Transcript Pacific Earthquake Engineering Research Center (PEER)

PEER Ground Motion Research Projects:
A Progress Report
Yousef Bozorgnia, Ph.D., P.E.
PEER Associate Director
PEER Annual Meeting,
January 20, 2007
San Francisco
PEER Ground Motions Projects
Ground Motion
Selection &
Modification
Other GM
projects
Upcoming
tasks/projects
PEER Annual Meeting, January 2007
Peak Horizontal Acceleration (g)
NGA, and its
associated
projects
1
0.1
Magnitude = 7.5 -- pga -- Strike Slip
Abrahamson & Silva (Vs30 = 760 m/s)
Boore & Atkinson (Vs30 = 760 m/s)
Campbell & Bozorgnia (Vs30 = 760 m/s)
Chiou & Youngs (Vs30 = 760 m/s)
Idriss [Vs30 = 450 -- 900 m/s]
0.01
0.1
1
10
Closest Distance (km)
100
Latest Developments in
Next Generation Attenuation (NGA)
Next Generation Attenuation (NGA) Is a Set of
Multidisciplinary Projects
Bringing together:
seismologists,
geologists,
geotechnical
engineers,
structural
engineers, and
users of ground
motion models
PEER Annual Meeting, January 2007
Step1: Compiled Strong-Motion Database
8
173 worldwide
earthquakes
> 10,000
uniformly
processed
records
Magnitude
7
6
5
4
0.1
1
10
100
Distance (km)
Previous Data New Data
0.5
A c c e l. (g )
4 7 De g .
0.0
-0.5
0.5
A c c e l. ( g )
V e r t ic a l
0.0
-0.5
0.5
A c c e l. ( g )
31 7 D eg .
0.0
-0.5
20
PEER Annual Meeting, January 2007
25
30
35
40
1000
NGA Database
There are more than 100 variables describing
the source/path/site conditions of a record:
6 types of distance
measures
4 site classification
schemes
Estimated VS30 for
most of recording
sites
FW/HW classes
The database is available to the public
PEER Annual Meeting, January 2007
NGA Model Developer Teams
NGA empirical ground motion model developers
 Abrahamson & Silva (updating their 1997
model)
 Boore & Atkinson (updating Boore et al., 1997
model)
 Campbell & Bozorgnia (updating their 1997,
2003 models)
 Chiou & Youngs (updating Sadigh et al., 1997
model)
 Idriss (updating his 1993 & 1996 models)
All developers started with a common database
PEER Annual Meeting, January 2007
NGA Attenuation Models
Ground motion
parameters:
Magnitude range:

Horizontal components
 5.0 - 8.5 (SS)
 5.0 - 8.0 (RV)

PGA, PGV, PGD
Distance range:


Pseudo spectral
acceleration at 5%
damping
Period: 0 - 10 sec
PEER Annual Meeting, January 2007
 0 – 200 km
Fault Mechanism:
 Strike-Slip
 Reverse
 Normal
Amplification
(Natural Log)
Depth (m) to Vs=1.5 km/s
NGA Models Were Constrained by Simulation
Period (sec)
To fill the gaps in data
Simulations of 3-D basin and 1-D rock motions
To model amplification due to sediment-depth
To constrain attenuation models
Nonlinear soil response analysis
Amplification factors for different soil profiles
subjected to a wide range of input motions
PEER Annual Meeting, January 2007
Example of Predicted Acceleration Spectra
(Campbell-Bozorgnia)
Strike Slip, RRUP = 10 km, VS30 = 760 m/s
Spectral Acceleration (g)
1
0.1
M=5.0
M=6.0
M=7.0
M=8.0
0.01
0.01
0.1
1
Period (sec)
PEER Annual Meeting, January 2007
10
Behavior at Long Periods
C&B06 Predicted Spectral Displacement
Strike Slip, RRUP = 10 km, VS30 = 760 m/s
Spectral Displacement (cm)
100
10
1
0.1
M 5.0
M 6.0
M 7.0
M 8.0
0.01
0.001
0.0001
0.0
PEER Annual Meeting, January 2007
1.0
2.0
3.0
4.0 5.0 6.0
Period (sec)
7.0
8.0
9.0 10.0
Standard Deviation vs. Magnitude
SA(1.0s), VS30=760
PGA, VS30=760
0.8
Standard Deviation (nat. log)
Standard Deviation (nat. log)
0.8
0.7
0.6
0.5
0.4
0.3
Campbell & Bozorgnia (2003)
Campbell & Bozorgnia (NGA M ay 06)
0.2
4.5
5.0
5.5
6.0
6.5
7.0
Moment Magnitude
PEER Annual Meeting, January 2007
7.5
8.0
8.5
0.7
0.6
0.5
0.4
0.3
Campbell & Bozorgnia (2003)
Campbell & Bozorgnia (NGA M ay 06)
0.2
4.5
5.0
5.5
6.0
6.5
7.0
Moment Magnitude
7.5
8.0
8.5
Comparison with Previous Models
Campbell & Bozorgnia (C&B) 06 vs. C&B03 –
SS, NEHRP B-C
PGA, Strike Slip, VS30=760
SA(1.0s), Strike Slip, VS30=760
0
0
10
10
Acceleration (g)
Acceleration (g)
RRUP=0.1
-1
10
10-2
10
50
4.5
10
10-2
5.0
5.5
6.0
6.5
7.0
7.5
8.0
10
50
Campbell & Bozorgnia (2003)
Campbell & Bozorgnia (NGA M ay 06)
200
10-34.0
RRUP=0.1
-1
Campbell & Bozorgnia (2003)
Campbell & Bozorgnia (NGA M ay 06)
200
10-34.0
8.5
Moment Magnitude
4.5
5.0
5.5
PGA, Strike Slip, VS30=760
6.5
7.0
7.5
0
10
10
10-1
M = 8.0
Campbell & Bozorgnia (2003)
Campbell & Bozorgnia (NGA May 06)
Acceleration (g)
Campbell & Bozorgnia (2003)
Campbell & Bozorgnia (NGA May 06)
10-1
M = 8.0
7.0
10
7.0
6.0
5.0
10-2 1
8.0
SA(1.0s), Strike Slip, VS30=760
0
Acceleration (g)
6.0
Moment Magnitude
100
Closest Distance to Rupture (km)
PEER Annual Meeting, January 2007
5.0
10-2 1
10
6.0
100
Closest Distance to Rupture (km)
8.5
Chiou & Youngs (C&Y) 06 vs. SAO97
SS, NEHRP C
Period = 0.01 (sec); Vs30 = 500 (m/s); SAO97 Rock
7.5
8.5
5.5
6.5
1
1
0.1
0.1
Sa (g)
Sa (g)
5.5
6.5
Period = 1 (sec); Vs30 = 500 (m/s); SAO97 Rock
0.01
0.01
0.001
0.001
1
10
Distance (km)
PEER Annual Meeting, January 2007
1
7.5
8.5
10
Distance (km)
Boore & Atkinson 06 vs. BJF97
NEHRP B-C (Unspecified Mechanism)
PEER Annual Meeting, January 2007
Comparison of Latest NGA Models
PGA – Strike-Slip, M 7.5, NEHRP B-C
Peak Horizontal Acceleration (g)
1
0.1
Magnitude = 7.5 -- pga -- Strike Slip
Abrahamson & Silva (Vs30 = 760 m/s)
Boore & Atkinson (Vs30 = 760 m/s)
Campbell & Bozorgnia (Vs30 = 760 m/s)
Chiou & Youngs (Vs30 = 760 m/s)
Idriss [Vs30 = 450 -- 900 m/s]
0.01
0.1
1
10
Closest Distance (km)
PEER Annual Meeting, January 2007
100
Sa(T=1.0s) – Strike-Slip, M 7.5, NEHRP B-C
Spectral Acceleration (g)
1
0.1
Magnitude = 7.5 -- T = 1 sec -- Strike Slip
Abrahamson & Silva (Vs30 = 760 m/s)
Boore & Atkinson (Vs30 = 760 m/s)
Campbell & Bozorgnia (Vs30 = 760 m/s)
Chiou & Youngs (Vs30 = 760 m/s)
Idriss [Vs30 = 450 -- 900 m/s]
0.01
0.1
1
10
Closest Distance (km)
PEER Annual Meeting, January 2007
100
NGA &
US National Seismic Hazard Maps
USGS is Adopting
NGA relations with documentation:



Boore & Atkinson
Campbell & Bozorgnia
Chiou & Youngs
With equal weighting
PEER Annual Meeting, January 2007
Courtesy: Art Frankel
0.2 sec Spectral Acceleration
2% P.E. in 50 years
Percentage of Change:
Using 3 NGA relations
Versus:
2002 Hazard Maps
Abrahamson and Silva (1997),
Sadigh et al. (1997),
Boore et al. (1997),
Campbell and Bozorgnia (2003),
Spudich et al. (1999) for extensional
areas
Using same set of fault
sources as 2002 maps;
Subduction zone and
deep earthquakes are
not included
Rock site condition
PEER Annual Meeting, January 2007
Courtesy: Art Frankel
1.0 sec Spectral Acceleration
2% P.E. in 50 years
Percentage of Change:
Using 3 NGA relations
Versus:
2002 Hazard Maps
Abrahamson and Silva (1997),
Sadigh et al. (1997),
Boore et al. (1997),
Campbell and Bozorgnia (2003),
Spudich et al. (1999) for extensional
areas
PEER Annual Meeting, January 2007
Reasons…
 “Some of the decrease of 1 sec Sa from the 2002
maps is caused by:
 Difference in the Vs30 assigned for “rock” sites in
the 2002 maps and the average Vs30 for rock sites
reported in NGA (shouldn’t be a factor in the
Campbell-Bozorgnia and Boore-Atkinson NGA
relations)”
 “Most of the decrease is from having additional data
from moderate and large earthquakes and improved
functional forms to fit the data”
PEER Annual Meeting, January 2007
Courtesy: Art Frankel
NGA Models Give Prediction on Geometric
Mean of Two Horizontal Components
Max and Min vs. Geometric Mean
1.5
CB06 - Min/G.M.
CB06 - Max/G.M. (Std. Dev.)
1.4
1.3
1.2
Ratio
You have to convert
these, if you need
prediction on:
Random horizontal
component
Maximum of two
horizontal
components
FN, FP components
1.1
1.0
0.9
0.8
0.7
0.01
Standard deviation can be slightly
larger than that for Geo-Mean
PEER Annual Meeting, January 2007
0.1
1
Period (sec)
10
Summary of NGA Significant
Accomplishments
Quantity and quality of data
Amount of time the developers spent on
models
Interactions among model developers
Number of independent variables
Availability of supporting ground motion
simulations
Public participation via workshops and
conferences
Formal peer review commissioned by USGS
PEER Annual Meeting, January 2007
And, it would have been much more difficult
without a national earthquake engineering center
For NGA, we have had
contracts with:





USGS (different
researchers)
San
Diego
State U.
UC
Santa
Barbara
UC
Berkeley
USGS
Student
U. Nevada
California Geological
Survey
SCEC (various
contracts)
Various universities
Several practitioners
PEER Annual Meeting, January 2007
Private
Industry
Caltrans
USGS
Private
Industry
PEER Ground Motion Selection &
Modification (GMSM) Working Group
Current state-of-the-practice in GMSM
Is confusing
Structural engineers rely on geotechnical
engineers
Geotechs/seismologists have little
understanding on how the selected THs will
be used by structural engineers
GMSM methods have different objectives
Methods can give VERY different results, even
if they have the same objective
PEER Annual Meeting, January 2007
The only possible improvement is:
To assemble a multi-disciplinary team of:



Seismologists
Geotechs
Structural engineers
They interact in frequent meetings &
workshops
They interact with other groups
Come up with an evaluation platform
PEER Annual Meeting, January 2007
PEER GMSM Working Group
PEER Annual Meeting, January 2007
GMSM Working Group Objectives
Examine existing GMSM methods,
Examine emerging GMSM methods,
Cast a platform for “objective”
evaluation and comparison of GMSM
methods
PEER Annual Meeting, January 2007
GMSM Methods to be Tested for
Nonlinear Analysis of:
Buildings
Bridges
Earth structures
Non-structural
PEER Annual Meeting, January 2007
So far
What GMSM Working Group Has Done:
Compiled and classified as
many as GMSM methods as
possible
Compiled existing building
models of:
 Different building sizes
 Different structural
systems
 OpenSees, DRAIN and
PERFORM models of
various buildings
PEER Annual Meeting, January 2007
Inelastic Spectra; Ductility=4
Structural Models Are Subjected to…
Hundreds of input
GMs
 Suppose you have
unlimited
resources
Compute statistics of
various structural
responses (EDPs)
1
0.9
total
conditioned on spectrum
0.8
0.7
0.6
Run the models with
the scaled/modified
GMs
Compare the results
PEER Annual Meeting, January 2007
0.5
0.4
0.3
0.2
0.1
0
0.002
0.01
0.027
0.1
GMSM Working Group Link to
Other R&D Projects
PEER NSF (“core”) research projects

Several of researchers have funding
from PEER NSF program
Tall Buildings Initiative
PEER Annual Meeting, January 2007
Upcoming Tasks in GMSM…
Upcoming tasks:




Other building structural models
Bridge structures
Earth structures
…
We are not done…
PEER Annual Meeting, January 2007
Other PEER GM Projects
Other On-Going GM Projects
Investigation of various
IMs
Strike-Slip Fault, Mw=7.5, RRUP=0.1, Vs30=760, Z2.5=2
1.2
Attenuation of inelastic
spectra
Collaboration with
PG&E/DOE on “Extreme
Ground Motion” at
Yucca Mountain nuclear
waste repository
More…
PEER Annual Meeting, January 2007
1.0
0.8
C y = Fy / W
Input motion for tall
buildings with large
embedded structure
= 1
=2
=4
=6
=8
0.6
C&B NGA Elastic
(Dec. 2006)
0.4
0.2
0.0
0.0
1.0
2.0
3.0
4.0
Period (sec)
5.0
6.0
Future GM Tasks and Projects
Maintaining and Updating GM Database
Processing records from recent EQs
Adding metadata for the EQs and sites
Adding “engineer-friendly” features to
the database
Adding utility software for engineering
applications

e.g., a module for selection and scaling
PEER Annual Meeting, January 2007
Upcoming GM Tasks
Attenuation models
 Incorporation of directivity into NGA
 Attenuation of vertical ground motion
 Completion of inelastic spectra
GM Selection and Modification:
 Various structural models
 Guidelines for tall buildings
PEER Annual Meeting, January 2007
An Important Issue to be Considered:
Near-fault Effects
2003 Bam EQ
0
-1
15
16
17
18
19
20
21
22
23
24
25
3
Spectral Acceleration (g)
Example: in CA,
there are 8,386
bridges within 10
km of faults
Accel. (g)
1
2
1
0
0.0
Technical aspects
of the issue:


Seismological
Engineering
implications,
especially for
nonlinear response
PEER Annual Meeting, January 2007
0.5
1.0
1.5
2.0
2.5
3.0
Possibly we have to initiate
a working group with focus
on “Near-Fault GM
Characterization and
Effects”
Thank You !