Seismic Investigation of Steel Pile Bents:

II. Retrot and Vulnerability Analysis

AymanA. Shama,
a)
M.EERI, John B. Mander,
b)
M.EERI, and Stuart S. Chen
c)
This paper is the second of a two-part study on the seismic vulnerability
of deck bridges supported on steel pile bents. A conceptual elastic cap/elasto-
plastic steel pile retrofit strategy is proposed in this part with the aim of
strengthening the connection and ensuring plastification takes place only in
the steel pile. An experimental program was carried out to assess the retrofit
strategy. On the basis of the experimental results for existing as well as ret-
rofitted connections, a seismic vulnerability analysis for bridges supported by
steel pile bents was performed. Fragility curves for such structures were de-
veloped using a simplified fundamental mechanics-based approach. The
study showed that the retrofitted connections exhibited superior energy ab-
sorptions with respect to the existing connections. Fragility curves also dem-
onstrated the effectiveness of the retrofit strategy proposed.
[DOI: 10.1193/1.1468250]
INTRODUCTION
Substructures consisting of steel pile bents have been widely used in the construction
of highway bridges throughout the United States. But the majority of these bridges were
built two to three decades ago, when design of structures for current high seismic loads
was not required. The experimental study performed on specimens simulating as-built
substructures indicated that the pile-to-pile cap connection, a connection primarily de-
signed for vertical loading, is susceptible to damage from cyclic lateral loading (Shama
et al. 2002). Therefore, in zones of moderate to high seismicity it is required to perform
any properly designed seismic strengthening for this class of connection and hence im-
proving the performance of bridges supported by such substructures during a potential
earthquake.
Two basic approaches are available for the retrofit of these substructures. The first
approach is to reduce the seismic forces that can be developed in the cap beam during
seismic events. This approach can be achieved by connecting the bent piles with a link
beam as shown in Figure 1. The location of the link beam can be adapted so that the
final moment at the connection will be less than both the nominal moment capacity of
the concrete beam and the plastic moment capacity of the steel H-pile. The merit of this
method is that it can be accomplished without traffic disruption. The approach has been
a)
Parsons Transportation Group Inc., 110 William Street, New York, NY 10038
b)
Department of Civil Engineering, University of Canterbury, New Zealand
c)
Department of Civil, Structural and Environmental Engineering, State University of NewYork at Buffalo, Buf-
falo, NY 14260
143
Earthquake Spectra, Volume 18, No. 1, pages 143160, February 2002; 2002, Earthquake Engineering Research Institute
used in California in the retrofit of Santa Monica Viaducts (Priestley et al. 1997). This
method will not be explored further here and is out the scope of the present study.
The second approach is to increase the cap beam strength to the level required for
shifting the plastic hinge location to the steel pile rather than the concrete cap. Hence,
ensure better ductile connection that can possess large deformation capability and permit
much more dissipation of seismic energy.
A conceptual elastic cap/elasto-plastic pile retrofit strategy, consistent with the sec-
ond approach, is adopted in the present research. This paper first sets forth the seismic
retrofit strategy and then goes on to present and compare the test results. The study is
concluded with seismic vulnerability analysis for highway bridges supported by such
substructures.
RETROFIT METHODOLOGY
A simplified stress distribution shown in Figure 2 is adopted to evaluate the addi-
tional retrofit depth required. It is assumed in this mechanism that the stress block force
couple C
m
will resist the external applied moment to the connection. Therefore, the
stress block force C
m
can be evaluated as
C
m
0.5f
c
b
f
l
emb
(1)
in which, and are stress block factors; f
c
compressive strength of concrete; b
f
width of the steel pile section; and l
emb
embedment of steel pile into the concrete cap
beam. The lever arm can be determined as
jdl
emb
10.5 (2)
Figure 1. Retrofit strategies for steel pile bents: (a) link beam, and (b) cap beam strengthening.
144 A. A. SHAMA, J. B. MANDER, AND S. S. CHEN
By assuming 1 (Shama et al. 2002), then jd can be taken as 0.5 l
emb
. The moment
strength of the connection can now be determined as
M
j
C
m
jd (3)
where strength reduction factor. Substitute Equations 1 and 2 in Equation 3, there-
fore
M
j
0.25f
c
b
f
l
emb
2
(4)
if is taken as 0.36, i.e., moderate damage, and 0.90, further simplification for
Equation 4 leads to
M
j
0.08f
c
b
f
l
emb
2
(5)
To ensure that plastic hinges will not occur at cap beam, the following condition
should be satisfied:
M
j
M
po
(6)
where M
po
is the overstrength plastic moment capacity of the steel H-pile section. There-
fore
0.08f
c
b
f
l
emb
2
M
po
(7)
and hence
l
emb

12.5M
po
f
c
b
f
(8)
Further simplifications of Equation 8 can lead to the following expression for the
total embedment depth:
Figure 2. Mechanism adopted for connection retrofit.
SEISMIC INVESTIGATION OF STEEL PILE BENTS: RETROFIT AND VULNERABILITY ANALYSIS 145
l
emb
d
p
3.5

f
su
f
c

t
f
d
p

(9)
where f
su
the ultimate stress of the steelpile section; t
f
flange thickness of the steel
pile section; d
p
depth of the steel pile section. The additional embedment depth re-
quired is
l
ad
l
emb
l
ab
(10)
where l
ab
the embedment depth for the as-built structure.
Additional design considerations are required to account for the effects of reversed
cyclic loading. Under cyclic loading a gap is expected to occur at the pile flange-
concrete interface. If the opening of that gap is not controlled, it may affect the hysteretic
response of the connection, and hence its energy absorption. It is, therefore, necessary to
provide sufficient horizontal reinforcing bars crossing the pile-concrete interface. This
steel is chosen such that its area has a yield force equal to the plastic shear capacity of
the pile section. Thus, the area of steel required is
A
s

V
p
f
yh
(11)
where V
p
the shear capacity of the pile section, f
yh
the yield stress of reinforcing re-
bars. Horizontal stirrups are provided to counteract shear forces. Shear forces are as-
sumed to provide a strut and tie actions. Concrete struts are assumed to act at 45 to
deliver the required shear force to a number of stirrups n. The force in each hoop can be
determined as
V
st

0.50V
p
n
(12)
Each stirrup is assumed to carry a force of A
v
f
yh
, where A
v
is the area of one stirrup,
therefore:
A
v

V
st
f
yh
(13)
CONSTRUCTION OF RETROFIT
The embedment depth of the retrofitted specimen was determined using Equation 9.
For a 35 MPa concrete and 315 MPa steel and using the dimensions of the HP 1042
steel pile section, the total embedment depth was determined as 625 mm. The original
specimen had an embedment depth300 mm. Therefore it was decided to use another
300 mm for the overlay depth.
The longitudinal reinforcement required to close the anticipated gap between the
steel section and the concrete cap beam during cyclic loading is determined according to
Equation 11:
146 A. A. SHAMA, J. B. MANDER, AND S. S. CHEN
A
s

V
P
f
yh

0.6F
y
d
P
t
w
f
yh

0.6315.224610.54
0.85413
1397 mm
2
.
Therefore 4 25-mm-diameter rebars (4 #8, A
s
1963 mm
2
) were used.
The size and number of stirrups that resist the shear force was determined by assum-
ing four stirrups will resist the shear force. Therefore, according to Equation 12
V
st

0.50V
p
n

0.500.6315.224610.54
41000
61.02 kN
Substitute in Equation 13 A
v
V
st
/f
yh
61.0210
3
/413148 mm
2
; therefore 12-mm-
diameter double-leg stirrups (A
s
226 mm
2
) were used.
To ensure joint confinement, additional transverse reinforcement (9.5-mm-diameter
17 galvanized wire rope) was added to the HP 1042 steel pile along the overlay
length, with a spiral pitch of 50 mm. To facilitate the top-down pouring of concrete in an
actual bridge, it was decided to increase the width of the pile cap by 100 mm on each
side. The additional width is required for practical purposes where the pile cap will be in
an inverted position in an actual scenario, and that extension of the width will facilitate
the concrete placing and compaction in the field. Additional reinforcement was provided
in the form of #4 rebars (13-mm diameter) every 150 mm for the longitudinal direction.
Diagonal 13-mm-diameter stirrups were used to improve joint shear resistance. The rest
of the stirrups were three-sided (U-shaped) #4 rebars. The retrofit construction was per-
formed without drilling big holes in the cap beam. Instead, seven 16-mm diameter (7 #5)
bars were provided as spacers along each of the two longitudinal sides of the cap beam
and attached to it by galvanized straps. The straps were fixed to the cap beam by
6 mm25 mm9.5 mm sleeve concrete anchors. Figure 3 illustrates the reinforcing de-
tails for the pile bent retrofit at the connections.
SCOPE OF THE EXPERIMENTAL PROGRAM
The specimens were tested in displacement control under incremental cyclic loading.
Two cycles of drift at levels ranged from 0.5% to 6%. The quasi-static displacement
function was sinusoidal with a one-minute period per cycle. Table 1 summarizes the
axial loads and drift information for each specimen tested in the retrofit study.
SPECIMEN ReS1
This specimen was tested without any axial load with two reversed cycles at drift
amplitude of 0.5%, 1%, 2%, 3%, 4%, and concluded with 26 cycles at 5%.
The complete hysteretic response of the specimen is shown in Figure 4. Yielding of the
pile section first occurred just prior to the 2% drift level, characterized by some diagonal
striations on the whitewashed flanges. Local buckling was initiated at the second cycle
of 3% drift and continued to grow as the drift amplitude increased.
Strain hardening was observed on the first cycle at 3% drift was attained. Beyond the
4% drift level, strength degradation of the steel material occurred as a result of the local
buckling. The specimen was exposed to a continued cycles of constant amplitude testing
SEISMIC INVESTIGATION OF STEEL PILE BENTS: RETROFIT AND VULNERABILITY ANALYSIS 147
at 5% drift, through which strength degradation of the specimen continued. Finally on
the 26th cycle, a visible horizontal fatigue crack was observed at the flange subjected to
compression when the lateral actuator is pushing.
Figure 3. Reinforcement details for retrofitted connections.
Table 1. Test program for retrofitted specimens
Spec.
ID
Gravity
Load
(kN)
Lateral
Actuator
Angle
Vertical Axial Load
Control (kN)
Total Axial
Load (kN)
Max.
Drift
No. of
Cycles at
Max. Drift
ReS1 0 0 0 0 5% 29
ReS2 150 0 150 150 6% 15
ReS3 120 24 1201.43 P
da
1202.01 V 5% 3
VP
da
cos P
da
force in diagonal actuator
148 A. A. SHAMA, J. B. MANDER, AND S. S. CHEN
Based on the crack initiation, and crack growth mechanism, it is evident that the fail-
ure of that specimen was due to low cycle fatigue. A photograph portraying such failure
is shown in Figure 5.
SPECIMEN ReS2
Specimen ReS2 was tested under a constant axial load of 150 kN with two reversed
cycles at drift amplitude of 0.5%, 1%, 2%, 3%, 4%, 5% and concluded
with 15 cycles at 6%. Figure 6 presents the force-deformation results.
Figure 4. Specimen ReS1: lateral load displacement.
Figure 5. Connection ReS1 after test.
SEISMIC INVESTIGATION OF STEEL PILE BENTS: RETROFIT AND VULNERABILITY ANALYSIS 149
This specimen exhibited inelastic behavior prior to 2% drift and experienced work
hardening in the pile steel material through both the 3% drift and the 4% drift cycles.
Local buckling of the steel pile flanges started to occur, along a 250-mm distance above
the cap beam concrete surface, during the second cycle at 4% drift. Strength degradation
of the steel pile followed and was observable at the beginning of the 5% drift.
It was decided to perform a constant cyclic high-amplitude test phase at the 6% drift
level up to the fatigue failure of the specimen. A horizontal crack occurred at the com-
pression flange after the 12th cycle was completed. The flaw continued to grow, propa-
gating vertically both sides of the flange as well as horizontally in the web and fracture
of the specimen was visible at the end of the 15th cycle. At that point the test was
stopped.
SPECIMEN ReS3
Specimen ReS3 was tested under variable axial load with two reversed cycles at each
drift amplitude of 0.5%, 1%, 2%, 3%, and 4% concluding with 3 cycles at
5%. This specimen also satisfied the main objective of the conceptual elastic cap/
elastic-plastic pile retrofit strategy proposed in this study. The specimen behaved in an
elastic manner prior to 2% drift. Yielding of the flanges in an area 250 mm above the
added concrete surface was noticed through both the 2% and the 3% drifts and diagonal
yield lines were visible on the whitewashed flanges. Strain hardening of the steel mate-
rial also occurred during both the 2% and 3% drifts. Local buckling was initiated during
the second cycle at 3% drift. This local buckling became more pronounced during the
4% drift and was characterized by observable strength degradation in force displacement
loops (see Figure 7). Due to shortcomings of the test setup, the test of pile specimen
ReS3 was not continued to reach the failure of the steel.
Figure 6. Specimen ReS2: lateral load displacement.
150 A. A. SHAMA, J. B. MANDER, AND S. S. CHEN
EVALUATION OF DAMPING
Effective damping (
eff
) in a structure can be viewed as a combination of equivalent
viscous damping
eq
, and viscous damping inherent in the structure
0
assumed 0.05 for
steel pile bents. The most common method for defining the equivalent damping
eq
is to
equate the energy dissipated in a response cycle of the actual structure to that of an
equivalent viscous system as (Chopra 1995):

eq

1
4
E
D
E
so
(14)
where E
D
energy dissipated by damping, and E
so
maximum strain energy. For a bi-
linear representation of cyclic loops E
D
is evaluated as
E
D
4
y

max
K
0
K
ef f
(15)
in which,
y
yield displacement,
max
maximum displacement, K
0
initial stiffness,
and K
ef f
the secant stiffness which can be written in terms of ductility ratio
(
max
/
y
) as
K
ef f
K
0

(16)
where post-yield stiffness ratio. The maximum strain energy can be quantified as
E
so
K
ef f

max
2
2
(17)
Assuming an overall bilinear response, as shown in Figure 8, the effective damping
due to hysteresis can be determined by substituting Equations 15 through 17 in Equation
14 and rearranging, the theoretical equivalent viscous damping can be quantified as
Figure 7. Specimen ReS3: lateral load displacement.
SEISMIC INVESTIGATION OF STEEL PILE BENTS: RETROFIT AND VULNERABILITY ANALYSIS 151

eq

1
1

1
(18)
where energy absorption efficiency factor defined as

E
cycle
E
EPP
(19)
where E
EPP
the energy absorbed by a 100% perfect elasto-plastic system, defined ac-
cording to the following relationship:
E
EPP
F
n

F
n

x
p

x
p

(20)
where F
n

the nominal capacity of the system in the push direction; F

n

the corre-
sponding value in the pull direction; x
p

the plastic component of the displacement in

the push direction; and x
p

the corresponding value in the pull direction.

In the present study elastic-perfect plastic behavior will be assumed and Equation 18
is simplified to

eq

1
1

(21)
The hysteretic energy absorbed by the system per cycle is given by
E
cycle

i1
n

F
i
F
i1
2

x
i
x
i1
(22)
where F
i
force in i-th step; and x
i
displacement of the same step.
The experimental equivalent viscous damping is defined according to the 1994 Uni-
form Building Code as

eq

1
2
E
cycle
F
max

max
(23)
Figure 8. Bilinear representation of cyclic loops.
152 A. A. SHAMA, J. B. MANDER, AND S. S. CHEN
where F
max
average of the maximum strength in the forward and reverse loading di-
rections, and
max
can be evaluated as the average of the maximum displacement in both
loading directions.
Figure 9 presents experimental results where the equivalent viscous damping is plot-
ted against the ductility amplitude for both as-built and retrofitted pile bents. The the-
oretical relationships according to Equation 21 are also plotted in the figures. Based on
the experimental results values of 0.47 and 0.75 are suggested for both existing
and retrofitted structures.
SEISMICVULNERABILITY OF BRIDGES SUPPORTED ON PILE BENTS
In light of the foregoing experimental findings for existing (Shama et al. 2002) as
well as retrofitted structures, the seismic vulnerability of deck bridges supported by steel
pile bents was investigated. NIBS (1997) defines a total of five damage states for high-
way system components. These are (1) none, (2) slight/minor, (3) moderate, (4) exten-
sive, and (5) complete. The quasi-static reversed cyclic loading experiments presented in
this study captured these damage states, summarized in Table 2.
EXPECTED SEISMIC STRUCTURAL RESISTANCE
The seismic demand of a bridge structure can be represented in terms of a design
code-like response spectrum; that is the seismic demand can be quantified by the lesser
of
C
d

2.5A
B
S
(24)
and
C
d

SA
T
ef f
B
L
(25)
where C
d
base shear demand; Apeak ground acceleration at the site; Ssoil type fac-
Figure 9. Evaluation of energy absorption efficiency factor for pile bents before and after ret-
rofit.
SEISMIC INVESTIGATION OF STEEL PILE BENTS: RETROFIT AND VULNERABILITY ANALYSIS 153
tor; T
ef f
effective period of vibration and B
L
, B
S
are spectral reduction factors used to
modify the elastic response spectrum for high damping, i.e., when the structure responds
in the inelastic range to account for hysteretic damping resulting from nonlinear effects.
Based on regression analysis on the values given by Newmark and Hall (1992), Cheng
and Mander (1997) suggested the following relationships for the spectral reduction fac-
tors as
B
s

ef f
0.05

0.5
and B
L

ef f
0.05

0.3
(26)
It is assumed here that the peak response of the nonlinear structure to be equal to the
displacement of a substitute SDOF system with an effective period (T
eff
) given by
T
ef f
2
M
K
2
W/g
F
y
/
2

C
c
g
(27)
where maximum displacement response; C
c
F
y
/Wbase shear capacity, in which
F
y
yield force of the pile bent; and Wtributary weight. Same as the capacity spec-
trum method, it is assumed that the intersection of the capacity and appropriately
damped demand curve at a point represents the inelastic displacement of the structure.
Table 2. Damage states for bridges supported by steel pile bents
Substructure Status Bending Axis Damage State* Drift Limit
Nonretrofitted Strong Axis Pre-Yield 0.01
Slight Damage
a1
0.02
Moderate Damage
a2
0.03
Extensive Damage
a3
0.04
Complete Damage
a4
0.05
Weak Axis Pre-Yield 0.01
Slight Damage
a1
0.02
Moderate Damage
a5
0.04
Extensive Damage
a6
0.06
Complete Damage
a4
0.07
Retrofitted Strong Axis Pre-Yield 0.01
Slight Damage
a1
0.02
Moderate Damage
a5
0.04
Extensive Damage
a6
0.06
a1Inelastic action on the steel pile section occur
a2Shear cracks occurs in the concrete cap beam
a3Damage is mainly concentrated in the cap beam with concrete spalling
a4Cap beam failure and slipping of the steel pile out of the socket
a5First occurrence of local buckling in the flanges of the steel pile section
a6Local buckling more pronounced in the steel pile flanges
*Damage states are based on HP 1042 steel piles. Pile cap width and depth are 700 mm and 600 mm, respec-
tively.
154 A. A. SHAMA, J. B. MANDER, AND S. S. CHEN
Therefore, replacing the base shear demand in Equations 24 and 25 with the structural
base shear capacity and rearranging, the peak ground acceleration can be determined by
taking the greater of
A0.4C
c
B
s
(28)
or
A
2
S

C
C

g
B
L
(29)
For deck bridges supported by flexible steel pile bents, Equation 29 invariably governs.
Analysis of Bridge Capacity
It will be assumed here that the piles within a cap are either oriented to deform along
its strong or weak axis. In other words, no combination of the two cases is permissible
within the cap. Consider the pile bent shown in Figure 10 subjected to a lateral force.
Assuming an admissible bent plastic mechanism shown in the same figure, the external
and internal virtual work can be equated as follows:
EWDIWD (30)
C
c
W
p
HnM
p

p
(31)
where
p
plastic rotation; Hdistance between the plastic hinges, i.e., distance from
the bottom surface of the cap beam to the second plastic hinge within the soil; n
number of piles in the bent; a fixity factor denoting 1 for fixed-pinned behav-
ior and (1) for fixed-fixed, where pile-to-cap connection efficiency defined as

M
j
M
P
(32)
Figure 10. Potential plastic mechanisms for pile bents.
SEISMIC INVESTIGATION OF STEEL PILE BENTS: RETROFIT AND VULNERABILITY ANALYSIS 155
where M
j
moment capacity of the concrete-pile connection (joint); and M
p
f
y
Z
p
nominal moment capacity of the pile.
From Equation 31 it is possible to solve for the base shear capacity coefficient as
C
c

1nM
p
WH
(33)
By assuming that the tributary load W is shared equally between the n piles such that
each of them carries an axial load P, the total tributary load can be expressed as
WnP (34)
Substituting Equation 34 into 33 one obtains
C
c

1M
p
PH
(35)
For small axial load, the plastic moment capacity of a steel H-pile for strong axis bend-
ing can be approximated as
M
p
f
y
t
f
b
f
d
p
(36)
where f
y
yield stress of the steel pile section; t
f
flange thickness of the steel pile; b
f
flange width of the steel pile. Similarly for weak axis bending:
M
p

f
y
t
f
b
f
2
2
(37)
The applied axial load on each pile can be expressed in terms of the axial yield load:
PP
y
2 f
y
b
f
t
f
(38)
where ratio of the axial applied load to the axial yield load. Also in terms of tributary
weight:
P
wBL
n
(39)
where wthe unit weight of the bridge deck, Bthe deck width, and Lthe deck span
length. Thus

P
P
y

wBL
2nf
y
b
f
t
f
(40)
By substituting Equations 36 and 38 into Equation 35, one can obtain an expression for
the base shear capacity coefficient of the strong axis bending pile:
C
c

1d
p
2H
(41)
156 A. A. SHAMA, J. B. MANDER, AND S. S. CHEN
By substituting Equation 41 into Equation 29 one can obtain an expression for the ex-
pected peak ground acceleration:
A
2
S

1d
p

i
2g
B
L
(42)
in which
i

i
/Hthe ith drift damage state.
Similarly for weak axis bending, by substituting Equations 37 and 38 into Equation
35, an expression for the base shear capacity of weak axis bending pile bents is obtained
as follows:
C
c

1b
f
4H
(43)
By substituting Equation 43 into Equation 29 one can obtain an expression for the ex-
pected peak ground acceleration for the motion along the weak axis bending of piles as
A
2
S

1b
f

i
4g
B
L
(44)
FRAGILITY CURVETHEORY
As shown by Equations 42 and 44, it is possible to deterministically predict, for a
given bridge, the level of ground motion necessary to attain a target level of a damage
state. However, there are lots of uncertainties associated with both the capacity and the
demand. The uncertainty in the demand arises from the fact that the code specified spec-
trum was established as a result of statistical analysis for a broad range of response spec-
tra of actual ground motions occurred in different locations having similar soil condi-
tions. At the site where the bridge is located, one can expect a series of seismic events
with associated response spectra that may follow, or not, the ensemble used to establish
the general code specified spectrum at this particular site. The uncertainty in capacity is
due to different sources, such as (1) variability of structural material properties, (2) in-
elastic energy absorption and damping, and (3) dispersion in calculating response due to
approximations in modeling. Therefore, there are several parameters, which invariably
have a measure of both randomness and uncertainty associated with them, entering the
problem. An increasingly popular way of characterizing the probabilistic nature of the
phenomena concerned is through the use of fragility curves.
By assuming that the structural capacity and seismic demand are random variables
that conform to either a normal or lognormal distribution, then following the central
limit theorem it can be shown that the composite performance outcome will be lognor-
mally distributed. Such a distribution only requires knowledge of a median value and a
suitable value of the logarithmic standard deviation (Kennedy et al. 1980) and can be
represented by the function
FS
a

lnS
a
/A

(45)
SEISMIC INVESTIGATION OF STEEL PILE BENTS: RETROFIT AND VULNERABILITY ANALYSIS 157
where (.)=standard normal cumulative distribution function; S
a
spectral acceleration
amplitude for a period of T1 s; Athe median (expected value) peak ground accelera-
tion for the given drift limit to be attained, as deterministically assessed by Equations 42
and 44; and
c
composite lognormal standard deviation which combines aspects of un-
certainty and randomness for both capacity and demand. Based on earlier studies on
variations in concrete and steel properties (Julian 1955; Shalon and Reintz 1955; Mirza
et al. 1979a, 1979b) and uncertainties arising from the paucity of data leading to sim-
plifications and assumptions (Jernigan 1996), Dutta and Mander (1999) quantified the
logarithmic standard deviation
p
corresponding to capacity as 0.20. In another study,
Pekcan (1998) used peak acceleration response of various SDOF systems subjected to a
wide range of ground motions on various types of soils in an attempt to identify the
randomness associated with the demand. His findings indicated that the random re-
sponse due to the disparities of the ground motion input resembles a lognormal distri-
bution
d
. He proposed a value of 0.55 for the long period ranges of the idealized spec-
tra given by Equation 25. Therefore the composite lognormal standard deviation
c
can
be obtained from the square root of the sum of squares (SRSS) of
p
and
d
as 0.6.
Baso

z and Mander (1999) validated this value against experiential fragility curves ob-
tained from data gathered from the 1994 Northridge and 1989 Loma Prieta earthquakes.
FRAGILITY CURVEAPPLICATIONTO STEEL PILE BRIDGES
The following assumptions have been made in the analysis:
The bridge is supported on steel pile bents with section HP 1042. Conse-
quently, d
p
and bf are taken as 246 mm and 256 mm, respectively.
The unit weight of the concrete decks is w7 kPa, with representative span
length and width of L8 m and B14 m.
The number of piles in a bent is taken as n5.
The median compressive strength of the deck concrete f
c
25 MPa.
The median yield stress of the steel pile bents f
y
320 MPa.
Based on the above assumptions, the ratio of the axial applied load to the axial yield
load was determined according to Equation 40 as 0.095.
Table 3 presents the different values for the parameters used in determining the ex-
pected peak ground acceleration of existing, as well as retrofitted structures, for differ-
ent damage states.
Fragility curves are displayed in Figure 11 for two damage states namely moderate
damage (DS3) and onset of irreparable (extensive) damage (DS4) for both the ex-
isting and retrofitted structures. It is shown that bridges retrofitted according to the ret-
rofit strategy proposed in this study are less prone to damage than existing structures. As
an example, at spectral acceleration of 0.6 g, 28% of the nonretrofitted bridges will ex-
hibit irreparable damage, whereas only 5% of the retrofitted bridges will experience
such damage.
158 A. A. SHAMA, J. B. MANDER, AND S. S. CHEN
CONCLUSIONS
A conceptual elastic cap/elasto-plastic steel pile retrofit strategy was proposed in this
study with the intent of strengthening the steel pile bents to cap beam connections that
behaved poorly when tested under cyclic lateral loading. The retrofitted connections
were tested under same loading conditions as of the as-built connections and showed
a superior performance in terms of energy absorption and ductility. In light of the ex-
perimental results for existing as well as retrofitted connections, a seismic vulnerability
analysis for bridges supported by steel pile bents was performed. Theoretical fragility
curves concluded this analysis provided some insight into the major factors that lead to
bridge damage and /or collapse. Fragility curves also demonstrated the effectiveness of
the retrofit strategy proposed in this study for pile-supported bridges.
ACKNOWLEDGMENTS
This research was carried out in the Department of Civil, Structural and Environ-
mental Engineering at the State University of New York at Buffalo. Financial support is
gratefully acknowledged from the Multidisciplinary Center for Earthquake Engineering
Research through contract with the Federal Highway Administration. The authors wish
to thank anonymous reviewers whose suggestions and comments led to great enhance-
ments in the clarification of the material presented.
Table 3. Expected peak ground motions for existing and retrofitted structures
DS
Connection
efficiency Drift
Spectral reduction
factors
B
L
Expected peak ground
acceleration
A
i
(g)
Pre-Ret. Post-Ret. Pre-Ret. Post-Ret. Pre-Ret. Post-Ret. Pre-Ret. Post-Ret.
1 1 1 0.01 0.01 1.00 1 0.32 0.32
2 0.80 1 0.02 0.02 1.50 1.69 0.66 0.77
3 0.40 1 0.03 0.04 1.60 1.88 0.75 1.20
4 0.30 1 0.04 0.06 1.67 1.93 0.85 1.53
5 0.10 1 0.05 0.07 1.69 1.95 0.90 1.66
Figure 11. Fragility curves for existing and retrofitted bridge piles for two damage states.
SEISMIC INVESTIGATION OF STEEL PILE BENTS: RETROFIT AND VULNERABILITY ANALYSIS 159
REFERENCES
American Association of State Highway and Transportation Officials (AASHTO), 1994. LRFD
Bridge Design Specifications, 1st Edition, Washington, DC.
Baso

z, N., and Mander, J. B., 1999. Enhancement of the Highway Transportation Lifeline Mod-
ule in HAZUS, final pre-publication report prepared for the National Institute of Building
Sciences.
Cheng, C-T., and Mander, J. B., 1997. Seismic Design of Bridge Columns Based on Control
and Repairability of Damage, Technical report, NCEER 97-0013, National Center for Earth-
quake Engineering Research, State University of New York at Buffalo, Buffalo, NY.
Chopra, A. K., 1995. Dynamics of Structures: Theory and Applications to Earthquake Engi-
neering, 1st ed., Prentice-Hall, Inc, Englewood Cliffs, NJ.
Dutta, A., and Mander, J. B., 1998. Seismic fragility analysis of highway bridge, INCEDE-
MCEER Center-to-Center Workshop on Earthquake Engineering Frontiers in Transportation
Systems.
Jernigan, J. H., Werner, S. D., and Hwang, H.H.M., 1996. Inventory of Bridges using GIS for
Seismic Risk Assessment of Highway Bridges, Research report, Center for Earthquake En-
gineering Research and Information, The University of Memphis, Memphis, TN.
Julian, O. G., 1955. Discussion of strength variation in ready-mixed concrete, J. Am. Concr.
Inst. 51 (12), 772-4772-8.
Kennedy, R. P. Cornell, C. A., Campbell, R. D., Kaplan, S., and Peria, H. F., 1980. Probabilistic
seismic safety study of an existing nuclear power plant, Journal of Nuclear Engineering and
Design 59, 315338.
Mirza, S. A., and MacGregor, J. G., 1979a. Variability of mechanical properties of reinforcing
bars, J. Struct. Div. ASCE 105, No. ST5, May 19211937.
Mirza, S. A., Hatzinikolas, M., and MacGregor, J. G., 1979b. Statistical descriptions of strength
of concrete, J. Struct. Div. ASCE 105, No. ST6, 10211037.
National Institute of Building Sciences (NIBS), 1997. Earthquake Loss Estimation Methodol-
ogy, HAZUS Technical Manual, prepared for the Federal Emergency Management Agency,
Washington, DC.
Newmark, N. M., and Hall, W. J., 1982. Earthquake Spectra and Design, Earthquake Engineer-
ing Research Institute, Berkeley, CA.
Pekcan, G., 1998. Design of Seismic Energy Dissipation Systems for Concrete and Steel Struc-
tures, Ph.D. dissertation, State University of New York at Buffalo, Buffalo, NY.
Priestley, M. J. N., Seible, F., and Calvi, G. M., 1996. Seismic Design and Retrofit of Bridges,
John Wiley & Sons, Inc., New York.
Shalon, R., and Reintz, R. C., 1955. Interpretation of strengths distribution as a factor in quality
control of concrete, Proceedings, Reunion Internationale des Laboratories dEssais et de
Recherches sur les Materiaux et les Constructions, 2, Laboratorio Naciano de Engenharia
Civil, Lisbon, Portugal, 100116.
Shama, A. A., Mander, J. B., Blabac, B. A., and Chen, S. S., 2002. Seismic investigation of steel
pile bents: I. Evaluation of performance, Earthquake Spectra 18 (1), 121.
(Received 4 January 2001; accepted 21 November 2001)
160 A. A. SHAMA, J. B. MANDER, AND S. S. CHEN