Vous êtes sur la page 1sur 88

CE5107 Pile Foundations

Lecture 2b
Axial capacity of pile group
Dr Helen Cheng
Jan 2011

Failure mechanism
Capacity of pile group may not be equal to the sum of
capacities of individual piles.
The ratio of the two capacity is termed the efficiency of the
pile group.
Pressure bulbs of neighbouring piles
tend to overlap, creating greater stress
concentration on the surrounding soil.
This leads to greater settlement of the pile group and is
termed group interaction.
Soil will fail in shear (local failure) or the pile group will
settle excessively (block failure)
2

PILE GROUP EFFECTS ON AXIAL CAPACITY


Efficiency:
= Group Capacity / Individual Pile
Capacities.
For groups in clay, usually < 1
For groups driven in sand, usually > 1
For groups (bored) in sand, ~ 0.67
For end bearing groups, usually ~ 1

BS 8004 recommends that


Friction piles
pile centre to centre spacing (s) > perimeter of square piles; or
> 3 times pile diameter of circular piles
End bearing piles
distance between surface of shafts of adjacent piles
> least width of the pile
4

PILE GROUP EFFECTS ON AXIAL CAPACITY


Local failure:
Pile group capacity QG = h m n Q
For clay, (after Converse-Labarre):

PILE GROUP EFFECTS ON AXIAL CAPACITY

Ultimate capacity of a pile group in cohesive soil (Whitaker 1957)

PILE GROUP EFFECTS ON AXIAL CAPACITY


Group capacity (PG) is lesser of:
Sum of individual pile capacities (P1)
Capacity of block containing piles + soil (PB)
Poulos & Davis (1980)

PILE GROUP EFFECTS ON AXIAL CAPACITY


Load transfer to soil from pile group (Tomlinson):

PILE GROUP EFFECTS ON AXIAL CAPACITY


Block failure:
Estimated using conventional theory:
For sand:

For clay:

B and L are the breath and Length of the pile group in plan
h is the embedded depth (Skempton 1951)

OTHER CASES FOR PILE GROUPS


GROUP WITH CAP ON SURFACE
Group capacity (PG) is lesser of:
Sum of individual pile capacities + capacity of net area of cap
Capacity of block containing piles & soil, + capacity of
portion of cap outside block perimeter.
GROUP ON PROFILE WITH UNDERLYING WEAK
LAYER
Take capacity as lesser of individual pile capacities, or
capacity of block.
EFFECT OF WEAKER UNDERLYING LAYERS CAN BE
VERY IMPORTANT!!
OPTIMISING PILE SUPPORT USE OF PILED RAFTS
(If interest, see Randolph M.F. (2003), Science and empiricism in
pile foundation design. Geotechnique 53, No. 10, 847-875.

10

Effect of Weaker Underlying Layer

11

Piled Rafts

Share load between raft and piles


Focus more on stiffness design
Piles as settlement reducers
Favourable circumstances
Where raft can provide reasonable
stiffness & load capacity e.g. Stiff clay, dense sands
Where induced soil movement do not occur

Unfavourable circumstances for piled rafts


Soft clay or loose sands near surface
Compressible layers at depth
Where consolidation settlement or swelling
movements of soil may occur

12

CE5107 Pile Foundations


Lecture 3 Settlements of Piles
and Pile Groups
Dr Helen Cheng
Jan 2011

13

Lecture 3 Learning Objectives


Elastic settlement of single piles
Analysis of pile-soil interaction
Settlement of pile groups
Settlement ratio method
Equivalent raft method
Equivalent pier method

Estimation of parameters
14

THE BASIC PROBLEMS

RA

Ap

d / 4
2

= 1, for solid pile

RA = area of pile section Ap


bounded by outer
circumference of pile

15

THE BASIC PROBLEMS


Equivalent Solid Cylindrical Pile:

16

SETTLEMENT OF SINGLE PILES


ELASTIC ANALYSIS
Closed Form Solutions
Randolph & Wroth (1979) uniform &
Gibson soil profiles, friction & end-bearing
piles

Chart Solutions
Poulos & Davis (1980) uniform soil profile
Poulos (1979) Gibson soil profile

17

Randolph & Wroth (1979)


Equations (Model)
Pt
wt

18

Randolph Method (1979)


Basic assumption is that shaft
and base settlements can be
considered separately.
Modification to assumption to
account for interaction between
upper and lower layers of soil
(AB), reducing the deformations
to negligible size at some radius
rm

19

Randolph Method (1979)


Deformation of Pile Shaft
The settlement of pile shaft is:

ws 0r0 / G

(1)

where ln(rm / r0 ) = ln0.25 (2 (1 ) 0.25 L / r0


The ratio of total shaft load to pile displacement is:
Ps
2 r0 L 0 2 L

Gr0 ws
Gr0 ws
r0

(2)

This ratio is non-dimensionalized by shear modulus (G) and pile


radius r0.
The values of ws and 0 are average values down the pile shaft.
20

Randolph Method (1979)


Deformation of Pile Base
The pile base acts as a rigid punch on the lower layer soil (Fig.1b).
The load-settlement ratio is from Boussinesq solution:
Pb
4

Gr0 wb (1 )

(3)

Combination of Shaft and Base Displacement


For rigid pile, the settlement of pile shaft is uniform and equal to that
at the pile base. The combined load-settlement ratio is:
Pt
P
P
4
2 L
b s

Gr0 wt Gr0 wb Gr0 ws (1 ) r0

(4)
21

22

Randolph Method (1979)


The solution for compressible piles and non-homogeneous soils
with stiffness increasing linearly with depth is:

Pt
4
2 L tanh(L)
4
1 L tanh(L)

/ 1

Gr0 wt h (1 )
r0
L h (1 ) r0
L

(5)

Where:
Soil Poisson ratio
h r0 / rb (ratio of radius of pile shaft / radius of pile base)
GL / Gb (ratio of Shear Modulus of soil at depth L / below soil
base)
ln(rm / r0 ) ln(influence radius/pile radius)
GL / 2 / GL (ratio of shear modulus at mid-pile depth to pile depth)

E p / GL (pile stiffness ratio)

L 2 / 0.5 L / r0
23

Randolph Method (1979)


1/ 2
For rigid piles, (that is wt / wb 1.2 and L / r0 (0.5 )

Eqn.(5) reduces to:

Pt
4
2 L


Gr0 wt h (1 )
r0

(6)

For compressible piles, the variation of settlement down the pile is


approximately:

w( z ) wb cosh[ ( L z )]

(7)

The ratio of settlement between top (z=0) and base of pile is:

wtop / wb cosh(L)

(8)
24

Randolph Method (1979)


For end bearing piles, with very stiff base ( 0.01) , The terms
containing (1 / h ) dominate in Eqn.(5), and the load-settlement ratio
reduces to:
Pt

Gr0 wt L / r0 tanh(L)

(9)

For stiff piles, tanh(L) L , and the load-settlement ratio is the


elastic compression of the pile:
wt

Pt
L
P
t L
GL r0 r0 E p Ap

(10)

25

Validation of Randolph Method

Test of Eqn.(5) for Compressible


Piles
Test of Eqn.(4 and 6) for Rigid
Piles
26

Validation of Randolph Method

Test of Eqn.(6 and 9) for End-bearing Piles


27

Design Chart from Eqn.(5) from Randolph Method


Floating Pile in Soft Clay

Contours of P/w become parallel to L-axis for large L/ro; this implies
inefficient design

For Floating piles, increasing length of long (compressible) pile show little
reduction in the P/w ratio
Limits of efficient design shown by dashed line

28

Design Chart from Eqn.(5) Randolph Method


End-bearing Pile in laterite

Contours of P/w become parallel to Ep-axis for Ep; this implies inefficient
design
For End-bearing piles, beyond a certain Ep, little shaft compression occurs,
thus nothing is gained by increasing Ep further
Limits of efficient design shown by dashed line

29

DesignChartforSinglePilebasedon
Randolf andWroth(1978)theory

END BEARING vs FRICTION PILE


SETTLEMENTS

30

Design Chart for pile stiffness

Essentially Rigid pile:


Very compressive pile:
31

What is minimum shaft settlement to mobilize full shaft friction?


From Eqn.(1), the shaft settlement is given by:

wt / r0 0 / G; typically 3
The ultimate shaft friction is: ( 0 ) f cu
Thus, the shaft settlement to cause full shaft friction may be taken as:

( wt / r0 ) f 3

G / cu

(11)

Typical values for Soft NC Clays: 1.0; G / cu 50


Typical values for Stiff OC Clays: 0.5; G / cu 150
Therefore the reasonable limits for shaft settlement to mobilize full
shaft friction are:
STIFF CLAYS 0.01 (wt / r0 ) f 0.06 SOFT CLAYS
32

ANALYSIS OF PILE-SOIL
INTERACTION

Simplified analytical solutions (Randolph)


Load-transfer (t-z) methods
Boundary element methods
Finite element methods

For given set of data, all these methods give similar


results.
Attention here is focused on solutions from finite
element method, using elastic continuum theory to
characterize soil behaviour.
Allowances can be made readily for non-linear soil
plastic behaviour.
33

ADVANTAGES OF ELASTICBASED ANALYSES


Continuous soil model; allows stress
transmission
Consistent model parameters understood
Can analyze group behaviour
Can modify to allow for non-linear and cyclic
loading effects.
Can use:
Parametric solutions (Poulos & Davis, Randolph &
Wroth, Butterfield & Banerjee)
Computer programs for problems involving layered
soils or non-uniform pile
34

WHEN MAY A COMPUTER


ANALYSIS BE NECESSARY?
When problem falls outside range of available
parametric solutions
When detailed information on load transfer is
desired
When soil profile is layered
When pile section is non-uniform
When require load-settlement curve to failure
For pile groups, when load and settlement
distributions are required
For examination of mechanisms of deformation
35

LOAD TRANSFER
CHARACTERISTICS

Effect of pile compressibility on


shear stress distribution along pile

Load distribution along pile effect


of stiffness of bearing stratum
36

Load Transfer to Pile Tip

0CK CV
Friction piles

(friction
pile)

0CK CV Cb
End bearing piles

37

PROPORTION OF BASE LOAD


ON PILE FACTORS CK, C

38

PROPORTION OF BASE LOAD


ON PILE FACTOR Cb

39

FEM Validation of Load Transfer


Load P=1000 kN
D=1m
L=25m
Es=10000 kPa
Use Dummy Plate to get
Axial Forces Distribution

Ep=1000Es ie K=1000

Plate has EA and EI 1E6


times smaller than real pile
solid properties

Cases: Eb/Es=1, 10, 100, 10000

ns=0.49

40

FEM Validation of Load Transfer


Eb/Es=10000

Eb/Es=100

d=2.85 mm

d=3.40 mm

Eb/Es=10

Eb/Es=1

d=5.74 mm

d=8.45 mm

41

FEM Validation of Load Transfer


P=1000 kN

P=1000 kN

Eb/Es=10000

Eb/Es=100

d = 2.85 mm

d = 3.40 mm

db = 0.001 mm

db = 0.66 mm

Compr = 2.85 mm

Compr = 2.74 mm

P=1000 kN

P=1000 kN

Eb/Es=10

Eb/Es=1

d = 5.74 mm

d = 8.45 mm

db = 3.51 mm

db = 6.73 mm

Compr = 2.23 mm

Compr = 1.72 mm

Pile Compression and Settlements

42

FEM Validation of Load Transfer


P=1000 kN

P=1000 kN

Eb/Es=10000

Eb/Es=100

Pb=820 kN

Pb=740 kN

Pb/P=0.82

Pb/P=0.74

P=1000 kN

P=1000 kN

Eb/Es=10

Eb/Es=1

Pb=370 kN

Pb=100 kN

Pb/P=0.37

Pb/P=0.10

Axial Force Distribution

43

FEM Validation of Load Transfer


P=1000 kN

P=1000 kN

Eb/Es=10000

Eb/Es=100

s = 2.3 kN/m2
Ps = sPiDL = 180 kN
Ps/P=0.18
Pb/P=0.82

s = 3.3 kN/m2
Ps = sPiDL = 260 kN
Ps/P=0.26
Pb/P=0.74

P=1000 kN

P=1000 kN

Eb/Es=10

Eb/Es=1

s = 8.0 kN/m2
Ps = sPiDL = 630 kN
Ps/P=0.63
Pb/P=0.37

Skin Friction Shear Stress Distribution

s = 11.5 kN/m2
Ps = sPiDL = 903 kN
Ps/P=0.90
Pb/P=0.10

44

Effects of Compressible Piles


K=Ep/Es=50

K=Ep/Es=1000

K=Ep/Es=5000

d = 22.28 mm

d = 8.45 mm

db = 2.51 mm

d = 7.51 mm

db = 6.73 mm

db = 7.15 mm

Compr = 19.77 mm

Compr = 1.72 mm

Compr = 0.36 mm

45

Effects of Compressible Piles


K=Ep/Es=50

K=Ep/Es=1000

K=Ep/Es=5000

d = 22.28 mm

d = 8.45 mm

d = 7.51 mm

db = 2.51 mm

db = 6.73 mm

db = 7.15 mm

Compr = 19.77 mm

Compr = 1.72 mm

Compr = 0.36 mm

Pile Compression and Settlements

46

Effects of Compressible Piles


K=Ep/Es=50

K=Ep/Es=1000

K=Ep/Es=5000

P = 1000 kN

P = 1000 kN

P = 1000 kN

Pb = 30

Pb = 100

Pb = 120

Pb/P = 0.03

Pb/P = 0.1

Pb/P = 0.12

Axial Force Distribution

47

Effects of Compressible Piles


K=Ep/Es=50

K=Ep/Es=1000

K=Ep/Es=5000

s = 12.5 kN/m2

s = 12.6 kN/m2

P/PiDL = 12.7 kN/m2

P/PiDL = 12.7 kN/m2

sP/PiDL = 1.0

sP/PiDL = 1.0

s = 40 kN/m2
P/PiDL = 12.7 kN/m2

sP/PiDL = 3.1

s = 4 kN/m2
P/PiDL = 12.7 kN/m2
sP/PiDL = 0.31

Skin Friction Shear Stress Distribution

48

Summary of Findings
Ratio of Load Transfer to pile base is strongly
dependent on ratio of Eb/Es; ie the relative
stiffness of soils around pile base compared to
soils around pile shaft
The shape of skin friction distribution is more
dependent upon Ep/Es; ie the relative stiffness
of pile material to soils around pile shaft
Soil strengths serve to limit the values of
maximum skin friction and end bearing that can
be mobilized
49

MAIN CHARACTERISTICS OF
BEHAVIOUR
Major part of settlement is IMMEDIATE SETTLEMENT
(typically >80%)
Effect of compressibility is important for long slender
piles. For long compressible piles, settlement is little
influence by soil stiffness at pile tip.
For piles of normal proportions in clay, the loadsettlement behaviour is largely linear at normal working
loads. Thus, elastic theory can be used directly.
Nonlinear effects are important when piles derive much
of their capacity from base resistance,
Examples are: Piles in sand
Piles with enlarged base
Large diameter bored piles

50

SETTLEMENT OF PILE GROUPS


METHODS OF ANALYSIS -Hand

Interaction factor method


Settlement ratio method
Equivalent raft method
Equivalent pier method

51

INTERACTION FACTOR
METHOD

52

Interaction factors -effect of various


parameters

53

Interaction factors -characteristics


Decreases as s/d increases


Decreases as K decreases
Decreases as L/D decreases
Less for end bearing than friction piles
Less for non-homogeneous than homogeneous
profiles
Increases as s increases
Increases if base enlarged
Decreases if soil between piles is stiffer
54

Interaction factors - effect of stiffer soil


between piles

NOT significant reduction in for < 1


55

Interaction factors warnings!!

56

Group analysis via interaction factors

57

Solution of group settlement for typical pile


head conditions
Known loads (flexible pile cap)
Settlement of each pile is calculated directly.
Will have differential settlements within the group

Rigid Pile Cap

Settlement of all piles equal


Pile loads unknown
Form equations for each pile in group
Solve equations for unknown loads & group settlement

Analysis requires:
Interaction factors
Settlement of single pile
58

Expression of analysis results

59

SETTLEMENT RATIO METHOD


FOR GROUP SETTLEMENT

60

Theoretical solutions for exponent w

61

SETTLEMENT OF PILE GROUPS


-EQUIVALENT RAFT METHOD

62

Settlement Analysis by the


Equivalent Raft/Footing Method
FILLS, etc.

G.W.

Equivalent Footing
placed at the Location
of the Neutral Plane

2:1 distribution

The compressibility in this


zone must be of soil and pile
combined

Settlement of the piled foundation is caused


by the compression of the soil due to increase
of effective stress below the neutral plane
from external load applied to the piles and, for
example, from fills, embankments, loads on
adjacent foundations, and lowering of
groundwater table.

Ecombined

Apile E pile

Asoil Esoil

Apile Asoil

2:1 distribution

63

Settlement of pile groups equivalent raft method

64

SETTLEMENT OF PILE GROUPS


-EQUIVALENT PIER METHOD

65

Settlement of pile groups equivalent pier method

66

Settlements due to underlying compressible


layers
Interaction factor approach
Require interaction factors for the appropriate soil
profile

Approximate Approach
S = settlement of group in founding layer + additional
settlement due to underlying layers (S)
Compute S from:
1-D analysis
Equivalent pier analysis, via calculation of settlement of
underlying layers from elastic theory

67

Settlement of pier due to underlying


compressible layers

68

COMPUTER PROGRAMS FOR


GROUP SETTLEMENT

3D Continuum Models: 3D FLAC or Plaxis 3D-Foundation


69

Lessons from comparisons


Assessment of soil modulus values is critical
The method of analysis is less critical (provided
it is sound)
Beware of analyzing very large groups of piles
with the interaction factor method. There is a
potential for significant over-estimation of
settlements
Equivalent raft and pier analyses are useful
checks on the order of group settlement and
should always be carried out in addition to
computer analyses
70

ESTIMATION OF SOIL PARAMETERS


Soil modulus (Es) is the key parameter.
Laboratory Testing
Not usually useful because of Differences between
stress paths in lab and field
Difficulty of accounting for installation effects

Interpretation of Load Tests


Fit theory to observed behaviour
Usually the most satisfactory method
Most correlations are for SECANT MODULUS at a typical design load
level
Can also correlate initial tangent modulus and degrade with stress level
71
e.g. hyperbolic curve

Estimation of soil parameters piles


in clay

72

Estimation of soil parameters piles


in sand

73

Estimation of soil parameters crude


correlations with CPT

74

Back-analysis of Pile Load Tests

75

Back-analysis of Pile Load Tests

Method is very useful for getting Elastic Soil Stiffness from Pile Load Tests Data
76

COMPUTER PROGRAMS FOR


GROUP SETTLEMENT
Program

Originator

Features

PIGLET

Randolph

Simplified linear analysis


with interaction factors

PGROUP

Banerjee

Linear boundary element


analysis

DEFPIG

Poulos

Simplified boundary
element analysis with
interaction factors; nonlinear capability

PGROUP

ONeill

Hybrid: t-z (&p-y)


analyses for single pile,
elastic theory for group
interaction

Coffey Geosciences

PILE GROUP CASES ANALYSED


PG

PG

h = 1.5L
s

E1

E1
E2

E2
(a) Floating pile group
Coffey Geosciences

(b) End bearing pile group

COMPARISON BETWEEN SOLUTIONS FOR


GROUP SETTLEMENT
1.4
1.2

Values of E 2 /E 1
0.2

SG E 1 d/PG

1.0

DEFPIG analysis
Equivalent pier analysis
Equivalent raft analysis

0.8
0.6

1.0

0.4
L/d
s
K
s/d

0.2
0
Coffey Geosciences

=
=
=
=

25
0.3
1000
3
2

5.0

3
n

APPLICATION TO CASE IN JAPAN

Coffey Geosciences

SENSITIVITY STUDY- CASE OF


ONEILL (1982)
Factors for decision:
Method of analysis & associated soil model
Idealization of soil profile
Geotechnical parameters

Coffey Geosciences

SENSITIVITY STUDY- EFFECT OF


ANALYSIS METHOD

Coffey Geosciences

SENSITIVITY STUDY- EFFECT OF


SOIL PROFILE IDEALIZATION

Coffey Geosciences

SENSITIVITY STUDY- EFFECT OF


SOIL MODULUS CORRELATION

Coffey Geosciences

CASE OF GOOSENS & VAN IMPE (1991) SOIL PROFILE & PARAMETERS

Depth (m)

10

20

30

Fill

Average cone
resistance. M Pa
<1
10

Youngs M odulus
E s M Pa
3
100

Clayey
sand

60

80

Medium stiff
clay
Rel. dense sand

1.5

11

12

42

Tertiary
clay

3.5

26

Very dense
sand

200
(assum ed)

40

50

Coffey Geosciences

COMPARISON BETWEEN SETTLEMENTS


0

Measurement Point
2
2A
3

Settlement (mm)

100

200

300

400

Coffey Geosciences

500

Com puted S TF for equivalent raft


M easured at 25-2-87
Com puted from DEFPIG group
interaction calculations
Com puted by M andolini
& Viggiani (1997)

LESSONS FROM COMPARISONS

Assessment of soil modulus values is critical


The method of analysis is less critical (provided it
is sound)
Beware of analyzing very large groups of piles
with the interaction factor method. There is a
potential for significant over-estimation of
settlements
Equivalent raft and pier analyses are useful
checks on the order of group settlement and
should always be carried out in addition to
computer analyses

Coffey Geosciences