Vol.7, No.

EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION

Earthq Eng & Eng Vib (2008) 7:403-414

December, 2008

DOI: 10.1007/s11803-008-1006-5

Experimental research and finite element analysis of

bridge piers failed in flexure-shear modes
Sun Zhiguo1, Si Bingjun2; Wang Dongsheng1 and Guo Xun3
1. Institute of Road and Bridge Engineering, Dalian Maritime University, Dalian 116026, China
2. School of Civil & Hydraulic Engineering, Dalian University of Technology, Dalian 116024, China
3. Institute of Engineering Mechanics, China Earthquake Administration, Harbin 150080, China

Abstract:

In recent earthquakes, a large number of reinforced concrete (RC) bridges were severely damaged due to
mixed flexure-shear failure modes of the bridge piers. An integrated experimental and finite element (FE) analysis study is
described in this paper to study the seismic performance of the bridge piers that failed in flexure-shear modes. In the first part,
a nonlinear cyclic loading test on six RC bridge piers with circular cross sections is carried out experimentally. The damage
states, ductility and energy dissipation parameters, stiffness degradation and shear strength of the piers are studied and
compared with each other. The experimental results suggest that all the piers exhibit stable flexural response at displacement
ductilities up to four before exhibiting brittle shear failure. The ultimate performance of the piers is dominated by shear
capacity due to significant shear cracking, and in some cases, rupturing of spiral bars. In the second part, modeling approaches
describing the hysteretic behavior of the piers are investigated by using ANSYS software. A set of models with different
parameters is selected and evaluated through comparison with experimental results. The influences of the shear retention
coefficients between concrete cracks, the Bauschinger effect in longitudinal reinforcement, the bond-slip relationship
between the longitudinal reinforcement and the concrete and the concrete failure surface on the simulated hysteretic curves
are discussed. Then, a modified analysis model is presented and its accuracy is verified by comparing the simulated results
with experimental ones. This research uses models available in commercial FE codes and is intended for researchers and
engineers interested in using ANSYS software to predict the hysteretic behavior of reinforced concrete structures.

Keywords: RC bridge piers; flexure-shear failure; seismic behavior; finite element; ANSYS software

1 Introduction
In recent earthquakes, including the 1995 Kobe
earthquake (Hashimoto et al., 2005), 1999 Chi-Chi
earthquake (Chang et al., 2000) and 2008 Wenchuan
earthquake, a large number of bridges were severely
damaged as a result of a mixed flexure-shear failure
of the RC (reinforced concrete) bridge piers. Figure
1 shows some examples of this kind of failure for the
Baihua Bridge and Huilan Overpass Bridge. For the
latter bridge, four piers suffered from severe flexureshear mode damage and minor-to-moderate flexureshear cracks were observed in many other piers.
The bridge piers that are susceptible to flexure-shear
failure are short columns with a shear span/depth ratio
Correspondence to: Wang Dongsheng, Institute of Road and
Bridge Engineering, Dalian Maritime University, 1 Linghai
Road, Dalian 116026, China
Tel: 86-411-84725098
E-mail: dswang@newmail.dlmu.edu.cn

PhD; Associate Professor; Professor

Supported by: National Natural Science Foundation of China
Under Grant No. 50878033 and National Special Foundation
of Earthquake Science of China Under Grant No.200808021
Received September 22, 2008; Accepted October 15, 2008

between 1.52.5, and their ultimate performance are

dominated by brittle shear capacity. Therefore, it is
important to investigate their seismic performance to
advance seismic assessment and retrofitting techniques.
Even though an increasing amount of research is
becoming available on the seismic performance of
RC bridge piers (Priestley and Park, 1987; Jaradat et
al., 1998, 1999; Wehbe et al., 1999; Yeh et al., 2002;
Lehman et al., 2004), studies on flexure-shear dominated
RC bridge piers is scare. The number of test specimens
available for flexure-shear dominated cantilever RC
columns in the PEER (Pacific Earthquake Engineering
.,
Research Center) database is only eight (Berry
2004).
In this paper, a nonlinear cyclic loading test on six
RC bridge piers with flexure-shear failure mode was
carried out. The damage states, ductility and energy
dissipation parameters, stiffness degradation and shear
strength of the piers, are studied and compared with
each other. Then, modeling approaches describing
the hysteretic behavior of the piers are investigated
by using ANSYS software. A set of models with
different parameters is selected and evaluated through
comparison with experimental results. The influences of
the shear retention coefficients between concrete cracks,

404

EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION

Vol.7

(a) Baihua Bridge

(b) Huilan Overpass Bridge

Fig. 1 Flexure-shear failure of RC piers

Constant axial load

400mm

Specimen A3

Specimen A4

6@60

Specimen A2

Specimen A3

1214

m
0m

m
0m

0
3

0
3

00

1214
6@60

Specimen A4

Specimen A1

6@60

1014

6@80

6@80

00

m
m

00

00

1014

814

850mm

750mm

Lateral load

700mm

Specimens A1, A2, A5 and A6

1014

Constant axial load

550mm

Lateral load

450mm

600mm

Lateral load

550mm 700mm

Constant axial load

6@40
Specimen A5

Fig. 2 Details of the pier specimens

Specimen A6

No.4

Sun Zhiguo et al.: Experimental research and finite element analysis of bridge piers failed in flexure-shear modes

405

Table 1 Pier design parameters

Longitudinal
reinforcement

Diameter,
D (mm)

Aspect
ratio,
M/VD

fc (MPa)

Quantity

t (%)

s (mm)

s (%)

A1

300

29.4

814

1.74

80

0.51

0.15

A2

300

32.2

1014

2.18

80

0.51

0.15

A3

300

1.5

29.4

1014

2.18

60

0.67

0.10

A4

300

2.5

30.1

1014

2.18

60

0.67

0.10

A5

300

27.3

1214

2.61

60

0.67

0.15

A6

300

32.2

1214

2.61

40

1.01

0.10

Specimen

the Bauschinger effect in longitudinal reinforcement,

the bond-slip relationship between the longitudinal
reinforcement and the concrete, and the concrete failure
surface on the simulated results are discussed. Finally, a
modified analysis model is presented and its accuracy is
verified by experimental results.

2.1 Specimen description

Six RC bridge pier specimens representing about
1/3 scale of the prototype bridge piers are designed
and constructed, which are designated as specimen
A1, A2, A3, A4, A5 and A6, respectively. The details
of the specimens are shown in Fig. 2 and their design
parameters are listed in Table 1. Measured yield strengths
of the 14 longitudinal bars and 6 spiral bars are 327.6
MPa and 511 MPa, respectively.
The confining reinforcement requirements of
AASHTO LRFD 2005, Caltrans 2001 and Eurocode
8 1994 code provisions for circular RC bridge piers
are expressed by Eqs. (1), (2), and (3), respectively.
Table 2 presents spiral reinforcement ratios obtained by
these expressions for the specimens. It is concluded that
only specimen A6 satisfies all these provisions.

fc
Ag
f
s = 0.45 1 c but not less than s = 0.12
f yh
Ac
f yh
(1)

wwd,r 2.436

Ag
Ac

Ag
Ac

1)

f c
1.25 P
(0.5 + )
f yh
f c Ag

Axial load ratio,

P/Agfc

mechanical reinforcement ratio defined as

wwd,r = s

f yh

(4)

f c

wwd,r 0.12 for ductile pier and wwd,r 0.08 for limited
ductile pier.
Table 2 Required spiral reinforcement ratios of the pier
specimens

2 Experimental study

s = 0.45(

Spiral reinforcement

(2)

(0.009 c + 0.17) k 0.098 ww,min

(3)
where s is the spiral reinforcement ratio, Ag is the gross
area of pier section, Ac is the cross-sectional area of the
concrete core, fyh is the yield strength of the spiral bars,
P is the applied axial load, c is the required curvature
ductility factor, 13 for ductile piers and 7 for limited
ductile piers, k is the axial load ratio, wwd,r is the

Specimen

s,exp/s,AASH

s,exp/s,Caltr

s,exp/s,Euro

A1

0.62

0.89

0.77

A2

0.57

0.82

0.70

A3

0.82

1.31

1.01

A4

0.80

1.26

0.98

A5

0.88

1.26

1.08

A6

1.12

1.80

1.39

2.2 Testing setup and loading sequence

The testing setup for each of the specimens is shown
in Fig. 3. The specimen bottom is bolted to a strong

Rolling shaft
Vertical actuator

Load cell

Hinge

Horizontal actuator
Specimen

Reaction
frame

Specimen bolted
to strong floor
Fig. 3 Testing setup

406

EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION

reinforced concrete foundation, and its top is held by one

vertical actuator to provide a constant axial load. Also,
the lateral force is applied at the top of the specimen
by two horizontal actuators which are mounted to the
reaction frame.
The lateral loading history presented in Fig. 4 is
applied to all the specimens. The loading cycles are
divided into two phases: load control and displacement
control. The load control phase is used to define the
piers experimental yield displacement y; Besides, a
displacement control loading sequence is used. The
displacement controlled loading history includes three
complete cycles each for u,,exp=1, 2, 3, , until the
shear capacity of the piers declined and to 85% of the
peak loads. Here, u, exp is the ratio of the applied lateral
Lateral force

displacement at the top of the piers over the yield

displacement y.
As shown in Fig. 4 (a), the first three cycles of the
lateral load is applied to 70% of the theoretical yield
load Fy, which is calculated based on fiber model and
measured material properties. The yield displacement
y is determined by extrapolating a straight line from
the origin through the measured point corresponding
to 0.7Fy to the theoretical yield load Fy. The average of
the displacement values in both positive and negative
loading directions is used as the yield displacement y.
2.3 Observed damage states
The progression of damage was similar for all the

Displacement

Load

Fy
0.7Fy
Displacement

b a
+

y =

-0.7Fy
-Fy

a =
Ka =

Fy

0.7 Fy
+
3

+ =

j =1

Fy

b =

Ka

+
j

0.7Fy
-0.7Fy

a + b
2

Kb =

Kb

0.7 Fy

j =1

Vol.7

6
5
4
3
2
1
1
2
3
4
5
6

max
y

Number of cycles

(a) Experimental definition of yield displacement

(b) Loading sequence

Fig. 4 Lateral loading sequence for pier specimens

specimens. Flexural cracks perpendicular to the pier

axis developed first in regions close to the bottom of the
specimens. At later stages of loading, the flexural cracks
became inclined and extended into the neutral axis of
the specimens due to the influence of shear. Then, initial
spalling of the concrete cover was observed, once the
cover concrete had completely spalled and the spiral and
longitudinal reinforcement was exposed, longitudinal
bar buckling and concrete core crushing initiated within
the next displacement cycles. The ultimate performance
of the piers is dominated by shear capacity due to
significant shear cracking and in some cases rupturing
of spiral bars. Figure 5 shows the final damage states of
the specimens at the end of the tests. Note that the initial
yielding of the longitudinal and spiral reinforcement was
also measured using a strain gauge, as shown in Table 3.
It is useful to study key damage states of the piers as
each damage state may be associated with one or more
engineering limit state. In this study, the first occurrence
of each key damage state, such as longitudinal
reinforcement yielding, initial spalling of the concrete
cover, spiral reinforcement yielding, exposing of
spiral and longitudinal reinforcement, longitudinal
reinforcement buckling, spiral fracture, is identified in

Table 3.
The lateral force-displacement responses of all
specimens are shown in Fig. 19. In these figures,
indicates lateral displacement at the top of the pier and F
is the lateral force acting on the specimen.
The concrete crack widths are of some importance
in assessing the damage level of bridge piers. In
this study, the flexural and shear crack widths are
measured using a reading microscope during the tests.
Fig. 6 shows the measured maximum crack widths
at each displacement ductility level as defined in
Fig.8. In general, the flexural and shear crack widths
are almost the same at small displacement levels. At
large displacement levels, shear cracks grow faster than
flexural cracks, and this may be an important feature of
flexure-shear dominated RC bridge piers.
In the 1995 Kobe earthquake, bridge piers with
residual inclination R (the displacement at zero lateral
force divided by the height of the pier) larger than 1
were demolished, since it makes placing girders difficult
and causes visual uneasiness (Fujino et al., 2005). The
residual inclinations R at each displacement ductility
level for each specimen are plotted in Fig. 7. It is
obvious from the figure that the residual inclinations

No.4

Sun Zhiguo et al.: Experimental research and finite element analysis of bridge piers failed in flexure-shear modes

A1

A2

A3

A4

A5

A6

407

Fig. 5 Failure patterns of the pier specimens at the end of the tests

Table 3 The damage parameters of the pier specimens

A1

Longitudinal reinforcement
Yielding
Buckling
4.07 mm / 0.8 33.0 mm / 6.5 / 2

Spiral reinforcement
Yielding
Fracture
16.7 mm / 3.3

A2

3.05 mm / 0.7

32.8 mm / 7.3 / 1

12.8 mm / 2.8

A3

2.37 mm / 0.7

22.1 mm / 6. 1/ 3

9.4 mm / 2.6

A4

5.62 mm / 0.8

50.6 mm / 7.5 / 1

16.6 mm / 2.5

A5

5.30 mm / 0.9

36.5 mm / 6.0 / 1

15.8 mm / 2.6

41.6 mm / 6.8 / 3

A6

4.00 mm / 0.6

44.4 mm / 7.0/ 3

16.8 mm / 2.7

49.5 mm / 7.8 / 2

Specimen

36.0 mm / 8.0 / 1

Concrete cover spalling

Exposing of
reinforcement

11.1 mm / 2.2 / 3

27.5 mm / 5.4 / 1

12.8 mm / 2.8 / 3

24.5 mm / 5.4 / 1

9.4 mm / 2.6 / 3

16.3 mm / 4.5 / 3

16.6 mm / 2.5 / 1

33.0 mm / 4.9 / 3

10.7 mm / 1.8 / 3

26.4 mm / 4.3 / 3

11.4 mm / 1.8 / 3

38.9 mm / 6.2 / 3

Note: In expression a mm / b / c, a denote the displacement at the top of the specimen, b is the displacement ductility factor
defined in Fig.8, and c is the cycle number, if applicable.

increase linearly with the displacement ductility factors,

and Specimens A5 and A6 have the largest longitudinal
reinforcement ratios and biggest residual inclinations at
each displacement level.
2.4 Ductility and dissipated energy
In this study, the ductility parameters suggested by
Sheikh and Khoury (1993), and Lgeron and Paultre

(2000) are used to evaluate the seismic performance

of the specimens. Fig. 8 describes various ductility
parameters including displacement ductility factors ,
cumulative displacement ductility ratios N, normalized
dissipated energy EN, work index IW, and work damage
indicator W. The and N represent the deformability
of the member, the EN and IW are used to assess
energy dissipation capabilities, whereas W estimates
toughness. Table 4 lists the ductility parameters for all

EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION

6 5

Flexural cracks
Shear cracks

2.0

2.5
2.0

1.5

1.5

1.0

1.0

0.5

0.5

0
0 1

5 6

A4

3.0

Flexural cracks
Shear cracks

2.5
2.0
1.5
1.0
0.5

0
7 6 5 4 3 2 1 0 1 2 3 4 5 6 7

5.5
A3
5.0
4.5
Flexural cracks
4.0
Shear cracks
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0
7 6 5 4 3 2 1 0 1 2 3 4 5 6 7

A2

(mm)

(mm)

2.5

3.5
(mm)

3.0

A1

Flexural cracks
Shear cracks

0
7 6 5 4 3 2 1 0 1 2 3 4 5 6 7

5.0
A6
4.5
4.0
Flexural cracks
Shear cracks
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0
87 6 5 4 3 2 1 0 1 2 3 4 5 67 8

5.5
A5
5.0
4.5
Flexural cracks
4.0
Shear cracks
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0
7 6 5 4 3 2 1 0 1 2 3 4 5 6 7

(mm)

(mm)

3.0

Vol.7

(mm)

408

Fig. 6 Crack width vs. displacement ductility level of the specimens

amplitude, , the stiffness for the piers is defined as

3.5
3.0

R (%)

2.5
2.0
1.5
1.0

A1
A2
A3
A4
A5
A6

K =
3

0.0
7 6 5 4

3 2 1 0 1 2 3 4 5 6 7

8 9

Fig. 7 Residual inclinations vs. displacement ductility level of

the specimens

tested specimens. Note that the ductility factors of the

tested specimens are in the range from 5.14 to 7.48,
and Specimen A6 with most transverse reinforcement
has the largest ductility and dissipated energy related
parameters.

(5)

j , max

j =1

0.5

9 8

K + K

j =1

+
j

, K

j , max

j =1

(6)

j =1

where Fj,max is the maximum lateral load and j is

the corresponding displacement within a cycle at the
displacement .
For comparison, the calculated stiffness for each
displacement is normalized with respect to the stiffness
of the first three cycles. The normalized stiffness versus
the displacement ductility factor for the specimens
is shown in Fig. 9, where indicates the ratio of the
stiffness for a subsequent displacement ductility level
to that for an initial displacement ductility level and
indicates displacement ductility factor. Note that
the stiffness degradation rate for all specimens is very
similar.
2.6 Shear strength

2.5 Stiffness degradation

For every three cycles under the same displacement

Four approaches to estimate the shear strength of

RC bridge piers were used in this study: Caltrans 2001

No.4

Sun Zhiguo et al.: Experimental research and finite element analysis of bridge piers failed in flexure-shear modes

Ki+

Fmax
0.85Fmax
0.75Fmax

Fi+

1+

u
1 =

-0.75Fmax
-0.85Fmax
-Fmax

+
2

+
u =
2
K1 =

K1

Ki =

K i+ + K i
2

u
1

i =1

Area wi

Fi

+
1

Fi + + Fi
2

EN =

+
u

Fi ,max =

i=m

+
1

i+ + i
2

N =

i =

i=m
i =1

W =

K +K
2

i
1

1 i =m
wi
Fmax 1 i =1

Iw =

Ki

409

Fi ,max i
Fmax 1
1

Fmax 1

i=m

Ki i

1 1

w K
i =1

Fig. 8 Definition of the ductility and dissipated energy parameters of the bridge piers
Table 4 Results of ductility and energy dissipation related parameters
Specimen

1(mm)

u(mm)

EN

IW

A1

5.1

34.4

6.75

48.6

95.0

43.7

240.3

A2

4.5

31.2

6.93

73.3

125.0

67.3

425.6

A3

3.6

18.5

5.14

53.1

98.0

84.2

411.8

A4

6.7

48.9

7.30

86.7

158.1

79.1

579.7

A5

6.1

37.6

6.16

69.3

110.7

63.0

384.7

A6

6.3

47.1

7.48

100.2

147.3

92.7

779.6

1.4
1.2
1.0
0.8
u

vc = 1 2

A1
A2
A3
A4
A5
A6

0.6

2 = 1 +

0.2

Vs =

0
1

(9)

0.025 1 = 0.08s f yh + 0.305 0.083 < 0.25 (10)

0.4

f c 0.33 f c

P
< 1.5
13.8 Ag

(11)

Ab f yh D
s
2

(12)

Fig. 9 Stiffness degradation with increasing displacement

ductility level of the bridge piers

where Vc and Vs are shear resistance provided by concrete

and spirals, respectively; Ab is cross section area of spiral
reinforcement; D is the distance between centers of the
peripheral spiral.
The second model is based on Eurocode 8 shear
design equation, and is expressed as

model, Eurocode 8 1994 model, Priestley model (1994),

and Bi model (2004).
The first model is based on Caltrans shear design
equation, which is expressed as

Vn = Vc + Vs

Vn = Vc + Vs
Vc = vc (0.8 Ag )

(7)
(8)

Vc = 0
Vc = 2.5 (0.035 f

if

2 / 3
c

Dsp 2
4

(13)

0.1
for

> 0.1

(14)
(15)

410

EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION

= P / ( Ac f c )

(16)

Vs = ( Ab / s ) 0.9 Df yh

(17)

Vn = 0.15C1C2 f c (0.8 Ag ) +

(18)

Vc = k f c (0.8 Ag )

(19)

Vp =

Vs =

Dc
P
2a

C2 =

(21)

0 1

2 3

4 5 6 7

8 9 10

A4

Test
Caltrans
Eurocode 8
Priestley
Bi & Fan

0 1

5 6

8 9 10

A2

F (kN)

F (kN)

Test
Caltrans
Eurocode 8
Priestley
Bi & Fan

240
210
180
150
120
90
60
30
0

Test
Caltrans
Eurocode 8
Priestley
Bi & Fan

0 1

270
240
210
180
150
120
90
60
30
0
0

2 3

4 5 6 7

8 9 10

A5

F (kN)

270
240
210
180
150
120
90
60
30
0

A1

F (kN)

F (kN)

F (kN)

where Vp is the shear resistance provided by the axial

load; k is the deterioration factor for concrete; c is the
depth of the compressive zone in a column; a=M/V, is
the ratio of moment to shear at critical section.
The fourth model, proposed by Bi (2004), is
expressed as

240
210
180
150
120
90
60
30
0

Test
Caltrans
Eurocode 8
Priestley
Bi & Fan

2 3

(22)
(23)

1
+ 2.0

(24)

A comparison of the measured shear strengthdisplacement ductility relationship and the predicted
values obtained by using different models is given in
Fig. 10, where the shear strength Fshear is defined as
Fshear=F, max, F, max is the lateral force of the specimen
corresponding to the lateral displacement in the first
cycle, and the lateral force in the last cycle must be equal
or less than 0.85F, max. The shear strength Fshear for each
specimen is indicated in Fig.10 by symbol X.
Note that for specimens A1 and A5, except for the
Priestley model that yields much larger shear capacity
than the test results, the other three models predict the
shear capacity with acceptable accuracy. Also, all four
models predict the shear capacity well for specimen A2,
but overestimate the shear capacity of specimen A6. For
specimen A3, all models predict the shear capacity very
well except for the Eurocode 8 model, from which the
specimen may fail in shear with much lower strength
than in the test results. For Specimen 4, the Eurocode 8
model yields shear capacity very close to the maximum
shear force from the tests, but the remaining three models
give greater shear capacity than in the test results.

(20)

Ab f yh D
cot 30
s
2

Ab f yh
D
2 s

C1 = 1 + 2.2k s

where
Dsp is the spiral diameter, D is circular column
diameter. The third model, the one proposed by Priestley
et al., is expressed as
Vn = Vc + Vp + Vs

Vol.7

4 5 6 7

8 9 10

270
240
210
180
150
120
90
60
30
00 1

360
330
300
270
240
210
180
150
120
90
60
30
0

A3

Test
Caltrans
Eurocode 8
Priestley
Bi & Fan

2 3

4 5 6 7

8 9 10

A6

Test
Caltrans
Eurocode 8
Priestley
Bi & Fan

0 1

2 3

Fig. 10 Shear strength-displacement ductility relationship for all specimens

4 5 6 7

8 9 10

Sun Zhiguo et al.: Experimental research and finite element analysis of bridge piers failed in flexure-shear modes

3 Numerical study
Next the capability of the commercially available
finite element analysis software ANSYS (2004) to model
the hysteretic behavior of RC bridge piers is evaluated.
The software ANSYS was used in the finite element
(FE) analysis of the pier specimens. First, a series of
FE models for specimen A3 are constructed using the
ANSYS software to evaluate the influence of material
models and their associated parameters on the hysteretic
response. Then, a modified analysis model is presented
and the model accuracy was verified by comparing the
calculated hysteretic curves with the experimental results.
Solid 65 elements which have crushing
(compressive) and cracking (tensile) capabilities were
used to model the concrete. All reinforcements were
modelled using Link 8 truss elements. Solid 45 elements
were used for the steel plates at the support and under
the load. The effect of bond-slip at the interface between
concrete elements and truss elements was simulated
using Combin 39 elements.
To account for the confinement effect, the Mander
model (Mander et al., 1988) for confined stress-strain
relationship with an assumption of perfectly plastic
after ultimate compression strength is used to define the
constitutive relationship of the concrete. Meanwhile, the
multilinear kinematic hardening relationship, using the
von Mises yield criterion, was also adopted. In addition,
the Willam and Warnke five parameter model is used as
the failure criterion of concrete under multiaxial stress
conditions, in which the failure surface is defined by
at least two constants: the concrete ultimate uniaxial
tensile strength, ft , and the ultimate uniaxial compressive
strength, fc.
Taking advantage of the symmetry of the geometry
and the reinforcement of the specimen, only half of the
specimens are modelled. Figure 11 shows the FE model
of specimen A3.
3.1 Influence of shear retention coefficient
After cracking, the tension stress in the concrete

411

element is set to zero in the direction normal to the

crack plane. The shear transfer coefficient t for open
cracks and c for closed cracks determine the amount
of shear transferred across the cracks. The value of the
shear transfer coefficient ranges from 0 to 1.0, with 0
representing no shear transfer at a crack section and 1.0
representing full shear transfer. In this study, the shear
transfer coefficient t for open cracks is assumed to
be 0.2, 0.3 and 0.5, while for closed cracks the shear
transfer coefficient c is assumed to be 0.5, 0.7 and 0.95,
respectively. As shown in Fig. 12, the simulated hysteretic
curves using different shear transfer coefficients are
almost the same. Therefore, it could be concluded that
the shear transfer coefficient does not have an obvious
influence on the simulated hysteretic response in this
study. This may be because the fixed crack model used
in the ANSYS software could not precisely reflect the
shear transfer capability across concrete cracks.
3.2 Influence of Bauschinger effect
To investigate the Bauschinger effect for the
longitudinal reinforcement in simulating the hysteretic
behavior of the pier specimen, two models are used:
the bilinear kinematic hardening (BKH) model without
Bauschinger effect (model 1), and the multilinear
kinematic hardening (MKH) model with Bauschinger
effect (model 2) (see Figs. 13 (a) and (b)). Figure
14 depicts the simulated hysteretic curves given by
models 1 and 2. Note that the Bauschinger effect in the
longitudinal reinforcement has a significant influence on
the pinching effect in the hysteretic response.
3.3 Influence of bond-slip effect
To investigate the effect of bond-slip between
longitudinal reinforcement and concrete on the
simulated hysteretic response of the pier specimen, two
FE models are used: model 3 and model 4. In model 3,
perfect bond between the concrete and the longitudinal
reinforcement is assumed, whereas in model 4 the bondslip is incorporated. As illustrated in Fig. 15, the bondA3

200
F (kN)

No.4

150
100
50

-15

-10

-5

0
0
-50

10
(mm)

-100

t = 0.2; c = 0.5

-150

t = 0.3; c = 0.7

-200
Fig. 11 FE model for specimen A3

15

t = 0.5; c = 0.95

Fig. 12 Influence of the shear retention coefficients on the

hysteretic behavior

412

EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION

Vol.7

fs

fs
E s=0.01Es

fy

Es=0.01Es

fy
1

0.7fy

Es=1/4Es

Es

Es
1

1
s

(a) Bilinear kinematic hardening model

(b) Multilinear kinematic hardening model

Fig. 13 Two constitutive models for longitudinal reinforcement

250

-15

-10

-5

250

A3

200

F (kN)

F (kN)

A3

150

200
150

100

100

50

50

0
0
-50
-100
-150

10
(mm)

15

BKH model
MKH model

-15

-10

-5

0
0
-50
-100
-150

-200

-200

-250

-250

10
(mm)

15

Bond-slip ignored
Bond-slip included

Fig. 14 Influence of constitutive models of reinforcing steel on

the hysteretic behavior

Fig. 16 Influence of the bond-slip between the longitudinal

reinforcement and concrete on the hysteretic behavior

slip model for the interface element has a simplified

linear relationship to the slip strength c with a constant
stress after the critical slip displacement c. Experimental
results of the bond-slip relationship between the concrete
and the longitudinal reinforcement were not available at
the time of this modelling. Therefore, representative
values for c and c are selected as 10 MPa and 0.1 mm.
Fig.16 depicts the simulated hysteretic curves by model
3 and 4. It is recognized that a certain degree of bondslip may have contributed to the pinching effect in the
hysteretic response. Also, model 4 predicts a low lateral
load at large lateral displacement as a result of the bondslip effect.

3.4 Influence of failure surface for concrete

Fig. 15 Relationship of the bond-slip between the longitudinal

reinforcement and concrete

It has been reported in the literature that if both

cracking and crushing capabilities for Solid 65 element
are activated in the ANSYS software, fictitious crushing
of the concrete may be caused due to the coupling of
excessive cracking strains to the orthogonal uncracked
directions through Poissons effect. This may be one of
the reasons that cause divergence of the solution at later
stages (Zhou et al., 2004). Therefore, in most previous
literature, the crushing capability of the concrete is
turned off and the crushing failure of the concrete is
ignored (Si et al., 2007).
Obviously, the FE model in which the crushing
capability of the concrete is turned off cannot predict
all the failure features of the specimens. However, it is
also found in this study that if fc is used to define the
concrete failure surface, the FE model fails prematurely
(Fig. 17).
In order to simulate the crushing failure of the
concrete in the conducted experiments, an enlarged
failure surface is adopted, i.e., 1.22 times of fc are
used to define the failure surface, but the normal stressstrain curves are still used to define the constitutive
relationship of the concrete. Figure 18 illustrates the
simulated hysteretic response by enlarged failure

Sun Zhiguo et al.: Experimental research and finite element analysis of bridge piers failed in flexure-shear modes

surface for concrete. It is concluded that the model using

an enlarged failure surface of concrete predicts the piers
load-displacement relationship well.
3.5 Modified FE model
A modified FE model is presented based on the
above analysis. In this model, a multilinear kinematic
hardening model is used to include the Bauschinger
effect in the longitudinal reinforcement under cyclic

250

250

-30

-20

A3

200

F (kN)

A3

150

-10

50

50

-150
-200

10

20
(mm)

-30

30

-20

-200

Fig. 18 Influence of enlarged failure surface for concrete to

the hysteretic behavior

50

100

0
10 20 30 40 50 -40 -30 -20 -10 0
(mm)
-50
-100

Experiment
Simulation

-150

20 30 40 -30
(mm)

-20

Experiment
Simulation

200

100

10

20
30
(mm)

Experiment
Simulation

200

A6

150

50

0
0
-60-50-40-30 -20 -10 0 10 20 30 40 50 60 -50 -40 -30 -20 -10 0
-50
(mm)
-50
-100
Experiment
Simulation
100
-150
-150

10

-200
A5

250
200
150
100
50
0
-10
0
-50
-100
-150
-200
-250

F (kN)

F (kN)

100

150

50

F (kN)

150

A3

F (kN)

F (kN)

F (kN)

200

-150
A4

30

Experiment
Simulation

-150

A2

50

100

20
(mm)

-250

100

-50

10

-100

Experiment
Simulation

Fig. 17 Influence of the crushing for concrete to the hysteretic

curves

150

0
0
-50

-10

-250

0
-50 -40 -30 -20 -10 0

150
100

-100

A1

200

100
0
0
-50

413

loading, the bond-slip relationship between longitudinal

reinforcement and concrete is accounted for by using
combin 39 element, and an enlarged failure surface for
concrete is used to simulate the crushing of concrete in
the tests, which can also prevent fictitious crushing of
the concrete in the simulation. The simulated hysteretic
curves are compared with the experimental results, as
illustrated in Fig. 19, where it is seen that the calculated
hysteretic curves correspond well with the experimental
ones.

F (kN)

No.4

150
100
50

0
10 20 30 40 50 -60-50 -40-30-20 -10 0 10 20 30 40 50 60
(mm)
(mm)
-50
Experiment
Simulation

-200
Fig. 19 Comparisons of the experimental and simulated hysteretic curves

-100
-150
-200

Experiment
Simulation

414

EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION

4 Conclusions
An integrated experimental and FE analysis study is
described in this paper to study the seismic performance
of bridge piers that failed in flexure-shear modes. In the
first part, a nonlinear cyclic loading test on six RC bridge
piers with circular cross sections was performed, and all
the piers failed in the flexure-shear mode. In the second
part, modeling approaches describing the hysteretic
behavior of the piers were investigated by using ANSYS
software. A set of models with different parameters
was selected and evaluated through comparison with
experimental results. Then, a modified analysis model
is presented and its accuracy was verified by comparing
the simulated results with the experimental ones. Based
on the studies presented in this paper, the following
conclusions can be made.
(1) The progression of damage is similar for all
the specimens: concrete flexural cracking, longitudinal
reinforcement yielding, concrete shear cracking,
concrete cover spalling, spiral reinforcement yielding,
reinforcement exposing, longitudinal reinforcement
buckling, and concrete core crushing. The ultimate
performance of the specimens is dominated by shear
capacity due to significant shear cracking, and in some
cases, rupturing of spiral bars.
(2) The observed flexural and shear crack widths are
almost the same at small displacement levels, but at large
displacement levels, the shear cracks grow faster than the
flexural cracks. The displacement ductility factors of the
tested specimens range from 5.14 to 7.48, and Specimen
A6 with most transverse reinforcement has the largest
ductility and dissipated energy parameters. All the
specimens have similar stiffness degradation rates. The
predicted accuracy of the current shear strength model
for the specimens with flexure-shear failure modes
needs to be further improved.
(3) Simulated hysteretic behavior of the pier
is strongly influenced by the Bauschinger effect in
longitudinal reinforcement, bond-slip relationship
between longitudinal reinforcement and concrete,
and choice of concrete failure surface. However, the
hysteretic response is not significantly affected by the
shear transfer coefficient between concrete cracks for
the considered pier specimens.
(4) The modified FE model using ANSYS software
predicts the piers hysteretic response well.

References
AASHTO LRFD (2005), AASHTO LRFD Bridge Design
Specifications, American Association of State Highway
and Transportation Officials, Washington, D.C.
ANSYS (2004), ANSYS Users Manual, ANSYS, Inc.,
Canonsburg, Pennsylvania, USA.
Berry M, Parrish M and Eberhard M (2004), PEER
Structural Performance Database Users Manual,
Version 1.0, University of California, Berkeley.
Bi Guiping (2004), Research on the Key Problems for
Seismic Behavior of Large-scale Complicated Overcross
Engineering, PhD. Dissertation, Tongji University. (in
Chinese)
Caltrans (2001), Seismic Design Criteria, California
Department of Transportation, Sacramento, California.

Vol.7

Chang KC, Chang DW, Tsai MH and Sung YC (2000),

Seismic Performance of Highway Bridges, Earthquake
Engineering and Engineering Seismology, 2(1): 5577.
Eurocode 8 (1994), Design Provisions for Earthquake
Resistance of Structures-Part 2: Bridges, European
Committee for Standardization, Brussels.
Fujino Y, Hashimoto S and Abe M (2005), Damage
Analysis of Hanshin Expressway Viaducts During 1995
Kobe Earthquake. I: Residual Inclination of Reinforced
Concrete Piers, Journal of Bridge Engineering, ASCE,
10(1): 4553.
Hashimoto S, Fujino Y and Abe M (2005), Damage
Analysis of Hanshin Expressway Viaducts During 1995
Kobe Earthquake. II: Damage Mode of Single Reinforced
Concrete Piers, Journal of Bridge Engineering, ASCE,
10(1): 5460.
Jaradat OA, Mclean DI and Marsh ML (1998),
Performance of Existing Bridge Columns Under Cyclic
Loading-Part 1: Experimental Results and Observed
Behavior, ACI Structural Journal, 95(6): 695704.
Jaradat OA, Mclean DI and Marsh ML (1999),
Performance of Existing Bridge Columns Under
Cyclic Loading-Part 2:Analysis and Comparisons with
Theory, ACI Structural Journal, 96(1): 5767.
Lgeron F and Paultre P (2000), Behavior of Highstrength Concrete Columns Under Cyclic Flexure and
Constant Axial Load, ACI Structural Journal, 97(4):
591601.
Lehman D, Moehle J, Mahin S, Calderone A and Henry
L (2004), Experimental Evaluation of the Seismic
Performance of Reinforced Concrete Bridge Columns,
Journal of Structural Engineering, ASCE, 130(6): 869879.
Mander JB, Priestley MJN and Park R (1988),
Theoretical Stress-strain Model for Confined
Concrete, Journal of Structural Engineering, ASCE,
114(8): 18041826.
Priestley MJN and Park R (1987), Strength and
Ductility of Concrete Bridge Columns Under Seismic
Loading, ACI Structural Journal, 84(1): 5175.
Priestley M J N, Verma R and Xiao Y (1994), Seismic
Shear Strength of Reinforced Concrete Columns,
Journal of Structural Engineering, ASCE, 120(8):
23102329.
Sheikh S A and Khoury S S (1993), Confined Concrete
Columns with Stubs, ACI Structural Journal, 90(4):
414431.
Si Bingjun, Sun Zhiguo and Ai Qinghua (2007),
Application of Solid65 Element in the Finite
Element Analysis of Concrete Structures, Industrial
Construction, 37(1): 8792. (in Chinese)
Wehbe NI, Saiidi MS and Sanders DH (1999), Seismic
Performance of Rectangular Bridge Columns with
Moderate Confinement, ACI Structural Journal, 96(2):
248258.
Yeh YK, Mo YL and Yang CY (2002), Full-scale Tests
on Rectangular Hollow Bridge Piers, Materials and
Structures, 35(2): 117125
Zhou S, Rizos DC and Petrou MF (2004), Effects of
Superstructure Flexibility on Strength of Reinforced
Concrete Bridge Decks, Computers & Structures,
82(1): 1323.