Geotechnical Earthquake

Engineering
by
Dr. Deepankar Choudhury
Humboldt Fellow, JSPS Fellow, BOYSCAST Fellow
Professor
Department of Civil Engineering
IIT Bombay, Powai, Mumbai 400 076, India.
Email: [email protected]
URL: http://www.civil.iitb.ac.in/~dc/

Lecture – 32
 Module – 8

Site Response Analysis

IIT Bombay, DC 2
 Site Response - The Problem
Predicts the response of a soil deposit due to earthquake
excitation

IIT Bombay, DC 3
 Site Response
Ideally, a complete ground response analysis should
include:

 Rupture mechanism at source of an earthquake

(source)
 Propagation of stress waves through the crust to the
top of bedrock beneath the site of interest (path)
 How ground surface motion is influenced by the soils
that lie above the bedrock (site)
IIT Bombay, DC 4
 Site Response
In Reality,

 Mechanism of fault rupture is very complicated and

difficult to predict in advance
 Crustal velocity and damping characteristics are generally
poorly known
 Nature of energy transmission between the source and site
is uncertain

IIT Bombay, DC 5
 Ground Response Analysis
In practice,

 Seismic hazard analyses (probabilistic or

deterministic) are used to predict bedrock motions at the
location of the site.

 Seismic hazard analyses rely on empirical attenuation

relationships to predict bedrock motion parameters.

 Ground response problem becomes one of determining

response of soil deposit to the motion of the underlying
bedrock.
IIT Bombay, DC 6
 Ground Response Analysis
Used to:

 Predict ground surface motions, Time histories, Response

spectra, Scalar parameters

 Evaluate dynamic stresses and strains, Liquefaction

hazards, Foundation loading

 Evaluate ground failure potential, Instability of earth

structures, Response of various geotechnical structures like
retaining wall, earth dam, pile, various foundations etc.
IIT Bombay, DC 7
 Ground Response Analysis
Definitions:

Rock outcropping motion the motion that would occur

where rock outcrops at a surface.

IIT Bombay, DC 8
 Ground Response Analysis
Definitions:
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.

IIT Bombay, DC 9
 Ground Response Analysis
Definitions:

Free surface motion – the motion that occurs at the surface of a

soil deposit

IIT Bombay, DC 10
 Ground Response Analysis
Common situations # 1

Rock outcrop motion is known – usually obtained from attenuation

relationship (based on database of rock outcrop motions)

Free surface motion is to be determined

IIT Bombay, DC 11
 Ground Response Analysis
Common situations # 2

Free surface motion is known – usually obtained from attenuation

relationship (based on database of soil outcrop motions)

Free surface motion is to be determined for site with different soil

conditions

IIT Bombay, DC 12
Ground Response Analysis

IIT Bombay, DC 13
 Linear Analysis
 A known time history of bedrock (input) motion is
represented as a Fourier series, usually using the FFT

 Each term in the Fourier series of the bedrock (input)

motion is then multiplied by the transfer function to
produce the Fourier series of the ground surface (output)
motion

 The ground surface (output) motion can then be expressed

in the time domain using the inverse FFT

IIT Bombay, DC 14
 Transfer Function
 The transfer function determines how each frequency in the
bedrock (input) motion is amplified, or de-amplified by the soil
deposit

 A transfer function may be viewed as a filter that acts upon

some input signal to produce an output signal

Transfer
Input function Output
(filter)

IIT Bombay, DC 15
Transfer Function Evaluation
Uniform Undamped Soil on Rigid Rock

Assume harmonic base motion,

Then, response should also be harmonic

Wave traveling in Wave traveling in

– z direction + z direction
(upward) (downward)
IIT Bombay, DC 16
 Transfer Function Evaluation
Uniform Undamped Soil on Rigid Rock

Displacement :

Stress:

At z = 0 (ground surface)
A=B

IIT Bombay, DC 17
 Transfer Function Evaluation
Uniform Undamped Soil on Rigid Rock

eikz e ikz

u z, t 2A ei t
2

i t
u z, t 2 A cos kz e

Defining a transfer function as the ratio of the displacement at

the ground surface to the bedrock displacement

u 0, t 2 Aei t 1
F
u H ,t 2 A cos kH ei t
cos kH

IIT Bombay, DC 18
 Transfer Function Evaluation
Uniform Undamped Soil on Rigid Rock

As kH = wH/Vs goes to zero, denominator goes to zero Transfer

function goes to infinity

Natural frequencies Fundamental period

VS n /H TS 2 4H
n 2 0 VS

IIT Bombay, DC 19
 Transfer Function Evaluation
Uniform Damped Soil on Rigid Rock

How do we handle damping?

Complex shear modulus
* * 2 * 2
G V
S /k
* 2 * 1/2
k /G Complex Wave
Number

* * * w
V S G / VS 1 i k k 1 i
VS*
IIT Bombay, DC 20
 Transfer Function Evaluation
Uniform Damped Soil on Rigid Rock

Repeat analysis as before

Transfer function becomes
u 0, t 2 Aei t 1
F
u H,t 2 A cos k * H ei t
cos k * H

IIT Bombay, DC 21
 Transfer Function Evaluation
Uniform Damped Soil on Rigid Rock

1 1
F w,
cos k * H
cos wH
V *S

1 1
F w,
2 2 2 2
cos kH kH cos wH / VS wH / VS

IIT Bombay, DC 22
 Transfer Function Evaluation
Uniform Damped Soil on Rigid Rock

Note:
Natural frequencies still exist
Low natural frequencies strongly amplified
High natural frequencies weakly amplified
Very high frequencies de – amplified
Amplification strongly frequency - dependent
IIT Bombay, DC 23
 Transfer Function Evaluation
Uniform Undamped Soil on Elastic Rock

us
i t ks*t i t k s*t
us zs , t Cs e Ds e
i t kr*t i t kr*t
ur z r , t Cr e Dr e
us z s H ,t ur z r 0, t

s zs H,t r zr 0, t
IIT Bombay, DC 24
 Transfer Function Evaluation
Uniform Undamped Soil on Elastic Rock
Maintaining equilibrium and compatibility of displacements
at the boundary, the amplitude of the transfer function can be
written as
1
F w, 0
2 2 2
cos ks H z sin ks H
*
v
s ss
z *
v
r sr

IIT Bombay, DC 25
 Transfer Function Evaluation
Uniform Undamped Soil on Elastic Rock
Note:
Even with no soil damping, resonance cannot occur
Why???
Energy removed from soil layer by transmission into rock
Form of radiation damping

IIT Bombay, DC 26
 Transfer Function Evaluation
Uniform Undamped Soil on Elastic Rock
Maintaining equilibrium and compatibility of displacements
at the boundary, the amplitude of the transfer function can be
written as
1
F w, 0
2 2 2
cos ks H z sin ks H
*
v
s ss
z *
v
r sr

IIT Bombay, DC 25
 Transfer Function Evaluation
Uniform Undamped Soil on Elastic Rock
Note:
Even with no soil damping, resonance cannot occur
Why???
Energy removed from soil layer by transmission into rock
Form of radiation damping

IIT Bombay, DC 26
 Transfer Function Evaluation
Layered, Damped Soil on Elastic Rock

IIT Bombay, DC 27
 Transfer Function Evaluation
Layered, Damped Soil on Elastic Rock
For layer j
ik *j z j ik *j z j
uj z j ,t Aj e B je ei t

From equilibrium
ik *j h j ik *j h j
Aj 1 Bj 1 Aj e Bje

From compatibility
G*j k *j iks*h j iks*h j
Aj 1 Bj 1 * *
Aj e Bje
G kj 1 j 1

IIT Bombay, DC 28
 Transfer Function Evaluation
Layered, Damped Soil on Elastic Rock
ik *j z j ik *j z j
For layer j uj z j ,t Aj e B je ei t

ik *j h j ik *j h j
Aj 1 Bj 1 Aj e Bje
From equilibrium

G*j k *j iks*h j iks*h j

From compatibility Aj 1 Bj 1 * *
Aj e Bje
G k j 1 j 1

If we know response at layer j (Aj and Bj are known), then

we have two equations with two unknowns (Aj+1 and Bj+1)
We can relate Aj+1 and Bj+1 to Aj and Bj by means of
recursive relationships

IIT Bombay, DC 29
 Transfer Function Evaluation
Layered, Damped Soil on Elastic Rock
Solving for the unknowns
1 * ik *j h j 1 * ik *j h j
Aj 1 Aj 1 j e Bj 1 j e
2 2

1 * ik *j h j 1 * ik *j h j
Bj 1 Aj 1 j e Bj 1 j e
2 2

Or, relating the coefficients to those at the ground surface

Aj 1 aj 1 A1 Bj 1 bj 1 B1

IIT Bombay, DC 30
 Transfer Function Evaluation
Layered, Damped Soil on Elastic Rock
Then, a transfer function relating the motion in layer i to
the motion in layer j can be written as

ai bi
Fij
aj bj

If we know the motion at any layer, we can use this

transfer function to compute the corresponding motion at
any other layer

IIT Bombay, DC 31
 Equivalent Linear Approach
The actual nonlinear hysteretic stress – strain behavior of
cyclically loaded soils can be approximated by equivalent linear
properties

Assume some initial strain and use to

estimate G and ξ
Determine peak strain and effective strain eff R max

IIT Bombay, DC 32
 Select properties based on updated strain level

Compute response with new properties and determine

resulting effective shear strain
IIT Bombay, DC 33
Repeat until computed effective strains are consistent with
assumed effective strains

IIT Bombay, DC 34
 Non – linear Approach
Solve Wave equation incrementally
2
u u
2
z t t
Approximate partial derivatives
i 1,t i ,t u ui ,t t ui ,t
z z t t
Finite difference form
i 1,t i ,t ui 1,t ui ,t
z t

IIT Bombay, DC 35
 Non – linear Approach
Solve Wave equation incrementally
then
t
ui ,t t ui ,t i 1,t i ,t
z

Velocity at time t+Δt can be calculated from velocity

and shear stress at time t

IIT Bombay, DC 36
 Non-Linear Approach
Solve wave equation incrementally

Start with initial stiffness, Gmax

Compute response for small time step, Δt
Compute shear strain amplitude at end of time step
Use stress-strain model to find Gtan for next time step
Compute shear strain amplitude at end of next time step
Continue stepping through time for entire input motion

IIT Bombay, DC 37
 Non – linear Approach
Solve Wave equation incrementally

Nonlinear response is simulated in

incrementally linear fashion

Material damping is taken care by

hysteretic response

Approach requires good model for

description of soil stress – strain
behaviour

IIT Bombay, DC 38
Non-Linear Stress-Strain Models
Two main types

Cyclic nonlinear models

Advanced constitutive models

IIT Bombay, DC 39
 Non – Linear Stress – Strain Models
Cyclic nonlinear models

Requires :
• Backbone curve
• Unloading – reloading rules
• Pore pressure model

IIT Bombay, DC 40
 Non – Linear Stress – Strain Models
Advanced constitutive models

Requires:
•Yield surfaces
•Hardening rule
•Failure surface
•Flow rule

IIT Bombay, DC 41
 Non-Linear Stress-Strain Models

Cyclic nonlinear models

Advantages:
Relatively simple
Small number of parameters
Disadvantages:
Simplistic representation of soil behavior
Cannot capture dilatancy effects

Advanced constitutive models

Advantages:
Can better represent mechanics of yield, failure
Disadvantages:
Many parameters
Difficult to calibrate
IIT Bombay, DC 42