You are on page 1of 48

Methodology for the Seismic

Analysis of Mercado de Santa


Anita
Juan M. Pestana, Sc.D., P.E.
Professor, University of California, Berkeley

Contact info: jmpestana@berkeley.edu

Lima, Peru. July 6, 2016 1/48


Outline of Presentation
Simulation of Seismic Structural Response
• Seismic Risk:
– Design Scenarios: 2 level evaluation
– Ground Motion Intensity Measures
– Selection of “Seed” & Modified Ground Motions
• Site Response Analysis:
– Soil Behavior & Parameters. 1D vs. 2D BV problems
• Soil-Structure Interaction Analysis
– Simplified Approaches
– 2D Dynamic Analyses 2/48
Simulation of Seismic Performance

3/48
Seismic Structural Response
• Inertial vs. Kinematic Response
– Above ground: Primarily Inertial Response:
Energy input to structure. Single degree of
Freedom System: Response Spectra.
– Below ground: Primarily Kinematic Response:
controlled by ground deformation, so Soil-
Structure-Interaction is essential
• For underground structures, Soil-Foundation-
Structure is the key to correctly describe
performance.
4/48
Outline of Presentation
Simulation of Seismic Structural Response
• Seismic Risk:
– Design Scenarios: 2 level verification.
– Ground Motion Intensity Measures
– Selection of “Seed” & Modified Ground Motions
• Site Response Analysis:
– Soil Behavior & Parameters. 1D vs. 2D BV problems
• Soil-Structure Interaction Analysis
– Simplified Approaches
– 2D Dynamic Analyses 5/48
Seismic Sources- Deaggregation

6/48
Seismic Hazard-
Example New Zealand

7/48
Ground Motion Intensity Measures
Selected Ground Motion Parameters
• Amplitude Parameters
– Peak Ground Acceleration (PGA)
– Peak Ground Velocity (PGV)
– Peak Ground Displacement (PGD)
• Frequency Parameters
– Ground Motion Spectra (Fourier, Power and Structural
Response Spectra) and mean period.
• Duration (several measures, e.g. bracketed > 0.05g)
• Other: Arias Intensity, Rate of Arias intensity,
presence of velocity pulse, period of pulse.
• Time Domain Response Spectrum.
8/48
Response Spectra
• Structural Response Spectra: Response of 1 DOF
system to ground motion. Therefore, it is a filter on
the ground motion. Maximum values do not occur at
the same time & does not quantify how many times
or how long the maximum value is sustained.
• Fourier vs. Response Spectra: Fourier is sometimes
reported when studying site response (in absence of
structure) because it is an accurate description of
motion.
• Please note that the Spectra is associated with a given
value of the damping ratio (typically 5% but not always)
9/48
Time Domain Response Spectra
(after Perri & Pestana, 2007)

Windowing Input Motion and Evaluating Structural


Response. Window length is a function of the
Fundamental Period of Structure being evaluated

10/48
Period Dependent Duration

11/48
Ground Motion Selection
• Key is to select ground motions that are representative from
the tectonic environment (source) and scenario (e.g.,
expected EQ magnitude and distance) of interest.
• Current state of practice uses Probabilistic Hazard
Assessment and Disaggregates the contribution to select
key scenarios (e.g., low M, short distance + high M, large
distance- example, San Francisco).
• Important characteristics: amplitude, frequency content,
duration, arias intensity, directivity (for near fault events).
• Issue of How many ground motions, matched or unmatched
(to the response spectra), (1, 3, 7, 11)

12/48
Selection of Ground Motions
Additional Resources & New Developments

• Ground Motions Prediction Equations:


– NGA: Next Generation Attenuation Relationships.
Different tectonic environments (NGA-West, NGA
East, Subduction)
• Check with Research Centers (some examples)
– PEER: Pacific Earthquake Engineering Research Center.
Large database for selection of ground motions. Reviewed.
– MCEER: Multidisciplinary Center for Earthquake Engng.
– Local Research Centers & Universities.
• Other: US. Geological Service, EERI, SCEC
13/48
Outline of Presentation
Simulation of Seismic Structural Response
• Seismic Risk:
– Design Scenarios: 2 level verification.
– Ground Motion Intensity Measures
– Selection of “Seed” & Modified Ground Motions
• Site Response Analysis:
– Soil Behavior & Parameters. 1D vs. 2D BV problems
• Soil-Structure Interaction Analysis
– Simplified Approaches
– 2D Dynamic Analyses 14/48
Dynamic Numerical Simulations
• Site Response Analysis: Assessment of maximum
shear stresses and strains.
• Boundary value problems where permanent
deformation is expected or 2D cases
• Liquefaction Problems: triggering, post-liquefaction
deformation, loss of free board for dams and levees,
lateral spread, soil improvement.
• Soil Structure Interaction (deep foundations, shallow
foundation, rocking and uplift)
• Problems are becoming increasingly complex, 1D, 2D
& 3D. (Structure-structure interaction- City block)
15/48
Site Response Analysis-Local Site Effects
Ground surface motions are affected by local site conditions

Site effects can influence:


• Amplitude - may amplify or de-amplify motion
• Frequency content - may shift to higher or lower frequency
• Duration - may extend duration of strong shaking

Shown by:
• Measured (recorded) surface motions
• Compilations of data on local site effects
• Measured
Correlation amplification functions
of Shaking
• Theoretical with
Amplification analyses
Site Conditions
16/48
Ground Response Analysis- Definitions
• Rock outcropping motion - the motion that would occur where rock
outcrops at a free surface
• Bedrock motion - the motion that occurs at bedrock overlain by a
soil deposit. Differs from rock outcrop motion due to lack of free surface effect.
• Free surface motion - the motion that occurs at the surface of a soil
deposit.
Free surface motion Rock outcropping motion

What is bedrock? For


earthquake engineering
purposes, it is usually Soil Rock
taken as material with vs
> 2,500 fps (750 m/s)
(soft rock). Sometimes,
relaxed but close. Bedrock motion
17/48
Input Ground Motion
(Prescribed Motion at the Base)
•Usually, we would like to model our problem all the way down
to bedrock, and use a bedrock motion as our input motion
•In cases where bedrock is at relatively shallow depths, we can
extend our finite element mesh all the way down to bedrock

Bedrock
18/48
Input Ground Motion (Prescribed Motion at the Base)
• In some cases, bedrock is so deep that it is impractical to model the
entire soil column. In such cases, we need to define an input motion at
a reasonable depth.
• That depth should be sufficient that the motion is not significantly
influenced by structures or topographic irregularities (i.e. it could be
computed from a 1-D analysis without significant error)

Ideally, the bottom layer


should not have
significant nonlinearity,
and it should be relatively
stiff and a reasonable
impedance contrast (if
applicable)

19/48
Input Ground Motion
The motion at the desired depth
can be computed from a 1-D
analysis, using PLAXIS or
other site response program

Input motion
Base of
mesh level

Bedrock motion
Bedrock

20/48
Dynamic Soil Properties
Propagation of Shear Waves controlled by:
1. Density (Unit Weight) - Mass / Vol (Weight / Vol)
2. Shear Stiffness and Damping- Dependency on strain level
3. Pore pressure generation (if applicable)

Comments/ Observations:
1. Density controls inertia- typically well bounded
2. Stress-strain properties of the soil will give rise to very
different soil responses which will be reflected primarily in
shear modulus reduction and damping curves. These are a
function of stress level, density and type of material and
mode of deformation. Currently few “realistic” models
implemented in PLAXIS but situation is rapidly changing.

21/48
Use of Realistic Shear Modulus Curves
• Example: HS Model with small strain stiffness: Two new
Parameters Goref and  0.7 , small strain stiffness (related to
shear wave velocity) and a reference strain at which the shear
stiffness is approximately 70% of the original value. The
stiffness is described by this relationship:

G 1

G0   
1  0.385  

 0.7 

22/48
Use of Realistic Shear Modulus &
Damping Relations
1.0
ref
Shear Modulus Reduction, G /G
sec

0.8
Confining
0.6 Pressure
(kPa)
Void
0.4 Ratio
Model Predictions 300
Model Predictions
G = 700, 0.80
G = 700, p = 100 kPa
0.2 b b
0.65
 = 1.0,  =1.0  = 1.0,  =1.0
s a s a
20
0.0
25
Toyoura sand 0.65
Toyoura sand
Damping Ratio (%)

20 (Kokusho 1980) 20 0.80


(Kokusho 1980)
Confining Pressure (kPa) Void
Void Ratio
15 ~20 0.793 Ratio
~50 300
10 0.742
~100 Increasing
0.696
~200 Confining
5 Pressure ~0.65
~300
0
0.0001 0.001 0.01 0.1 1 0.0001 0.001 0.01 0.1 1
Single Amplitude Shear Strain,  (%) Single Amplitude Shear Strain,  (%)
23/48
eservoir, Dynamic Deformation

Example of Dynamic Numerical Simulations:


San Pablo Dam (California)
340
Hydraulic Fill Shell WSEL +314 ft
300
Dow nstream Buttress (1967)
Elevation

260 Upstream Buttress (1979)

220 Hydraulic Fill Shell


Puddle Core
180
Foundation Alluvium
140

100
-0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

Distance (ft) (x 1000)

Courtesy of F. Makdisi
Crest Settlement ~ 35 ft

24/48
Example of Dynamic Numerical Simulations:
Success Dam
– 145-foot high zoned embankment dam
– Analyzed with TARA-FL, Quad4, and 3 FLAC approaches
– Recent alluvium beneath upstream and downstream shells
• Potentially liquefiable
• Modeled with FLAC – UBCSAND

Courtesy of Dr. Vlad Perlea, USACOE


25/48
Example of Dynamic Numerical
Simulations: Success Dam
Existing Embankment Evaluation (MCE):
Courtesy of Dr. Vlad Perlea, USACOE

Vertical
Displacement
(Contours at 5-
foot intervals)

Maximum
Shear Strain
Increment

26/48
Example of Dynamic Numerical Simulations: 0.500

0
+02
+02
Submerged BART Tube (California) -0.500

Stiff Clay (MPSA-C)

Stiff Clay (OBM)


-1.500

Dense Sand (UAM) -2.500


0.6
Hard Clay (LAA)
0

Input Motion -0.6


0 20 40 60 -3.500
 Fully coupled dynamic analyses
 Finite difference method (FLAC2D)
-4.500

 UBCSAND (Prof. Byrne) for liquefiable soils


2E 2
 Calibration through simulation of CSS tests
-3.000 -2.000 -1.000 0.000 1.000 2.000 3.000
 Alternative model in finite elements (OpenSees)
(*10^2)

Courtesy of T. Travasarou
27/48
Representative Soil Properties
Fill Soil Type ρ qc1N Dr % k (cm/s)
Ordinary Sand ~1.9 ρw 50 40 +/- 5 0.01
Special Gravel ~2.1 ρw 32 35 +/- 5 1.0

Tunnel
ρ=~ 1.1 ρw

Centrifuge
Model
(scale 1 : 40)

Kutter et al. (2008) 28/48


Simplified Analysis- “Raking” Approach

29/48
Example of Dynamic Numerical Simulations:
Doyle Drive Replacement (California)

Analyses of Ground Improvement and Cut and Cover Performance


Pestana, ARUP 2008

30/48
Interpretation of Results
• Plots of curves:
• Plot various quantities vs. time: Time Histories
Acceleration, velocity, displacement, Shear stress, shear strain,
Effective stress, pore pressures
• Plot various quantities vs. spatial position:
Displacement : Along section, Deformed mesh
Shear stress, shear strain: Contours, Shading
• Plot temporal and spatial variation together
Very important for proper
Helps analyst interpretation of results of
understand how dynamic analysis
system is behaving Helps identify errors in
Animation
input by providing
“reality check”
31/48
Application of Methodology to the Seismic Performance 2500 return
Period event.

MERCADO DE SANTA ANITA

32/48
Seismic Sources
• Seismic Risk arising from three potential
sources:
– Transform Faulting Seismicity
– Subduction Faulting Seismicity
– Regional Seismicity
• According to Seismic Risk Study, risk is
dominated for Subduction Faulting Seismicity
with equal contributions from events at
distances of 40 km and 130 km
33/48
Seismic Hazard- 2500 years Return Period

34/48
Seed Motions
• Selected from similar tectonic environment :
subduction zone, distance and ground conditions
Compared to Transform Fault, no database is
available. (see Carlton, Pestana and Bray, 2015).
– Maule, Chile 2010; Santiago Puente Alto station
Mw= 8.8, Rrup= 75 km, Vs= 540 m/s
– Maule, Chile 2010; Cerro Santa Lucia station
Mw= 8.8, Rrup= 75 km, Vs= 540 m/s
– South Peru, 2001; Moquegua station.
Mw= 8.4, Rrup= 71 km, Vs= 573 m/s
35/48
Spectrally Matched Ground Motion
Maule, Chile 2010 Santiago Puente Alto

36/48
Ground Motion for Numerical Analyses

37/48
Short Term Response Spectrum
(after Perri and Pestana, 2007)

38/48
Site Response-
Shear Wave Velocity Profile

39/48
Nonlinear Soil Properties

Example:
Dense Gravels

40/48
Free surface motion Rock outcropping motion

Deconvolution
Soil Rock

Bedrock motion

41/48
42/48
Numerical Model of Underground
Station- Mercado Santa Anita

Close-up
Structural detail

43/48
Seismic Response of Station

44/48
Relative Displacement
DERIVAS CHILE
(MAULE 2500)
12.00

10.00

8.00

6.00

4.00

2.00

0.00

-2.00
Deriva [mm]

-4.00

-6.00

-8.00

-10.00

-12.00

-14.00

-16.00

-18.00

-20.00
0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00 45.00
Tiempo [s]

A-C A-B B-C

45/48
Simplified Pseudostatic
vs. Dynamic Modeling
• Analyses using the raking approach give
smaller structural demands than the Dynamic
Modeling using PLAXIS.
• Demands from PLAXIS were selected for
verification (for 475 and 2500 years).

46/48
Structural 0
LEYES DE ESFUERZOS FLECTORES EN ELA

Demand- -1

-2
S09.2500E:
Mmax
[kNm/m]
-3

Moments in -4

-5

retain wall. -6

-7
S09.2500E:
Mmin [kNm/m]

-8

TR=2500 years. -9

PROFUNDIDAD (m)
-10 S09.2500E:
Mmax
-11 [kNm/m]
-12

-13

-14
S09.2500E:
-15 Mmin [kNm/m]

-16

-17

-18

-19 T2.1_Rd : Mmax


[kNm/m]
-20

-21

-22

-23
T2.1_Rd : Mmin
[kNm/m]
-24

-25
-2,200
-2,000
-1,800
-1,600
-1,400
-1,200
-1,000
-800
-600
-400
-200
0
200
400
600
800
1,000
1,200
1,400
1,600
1,800
2,000
ESFUERZOS EN m·kN/ml.
47/48
CONCLUSIONS
• Given the methodology described in here:
– Ground Motions
– Site Response
– Soil-Structure Analysis: Simplified Raking vs.
Numerical Dynamic Analysis.
• Not presented here: Evaluation for 475 years is
similar- Station performs in “elastic range”
• Evaluation for the scenario 2500 years return
period, the station does not collapse.
48/48

You might also like