Seismic behavior of underground structures and site response

Y.M.A. Hashash

University of Illinois at Urbana-Champaign, Urbana, IL, USA

ABSTRACT: The past 20 years have seen major progress in understanding of seismic response of underground structures. Seismic design loads for underground structures are characterized in terms of deformations
and strains imposed by the surrounding ground. In contrast, above-ground structures are designed for inertial
forces imposed by ground accelerations. Design methods have been developed to estimate seismic loads on
underground structures. These methods include pseudo-static analysis whereby free-field motion is imposed
on the underground structure as well as dynamic soil-structure interaction analysis. Free-field ground deformations and velocities due to a seismic event are estimated using site specific response analysis that accounts
for local geology. Historically, underground facilities have experienced a lower rate of damage than aboveground structures. However, recent failures of underground structures during recent earthquakes were reminders of the need to include seismic considerations in design details of underground structures.
1 INRODUCTION
Underground facilities are an integral part of the infrastructure in an increasingly urban world. Underground space is used for subways, highways, material storage and water and sewage transport.
Underground facilities built in areas subject to
earthquake activity must withstand both seismic and
static loading. Historically underground facilities
have experienced a lower rate of damage than have
above-ground structures. Nevertheless, some underground structures have experienced significant damage in recent large earthquakes.
Underground structures have features that make
their seismic behavior distinct from most aboveground structures, most notably (1) their complete
enclosure in soil or rock and (2) their significant
length (i.e., tunnels). This paper focuses on relatively large underground facilities commonly used in
urban areas. This paper does not discuss pipelines or
sewer lines that are conventionally labeled lifelines.
Large underground structures can be grouped into
three broad categories, each having distinct design
features and construction methods: (1) bored or
mined tunnels, (2) cut-and-cover tunnels, and (3)
immersed tube tunnels (Power, et. al, 1996). These
structures are commonly used for metro structures,
highway tunnels, and large water and sewage transportation ducts.

2 FRAMEWORK FOR CHARACTERIZING

SEISMIC BEHAVIOR OF UNDERGROUND
STRUCTURES
Earthquake effects on underground structures can be
grouped into two categories: (1) ground shaking and
(2) ground failure such as liquefaction, fault displacement, and slope instability. Ground shaking,
which is the primary focus of this paper, refers to deformation of the ground produced by seismic waves
propagating through the earth's crust. The major factors influencing shaking damage include: (1) the
shape, dimensions, and depth of the structure; (2) the
properties of the surrounding soil or rock; (3) the
properties of the structure; and (4) the severity of the
ground shaking (Dowding and Rozen, 1978, St. John
and Zahrah, 1987).
Seismic design of underground structures is
unique in several ways. For most underground structures, the inertia of the surrounding soil is large relative to the inertia of the structure. Measurements
made by Okamato, et al. (1973) of the seismic response of an immersed tube tunnel during several
earthquakes show that the response of a tunnel is
dominated by the surrounding ground response and
not the inertial properties of the tunnel structure itself. The focus of underground seismic design,
therefore, is on the free-field deformation of the
ground and its interaction with the structure. This
concentration on displacement is in stark contrast to

the design of above-ground structures, which focuses on inertial effects of the structure itself.
Three types of deformations (Owen and Scholl,
1981) express the response of underground structures to seismic motions: (1) axial extension and
compression, (2) longitudinal bending, and (3) ovaling/racking (Figure 1). Axial deformations in tunnels
are generated by the components of seismic waves
that produce motions parallel to the axis of the tunnel and cause alternating compression and tension.
Bending deformations are caused by the components
of seismic waves producing particle motions perpendicular to the longitudinal axis. The design of a
tunnel lining to accommodate axial and bending deformations generally concentrates on the direction
along the tunnel axis (Wang, 1993).
Ovaling or racking deformation in a tunnel structure develops when shear waves propagate normal or
nearly normal to the tunnel axis, resulting in a
distortion of the cross-sectional shape of the tunnel
lining. Design considerations for this type of
deformation are in the transverse direction. The general behavior of the lining may be simulated as a
buried structure subject to ground deformations under a two-dimensional plane-strain condition.
Diagonally propagating waves subject different
parts of the structure to out-of-phase displacements
(Figure 1), resulting in a longitudinal compressionrarefaction wave traveling along the structure. In
general, larger displacement amplitudes are associated with longer wavelengths, while maximum curvatures are produced by shorter wavelengths with
relatively small displacement amplitudes (Kuesel,
1969).
3 SITE RESPONSE AND FREE FIELD
DEFORMATIONS
Deformations and velocities in the surrounding
ground dominate the estimate of seismic demand on
an underground structure. A site-specific response
analysis is needed to estimate expected free field
ground deformations and velocities. It has long been
recognized by researchers that local soils modify the
ground motion.
One-dimensional site response analysis is used to
solve the problem of vertical propagation of horizontal shear waves (SH waves) through a horizontally
layered soil deposit. Seed, Idriss and co-workers introduced the equivalent linear approximation method
to capture non-linear cyclic response of soil. For a
given ground motion time series (T.S., also referred
to as time history) and an initial estimate of modulus
and damping values, an effective shear strain (equal
to about 65% of peak strain) is computed for a given
soil layer. Modulus degradation and damping curves
are then used to obtain revised values of shear
modulus and damping. Solution of wave propagation

equations is performed in the frequency domain. An

iterative scheme is required to arrive at a converged
solution. This approach has provided good results
compared with field measurements and is widely
used in engineering practice (e.g. SHAKE, Schnabel
et. al., 1972, FLUSH, Lysmer et. al., 1975). More
recently Sugito (1995) and Dominic et al. (2000) extended the equivalent linear approach to include frequency and pressure dependence of soil properties.
The equivalent linear approach is computationally
easy to use and implement. However, it does not
capture the full range of cyclic behavior of soil, including modulus degradation due to number of loading cycles, permanent (residual) straining of soil and
excess pore pressure generation. Non-linear analysis
is used to capture these important aspects of soil behavior. In this approach equations of motion and
equilibrium are solved in discrete time increments in
time domain. A constitutive model is used to represent the cyclic behavior of soil.
The earliest constitutive relations use a simple
model relating shear stress to shear strain, whereby
the backbone curve is represented by a hyperbolic
function. Lee and Finn (1978) developed onedimensional seismic response analysis program using the hyperbolic model. Matasovic and Vucetic
(1995) further extended the model with modification
of the hyperbolic equation. Plasticity models have
also been used to represent cyclic soil behavior.
Borja et al. (1999) used a bounding surface plasticity
model to represent cyclic soil response at Lotung
Site in Taiwan.
Hashash and Park (2000) introduce an extension
of the model by Matasovic to account for the influence of confining pressure on soil modulus and
damping properties. Hashash and Park (2001) show
that confining pressure has an important impact in
computed ground motion in a site response analysis.
They show that significant portions of high frequency components of ground motion are propagated through deep soil deposits and that Propagation of seismic waves through very deep deposits
result in the development of long period ground motion. Spectral amplitudes of propagated ground motions are higher than what would be obtained using
conventional wave propagation analyses.
One dimensional site response analysis provides
data useful in the analysis of racking deformations in
an underground structures and is usually sufficient
for analysis of short structures such as subway station. Three dimensional wave propagation analysis
is required to develop ground deformations along the
length of a long tunnel to properly account for variability in ground conditions and it influence on
ground motion incoherency, phase shift and arrival
delay times. These type of analyses are not commonly performed due to their relative complexity.

Tunnel during
wave motion
Compression

Wave front

Tension

Tunnel cross section

before wave motion

Tunnel

Compression-extension

Compression of tunnel section

Positive
curvature

Tunnel

Top

(Comp.)

(Tension)

(Comp.)

(Tension)

(Comp.)

(Tension)

Tunnel

Negative
curvature

Bottom

Longitudinal bending
B

Diagonally-propagating wave

Tunnel Before
Wave Motion

diff

Tunnel During
Wave Motion

Tunnel During
Wave Motion

Tunnel Before
Wave Motion
Free Field Soil
Deformations

Shear Wave Front

Shear Wave Front

Ovaling of tunnel section

Racking of tunnel section

Figure 1. Deformation modes of tunnels due to seismic shaking

Soil Element,
Springs

a) Frame analysis with equivalent soil spring

Free Field
Racking

b) Racking analysis for Alameda Tubes, 12 m diameter

Figure 2. Pseudo static analysis approach

4 DESIGN AND ANALYSIS METHODS

Extensive reviews of seismic design procedures of
underground structures are presented in St. John &
Zahra (1987) and Wang (1993). The study by Wang
(1993) is an unpublished monograph by Parsons
Brinckerhoff used by many designers. A report under preparation by Hashash et. al. (2000) presents an
extensive review of seismic design approaches for
underground structures.
4.1 Free field deformation method
Early solutions developed to compute tunnel deformations due to seismic wave propagation assume
that the tunnel conforms to free field ground deformations. This may lead to an over estimate of structure deformations especially in soft soils where the
tunnel maybe stiffer than the surrounding soil.
4.2 Pseudo-static soil-structure interaction method
A psudo-static analysis method is commonly used to
account for soil-structure interaction effects. In a
simplified procedure described by St. John and Zahrah (1987) the tunnel is simplified as an elastic beam
on an elastic foundation representing the surrounding soil. Closed form solutions have been developed
to account for lining stiffness effect on ovaling deformations of circular tunnels. These solutions use
the concept of flexibility ratio that relates elastic soil
stiffness to elastic tunnel lining stiffness (Einstein
and Schwartz, 1979, and Peck et. al, 1972). Wu and
Penzin (1997) present closed-form solutions for circular tunnels. For rectangular sections the structural
racking deformations are computed by applying free
field racking deformations in a frame analysis. Soil
is represented by elastic springs. Deformations can
be applied as boundary displacements in a finite
element or finite difference analysis as shown in
Figure 2.
Figure 3 illustrates the use of the pseudo static
approach in the longitudinal analysis of a tunnel. In
this analysis the forcing function is the free field
displacement time history. Tunnel soil interaction is
represented as the interaction between longitudinal
beam elements and soil spring.
The pseudo static approach may be valid for
weak ground motion where deformations are elastic.
Ground response is highly non-linear for strong
ground motion and motions with fling effect.
4.3 Dynamic soil-structure interaction method
There is an increasing use of dynamic soil structure
analyses for underground structues whereby the soils
as well as the underground structures are represented
using finite elements. Wang (1993) used the program FLUSH to analyze rectangular tunnel sections.

An unpublished analytical study was performed for

the stations for the future East Side extension of the
LA Metro Red Line using time domain non-linear
seismic soil structure interaction analysis. Figure 4
shows a typical mesh used in the analysis. The
analyses show that structural racking displacements
exceeded the free field racking displacements
(Figure 4). That result was surprising and contrary to
conventional design approaches. The analyses imply
that the ground displacements were amplified due to
the presence of the station.
4.3.1 Wave scattering solutions
A class of solutions has been developed to solve the
problem of scattering and diffraction of elastic plane
waves by an arbitrarily shaped opening below
ground surface in an elastic half space. Manoogian
(1998) analyzes deep and shallow tunnels of circular, rectangular and elliptical cross sections using
this approach. Analyses show that for shallow tunnels, and for higher wave frequency components,
surface displacement amplitudes are amplified by a
factor of up to 3.5 within a distance of 3 tunnel radii
from tunnel centerline. The study by Manoogian has
important ramifications, it implies that placement of
tunnels amplifies anticipated ground motion, and
adversely affects adjacent structures.
5 OBSERVED SEISMIC RESPONSE OF
UNDERGROUND STRUCTURES
Dowding and Rozen (1978), Owen and Scholl
(1981) and Sharma and Judd (1991) present summaries of case histories of damage to underground facilities. Damage to tunnels is greatly reduced with
increased overburden, but damage is greater in soils
than in competent rock. More recently several large
earthquakes resulted in damage to modern underground structures in major urban centers.
5.1 Underground structures in the US
Bay Area Rapid Transit System, San Francisco, CA:
The BART system was one of the first underground
structures to be designed with seismic considerations
(Kuesel, 1969). On the San Francisco side, the system consists of below ground stations and tunnels in
fill and soft Bay Mud deposits, and is connected to
Oakland via the transbay tube. During the 1989
Loma Prieta Earthquake flexible joints connecting
the transbay tube to the San Francisco and Oakland
vent buildings displaced few inches (PB 1991). Peak
ground accelerations (PGA) experienced did not exceed 0.3 g. The BART system is now slated for
seismic retrofit design for anticipated earthquake
events with PGA~0.8g and considerable near field
effects such as high velocity pulses.

A la m e d a
P o rta l

V e rtic a l
O a kla n d
P o rta l

T ra n sv e rse
S o il S p rin g

V e r tic a l

Tube

20

L o n g itu d in a l

0
-2 0
-4 0
-6 0
-8 0
0

T r a n s v erse

10

15

20

25

30

35

40

D isp la c e m e n t T im e
H isto ry In p u t

T u b e S e c tio n

Figure 3. Simplified three-dimensional model for analysis of the global response of an immersed tube tunnel (Hashash et al. 1998)

a) Station dimensions 60 ft wide, 50 ft high

b) Free field racking (max 3.9 in.)

c) Structural racking (max 6 in.)

Figure 4. Dynamic analysis of little Tokyo subway station, LA Metro

Alameda Tubes, Oakland-Alameda, CA: The

Alameda tubes connect Alameda Island to Oakland
through a pair of immersed tube tunnels built in
1930s-1950s. The tubes, and vent buildings experienced some structural cracking that caused an
increased water leakage into the tunnels during the
Loma Prieta Earthquake.
L.A. Metro, Los Angeles, CA: The Los Angeles
Metro is being constructed in several phases, some
of which were complete and operational during the
1994 Northridge Earthquake. PGA near the affected
sections was around 0.3 g. Concrete lining of bored
tunnels experienced some cracking.
5.2 Underground structures in Kobe, Japan, 1995
Hyogoken-Nambu Earthquake
The 1995 Hyogoken-Nambu earthquake caused a
severe damage to Rokko, Bantaki and NaruoMikage Tunnels and resulted in a major collapse of
Daikai subway station in Kobe, Japan (Nakamura,
et. al, 1996). The collapse experienced by center
columns of the station was accompanied by collapse
of the ceiling slab and settlement of soil cover by
more than 2.5 m. The station design in 1962 did not
include specific seismic provisions.
5.3 Other tunnels
Tunnel performance in Taiwan, September 21, 1999
Chi-Chi earthquake: Most tunnels in Central Taiwan
survived the severe shaking intact. Limited spalling

and cracking was observed. The main damage was

limited to slope failure at tunnel portals which resulted in blockage of tunnels.
Tunnel performance in Turkey during the 1999
Koceli earthquake: A 100-m section of twin shotcrete supported tunnels collapsed during the earthquake. The tunnels were under construction. The
cause of the collapse is currently being investigated
(Mair, 2000).
6 CURRENT CHALLENGES AND AREAS OF
FUTURE RESEARCH
Underground structures in the US have experienced
limited damage during Loma Prieta and Northridge
earthquakes, but shaking levels were much lower
than maximum anticipated events. More recent large
earthquakes in Kobe, Japan and Koceli, Turkey have
resulted in severe damage to some tunnel structures.
This may point to limitations in our understanding of
seismic behavior of underground structures. There is
a need to enhance our knowledge to design earthquake resistant underground structures for high levels of shaking.
As we look into the future, development plans in
urban areas in the USA, China as well as other countries include the construction of major underground
subway/metro systems. Seismic consideration will
have to be included in the design of such systems.
There are still many issues that are not well understood as they relate to the behavior of underground

structures. Research in the future will have to answer

some important questions in our quest to reduce
earthquake hazard. The following are some proposed
areas that will contribute to reduction of uncertainty
and increase in our knowledge:
1. Instrumentation of tunnels and underground
structures to measure their response during ground
shaking: These instruments would include measurement of vertical and lateral deformations along
the length of the tunnel. This will be useful to understand the effect of spatial incoherencey and directivity of the ground motion on tunnel response as well
as structural deformations.
2. Study of the influence of high vertical accelerations on the generation of large compressive loads
in tunnel linings and subway station columns: Large
vertical forces may have been a factor in the collapse
of the Daikai Subway station as well as other tunnel
structures in Japan and Turkey.
3. Development of improved numerical models to
simulate the dynamic soil structure interaction problem of tunnels, as well as portal and subway structures: These models will be useful in studying the
effect of high velocity pulses (fling effect) generated
near fault sources on underground structures.
4. Ground Motion Incoherence: Evaluation of the
significance of ground motion incoherence on the
development of differential movement along the
length of a tunnel (Power et al., 1996). Ground motion incoherence is particularly important in soft
soils and shallow tunnels where the potential for
slippage between the tunnel and soil is high.
5. Effect of the underground structure on the local ground motion: Evaluation of the influence of
underground structures on local amplification or attenuation of propagated ground motion. Change in
ground motion due to shielding and amplification effects.
6. Tunnel Components: Research into the application of non-conventional lining, bolting, and water
insulation materials that can be used for seismic
joints to enhance seismic performance of a tunnel.
7. Site Response: Improved resolution of local
site response will be important for estimating ground
deformation around an underground structure.
REFERENCES
Borja, R.I., Chao, H-Y, Montans, F.J., & Lin, C-H. 1999.
Nonlinear ground response at Lotung LSST site. J. Geotechnical and Geoenvironmental Engineering, Vol. 125,
No. 3, 187-197.
Dowding, C. H. & Rozen, A. 1978. Damage to rock tunnels
from earthquake shaking. Journal of the Geotechnical Engineering Division, ASCE, Vol. 104, GT2, 175-191.
Dominic, A., Kausel, E. & Whittle, A. 2000. A model for dynamic shear modulus and damping for granular soils ,
Journal of Geotechnical and Geoenvironmental Engineering, ASCE, Nov. 2000.
Einstein, H. H. & Schwartz, C. W. 1979. Simplified analysis

for tunnel supports. Journal of the Geotechnical Engineering Division, ASCE, Vol. 105, GT4.
Hashash, Y. M. A., Tseng, W. S., & Krimotat, A. 1998. Seismic soil-structure interaction analysis for immersed tube
tunnels retrofit. Geotechnical Earthquake Engineering and
Soil Mechanics III, ASCE Geotechnical Special Publication
No. 75, Vol. 2, 1380-1391.
Hashash, Y.M.A. & Park, D., 2001. Non-linear onedimensional seismic ground motion propagation in the Mississippi Embayment, to be published, Engineering Geology.
Kuesel, T. R. 1969. Earthquake Design Criteria for Subways.
Journal of the Structural Division, ASCE, ST6, 1213-1231.
Lee, M.K.W. & Finn, W.D.L. 1978. DESRA-2, Dynamic
effective stress response analysis of soil deposits with
energy transmitting boundary including assessment of
liquefaction potential. Soil Mecahnics Series No. 36,
Department of Civil Engineering, University of British
Columbia, Vancouver, Canada.
Lee, V.W. & Trifunac, M.D. 1979. Response of tunnels to incident SH-waves, Engineering Mechanics Division, ASCE,
Vol. 105, 643-659.
Lysmer, J., Udaka, T., Tsai, C. F., & Seed, H. B. 1975.
FLUSH: A computer program for approximate 3-D analysis
of soil-structure interaction problems. Report No. EERC
75-30, Earthquake Engineering Research Center.
Mair, R. Personal communications, 2000.
Manoogian, M.E. 1998. Surface motion above an arbitrarily
shaped tunnel due to elastic SH waves, ASCE, Geotechnical Special Publication No. 75, Geotechnical Earthquake
Engineering and Soil Dynamics III, 754-765, Seattle.
Matasovic, N. & Vucetic, M. 1995. Seismic response of soil
deposits composed of fully-saturated clay and sand layers.
Proc. 1st International Conference on Earthquake Geotechnical Engineering, JGS, Vol. 1, 611-616, Tokyo, Japan.
Okamoto, S., Tamura, C., Kato, K., & Hamada, M. 1973. Behaviors of submerged tunnels during earthquakes. Proceedings of the Fifth World Conference on Earthquake Engineering, Rome, Vol. 1, 544-553.
Owen, G. N. & Scholl, R. E. 1981. Earthquake engineering of
large underground structures. Report No. FHWA/RD80/195, Federal Highway Administration and National Science Foundation.
PB, 1991. Trans-bay tube seismic joints post earthquake
evaluation, Bay Area Rapid Transit District, report prepared
by Parsons Brinckerhoff Quade & Douglas, Nov.
Peck, R. B., Hendron, A. J., & Mohraz, B. 1972. State of the
art in soft ground tunneling. Proceedings of the Rapid Excavation and Tunneling Conference, American Institute of
Mining, Metallurgical, and Petroleum Engineers, New
York, 259-286.
Power, M. S., Rosidi, D., & Kaneshiro, J. 1996. Volume III
Strawman: Screening, evaluation, and retrofit design of
tunnels. Report Draft, National Center for Earthquake Engineering Research, Buffalo, New York.
Sharma, S. & Judd, W.R., 1991. Underground Opening damage
from earthquakes, Engineering Geology, 30.
Schnabel, P. B., Lysmer, J., and Seed, B. H. 1972. SHAKE A
computer program for earthquake response analysis of horizontally layered sites. Report No. EERC 72-12, University
of California at Berkeley, California.
St. John, C. M. & Zahrah, T. F. 1987. Aseismic design of underground structures. Tunneling and Underground Space
Technology, Vol. 2, No. 2, 165-197.
Sugito, M., 1995. Frequency-dependent equi-linearized technique for seismic response analysis of multi-layered
ground. Doboku Gakkai Rombun-Hokokusho/Proceedings
of the Japan Society of Civil Engineers, No. 493, 3-2, 49-58
Wang, J.-N. 1993. Seismic Design of Tunnels: A State-of-theArt Approach. Monograph 7, Parsons Brinckerhoff Quade
& Douglas, Inc., New York.

Wu, C. L. & Penzien, J. 1997. Stress analysis and design of

tunnel linings. Proceedings of the 1997 Rapid Excavation
Tunneling Conference, 431-455.