A Study of Nonlinear Time History Analysis vs.

Current Codes Analysis

Procedure of Comparing Linear Dynamic Demand with Nonlinear Static
Capacity for Ordinary Standard Bridge
Downloaded from ascelibrary.org by New York University on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

Xiaoyun Wu1
1
Ph.D., P.E., IDC Consulting Engineers Inc, 300 S Harbor Blvd, Ste 710, Anaheim,
CA 92805; PH (714) 520-9070; FAX (714) 520-9068; email: [email protected]

ABSTRACT

Nonlinear time history analysis is known simulating a structure behavior under severe
earthquake more proper than other methods. However for simplicity, most of bridges
in category of Ordinary Standard Bridge (OSB) are being analyzed by a combined
procedure which consists of a linear ARS analysis for earthquake response (demand)
and a static nonlinear pushover for ultimate displacement (capacity) per guideline of
many transportation agencies worldwide. The demand and capacity are then
compared to determine the safety of the bridge. For single degree freedom (SDF)
system this procedure has been proven an effective method with satisfactory
accuracy. For bridges in the category of OSB but with noticeable characteristics of
MDF (multi-degree of freedom) system, large discrepancies between deformation
patterns from linear analysis and nonlinear pushover are often observed by engineers,
so the accuracy of conclusion from this procedure is questioned. To explore nonlinear
dynamic behavior of these bridges and investigate the adequacy of the popular
combined linear with nonlinear analysis procedure, a series bridges within category of
OSB ranging from slight to severe mass and stiffness unbalance were analyzed. The
analysis methods used for each bridge include linear and nonlinear time history
analysis, linear ARS analysis and nonlinear static pushover. To ensure valid results
comparison, a ground acceleration time history is used for both linear and nonlinear
time history analysis, its frequency domain ARS curve is used for ARS analysis.
Selected bridges model, ground acceleration and analysis methods, procedures,
results, comparison, discussion, conclusions and suggestions are all presented in the
paper.

INTRODUCTION

Due to the complicity and time consuming of bridge nonlinear time history analysis,
most of bridges in OSB category are currently designed based on a simplified seismic
analysis procedure which consists of a linear ARS analysis for earthquake response
(demand) and a static nonlinear pushover for ultimate displacement (capacity) per
guideline of many transportation agencies worldwide. The demand from elastic
analysis is compared to the capacity from nonlinear analysis to determine the safety
of the bridge. This simplified procedure is an extended application of a proven
method for single degree freedom (SDF) system. The accuracy of this simplified

467

Challenges and Advances in Sustainable Transportation Systems

 468 CHALLENGES AND ADVANCES IN SUSTAINABLE TRANSPORTATION SYSTEMS

procedure is often questioned when engineers are working on the bridge which
possesses characteristics of MDF (multi-degree of freedom) system even they do
belong to category of OSB, because noticeable discrepancies between deform pattern
Downloaded from ascelibrary.org by New York University on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

from linear analysis and nonlinear pushover are observed. To explore nonlinear
dynamic behavior of these bridges and investigate the adequacy of the combined
linear with nonlinear analysis procedure, a series bridges within category of OSB
ranging from symmetric, slight to severe mass and stiffness unbalance are analyzed.
The analysis methods used for each bridge include linear and nonlinear time history
analysis, linear ARS analysis and nonlinear static pushover. A selected ground
acceleration time history is used for both linear and nonlinear time history analysis, its
frequency domain ARS curve is used for ARS analysis. Two single column models
subjected to earthquake inputs of different strengths are also analyzed to show how
well the aforementioned simplified procedure works for SDF system and how it is
changed in MDF system. This paper presents the investigation analysis which
includes selected bridges model, ground acceleration, analysis methods, procedures,
results, comparison, discussion, conclusions and suggestions.

SAMPLE STRUCTURES AND ANALYSIS MODELS

Analyzed Bridges and Systems. The investigation and analysis include total four
bridge structures and two single columns, namely Bridge-1, Bridge-2, Bridge-3,
Bridge-4, Column-A and Column-B. All bridges and columns are concrete structures.
The table below listed the geometry and type of the analyzed bridges.
Table 1. Geometry Data and Type of the Analyzed Bridges
Bridge Bridge Span Skewed
# of Column Superstructure Substructure
Name Length Width Length Angle
Span Length type type
(ft) (ft) (ft) (degree)
60.53 A1: 0 - 5'-0' column on
Concrete precast
pile group, fix-
Bridge-1 3 202.30 27'-8" 81.24 Bnts: 0 B2: 26'-5" P/S I-Girder fix joints. Seat
with CIP Deck
60.53 A4: 0 B3: 27'-1" type abutment
60.53 A1: 45 - 5'-0' column on
Concrete precast
pile group, fix-
Bridge-2 3 202.30 27'-8" 81.24 Bnts: 45 B2: 26'-5" P/S I-Girder fix joints. Seat
with CIP Deck
60.53 A4: 45 B3: 27'-1" type abutment
70.00 A1: 19 - 5'-0' column on
Concrete precast
pile group, fix-
Bridge-3 3 214.70 33'-3" 82.00 Bnts: 19 B2: 26'-0" P/S I-Girder fix joints. Seat
with CIP Deck
62.70 A4: -23 B3: 26'-0" type abutment
67.54 - -
42' at 97.92 - B2: 64'-3" 6'-0" column on
Abut-1 Concrete precast
97.92 - B3: 67'-3" 8'-0" CIDH pile
Bridge-4 6 540.78 and P/S I-Girder
shaft. Seat type
taped to 97.80 A1: 25.05 B4: 70'-2" with CIP Deck
abutment
15' 105.73 Bnts: 25.05 B5: 72'-5"
73.88 A7: 25.05 B6: 74'-8"
Note: A1 is Abutment 1, B2 is Bent 2, and so on.

Challenges and Advances in Sustainable Transportation Systems

 CHALLENGES AND ADVANCES IN SUSTAINABLE TRANSPORTATION SYSTEMS 469

It can be seen from the above table that Bridge-1 is a symmetric structure
system. The skewed angle and un-uniformity increased from Bridge-2 to Bridge-3
and Bridge-4 is the most irregular one among all four bridges.
Downloaded from ascelibrary.org by New York University on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

Column-A and Column-B both are 5’-0” diameter reinforced concrete

cantilevered columns with a clear depth of 21’-6”. Column-A supports a concentrated
weight of 600 kips at top of the column to represent superstructure while Column-B
supports 750 kips. Both columns represent SDF systems with different natural
frequency due to difference of the supporting mass.
The computer models of all the analyzed bridges are shown below in Figure 1
through Figure 4.

Figure 1. Computer Model of Bridge-1 Figure 2. Computer Model of Bridge-2

Figure 3. Computer Model of Bridge-3 Figure 4. Computer Model of Bridge-4

Ground Motions. Imperial Valley earthquake record IMPVALL/H-E07239 with 5%

damping and 0.463g PGA recorded on October 1979 is selected as seismic excitation
for the study. The 30 seconds acceleration record is used in both linear elastic time
history analysis and nonlinear time history analysis. The frequency domain ARS
curve of the record is also used for ARS analysis of all bridge models for obtaining
demand displacements as specified in most of the bridge seismic design guidelines.
Kobe earthquake record of year 1995, KOBE/KJM000 with 5% damping and 0.821g
PGA is also applied as input to some models for both linear and nonlinear analysis to
see the sensitivity of the structure to different ground motions and to have better
universality of the conclusions drawn from the study. The ground motions are from
PEER (Pacific Earthquake Engineering Research Center) with headquarters at
University of California, Berkeley.
To investigate bridge seismic response in different post yielding stages,
Imperial Valley earthquake record is multiplied by a series of constants to create
several seismic excitations with different strengths. In order to clearly identify the
structural 3-D response effect, single direction seismic input is used in each analysis

Challenges and Advances in Sustainable Transportation Systems

 470 CHALLENGES AND ADVANCES IN SUSTAINABLE TRANSPORTATION SYSTEMS

case. The acceleration records and their frequency domain ARS curves are shown in
Figure 5 and Figure 6.
Downloaded from ascelibrary.org by New York University on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

Figure 5. Acceleration Record and ARS Curve of Imperial Valley Earthquake,

1979

Figure 6. Acceleration Record and ARS Curve of Kobe Earthquake, 1995

Analysis Method and Analysis Cases. For each bridge model elastic ARS analysis,
static nonlinear pushover analysis, linear time history analysis, and nonlinear time
history analysis are carried out. In summary, total 140 cases’ analysis are carried out
for this study with the computer program of SAP2000. The detail models and cases
are listed in the Table 2 through Table 4 below for reference.
Table-2. Summary of Analysis Cases for SDF System (Single Columns)
Ground Motion: Imperial Valley, 1979 Ground Motion: Kobe, 1995
Excitation Cloumn-A Cloumn-B Excitation Cloumn-A
Muliplier Analysis Case Analysis Case Muliplier Analysis Case
ARS ARS ARS
Static Pushover Static Pushover Static Pushover
1.00 0.50
Linear Time History Linear Time History Linear Time History
Nonlinear Time History Nonlinear Time History Nonlinear Time History
ARS ARS ARS
Static Pushover Static Pushover Static Pushover
1.25 0.75
Linear Time History Linear Time History Linear Time History
Nonlinear Time History Nonlinear Time History Nonlinear Time History
ARS ARS ARS
Static Pushover Static Pushover Static Pushover
1.50 1.00
Linear Time History Linear Time History Linear Time History
Nonlinear Time History Nonlinear Time History Nonlinear Time History

Challenges and Advances in Sustainable Transportation Systems

 CHALLENGES AND ADVANCES IN SUSTAINABLE TRANSPORTATION SYSTEMS 471

Table 3. Summary of Analysis Cases for Bridge-1 & Bridge-2

Bridge-1 Bridge-2
Ground Incidence
Downloaded from ascelibrary.org by New York University on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

Excitation Excitation
Motion Direction Analysis Case Analysis Case
Muliplier Muliplier
ARS ARS
Static Pushover Static Pushover
L-Dir
Linear Time History Linear Time History
Nonlinear Time History Nonlinear Time History
1.00 1.00
ARS ARS
Static Pushover Static Pushover
T-Dir
Linear Time History Linear Time History
Nonlinear Time History Nonlinear Time History
ARS ARS
Imperial Valley, 1979

Static Pushover Static Pushover

L-Dir
Linear Time History Linear Time History
Nonlinear Time History Nonlinear Time History
1.10 1.10
ARS ARS
Static Pushover Static Pushover
T-Dir
Linear Time History Linear Time History
Nonlinear Time History Nonlinear Time History
ARS ARS
Static Pushover Static Pushover
L-Dir
Linear Time History Linear Time History
Nonlinear Time History Nonlinear Time History
1.20 1.20
ARS ARS
Static Pushover Static Pushover
T-Dir
Linear Time History Linear Time History
Nonlinear Time History Nonlinear Time History
ARS
Static Pushover
L-Dir
KOBE, 1995

Linear Time History

Nonlinear Time History
1.00 -
ARS
Static Pushover
T-Dir
Linear Time History
Nonlinear Time History

For both linear and nonlinear time history analysis, direct integration with
Hilber-Hughes-Taylor method is used. Rayleigh damping with coefficients defined by
structure nature periods is applied. Nonlinear P-∆ effect is also included in the
nonlinear time history analysis and the analysis starts after the structure dead load is
fully loaded. For static nonlinear pushover, two lateral loading cases are pushed for
each model. One is associated with uniform lateral gravity loading and the other is a
set of concentrated force at top of each column which equals to column elastic shear
force from the bridge’s ARS analysis. Both sets of lateral load have very closed
pushover results. All analyzed bridges are detailed to have nonlinear plastic hinges
concentrated at substructure columns, and abutment shear keys are designed to be

Challenges and Advances in Sustainable Transportation Systems

 472 CHALLENGES AND ADVANCES IN SUSTAINABLE TRANSPORTATION SYSTEMS

fused when bridge subjects to design earthquake per current AASHTO and Caltrans
seismic design criteria. Therefore in the models the structure nonlinear plastic
components are column plastic hinges represented by PMM yielding surface at top
Downloaded from ascelibrary.org by New York University on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

and bottom of the columns as well as the shear key residual resistance represented by
lateral nonlinear links. The shear key residual elastic stiffness and maximum
resistance is assigned a value of about 10% of the adjacent column bent to prevent
girder’s swinging effect per code. The soil passive resistance at abutment is ignored
for better observation of the substructure columns’ dynamic behavior. All bridges in
the study belong to OSB category bridges. The 1st natural period of each bridge is
greater than 0.7sec and the difference of stiffness/mass ratio of adjacent bent is within
25%. The difference of this ratio of any two bents in a frame is less than 50%.

Table 4. Summary of Analysis Cases for Bridge-3 & Bridge-4

Bridge-3 Bridge-4
Ground Incidence
Excitation Excitation
Motion Direction Analysis Case Analysis Case
Muliplier Muliplier
ARS ARS
Static Pushover Static Pushover
L-Dir
Linear Time History Linear Time History
Nonlinear Time History Nonlinear Time History
1.00 1.00
ARS ARS
Static Pushover Static Pushover
T-Dir
Linear Time History Linear Time History
Nonlinear Time History Nonlinear Time History
ARS ARS
Imperial Valley, 1979

Static Pushover Static Pushover

L-Dir
Linear Time History Linear Time History
Nonlinear Time History Nonlinear Time History
1.25 1.25
ARS ARS
Static Pushover Static Pushover
T-Dir
Linear Time History Linear Time History
Nonlinear Time History Nonlinear Time History
ARS ARS
Static Pushover Static Pushover
L-Dir
Linear Time History Linear Time History
Nonlinear Time History Nonlinear Time History
1.50 1.50
ARS ARS
Static Pushover Static Pushover
T-Dir
Linear Time History Linear Time History
Nonlinear Time History Nonlinear Time History

ANALYSIS RESULTS AND DISCUSSION

SDF Systems. Table 5 below listed the displacements of Column-A and Column-B
from ARS analysis, LTH (linear time history) analysis, NLTH (nonlinear time
history) analysis and static nonlinear pushover analysis. It can be seen that the ratios
of the maximum displacement of ARS analysis to LTH analysis are about 1.0. This is

Challenges and Advances in Sustainable Transportation Systems

 CHALLENGES AND ADVANCES IN SUSTAINABLE TRANSPORTATION SYSTEMS 473

expected because both are linear dynamic response to the same earthquake record.
The ratios of the maximum displacement of LTH analysis to NLTH analysis are also
very close to 1.0 disregarding the change of the ground motion and structure. This
Downloaded from ascelibrary.org by New York University on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

result re-confirms that the popular simplified seismic analysis procedure which
treating the displacement from linear elastic dynamic analysis as real structure
displacement response including nonlinear behavior works great for SDF. However a
few cases among all 12 analyzed cases the NLTH displacement is smaller than LTH
displacement with a difference about 35%. These special cases require more study.

Table 5. Top of Column Displacement of SDF Sytem Models

Ground Motion: Imperial Valley, 1979 Ground Motion: Kobe, 1995
Column-A Column-B Column-A
Analysis
1st Mode T = 0.8094 sec 1st Mode T = 0.9036 sec 1st Mode T = 0.8094 sec
Method
1.0EQ 1.25EQ 1.5EQ 1.0EQ 1.25EQ 1.5EQ 0.5EQ 0.75EQ 1.0EQ
∆max (ft) ∆max (ft) ∆max (ft) ∆max (ft) ∆max (ft) ∆max (ft) ∆max (ft) ∆max (ft) ∆max (ft)
Yield 0.13 0.13 0.13 0.14 0.14 0.14 0.13 0.13 0.13
Static NL
Pushover 1.21 1.21 1.21 1.19 1.19 1.19 1.21 1.21 1.21
ARS 0.70 0.87 1.05 0.67 0.84 1.01 0.51 0.76 1.01
LTH 0.70 0.88 1.05 0.67 0.83 1.00 0.48 0.73 0.97
NLTH 0.63 0.85 1.08 0.63 0.86 1.15 0.46 0.73 0.89
Ratio of
ARS/LTH 1.00 1.00 1.00 1.01 1.01 1.01 1.04 1.04 1.04
Ratio of
NLTH/LTH 0.90 0.97 1.03 0.94 1.03 1.15 0.96 1.00 0.92
Ratio of
NLTH/Push 0.52 0.70 0.89 0.53 0.72 0.97 0.38 0.60 0.73
Ratio of
NLTH/Yield 4.69 6.33 8.07 4.54 6.19 8.26 3.46 5.45 6.64
Column
Status Stable Failed Failed Stable Failed Failed Stable Failed Failed
Note: ∆max: Maximum displacement at top of the column.
Yield: 1st yielding displacement of the system
NL Pushover (or Push) is nonlinear static pushover analysis
ARS: Acceleration response spectrum analysis
LTH: Linear time history analysis
NLTH: Nonlinear time history analysis

An interesting observation comes from the ratios of the max NLTH

displacement to the column yielding displacement (ductility demand) and the ratios of
NLTH displacement to capacity displacement from nonlinear static pushover (D/C
ratio of the displacement). It is seen that in the cases analyzed, when the ductility
demand greater than 5 the column status is failed, while the D/C ratio of the
displacement is less than 1, some are even far less than 1. This observation tells that
simply limiting the demand displacement to be less than static capacity displacement
may not be enough to ensure the safety of the structure. Dynamic capacity
displacement can be much smaller than static capacity displacement therefore limiting
the demand ductility to certain number is important when using simplified analysis

Challenges and Advances in Sustainable Transportation Systems

 474 CHALLENGES AND ADVANCES IN SUSTAINABLE TRANSPORTATION SYSTEMS

procedure for design. Caltrans’ SDC (Seismic Design Criteria) sets a limit of 4 for
single column bents supported on fixed foundation. The number came out from this
study (< 5) shows it is an adequate threshold.
Downloaded from ascelibrary.org by New York University on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

Bridge Models. Table 6 through Table 9 below listed the analysis results for Bridge-
1, Bridge-2, Bridge-3 and Bridge-4.

Table 6. Top of Column Displacement of Bridge-1 (T1 = 0.8094 sec)

Ground Motion: Imperial Valley, 1979
T-Dir Displacement with T-Dir Incidence L-Dir Displacement with L-Dir Incidence
Analysis 1.0EQ 1.1EQ 1.2EQ 1.0EQ 1.1EQ 1.2EQ
Method B2 B3 B2 B3 B2 B3 B2 B3 B2 B3 B2 B3
∆max ∆max ∆max ∆max ∆max ∆max ∆max ∆max ∆max ∆max ∆max ∆max
(ft) (ft) (ft) (ft) (ft) (ft) (ft) (ft) (ft) (ft) (ft) (ft)
Yield 0.26 0.28 0.26 0.28 0.26 0.28 0.17 0.17 0.17 0.17 0.17 0.17
NL Pushover 1.24 1.57 1.24 1.57 1.24 1.57 1.21 1.25 1.21 1.25 1.21 1.25
ARS 0.58 0.60 0.64 0.66 0.70 0.71 0.62 0.62 0.68 0.68 0.75 0.75
LTH 0.58 0.59 0.64 0.65 0.69 0.71 0.63 0.63 0.69 0.69 0.75 0.75
NLTH 0.64 0.63 0.69 0.69 0.76 0.77 0.43 0.43 0.45 0.45 0.50 0.50
ARS/LTH 1.00 1.00 1.00 1.00 1.00 1.00 1.01 1.01 1.01 1.01 1.01 1.01
NLTH/LTH 1.11 1.06 1.09 1.05 1.10 1.08 0.68 0.68 0.65 0.65 0.66 0.66
NLTH/Push 0.52 0.40 0.56 0.44 0.61 0.49 0.35 0.34 0.37 0.36 0.41 0.40
NLTH/Yield 2.44 2.25 2.63 2.45 2.89 2.74 2.45 2.45 2.58 2.59 2.86 2.86
ARS/Push 0.47 0.38 0.51 0.42 0.56 0.46 0.51 0.50 0.57 0.55 0.62 0.60
ARS/Yield 2.21 2.13 2.43 2.34 2.65 2.55 3.58 3.59 3.94 3.95 4.30 4.31
Column
Status Stable Stable Stable Stable Stable Stable Stable Stable Stable Stable Stable Stable
Ductility Capacity = Static NL Pushover/Yield = 5.61 (T-Dir) 6.97 (L-Dir)

Table 7. Top of Column Displacement of Bridge-2 (T1 = 0.9638 sec)

Ground Motion: Imperial Valley, 1979
T-Dir Displacement with T-Dir Incidence L-Dir Displacement with L-Dir Incidence
Analysis 1.0EQ 1.25EQ 1.5EQ 1.0EQ 1.25EQ 1.5EQ
Method B2 B3 B2 B3 B2 B3 B2 B3 B2 B3 B2 B3
∆max ∆max ∆max ∆max ∆max ∆max ∆max ∆max ∆max ∆max ∆max ∆max
(ft) (ft) (ft) (ft) (ft) (ft) (ft) (ft) (ft) (ft) (ft) (ft)
Yield 0.28 0.30 0.28 0.30 0.28 0.30 0.16 0.16 0.16 0.16 0.16 0.16
NL Pushover 1.53 1.58 1.53 1.58 1.53 1.58 1.01 1.41 1.01 1.41 1.01 1.41
ARS 0.54 0.55 0.67 0.68 0.77 0.79 0.62 0.62 0.77 0.77 0.90 0.90
LTH 0.53 0.55 0.65 0.68 0.76 0.78 0.62 0.62 0.78 0.78 0.90 0.90
NLTH 0.67 0.63 0.83 0.81 0.99 0.97 0.43 0.43 0.65 0.65 0.79 0.79
ARS/LTH 1.02 0.99 0.98 0.99 0.98 0.99 1.01 1.01 1.00 1.00 1.00 1.00
NLTH/LTH 1.26 1.15 1.27 1.20 1.31 1.24 0.69 0.69 0.84 0.84 0.88 0.88
NLTH/Push 0.44 0.40 0.54 0.51 0.65 0.62 0.42 0.30 0.64 0.46 0.78 0.56
NLTH/Yield 2.43 2.13 3.02 2.74 3.61 3.30 2.65 2.64 4.02 4.00 4.89 4.87
ARS/Push 0.35 0.35 0.44 0.43 0.50 0.50 0.61 0.44 0.76 0.55 0.89 0.64
ARS/Yield 1.96 1.87 2.42 2.31 2.81 2.68 3.84 3.82 4.79 4.76 5.55 5.53
Column
Status Stable Stable Stable Stable Failed Failed Stable Stable Failed Failed Failed Failed
Ductility Capacity = Static NL Pushover/Yield = 5.35 (T-Dir) 6.27 (L-Dir)

Challenges and Advances in Sustainable Transportation Systems

 CHALLENGES AND ADVANCES IN SUSTAINABLE TRANSPORTATION SYSTEMS 475

Table 8. Top of Column Displacement of Bridge-3 (T1 = 1.0406 sec)

Ground Motion: Imperial Valley, 1979
T-Dir Displacement with T-Dir Incidence L-Dir Displacement with L-Dir Incidence
Downloaded from ascelibrary.org by New York University on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

Analysis 1.0EQ 1.25EQ 1.5EQ 1.0EQ 1.25EQ 1.5EQ

Method B2 B3 B2 B3 B2 B3 B2 B3 B2 B3 B2 B3
∆max ∆max ∆max ∆max ∆max ∆max ∆max ∆max ∆max ∆max ∆max ∆max
(ft) (ft) (ft) (ft) (ft) (ft) (ft) (ft) (ft) (ft) (ft) (ft)
Yield 0.26 0.26 0.26 0.26 0.26 0.26 0.17 0.17 0.17 0.17 0.17 0.17
NL Pushover 1.44 1.44 1.44 1.44 1.44 1.44 1.14 1.14 1.14 1.14 1.14 1.14
ARS 0.52 0.47 0.65 0.58 0.75 0.68 0.65 0.65 0.81 0.81 0.94 0.94
LTH 0.51 0.48 0.63 0.59 0.73 0.69 0.65 0.65 0.81 0.81 0.94 0.94
NLTH 0.65 0.66 0.85 0.81 1.02 0.97 0.44 0.44 0.58 0.58 0.69 0.69
ARS/LTH 1.02 1.02 0.98 1.02 0.98 1.02 1.00 1.00 1.00 1.00 1.00 1.00
NLTH/LTH 1.28 1.39 1.34 1.36 1.39 1.41 0.67 0.67 0.71 0.71 0.73 0.73
NLTH/Push 0.45 0.46 0.59 0.56 0.71 0.68 0.38 0.38 0.51 0.51 0.61 0.61
NLTH/Yield 2.46 2.57 3.22 3.13 3.86 3.78 2.55 2.55 3.41 3.41 4.05 4.05
ARS/Push 0.36 0.32 0.45 0.40 0.52 0.47 0.57 0.57 0.72 0.72 0.83 0.83
ARS/Yield 1.96 1.81 2.45 2.26 2.84 2.62 3.82 3.82 4.78 4.78 5.54 5.54

Column Status Stable Stable Stable Stable Failed Failed Stable Stable Failed Failed Failed Failed
Ductility Capacity = Static NL Pushover/Yield = 5.58 (T-Dir) 6.68 (L-Dir)

Table 9. Top of Column Displacement of Bridge-4 (T1 = 2.8043 sec)

Ground Motion: Imperial Valley, 1979
T-Dir Displ with T-Dir Incidence L-Dir Displ with L-Dir Incidence
Analysis 1.0EQ 1.0EQ
Method B2 B3 B4 B5 B6 B2 B3 B4 B5 B6
∆max ∆max ∆max ∆max ∆max ∆max ∆max ∆max ∆max ∆max
(ft) (ft) (ft) (ft) (ft) (ft) (ft) (ft) (ft) (ft)
Yield 2.29 1.33 2.91 2.65 2.13 1.38 1.38 1.49 1.48 1.20
NL Pushover 4.43 4.83 5.33 5.44 5.72 4.07 4.28 4.28 4.61 4.60
ARS 1.97 1.53 1.59 2.09 3.01 1.62 1.62 1.63 1.63 1.63
LTH 1.70 1.65 1.74 2.07 2.95 1.64 1.64 1.64 1.64 1.63
NLTH 1.82 1.78 1.88 2.14 2.65 1.73 1.73 1.73 1.73 1.72
ARS/LTH 1.15 1.08 1.09 0.99 0.98 1.01 1.01 1.01 1.00 1.00
NLTH/LTH 1.07 1.07 1.08 1.03 0.90 1.06 1.06 1.06 1.06 1.05
NLTH/Push 0.41 0.37 0.35 0.39 0.46 0.43 0.40 0.40 0.37 0.37
NLTH/Yield 0.79 1.33 0.65 0.81 1.24 1.26 1.26 1.16 1.17 1.44
ARS/Push 0.44 0.32 0.30 0.38 0.53 0.40 0.38 0.38 0.35 0.35
ARS/Yield 0.86 1.15 0.55 0.79 1.41 1.17 1.18 1.09 1.10 1.36
Column
Status Stable Stable Stable Stable Stable Stable Stable Stable Stable Stable
Substructure of Bridge-4 is column on large diameter CIDH pile shaft. The displacement
listed is the absolute displacement including displacement of CIDH pile shaft.

Figure 7 through Figure 10 showing the linear time history displacement and
nonlinear time history displacement of Bridge-1 and Bridge-2 models subjects to
ground motion of Imperial Valley EQ, 1979.

Challenges and Advances in Sustainable Transportation Systems

 476 CHALLENGES AND ADVANCES IN SUSTAINABLE TRANSPORTATION SYSTEMS
Downloaded from ascelibrary.org by New York University on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

Figure 7. Linear Column Displacement Time History of Bridge-1 (LTH)

Imperial Valley EQ, 1979, Incidence Constant = 1.0

Figure 8. Nonlinear Column Displacement Time History of Bridge-1 (NLTH)

Imperial Valley EQ, 1979, Incidence Constant = 1.0

Challenges and Advances in Sustainable Transportation Systems

 CHALLENGES AND ADVANCES IN SUSTAINABLE TRANSPORTATION SYSTEMS 477
Downloaded from ascelibrary.org by New York University on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

Figure 9. Linear Column Displacement Time History of Bridge-2 (LTH)

Imperial Valley EQ, 1979, Incidence Constant = 1.25

Figure 10. Nonlinear Column Displacement Time History of Bridge-2 (NLTH)

Imperial Valley EQ, 1979, Incidence Constant = 1.25

Challenges and Advances in Sustainable Transportation Systems

 478 CHALLENGES AND ADVANCES IN SUSTAINABLE TRANSPORTATION SYSTEMS

The discussion of the bridge models analysis results are organized and
presented below.
Downloaded from ascelibrary.org by New York University on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

A. Linear Seismic Displacement vs. Nonlinear Time History Displacement

From above tables it can be seen that the ratios of the maximum displacement of ARS
analysis to LTH analysis are all about 1.0 as expected. However unlike in the SDF
system, the ratios of the maximum displacement of LTH analysis to NLTH analysis
are no longer close to 1.0 in most of the analysis cases. By comparing this ratio for
Bridge-1 through Bridge-4 in the tables, it is seen that three factors affect this ratio.
They are similarity to SDF system, structure regularity, and substructure nonlinearity
in seismic response. Bridge-1 is a bridge symmetric in both longitudinal and
transverse directions. Its single column bent acts similar to a cantilever beam in
transverse direction while multi-bent substructure with superstructure forms a frame
in longitudinal direction. Therefore its transverse seismic response is similar to a SDF
system and the ratio of NLTH/LTH is seen to be closed to 1.0 while MDF system
type response in longitudinal direction has this ratio much smaller than 1.0. Both
Bridge-2 and Bridge-3 have torsional dynamic response due to skew and span
unbalance irregularity, the ratios of NLTH/LTH have noticeable difference from 1.0.
Bridge-4 is a most irregular bridge system among all the analyzed models however its
ratios of NLTH/LTH is closed to 1.0 with the difference less than 10% in both
longitudinal and transverse direction. The reason is that Bridge-4 substructure
columns do not experience much nonlinearity in seismic response to the selected
ground motion which can be seen from the ratio ARS/Yield. The ratios of ARS/Yield
are all less than 1.5 in table of Bridge-4.

B. Displacement Ductility Ratios ARS/Yield and NLTH/Yield

As discussed earlier in section for SDF system, the ratio ARS/Yield and NLTH/Yield
are the indication of the structure demand ductility, it is important in predicting the
column nonlinearity in responding to design seismic excitation and the safety. Unlike
in SDF system, the ratio of ARS/Yield and NLTH/Yield may have big discrepancy in
bridge which has obvious 3-D responding effect, for example, in Bridge-2 and
Bridge-3 as shown in Table 7 and Table 8 above. In some cases the ratio NLTH/Yield
is larger than ratio ARS/Yield, and reversed in other cases.
Per SDC design criteria with simplified analysis procedure, both Bridge-2 and
Bridge-3 are considered as safe in transverse direction when responding to ground
motion of Imperial Valley, 1979 with an excitation amplify constant less or equal to
1.5. Because both displacement capacity and ductility capacity is larger than demand,
ie. ARS/Push < 1.0 and ARS/Yield < 4.0. However the NLTH analysis shows that the
columns are failed during the seismic event with amplify constant equal to 1.5. In
longitudinal EQ response the case with excitation amplifier of 1.25 has the same
situation. This observation may alert design engineers that when bridge has high
skewed angle or obvious unbalanced span simplify procedure may not be enough for
seismic analysis, nonlinear time history analysis may need to be carried out.

Challenges and Advances in Sustainable Transportation Systems

 CHALLENGES AND ADVANCES IN SUSTAINABLE TRANSPORTATION SYSTEMS 479

C. Linear Seismic Response vs. Nonlinear Seismic Response

In the analysis of this study the bridge deform pattern differences among linear
seismic analysis, nonlinear time history analysis and static nonlinear pushover
Downloaded from ascelibrary.org by New York University on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

analysis are observed in bridges with noticeable 3-D effect. By comparing the
displacement time histories (or step history in static pushover), the cause of the
discrepancy seem can be explained.
From Figure 7 and Figure 9, it is seen that the linear transverse displacement
time history curves of two column bents have magnitude difference in high peak
region and are almost overlap each other in rest of the low magnitude time zone. This
means the bridge has torsional 3-D movement when it is in strong seismic response
time period, and the response reduces into an in-plan movement in the subsequent
weaker response time zone. In nonlinear time history analysis the opposite
phenomena is seen. The displacement curve of Joint-37 (Bent-2) and Joint-38 (Bent-
3) is obviously separated in lower response magnitude zone from Figure 8 and Figure
10. This implies that the bridge has relatively big torsional movement during
nonlinear seismic response. The reason is that when one of the columns is yielded
first, it further causes unbalance of the stiffness in substructure system and induces
the 3-D or torsional response effect. Obviously this makes the deform pattern
different between linear seismic analysis and nonlinear time history analysis. One
interesting observation is that even the torsional effect in nonlinear seismic response
is large the maximum torsional effect of the linear dynamic analysis is larger than
maximum effect in the nonlinear dynamic analysis. The nonlinearity of the system
tends to reduce the extreme displacement at corner or end columns. This is so because
the corner or end column is soften by the formation of the plastic hinge due to large
movement caused by torsion response and this in turn elongates the vibration period
of the nonlinear portion of the bridge. The plasticity also limits/reduces the inertial
force that transfers through the corner or end columns. The longer structure vibration
period also reduces the bridge’s response in the subsequent dynamic movement while
in the elastic analysis the unbalanced stiffness distribution and value is kept through
the whole response history and ends up to have the extreme displacements at corner
or end columns. This can be seen from the ratio of Bent-2 displacement to Bent-3
displacement (∆2/∆3) for Bridge-3. This ratio from nonlinear time history analysis is
more close to 1.0 than the ratio from linear dynamic analysis.
From the study it can be concluded that the factors cause the discrepancy
between deform pattern of the static nonlinear pushover analysis and linear ARS
dynamic analysis is mainly the 3-D effect and the nonlinearity. In ARS analysis there
is no stiffness redistribution occurs. In nonlinear static pushover analysis when one or
some substructure columns form plastic hinge, the nonlinearity changes the stiffness
distribution but the preset lateral force distribution does not change in entire pushover
process. Therefore when one or some columns form plastic hinge, the same lateral
push force is kept at top of the ‘soften’ columns and push them even further and the
lateral displacement becomes even non-uniformly distributed along the bridge length.
In severe case the displacement pattern of the bridge from ARS dynamic analysis can
be very different from nonlinear static pushover analysis, as showed in Figure 11.
When the bridge is regular all columns form plastic hinge at about the same time and
the small discrepancy can be ignored.

Challenges and Advances in Sustainable Transportation Systems

 480 CHALLENGES AND ADVANCES IN SUSTAINABLE TRANSPORTATION SYSTEMS
Downloaded from ascelibrary.org by New York University on 07/07/15. Copyright ASCE. For personal use only; all rights reserved.

Figure 11. Deform Patterns of ARS Analysis and Nonlinear Static Analysis of
Bridge-4

CONCLUSION

Based on above discussions some conclusions can be drawn as following.

1. Popular simplified bridge seismic analysis procedure in current engineering
practice works very well in SDF system and it is adequate for design of the
OSB category bridge which similar to a SDF system. For these bridges SDC
demand ductility limit of 4 for single column bent and 5 for multi-column bent
is seen a good threshold.
2. The bridge which possesses MDF system characterizes, elastic linear seismic
displacement may not equal or close to its nonlinear dynamic displacement.
3. For bridges in OSB category with high skewed angle or obvious 3-D effect the
simplified seismic analysis procedure may not be adequate, and nonlinear time
history analysis is strongly recommended. In case the nonlinear time history
analysis is not available, reducing the limit of the demand ductility in design
may need to be considered.
4. When bridge deform pattern from ARS analysis is very different from its
nonlinear static pushover, strong 3-D dynamic effect is implied and nonlinear
time history analysis is strongly recommended. In this case design the bridge
to have demand ductility close to code limit may overestimate the structural
dynamic displacement capacity.
5. Bridge substructure nonlinearity tends to reduce the extreme displacement due
to torsional dynamic response at corner or end columns. However it needs
more analysis and studies for bridge on the soft soil site with low shear wave
velocity.

REFERENCES

American Association of State Highway and Transportation Officials (2010).

AASHTO LRFD Bridge Design Specifications
Aviram, A., Mackie, K. R., and Stojadinovic, B. (2008). “ Guidelines for Nonlinear
Analysis of Bridge Structures in California”
Caltrans. (2010). Caltrans Seismic Design Criteria, Version 1.6

Challenges and Advances in Sustainable Transportation Systems