Site Response Analysis

Report
An Evaluation of Current Site Response
Analysis Methods
Chandrakanth Bolisetti
Graduate Student Researcher
Dr. Andrew Whittaker
Professor and Chair
Department of Civil, Structural and Environmental Engineering
University at Buffalo, SUNY
The City Block Project
Acknowledgments
• National Science Foundation, CMMI 0830331
• Dr. Amjad Aref, University at Buffalo
• Ibrahim Almufti and Dr. Michael Willford, ARUP San Francisco
• Dr. Boris Jeremic, UC Davis
• Dr. Ben Mason, Oregon State University
Overview
• Soil-structure interaction analysis for performance assessment
of buildings and nuclear power plants
– Detailed 3D analyses
– Nonlinear analyses for high intensity ground motions
• Evaluation of existing industry-standard numerical tools
– Site response analysis (pre-requisite for SSI analysis)
– SSI analysis
• SSSI analysis
Overview
• Soil-structure interaction analysis for performance assessment
of buildings and nuclear power plants
– Detailed 3D analyses
– Nonlinear analyses for high intensity ground motions
• Evaluation of existing industry-standard numerical tools
– Site response analysis (pre-requisite for SSI analysis)
– SSI analysis
• SSSI analysis
Outline
• Introduction
• Numerical Tools
• Numerical Analysis
• Sample Results
• Conclusions and future research
Introduction
Site Response Analysis
1D site response analysis
• Purposes
– Site effects for seismic hazard analysis
– Soil-structure interaction analysis
Introduction
Site Response Analysis
• State-of-the-art
– Frequency domain equivalent linear analysis
• SHAKE, DEEPSOIL
– Time domain nonlinear analysis
• DEEPSOIL nonlinear, LS-DYNA
– Mostly 1D
• Limitations
– Mostly developed for characterizing site effects
– The 1D assumption
• Horizontal ground motion components are usually not uncorrelated
• Not sufficient for high fidelity SSI analyses required for performance
assessment of NPPs (Jeremic, 2011)
Numerical Tools
Frequency Domain
• The equivalent linear approach: SHAKE and DEEPSOIL
– Seed and Idriss (1969)
– Iterative procedure
– Modulus reduction and damping
curves
• Effective shear strain ratio
R 
M 1
10
– An empirical value of 0.65
is recommended
Hashash et al, 2010
Numerical Tools
Time Domain
• DEEPSOIL nonlinear
– MKZ model (Matasovic, 1993)

 G0
 
1   
 r 
s
– Extended Masing rules
define the stress-strain
hysteresis
– Outcrop input using the
Joyner and Chen (1975)
method
Hashash and Park (2001)
Numerical Tools
• LS DYNA nonlinear
– General finite element analysis
– Column of solid elements
constrained to move in shear
– MAT_HYSTERETIC model (MAT_079)
– Outcrop input using the Joyner
and Chen (1975) approach
– ARUP, San Francisco
Time Domain
Numerical Analyses
Site E1
Site E2
Site Selection
Site W1
Site W2
300m/s
2500m/s
1000m/s
300m/s
2500m/s
Bed Rock
2500m/s
Bed Rock
2500m/s
Bed Rock
1000m/s
Bed Rock
1000m/s
100m
Numerical Analyses
WUS Ordinary motions
Event
Station
PGA (g)
Northridge, 1994
Vasquez Rocks Park
0.15
Northridge, 1994
Wonderland Ave
0.17
San Fernando, 1971
Lake Hughes #4
0.19
WUS ordinary ground motions
0.8
GM-1
GM-2
GM-3
Site-W1
Site-W2
Acceleration (g)
0.6
0.4
0.2
0
0.01
0.1
1
P eriod (sec)
10
Numerical Analyses
WUS Pulse motions
Event
Station
PGA (g)
Tp (sec)
Landers, 1992
Lucerne
0.73
5.1
Northridge, 1994
Rinaldi Receiving Stn.
0.83
1.5
Chi Chi, Taiwan, 1999
TCU 128
0.19
9.0
Acceleration resp onse sp ectra for selected pulse motions
3
Acceleration (g)
LCN260 T p = 5.12 sec
RRS228 T p = 1.51 sec
T CU128 T p = 9.00 sec
Site-W1
Site-W2
2
1
0
0.01
0.1
1
P eriod (sec)
10
Numerical Analyses
CEUS motions
Event
Station
PGA (g)
Virginia, 2011
Charlottesville
0.10
New Hampshire, 1982
Franklin Falls Dam
0.31
Saguenay, CA, 1988
Dickey
0.09
CEUS ordinary ground mot ions
1
CVA090
FFD315
SNY090
Site-E1
Site-E2
Acceleration (g)
0.75
0.5
0.25
0
0.01
0.1
1
P eriod (sec)
10
Sample Results
Site E1, Charlottsville
Peak acceleration profiles
Comparison of acceleration response spectra at the surface Peak strain profiles
0
0
0.4
Shake
Shake
Mat Hyst eretic
Deepsoil
Shake
Mat Hysteretic
Mat Hyst eretic Ramberg Osgood
Deepsoil
Deepsoil
0.3
 25
 50
Depth below surface (m)
Acceleration (g)
Depth below surface ( m)
 25
0.2
 50
0.1
 75
 75
0
0.01
0.1
1
10
P eriod (sec)
 100
 100
0
0
0.025
0.05
0.075
110
4
4
210
0.1
P eak strain (%)
P eak acceleration (g)
4
310
410
4
Sample Results
Site W1, Vasquez Park
Peak acceleration profiles
Peak strain profiles
0
0
Shake
Mat Hyst eretic
0.8
Deepsoil
Shake
Mat Hyst eretic
Deepsoil
Comparison of acceleration response spectra at the surface
Shake
Mat Hyst eretic
Deepsoil
 25
 25
 50
Depth below surface ( m)
Acceleration (g)
Depth below surface ( m)
0.6
0.4
 75
 50
 75
0.2
 100
0
0.05
0
0.01
0.1
P eak acceleration (g)
0.15
 100
0.2
0.1
0
0.01
1
P eriod (sec)
0.02
P eak strain (%)
0.03
0.04
10
Sample Results
Site W1, Rinaldi
Peak acceleration profiles
Peak strain profiles
0
0
Shake
Mat Hyst eretic
Deepsoil2
Shake
Mat Hysteretic
Deepsoil
Comparison of acceleration response spectra at the surface
Shake
Mat Hyst eretic
Deepsoil
 25
 25
 50
Depth below surface ( m)
Acceleration (g)
Depth below surface ( m)
1.5
1
 75
 50
 75
0.5
 100
0
0.375
0
0.01
0.75
P eak acceleration (g)
1.125
 100
1.5
0.1
0
0.5
1
P eriod (sec)
1
P eak strain (%)
1.5
10
Conclusions
• Good match for low soil strains but large differences at high
soil strains (close to 1%)
• Peak strain values are underestimated in SHAKE, especially for
intense motions
– Effective shear strain ratio?
• Accelerations are underestimated in SHAKE
– Large values of damping ratio?
• Implications for SSI analysis
– Need to be cautious when large strains are expected
– 1D analysis insufficient (Jeremic, 2011)
– Materials not suitable for full SSI analyses
Conclusions
• High frequency ‘noise’ in time-domain analysis results
–
–
–
–
Piecewise nonlinearity (LS DYNA only)
Internal wave reflections due to impedance changes
Joyner and Chen (1974)
Cautious site layering, or filtering of the response
• SHAKE response for pulse motions
– Convergence issues
– Smaller value of effective shear strain ratio needs to be used
Contacts
Chandu Bolisetti: [email protected]
Dr. Andrew Whittaker: [email protected]

similar documents