Is VS30 an Effective Parameter for Site Characterization? Norm Abrahamson PG&E

Transcription

1 Is VS3 an Effective Parameter for Site Characterization? Norm Abrahamson PG&E

2 Before VS3 Generic Site Categories Used in Western U.S. GMPEs Soft soil Soil Requires special consideration Deep soil excluding soft-soil Rock Not deep soil Sometimes separated into hard-rock and soft-rock

3 VS3 VS3 is not the fundamental physical parameter. Index of the velocity profile Provides a quantitative parameterization of site Less subjective than site categories Provides a smooth transition between site categories For typical strong motion sites in California, VS3 correlated with deeper Vs profile Most soil sites are in alluvial basins (deep soils) Need soil depth for shallow soil sites

4 VS3=3 Correlation of VS Profile and VS3 VS3=8

5 Site Amplification (LN Units) Does VS3 Work? PGA Amplification T=. Amplification Vs3 (m/s) - 3 Vs3 (m/s) T=.5 Amplification T=.75 Amplification Site Amplification (LN Units) Vs3 (m/s) - 3 Vs3 (m/s)

6 Site Amplification (LN Units) Does VS3 Work? T=. Amplification T=.5 Amplification Vs3 (m/s) - 3 Vs3 (m/s) T=. Amplification T=3. Amplification Site Amplification (LN Units) Vs3 (m/s) - 3 Vs3 (m/s)

7 Does VS3 Work? Site Amplification (LN Units) T=5. Amplification T=8. amplification Site Amplification (LN Units) Vs3 (m/s) Vs3 (m/s) B PGV amplification B B B B B B BB B B B B B B B B B B B B B B B Vs3 (m/s) Vs3 (m/s) B PGD amplification B

8 Does VS3 Work? Short periods V S3 captures general site amplification for soil and soft-rock sites Extrapolating V S3 scaling to hard-rock sites may not work Does not capture effects of kappa Most GMPEs put an upper limit on the VS3 used (-5 m/s) Long periods V S3 captures general amplification for soil sites V S3 is not correlated with long period amplification for rock sites

9 How Well Does VS3 Work? At long periods, explains factor of -3 differences between soil and rock sites Look at the remaining variability of the site response after using VS3 Components of variability

10 Components of Ground Motion Variability Sa(M i,r ik,vs3 k )= Site effect not explained by VS3 F(M i,r ik,vs3 k )+"B i + "WS ik + ˆ S k " = # + $ SS + $ S Between Event Within Event

11 Components of Variability from Taiwan Data σ τ φ S φ SS PGA T= T= T= T= T= LN units (From Lin et al, )

12 Comparison of φ S based on V S3 California Taiwan PGA.44.8 T=..36 T= T=.5.3 T= T=3..43 LN units

13 Beyond VS3 Add soil depth ( parameters to VS profile) Allows separation of shallow and deep soil sites Add Kappa Allows extrapolation of VS3 scaling to hardrock sites Only inputs the high frequencies (> Hz) Add site period Difficult, sites may have more than one site period (shallow and deep parts of the profile)

14 Effect of Soil Depth AS8 model V S3 =7 m/s

15 Plan for Update of AS8 NGA In addition to VS3: (NGA-west) Soil depth (depth to VS= km/s) Kappa Difficulty in estimating additional parameters Default values will be provided if additional parameters are not known.

16 Use and Misuse of VS3 Intended Use of VS3 in NGA models Clear hand-off between ground motion and site response for definition of the input motions Consistent definition of rock or input motion VS3 used in the GMPE should be for the 3 m below the level of the input motion Use for deep soil sites that have typical profiles e.g deep soil site in LA basin Misuse Replace site-specific analysis (for sites with profiles not typical as contained in GM data base) VS3 was never intended to provide a site-specific site amplification

17 Conclusions V S3 remains a useful parameter for general site classification Better than the old categories of soil and rock Less ambiguous Avoids jumps in ground motion from one category to the next) Allows for defining motion for the appropriate layer for the input motion I have not seen an other single parameter that works better

18 Conclusions (cont) Additional site parameters can be added to GMPEs Should address common conditions with large effects Shallow soils (not deep alluvial basins) Depth to rock Hard rock conditions (e.g. EUS) Kappa Adding a few additional site parameters will not replace site-specific response In my opinion, it is a mistake to try to build detailed site-specific effects into GMPEs This should be handled by site-specific analysis

19

20 Comparison of φ S based onv S3 California Japan Taiwan PGA T= T= T= T= T=3..43