
and soil friction. The snaking takes place at a certain distance from the pipeline’s unrestrained
end and gradually increases towards the restraint. It is shown that longitudinal seismic faults
have less effect on a straight pipeline than a snaked pipeline. The new seismic analysis demon-
strates that an increase of ground displacement causes a very small change in bending and
longitudinal stresses. The new approach results in a safe, subsea pipeline construction and
operation with a significant cost reduction. 1999 Elsevier Science Ltd. All rights reserved.
1. Introduction
* Corresponding author.
0029-8018/00/$ - see front matter 1999 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 2 9 - 8 0 1 8 ( 9 8 ) 0 0 0 7 9 - 1
474 N.Y. Kershenbaum et al. / Ocean Engineering 27 (2000) 473–487
2. Pipeline snaking
The purpose of the pipeline snaking analysis is to determine the unburied pipeline
shape preceding a seismic impact. As it was mentioned above, pipeline seismic
stresses will depend on the initial shape of the pipe axis. Kershenbaum et al. (1996)
N.Y. Kershenbaum et al. / Ocean Engineering 27 (2000) 473–487 475
assumed that the pipeline was initially straight laying on a horizontal bottom foun-
dation. Internal pressure and pumping product temperature would cause a linear
expansion of the pipeline. However, the frictional forces between the pipeline and
the seabed would resist this expansion. These longitudinal frictional resisting forces
could cause pipeline buckling due to Euler’s instability. The authors assumed that for
a long enough pipeline, the cumulative friction resistance to pipeline axial expansion
stipulates the sinusoidal form of lateral deflection/snaking (see Fig. 1):
冋
Z ⫽ A(x) sin n(x)
x
L 册 (1)
where Z is the pipeline lateral deflection in the horizontal plane, n(x) is the mode
number, x is the coordinate along the axis of the pipeline counted from the restrained
end to unrestrained end, and L is the total length of the pipeline. The pipeline total
energy balance work, W, is as follows:
W ⫽ We ⫺ Wl ⫹ Wf (2)
where We, Wl, and Wf are pipe bending strain energy, longitudinal, and friction work
in the lateral direction, respectively. Substituting into Eq. (2), the following equation
is used for the equilibrium condition:
dW
⫽0 (3)
dn(x)
The amplitude A(x) is obtained from the following expression:
2l(x)
A(x) ⫽ (4)
√⑀(x)
where l(x) is half a single wave length (see Fig. 1) of the pipeline at location x.
Kershenbaum et al. (1996) used a different approach to determine the amplitude
A(x) as follows:
1
A(x) ⬇ (5)
EI 5
⫹
16l4(x)ws f
冪A
I
2
s
3. Geotechnical aspects
Fig. 2. Subsea unburied pipeline behavior at operation (dn ⫽ 300 mm, L ⫽ 2000 m, ⌬P ⫽ 14 MPa, Ti
⫽ 71°C).
N.Y. Kershenbaum et al. / Ocean Engineering 27 (2000) 473–487 477
Table 1
Values of a and b coefficients and ground displacement, D
Ground fault type Coefficients (a) and (b) Ground displacement, D for
M ⫽ 6/7
Fig. 4. Oil pipeline under mid-point strike-slip fault during operation (no initial tension) (dn ⫽ 300 mm,
L ⫽ 2000 m, ⌬P ⫽ 14 MPa, M ⫽ 6).
Fig. 5. Oil pipeline under mid-point strike-slip fault during operation (no initial tension) (dn ⫽ 300 mm,
L ⫽ 2000 m, ⌬P ⫽ 14 MPa, M ⫽ 7).
stresses, ls, due to bottom slab strike-slip. These analyses were performed on a
300-mm nominal-diameter subsea pipeline, of 358.6 MPa SMYS, in operational and
shutdown conditions for 6 and 7 seismic fault magnitudes. The results are presented
in Table 2. The largest compressive axial load of 628 kN due to pressure/temperature
impact during the operation is unchanged because it represents the snaked pipeline
Euler’s instability force. The data representing pipeline shutdown condition indicates
an increase in the initial axial tension. However, the strike-slip fault has a small
effect on the bending and longitudinal stresses of the pipeline. Pipeline stress ranges
due to strike-slip fault for 6 and 7 magnitudes are far from critical conditions. The
maximum bending stress is less than 30% SMYS and the total longitudinal stress
value is not more than 51% SMYS (see Table 2).
The results of the analysis demonstrate the low sensitivity of pipelines to the
480 N.Y. Kershenbaum et al. / Ocean Engineering 27 (2000) 473–487
Table 2
Pipeline stress ranges due to seismic faults
Inlet ta Distance from Pipe axial tension Ground Pipe maximum Total maximum
restraint initial/after fault displacement bending stress longitudinal
due to fault stress
(°C) (m) (KN) (m/Mb) (% SMYS) (% SMYS)
Strike-slip/operational conditions
71 1000 0.0/52.0 0.120/6 7.4 31.2
71 1000 0.0/178.5 0.955/7 23.3 47.2
Strike-slip/shutdown conditions
⫺2 1000 135/172 0.120/6 7.4 31.2
⫺2 1000 135/290 0.955/7 27.0 50.9
Normal/reverse slip, vertical component/operational conditions
71 1000 ⫺ 628.0/ ⫺ 502.7 0.390/6 36.1 47.9
71 1000 ⫺ 628.0/ ⫺ 85.0 0.645/7 42.9 57.5
Normal/reverse slip, vertical component/shutdown conditions
⫺2 1000 135.0/146.8 0.390/6 35.5 44.3
⫺2 1000 135.0/155.7 0.645/7 45.2 48.2
Normal/reverse slip, longitudinal component/operational conditions
71 30 ⫺ 628.0/ ⫺ 628.0 0.390/6 84.3 84.9
71 30 ⫺ 628.0/ ⫺ 628.0 0.645/7 86.3 87.4
71 1000 ⫺ 628.0/ ⫺ 628.0 0.390/6 72.4 77.1
71 1000 ⫺ 628.0/ ⫺ 628.0 0.645/7 74.0 78.0
Normal/reverse slip, longitudinal component/operational conditions
93 30 ⫺ 524.0/ ⫺ 524.0 0.390/6 83.6 85.2
93 30 ⫺ 524.0/ ⫺ 524.0 0.645/7 86.2 87.4
Oblique-slip combined effect/operational conditions
71 30 ⫺ 628.0/ ⫺ 502.7 0.550/6 84.5 85.3
71 30 ⫺ 628.0/ ⫺ 85.0 0.912/7 89.6 89.7
71 1000 ⫺ 628.0/ ⫺ 502.7 0.550/6 78.4 80.9
71 1000 ⫺ 628.0/ ⫺ 85.7 0.912/7 81.4 81.7
a
t: temperature; bm/M: meters/Magnitude.
strike-slip fault magnitude. When the fault magnitude increased from 6 to 7, strike-
slip total displacement increased almost 8 times but pipe total longitudinal stress
increased only from 31.2% SMYS to 51% SMYS. The total longitudinal stress was
used as a criteria of an unburied pipeline under seismic loads. In respect to pipeline
occasional loads, such as an earthquake, ASME B31.4 code requires the sum of
longitudinal stress to be below 80% SMYS to resume operation.
was used to estimate normal and reverse-slip fault impacts on the subsea pipeline
movements and stresses.
where Mx is current pipe bending moment within each span (Mx1, Mx2),
wsx2
Mx1 ⫽ R1x ⫺ ⫹ Ty (9b)
2
Fig. 6. The fault rupture line divides the pipeline into two unequal portions.
482 N.Y. Kershenbaum et al. / Ocean Engineering 27 (2000) 473–487
wsx2
Mx2 ⫽ ⫺ M ⫹ R2bx ⫺ ⫹ Ty (9c)
2
wsl2 M Th
R2b ⫽ ⫹ ⫹ (9d)
2 l2 l2
My is the bending moment at yield, Ky is the pipe curvature, y is the pipe axis
elevation, A and B are Ramberg–Osgood coefficient and exponent, respectively, ws
is the pipe submerged weight, T is the final or current pipe tension (initial and fault
induced tension), and h is the soil vertical displacement due to the seismic fault
(Fig. 6).
The boundary conditions for each span are as follows:
first span
x ⫽ 0, y ⫽ 0, y⬘ ⫽ 0 (10a)
x ⫽ l1, y ⫽ 0 (10b)
second span
x ⫽ 0, y ⫽ 0 (11a)
x ⫽ l2, y ⫽ ⫺ h, y⬘ ⫽ 0, y⬙ ⫽ 0 (11b)
The results are presented in Table 2 for a 300-mm nominal-diameter subsea pipe-
line, of 358.6 MPa SMYS, in operational and shutdown conditions for 6 and 7 fault
magnitudes and 71°C temperature. The initial compressive axial load of 628 kN due
to pressure/temperature impact during the operation was reduced to 503 and 85 kN
in compression after 6 and 7 magnitude faults, respectively. Similar results were
obtained for pipeline shutdown conditions: pipe initial tension of 135 kN was
increased to 156 and 467 kN in tension due to the faults of 6 and 7 magnitudes,
respectively. However, the total longitudinal stresses (non-linear formulation) for
magnitudes of 6 and 7 are different by about 10 to 20%, and the maximum value
does not exceed 57% SMYS when the fault displacements increase by 65% (from
0.390 to 0.645 m) as seen in Table 2 for normal/reverse-slip vertical component data.
In other words, the pipeline has low sensitivity to vertical fault.
The pipeline expansion results (see Fig. 2) show the pipe string snaking phenom-
enon during operation. Longitudinal ground displacement produces additional
dynamic bending moments along the snaked section of the pipeline as a result of
longitudinal distributed friction load created by axial movement. Hence, the pipeline,
which is already slightly bent on the bottom under the operational expansion load
and lateral friction loads, is additionally loaded longitudinally and consequently will
be bent dynamically. The following additional assumptions and analytical back-
ground are considered:
N.Y. Kershenbaum et al. / Ocean Engineering 27 (2000) 473–487 483
앫 Prior to any fault motion, the pipeline at any location has bending moments and
forces in balance with external friction loads in the horizontal plane. This can be
represented by a sinusoidal shape function as follows:
Zo ⫽ Ao sin 冋 册
n x
L
(12)
where Zo is the pipeline axis deflection in the horizontal plane, Ao is the maximum
amplitude (conservative case), n is the corresponding snaked pipeline mode number,
x is the coordinate, and L is the total length of the pipeline.
앫 Dynamic bending of the pipeline as a result of fault motion is superposed on the
normal operational snaking.
앫 Total friction force acting on a pipe is opposite to the velocity direction with
respect to the ground, consequently:
ftVl
fl ⫽ (13a)
√V2l ⫹ V2a
ftVa
fa ⫽ (13b)
√V ⫹ V2a
2
l
where fl and fa are the lateral and axial friction factors, respectively. ft is the total
friction factor assumed to be 0.7. Vl and Va are the current velocities of the lateral
and axial pipe sections, respectively.
앫 The ground maximum axial velocity is determined by the following expression
(based on the predicted longitudinal ground displacement, of ⌬X, and assumed
half-way constant acceleration/deceleration, of Jx):
Va ⫽ √⌬XJx (14)
앫 Ground fault with constant acceleration allows estimation of ground motion dur-
ation, t1, as follows:
⌬X
t1 ⫽ 2 冪J x
(15)
Dynamic lateral deviation of the pipe segment adjacent to the fault section (highest
loaded section) is described by the following equation (Timoshenko and Gere, 1961):
EId 4(Z ⫺ Zo) d2Z ws d2Z
⫽ ⫺ (Fl ⫺ Fe) ⫺ (16a)
dx4 dx2 g dt2
with the following boundary conditions:
dZ
t ⫽ 0, Z ⫽ Zo, ⫽0 (16b)
dt
where Fl is defined by Eq. (18a), Fe is Euler’s critical compressive load, g is gravity
484 N.Y. Kershenbaum et al. / Ocean Engineering 27 (2000) 473–487
acceleration, and ws is the pipe’s submerged weight. The solution is assumed to have
the following form:
Z ⫽ A(t) sin 冋 册
n X
L
(17)
where A(t) is the dynamic amplitude of the pipe lateral deviation. F1, l, and Fe are
given by the following:
Fl ⫽ Fe ⫹ ws f1l (18a)
L
l⫽ (18b)
n
n22EI
Fe ⫽ (18c)
L2
Pipe bending stress, bl, due to ground longitudinal displacement is determined by
the Hutchinson relationship:
bl ⫽ p 冋 i⑀
⑀p
⫺i⫹1 册 1/i
(19a)
where
d
⑀⫽ (19b)
2
l2
⫽ (19c)
Amax2
Amax ⫽ max[A(t)], if fag ⬍ JX,
冪
4l⌬X
Amax ⫽ 2
⫹ A2o, if fag > JX (19d)
Fig. 7. Oil pipeline under mid-point reverse strike-slip fault during operation (dn ⫽ 300 mm, To ⫽
628 kN, ⌬P ⫽ 14 MPa, M ⫽ 6).
ing stress, bt, compressive/tensile axial stress, f, due to operational and fault
resulting axial force and pressure “cap” stress, c, with corresponding sign (plus or
minus), hence:
l ⫽ max(bt ⫹ f ⫹ c) (21)
The results of the pipe stress/displacement analysis using non-linear pipe proper-
ties are presented in Table 2 (see normal/reverse slip, vertical and longitudinal
components) and also in graph form in Figs. 7 and 8. Extreme stresses are caused
by the longitudinal component of the fault displacement near the restraint and are
associated with the original snaked section of the pipeline. Only simultaneous occur-
Fig. 8. Oil pipeline under mid-point reverse strike-slip fault at shutdown (dn ⫽ 300 mm, To ⫽ 135 kN,
⌬P ⫽ 0 MPa, M ⫽ 6).
486 N.Y. Kershenbaum et al. / Ocean Engineering 27 (2000) 473–487
rence of the above conditions provides over-stress but not rupture of the pipe in
service. It is necessary to remember that the case considered is an extreme emerg-
ency: seismic fault happening directly within pipeline location, near restraint, and
its front is perpendicular to the pipeline axis.
Then, the total longitudinal stress, , due to operational loading and oblique-slip
fault displacements is calculated as the sum of oblique fault bending stress, b, nor-
mal (reverse), na, and strike, sa, faults axial stresses, and also, “cap” pressure stress,
c, and operational stress, o, due to pipeline Euler’ instability.
⫽ b ⫹ na ⫹ sa ⫹ c ⫹ o (23)
All the above calculations were performed using non-linear stress–strain relationship
of the pipe material. The results presented in Table 2 demonstrate very high seismic
viability of an unburied pipeline. Hence, the pipe is not very sensitive to an earth-
quake magnitude increase: oblique-slip fault total longitudinal stress has increased
by approximately 5% when the fault magnitude was changed from 6 to 7 and bottom
displacement increased by 66%. There is a pipe over-stress situation only at the
pipeline restraint vicinity and for already snaked pipeline. This is not surprising since
it was true for normal/reverse-slip. The non-snaked pipeline is not sensitive to the
longitudinal component of the bottom displacement due to seismic fault. It is
important to point out that the fault longitudinal component of the displacement,
generates the major part of the pipeline stress magnitudes when snaked. The last
two rows of Table 2 illustrate the transition case when snaked pipe becomes straight
due to the unrestrained section proximity. The stress values reduce significantly
because of no snaking. However, in reality, there are a lot of pipeline longitudinal
axis shape imperfections that have to stipulate the snaked pipeline approach for seis-
mic stress analysis.
7. Conclusions
1. Unburied pipeline takes place at a certain distance from the pipeline unrestrained
end and gradually increases towards the pipeline restraint.
N.Y. Kershenbaum et al. / Ocean Engineering 27 (2000) 473–487 487
2. Snaking phenomenon does not magnify the resulting longitudinal stress due to
fault compared with initially straight pipeline.
3. The normal or reverse-slip fault’s vertical component does not produce a signifi-
cant amplification of the stress.
4. The major contribution to pipeline seismic loading is generated by longitudinal
displacementof the ground. This effect occurs only for already snaked pipeline
and in proximity to the pipeline restrained section.
5. The straight pipeline is not sensitive to the longitudinal component of the fault.
However, due to imperfections imposed by pipe laying and seabed surface profile,
it is recommended to consider all pipelines as snaked.
References
Dunlap, V.A., Bhojanala, R.P., Morris, D.V., 1990. Burial of vertically loaded offshore pipelines in weak
sediments. 22nd Annual Offshore Technology Conference OTC6375, 263–270.
Figarov, N.G., Kamyshev, A.M., 1996. Seismic stability of offshore pipelines. 6th International Offshore
and Polar Engineering Conference 2, 477–481.
Hobbs, R.E., 1984. In-service buckling of heated pipelines. Journal of Transportation Engineering, ASCE
110, 175–189.
Ju, G.T., Kyriakides, S., 1988. Thermal buckling of offshore pipelines. Journal of Offshore Mechanics
and Arctic Engineering, ASME 110, 355–364.
Kershenbaum, N.Y., Harrison, G.E., Choi, H.S., 1996. Subsea pipeline lateral deviation due to high tem-
perature product. 6th International Offshore and Polar Engineering Conference 2, 74–79.
Nyman, D.J., 1972. Guidelines for the Seismic Design of Oil and Gas Pipeline Systems. American Society
of Civil Engineers, New York.
Rajani, B.B., Robertson, P.K., Morgenstern, N.R., 1993. A simplified design method for pipelines subject
to transverse soil movements. 12th International Conference on Offshore Mechanics and Arctic Engin-
eering 5, 157–165.
Timoshenko, S.P., Gere, J.M., 1961. Theory of Elastic Stability McGraw-Hill, New York.
Trigg, A., Rizkalla, M., 1994. Development and application of a closed form technique for the preliminary
assessment of pipeline integrity in unstable slopes. 13th International Conference on Offshore Mech-
anics and Arctic Engineering 5, 127–139.
Wells, D.L., Coppersmith, K.J., 1994. New empirical relationships among magnitude, rupture length,
rupture width, rupture area, and surface displacement. Bulletin of the Seismological Society of America
84 (4), 974–1002.
Zhow, Z., Murray, D.W., 1993. Behavior of buried pipelines subjected to imposed deformations. 12th
International Conference on Offshore Mechanics and Arctic Engineering 5, 115–126.