Encyclopedia of Earthquake Engineering

DOI 10.1007/978-3-642-36197-5_106-1
# Springer-Verlag Berlin Heidelberg 2014

Seismic Design of Pipelines

Hany El Naggar*
Department of Civil and Resource Engineering, Dalhousie University, Halifax, NS, Canada

Synonyms
Earthquakes; Lifelines; Liquefied soil; PGD; Pipeline breakings; Seismic-induced settlement

Introduction
Pipelines are a cost-effective means for the transportation of water supply and sewers, in addition to
commercial fluids such as oil and gas. The designation of pipeline systems as “lifelines” indicates
that they should be designed to function and operate at all times for public safety and well-being and
also for economic reasons. Engineers must consider the different types of loads that are going to be
imposed on the pipeline, the environment that the pipeline will travel in, and the type of material that
the pipeline is going to convey. Consequently, pipelines must be designed for different loads such as
stresses due to pressure generated by the flow (internal pressure), external pressure generated by the
weight of earth and by live loads for buried pipelines, or external pressure generated by fluid if the
pipe is submerged underwater. Also, seismic loads (earthquake loads) in medium and high seismic-
ity zones are detrimental and should be considered. In general, pipeline design includes several
general steps: (1) load determination, (2) critical performance evaluation (determining the critical
stress and/or deformation), (3) comparison of the critical performance with the limiting criteria
established by adopted codes and standards, and finally (4) selection of the pipe.
Over the past century, several catastrophic earthquakes caused severe damage to buried pipelines.
In the 1906 San Francisco earthquake, one of the main reasons that caused the extensive damage was
due to the failure of several water pipelines which hindered firefighting efforts. Several years later, in
the 1971 San Fernando earthquake, about 80 % of the reported destruction were in welded steel
buried pipelines. Following the 1989 Loma Prieta earthquake, the East Bay Municipal Utilities
District identified over 120 water pipeline breaks, and the San Jose Water Company reported another
155 pipe breaks. However, most of the serious damage reported by Pacific Gas and Electric
Company (PG&E) occurred to natural gas mains and service lines. During the Northridge earth-
quake (1994) in California, some natural gas pipelines were severely damaged and the leak of
containment fluid caused a large explosion in the Northridge town resulted in several deaths (Lau
et al. 1995; O’Rourke and Palmer 1996). In the 1995 Hyogo-Ken Nanbu earthquake (in Japan), the
natural gas leakage from buried pipelines resulted in numerous number of fires (531 cases reported)
which started primarily due to gas release and electricity sparks affecting areas of over one square
kilometer totally burnt (Scawthorn and Yanev 1995). More recently, the 1999 Chi-Chi earthquake
(in Taiwan) also caused severe damage to natural gas distribution systems. More than 100,000
customers were affected after the earthquake, and the estimated economic loss of five major natural
gas companies was approximately US$ 25 million (Chen et al. 2000). In the 1999 Izmit earthquake
(in Turkey), Tupras refinery suffered serious fire damage as the pipeline conveying water from

*Email: [email protected]

Page 1 of 20
Encyclopedia of Earthquake Engineering
DOI 10.1007/978-3-642-36197-5_106-1
# Springer-Verlag Berlin Heidelberg 2014

Fig. 1 Main types of faults

neighbor lake was damaged, so the refinery was dependent on internal water reservoirs which were
insufficient. In 2001 two earthquakes occurred (1 month apart) in San Salvador; several pipelines in
rural areas experienced severe destruction caused by huge landslides. Till now, in several very recent
earthquakes occurred in Chile (2010), New Zealand (Darfield, 2010), Japan (Tohoku, 2011), and
Italy (earthquake of Emilia, 2012), serious damages to pipelines have been witnessed.
Accordingly, pipelines should be designed to function and operate during and following design
earthquakes for life safety and economic reasons. As mentioned above several earthquakes in the last
few decades resulted in too many pipeline breakings, and that extensive pipe breakage has the
potential to lead to great economic harm to our urban communities.
The main focus of this chapter will be centered on the seismic analysis and design of pipelines.

Permanent Ground Deformation and Seismic Wave Propagation Hazards

In seismic events, buried pipelines can be damaged mainly by either the permanent ground
deformations (PGD) or by the transient seismic wave propagation. PGD movements include
faulting, landslides, lateral spreading due to liquefaction, and seismic settlement. Even though
PGD risks are usually restricted to small regions within the pipeline, the chances of them causing
severe damage is substantial since they impose large deformations. On the other hand, the seismic
wave propagation risks typically affect the whole pipeline, but with lower damage rates (as the total
deformations are general less and not permanent). For example, only 5 % of the affected area
experienced lateral spreading during the 1906 San Francisco earthquake; approximately 52 % of all
pipeline breaks occurred within one city block of the lateral spreading (O’Rourke et al. 1985).

Permanent Ground Deformation (PGD)

This section describes in details the four different forms of permanent ground deformation and
presents methods to calculate the amount of PGD as well as the extent of the PGD zone. Equations to
quantify the amount of PGD are provided. Also, useful observations and notes to determine the
extent of the PGD zone are presented and discussed.

Fault
Stresses in the earth’s crust push the two sides of the fault. Eventually enough stress builds up and the
rocks slip suddenly releasing energy in waves that travel through the rock to cause earthquake.
Accordingly, earthquakes occur on faults. A fault is a thin zone of crushed rock separating blocks of
the earth’s crust. When an earthquake occurs on one of these faults, the rock on one side of the fault

Page 2 of 20
Encyclopedia of Earthquake Engineering
DOI 10.1007/978-3-642-36197-5_106-1
# Springer-Verlag Berlin Heidelberg 2014

slips with respect to the other. The fault surface can be vertical, horizontal, or oblique to the surface
of the earth. If the earthquake magnitude is large enough, the offset along the fault will propagate all
the way to the earth’s surface causing surface rupture (fault offset). Figure 1 below shows the main
types of faults. In normal and reverse faults, the major ground displacement is vertical with a minor
horizontal displacement. These ground displacements pose axial tension/compression and bending
stresses in the pipeline depending on the direction of movement. On the other hand, in strike slip
fault the main deformation (the offset) occurs in the horizontal plane, which poses axial tension/
compression and bending stresses in the pipeline depending on the intersection angle of the pipeline
and the fault.
The following empirical equations by Wells and Coppersmith (1994) are the most recognized
equations to estimate the average fault displacement relative to the size of the considered earthquake:

For the normal fault ! log df ¼  4:45 þ 0:63 M (1)

For the reverse fault ! log df ¼  0:74 þ 0:08 M (2)

For the strike slip fault ! log df ¼  6:32 þ 0:90 M (3)

where df is the average fault displacement in meters and M is the moment magnitude of the
earthquake.

Landslides
Seismically induced landslide involves a wide range of downslope mass ground movements, which
can occur in offshore, coastal, and onshore environments.

Offshore Landslides In offshore landslides, the response of the pipeline is mainly governed by the
orientation of the pipeline to the direction of ground movement. The imposed displacements for
undersea slides are so large that pipeline response is likely controlled by the maximum force
available at the soil-pipe interface.

Onshore Landslides There are several types of onshore landslides based on soil movements,
geometry of the slide, and the types of material involved. The main types are rock falls, rock
topples, slides, and lateral spreads. Rock fall and rock topple can cause direct damage to above-
ground pipelines by the impact of falling rock fragments. In an earth slide the earth moves relatively
as a block; they typically develop along natural slopes and embankments.
Based on Newmark’s Block model for landslides (Newmark 1965), the critical acceleration at
which the slide will be triggered ac can then be determined from

ac ¼ g ðFOS  1Þ sin a (4)

where g is the acceleration due to gravity, FOS is the factor of safety, and a is the angle of the slope.
Then the displacement of the block can be calculated by double integrating the ground acceler-
ation. Jibson and Keefer (1993) proposed the following equation to estimate the Newmark displace-
ment, ds, in centimeters as

Log dS ¼ 1:460 log I a  6:642 ac þ 1:546 (5)

Page 3 of 20
Encyclopedia of Earthquake Engineering
DOI 10.1007/978-3-642-36197-5_106-1
# Springer-Verlag Berlin Heidelberg 2014

B B

A
A Y = 100 A/B
S = 100 A/B Slip surface

Ground slope Free face

Fig. 2 Sketch showing the ground slope and the free face lateral spreads

where ac is the critical acceleration in gs and Ιa is the Arias Intensity in m/s.

Arias Intensity can be calculated using the following relationship developed by Wilson and
Keefer (1983):

Log I a ¼ M  2 log R  4:1 (6)

where M is the earthquake magnitude and R is source distance in kilometers.

Lateral Spreading Due to Liquefaction

Seismic shaking may trigger the liquefaction of saturated loose cohesionless soils. The liquefaction
process increases the pore water pressure in the ground to a level where the effective stress
approaches zero at which point the soil loses entirely its shear strength, which in turn results in
the lateral movement of the soil (lateral spreading). In past major earthquakes, large number of
pipelines suffered massive damage caused by liquefaction-induced ground failures (Zhang
et al. 2004).
Youd et al. (2002) proposed two empirical equations to approximately estimate the PGD due to
liquefaction:
For lateral spreads down gentle ground slopes (GS),

LogdL ¼  16:213 þ 1:532 M  1:406 log R  0:012 R þ 0:338 log S þ 0:54 log T 15 þ
(7)
3:413 logð100  F 15 Þ  0:795 logðD5015  0:1Þ

For lateral spreads at a free face (FF),

Log dL ¼  16:710 þ 1:532 M  1:406 log R  0:012 R þ 0:592 log Y þ 0:54 logT 15 þ
(8)
3:413 logð100  F 15 Þ  0:795 logðD5015  0:1Þ

where dL is the PGD in meters, M is the earthquake magnitude, S is the ground slope %, g is the free
face ratio (in % see Fig. 2), T15 is the thickness in meters of the saturated cohesionless soil layer with
a corrected standard penetration test (SPT value of less than 15), F15 is the percentage of average
fines contents in T15 (in %), D5015 is the mean grain size in mm in T15, and R* is an adjusted distance
parameter in kilometers given by

R ¼ R þ 10 eð0:89M 5:64Þ (9)

where R is the horizontal distance in kilometers from the site of interest to the nearest bound of the
seismic energy source (do not use less than 0.5 km).

Page 4 of 20
Encyclopedia of Earthquake Engineering
DOI 10.1007/978-3-642-36197-5_106-1
# Springer-Verlag Berlin Heidelberg 2014

Seismic Settlement
Seismic-induced settlement may be caused by densification of cohesionless soils, consolidation of
cohesive soils, or consolidation of liquefied soil. In this section we will only discuss liquefaction-
induced ground settlement as it can cause larger settlement and hence higher potential for damage to
buried pipelines.
Takada and Tanabe (1988) proposed the two following empirical equations to calculate
liquefaction-induced settlement at embankments and leveled sites:
For embankments
amax
dGS ¼ 0:11 H 1 H 2 þ 20 (10)
N
For leveled sites
amax
dGS ¼ 0:30 H 1 þ2 (11)
N
where dGS is the liquefaction-induced settlement in centimeters, H1 is the thickness of saturated
cohesionless soil layer in meters, H2 is the height of embankment in meters, N is the SPT N-value in
the cohesionless layer, and amax is the ground acceleration in gals.

Seismic Wave Propagation

For the seismic analysis and design of buried pipelines, the effect of the seismic wave propagation on
the pipeline is usually characterized by the induced ground strain and curvature. Newmark (1967)
developed a straightforward method to estimate the ground strain. The general form of a traveling
wave in Newmark’s method is given by

t x
U ¼f þ (12)
T l
where U is the function of the separation distance between the two points, x, and the speed of the
seismic wave; T is the period of the repeating motion; and l is the wavelength.
For particle motion parallel to the direction of propagation (R-waves), eg, the ground strain along
the direction of propagation can be calculated as

V max
eg ¼ (13)
CR

where Vmax is the maximum horizontal ground velocity in the direction of wave propagation and CR
is the propagation velocity of the R-wave.
For particle motion perpendicular to the direction of propagation (S-waves), Cg, the curvature can
be calculated as

Amax
Cg ¼ (14)
CS 2

where Amax is the maximum ground acceleration perpendicular to the direction of wave propagation
and Cs is the propagation velocity of the S-wave.

Page 5 of 20
Encyclopedia of Earthquake Engineering
DOI 10.1007/978-3-642-36197-5_106-1
# Springer-Verlag Berlin Heidelberg 2014

Wave propagation with particle motion parallel to the pipeline (R-wave) would induce large axial
strain in the pipeline. On the other hand, S-waves propagating parallel to the pipeline with particle
motion perpendicular to the pipeline direction would induce only bending strains.
For S-waves traveling in a horizontal plane and at an angle with respect to the pipeline, the S-wave
particle motion (perpendicular to its direction of propagation) would have one component parallel to
the pipeline inducing axial strain, eg, and another component perpendicular to the pipeline inducing
bending strain, eb, which can be calculated as

V max
eg ¼ (15)
2 Cs

and

pD V max
eb ¼ (16)
l Cs

where eb is the upper bound bending strain (maximum), Vmax is the peak ground velocity, Cs is the
apparent propagation velocity of the S-wave, l is the wavelength, and D is the pipe diameter.
The bending strains induced in a pipeline due to traveling waves (S-waves, L-waves, or the
vertical component of R-waves) are generally small compared to the axial strains induced in the
ground by the traveling waves.

Pipelines Response to Faulting

This section presents the response of continuous pipelines subject to fault offsets. PGD due to
faulting can be resolved into two components: longitudinal PGD (parallel to the pipeline) and
transverse PGD (perpendicular to the pipeline) axis. In the case of “normal fault” type (see Fig. 1),
the pipeline will be subjected to bending axial tensile force, caused by the transverse and longitu-
dinal components, respectively. In this case, tensile rupture would be the most probable failure
mechanism. In the “reverse fault”-type case, the pipeline will be subjected to bending axial
compressive force, caused again by the transverse and longitudinal components, respectively. In
this case, buckling would be the most likely failure mechanism. In the last case of “strike slip fault,”
the pipeline can be subjected to either tension or compression depending on the intersection angle
between the pipeline and the fault and pipe and the relative movement at the fault.
Kennedy et al. (1977) proposed a simplified method to analyze the tensile and bending behavior
of pipelines due to fault movements. Figure 3 shows the Kennedy et al. model for one side of the
fault.
According to Kennedy et al. (1977), the total strain in the pipe (bending + tensile) is given by

DL D
e ¼ ea þ eb ¼ þ (17)
L 2Rc

where DL is the total elongation of the pipeline, L is the total length of the pipeline, D is the pipe
diameter, and Rc is the radius of curvature of the curved portion. Rc can be estimated by

Page 6 of 20
Encyclopedia of Earthquake Engineering
DOI 10.1007/978-3-642-36197-5_106-1
# Springer-Verlag Berlin Heidelberg 2014

Lc

Tu d f sin b /2

Pu

Rc

Fig. 3 Kennedy et al. (1977) model for one side of the fault

spDt
Rc ¼ (18)
Pu

where s is the axial stress at fault crossing and Pu is the peak lateral pipeline-soil interaction force per
unit length which can be calculated using the 1984 ASCE guideline using the following relations for
sand and clay, respectively:
For sand

Pu ¼ g H N qh D (19)

For clay

Pu ¼ cu N ch D (20)

where g is the unit weight of the soil, H is the embedment depth of the pipeline, and Nqh and Nch are
the horizontal bearing capacity factors for sand and clay, respectively. Figure 4 presents horizontal
bearing capacity factors for sand and clay after Hansen (1961).
The total elongation of the pipeline, DL, can be estimated using the following equation:
 2
df sin b
DL ¼ df cos b þ (21)
3Lc

where df is the average fault displacement, b is the fault angle, and Lc is the horizontal projection
length of the laterally deformed pipeline (see Fig. 3).
Lc can be approximately calculated using the following simplified equation:
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Lc ¼ Rc df sin b (22)

Pipelines Response to Longitudinal PGD

This section presents the response of continuous pipelines subject to longitudinal permanent ground
deformations, PGD, where the soil movement is in the same direction as the pipeline. O’Rourke
et al. (1995) proposed an inelastic pipeline model to analyze the response of pipelines to longitudinal

Page 7 of 20
Encyclopedia of Earthquake Engineering
DOI 10.1007/978-3-642-36197-5_106-1
# Springer-Verlag Berlin Heidelberg 2014

Nqh Nch
2 10 50 2 10
0

Depth to Diameter Ratio H/D 4

8
j = 20° j = 0°
10
25°
12
30°
14
35°
16
40°
18
45°
20
Sand Clay

Fig. 4 Horizontal bearing capacity factors for sand and clay (Reproduced after Hansen 1961)

PGD. Figure 5 shows the considered model. To develop this model O’Rourke et al. utilized an
idealized block pattern in which a mass of soil having length L moves down a slight incline. The soil
displacement on either side of the PGD zone is zero, while the soil displacement within the zone is a
constant value d.
In O’Rourke et al. model, a block of soil between Points B and D moves to the right pulling the
pipe laterally with it with the soil forces acting on the pipeline within the PGD zone to the right. On
the other hand, the soil between Points A and B and the soil between Points D and E resist the
pipeline movement and these soil restraint forces are directed to the left. The combined pipeline-soil
interaction forces result in a region of pipe axial tension between points A and B and a region of axial
compression between points D and E. The conditions outlined in Fig. 5 correspond to the case where
the PGD, d, is comparatively large and the length of the PGD zone, L, is comparatively short. In that
case, the maximum pipe displacement is less than the ground displacement and the pipe strain is
controlled by L.
Figure 6 presents the other possibility of the O’Rourke et al. model where the length of the PGD
zone is relatively large while the amount of PGD is comparatively short. Also in this case, there is
still axial pipe tension between points A and B and axial compression between points D and E;
however, the zone is long enough that the pipe displacement matches that of the ground between
Points C and D where the axial force and strain in the pipe are zero.
As it can be seen from Figs. 5 and 6, the axial force in the pipeline in the segment AB is linearly
proportional to the distance from Point A. Accordingly, the pipelines strain and displacement can be
evaluated using the following Ramberg-Osgood model relations:
 
bp x n bp x r
eðxÞ ¼ 1þ (22)
E 1 þ r sy

Page 8 of 20
Encyclopedia of Earthquake Engineering
DOI 10.1007/978-3-642-36197-5_106-1
# Springer-Verlag Berlin Heidelberg 2014

Fig. 5 O’Rourke et al. model to analyze the response of pipelines to longitudinal PGD

 
bp x2 2 n bp x r
dðxÞ ¼ 1þ (23)
E 2 þ r 1 þ r sy

where n and r are Ramberg-Osgood parameters (given in Table 1 below), E is the modulus of
elasticity of steel, sg is the effective yield stress, and bp is the pipe burial parameter, defined as the
friction force per unit length tu divided by the pipe cross-sectional area A.
The pipe burial parameter bp can be obtained from the following:
For sand

tan ’ g H
bp ¼ (24)
t

Page 9 of 20
Encyclopedia of Earthquake Engineering
DOI 10.1007/978-3-642-36197-5_106-1
# Springer-Verlag Berlin Heidelberg 2014

Fig. 6 O’Rourke et al. model to analyze the response of pipelines to longitudinal PGD (PGD zone is relatively large,
while the amount of PGD is comparatively short)

Table 1 Yield stress and Ramberg-Osgood parameters for mild steel and X-grade steel
Grade B X-42 X-52 X-60 X-70 X-80
Yield stress (MPa) 241 290 359 414 483 552
n 10 15 9 10 5.5 16
r 100 32 10 12 16.6 16

Page 10 of 20
Encyclopedia of Earthquake Engineering
DOI 10.1007/978-3-642-36197-5_106-1
# Springer-Verlag Berlin Heidelberg 2014

Fig. 7 Adhesion factors (Reproduced after Honegger and Nyman 2004)

For clay
a cu
bp ¼ (25)
t
where ’ is the angle of shear resistance, g is the effective unit weight of the soil, H is the embedment
depth of the pipeline, a is the adhesion factor for clay, cu is the undrained cohesion of the clay (see
Fig. 7 below), and t is the pipe wall thickness.

Wrinkling of the Pipe Wall in Compression

Substituting a critical local buckling strain into Eq. 22, one can obtain the critical length of PGD
zone Lcr. This can then be used to calculate the critical ground movement dcr from Eq. 23. The
critical strain in compression may be taken as 0.175 t/R.

Pipeline Response to Transverse PGD

This section presents the response of continuous pipelines subject to transverse permanent ground
deformations, PGD, where the soil movement is perpendicular to the pipeline. When subjected to
transverse PGD, the pipeline will stretch and bend as it attempts to conform to the transverse ground
movement profile. In this case, the failure mechanism of the pipeline will be governed by the relative
magnitude of the axial tensile strain and the bending strain. If the tensile strain is relatively small, the
pipe wall may buckle in compression due to excessive bending. Instead, if tensile strain is relatively
large, the pipe may rupture in tension due to the combined effects of the tensile and bending stresses.
In general, the response to transverse PGD is a function of the magnitude of PGD, the width of the
PGD zone, and the pattern of ground deformation. Two types of transverse ground deformation
patterns are discussed here: the spatially distributed transverse PGD pattern and the abrupt trans-
verse PGD pattern. Figure 8 shows sketch of the considered patterns.

Page 11 of 20
Encyclopedia of Earthquake Engineering
DOI 10.1007/978-3-642-36197-5_106-1
# Springer-Verlag Berlin Heidelberg 2014

Fig. 8 Sketch of the considered patterns

Fig. 9 O’Rourke (1989) model to analyze the response of pipelines to spatially distributed transverse PGD

Spatially Distributed Transverse PGD

O’Rourke (1989) proposed a simple model to analyze the response of pipelines to spatially
distributed transverse PGD. In this model O’Rourke considered two types of response (wide and
narrow width PGD zones) as shown in Fig. 9 below. In the wide width PGD zone, the pipeline is
relatively flexible and its lateral displacement is assumed to closely conform to the soil outline.
Accordingly, the pipeline strain is expected to be mainly due to the ground curvature (i.e.,
displacement controlled). On the other hand, for the narrow width PGD case, the pipeline is
relatively stiff and the pipeline lateral displacement is significantly smaller than that of the soil.
Hence, the pipeline strain is anticipated to be due to loading at the pipeline-soil interface (i.e., load
controlled).

Page 12 of 20
Encyclopedia of Earthquake Engineering
DOI 10.1007/978-3-642-36197-5_106-1
# Springer-Verlag Berlin Heidelberg 2014

The maximum bending strain, eb, in the pipeline is given by the following:

– For the wide width PGD zone

p2 dD
eb ¼  (26)
W2

– For the narrow width PGD zone

pu W 2
eb ¼  (27)
3pEtW 2

And the average axial tensile strain, ea, in the pipeline is estimated by the following:

– For the wide width PGD zone

p2 d 2
ea ¼ (28)
2 W

– For the narrow width PGD zone

The axial tension in this case is small and neglected.

where d is the magnitude of the PGD, D is the pipe diameter, W is the length of the PGD zone, Pu
is the maximum lateral force per unit length at the pipeline-soil interface, E is the elastic modulus of
the pipeline material, and t is the pipe wall’s thickness.

Abrupt Transverse PGD

Parker et al. (2008) proposed a simple model to analyze the response of pipelines to transverse
PGD. Figure 10 presents the geometric and force details of the model. In this model the width of the
abrupt transverse PGD is 2W1. Within this width the pipeline is subject to a lateral force per unit
length Pu1. This lateral load is resisted by soil resistance forces Pu2 over a distance W2 on each side of
the abrupt transverse PGD zone (see Fig. 10). Therefore, from horizontal equilibrium in the direction
of pipeline, we get

Pu1 W 1 ¼ Pu2 W 2 (29)

The tensile force in the pipeline is assumed to be a constant value To within the PGD zone. Beyond
the margins the pipeline axial tension decreases linearly at Points C and E to zero at Points A and
G (see Fig. 10).
According to Parker et al. (2008), the total elongation due to the pipeline deformation (Point B to
Point F) can be estimated using the following equation:

Page 13 of 20
Encyclopedia of Earthquake Engineering
DOI 10.1007/978-3-642-36197-5_106-1
# Springer-Verlag Berlin Heidelberg 2014

Fig. 10 Geometric and force details of the Parker et al. (2008) model

1 2
DL ¼ P W 3 ð1 þ Pu1 =Pu2 Þ =T o 2 (30)
3 u1 1
Thus, the axial pipe strain is calculated as

DL
ea ¼ (31)
L
and the bending strain in the pipeline can be evaluated as

D Pu1
eb ¼ (32)
2 To

Pipelines in Liquefied Soil

It is common practice that pipelines are buried at depths of 2 m or less from the ground surface. Thus,
the top of the susceptible liquefiable soil layer is commonly located below the bottom of the pipeline.
However, in some cases when the pipeline is buried at a river bed in saturated sand, for example, the
soil surrounding the pipeline may liquefy during strong seismic shaking event. In this case, the
pipeline will probably deform laterally following the flow of the liquefied soil downward a mild
slope or move upward due to buoyancy, particularly when something restrains the pipeline at one
point or a compressive load acts on the pipeline.

Horizontal Movement
When a pipeline is surrounded by liquefied soil, the pipeline may move laterally due to the flow of
liquefied soil downslope. The response of a buried pipe surrounded by liquefied soil subject to
spatially distributed transverse PGD can be analyzed using the O’Rourke (1989) method presented
earlier.

Page 14 of 20
Encyclopedia of Earthquake Engineering
DOI 10.1007/978-3-642-36197-5_106-1
# Springer-Verlag Berlin Heidelberg 2014

Fig. 11 Profile of pipeline crossing liquefied zone

Vertical Movement
When a pipeline is surrounded by liquefied soil, the pipeline may uplift due to the buoyancy and
moves upward. Hou et al. (1990) proposed analytical method to analyze the response of pipelines
subjected to vertical movements. According to Hou et al. (1990), the uplifting force per unit length,
Puplift, acting on the pipeline within the liquefied zone is given by

p D2
Puplift ¼ ðgsoil  gcontents Þ  p D t gpipe (33)
4
where D is the pipe diameter, gsoil is the unit weight of the liquefied soil, gcontents is the unit weight of
the contents inside the pipe (water, oil, gas, etc.), gpipe is the unit weight of the pipe material, and t is
the pipe wall’s thickness.
The maximum uplift displacement and/or the spacing for pipeline restraints is given by (see
Fig. 11 below)

16Puplift W s 4
A dmax 3
þ 16 I dmax  (34)
E p5
and the maximum strain in the pipeline is then given by

p2 dmax D p2 dmax 2
emax ¼ þ (35)
W s2 4 W s2

where A is the cross-section area, I is the moment of inertia, Ws is the spacing of the restraints, and
E is the elastic modulus of the pipe’s material.

Seismic Design Guidelines and Pipeline-Soil Interaction

The pipeline-soil interaction effect exerts external forces on the buried pipeline by the adjacent soil
when seismic motions are provided. The type of soil around the pipeline plays a significant role in its
seismic behavior. In general, the soil displacement will produce friction like forces at the pipeline-
soil interface (see Fig. 12 below). The overall seismic performance of buried pipeline is strongly
related to the force-deformation relationship at the pipeline-soil interface (the p-y curves). For
cohesionless soils, the probability of liquefaction becomes higher in loose materials. In cohesive
soils, softer soils will undergo greater differential settlement due to consolidation and higher
amplification effects and accordingly greater interaction forces.

Page 15 of 20
Encyclopedia of Earthquake Engineering
DOI 10.1007/978-3-642-36197-5_106-1
# Springer-Verlag Berlin Heidelberg 2014

Fig. 12 External forces on the buried pipeline by the adjacent soil during seismic shaking

Fig. 13 Idealized elastoplastic models for the force-deformation relationship at the pipeline-soil interface (Reproduced
after ASCE 1984)

Pipelines in Competent Non-liquefied Soil

The 1984 ASCE Guidelines suggest for the purpose of analysis idealized elastoplastic models for the
force-deformation relationship at the pipeline-soil interface (see Fig. 13 below).

Longitudinal Movement
In this case the relative movement is parallel to the pipeline which results in axial forces at the
pipeline-soil interface. The1984 ASCE Guidelines provide relations for both cohesionless and
cohesive soils. For cohesionless materials, the longitudinal resistance is due to the friction in the
longitudinal direction at the pipeline-soil interface. The normal pressure which leads to the axial
friction is the overburden and the lateral soil pressures. In the 1984 ASCE Guidelines, the normal
pressure is taken as the average of the vertical and at rest lateral soil pressures acting on the pipeline.
(The 1984 ASCE Guidelines assumes X u ffi 2:5  5:0 mm).

Page 16 of 20
Encyclopedia of Earthquake Engineering
DOI 10.1007/978-3-642-36197-5_106-1
# Springer-Verlag Berlin Heidelberg 2014

 
1 þ ko
tu ¼ p D g H tan k’ (36)
2

For cohesive materials, the longitudinal resistance is proportional to the adhesion at the pipeline-soil
interface:

t u ¼ p D a cu (37)

where D is the pipe diameter, g is the effective unit weight of the soil, H is the depth of the pipeline, ’
is the angle of shear resistance, ko is the coefficient of lateral soil pressure at rest, k is a friction factor,
a is the adhesion factor (given in Fig. 7), and cu is the undrained shear strength of the soil.

Transverse-Horizontal Movement
In this case the relative movement is perpendicular to the pipeline which results in transverse-
horizontal forces at the pipeline-soil interface. The 1984 ASCE Guidelines provide relations for both
cohesionless and cohesive soils. For cohesionless materials, the maximum soil resistance in
horizontal transverse direction may be calculated using the following equation:

Pu ¼ g H D N qh (38)

and the maximum elastic relative displacement in horizontal transverse direction is

8
<ð0:07  0:10ÞðH þ D=2Þ f or loose sand
>
yu ¼ ð0:03  0:05ÞðH þ D=2Þ f or medium sand (39)
>
:
ð0:02  0:03ÞðH þ D=2Þ f or dense sand

For cohesive materials, the maximum soil resistance in horizontal transverse direction may be
evaluated using the following equation:

Pu ¼ cu N ch D (40)

The maximum elastic relative displacement in horizontal transverse direction is

yu ¼ ð0:03  0:05ÞðH þ D=2Þ (41)

where g is the unit weight of the soil, H is the embedment depth of the pipeline, and Nqh and Nch are
the horizontal bearing capacity factors for sand and clay, respectively (Fig. 4).

Pipelines in Liquefied Soil

The response of continuous pipelines buried in liquefied soil layer is very sensitive to the stiffness of
the soil (Suzuki et al. 1988; Miyajima and Kitaura 1989). Based on several experimental results, it is
recommended that the stiffness of liquefied soil ranges from 1/100 to 3/100 of that for non-liquefied
soil (Yoshida and Uematsu 1978; Matsumoto et al. 1987; Yasuda et al. 1987; Tanabe 1988).
Accordingly, the reduced stiffness will be used at the pipeline-soil interface. Analysis is then
performed using the same procedures as in the competent non-liquefied soil case (on the
conservative side).

Page 17 of 20
Encyclopedia of Earthquake Engineering
DOI 10.1007/978-3-642-36197-5_106-1
# Springer-Verlag Berlin Heidelberg 2014

Summary
This chapter focused on the seismic analysis and design of pipelines. In seismic events, buried
pipelines can be damaged mainly by either the permanent ground deformations (PGD) or by the
transient seismic wave propagation. Hence, different analysis methods for pipelines subjected to
permanent ground deformation (PGD) or transient seismic wave propagation hazards were
presented and discussed in details. In addition, the response of continuous pipelines subject to
fault offsets was discussed. Also, several methods were introduced to predict the response of
pipelines to either longitudinal or transverse PGD.
Earthquake shaking may trigger the liquefaction of saturated loose cohesionless soils. The
liquefaction process increases the pore water pressure in the ground to a level where the effective
stress approaches zero at which point the soil loses entirely its shear strength. A separate section in
this chapter presented and discussed methods of analyzing and designing pipelines buried in
liquefiable soils.
The overall seismic performance of buried pipelines is strongly related to the pipeline-soil
interaction. Consequently, this chapter dedicated a section to provide guidelines for the seismic
design of pipelines considering the pipeline-soil interaction effects as it plays a significant role in its
seismic behavior.

Cross-References
▶ Earthquake Magnitude Estimation
▶ Earthquake Mechanisms and Tectonics
▶ Earthquake Response Spectra and Design Spectra
▶ Liquefaction: Lateral Spreading
▶ Liquefaction: Performance of Buried Pipelines
▶ Seismic Vulnerability Assessment: Lifelines
▶ Soil-Structure Interaction

References
American Society of Civil Engineers (ASCE) (1984) Guidelines for the seismic design of oil and gas
pipeline systems. Committee on Gas and Liquid Fuel Lifelines, ASCE
Chen WW, Shih BJ, Wu CW, Chen YC (2000) Natural gas pipeline system damages in the Ci Ci
earthquake (The City of Nantou). In: Proceedings of the 6th international conference on seismic
zonation, Palm Springs, California.
Hansen JB (1961) The ultimate resistance of rigid piles against transversal forces. Bulletin 12, Dan-
ish Geotechnical Institute, Copenhagen
Honegger, D.G. and Nyman, D. (2004). PRCI guidelines for the seismic design and assessment of
natural gas and liquid hydrocarbon pipelines. Pipeline Research Council International, Published
by Technical Toolboxes, Houston, Texas
Hou Z, Cai J, Liu X (1990) Response calculation of oil pipeline subjected to permanent ground
movement induced by soil liquefaction. In: Proceedings of the China-Japan symposium on
lifeline earthquake engineering, Beijing, pp 107–114

Page 18 of 20
Encyclopedia of Earthquake Engineering
DOI 10.1007/978-3-642-36197-5_106-1
# Springer-Verlag Berlin Heidelberg 2014

Jibson RW, Keefer DK (1993) Analysis of the seismic origin of landslides: examples from the new
Madrid seismic zone. Geol Soc Am Bull 105:521–536
Kennedy RP, Chow AW, Williamson RA (1977) Fault movement effects on buried oil pipeline.
J Transp Eng Div ASCE 103(TE5):617–633
Lau DL, Tang A, Pierre J-R (1995) Performance of lifelines during the 1994 Northridge earthquake.
Can J Civ Eng 22:438–451
Matsumoto H, Sasaki Y, Kondo M (1987) Coefficient of subgrade reaction on pile in liquefied
ground. In: Proceedings of the second national conference on soil mechanics and foundation
engineering, Tokyo, Japan, pp 827–828 (in Japanese)
Miyajima M, Kitaura M (1989) Effects of liquefaction-induced ground movement on pipeline. In:
Proceedings of the Second U.S.-Japan workshop on liquefaction, large ground deformation and
their effects on lifelines, Buffalo, Technical report NCEER-89-0032, Multidisciplinary Center for
Earthquake Engineering Research, Buffalo, pp 386–400
Newmark NM (1965) Effects of earthquakes on dams and embankments. Geotechnique
15(2):139–160
Newmark NM (1967) Problems in wave propagation in soil and rocks. In: Proceedings of the
international symposium on wave propagation and dynamic properties of earth materials,
University of New Mexico Press, pp 7–26
O’Rourke MJ (1989) Approximate analysis procedures for permanent ground deformation effects
on buried pipelines. In: Proceedings of the second U.S.-Japan workshop on liquefaction, large
ground deformation and their effects on lifelines, Buffalo. Technical report NCEER-89-0032,
Multidisciplinary Center for Earthquake Engineering Research, Buffalo, pp 336–347
O’Rourke TD, Palmer MC (1996) Earthquake performance of gas transmission pipelines. Earthq
Spectra 12(3):493–527
O’Rourke TD, Grigoriu MD, Khater MM (1985) Seismic response of buried pipes. Pressure vessel
and piping technology – a decade of progress, ASME, pp 281–323
O’Rourke MJ, Liu XJ, Flores-Berrones R (1995) Steel pipe wrinkling due to longitudinal permanent
ground deformation. J Transp Eng 121(5):443–451
Parker E, Moore R, Evans T, Usher N (2008) Evaluation of landslide impact on deepwater
submarine pipelines. OTC 19459, Offshore Technology Conference, Houston
Scawthorn C, Yanev PI (1995) Preliminary report 17 January 1995, Hyogoken Nambu, Japanese
earthquake. Eng Struct 17(3):14657
Suzuki H (1988) Damage to buried pipes caused by large ground displacement. In: Proceedings of
the first Japan-U.S. workshop on liquefaction, large ground deformation and their effects on
lifeline facilities, Tokyo, pp 127–132
Takada S, Tanabe K (1988) Estimation of earthquake induced settlements for lifeline engineering.
In: Proceedings of the ninth world conference earthquake engineering, vol VII, pp 109–114
Tanabe K (1988) Fundamental study on seismic assessment and design of buried pipelines subjected
to ground failure during earthquake. Doctoral dissertation, Kobe University (in Japanese)
Wells DL, Coppersmith KJ (1994) New empirical relationships among magnitude, rupture length,
rupture width, rupture area, and surface displacement. Bull Seismol Soc Am 84(4):974–1002
Wilson RC, Keefer DK (1983) Dynamic analysis of a slope failure from the 6 August 1979 Coyote
Lake, California earthquake. Bull Seismol Soc Am 73(3):863–877
Yasuda S, Saito K, Suzuki N (1987) Soil spring constant on pipe in liquefied ground. Proceedings of
the nineteenth JSCE conference on earthquake engineering, Tokyo, Japan, pp 189–192
(in Japanese)

Page 19 of 20
Encyclopedia of Earthquake Engineering
DOI 10.1007/978-3-642-36197-5_106-1
# Springer-Verlag Berlin Heidelberg 2014

Yoshida T, Uematsu M (1978) Dynamic behavior of a pile in liquefaction sand. Proceedings of the
fifth Japan earthquake engineering symposium, Tokyo, Japan, pp 657–663 (in Japanese)
Youd TL, Hansen C, Bartlett S (2002) Revised multilinear regression equations for prediction of
lateral spread displacement. J Geotech Geoenviron Eng ASCE 128(12):1007–1017
Zhang G, Robertson PK, Brachman RWI (2004) Estimating liquefaction-induced lateral displace-
ments using the standard penetration test or cone penetration test. J Geotech Geoenviron Eng
ASCE 130:861–871

Page 20 of 20