GEOTECHNICAL EARTHQUAKE ENGINEERING

2010 - 2011

Fabrice EMERIAULT
GENERAL OUTLINE

Introduction

Elements of engineering seismology

Cyclic and dynamic behaviour of soils

Liquefaction of soils

Seismic behaviour of retaining walls

Seismic slope stability and design of earth dams

Seismic behaviour of foundations

Seismic behaviour of buried structures

LIQUEFACTION OF SOILS
LIQUEFACTION OF SOILS

Introduction

Physical evidence

Behaviour of soils: drained or undrained

monotonic loading

Behaviour of soils: drained or undrained cyclic
loading

Prediction of liquefaction by laboratory tests

Prediction of liquefaction by in situ tests

INTRODUCTION

Soil is considered as a multiphasic material (bi-
phasic in the case of saturated soil)

Earthquake =

essentially shear loading of the soil (S-waves or
surface waves)

cyclic loading

volumetric effect is important
PHYSICAL EVIDENCE
Loss of bearing capacity due to liquefaction

7
Niigata , 1964 http://www.youtube.com/watch?v=KLZFlnND0hA
Caracas, 1967 Izmit, 1999

http://www.youtube.com/watch?v=ngxG49Lf6co&NR=1
Failure of slope or embankments due to liquefaction

San Fernando dam, 1971

Liquefaction of the part of the dam

created by hydraulic fill
Lateral spreading due to liquefaction

Kobé, 1995
Sand boils due to liquefaction
BEHAVIOUR OF SOILS: DRAINED
OR UNDRAINED MONOTONIC
LOADING
DEVIATORIC OR SHEAR LOADING
+ +
Mean pressure: p = σ1 σ 2 σ 3 p' = p − u
3
Deviatoric stress: q = σ1 −σ 3 q' = q

Isotropic axis
σ3
sII More generally the deviatoric
stress can be written as:
π
p
s II = (σ 1− p )2 + (σ 2 − p )2 + (σ 3− p )2
2
= q
σ2 3

σ1
DRAINED TRIAXIAL TEST = MONOTONIC
SHEAR LOADING

∆σ
σ 1 = σ 0 + ∆σ = σ '1 p =σ0 +
3
σ 2 = σ 3 = σ '2 = σ '3 = σ 0 q = ∆σ
= 3( p − σ 0)

compression Stress path in the p – q plane

q = 3( p − σ 0)
σ0
p Straight line with a slope equal to 3
extension
EFFECT OF THE INITIAL DENSITY
Loose material: e > ecr Contractancy

q
εv
Critical state

ε1 ε1

Dense material: e < ecr Dilatancy

q
εv
Critical state Characteristic state
ε1

ε1

Characteristic state: change of evolution of volumetric strains ε&v = 0

EXAMPLE OF DRAINED TEST ON LOOSE
SAND

Reference:
Drained triaxial test on
loose sand
Relative density = 30 %
σ3 = 400 kPa
UNDRAINED MONOTONIC SHEAR LOADING
OF LOOSE SOILS
Saturated soil:
σ 1 = σ + ∆σ
0
Undrained test = constant volume
σ2 =σ3 =σ0 Pore pressure is not constant
q
σ'3

ε1 ε1

∆u > 0
ε1
u0 σ'0 p’

In red, the case of very loose material

EXAMPLE OF DRAINED TEST ON LOOSE
SAND

Static liquefaction:
Strain goes from 1% to
25% in 0,2 s
Strength is almost
reduced to 0
EXAMPLE OF DRAINED TEST ON MEDIUM
LOOSE SAND

Static liquefaction but

less pronounced:
Strain goes from1,5%
to 18% in 0,4 s
Strength is low
UNDRAINED MONOTONIC SHEAR LOADING
OF DENSE SOILS
q
σ'3

ε1 ε1

q
u
∆u < 0
u0 ε1

∆u > 0
σ'0 p’

Effect of the constrained dilatancy of dense soils

EXAMPLE OF DRAINED TEST ON DENSE
SAND

The constrained
dilatancy leads to a
decrease of the pore
pressure, thus to an
increase in effective
stress and consequently
of the strength
BEHAVIOUR OF SOILS: DRAINED
OR UNDRAINED CYCLIC LOADING
MAIN MECHANISM = CHARACTERISTIC STATE

24
 EXAMPLE OF A DRAINED CYCLIC SHEAR
TEST
20 cycles with 0.2 MPa of amplitude
q q

0.8 MPa

Limit of the
characteristic zone
in compression

1%
εv
10 %
ε1
Limit of the
characteristic zone Densification
in extension
Dilatancy
EXAMPLE OF UNDRAINED CYCLIC TEST ON
A SATURATED LOOSE SAND

Stress controlled cyclic test ∆σ

Phénomène de liquéfaction

Pression
interstitielle

Cisaillement
 29

Stress controlled cyclic test ∆σ

 30

Liquefaction / cyclic mobility

 31

Strain controlled cyclic test ∆ε

PREDICTION OF LIQUEFACTION
BY LABORATORY TESTS
PREDICTION OF LIQUEFACTION BY
LABORATORY TESTS

Basic soil analysis

Water content

Seive analysis

Atterberg limits

Mechanical tests

Cyclic triaxial test

Cyclic direct shear test

Cyclic torsional test
BASIC SOIL ANALYSIS
Liquefiable soils (according to the french code PS 92)
Qualitative analysis of the potentiel of liquefaction of the soil

Type I: Sands and silts

- Sr close to 100 % (saturated soil)
- uniform granulometry (Cu = D60/D10< 15)
- 0.05 < D50 < 1.5 mm
- Final effective stress smaller than 200 kPa or 300 kPa (depending on
the zone)
Type II: Clays
- D15 > 5 µm
- liquid limit wL < 35 %
- in situ water content wnat > 0.9 wL
- Point (wL , PI) above line A in Casagrande abaqus

Soils not prone to liquefaction: - Soils with D10 > 2 mm

- Soils with D70 < 74 µm and PI > 10 %
38
 MECHANICAL TESTS

Methodology:

Perform several cyclic tests (stress controlled)

- on undisturbed samples of the same soil
- under the same initial effective stress
- but with different values of ∆σ

For each test determine the number of cycles before

initiation of the liquefaction

The Cyclic Stress Ratio CSR = ∆σ / σ’0 can be plotted

against N

On the example: N = 24 for ∆σ = 30 kPa

39

Cyclic stress ratio CSR vs N

CSR

For the same initial conditions, liquefaction can be obtained with a smaller
deviatoric stress amplitude with a larger number of cycles 40

Log scale for N


Effect of K0 on CSR

Effect of
Densité initiale relative density Dr on CSR

Increasing Dr
PREDICTION OF LIQUEFACTION

Evaluate the cyclic shear stress created by the

earthquake

Evaluate an equivalent number of cycles

Plot in the CSR vs N curve

If the point is above the curve, liquefaction will occur

If the point is below the curve, liquefaction will not be
observed

43

Evaluate the cyclic shear stress created by the
earthquake

τave/σ’0 = 0.65 amax / g σ0/σ’0 rd

amax maximum acceleration

σ0 total vertical stress at the considered depth

σ’0 effective vertical stress at the considered depth

rd reduction coefficient function of the depth
rd = 1 – 0,015 z (z in m)
44

Evaluate an equivalent number of cycles

Statistical analysis have shown that a number of cycles of

amplitude τave equivalent to the EQ can be proposed and
that it is related to the magnitude of the EQ.

Magnitude Number of cycles

5.5 – 6 5
6.5 8
7 12
7.5 15
8 20
45

Plot in the CSR vs N curve

If the point is above the curve, liquefaction will occur

If the point is below the curve, liquefaction will not be
observed

46

Example of application

Case of Niigata site (EQ in 1964)

The analysis « predicts » the liquefaction of a certain
layer of soil
PREDICTION OF LIQUEFACTION
BY IN SITU TESTS
PREDICTION OF LIQUEFACTION BY IN SITU
TESTS

Standard Penetration Test SPT

Cone Penetration Test CPT

Sismocone

Pressumeter test

Standard Penetration Test SPT
The resistance of the soil is characterized by N1

Empirical relation between N1 and CSR

Corresponds to a certain magnitude (7.5)

and thus to a certain number of equivalent
cycles
Liquefaction
For different magnitudes, CSR is
multiplied by a coefficient:

Magnitude Coefficient
No liquefaction
5.25 1.5
6 1.32
6.75 1.13
7.5 1
8.5 0.89

50
The methodology is then the same as with
lab tests results

Cone Penetration Test CPT
The resistance of the soil is characterized by qc

Same approach as for the SPT: empirical

relation between qc and CSR

Or correlation between N1 and qc

51

Cone Penetration Test CPT
The resistance of the soil is characterized by qc

Same approach as for the SPT: empirical

relation between qc and CSR

Or correlation between N1 and qc

for fine sands (0.1 mm) qc is

approximately equal to 4 N1

52

Cone Penetration Test CPT
Other empirical approaches based on qc and Rf
measured in the CPT

For Robertson & Campanelle (1985):

Definition of zone A where soils are prone

to liquefaction

53
For Olsen et al.(1996):

54

Sismocone
Penetration test with measure of the S-wave
velocity Vs

55
TREATMENT OF LIQUEFIABLE
SOILS
TREATMENT OF LIQUEFIABLE SOILS

Vibroflottation

Dynamic compaction

Stone columns

Jet-Grouting

Drainage

…

These techniques will be presented in an other

session