Seismic Design of Piled Raft

Supported Structure:
Probabilistic Approach

Unfortunately, soils are made by nature

and not by man, and the products of
nature are always complex As soon as
we pass from steel and concrete to earth,
the omnipotence of theory ceases to
exist. Natural soil is never uniform. Its
properties change from point to point
while our knowledge of its properties are
limited to those few spots at which the
samples have been collected
~Karl Terzaghi (1936)
2

Contents
Introduction
Design methodologies
Influence of variability
Foundation design components
Concept of probabilistic analysis
Seismic design of piled raft supported structure
Codal provisions
Incorporation of uncertainty
Case example

Concluding remarks
Future scope

Introduction

Courtesy: Lacasse 2004

El Salvador earthquake

Courtesy: Lacasse 2004

Courtesy: Lacasse 2004

Failure due to a combination

of factors: Gaustad slide, Oslo

What went wrong ?

Failure due to extreme value of single
parameter
Failure due to gross error
Communication (complex project
organisation)

Design Methods
Concept of Load and Resistance
Resistance : Pile load carrying
capacity

Working Stress Design (WSD)

Courtesy: Becker 1996a

Limit State Design

(LSD)

Ultimate Limit State (ULS)

Serviceability Limit State (SLS)

Courtesy: Becker 1996a

10

Courtesy: Becker 1996a

Ultimate Limit State

11

LSD: Eurocode 7 (2004)

Partial factors for resistance

Partial factors for load

12

Influence of variability in
Geotechnical design
Tejchman and Gwizdala (1988)

Allowable axial load varies from

3330 kN-5200 kN
Wide variation of allowable load -

though from the same test site

Attributed to the inherent

variability of soil - due to geologic

process

4 pile axial load tests in a same site

Pile length = 12m,
Pile diameter = 1.5m
Silty medium dense to dense sand

13

Foundation Design Components

(Becker 1996b)

DESIGN CRITERIA
(CLIENT NEEDS)
PROJECT DESCRIPTION

How much
we are sure?

DESIGN ISSUE

SAFETY
FACTORS

Any
quantification
tools?

LOAD EFFECTS

CALCULATION MODEL
Rn = cNc + 0.5BN + qNq

ENGINEERING
EXPERIENCE AND
JUDGEMENT

How many design

offices use it
appropriately?

ET
AT
IO
N

SITE INVESTIGATION
REQUIREMENTS

IN
TE
RP
R

DESIGN DECISION

FoS without knowing

the sources and
amount of
uncertainty involved
in the evaluation of
design parameters

GEOTECHNICAL ANALYSIS FOR

CAPACITY AND SETTLEMENT
GEOTECHNICAL MODEL

Model Uncertainty

14

How to reduce risk?

Reducing gross error
Improving quality control

Extreme Value
Threshold level based on experience
Trade-off safety and cost

Combined effects
Identify most frequent cause of failure
Use reliability analysis and risk assessment
15

Concept of Limit State

X : Load
Y : Resistance

Or,

The failure surface or the limit state is

defined as the boundary between safe
and unsafe regions
16

Probability of Failure
Mean simplified Profile

FOS = R / L

R1

Variable Strength Profile

R1
R = R1 + R2

R2

Performance Function : G = R (Resistance) L

(Load)
Probability of Failure event : pf = P(R<L)

PD
F

Mean Resistance = 80
Standard dev. = 30
15

PD
F

Mean Load = 20
Standard dev. = 20
10

pf = 0.0005
pf = 0.05

Safety margin
17(R-L)
R, L

Methods of Reliability Analysis

Reliability : Based on specific performance criteria

First Order Reliability Methods (FORM)

Cornell Reliability
Index
Hasofer-Lind Reliability
Index

Monte Carlo Simulation (MCS)

A particular set of values of random variables are

generated with the known probability distributions
Nf is the number of simulations such that G() gives less
than zero

18

 Distribution of safety margin, Z = R-L (Melchers

2002)
19

Variability Model

Aleatory : Natural variability of soil

(Inherent variability)
Epistemic : Lack of knowledge (Variability
due to simplified and idealized design
calculation and statistical models)
[Model/transformation uncertainty]
Measurement Error : During
field/laboratory tests

20

Courtesy: Phoon (1995)

21

22

Seismic Design of
Piled Raft Supported
Structure

Pile Foundations : Applications

Bridges

Offshore Structures

High-rise Structures

Storage Tanks

CODAL PROVISIONS ON DESIGN

PRC 1989
UBC 1997
ASCE 4 1998
IS 1893- Part I 2002
Eurocode 8 Part 5 - 2003
JSCE 2000

In general
perception
False
perception

FEMA 440 2005

Neglecting SSI in the analysis
may lead to unsafe design for
both the superstructure and the
foundation
(Mylonakis
and
Gazetas,2000a).

EFFECT OF SSI: A CASE STUDY

Case I: Failed route 3 section of the Hanshin Expressway after 1995 Kobe earthquake
# Computed
response is
25% higher
than the
response of
the fixed
base
structure
neglecting
interaction
effects
which lead to
increase the
ductility
demand on
the pier by
twice due to
increased
base shear.

Effects of dynamic SSI on acceleration

response

Collapse of Hanshin expressway

supported by central columns,
1995 Kobe earthquake.
http://www.eqe.com/publications/kobe/ko
be.htm

Gazetas and Mylonakis, 1998

EFFECT OF SSI: A CASE STUDY

Case II: Typical example of tall building failures on lake bed area of Mexico City.

1985 Mexico City Response

Spectra

Observation

1985 Mexico City earthquake Juarez

Hospital

In Mexico city earthquake the time period of

10-12 storied buildings founded on soft clay

layer were increased from 1.0 to 1.5 seconds to
2.0 sec bringing it closer to resonance
conditions(Resendiz and Rosset 1985, BazanZurita and Bielak 1994).

OF SSI: AA
CASE
STUDY STUDY
EFFECT EFFECT
OF SSI:
CASE
Case III: Response spectra developed for various earthquakes to illustrate the effect
of SSI (Mylonakis and Gazetas 2000 and FEMA 451 2001)

Response spectra developed after

different earthquakes

Spectrum from 1989 Loma Prieta at

Deep Soft Soil Site, Treasure Island

Observation
Yashinsky, 1998cites damage in number of pile-supported bridge structures due to SSI
effect inLoma Prieta EarthquakeinSan Franciscoin 1989.

Sources of Uncertainty
Inherent uncertainties in system properties
Nonlinear behavior and ground motion characteristics
Difficulty in assessing the effects of seismic soil
foundationstructure interaction

Assessment of Safety Margin

Identification, quantification and modelling of the
sources of uncertainty
Computation of the structural response for a wide range
of realistic scenarios
Assessment of SFSI effects on the structural response
and their likelihood of occurrence

Components of probabilistic
analysis

Modelling of System

Idealization of soil-pile raft-structure system

Raft-soil modeling

KxG1 =

KxG2 = [2GL /(1 )](0.73 1.54 0.75 ) [0.2 /(0.75 )]GL[1 ( B / L)]

Pile-soil modeling

Details of soil and pile properties

Tfixed = 0.4 sec (short period) and 2 sec

(long period)

Modelling of Inherent Variability of

Soil

Gi is a zero mean, unit variance, Gaussian random numbers

COVcu : 20 - 50% (Phoon and Kulhway 1999)

Responses
Tssi

Tfixed

VB, col

VB, fixed

VB, pile

Fixed base
system
Flexible base
system

Mean value of SSI period ( Tssi, mean)

Convergence study

No. of Monte Carlo trials

Realizations of soil stiffness for

Monte Carlo Analysis
Mean stiffness = 2700
kN/m
COVcu = 50%

Responses
Realizations of Tssi

Tssi (mean)

Tssi(mean) : Sample mean of time

period of piled raft supported
structure
Tssi : Realizations of time
period from Monte Carlo
simulations

Tssi(mean) : Sample mean of time period of piled raft supported

structure
Tssi : Realizations of time period from Monte Carlo simulations

Probability distribution for system responses

Tssi
COV of Tssi : 7.2% and 9.7% for COVcu = 20%.
COV of VB, pile: 6.2 % for COVcu = 20%.
COV of VB,col: 9.7 % for COVcu = 20%.

(VB,pile)ssi

Results
Normalised Period of structure

Tssi (mean)/Tssi
(det)

Ep /Es = 10000

Ep/Es = 5000

Ep/Es = 1500
Tfixed/Tssi
(det)

COVCu
Tfixed = 2 sec

Ep /Es = 10000, 5000, 1500

(VB, col)ssi (mean)/(VB,

col)ssi (det)
Ep/Es = 5000
Ep /Es = 10000

(VB)fixed/(VB, col)ssi (det)

Ep /Es = 10000, 5000, 1500

Ep/Es = 1500
Normalised Shear force at Pile head

Normalised Shear force at Column

Results
(VB)fixed/(VB, pile)ssi
(det)

(VB, pile)fixed/(VB, pile)ssi

(det)

Tfixed = 2 sec

Ep/Es = 1500
Ep/Es = 5000
Ep /Es = 10000

COVCu

COVCu

Ep /Es = 10000
Ep /Es = 5000
Ep /Es = 1500

Conclusions
Effect of inherent variability of soil may have serious implication in
seismic design of column and pile for piled raft supported structures.
Shear force at column significantly increased due to consideration of
soil inherent variability.
Shear force at pile head obtained from probabilistic analysis yields a
subdued value compared to shear obtained from deterministic SSI
which indicates an over-safe design of pile.
The effect of inherent soil variability is also dependent on E p/Es ratio.
This study shows the importance of reliability based design for piled
raft supported heavy structures with an emphasis to carry out a
detailed study in this direction.

Various applications of probabilistic

analysis
Pile supported structure in liquefiable soil

Fragility curves

Fragility is the probability of attaining or exceeding a certain damage state

It is better to be
probably
right
than
to be
exactly
wrong!

Thank You