See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/326129563

Seismic Response of Bridge Superstructures Considering Vertical Ground

Accelerations

Conference Paper · June 2018

CITATION READS

1 289

2 authors:

Hartanto Wibowo Sri Sritharan

Iowa State University Iowa State University
13 PUBLICATIONS   80 CITATIONS    219 PUBLICATIONS   2,568 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

UHPC research program View project

Tall Concrete Wind Turbine Tower View project

All content following this page was uploaded by Hartanto Wibowo on 02 July 2018.

The user has requested enhancement of the downloaded file.

 Eleventh U.S. National Conference on Earthquake Engineering
Integrating Science, Engineering & Policy
June 25-29, 2018
Los Angeles, California

SEISMIC RESPONSE OF BRIDGE

SUPERSTRUCTURES CONSIDERING
VERTICAL GROUND ACCELERATIONS
H. Wibowo1 and S. Sritharan2

ABSTRACT

In current design practice, the effects of vertical ground acceleration are not always given
consideration. This is due to the smaller amplitude of the vertical ground acceleration compared
to the horizontal ground accelerations, time lag between the maximum peak horizontal and peak
vertical accelerations, and the notion that the structures are relatively strong in the vertical
direction. However, some recent earthquakes showed that the peak vertical acceleration can be
large and thus significant impact can also be observed on the structural response. While previous
research has focused on substructure, especially on the response of columns, this study emphasizes
on the superstructure responses, which has not received as much attention. This paper presents the
results of an analytical study on the effects of vertical ground acceleration on the seismic response
of medium- and long-span bridges. It is shown that the amplification in superstructure moment
demand when vertical ground acceleration is taken into account can be significant. Based on the
analyses outcomes, the current practice that addresses the effects of vertical ground accelerations
is also evaluated.

1
Post-Doc. Res. Assoc., Dept. of CCEE, Iowa State University, Ames, IA 50011-1066 (hwibowo@iastate.edu)
2
Wilkinson Chair Prof., Dept. of CCEE, Iowa State University, Ames, IA 50011-1066 (sri@iastate.edu)

Wibowo H, Sritharan S. Seismic response of bridge superstructures considering vertical ground accelerations.
Proceedings of the 11th National Conference in Earthquake Engineering, Earthquake Engineering Research Institute,
Los Angeles, CA. 2018.
 Eleventh U.S. National Conference on Earthquake Engineering
Integrating Science, Engineering & Policy
June 25-29, 2018
Los Angeles, California

Seismic Response of Bridge Superstructures Considering Vertical

Ground Accelerations

H. Wibowo1 and S. Sritharan2

ABSTRACT

In current design practice, the effects of vertical ground acceleration are not always given
consideration. This is due to the smaller amplitude of the vertical ground acceleration compared to
the horizontal ground accelerations, time lag between the maximum peak horizontal and peak
vertical accelerations, and the notion that the structures are relatively strong in the vertical direction.
However, some recent earthquakes showed that the peak vertical acceleration can be large and thus
significant impact can also be observed on the structural response. While previous research has
focused on substructure, especially on the response of columns, this study emphasizes on the
superstructure responses, which has not received as much attention. This paper presents the results
of an analytical study on the effects of vertical ground acceleration on the seismic response of
medium- and long-span bridges. It is shown that the amplification in superstructure moment demand
when vertical ground acceleration is taken into account can be significant. Based on the analyses
outcomes, the current practice that addresses the effects of vertical ground accelerations is also
evaluated.

Introduction

Earthquake ground accelerations are typically recorded in three perpendicular directions: two
horizontals and one vertical. However, in design practice, the vertical ground motion is commonly
ignored due to three main reasons: (1) the amplitude of vertical ground acceleration is normally
smaller than its horizontal counterparts, (2) the peak vertical acceleration generally does not occur
at the same time as the peak horizontal accelerations, and (3) the acceptable notion that the
structures are strong in the vertical direction. The first two reasons are fair, but not always true for
ground motions from every earthquake. It has been shown in some past major earthquakes such as
the 1994 Northridge and 2011 Christchurch Earthquakes, among others, that these were not the
case. It is also believed that near-fault earthquake events generates larger vertical acceleration
compared to far-field earthquake events. When the effects of vertical ground acceleration are large,
1
Post-Doc. Res. Assoc., Dept. of CCEE, Iowa State University, Ames, IA 50011-1066 (hwibowo@iastate.edu)
2
Wilkinson Chair Professor, Dept. of CCEE, Iowa State University, Ames, IA 50011-1066 (sri@iastate.edu)

Wibowo H, Sritharan S. Seismic response of bridge superstructures considering vertical ground accelerations.
Proceedings of the 11th National Conference in Earthquake Engineering, Earthquake Engineering Research Institute,
Los Angeles, CA. 2018.
it can increase the demands in the structure that may potentially lead to subsequent failure of the
components and/or collapse of the system. Thus far, the research on vertical ground accelerations
were limited to its effects on bridge substructures. This study instead is focused on its effects on
bridge superstructures with an emphasis on bridges with drop pier caps, which can largely be found
across the nation.

Literature Review

As previously mentioned, many studies on vertical ground accelerations focused on the column or
substructure response. This was mostly triggered by some damage observations on buildings and
bridges in past earthquakes including the 1971 San Fernando, 1994 Northridge, 1995 Kobe, 2008
Wenchuan, and 2011 Christchurch Earthquakes. But it is generally very difficult to attribute
damage directly to vertical ground motions since the earthquake ground motions are coupled in
multiple directions. Analyses on several regular bridges by Saadeghvaziri and Foutch [1] showed
that vertical ground accelerations increased the force and moment demands as well as the extent
of damage in the columns and abutments. The vertical ground accelerations were deemed to cause
fluctuations of the axial force in the columns or piers that in turn reduced the shear capacity [2],
especially in the plastic hinge regions, which had also been shown experimentally [3, 4, 5]. It was
also shown that as the peak vertical-to-horizontal acceleration ratio increased, the variation in the
axial and moment demands increased [6]. A study by Gulerce et al. [7] showed that the vertical
ground motion has an effect on the axial force demand and moment demands at the face of the pier
cap and middle of the span. Another study by Chen et al. [8] showed that vertical ground
accelerations generate higher axial demand in the columns and larger moment in the deck as well
as more frequent footing uplift. For superstructure demand, a study on typical two- and four-span
bridges in California by Kunnath et al. [9] and Sheng and Kunnath [10] concluded that the vertical
ground acceleration can increase the axial force demand in bridge columns that translates to a
higher moment demand in the girder at the face of the pier cap. Further, it can also increase the
superstructure moment demand at midspan.

In seismic design of bridges, AASHTO does not provide a direct requirement as to when
and how to include vertical ground acceleration; it only states that all components of the ground
motion need to be included in the nonlinear response history analysis [11, 12]. To account for the
effect of vertical acceleration in design of ordinary standard bridges where the site PGA is 0.6 g
or greater, Caltrans [13] suggests that a uniformly distributed vertical force of 25% of the dead
load be applied on the superstructure in both downward and upward directions independently. This
is to account for the axial force variation and increase in moment caused by the vertical ground
acceleration. However, there is no rational justification for using this 25% factor and whether this
will lead to an adequately safe design. Vertical ground motion spectra are also not provided in the
USGS National Seismic Hazard Mapping program. To develop a vertical design response
spectrum, Marsh et al. [14] suggested that the magnitude and distance data from deaggregation of
hazard curve and an appropriate ground motion prediction equation (GMPE) can be used to obtain
a period-dependent V/H ground motion spectrum ratio, which can then be used as a multiplier for
the horizontal design spectrum. When such GMPE does not exist, the horizontal design response
spectrum can be multiplied by a factor of 2/3 to develop the vertical design response spectrum.
 Finite Element Models and Ground Motions

In this study, three-span straight, skewed, and curved bridges with medium- and long- spans were
analyzed. The medium-span bridges have the span lengths of 105 ft, 105 ft, and 105 ft whereas the
long-span bridges have the span lengths of 105 ft, 152.5 ft, and 105 ft. The basic properties of
these prototype bridges are given in Table 1. The bridges consisted of reinforced concrete
substructure and steel girder superstructure with reinforced concrete deck and V-shape cross-
frames shown in Figure 1. Finite element models of the bridges were developed using CSi Bridge
2017 [15]. The bridge substructure was modeled as frame elements with fiber plastic hinge model
located at the top and bottom portions of the pier. The bridge girders were also modeled as frame
elements whereas the deck was modeled as shell elements. The boundary conditions at the
abutments are simply supported with shear keys in the transverse (or radial for curved bridges)
direction, while at the bottom of the columns are fixed. This way, the bridges were free in the
longitudinal (or tangential) direction and restrained in the transverse (or radial) direction. The
superstructure rested on pin supports on top of the pier cap. An example of the developed
numerical models is depicted in Figure 2. The bridge model has been verified with experimental
results in different studies [16, 17]. The nonlinear response history analyses were carried out
assuming a 5% equivalent viscous damping for the structure.

Table 1. Properties of the Bridges

Parameter Long-Span Medium-Span

Total Length 362.50 ft 315 ft
Span Lengths 105 ft – 152.50 ft – 105 ft 105 ft – 105 ft – 105 ft
Skew Angle
45°
(for skewed bridges)
Subtended Angle
104° (1.8 rad)
(for curved bridges)
Radius at Centerline
200 ft
(for curved bridges)
Plate Girders 3 @ 11 ft
Deck Width 30 ft
Deck Thickness 8⅛ in
Pier Type Single circular column with drop cap
Pier Cap Length 28 ft
Pier Cap Width 6 ft
Pier Cap Depth Tapered, 5 ft at column face, 3.33 ft at the ends
Column Height 20 ft
Column Diameter 5 ft
Column Longitudinal
20 #11
Reinforcement
Column Transverse
#8 @8 in spiral
Reinforcement
Rebar Yield Strength 60 ksi ASTM A706 rebars
Concrete Yield
4 ksi for deck, 5.5 ksi for others
Strength
 Figure 1. Typical Bridge Superstructure Cross-Section

Twelve sets of earthquake ground motions were selected for this study: three non-pulse-
like sets and nine pulse-like sets, representing both free-field and near-fault motions, respectively.
The bridges were subjected to both longitudinal and transverse ground motion components along
with the vertical ground motion component. These ground motion records were scaled for a Site
Class B in an AASHTO Seismic Zone 3 area with a PGA of 0.47 g, Ss of 1.14 g, and S1 of 0.41 g.
The scaling method was based on equal spectral accelerations of the longitudinal ground
acceleration at the period of 1.0 s, which is also within the range of the fundamental periods of the
prototype bridges. The scaling factors as well as some basic characteristics of the earthquake
records used in this study are presented in Table 2.

Table 2. Ground Motion Records

Three-
Magnitude, Record Scaling
Letter Earthquake
Mw Station Factor
Name
ELC 1940 El Centro 6.9 El Centro 0.50
MAN 1972 Managua 6.2 Station 1 1.26
GIL 1989 Loma Prieta 6.9 Gilroy #1 1.31
PAC 1971 San Fernando 6.6 Pacoima Dam 0.51
IMP 1976 Imperial Valley 6.5 El Centro #6 0.94
PET 1992 Cape Mendocino 7.0 Petrolia 0.42
LUC 1992 Landers 7.3 Lucerne 1.36
SYL 1994 Northridge 6.7 Sylmar 0.48
TAK 1995 Kobe 6.9 Takatori 0.36
IZM 1999 Kocaeli 7.5 Izmit 1.54
CHI 1999 Chi-Chi 7.6 TCU068 0.51
CHR 2011 Christchurch 6.2 D07C 0.39
 Figure 2. Example of the Developed Finite Element Model

Analysis Results

The flexural response of the bridges is highly influenced by the amplitude of the vertical ground
acceleration. The maxima and minima are obtained from the envelope response along the length
of the bridge; an example of this process is illustrated in Figure 3. The maximum differences of
the maximum positive and negative moments with and without the vertical ground acceleration
were found to be over 300% for the case of straight medium-span bridge under IMP ground motion
record. The differences of the maximum positive moments for all cases are presented in Figure 4.
The average increase of the maximum positive moment is approximately 38% with a 75th
percentile value of approximately 51%. Similar observations were found for the maximum
negative moments although not presented herein. The vertical displacement at midspan was also
affected by the vertical ground acceleration, as illustrated in Figure 5. This vertical displacement
in turn caused the fluctuation of axial forces in the bridge piers.

100,000

50,000
Moment (kip.in)

-50,000

-100,000

-150,000
0 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000
Location (in)

W/o Vert. Acc. - Maxima W/o Vert. Acc. - Minima

W/ Vert. Acc. - Maxima W/ Vert. Acc. - Minima
Figure 3. Superstructure Moment Envelope for Case of Straight Medium-Span Bridge Subjected
to CHR Record
 350
300
250
Difference (%)

200
150
100
50
0
-50
GIL

PET

GIL

PET
PAC

CHR

PAC

CHR
IMP

IMP
CHI

CHI
MAN

TAK

MAN

TAK
SYL

SYL
IZM

IZM
ELC

LUC

ELC

LUC
Long-Span Medium-Span

Straight Skewed Curved Average 75-th Percentile

Figure 4. Differences of the Maximum Positive Moments for All Cases

0.4
Displacement (in)

0.2

-0.2

-0.4

-0.6
25 30 35 40 45 50
Time (s)

With Vert. Acc. W/o Vert. Acc.

Figure 5. Vertical Displacement at Midspan for Straight Medium-Span Bridge Subjected to CHR
Record

The maximum moment ratio was taken as the maximum moment with vertical ground
acceleration divided by the maximum moment without vertical ground acceleration. Interestingly,
the maximum moment ratio was found to be well correlated with the vertical spectral acceleration.
The vertical spectral acceleration was obtained from the vertical response spectrum of a particular
earthquake for a given period corresponding to the vertical mode of the bridge. An example of
such correlation is presented in Figure 6. Similar correlations were found for the vertical
displacement at midspan. These correlations, however, were found to be weaker for more irregular
bridges. This is because the modes of vibration were more complex and the vertical mode was
coupled with other modes. It can also be seen that pulse-like records generally caused higher
demand on the structures. However, non-pulse-like records may also have high amplitude vertical
ground acceleration, as shown in Figure 6 to be the second most demanding record for this
particular bridge.

6
Max. Positive Moment Ratio

y = 0.7672x + 1
5
R² = 0.9694
4
3
2
1
0
0.0 1.0 2.0 3.0 4.0 5.0
Sa,vertical (g)

Non-Pulse Pulse
Figure 6. Maximum Positive Moment Ratio vs. Vertical Spectral Acceleration for Straight
Medium-Span Bridge

Conclusions

From the analysis results presented in this paper, it can be concluded that vertical ground
accelerations can have significant effects on bridge superstructure demands such as displacement
and moments. Thus, ignoring vertical ground acceleration, especially when the amplitude is large,
will lead to a non-conservative design. Generally, near-fault earthquakes would generate higher
vertical ground accelerations due to its pulse-like motion but far field earthquakes can also be as
demanding. For the cases presented herein, the average increase in the positive moments when
vertical ground acceleration was considered was found to be approximately 38% with the 75th
percentile value of approximately 51%. Therefore, the current design practice to uniformly apply
vertical force of 25% of the dead load on the superstructure may not be conservative and needs to
be revisited.

References

1. Saadeghvaziri MA, Foutch DA. Dynamic behaviour of R/C highway bridges under the combined effect of vertical
and horizontal earthquake motions. Earthquake Engineering and Structural Dynamics 1991; 20 (6): 535-549.
2. Papazoglou AJ, Elnashai AS. Analytical and field evidence of the damaging effect of vertical earthquake ground
motion. Earthquake Engineering and Structural Dynamics 1996; 25 (10): 1109-1137
3. Kim SJ, Holub CJ, Elnashai AS. Experimental investigation of the behavior of RC bridge piers subjected to
 horizontal and vertical earthquake motion. Engineering Structures 2011, 33 (7): 2221-2235.
4. Matsuzaki H, Kumagai Y, Kawashima K. Effect of strong vertical excitation on the seismic performance of RC
bridge columns. Proceedings of 15th World Conference on Earthquake Engineering, Lisbon, Portugal. 2012.
5. Lee H, Mosalam KM. Seismic evaluation of the shear behavior in reinforced concrete bridge columns including
effect of vertical accelerations. Earthquake Engineering and Structural Dynamics 2014, 43 (3): 317-337.
6. Kim SJ, Holub CJ, Elnashai AS. Analytical assessment of the effect of vertical earthquake motion on RC bridge
piers. Journal of Structural Engineering 2011; 137 (2): pp. 252-260.
7. Gulerce Z, Erduran E, Kunnath SK, Abrahamson NA. Seismic demand models for probabilistic risk analysis of
near fault vertical ground motion effects on ordinary highway bridges. Earthquake Engineering and Structural
Dynamics 2012, 41 (2): 159-175.
8. Chen Y, Kun C, Larkin T, Chouw N. Impact of vertical ground excitation on a bridge with footing uplift. Journal
of Earthquake Engineering 2016, 20 (7): 1035-1053.
9. Kunnath SK, Erduran E, Chai YH, Yashinsky M. Effect of near-fault ground motions on seismic response of
highway overcrossings. Journal of Bridge Engineering 2008, 13 (3): 282-290.
10. Sheng L-H, Kunnath S. Effect of vertical acceleration on highway bridges. Proceedings of the Fourth US-Taiwan
Bridge Engineering Workshop, Princeton, NJ. 2008.
11. American Association of State Highway and Transportation Officials, AASHTO. AASHTO Guide Specifications
for LRFD Seismic Bridge Design, 2nd Ed. AASHTO, Washington, DC, 2011.
12. American Association of State Highway and Transportation Officials, AASHTO. AASHTO LRFD Bridge Design
Specifications, Customary U.S. Units, 7th Ed. AASHTO: Washington, DC, 2014.
13. California Department of Transportation, Caltrans. Seismic Design Criteria Version 1.7. Caltrans: Sacramento,
2013.
14. Marsh ML, Buckle IG, Kavazanjian Jr. E. LRFD Seismic Analysis and Design of Bridges Reference Manual,
NHI Course No. 130093 and 130093A, Report No. FHWA-NHI-15-004. FHWA: Washington, DC, 2014.
15. Computers and Structures, Inc., CSI. CSi Bridge 2017 v.19.1.0. 2017. (Computer Software)
16. Levi MJ, Sanders DH, Buckle IG. Seismic response of columns in horizontally curved bridges, Report No.
CCEER-11-04. Center for Civil Engineering Earthquake Research, Department of Civil and Environmental
Engineering, University of Nevada, Reno: Reno, 2011.
17. Wibowo H, Buckle IG, Sanders DH. Experimental and analytical investigations on the effects of live load on the
seismic performance of a highway bridge, Report No. CCEER-13-15. Center for Civil Engineering Earthquake
Research, Department of Civil and Environmental Engineering, University of Nevada, Reno: Reno, 2013.

View publication stats