You are on page 1of 41

SEISMIC PERFORMANCE OF A

NOVEL PRECAST SEGMENTAL


CONCRETE BRIDGE
Amjad J. Aref, Ph.D.
Professor, Department of Civil, Structural,
and Environmental Engineering
University at Buffalo State University of New York
September 20, 2010
PROJECT TEAM MEMBERS
Petros Sideris
Ph.D. Candidate, Dept. of CSEE at UB
Myrto Anagnostopoulou
Senior Structural Engineer, SEESL at UB
Amjad J. Aref
Professor, Dept. of CSEE at UB
Andre Filiatrault
Professor, Dept. of CSEE at UB
OVERVIEW
INTRODUCTION
EXPERIMENTAL INVESTIGATION
Prototype Bridge
Experimental Bridge Specimen
Test Motions Seismic Hazard
Test Execution
NUMERICAL MODELING
Modeling Approaches for Segmental Systems
Multi-Element Approach (as proposed herein)
Simplified 2D Model of Bridge Specimen in Ruaumoko
Simplified 3D Model of Bridge Specimen in SAP2000
Preliminary Comparison of Numerical with Experimental Results
Conclusion
Experimental Investigation
Numerical Modeling for Segmental Systems
INTRODUCTION
Precast Concrete Segmental Bridges
Brief History:
First cast-in-place segmental concrete bridge, Germany (1950) to
cross the Lahn River
First precast segmental concrete bridge, France (1962) to cross the
Seine River
First application in the United States: John F. Kennedy Memorial
Causeway in Corpus Christi, Texas (1973).
Advantages:
Higher construction quality (precast plants)
Rapid construction (Accelerated Bridge Construction - ABC)
Concerns:
Effects of segmental joint response on global system stability
Reliability of existing analysis tools in predicting the 3D dynamic
response of such systems (Performance-Based Design)
Analytical Studies + Large-scale bridge experiments
Investigate these concerns and propose feasible alternatives
Superstructure Section
EXPERIMENTAL INVESTIGATION
Prototype Bridge (Megally et al. 2002)
Single-cell box girder bridge consisting of 5 spans
Each span is post-tensioned with a harped shape tendon
The piers are square hollow sections of ~30 feet height
The Span-by-Span construction method has been assumed
EXPERIMENTAL INVESTIGATION
Front Elevation
EXPERIMENTAL INVESTIGATION
Experimental Bridge Specimen
Single-span bridge - Both of its supports
overhanging at equal lengths (25% of span)
Large-scale model (S
L
=2.4)
Consists of:
Post-tensioned Deck (12 tendons):
8 segments
Post-tensioned Pier (8 tendons):
5 segments
Cap beam
Foundation block
Deck simply supported on cap beams
Sand bags to simulate additional loads
All post-tensioned together
EXPERIMENTAL INVESTIGATION
Experimental Bridge Specimen
Dimensions:
61.875 ft long (pier-to-pier distance: 41.875 ft)
14.125 ft high including cap beam and foundation block
(16.489 ft including deck as well)
Design according to:
AASHTO LRFD Bridge Design Specifications (2007)
PCI Bridge Design Manual (2003) (partial assistance)
Response modifications factors:
Superstructure: R=2.5
Substructure: R=3.75 (=1.5*2.5)
Cap beam and foundation block designed to remain elastic
under any load combination (capacity design)
EXPERIMENTAL INVESTIGATION
Experimental Bridge Specimen
Components (out of scale)
Deck Segment
Pier Segment
Cap Beam Foundation Block
Structural Engineering and Earthquake Simulation Laboratory (SEESL) at UB
EXPERIMENTAL INVESTIGATION
Lateral restrainer

Sand Bag
Cap Beam
Pier
Foundation
Block
Deck
EXPERIMENTAL INVESTIGATION
Post-tensioning system (Deck and piers)
2
'-
4

3
/
8
"
20'-111/4"
10'-0"
1
'-
1
0

1
/
2
"
1
0
'-
0
"
2
'-
3
"
T. 6
T. 7
T. 8 T. 10
T. 9
T. 5
T. 4
T. 3 T. 1
T. 2 T. 11
T. 12
T. 14
T. 13
5'-93/4" 8'-41/2" 8'-41/2" 8'-41/2"
30'-111/4"
1'-27/8"
9"
1'-27/8"
9"
9"
Y
X
9"
1
0
'
-
0
"
2
'
-
3
"
Actuator bolt holes
Ducts
9'-0"
C.M.
11'-0 3/8"
4
'
-
2

7
/
8
"
Stopper against lateral sliding
EXPERIMENTAL INVESTIGATION
Novelties:
Internal unbonded tendons (super- and substructure):
Strain distributed over larger lengths
Higher ductility
Enhanced self-centering capabilities
Moment arm is maintained Stability
Segmental joints (super- and substructure):
Simple plane surface-to-plane surface contact
Negligible/No tensile strength gap opening is allowed
Relative segment sliding
(i) Multi-level seismic isolation
(ii) Restoring force provided by the tendons (dowel effect)
Vertical seismic design spectrum
Considered in MCEER/ATC Joint Venture (MCEER-03-SP03, 2003),
but not in AASHTO 2007
Vertical design spectrum 2/3 of Horizontal
EXPERIMENTAL INVESTIGATION
Test Motions Seismic Hazard
Ground motion (GM) ensembles
Subset of FEMA P695 Far-field GM set (5 out of 22 motions)
Subset of FEMA P695 Near-field GM set (6 out of 28 motions
3 with and 3 without pulse)
Subsets were selected to be representative of the full sets
Both GM sets scaled to Seismic Hazard Levels:
MCE (2% in 50 yrs)
DBE (10% in 50 yrs)
Intermediate DBE (R=2.5 57% in 50 yrs)
Low DBE (R=3.75 88% in 50 yrs)
Asynchronous base excitation
Time delay between the time instants that the seismic wave
reaches each piers
t
1
=0.05 sec (based on soil profile at the site)
t
2
=0.5 sec (amplified)
EXPERIMENTAL INVESTIGATION
Test Motions Seismic Hazard
AASHTO 2007: Sa(T) for 10% in 50 yrs (DBE)
Different hazard levels using FEMA 356 (at short periods)
Scaled subset geometric mean spectra fitted to:
Real System Domain:
Far-Field Subset fitted to AASHTO Spectra
0.0
0.5
1.0
1.5
2.0
2.5
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
T (sec)
S
a

(
g
)
2% in 50 yrs
10% in 50 yrs
88% in 50 years (4.1% per yr)
57% in 50 years (1.6% per yr)
Model Domain:
Far-Field Subset fitted to AASHTO Spectra
0.0
1.0
2.0
3.0
4.0
5.0
6.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
T (sec)
S
a

(
g
)
2% in 50 yrs
10% in 50 yrs
88% in 50 years (4.1% per yr)
57% in 50 years (1.6% per yr)
Real Structure Domain:
Near-Field Subset fitted to AASHTO Spectra
0.0
0.5
1.0
1.5
2.0
2.5
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
T (sec)
S
a

(
g
)
2% in 50 yrs
10% in 50 yrs
88% in 50 years (4.1% per yr)
57% in 50 years (1.6% per yr)
Model Domain:
Near-Field Subset fitted to AASHTO Spectra
0.0
1.0
2.0
3.0
4.0
5.0
6.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
T (sec)
S
a

(
g
)
2% in 50 yrs
10% in 50 yrs
88% in 50 years (4.1% per yr)
57% in 50 years (1.6% per yr)
Link: GM
Ensemble
Tables
EXPERIMENTAL INVESTIGATION
Test Execution
Instrumentation:
~ 220 channels
~ 40 Load Cells Post-tensioning forces
~ 70 accelerometers Accelerometer
~ 90 LVDTs gap opening and relative sliding
~ 20 String Pots global system displacements
Achieved Test Protocol
142 seismic tests
176 identification tests (white noise: PGA~0.1g)
EXPERIMENTAL INVESTIGATION
Test Execution
Achieved Test Protocol:
Unloaded Specimen no lateral restrainers:
Far-field ensemble:
Low DBE
Intermediate DBE
DBE Vertical only
MCE Vertical only
Ultimate loading: MCE for Vertical + 62.5% DBE (R=1.6) for horizontal
Near-field ensemble:
Intermediate DBE
Unloaded specimen with Lateral restrainers in contact with deck:
Far-field ensemble:
Low DBE
Fully Loaded specimen Lateral restrainers at distance from deck:
Far-field:
DBE Vertical only
MCE Vertical only
Low DBE
Low DBE Asynchronous motion: Time delay of t=0.05 sec
Low DBE Asynchronous motion: Time delay of t=0.5 sec
Partially Loaded specimen Lateral restrainers at distance from deck:
Far-field:
DBE Vertical only
EXPERIMENTAL INVESTIGATION
Test Execution
Recorded Dynamic Response
1979 Imperial Valley Earthquake
Delta UNAMUCSD
M
w
=6.5
Appropriately scaled in accordance with similitude
assumptions
Far-Field Motion 2 Ultimate Loading Unloaded
Specimen
Videos
General view - Test ABC_S1_SC_M2_XYZ
Base of west pier - Southwest corner - Test
ABC_S1_SC_M2_XYZ_V05
Base of east Pier - Northwest corner - Test
ABC_S1_SC_M2_XYZ_V11
Test Execution
Dynamic Response
Far-Field Motion 2 Ultimate Loading Unloaded
Specimen:
General view - Test ABC_S1_SC_M2_XYZ
Base of west pier - Southwest corner - Test
ABC_S1_SC_M2_XYZ_V05
Base of east Pier - Northwest corner - Test
ABC_S1_SC_M2_XYZ_V11
Far-Field Motion 4 MCE (Vertical only) Unloaded
specimen:
General view - Test_ABC_S1_FF4_M4_Z_e
Deck mid-joint - Test_ABC_S1_FF4_M4_Z_V13
Far-Field Motion 4 MCE (Vertical only) Fully loaded
Specimen:
General view Test_ABC_S3b_FF4_M4_Z_e
Deck mid-joint - Test_ABC_S3b_FF4_M4_Z_V13
EXPERIMENTAL INVESTIGATION
Link:
Accelerograms
for SC_M2
Link:
Accelerogram
for FF4_M4
EXPERIMENTAL INVESTIGATION
Test Execution
Recorded Dynamic Response
1979 Imperial Valley Earthquake
Preliminary Conclusions:
Maximum acceleration appears to be bounded
Negligible permanent displacements
Total Acceleration - Longitudinal Direction (X) - West Pier
-1.5
-1
-0.5
0
0.5
1
1.5
0 10 20 30 40 50 60
Time
A
c
c
e
l
e
r
a
t
i
o
n

(
g
)
Foundation
Deck
Total Acceleration - Lateral Direction (Y) - West Pier
-1.5
-1
-0.5
0
0.5
1
1.5
0 10 20 30 40 50 60
Time
A
c
c
e
l
e
r
a
t
i
o
n

(
g
)
Foundation
Deck
Relative Displacement - Lateral Direction (Y) - West Pier
-6
-4
-2
0
2
4
6
0 10 20 30 40 50 60
Time
D
i
s
p
l
a
c
e
m
e
n
t

(
i
n
)
Cap-beam
Deck
EXPERIMENTAL INVESTIGATION
Test Execution
Recorded Dynamic Response
1992 Landers Earthquake
Coolwater - SCE
M
w
=7.3
Appropriately scaled in accordance with similitude
assumptions
Far-Field Motion 4 MCE (Vertical only) Unloaded
specimen:
EXPERIMENTAL INVESTIGATION
Test Execution
Recorded Dynamic Response
1992 Landers Earthquake
Preliminary Conclusions:
Resonance, but response is bounded
Negligible permanent displacements
Vertical Total Acceleration - Mid-span
-6
-4
-2
0
2
4
6
0 5 10 15 20
Time
A
c
c
e
l
e
r
a
t
i
o
n

(
g
)
Deck
Foundation
Vertical Total Acceleration - Mid-span (Zoom in)
-6
-4
-2
0
2
4
6
6 7 8 9 10 11 12 13 14
Time
A
c
c
e
l
e
r
a
t
i
o
n

(
g
)
Deck
Vertical Relative Displacement - Mid-span
-2.5
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
2.5
0 2 4 6 8 10 12 14 16 18 20
Time
D
i
s
p
l
a
c
e
m
e
n
t

(
i
n
)
Deck
NUMERICAL MODELING
Modeling Approaches for Segmental Systems
Finite Element Method
Approach
3D solid elements and 2-node elements
Nonlinear material properties
Contact interfaces friction
Sequential loading
Good accuracy, if convergence can be achieved
Excessive computational resources
Analysis time - Storage capacity
Difficult to use
Beam-Column Elements:
Approach:
Equivalent plastic hinges (P
x
-M
y
-M
z
Interaction) to model joint opening
and sliding
Sequential loading
Easy to use
May provide fairly good results, but cannot provide a general
framework
NUMERICAL MODELING
Modeling Approaches for Segmental Systems:
Beam-column elements with zero-length hertzian contact
springs
Placed at the ends of beam-column element
Distributed the edge of the end cross-sections
Easy to use
May provide fairly good results in some cases
It cannot provide a general framework, since concrete crushing
is not considered
Macro-element approach:
It could provide an optimum balance amongst:
Desired accuracy
Computational resources
Simplicity
Currently under development (not presented herein !)
NUMERICAL MODELING
Modeling Approaches for Segmental Systems:
Multi-Element Approach (as proposed herein)
Concept:
Use elements from existing structural analysis software
Available to practicing engineers and small/medium
construction firms
Easy to use
Sequential loading
Better performance than beam-column element approach
(???)
This presentation will illustrates some partial
findings of this research for the multi-element
approach:
Structural analysis software:
Ruaumoko 2D model
SAP2000 3D model
NUMERICAL MODELING
Multi-Element Approach (as proposed herein)
Segment:
Beam-column element for interior part of a segment
End plastic hinges with Px-My-Mz Interaction must be
considered for consistency (not for joint opening and/or sliding!!!)
Fiber Springs to model end regions of segments
Combination of 2-node friction/contact elements and
hysteretic elements
Appropriately distributed over segment cross-section in
parallel with each other and the neutral axis of the segment
Post-tensioning Strands:
Tension-only truss elements with inelastic behavior
Post-tensioning is applied as: initial stress or initial strain or
initial temperature change
Appropriate inter-element boundary conditions:
Rigid links:
Connect beam-column with fiber springs
Keep tendons at proper position
Link: Schematic derivation of multi element modeling
NUMERICAL MODELING
Multi-Element Approach (as proposed herein)
Criteria for Distribution of Fiber Springs
Equivalent Cross-Section moduli:
Maximum distance between fiber springs Resultant axial
force should be able to run smoothly over the area of the
cross-section Partial contact should be allowed
Symmetric distribution for symmetric sections
2 2
1
2 2
1
1
N
y i i i
i
A
N
z i i i
i
A
N
i i i
i
A
EI Ez dA z E A
EI Ey dA y E A
EA E dA E A
=
=
=

Recommendation:
Acceptable Error < 1%
1
0
N
i i
i
i
i
E A
y
L
=


1
0
N
i i
i
i
i
E A
z
L
=


and
NUMERICAL MODELING
Multi-Element Approach (as proposed herein)
Length of fiber springs:
~0.9d (from section depth), or ~1-1.5 t (from wall thickness)
Further calibration of Fiber Springs
Axial Properties:
Stiffness:
Yield Force:
Shear Properties: According to friction properties at the
segment-to-segment interface and segment shear properties
Bending Properties: Correspond to global segment properties
as if is was modeled as a beam-column element
Torsional Properties: Correspond to global segment properties
as if is was modeled as a beam-column element
,
1
N
Beam Column fiber
y y i
i
F F

=
=

1
N
Beam Column fiber
i
i
K K

=
=

NUMERICAL MODELING
Multi-Element Approach (as proposed herein)
Simplified 2D Model of Bridge Specimen in Ruaumoko
Lateral direction
Relative joint sliding is not considered
Shear is transferred by a pinned connection between the two
segments. Pin is located at segmental interface
Rigid links
Contact
element
Fiber
Springs
Tendons
No
sliding
Deck mass
Beam-Column
element
NUMERICAL MODELING
Multi-Element Approach (as proposed herein)
Simplified 2D Model of Bridge Specimen in Ruaumoko
Short example - Fully loaded specimen
Segment interior: Beam-column elements with end plastic
hinges (P
x
-M
y
-M
z
Interaction)
Segment End discretization: 9 compression-only bilinear
hysteretic Fiber springs of length (2x6=12)
Tendons: Tension-only bilinear elements with slackness and
initial loading
Applied motion:
N-S Component of 1940 El Centro record
Properly scaled (similitude) in time (x 1/2.388) and
amplitude (x 2.388) PGA=1.62g (versus the original
0.34g)
Example results
Modal analysis: T
1
=0.224sec, T
2
=0.018sec and T
3
=0.009sec
NUMERICAL MODELING
Multi-Element Approach (as proposed herein)
Simplified 2D Model of Bridge Specimen in Ruaumoko
Example results
Dynamic analysis (=3%):
Preliminary Conclusions:
Total deck acceleration seems to be limited at 0.5 -0.6 g
System, after minor concrete crushing at pier base, returns
to its original position
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
0 2 4 6 8 10 12
t (sec)
T
o
t
a
l

A
c
c
e
l
e
r
a
t
i
o
n

(
g
)

Base
Deck
-2.50
-2.00
-1.50
-1.00
-0.50
0.00
0.50
1.00
1.50
2.00
2.50
0 2 4 6 8 10 12
t (sec)
R
e
l
a
t
i
v
e

D
e
c
k

D
i
s
p
l
a
c
e
m
e
n
t


(
i
n
)


.
NUMERICAL MODELING
Multi-Element Approach (as proposed herein)
Simplified 3D Model in SAP2000
Segment ends:
Definition of Fiber Spring:
Two 2-node Nonlinear Links in series
(i) Friction Isolator (lateral response)
(ii) Multi-linear Plastic Spring with Kinematic
Hardening (axial response)
Length: 2x6=12 for piers, and 2x7.5=15 for deck
Cross-section discretization:
Deck Pier
Link: Schematic derivation of multi element modeling
NUMERICAL MODELING
Multi-Element Approach (as proposed herein)
Simplified 3D Model in SAP2000
Segment Interior:
Beam-column element with end plastic hinges (P
x
-M
y
-M
z
Interaction)
PT Tendons:
Beam-Column Steel Element, with initial strain to induce PT
forces (I
yy
=I
zz
=J=0 ~ truss element)
Material properties:
Concrete: f
c
=6000 psi (unconfined)
Steel: F
y
=50 ksi
Segmental Joint and Deck-to-Cap beam interface
Friction: =0.3
Model development:
Sequential load application (post-tensioning, dead loads, live
loads)
Beam Element passing
through the center line of
the deck cross-section
Extrude View
Beam Element passing
through the center line of
the pier cross-section
Extrude View
Segment-to-segment
contact
Segment-to-segment
contact
NUMERICAL MODELING
Multi-Element Approach (as proposed herein)
Simplified 3D Model in SAP2000
Friction Element
Hysteretic Element
Beam - Column Elements
(for tendons)
NUMERICAL MODELING
Multi-Element Approach (as proposed herein)
Simplified 3D Model in SAP2000
Modal analysis:
Difference in the fundamental mode of the SAP2000 model
with the Ruaumoko Model, mainly due to:
2D versus 3D model
Infinite shear stiffness at segmental joint in 2D model
Mode T (sec) Mode Shape / Deformation Characteristics
1
st
0.284 Uniform lateral Pier bending
2
nd
0.163 Anti-symmetric lateral
3
rd
0.150 Longitudinal Pier and deck bending
4
th
0.084 Vertical Deck Bending
5
th
0.040 Lateral Lateral Deck Bending / Torsion of piers
NUMERICAL MODELING
Multi-Element Approach (as proposed herein)
Simplified 3D Model in SAP2000
Dynamic analysis:
Applied motion:
Motion 3 from FEMA P695 Far-Field GM Subset (Nishi-
Akashi components - 1995 Kobe earthquake)
DBE Hazard Level
Properly scaled in time (x 1/2.388) and amplitude (x
2.388), due to similitude requirements
After all scaling: PGA
x
=251g, PGA
y
=2.47g,
PGA
z
=1.81g
Rayleigh damping of =3% assigned to the 1
st
and 4
th
mode
NUMERICAL MODELING
Multi-Element Approach (as proposed herein)
Simplified 3D Model in SAP2000
Dynamic analysis:
Computed Response
-3
-2
-1
0
1
2
3
0 5 10 15
Time (sec)
T
o
t
a
l

A
c
c
e
l
e
r
a
t
i
o
n

X

(
g
)
Base Acceleration
Deck
-3
-2
-1
0
1
2
3
0 5 10 15
Time (sec)
T
o
t
a
l

A
c
c
e
l
e
r
a
t
i
o
n

Y

(
g
)
Base Acceleration
Deck
-5
-4
-3
-2
-1
0
1
2
3
4
5
0 5 10 15
Time (sec)
R
e
l
a
t
i
v
e

D
i
s
p
l
a
c
e
m
e
n
t

(
i
n
)
Longitudinal
Lateral
Vertical
NUMERICAL MODELING
Multi-Element Approach (as proposed herein)
Simplified 3D Model in SAP2000
Dynamic analysis:
Preliminary conclusions:
Total acceleration seems to be limited, mainly due to:
(i) Concrete crushing and relative segment sliding
at joints
(ii) Sliding of deck on cap beams
Lateral total acceleration does not exceed 0.3g
0.5g
Longitudinal total acceleration does not exceed 0.9g -
1.1g
System returns to its original position
NUMERICAL MODELING
Preliminary Comparison of Numerical with
Experimental Results
Modal analysis
Fundamental frequency comparison (Loaded???)
Deviation may be mainly attributed to Silicone sealant used
at the segmental joint.
If shear spring included:
T1=0.35 sec, T3=0.31 sec, T4=0.12 sec
NUMERICAL MODELING
SAP2000 Experiment
Mode T (sec) T (sec) Mode Shape / Deformation Characteristics
1
st
0.284 0.36 Uniform lateral Pier bending
3
rd
0.150 0.27 Longitudinal Pier and deck bending
4
th
0.084 0.14 Vertical Deck Bending
CONCLUSIONS
Experimental Investigation
A novel bridge system consisting of post-tensioned
superstructure and substructure was tested
System was subjected to severe ground motions:
Deck: Survived with minor concrete crushing several MCE motions
Piers: Survived severe Far-field and Near-field motions
General characteristics of the response of segmental systems
(associated with the novel structural concepts of this study),
which originally observed from numerical analyses, proved to
be valid experimentally as well:
Segmental joint opening and relative sliding have been observed to
provide the system with enhanced self-centering capabilities and
higher ductility capacity
Damage on concrete segments was mainly spalling of the rebar
cover and some crushing of concrete at base segments
CONCLUSIONS
Numerical Modeling for Segmental Systems
A method to model efficiently segmental systems using existing
structural analysis software widely available to practicing
engineers
Similar approaches have been used in literature; however in
this study, a general framework is attempted to be established
consisting of rules and recommendations based on general
principles of classical structural analysis
The proposed framework was used with two structural analysis
programs: Ruaumoko and SAP2000. Both models appeared to
capture the general trends of the response (which also
observed experimentally)
A comparison of the 3D SAP2000 model with some modal
experimental results clearly showed the need for further
refinement of this technique Especially to capture sliding
ACKNOWLEDGEMENTS
Federal Highway Administration of the U.S.
Department of Transportation
Bodossaki Foundation
SEESL Personnel (University at Buffalo)
Joe Salvadori (DSI)
Curt Haselton (California State University,
Chico)
David Welch (University at Buffalo)
Thank you!!!
Questions?

You might also like