Professional Documents
Culture Documents
Validation of Seismic Soil Foundation Structure Interaction
Validation of Seismic Soil Foundation Structure Interaction
A Project
MASTER OF SCIENCE
in
Civil Engineering
by
SPRING
2013
© 2013
ii
VALIDATION OF SEISMIC
A Project
by
Approved by:
____________________________
Date
iii
Student: Thomas Albert Mar
I certify that this student has met the requirements for format contained in the University format
manual, and that this project is suitable for shelving in the Library and credit is to be awarded for
the project.
iv
Abstract
of
VALIDATION OF SEISMIC
by
shell elements for the bridge deck. The column and piles were modeled using frame
elements. Abutment-backfill and ground soil were simulated using nonlinear springs.
earthquake motions. The resulting dynamic response of the bridge model was found to
be in close agreement with motions recorded at various locations on the bridge. This
validates the practical application and methodology of this project and may be used for
____________________________
Date
v
ACKNOWLEDGEMENTS
I would like to express my upmost gratitude to Dr. Anoosh Shamsabadi for the
many hours he spent sharing his expertise and advising me on this project. I also feel
very grateful toward my committee members: Professor Benjamin Fell and Professor
Matthew Salveson, whose instruction and insight has prepared me for a future in
engineering. And last but not least, I would like to thank my friends and family for their
vi
TABLE OF CONTENTS
Page
Acknowledgements ........................................................................................................ vi
Section
1. INTRODUCTION ..................................................................................................... 1
2. BRIDGE DESCRIPTION........................................................................................... 2
Appendix ...................................................................................................................... 32
References ..................................................................................................................... 39
vii
LIST OF TABLES
Tables Page
viii
LIST OF FIGURES
Figures Page
17. Lateral force-displacement curves at depths x = 0’, 4.7’, 9.4’, 14.1’, 18.8’,
23.5’ ......................................................................................................... …… 20
ix
18. Lateral force-displacement curves at depths x = 28.2’, 32.9’, 37.6’, 42.3’,
19. Vertical force-settlement curves at depths x = 0’, 4.7’, 9.4’, 14.1’, 18.8’,
x
36. 1979 Imperial Valley earthquake motions in longitudinal, transverse,
xi
1
1. Introduction
1994 Northridge, California, the 1995 Kobe, Japan, and the 1999 Chi-Chi, Taiwan
earthquakes, have increased the importance of nonlinear seismic analyses employing soil-
from these records that SFSI plays a significant role in the global response of highway
history analysis employing SFSI on a typical California highway bridge structure. The
analysis will use a direct approach in which nonlinear soil foundation properties are
explicitly included in a global finite element model to account for both the geotechnical
and structural responses during a seismic event. Many bridges are supported on deep pile
foundations that penetrate multiple soil layers with varying stiffness and strength
along the entire pile length. The effects of depth-varying foundation properties will be
rigorously addressed by including each individual pile (with distributed soil springs along
their lengths) in the global bridge model. Recorded earthquake motions will be applied
to the model and the resulting displacement response will be compared to real earthquake
2. Bridge Description
Meloland Road Overcrossing (MRO) was chosen for this investigation due to its
relatively simple design and heavy seismic instrumentation. MRO was constructed in
from the Imperial Valley fault rupture. The structure is a two-span prestressed reinforced
concrete box-girder bridge supported on single column bent and integral abutments, as
shown in Figure 1.
The As Built bridge plans indicate the bridge deck to be 208 ft long and 34 ft
wide with each span measuring 104 ft, as shown in Figure 2. The depth of the deck is 5.5
timber piles with a square concrete cap, as shown in Figure 3 and Figure 4. The
3
monolithic abutment backwalls have a height of the approximately 13 ft. Each abutment
Recorded time-histories from these instruments are available for several earthquakes,
accelerometers are located at a free-field site 30 ft from the bridge column. The location
of the instruments that measure free-field motions (Channels 14, 15 and 24) and structure
from the 2010 Baja California earthquake were used as the input motions for this project.
Each record lasts approximately 88 seconds. The strongest component of the ground
motion is in the longitudinal (N-S) direction. The acceleration time histories are shown
in Figure 7. Peak ground accelerations, velocities, and displacements of the three input
8 13 7 27 3
26.5
18
30 1
4
2
29 22 21 18 31 28
32
34
25 23 20 16 10
30
12
0.1
0.05
0
-0.05
-0.1
-0.15
-0.2
-0.25
0 10 20 30 40 50 60 70 80 90
Time (sec)
0.05
0
-0.05
-0.1
-0.15
-0.2
0 10 20 30 40 50 60 70 80 90
Time (sec)
0.05
-0.05
-0.1
-0.15
0 10 20 30 40 50 60 70 80 90
Time (sec)
A Log of Test Boring sheet provided with the As Built bridge plans shows the
type and strength description of each layer of soil. Idealized soil profiles were developed
for this investigation to simplify calculations. The idealized soil profile at the bent is
shown in Figure 8.
The piles penetrate a combination of sandy and clay soil layers. The water table
is located roughly 9 ft below the ground surface. The soil parameters relevant to this
project, including effective unit weight, friction angle, and cohesion, are summarized in
Table 2.
10
Soil Properties
Layer Soil Type
γ' (pcf) ϕ (deg) c (psf)
1 Sand 120 32 --
2 Sand 57.6 32 --
3 Clay 55.6 -- 1500
4 Sand 57.6 32 --
5 Clay 55.6 -- 1500
Notes: γ' = Effective Unit Weight, ϕ = Friction Angle, c = Cohesion
The pile foundations of the abutments penetrate to approximately the same depth
as those of the bent. Therefore, the soil at the abutments features a very similar profile as
shown in Figure 9.
The approach for modeling MRO was to split the complete SFSI system into two
groups connected by a foundation interface. The first group included the superstructure
and column. The second group included the pile foundations and surrounding soil. The
abutment backwalls and pile cap acted as an interface for the two groups. Figure 10
Structure
Interface
Soil Springs
Founda on & Soil
In this direct approach for modeling, all structural components, foundation components
and soil support springs are explicitly included in the bridge model.
Midas Civil was used to build the finite element model of MRO. Figure 11 shows
12
the complete global model with pile foundations. The bridge deck, abutment backwalls,
and pile cap were modeled using shell elements with applicable structural properties.
Frame elements were used for the column and piles. In addition, the column elements
were modified for cracked sectional properties. The typical material properties used for
The cracked moment of inertia for the column was determined from moment-
curvature analysis. This analysis was performed using the software’s section designer.
By indicating the column’s material properties, geometry, and loading, the idealized
Material Properties
Material Bridge Components Unit Weight Modulus of Elasticty Yeild Strength
(pcf) (ksi) (ksi)
Deck, Backwalls,
Concrete 150 3645 5
Column, Pile Cap
Timber Piles 50 1500 1.25
13
(Φy , M p)
120000
100000
Moment (kip-in)
80000
Actual
60000
Idealized
40000
20000
0
0 0.0001 0.0002 0.0003 0.0004 0.0005 0.0006
Curvature (rad/in)
curvature, Φy , was calculated to be 0.00009 rad/in. The cracked moment of inertia, Icr,
The cracked moment of inertia was roughly half of the gross moment of inertia.
Therefore, a section modifier of 0.5 was used for the moment of inertia in each direction
of the column.
14
The interaction between the bridge deck, abutment wall and embankment soil has
a hyperbolic relationship that has been experimentally observed and verified with finite
element analysis. However, including abutment backfills in a global bridge model would
Stewart, and Taciroglu (2010) proposed the use of a simplified relationship between the
lateral load per unit with of abutment backwall, F, and the lateral displacement, y. The
( )
where Fult is the maximum abutment force per unit width of backwall developed at
displacement ymax . K50 is the average abutment stiffness at half the maximum abutment
force.
the abutment wall was modeled using a nonlinear, inelastic spring. The spring represents
interface. To develop the backbone curve of the spring, the following simplified HFD
H is the height of the backwall in feet and y is the lateral displacement in inches.
Given an abutment height of 13 ft and width of 34 ft, the backbone curve for the
Soil-Abutment Interaction
4500
4000
3500
Lateral Force (kip)
3000
2500
2000
1500
1000
500
0
0 1 2 3 4 5
Displacement (in)
interaction was divided into 16 smaller springs in parallel and uniformly applied across
the abutment face as shown in Figure 14. Each spring was also connected to a
completely fixed ground node. It should be noted that these springs only work in
For skewed bridges a uniformly distributed pattern of springs may not accurately
abutments are expected to “purely push” into the embankment soil during an earthquake,
The interaction between pile foundations and their surrounding soil has three-
dimensional complexity. Nonlinear, inelastic springs along the length of the piles were
used to represent this interaction. Figure 15 shows how support springs are used for
corresponding pile reactions. Three soil reactions were approximated for this project
with the use of support springs: lateral, axial and tip resistance. Due to the geometry of
the bridge structure, significant twisting of the column and foundations was not expected.
Figure 15: Nonlinear springs represent interaction between soil and pi les (S hamsabadi et al. 2010)
The following three sections discuss the development of the nonlinear springs that model
Using the computer program Ensoft LPile, lateral pile-soil support curves (called
p-y curves) were generated using the parameters given in Table 2 and depths shown in
Figure 8. LPile utilizes the following equation to develop a backbone curve that
represents the relationship between lateral soil resistance and lateral deflection:
( )
{( )
For the bent piles, p-y curves were calculated at 11 depths at equal increments of
4.7 ft from pile top to the pile tip at 47 ft, shown in Figure 16.
P-y Curves
14000
Soil Resistance, p (lbs/in)
12000
10000
8000
6000
4000
2000
0
0 0.5 1 1.5 2 2.5
Lateral Deformation, y (in)
0 ft 4.7 ft 28.2 ft 32.9 ft 37.6 ft 9.4 ft, 14.1 ft, 18.8 ft, 23.5 ft, 42.3 ft, 47 ft
Soil resistance of sand is dependent on the overburden pressure due to the weight
of overlying material. Hence, the p-y curves corresponding to nodal depths in sand layers
increase with depth. Soil resistance of clay, on the other hand, is more heavily
dependent on its cohesions. The p-y curves corresponding to nodes in clay layers are
therefore the same because they are calculated using the same cohesion factor for this
In order to implement the p-y curves as soil springs for the finite element model,
done by simply multiplying the resistance values by the tributary pile length of each
node. The resulting curve is then mirrored about its axes so that the spring will behave
the same in tension and compression. The resulting force-displacement curves are shown
150
100
Lateral Force, F (kip)
50
0
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
-50
-100
-150
-200
Lateral Deflection, Y (in)
0 ft 4.7 ft 9.4 ft, 14.1 ft, 18.8 ft, 23.5 ft
Figure 17: Lateral force-displacement curves at depths x = 0’, 4.7’, 9.4’, 14.1’, 18.8’, 23.5’
600
400
Lateral Force, F (kip)
200
0
-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5
-200
-400
-600
-800
Lateral Deflection, Y (in)
28.2 ft 32.9 ft 37.6 ft 42.3 ft 47 ft
Figure 18: Lateral force-displacement curves at depths x = 28.2’, 32.9’, 37.6’, 42.3’, 47’
21
Axial skin resistance support curves (called t-z curves) were developed using
hyperbolic functions similar to the HFD model. The following function fits the trend
Here, S is the pile settlement ratio and Q is the side load transfer ratio where
( )
The resulting t-z curves using the properties of Table 2 were multiplied by the tributary
surface area of each corresponding node to produce the force-settlement curves shown in
Figure 19 and Figure 20. Theses curves were used for the vertical soil springs in the
5
0
-0.2 -0.15 -0.1 -0.05 -5 0 0.05 0.1 0.15 0.2
-10
-15
-20
-25
Settlment, Z (in)
0 ft 4.7 ft 9.4 ft, 14.1 ft, 18.8 ft, 23.5 ft
Figure 19: Vertical force-settlement curves at depths x = 0’, 4.7’, 9.4’, 14.1’, 18.8’, 23.5’
30
20
10
Side Force, F (kip)
0
-0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2
-10
-20
-30
-40
Settlement, Z (in)
Figure 20: Vertical force-settlement curves at depths x = 28.2’, 32.9’, 37.6’, 42.3’, 47’
23
Finally, a tip resistance curve (called a q-w curve) for the piles was developed.
The q-w curve was also developed using a hyperbolic function developed from empirical
data. Since the pile tip is embedded in a clay soil layer, the following equation was
S is now the Base Settlement Ratio and Q is the End Bearing ratio where
provided the pile has been driven at least 5 times the diameter of the pile into the strata.
Therefore, the Ultimate end bearing for the piles, given the cohesion in Table 2, is 13500
psf. The End Bearing values are multiplied by the area of the tip resulting in the tip
20
10
Tip Force (kip)
0
-1 -0.5 0 0.5 1
-10
-20
-30
Settlement (in)
Since each pile penetrates to approximately the same depth and same soil, one q-
w curve was adequate for each pile. Furthermore, all of the support springs developed
for the bent piles were also used for the abutment piles at the appropriate depths.
Although the bent piles have a slightly different soil profile, this simplification was
The mode shape shown in Figure 22 displays the primary vertical mode. This
mode has a period of 0.264 seconds. Subsequent mode shapes are shown in Figures 23 to
27, with Figure 23 showing the primary transverse mode and Figure 27 showing the
The earthquake motions were applied to the completed bridge model. Figures 28
instrument-recorded responses for major points on the bridge. The results of the model
match the records very well at each location. The small discrepancies are most likely due
to the idealization of the model, but are insignificant to the global response. Further
investigation into the behavior of the bridge model with design basis earthquake and
28
maximum credible earthquake level input motions can be found in the appendix section.
Channel 13 - Transverse
5
4
3
2
Displacement (in)
1
0
-10 10 30 50 70 90
-1
-2
-3
Recorded CH 13
-4
MIDAS Model
-5
Time (sec)
Channel 19 - Vertical
2
1.5
1
Displacement (in)
0.5
0
-10 10 30 50 70 90
-0.5
-1
Recorded CH 19
-1.5
MIDAS Model
-2
Time (sec)
Channel 27 - Longitudinal
8
4
Displacement (in)
0
-10 10 30 50 70 90
-2
-4
Recorded CH 27
-6
MIDAS Model
-8
Time (sec)
Channel 1 - Vertical
2
1.5
1
Displacement (in)
0.5
0
-10 10 30 50 70 90
-0.5
-1
Recorded CH 1
-1.5
MIDAS Model
-2
Time (sec)
Channel 2 - Transverse
5
4
3
2
Displacement (in)
1
0
-10 10 30 50 70 90
-1
-2
-3
Recorded CH 2
-4
MIDAS Model
-5
Time (sec)
Channel 17 - Vertical
2
1.5
1
Displacement (in)
0.5
0
-10 10 30 50 70 90
-0.5
-1
Recorded CH 17
-1.5
MIDAS Model
-2
Time (sec)
the finite element model given an actual earthquake excitation was in very close
agreement with recorded displacements. This remarkable match achieved between the
model and recorded motions is due to the additional steps taken to realistically estimate
passive soil capacities of the abutments and foundation. The results suggest that
methodology and procedures performed in this project may be applicable to any ordinary
While detailed structural bridge models consisting of complete SFSI systems can
be used in high profile projects where plenty of resources are available, this kind of
bridge model becomes economically impractical for many ordinary bridges where
and soil capacities are kinematically condensed into a single support, can reduce
Appendix
The MRO finite element model was exercised with stronger ground motions to
further investigate its behavior. Two additional three-component sets of ground motions
were implemented into the model including a design basis earthquake (DBE) level set
and a maximum credible earthquake (MCE) level set. The acceleration spectra assuming
2
Design Spectrum
1.5
Sa (g)
MCE Spectrum
1
0.5
0
0 0.5 1 1.5 2 2.5
T (sec)
The design and MCE spectra were calculated based on the bridge’s location at 32.773°
latitude and -115.448° longitude. The spectra were also calibrated for a Site Class D soil.
the location’s average hazard. Given the bridge’s location, spectral acceleration period of
0.26 sec, and an average shear wave velocity of 630 ft/sec in the first 100 ft of soil, the
33
mean magnitude, M, and mean site-to-source, R, for a DBE and MCE are shown in Table
4.
Ground motions were selected from the PEER Ground Motion Database that
coincidentally, the motions chosen were from MRO during the 1979 Imperial Valley
Figure 35. The three-component recorded motions are shown in Figure 36.
0.1
Sa (g)
0.01
0.001
0.01 0.1 1 10
T (sec)
0.1
Acceleration (g)
0
0 5 10 15 20 25 30 35 40
-0.1
-0.2
-0.3
-0.4
Time (sec)
0.1
0
0 5 10 15 20 25 30 35 40
-0.1
-0.2
-0.3
-0.4
Time (sec)
0.1
0
0 5 10 15 20 25 30 35 40
-0.1
-0.2
-0.3
Time (sec)
Figure 36: 1979 Imperial Valley earthquake motions in longitudinal, transverse, and vertical directions
35
The maximum spectral acceleration for a period of 0.26 sec is 0.56g. Therefore,
the three-component motions were scaled up to reach DBE and MCE levels. Scale
factors of 2.4 to reach a Sa of 1.35g and 3.6 to reach a Sa of 2.03g were used for DBE
1979 Imperial
DBE MCE
Valley
Sa (g) 0.56 1.35 2.03
SF -- 2.4 3.6
The scaled recorded motions were implemented into the model. The
displacement response at the abutments for the DBE is shown in Figure 37 and for the
20
Displacement (in)
10
0
0 5 10 15 20 25 30 35 40
-10
CH 27
-20
CH 19
-30
CH 13
-40
Time (sec)
10
0
0 5 10 15 20 25 30 35 40 45
-10
-20 CH 13
-30
CH 19
-40
CH 27
-50
Time (sec)
From Figures 37 and 38, it appears the global system does not undergo significant
inelastic deformations. That is not to say there are none, but they are very minor as
120
100
80
60
40
20
0
0.9 0.95 1 1.05 1.1 1.15
Displacement (in)
1
Displacement (in)
0.5
0
0 5 10 15 20 25 30 35 40
-0.5
CH 13
-1 CH 19
-1.5 CH 27
Time(sec)
It is clear that there is very little relative displacement of the bridge during the
characteristic of many bridges with integral abutments. With this configuration, the
bridge is essentially fixed to the surrounding soil. Therefore, during a seismic event, the
bridge will move along with the surrounding soil. This is further illustrated by plotting
the relative displacement at the top and bottom of the column during the MCE.
38
1
Displacment (in)
0.5
0
0 5 10 15 20 25 30 35 40
-0.5
Top
-1
Bottom
-1.5
Time (sec)
It can be seen in Figure 41 that the bottom of the column basically moves with the
ground. Although the top of the column does have noticeable relative displacement, it is
References
ASCE (2005). Minimum Design Loads for Buildings and Other Structures, Standards
http://www.conservation.ca.gov/cgs/smip.
Ensoft (2013). LPile: A Program for the Analysis of Piles and Drilled Shafts Under
MIDASoft (2013). Midas Civil: Integrated Solution System for Bridge and Civil
PEER (2010). PEER Ground Motion Database (Beta). Pacific Earthquake Engineering
(http://peer.berkeley.edu/peer_ground_motion_database)
Reese, L., Isenhower, W., Wang, S.T. (2006). Analysis and design of shallow and deep
Shamsabadi, A., Khalili-Tehrani, P., Stewart, J.P. and Taciroglu, E. (2010). Validated
simulation models for lateral response of bridge abutments with typical backfills.
(http://geohazards.usgs.gov/deaggint/2008/)