predicting site response. based on theoretical calculations –1-d equivalent linear, non-linear...

68
Predicting Site Response

Upload: amice-andrews

Post on 13-Jan-2016

216 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Predicting Site Response

Page 2: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Predicting Site Response

• Based on theoretical calculations– 1-D equivalent linear, non-linear– 2-D and 3-D non-linear

• Needs geotechnical site properties

Page 3: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

THEORETICAL PREDICTION OF SITE RESPONSE

Page 4: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Site Response

• Observations of Site Response and Implications for Predicting Response (with a major emphasis on variability of site response)

• Methods for Predicting (Estimating) Response• Other topics

– Nonlinear response– Spatial variability– Topographic amplification

Page 5: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Predicting Site Response

• Empirical studies and blind prediction experiments lead to the conclusion that site-specific, earthquake-specific predictions may be very uncertain.

• Despite this pessimistic conclusion, I will now proceed to discuss site-specific, earthquake-specific predictions.

• But always keep in mind the possible uncertainties in the predictions of site response

Page 6: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Predicting Site Response

• Theoretical• Full resonance• Simplified (square-root impedance)

Page 7: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

0 10 20 30 40-600

-400

-200

0

200

400

600

Time (sec)

Acc

ele

ratio

n(c

m/s

ec2

)

Surface Motion

0 10 20 30 40-600

-400

-200

0

200

400

600

Time (sec)

Acc

ele

ratio

n(c

m/s

ec2

)

Input Motion

File

:C

:\m

etu

_0

3\r

ec_

pro

c_st

ron

g_

mo

tion

\site

_re

spo

nse

_e

xam

ple

_4

pp

t.d

raw

;D

ate

:2

00

3-0

9-1

5;

Tim

e:

10

:41

:13

1 2 3 4 5 6 70

50

100

150

200

250

Slowness (msec/m)

De

pth

(m)

Slowness Model

0.01 0.1 1 10 1000.4

1

2

3

Frequency (Hz)

Site

Re

spo

nse

Transfer Function

100 200 300 400 500 600 7000

50

100

150

200

250

Velocity (m/sec)

De

pth

(m)

Velocity Model

Predicting site response: given input motion and velocity model, predict surface motion

Page 8: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Theoretical basis to understand site effects

• Layer over a half space• Multiple layers• Effects of basins• Effects of topography• Effects of manmade structures

Page 9: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Site amplification due to resonance effects

• Trapped waves reverberate due to multiple reflections

• Constructive interference causes resonance, which depends on thickness of layer and elastic properties

• Can occur even if NO discontinuities in seismic impedances

Page 10: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Resonance effects (cont.)• Attenuation in layer modifies resonance,

especially at high frequencies• Typically see fundamental resonance and

perhaps one more peak (higher frequency peaks damped by attenuation)

Page 11: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties
Page 12: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties
Page 13: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Use program nrattle in site_amp collection of programs on my web site

Page 14: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Multiple flat layers

• Haskell (1960’s) found an exact solution for the effect of multiple layers on a vertically incident S-wave.

• Subsequently improved for greater numerical stability.

• Each layer boundary causes some amplification. Net effect can be much greater than a factor of two.

• Multiple resonances exist.

Page 15: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

In some cases the velocity profile is more of a gradient with no significant jumps in impedance. How is the amplification computed in such situations?

Page 16: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

The gradient can be replaced by a stack of constant velocity layers, with thicknesses and velocities chosen so that the travel time through the profile is the same as for the continuous profile (using my program site_amp is a convenient way of doing this).

Page 17: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Full resonance amplifications can be obtained for the model made up of a stack of constant velocity layers (using nrattle in this case; note that the amplifications approach unity for low frequencies)

There is another method for approximating the amplification, which I now discuss.

Page 18: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Impedance effects• Seismic Impedance (Z) = velocity * density• Conservation of energy along a ray tube due to impedances predicts an

amplification of

where ZR is the impedance at the source and Z is the average impedance at points along the ray tube

• Shear-wave velocity and density increase with increasing depth in the crust• Typical velocity and density at seismogenic depths (10km) are 3.6 km/sec

and 2.8 gm/cm**3• Typical velocity near surface depends on NEHRP site class • Rock sites range from 2.8 km/sec (ENA hard rock) to as low as 600 m/sec

(Calif. Soft rock)• Thus all seismic waves will be amplified by impedance effects as they

travel to the surface

RA Z Z

Page 19: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Steps in computing square-root-impedance amplifications:

Obtain velocity-depth profile:

Page 20: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Steps in computing square-root-impedance amplifications:

Compute travel time as a function of depth:

Page 21: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Steps in computing square-root-impedance amplifications:

For each depth, compute the average velocity from the surface to that depth:

( )SZV z tt z

I use z=0.4 km only as an example; the procedure is repeated for all depths

Page 22: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Steps in computing square-root-impedance amplifications:

• Compute the average velocity VSZ for all depths z:

Page 23: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Steps in computing square-root-impedance amplifications:

• For each depth z, replace the real velocity with a layer of velocity equal to the average velocity to that depth (the example here is for z=0.4 km):

Page 24: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Steps in computing square-root-impedance amplifications:

• Now use the square-root impedance method to compute an amplification for this new model. The amplification will only be for the frequency corresponding to the resonant frequency of the constant velocity layer

Page 25: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

4SVfH

Recall the equation for resonance frequency of a single layer of thickness H and shear-wave velocity VS

This is a quarter-wavelength condition, as it states that resonance will occur for a period for which the wavelength (VS/f) is 4 times the layer thickness

Page 26: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Now replace the material down to depth z with a layer of constant velocity VSZ. Then the quarter-wavelength period for this layer will be

1 4 4SZVfz

Thus for each depth z we can replace the material above that depth with a constant velocity layer with velocity VSZ and associate a quarter-wavelength frequency f1/4λ with each depth. The next slide shows Vs, VSZ, and f1/4λ as a function of depth z.

For the example being used (z=0.4), the value is 1.45/(4*0.4)=0.9 Hz

Page 27: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties
Page 28: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

We can also estimate the amplification for this new velocity model for each depth z, using the square-root impedance (SRZ) amplification:

where the impedance is given by:

“R” indicates a reference depth, and the overbar indicates that the impedance and density are averaged over the depth z. Because we also can associate a quarter-wavelength frequency with each depth, we then obtain an approximation of amplification as a function of frequency.

( ) ( )RA z Z Z z

SZZ V

Page 29: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

For the example (z=0.4 km), the amplification is sqrt(3.5/1.45)=1.55, ignoring changes in density. Here is the square-root-impedance amplification for the BJ97 generic rock velocity model (which does include the density variation).

Page 30: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Here is the square-root-impedance amplification for the BJ97 generic rock velocity model, compared to full resonance amplifications

Page 31: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Square-root impedance method (aka Quarter-wavelength method)

• Advantages – Simple, quick– No assumptions needed about deeper velocities

(except for source region)– Provides smooth version of amplification

• Uses– amplification at “generic” sites – First approximation/ guidance– Assessing depths that control amplification– Comparing velocity profiles

Page 32: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Square-root impedance method (aka Quarter-wavelength method)

• Disadvantages – Underpredicts resonance peaks (but maybe OK for deep

sediment-filled basins, for which fundamental mode is lower than frequencies of most engineering frequencies)

– Purely linear calculations

Page 33: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Now compare amplifications for some simple models.

All models have the same travel time to 37.5 m

Use nrattle for full resonance amps and site_amp for square-root-impedance (SRI) amps

Page 34: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Note shift of first resonant peak

SRI amps always underestimate fundamental mode, but provide average of higher mode amplifications

Page 35: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

0 20 40 60

Q

0 2000 40000

200

400

600

Shear Velocity (m/s)

De

pth

(m)

File

:C

:\m

etu

_0

3\r

ec_

pro

c_st

ron

g_

mo

tion

\BJ_

19

91

_d

ee

pso

il_ve

l.dra

w;D

ate

:20

03

-09

-15

;Tim

e:

11:2

2:2

2

Stringent conditions are required for perfect constructive and destructive interference (the essence of resonant response)

Page 36: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

0.1 0.2 1 2 100

1

2

3

4

Frequency (Hz)

So

ilR

esp

on

se

from square-root impedance method

d

1b

2b

> 10b

0 5 10Time (sec)

Impulse Response

d

1b

2b

5b

File

:C

:\m

etu_

03\r

ec_p

roc_

stro

ng_m

otio

n\sp

ect_

reso

nanc

e_bu

ildup

.dra

w;

Dat

e:20

03-0

9-15

;Ti

me:

11:0

1:24

The amplitude spectrum of a resonant systemrequires precise constructive interference of multiple arrivals. In practice this probably will not happen because of geologic complexity or nonlinear changes in seismic velocity, and thus observations may not show a well-defined set of resonant peaks.

Page 37: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Nonlinear response

• Amplitude-dependent amplification• Shift of resonant peaks to lower frequencies with

increased shaking• Loss of high frequencies in records• Gain of high frequencies in records (later cusp-like

arrivals)

Page 38: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Nonlinear soil behaviour:If soil is linear it obeys Hooke’s Law:

stress = G * strainwhere G=shear modulus

But real soil shows hysteretic behaviour under high strains

In hysteretic behaviour, loading and unloading curves are not the same (energy is lost). Effective slope G decreases at high strain. Since the shear-wave velocity is given by vs=sqrt(G/density), this means Vs decreases with strain.

Page 39: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Consequences of nonlinearity

• Vs decreases at large strains• Resonance frequency decreases, since f=Vs/4H• Amplification decreases due to damping (loss of energy) in

hysteresis loop (proportional to loop area)• Thus nonlinearity will result in lower soil amplification, and a

shift to lower frequencies• There can also be a shift of energy to higher frequencies

Page 40: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Nonlinear site response

• Under weak motions, to a good approximation, the stress-strain relationship in the soil is linear.

• Under strong motion, that linearity fails.• A common approximation to the effect of

nonlinearity is to decrease the velocity where the material is nonlinear, and increase the energy loss Q.

Page 41: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

“Equivalent Linear” Soil Response (SHAKE)

• The world’s most commonly used program for approximating nonlinear soil response

• Highly recommended: Strata, free from http://nees.org/resources/strata

• Strata can compute equivalent linear amplifications both using time-domain and frequency-domain calculations.

Page 42: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties
Page 43: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Equivalent Linear

• Use G/Gmax and damping vs. strain to approximate nonlinear response

• Do iterative linear analysis, adjusting the modulus and the damping to be consistent with the strain in various layers

Page 44: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Equivalent Linear• Consequences

– Shifts resonant peaks to longer periods– Large attenuation of high frequencies

• Strengths and limitations– can use RVT to avoid needing a suite of input time

series– Fast– OK for period range of most interest– Too much attenuation at short periods (some people

make the damping vs. strain period dependent to overcome this)

Page 45: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Example

• Use a 1 layer case• 1940 El Centro motion as input, scaled to

0.01g and 1.0g

Page 46: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

0.01 0.1 1 10

1

2

3

4

Per (sec)

Am

p

1 layer

File

:C

:\ro

se\la

bs\1

0_si

te_a

mps

\nra

t_1l

amps

_log

.dra

w;

Dat

e:20

05-0

5-06

;Ti

me:

13:3

9:11

Fundamental mode 0.5 sec

Linear amplification as function of period period (not frequency)

Page 47: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Modulus reduction and damping curves

10-4 0.001 0.01 0.1 1 10

0

0.2

0.4

0.6

0.8

1

Strain (%)

G/G

max

10-4 0.001 0.01 0.1 1 10

0

5

10

15

20

25

30

Strain (%)

dam

ping

(%)

File

:C

:\edu

shak

e\R

UN

1LM

D.d

raw

;D

ate:

2005

-05-

06;

Tim

e:13

:26:

27

Page 48: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Effective strain vs. depth

10-4 0.001 0.01 0.1 1 10

0

5

10

15

20

25

30

Effective Strain (%)

Dep

th(m

)

input: 0.01ginput: 1.00g

File

:C

:\ed

usha

ke\r

un1l

_str

ain_

vs_d

epth

.dra

w;

Dat

e:20

05-0

5-06

;Ti

me:

13:0

7:04

Page 49: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

G/Gmax.& D values for strains obtained during soil response

10-4 0.001 0.01 0.1 1 10

0

0.2

0.4

0.6

0.8

1

Strain (%)

G/G

max

10-4 0.001 0.01 0.1 1 10

0

5

10

15

20

25

30

Strain (%)

dam

ping

(%)

input: 0.01g

input: 1.00g

File

:C

:\edu

shak

e\R

UN

1LM

D_w

ith_d

ots.

draw

;D

ate:

2005

-05-

06;

Tim

e:13

:15:

01

Page 50: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

0.01 0.1 1 10

0.01

0.1

1

T (sec)

PS

A(g

)

soil surface: 1.0g inputsoil surface: 0.01g inputrock: input spectrum

period shift

File

:C

:\ed

usha

ke\r

un1l

_rs.

draw

;D

ate:

2005

-05-

06;

Tim

e:13

:52:

24

Input and surface response spectra. Note differences at high frequency for low- and high-strain cases and shift of resonant peak to longer period for high strain case

Page 51: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Amplification

0.1 0.2 1 20.1

0.2

1

2

T (sec)Am

plif

ica

tion

,re

lativ

eto

inp

ut

as

ab

ed

rock

ou

tcro

p

Input: 0.01gInput: 1.0g F

ile:

C:\

ed

ush

ake

\ru

n1

l_rs

_a

mp

.dra

w;

Da

te:

20

05

-05

-25

;T

ime

:0

9:4

9:3

0

A limitation of method: severe overdamping of short period motions often occurs (because strain is determined by longer periods, but the strain-dependent damping applies at all periods). Using period-dependent modulus and damping values is an attempt to retain advantages of the equivalent linear approach, without overdamping. But some cases a truly nonlinear calculation is necessary.

Page 52: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Truly Nonlinear• Time-domain integration, following stress-strain

relation• A number of programs, based on different

assumptions and requiring varying amounts of detail regarding the soil behavior– Pore pressure effects accounted for in some codes

Page 53: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties
Page 54: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Empirical evidence of soil nonlinearity

• Liquefaction• Soil-vs-rock differences in spectra and time series• Period shifts and high frequency spikes

Page 55: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Evidence of liquefaction

Page 56: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

0 10 20 30 40-200

-100

0

100

200

Acc

eler

atio

n(c

m/s

ec2)

1989 Loma Prieta earthquake

0 10 20 30 40

-50

0

50

Time (sec)

Acc

eler

atio

n(c

m/s

ec2)

File

:C

:\met

u_03

\rec

_pro

c_st

rong

_mot

ion\

tri_

ybi.d

raw

;D

ate:

2003

-09-

17;

Tim

e:10

:36:

10

Treasure Island (EW)

Yerba Buena Island (EW)

• At Treasure Island, relative to YBI:

• loss of high frequencies• increased amplitudes

Page 57: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

20 30 40 50-300

-200

-100

0

100

200

300

Acc

ele

ratio

n(c

m/s

ec2

)

2001 Nisqually, Washington, earthquake (M 6.8)

20 30 40 50

-50

0

50

20 30 40 50-300

-200

-100

0

100

200

300

Time (sec)

Acc

ele

ratio

n(c

m/s

ec2

)

20 30 40 50

-50

0

50

Time (sec)

File

:C

:\m

etu

_0

3\r

ec

_p

roc

_s

tro

ng

_m

oti

on

\sd

s_

sd

n_

un

filt

_fi

lt.d

raw

;Da

te:2

00

3-0

9-1

7;

Tim

e:

11:0

4:4

1

SDS (NS): unfiltered SDS: filtered from 10--20 Hz

SDN: filtered from 10--20 HzSDN (NS): unfiltered

• SDS within 200 m of SDN• liquefaction at SDS, not at SDN• Note cusps at SDS and increased amplitude at high frequencies

(after Frankel et al., 2002)

Page 58: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

0.1 0.2 1 2 10 200

5

10

15

Frequency (Hz)

Sp

ect

ral

Ra

tio

M 6.8 mainshockM 3.4 aftershock

SDS (geometric mean, horizontal components)

File

:C

:\m

etu_

03\r

ec_p

roc_

stro

ng_m

otio

n\sd

s_ra

tio_m

s_as

_log

x_lin

y.dr

aw;

Dat

e:20

03-0

9-17

;Ti

me:

14:5

5:44

Note shift of amplification peaks for the mainshock to lower frequencies and the increase in high frequency motion for the mainshock.

(after Frankel et al., 2002)

Page 59: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Effects of Topography

• Surface (amplification, deamplification)– Ridges, peaks– Canyons

• Buried (focusing, defocusing)– Example from Northridge earthquake (Frankel et

al.) (not shown here)

Page 60: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties
Page 61: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties
Page 62: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

(Boore, 1972)

Page 63: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

(Boore, 1972)

Page 64: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Basin Effects

• Sediment filled basins• Trapped body waves • Surface waves generated at basin edges• Increased duration• Two-dimensional models

Page 65: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

Synthetic basin models

• Note surface waves generated at the basin edge.• Duration of shaking in the basin is greatly increased.

Page 66: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties
Page 67: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties
Page 68: Predicting Site Response. Based on theoretical calculations –1-D equivalent linear, non-linear –2-D and 3-D non-linear Needs geotechnical site properties

End