MONOGRAPH ON

GRANULAR PILES AND GRANULAR PILE ANCHORS

1
Madhav Madhiraand 2Vidyaranya Bandi
1 Prof. Emeritus, JNT University and Visiting Professor, IITH, Hyderabad, 500085, India
2 Engineering Manager, L&T TI-IC, Mumbai 400097, India.

1. INTRODUCTION

Granular piles are often constructed through soft soils fully penetrating to an end

bearing stratum or as floating piles in deep deposits, the tips restingat depths where the

strength of the soil is adequate.

Failure mechanisms (Fig. 1) for a single granular pile are bulging, general shear and pile

failure though probable failure is by bulging or pile failure. Methods to estimate the

ultimate capacity of granular piles corresponding to general shear, bulging and pile failures

are presented in Table 1. In pile failure mode, the total load applied on the granular pile is

resisted by shaft resistance generated along the shaft length and the bearing resistance at

the base of the GP while the resistance generated by lateral confinement of the granular fill

material near the top in GP limits its bulging capacity.

The functional utility of the granular pile to carry the compressive load is extended to

resist the uplift or pullout forces generated in foundations by a simple modification of

connecting the base of the foundation to a plate, pedestal or geogrid at the tip of the

granular pile by a cable or rod to transfer the pullout load (Fig. 2).

1
Fig. 1 (a) Bulging, (b) General Shear and (c) Pile Failure Mechanisms for Single
Granular Pile

Uplift Force, Po
G.L

Cable
Dense
Granular Fill
L d
In Situ Soil

Plate/Pedestal

Fig. 2 Granular Pile Anchor.

2
 Table 1 Estimation of Ultimate Load (Aboshi and Suematsu, 1985)
Mode of
Derived Formula References
Failure
sin 

q ult   c zk pc  2c o k pc 11  sin 
s
,
s
Greenwood
where kpc is the passive earth pressure coefficient of column (1970)
and s is the frictional resistance of soil


qult  Fc1C o  Fq1 Qo 11  sin 
sin 
s
, Vesic (1972),
s
Datye &
where Fc1 and Fq1 are the cavity expansion factors and Qo Nagaraju
is the surcharge stress (1975)
Bulging
1  sin  s Hughes and
q ult   ro  4C o  ,
1  sin  s Withers
(1974)
1  sin  s
qult 
1  sin  s
 2 2
4C o   ro  K o q s  W B  1  W B q s
 
 
Madhav et al.
where W and B are diameters of stone column and footing (1979)
respectively.

1 
q ult  C o N c    c BN     c D f N q ,
2  Madhav and
where Nc, Nq and Nγ are the dimensionless parameters that Vitkar (1978)
General depend on the trench and soil parameters.
Shear 1 
qult    c B tan 3    2Co tan 2   2(1  a s )C o tan
2  Barksdale and
tan  s a s tan  s 
1
Bachus (1983)
  45 0 
2

 
q ult  1  a s C o   s z   s z a s tan  s cos 2 
n ,
Sliding s  Aboshi et al.
1  n  1a s
Surface (1979)
where ar is area replacement ratio and s & γs are column
parameters

3
2. ULTIMATE CAPACITY OF GRANULAR PILE (GP) IN HOMOGENOUS
GROUND

Ultimate capacity of granular pile in compression is estimated for homogenous ground,

i.e., the undrained strength, cu, of the in situ soil is constant with depth.

2.1 Single Granular Pile in Compression

A granular pile of diameter, d, and length L, is considered (Fig.3). The saturated unit

weight, the undrained strength and the shear modulus of the in situ soil assumed constant

with depth are s, cu and G respectively while  gp and gp are respectively the angle of

shearing resistance and unit weight of the granular pile material.

The ultimate pile capacity, Pcomp, is limited by the interface shear stresses, τ, acting on

the cylindrical boundary and the ultimate bearing stress, q b, at the base of the GP (Fig. 4a).

The ultimate shear stresses,  equals the undrained shearstrength, cu, while the limiting

bearing stress, q b, equals Nc.cu (Fig. 4b).The ultimate capacity, P ult, of GP in compression

 .d 2
by pile capacity after normalization with .cu reduces to
4

* L
Ppf  4  Nc (1)
d
where Ppf * = Pult, pf/{d2/4}cu and Nc – bearing capacity factor that varies from 6.2 to 9

for L/d increasing from 0 to 5 or more.

4
 Compressive Load, P

G.L

gp, gp

L d
G, s, cu

Fig. 3 Granular Pile under Compression.

Compressive Load, P

Base Resistance, qb

(a) (b)
Fig. 4(a) Pile & (b) Bulging Failures for GP.

5
For bulging failure, following Gibson and Anderson (1961), Hughes and Withers (1974)

and Hughes et al. (1975), for expansion of a cavity near the top (at a depth of d/2 from the

top) of the GP, Pult, bf is

.d 2
Pult, bf  N cu .Nc*   ho  (2)
4

where the lateral confining pressure  h o , is the horizontal total stress at depth equal to half

(1  sin  gp )
the diameter (d/2) of GP, N c*  1  ln  G  and N   . Normalizing Pult, bf
c  (1  sin  gp )
 u 

 .d 2
with .cu Eq. (2) reduces to
4

4 Pcomp
P* 
d 2 cu

 N  N *c    (3)

  w .d  K o . sub 
where      1
 cu   w 

The critical length, (L/d)cr defined as is the length at which the ultimate capacities by

pile and bulging failures equal. The ultimate capacity is governed by pile failure for L/d

smaller than the critical length and by bulging falure for L/d greater than the critical length.

2.2 Ultimate Pullout Capacity of Granular Pile Anchor (GPA) – Homogenous Ground

The applied pullout load is transferred to the base through the cable or steel rod attached

to the base plate, pad or sheet placed prior to the installation of the granular pile material

(Fig. 5).

The ultimate pullout capacity of the GPA is the lesser of the loads at which it is either

pulled out by pile (Fig. 6a) or by bulging (Fig. 6b) failure. The normalized ultimate

capacity, P* of GPA by pile capacity is

6
 Pullout Load, Pult

G.L

gp, gp

d
L
G,s, cu

Fig. 5 GPA under Pullout

τ Pult
τ
τ
W ‘ = gp.V
‘
Pult ‘ Pullout Load
L-d/2
‘
‘
‘
τ
τ σh
(a) (b)

Fig. 6 Pullout (a) and Bulging (b) Failures of GPA.

4 Pult L
P*  2
 (4   ) (4)
d cu d

7
  gp.d
where 
cu
Bulging is considered likely to occur at a distance of half-diameter of the GPA from the

tip instead of from the top as was considered for bulging capacity of granular piles in

compression. The bulging capacity of the GPA is

 .d 2
Pult  .N cu .Nc*   ho  (5)
4

The total horizontal stress, h0, is considered at depth z  L  d  2

 assuming groundwater
level to be at ground level. The normalized ultimate pullout load by bulging, P* is

4 Pult   L 1 
P*  2
 N   N c*   .    (6)
d cu   d 2 

where    w .d  K . sub  1 - a lateral confining stress parameter that depends particularly
cu   w 

on lateral earth pressure coefficient, K, of the in situ soil. The critical length, (L/d)cr

defined as is the length at which the ultimate capacities by pile and bulging failures equal.

The ultimate capacity is governed by pile failure for L/d smaller than the critical length and

by bulging falure for L/d greater than the critical length.

2.3 Results

The ultimate capacity of GP in compression and the ultimate pullout capacityof GPA

are estimated for both the pile and bulging failure mechanisms using Eqs. 1 & 3 for GP and

4 & 6 for GPA for the following ranges of the parameters: s: 14 to 16 kN/m3; gp: 18 to 21

kN/m3; cu: 10 to 60 kPa; L/d: 1 to 25; gp: 300 to 450; G/cu: 50 to 500; sd/cu: 0.1-2;

(=gpd/cu ): 0.1-2.5; γsubd/cu: 0.03 to 0.7; γwd/cu: 0.08 to 1.2, β = 0.1 - 1.6 and K0=0.5-1.0.

8
 The ultimate capacity of GP is presented in Fig. 7 as a function of L/d for φgp in the

range of 300 to 450, for G/cu = 100 & 200, and β = 1.0.

45

450

400
30

350
P*

gp=300

15
G/cu=100
200

0
0 10 L/d 20 30

Fig. 7 Ultimate capacity, P* of GP - Effect of φgp for G/cu = 100 and 200 & β = 1.0.

9
 30

2.6

1.9
P*

1.45
25 1
0.55
 = 0.1

20
0 15 30
L/d

Fig. 8 Ultimate capacity, P*of GP - Effect of for gp = 350 & G/cu=200.

Fig. 8 depicts the effect of the lateral stress parameter, , on the ultimate capacity of the
GP.

10
450

400
(L/d)cr

5 350

gp =300

0
0 300 600
G/cu

Fig.9 Critical length, (L/d)cr for GP in compression – Effect of gp for β = 1.0.
10
The effects of G/cu and gp on the critical length, (L/d)cr of GP are shown in Fig. 9.

The variation of ultimate pullout capacity, P * with L/d showing the effect of for = 1,

G/cu=200 & φgp=350 is depicted in Fig. 11.

The ultimate pullout capacity of GPA is presented in Fig. 10 as a function of L/d and

includes the effects of G/cu &  gp for 1.3 & =1.0. It may be noted that the ultimate

pullout capacity increases with L/d even for bulging failure mode since bulging is expected

to occur near the tip of GPA.

200

450

400
P*

350
100
gp=300

G/cu =50
200
500
0
0 15 30
L/d

Fig. 10 Ultimate pullout capacity, P* vs L/d for GPA – Effect of G/cu & gp for 1.3 &
 = 1.0.

11
 120

=0.1

Bulging Failure
P*

60 0.7
1.3
1.9
2.5

0
0 15 30
L/d

Fig. 11 Ultimate pullout capacity, P* vs L/d for GPA – Effect of for = 1, G/cu=200 &
φgp=35 0.

120
1
p*

0.5
60

=0.1

0
0 15 30
L/d

Fig. 12 Ultimate pullout capacity, P* vs L/d for GPA – Effect of  for = 1.3, G/cu=200 &
φgp=35 0.
12
 The effect of the lateal stress parameter, , on ultimate pullout capacity of GPA is

depicted in Fig. 12.

20

350
(L/d)cr

10
gp=300

0
0 300 600
G/cu

Fig. 13 Critical length, (L/d)cr vs G/cu for GPA -Effect of gp for  1.0 in
GPA.

13
 30

=0.7
(L/d)cr

15 1.3

1.9
2.5

0
0 300 600
G/cu

Fig. 14 Critical length, (L/d)cr vs G/cu for GPA - Effect of for &gp

16
1
(L/d)cr

8 0.55

=0.1

0
0 300 600
G/cu

Fig. 15 Critical length, (L/d)cr vs G/cu for GPA - Effect of for gp 

14
The variation of critical length, (L/d)cr of GPA with G/cu for gp varying from 300 to 450 is

shown in Fig. 13 while the variations with  and  in Figs 14 and 15 respectively.

3. NON-HOMOGENOUS GROUND

The undrained shear strength of in situ soil is considered (Fig.16) to increase linearly

with depth (non-homogenous ground), and the ultimate capacities of the GP and GPA

estimated. The variation of undrained shear strength of normally consolidated soil with

depth normalized with length of the granular pile, is expressed as

 z
cu ( z )  cuo 1   c  (7)
 L

where c, non-homogeneity strength parameter expresses the rate of increase of undrained

shear strength with depth.

cu

Non-homogenous
Ground
 z
cu  cuo 1   c 
 L

z
L

Fig. 16 Profile of undrained shear strength of the soil with normalized depth

15
3.1 Ultimate Capacity of Granular Pile (GP)

The ultimate compressive capacity, Pult of GP by pile capacity normalized with

 .d 2
.cuo is
4

 L   
P *   4.  N c .1  c  (8)
 d  2 

where the normalized compressive capacity, P*,of GP, P*  Pult is

.d 2 
 cuo
 4 
The normalized ultimate capacity, P*, of GP for bulging failure in non-homogeneous

ground is

4 Pcomp  d  * 
P*  2
 N  1   c  N c  0.5.  (9)
d cu 0  2.L  

3.2 Ultimate Pullout Capacity of Granular Pile Anchor (GPA)

The normalized of ultimate pullout capacity, P*of GPA for pile failure is

L
P*  {4(1  0.5. c )  } (10)
d

 gp.d
where   - function of the density of the granular fill material.
cuo

The undrained strength of the soil at distance d/2 from the tip of GPA where bulging is

expected to occur, is

 ( L  d / 2)   c  L 1 
cu  cuo .1   c   cuo .1  .    (11)
 L   (L / d )  d 2  

The normalized ultimate pullout load, P*, of GPA for bulging failure, is

16
 4 Pult  c  L 1  *  L 1 
P*   N   1  .    N c   .    (12)
 d 2 c uo  (L / d )  d 2  d 2 

where    w .d   sub . K o  1 - lateral confining pressure parameter

c uo  w 

3.3. Results

The ultimate compressive and pullout resistances of GP & GPA in non-homogenous

ground are estimated for both the pile and bulging failure mechanisms using Eqs. 8 &10,

and 12 & 14 respectively for the following ranges of the parameters: s: 14 to 16 kN/m3;

gp: 18 to 21 kN/m3; cuo: 10 to 60 kPa; L/d: 1 to 25; gp: 300 to 450; G/cuo: 50 to 500;

sd/cuo: 0.1-2; (=gpd/cuo): 0.1-2.5; γsubd/cuo: 0.03 to 0.7; γwd/cuo: 0.08 to 1.2, αc =0.5 –

1.0, β = 0.1 - 1.6 and K0=0.5-1.0.

500

26 200

100
P*

G/cuo=50

10
0 15 30
L/d

Fig. 17 Ultimate compressive capacity, P* for GP vs. L/d - Effect of G/cu for φgp = 350, β
= 1.0 & c=0.5 in non-homogenous ground.

17
The variations of ultimate capacity of GP in non-homogeneous ground (undrained strength

increasing linearly with depth) with L/d for different G/cu and non-homogeneity parameter,

c are given in Figs. 17 and 18.

30

1
P*

0.7
27
0.5

c=0

24
0 12.5 25
L/d

Fig. 18 Ultimate compressive capacity, P* for GP vs L/d– Effect of c for G/cuo = 200, φgp
= 350 & β = 1.0 in non-homogenous ground.

18
 5.5

c=0
(L/d)cr

0.5
0.7
4
1

2.5
0 300 600
G/cuo

Fig. 19 Critical length, (L/d)cr vs. G/cuo for GP – Effect of cfor φgp = 350 & β = 1.0 in
non-homogenous ground.

8
c=0

0.5
0.7
(L/d)cr

2
28 37 46
Angle of Shearing Resistance,  gp

Fig. 20 Critical length, (L/d)cr vs. gpfor GP– Effect of c for G/cuo=200 & β = 1.0 in
non-homogenous ground.

19
The effects of non-homogeneity parameter, c, as effecting the variations of critical length,

(L/d)cr of GP with G/cuo and φgp are given in Figs. 19 and 20.

4.8

c=0

4.2 0.5
0.6
0.7
(L/d)cr

3.6

3
0 1.5 3


Fig. 21 (L/d)cr vs.  for GP–Effect of cfor G/cuo=200 &φgp=35 0 in non-homogenous

ground.

Fig. 21 depicts the effect of non-homogeneity parameter, con the variation of (L/d)crfor

GP with  for G/cuo=200 & φgp=35 0.

20
 150

500
200
100
P* G/cuo=50

75

0
0 15 30
L/d

Fig. 22 Ultimate pullout capacity, P* for GPA vs. L/d -Effect of G/cu for φgp=350, =1.3, β
= 1.0 & c=0.5 in non-homogenous ground.

Figs. 22 and 23 present the variations of ultimate pullout load of GPA with L/d for

different G/cu and c.

21
 140
1
0.8
0.5
c=0
P*

70

0
0 15 30
L/d

Fig. 23 Ultimate pullout capacity, P* for GPA vs. L/d– Effect of c for G/cuo=200, φgp=350,
λ=1.3 & β = 1.0 in non-homogenous ground.

Figs. 24 and 25 show variations of (L/d)cr with (G/cu0) and φgp and show the effect of non-

homogeneity parameter, c.

22
 16
c=0

0.5
0.7
(L/d)cr 1

12.5

9
0 300 600
G/cuo

Fig. 24 Critical length, (L/d)cr vs. G/cuo for GPA – Effect of cfor φgp = 350, λ=1.3 & β =
1.0 in non-homogenous ground.

27
c=0
0.5
0.7
(L/d)cr

17

7
28 35 42
Angle of Shearing Resistance,  gp

Fig. 25 Critical length, (L/d)cr vs. gp for GPA–Effect of c for G/cuo=200, λ=1.3 & β = 1.0
in non-homogenous ground.

23
 27
c=0

0.5
(L/d)cr
0.7

1
17

7
0 1.25 2.5


Fig. 26 Critical length, (L/d)cr vs.  for GPA –Effect of cfor G/cuo =200, φgp =350 &
β=1.0 in non-homogenous ground.

The variations of the critical length, (L/d)cr of GPA with the parameters and β for

different non-homogeneity parameter, c, can be seen in Figs. 26 and 27 respectively.

24
 50

c=0

0.5
(L/d)cr

25 0.7
1

0
0 0.8 1.6


Fig. 27 Critical length, (L/d)cr vs. for GPA –Effect of c for G/cuo =200, φgp =35 0 &
=1.3 in non-homogenous ground.

4. LOAD - DISPLACEMENT RESPONSE OF GP ANDGPA

Settlement of a granular pile under working loads is similar to that ofincompressible

floating pile in a half space, with correction for the effect of pile compressibility and is

given (Poulos and Davis 1980) as

P .I
  (13)
K s .d

where I  I o . R k . R h . R  , ρ - settlement at top of GP, P - applied axial load,

Io- settlement influence factor for incompressible pile in semi-infinite mass, for νs=0.5, RK

- correction factor for pile compressibility, Rh - correction factor for finite depth of layer on

25
a rigid base, and Rν - correction factor for soil Poisson’s ratio, νs. Plot of Io is given in Fig.

28 while those for RK, Rh, and Rv in Figs. 29, 30 & 31.

1.00

db/d=1
2
Io

0.10
3

0.01
0 10 20 L/d 30 40 50

Fig. 28 Settlement Factor Io for L/d=100 and νs=0.5 for Incompressible Pile (Poulos and
Davis, 1980)

26
 3

Values of

100
50
RK 2
25

10

2
1
1
10 100 1000 10000
K

Fig. 29 Correction factor for compressibility, RK for νs=0.5 (Poulos and Davis, 1980)

1
Values of

0.8 50

25
0.6 10
Rh 5
0.4

2
0.2
1

0
1 2 0
0.5
Fig. 30 Correction factor for Finite Layer, Rh (Poulos & Davis, 1980).

27
 1

0.95

0.9
K=100
R
500
0.85
2000
1000
0.8

0.75
0.0 0.1 0.2 0.3 0.4 0.5
s

Fig. 31 Correction factor for Poisson’s ratio, R (Poulos & Davis, 1980).

Displacements, u, of GPA under working loads are estimated in a similar manner as

that for GP, i.e., treating it as a compressible pile subjected to pullout.

P .I
 u  (14)
E s .d

where P is the pullout load and I the influence coefficient for upward displacement. The

top ρu0 and tip displacements, ρuL are represented respectively by the displacement

influence coefficients IUO and IUL.

28
 2

50

25

1.5

10
IUO

L/d=5
1

0.5
10 100 1000
K
Fig. 32 Normalized tip displacement, IUO vs. K, for νs=0.5 – Effect of K.

13
IUL

6.5
50

25

10

L/d=5
0
10 100 1000
K
Fig. 33 Normalized top displacement, IUL, vs. K, for νs=0.5 – Effect of L/d.

29
The variations of displacement influence coefficients of GPA at the tip , IUO and at the top,

IUL, with relative pile stiffness, K for different L/d are given in Figs. 32 and 33.

CONCLUSIONS

Solutions and results for the ultimate capacities of GP in compression and GPA in

pullout are presented for homogenous (undrained shear strength constant with depth) and

non-homogenous (undrained shear strength increasing linearly with depth) ground

conditions. The ultimate capacities are reported as the lesser of the pile and bulging

capacities. The ultimate capacities of GP and GPA are functions of granular pile and in-

situ ground properties, viz., the unit weight, gp and angle of shearing resistance, gp of

granular pile material; and unit weight, s, undrained strength, cu, the rigidity modulus, G

and the non-homogeneity strength parameter, c of the soil.

Variations of ultimate capacities in GP and GPA with L/d are presented as functions of

G/cu, gp, unit weight parameter, and lateral confinement pressure parameter, for

homogenous and non-homogeneous ground conditions. The transition from pile to bulging

capacity with the L/d is termed as the critical length, (L/d)cr. Variations of (L/d)cr as

functions of relevant parameters including the non-homogeneity strength parameter, c are

presented.

Displacements of GP and GPA are presented considering them as compressible pile, in

situ soil to behave linearly and the in situ ground to be homogeneous. The elastic

continuum approach of Poulos & Davis (1980) is extended to predict displacement

responses of GP and GPA. The variations of normalized displacement influence

coefficients with L/d and K are presented.

30
REFERENCES

Aboshi H., Ichimoto E., Enoki M. and Harda K. (1979).The Composer-A Method to
Improve Characteristics of Soft Clays by Inclusion of Large Diameter Sand Columns.
Proceedings of International Conference on Soil Reinforcement: Reinforced Earth and
Other Techniques, Paris, Vol.1: pp. 211-216.

Aboshi, H. and Suemastu N. (1985). The Sand Compaction Pile Method: State-of-The-Art-
Paper. Proceedings of 3 rd International Geotechnical Seminar on Soil Improvement
Methods, Nanyang Technological Institution, Singapore.

Barksdale R. D. and Bachus R. C. (1983).Design and Construction of Stone Columns.

Report No. FHWA/RD-83/026, U. S. Department of Transportation, Federal Highway
Administration, Washington, D. C., pp. 194.

Datye K.R. and Nagaraju S.S. (1975).Installation and Testing of Rammed Stone Columns.
Proceedings of Indian Geotechnical Society Specialty Sessions, 5th ARC on SMFE,
Bangalore, pp. 101-104.

Gibson R. E. and Anderson W. F. (1961).In Situ Measurement of Soil Properties with the
Pressuremeter. Civil Engineering Publication Works Revision 56, pp. 615-618.

Greenwood D.A. (1970).Mechanical Improvement of Soft Soils below Ground Surface.

Proceedings of the Ground Engineering Conference, Institute of Civil Engineers,
London, June 11-12, pp. 9-20.

Hughes J.M.O. and Withers N.J. (1974). Reinforcing of Soft Cohesive Soils with Stone
Columns. Ground Engineering Journal, Vol. 7, No. 3, pp. 42-49.

Hughes J.M.O., Withers N.J. and Greenwood D.A. (1975). A Field Trial of the Reinforced
Effect of Stone Column in Soil. Geotechnique, Vol. 25, No.1, pp. 31-44.

Madhav, M.R. and Vitkar, P.P. (1978). Strip Footing on Weak Clay Stabilized with
Granular Trench or Piles. Canadian Geotechnical Journal, Vol.15, No.2, pp. 605-609.

Madhav M. R., Iyengar N. G. R., Vitkar P. P. and Nandi S. A. (1979). Increased Bearing
Capacity and Reduced Settlements due to Inclusions in Soil. Symposium on
Reinforcements of Soils, Paris, Vol. 2: pp. 329-333.

Vesic A.S. (1975). Bearing Capacity of Shallow Foundations. Foundation Engineering

Handbook, Van Nostrand Reinhold, pp. 121-144.

31