Vous êtes sur la page 1sur 34


3.1 INTRODUCTION shaft and base resistances, less the weight of the pile; that
There arc two usual approaches to the Galculation of the
ultimate load capacity of piles: the "static" approach,
which uses the normal soil-mechanics method to calculate
the load capacity from measured soil properties; and the
"dynamic" approach, which estimates the load capacity
Psu ultimate shaft resistance
of driven piles from analysis of pile-driving data. The
Pbu ultimate base resistance
first approach will be described in detail in this chapter,
W = weight of pile
and the second in Chapter 4.
In this chapter, a general expression for the ultimate
Psu can be evaluated by integration of the pile-soil
load capacity of a single pile is given and its application to
shear strength Ta over the surface area of the shaft. Ta
piles in clay and sand is described. Approaches for groups
is given by the Coulomb expression
of piles in clay and sand will then be outlined. Other topics
include the design of piles to rock, the use o.f in-situ tests
Ta = Ca + On tan rPa (3.2)
such as the standard penetration test and the static cone
to estimate pile-load capacity, the calculation of uplift
resistance of piles and grou'ps, and the load capacity of
bent piles.
Ta pile-soil shear strength
Ca = adhesion
3.2 ULTIMATE LOAD CAPACITY OF SINGLE P1LES On = normal stress between pile and soil
rPa angle of friction between pile and soil
;,.2.1 General Expression It is an implicit assumption of Eq. 3.1 that shaft and base
resistance are not interdependent. This assumption cannot be
The net ultimate load capacity, Pu , of a. single pile is strictly correct, but there is !ittk doubt that it is correct enough
generally accepted to be equal to the sum of the ultimate for practical purposes for all normal-proportion piles and piers.

an is in turn frequently related to the vertical stress av , base resistance of piles, reliance has to be placed on approx
as imate theoretical or semiempiricaJ methods. With regard
to: sands, these methods have been reviewed by Vesic
(1967), who fmrnd that the solution of Bcrezantzev et al.
(I 96 l ) generally fitted experimental results best.
where From Eqs. (3.1), (3.5), and (3.6),

Ks coefficient of lateral pressure
Pu Cf ca + avKs tan <Pa) dz
+ A b (CNc + O v bNq + 0.5-ydN"f) W (3.7)
Equation (3.7) is a general expression for the ultimate
and load capacity of a single pile. If the undrained or short
term ultimate load capacity is to be computed, the soil
parameters c, , Ca, and r should be values appropriate
to undrained conditions, and a v and av b should be the
L total stresses. If the long-term ultimate load capacity of
"" f O C(c a + avKs tan <Pa )dz (3.5)
piles in sand is required, the soil parameters should be
drained values, and a and a b the effective vertical
where stresses. The vertical stresses are usually taken to be the
overburden stresses, and for clays, this is probably true
C pile perimeter enough, even close to the piles. However, for sands, there is
L length of pile shaft now clear evidence implying that the vertical stress near the
pile may be less than the overburden. This matter is dis
It is usually accepted that the ultimate resistance Pb u
cussed in greater detail in Section 3.2.3.
can be evaluated from bearing-capacity theory as
For steel H-piles, two modes of failure of the shaft
are possible: (a) the development of the limiting pile
soil shear strength along the entire surface area of the
pile; and (b) the development of the limiting pile-soil
shear strength along the outer parts of the flanges, plus
the development of the full shear strength of the soil along
Ab = area of pile base
the plane joining the tips of the flanges-that is, the soil
c cohesion of soil
within the outer boundaries of the pile effectively forms
av b = vertical stress in soil at level of pile base
part of the pile shaft. Therefore, when using Eq. (3.7),
r unit weight of soil
the ultimate skin resistance, Psu , should be taken as the
d pile diameter
lesser of the two values.
Ne , Nq , N'Y bearing capacity of factors, which are
primarily functions of the angle of
internal friction <f; of the soil, the relative 3 .2.2 Piles in Clay
compressibility of the soil and the pile

Rigorous solutions for the bearing capacity of surface For piles in clay, the undrained load capacity is generally
footings using the methods of classical plasticity are now taken to be the critical value unless the clay is highly
well-established (Prandtl, r 1923; Sokolovskii, 1965; Cox, overconsolidated. (Burland, 1973, contends, however,
1962; Lundgren and Mortensen, 1953; Davis and Booker, that an effective stress-drained analysis is more appropriate,
1971), and tl1e only doubts regarding the practical validity as the rate of pore-pressure dissipation is so rapid that
of these solutions lie in the possible effects of the differ for normal rates of load application, drained conditions
ences between the behavior of real soil and that of the ideal generally prevail in the soil near the pile shaft.) If the clay
material assumed in the theory. At the present time, there is .saturated, the undraiQed angle of friction <Pu is zero,
are few if any classical plasticity solutions that are relevant and <Pa may also be taken as zero. In addition, Nq l
to a buried footing, and therefore, for the calculation of and N"f O for= 0, so that Eq. (3.7) reduces to

(1974); are shown in Fig. 3.1. It is generally agreed that

for soft clays (cu 24 kPa), ca/cu is 1 (or even greater*);
however, for driven piles in stiff clays, a wide scatter of


Cu == undrained cohesion of soil at le_vel of pile base
Ca undrained pile-soil adhesion

Further simplification is possible in many cases, since Penetration

for piles without an enlarged base, Abavb = W, in which Case Soil Conditions Ratiob Ca/Cu
case Sands or sandy soils <20 1.25
overlying stiff
(3.10) cohesive soils >20 See Fig. 3.2

II Soft clays or silts <20 (>8) 0.40

Undrained Pile-Soil Adhesion Ca overlying s1iff
The undrained pile-soil adhesion Ca varies considerably cohesive soils >20 0.70

with many factors, including pile type, soil type, and Ill Stiff cohesive soils <20 (>8) 0.40
method of installation. Ideally, Ca for a given pile at a without overlying
given site should be determined from a pile-loading test, strata >20 See Fig. 3.3
but since this is not always possible, resort must often be
a After Tomlinson (1970).
made t_o empirical values of Ca . Many attempts have been
. r _ Dep1h of penetration in stiff clay
made to correlate Ca with undrained cohesion Cu , notably b p enet ratwn ra 10 - Pile diameter
Tomlinson (1957, 1970), Peck (1958), Woodward et al. NOTE]: Adhesion factors not applicable to H-section piles.
(1961), Coyle and Reese (1966), Vesic (1967), Morgan NOTE 2: Shaft adhesion in overburden soil for cases I and
II must be calculated separately.
and Poulos (1968), McClelland et al. (1969), McClelland
(1972), and McClelland (1974).
For driven piles, the rapid dissipation of excess po_re pressures
For driven piles, typical relationships between ca /cu due to driving may result in a locally overconsolidated condition,
and Cu , based on the summary provided by McClelland and hence a value of c a even greater than c u for the unaffected soil.

25 50 75 100 125 150 175

1.0 -----.------------------------,

K eme

..........__ -

' ',
'- Woodward

', ',, - .............. ,.........



0 '-----'-----'------'------'------'------'----'------'
0 0.5 10 1.5 2.0 2.5 3.0 3.5 4.0
Undrained Cohesion c 0 kips/sq ft
FIGURE 3.1 A dhesion factors for driven piles in cl ay (after McClelland, 1974).

Undrained shear strength (c, I kN/m 2

50 100 150 200

o 21
Figures denote penetration ratio
Depth of penetration in clay
o 21 Pile diameter
1.5 Key
Steel tube piles
Precast concrete
10 piles

o 27
170 23
44. O 0
0.5 60 ... 56

Design curve for
penetration ratio 20

o------------------,- ------"---'
1000 2000 3000 4000
Undrained shear strength !c 0) lu/ft2

FIGURE 3.2 Adhesion factors for case I (sands and gravels overlying stiff to very stiff cohesive soils) (Tomlinson, 1970).

Undrained shear strength le,) kN/m2

50 100 150 200 250
Figures denote penetration ratio
Depth of penetration in clay

1.5 Pile diameter

Steel tube piles
Precast concrete
1.0 O piles

Design curve for

penetration ratio > 20
1000 2000 3000 4000 5000

Undrained shear strength {c 0 I lb/ft 1

FIGURE 3.3 Adhesion factors for case III (stiff to very stiff clays without overlying strata) (Tomlinson, 1970).

values of ca/c u is evident. This scatter is often attributed be little data on appropriate values of Ca for driven piles
to the effects of "whip" during driving. A more complete founded in very sensitive clays, and the extent to which
investigation of adhesion for driven piles in stiff clay has "set-up" compensates for remolding can at present only
been made by Tomlinson (1970), who found that ca/cu be determined by a load test.
may be markedly influenced by the soil strata overlying For bored piles, the available data on ca/cu is not as
the clay, as well as by the value of Cu . Tomlinson has sug extensive as for driven piles, and much of the data that is
gested the adhesion factors shown in Table 3.1 and Figs. available is related to London clay.cTable 3.2 gives a sum
3.2 and 3.3 for cu > 1000 lb/sq ft (48 kPa). The most mary of adhesion factors, one of which is expressed in
notable feature of Tornlinson's results are the high values terms of remolded strength, c,, as well as the undisturbed
of ca/cu for case I, where sand or sandy gravel overlies undrained strength, cu . Results obtained by Skempton
the clay, because of the "carrying down" of a skin of the (1959) and Meyerhof and Murdock (1953) suggest that
overlying soil into the clay by the pile. There appears to an upper limit of Ca is 2000 lb/sq ft (96 k.Pa).

TABLE 3.2 ADHESION FACTORS FOR BORED PILES IN CLAY A somewhat different approach to the calculation of
the ultimate shaft capacity Psu has been adopted by Vijay
Adhesion vergiya and Focht (1972) for steel-pipe piles. From an
Soil Type Factor Value R-eference examination of a number of loading tests on such piles,
they concluded that Psu can be expressed as follows:
London clay 0.25-0 7 Golder and Leonard
Average, ( 1954)
0.45 Tomlinson (195 7) (3.lOa)
Skempton (195?)
Sensitive cby Golder (1957)
o 1 mean effective vertical stress between ground
Highly expnsive Ca/cu 0.5 Mohan and Chandra
clay (1961) surface and pile tip
Cm = average undrained shear strength along pile.

0 0.1 0.2 0.3 0.4 0.5

0r- -----------,r--------,-------,r----::=""I



- 100

Loec:tion Syn;Ljo/ Source

Detroit D , Housel
Morganz.a Mansur
a:: 0 Cleveland 0 Peck
Dray:on X Peck
North Sea 6 Fox

Lemoo"e 0 Woodward
Stan;riore Torr:linson
i50 ' New Orleans Blessev
Venice McClelland

0 Alliance v' McClelland
oriaidsonville Darragh
MSC Hosston McCle!lano
175 San Francisco Seed
Bri:ish Columbia


22:5 ,_______,___,_____ ------'------------'

flG!JRE J.4 Fiictional capacity coefficient ii. vs. pile penetration (Vijayvergiya and Focht, 1972).

10 --,---,---,---,----,---,------,-----,----,
strain behavior of the soil. From an analysis of the expan
9 sion of a cavity in a mass, Ladanyi (1963) found that
for insensitive clays, 7.4 < Ne < 9 .3, depending on the
8 stress-strain behavior of the soil. This analysis broadly
confirmed the earlier analysis of Bishop et al. (1945), which
7 gave the following result f or a circular base (as quoted by


Rectangular base Lr I.. Br

For piles in stiff, overconsolidated clays, the drained loaJ
2 capacity, rather than the undrained, may be the critical
. { Lr Br)
value, and Vesic (1967, 1969) and Chandler (1966, 1968)
Br] have advocated an effective-stress approach in such cases.
[0 84 + 0 16 -;---
If the simplifying assumption is made that the drained
0 2 3 4 pile-soil adhesion t:' is zero and that the tenus in Eq.
(3. 7) involving the bearing capacity factors Ne and
can be ignored, the drained ultimate load capacity fr om
FIGURE 3.5 Bearing-capacity factors for foundations in clay ( (3 .7) may be expressed as
0) (after Skempton, 1951).
Pu -.. fL cOv, r,.vs tan a, dz 13J2,
A s = pile surface area +
A = dirnensionlt:ss coefficient

In effect, the average pile-soil adhesion factor is then where

OvJ effective vertical stress at depth z
(3.1 Ob) <Jvb effective vertical stress at level of pile base
, drained angle of friction between pile and soil

/\ was found to be a function of pile penetration and is Burland (1973) discusses appropriate values of the
plotted in Fig. 3.4. combined parameter {3 Ks tan if, and demonstrates that
Equation (3. lOa) has been used extensively to predict a lower limit for this factor for normally consolidated clay
the shaft capacity of heavily loaded pipe-piles for offshore can be given as
(3 = (1 . sin ') tan ' (3.13)
Bearing Capacity Factor Ne
The value of Ne usually used in design is that proposed by where
Skempton (195 I) for a circular area, which increases from
6 .14 for a surface foundation to a limiting value of 9 for ',' = effective stress friction angle for the clay
length ? 4 diameters (Fig. 3.5). The latter value of Ne
9 has been confirmed in tests in London clay (Skempton, For values of , in the range of 20 to 30 degrees, Eq.
19 59) and has been widely accepted in practice. However, (3.13) shows that ;3 varies only between 0.24 and 0.29.
differing values have been found by other investigators; This range of values is consistent with values of i3 Ks
for example, Sowers (1961) has found 5 < Ne < 8 f or tan deduced from measurements of negative friction
model tests, and Mohan (1961) has found 5.7 < Ne < on piles in soft clay (see Figs. 11.26 and 11.2 7). Meyerhof
8.2 for expansivt: clays. The variations in the value of (1976) also presents data that suggests similar values of
Ne may well be associated with the influence of the stress- (3; however, there is some data to suggest that /3 decrea1.es

Average skin resistance (lb/in. 2)

0 2 3 4 5 6
with increasing pile length, and that for long piles (in
excess of about 60 m), {3 could be as low as 0.15.
For piles in stiff days, Burland suggests that taking
Ks = K0 and = the remolded friction angle, gives an
upper limit to the skin friction for bored piles and a lower
limit for driven piles. Meyerhof (1976) presents data
indicating that Ks for driven piles in stiff clay is about
1.5 times K0 , while Ks for bored piles is about half the
value for driven piles. For overconsolidated soils, K0
can be approximately estimated as
g' 60 f----------lho-----+--+- -
.5! K0 = (1 sin') ,,/cicR (3.14)
OCR overconsolidation ratio
Field :ests I loose. It is inferred that can be taken as ', the drained fric
1001--- -ll-J-a.-+-,moist, sand)-(G-4) --------1.'-----l
- i

tion angle of the clay.
In the absence of contrary data, a and ab may be
Medium dense sand (G-21
taken as the effective vertical overburden stresses. Values
120 .... of l'v'q may be taken to be the same as for piles in sand;
these values are plotted in Fig. 3 :11.
Loose\and ( G- 1 l :
140 ._____.___,____,___.__._____,__,___....__,___,
0.1 0.2 0.3 0.4 3.2.3 Piles in Sand
Average skin resistance ltons/f: 2 )
\ Conventional methods of calculation of the ultimate load
FIGURE 3.6 Varial!on of skin 1esistance with pile length (Vesic,
l96Ti. capacity of piles in sand (Broms, 1966; Nordlund, 1963)
Point resistance (lb/in.2)
0 100 200 300 400 500 600 700 800 900

Pile diameter 4.0 in.



I 60 \

-s \

Field tests
so: ,\
I loose, moist, sandI
\ G-41

Loose Medium dense sand
sand IG2)
(G-1) I I
120 I

10 20 30 40 50 60
Point resistance (ton/ft2 )

FIGURE 3.7 Variation of point resistance with pile length (Vesic, 1967).

assume that the vertical stresses a,, and Ov b in Eq. (3.7)
are the effective vertical stresses caused by overburden.
However, extensive research by Vesic (1967) and Kerisel ---,-- WT __
(1961) has revealed that the unit shaft and base resistances z,
of a pile do not necessarily increase linearly with depth,
but instead reach almost constant values beyond a certain ----
o' vc
- -
depth (Figs. 3.6 and 3.7). These characteristics have been
confinned by subsequent research (e.g., BCP Comm., L
1971; Hanna and Tan, 1973). Vesic also found that the
ratio of the limiting unit point and shaft resistances, fb ffs ,
cf a pile at depth in a homogeneous soil-mass appears to
be independent of pile size, and is a function of relative
density of the sand and method of installation of the f--;-d

piles. Relationships between fb lfs and angle of internal FIGURE 3.9 Simplified distribution of vertical stress adjacent to
friction ('), obtained by Vesic, are shown in Fig. 3.8. pile in sand.

The above research indicates that the vertical effective

stress adjacent to the pile is not necessarily equal to the
Some design approaches have effectively incorpo,ated
effective overburden pressure, but reaches a limiting
value at depth. This phenomenon was attributed by Vesic Vesic's findings by specifying an upper limit to the shaft
and base resistances. Eor example, McClelland et al.(1969)
to arching and is similar to that considered by Terzaghi
(1943) in relation to tynnels. There are however other have suggested, for medium-dense clean sand the following
hypotheses, such as arching in a horizontal plane, which design parameters:= 30 ;Ks 0.7(compression loads)
might explain the phenomena shown in Figs. 3.6 and 3.7. or 0.5 (tension loads), with a maximum value of shaft
resistance fs of 1 ton/ft 2 {9t kNjm2 ); and Nq 41, with
500.-------------- a maximum base resistancefb of 100 ton/ft 2 (9.6 MN/m2).
4001-----+-----+-----+- ---l--------! However, such approaches take little account of the nature
of the sand and may not accurately reflect the variation
Ves:c's tests
of pile capacity with pile penetration, as the limiting
in sand resistances generally will only become operative at relative
2001-----+--- - -+----4-'<-- --- ly large penetrations(of the order of 30 to 40 m).
In order to develop a method of ultimate load pre

diction that better represents the physical reality than
the conventional approaches, and yet is not excessively
100 ---.. - --+-
- ---+------+--
complicated, an idealized distribution of effective vertical
,; stress a with depth adjacent to a pile is shown in Fig.
3.9. a is assumed to be equal to the overburden pressure
u 501-- ---+----+ ----+.',/L--t---.., to some critical depth Zc , beyond which a remains con
"' 40t-------i-- -- --+---- --c"',-----t---, stant. The use of this idealized distribution, although
/ simplified, leads to the two main characteristics of behavior
G. /
observed by Vesic: namely, that the average ultimate skin
;:, resistance and the ultimate base resistance become con
stant beyond a certain depth of penetration.
Georgia Hwy. Dept.
tests in
If the pile-soil adhesion Ca and the term cNc are taken
rnicaceous silts as zero in Eq. (3.7), and the term 0.5,d Ny is neglected as
10- --+-- -- -+----+--- -+----l I being small in relation to the tenn involving Nq , the ulti
Range of observed mate load capacity of a single pile in sand may be expressed
values in saturated as follows:

0 10 20 30 40 (3.15)
Ang!e of shearing resistance I degrees)

FIGURE 3.8 Variation of f0/fs with q, (Vesic, 1967). where


a:, = effective vertical stress along shaft results, it may be possible to derive different relationships
- effective overburden stress for z ,;;; Zc or limit for different pile materials.
ing value a ;,c for z > z c For bored or jacked piles, the values of Ks tan in
a b - effective vertical stress at level of pile base Fig. 3.10b are considered to be far too large, and it is sug
Fw - co rrection factor for tapered pile (= I for gested that values derived from the data of Meyerhof
uniform diameter pile) (I 976) are more appropriate fo r design. These values are
shown in Fig. 3.10c, and have been obtained by assuming
On the basis of the test results of Vesic (1967), values <!> 0.75'. The values for bored piles are reasonably
of Ks tan and the dimensionless critical depth z c/d consistent with, althoug.11 more conservative than, those
have been evaluated. Vesic's results are presented in terms recommended by Reese, Touma, and O'Neill (1976).
of the relative density Dr of the sand, but the results may Also shown are values of Ks tan; tor driven piles, derived
also be expressed in terms of the angle of internal friction from Meyerhofs data; these latter values are considerably
', by using a relationship such as that suggested by Meyer smaller (typically about one half) of the values given in
hof (1956): Fig. 3 .1 Ob. Some of this difference may lie in the method
of interpretation of the data of Vesic and others by Meyer
' == 28 + lSDr (3.16) hof, which leads to smaller values of Ks tan associated
with larger values of .,;c/d.
Relationships between Ks tan and ,p, and Zc /d and The bearing capacity factor Nq is plotted against
', are shown in Fig. 3.10. In a layered-soil profile, the in Fig. 3.11, these values being based on those derived by
critical depth Zc refers to the position of the pile embedded Berezantzev et al. (1961). Vesic (1967) has pointed out
in the sand. It should be emphasized that these relation that there is a great variation in theoretical values of Nq
ships may be subject to amendment in the light of further derived by different investigators, but the values of
test results. For example, at present, the dependence of Berezantzev et al. appear to fit the available test data best.
I<s tan on the pile material is not defined. Vesi,:;'s tests The solutions given by Berezantzev et al. indicate only
were carried out on steel tube piles, but the values of Ks a small effect of relative embedment depth L/d, and the
tan derived from these tests appear to be applicable to curve in Fig. 3.11 represents an average of this small range.
other pile materials. However, in the light of future test The curves given by Meyerhof (1976) show a larger effect

For driv<2n p1l<2s 0 = 1/,i 0'1+10 (Fig.3.10a,F,g.3.10b)
For bored piles,0 =01 -3 (F19.310a), 0=01 (Fig 3.10c)
where ',2)1 angl<Z of internal f iction prior to

1nstallot1on of pile

(a) Z C /d VS /l) {b) K5 tan 0'0 vs 0 0 Based

{ c) Values of K5 ton 0
{ Driven Piles) on Meyarhol ( 1976)
20 30

15 25 1 2 1---l----4-

20 I
10 0-8

5 15

35 40
0 0'1

FIGURE 3.10 Values of and Ks tan o for piles in sand.


c;'J', +40 (a) Driven Piles

For driven pil<2s , (l) = - -
(a) For the determination of Nq , the values of beneath
F or bored piles, 0 = 0'1 3
the pile top should be taken as the final value subsequent to
wh<2n2 0' = ongl<2 of int,;,rnol friction
prior to 1nstallat1ori of p,I
driving, as given by Kishida (1967):
1000 1
1 + 40
rt> = (3.17)


Nq 100
/ rt>
1 = angle of internal friction prior to installation of
the pile

(b) For. the determination of Ks tan and z c/d, the

value of along the pile shaft should be taken as the mean
of the values prior to, and subsequent to, driving; that is,
25 30 35 40 45
FIGURE 3.11 Relationship between Nq and (after Berezantzev
et al., 1961).
(b) Bored Piles
of L/d; however, the curve of Fig. 3.11 also lies near the (a) For the determination of Nq and zc/d, it is suggested
middle ofMeyerhofs range. that the value of ,P be taken as rt> 'i 3, to allow for the
Values of the taper correction factor Fw are plotted possible loosening effect of installatipn (see Section 2.4).
against <P in Fig. 3.12 and have been derived from the re (b) For Ks tan ,P, Fig. 3.I0c should be used, taking the
sults of the analysis developed by Nordlund (1963). value of\ directly.
In applybg the results in Fig. 3.10 to Fig. 3.12, it is
suggested that the following values of be used to allow The above suggestions may also require modification
for the effects of pile installation. in the light of future investigations. Furthem10re, if jetting
is used in conjunction with driving, the shaft resistance may
decrease dramatically and be even less than the value for
a corresponding bored pile.
McClelland (1974) has reported tests in which the use
of jetting with external return flow followed by driving
reduced the ultimate shaft capacity by about 50%, while
installation by jetting alone reduced the ultimate shaft
capacity to only about 10% of the value for a pile installed
by driving orJy.
Another case in which caution should be exercised is
when piles are to be installed in calcareous sands. Such

sands may show friction angles of 35 or more, but have
beer:i found to provide vastly inferior support for driven
piles than normal silica sands. In such cases, McClelland
(1974) suggests limiting the skin resistance to 0.2 tons/ft2
(19 kN/m 2) and base resistance to 50 tons/ft2 (4800
kN/m 2 ). In such circumstances, drilled and grouted piles
may provide a more satisfactory s_olution than wholly
0.5 1.0 1.5 2.0 driven piles.
Pile !aper angle w"
In many practical cases, only standard penetration
FIGURE 3.12 Pile taper factor F1.,J (after Nordlund, 1963). test data may be available. The value of I/> may be esti-

3 0;. 40
For drMrn pil<Zs, 0 = 01+ 10 For drivczn pilczs. 0 = -- -
wh(lr<Z. 0; angle of internal frlction
For bored p1lczs, 0 = 0; -3
prior to 1nstallaf1on of pilcz
whczre 01 -angle of intczrnal tristion
pr 1or to installation of pile

Volu<Zs of L/d


100 l..-.l.....J--1....i......l-J......L....l......J_..l-.L-..J,__I............
28 32 36 40 44
FIGURE 3.14 Dimensionless ultimate base-load capacity for pile
in uniform sand.

10 against in Fig. 3.14. The value of L/d does not generally

32 36 40 44
have a marked effect on the ultimate base load unless
is relatively large, that is, for dense sands.
FIGURE 3.13 Dimensionless ultimate shaft-load capacity for pile
in uniform sand. The use of a high value of, for very dense sands (say,
, > 40 ) simultaneously for the shaft and the base, should

also be treated with caution, since the full base resistance

mated from a correlation such as that given by Peck, may well only be mobilized after a movement sufficient
Hansen, and Thorburn (1974), or by the following em for the operative value of , along the shaft to be signifi
pirical relationship suggested by Kishida (1967): cantly less than the peak.
If the pile is founded in a relatively thin, firm stratum
(3 .19) underlain by a weaker layer, the ultimate base load may
be governed by the resistance of the pile to punching into
where the weaker soil. Meyerhof (1976) shows that if the weaker
layer is situated less than about 10 base diameters below
N standard penetration number the base, a reduction in base capacity can be expected;
he suggests that in such cases, the ultimate point resistance
A more detailed discussion of the relationship between can be taken to decrease linearly from the value at lOdb
'1 and N, and also '1 and relative density D,, is given above the weaker layer to the value at the surfac of the
by de Mello (1971). weaker layer.
For the case of a driven pile in a uniform layer of The suggested approach of ultimate load calculation
sand, dimensionless values of the ultimate shaft load and has been applied to 43 reported load-tests on driven piles.
ultimate base load may be derived using Eq. (3.15) and The details of the parameters chosen for the calculations
Figs. 3.10, 3.11, and 3.12. In Fig. 3.13, the dimensionless are given in Table 3.3, and the comparison between cal
ultimate shaft load Psu frd3 is plotted against for various culated and measured ultimate loads is shown in Fig. 3.15.
values of L/d; "i is the effective unit weight of the soil The mean ratio of calculated to observed ultimate loads is
above the critical depth Zc, The marked increase in ultimate 0.98, with a standard deviation of 0.21. It should be
shaft load with increasing L/d and is clearly shown. The pointed out that the ultimate load of all piles considered
dimensionless ultimate base load Pbu ftdA b is plotted in the comparison is less than 300 tons. The use of this

Tests reported by Nordlund (1963)

o Power plant site-area I
+ Mojave River bridge CAPACITY OF PILES IN SAND
X Port Mann
'v Buffalo Bayou
9 0 St. viaduct Reference Case Remarks
300 Field tests reported by Vesi (1967)
-0- Ultimate skin loads
D Ultimate base loads Nordlund (1963) Power Plant Values of</)\ suggested
Site-Areas by Nordlund used
Field tests reported by Desai ( 1973)0
Field tests by Tavenas (1971) I & II
250 X
Mojave River Upper 14 ft of sand
Bridge assumed to have lower
</); (38 ) than lower
., depths (,p11 40 )
200 --0-

U) 0 because of jetting
2 0 during installation
a: 150
Port Mann Values of ,p suggested
9 Xo
by Nordlund used
:i if' 0 Buffalo Bayou Values of ,p; suggested
by Nordlund used
100 Interchange Vertical stress due to
soil above excavation
level ignored
0 Street As above;H-pile treated
Viaduct as a sql!are pile

Vesic (1967) PilesHll From reported N values,

0 l<____L.__--1.___,...___..J_____L_____,
H16 & H2 following values of ,p
0 50 100 150 200 250 300
chon: 0-12 ft, 1,
Measured Pu (U.S. tons)
33 ; 12-30 ft,</); 38 .

FIGURE 3.15 Comparison between calculated and measured ulti 30+ ft,,p', 42
mate load capacity of driYen piles in sand.
Desai (1973) Piles 2,3,10 ,P , assumed to be 33 ,
constant with depth
approach for piles of much larger capacity-those used
Tavenas (1971) PilesH2-6, ,P ', assumed to be 3 3 ,
for offshore structures for example-should be treated constant with depth
with caution. Indeed, for relatively short, larger-diameter
piles, the average values of shaft resistance given by this
approach are considerably larger than those normally
adopted for design purposes (for example, the values sug standard penetration value, N, of about I 6. The water table
gested by McClelland et al., 1969). These high values was 3.4 m below the surface. On the basis of the available
data, the following values were adopted:
arise because of the combination of high values of Ks
tan c/J (Fig. 3. lOb) with a relatively large critical depth. (a) Bulk unit weight above water table 17.3 kN/m 3
In such ,;ases, a more conservative estimate of shaft resis (b) Submerged unit weight below water table = 7.8
tance may be desirable for design, based on the values of kN/m3
Ks tan c/J derived from Meyerhof (1976) and shown in (c) Angle of internal friction angle prior to installation:
Fig. 3.10c.

To illustrate the application of the suggested method c/J'1 = 25 (0 2.4 m)
l/J'1 = 32

of calculation, the following example details calculations (2.4 - 18.3 m)
c/J '1 = 30

(18.3 - 20.8 m)
for two of the pile tests reported by Nordlund (1963).
c/J '1 33 (> 20.8 m)

Illustrative Example Considering first the ultimate skin resistance, the

The piles considered are Piles B and A from the Power values of c/J given by Eq. (3.18) are as follows:
Plant Site, A.rea I, Helena, Ark. Pile B was a closed-end

steel-pipe pile, 24.4 m long and 0.32 m in diameter, driven c/J = 28.75 (0-2.4 m)

into fine sand grading to coarse and having an average c/J = 34 (2.4-18.3 m)

, 32.5 (I 8.3-20.8 m) 0+41.52

Pu = ( . 2 ) X7TX0.51 X3.35 X1.00X2.4'

, 34.75 (> 20.8 m)

From Fig. 3.l 0b, the values of Kstan,_are 1.00 (0-2.4 +

[<41.s2;44l 2)X 0.15 + 44.12 X(10 2.55)]
m), 1.30 (2.4-18.3 m), 1.18 (18.3-20.8 m), 1.35 (20.8 m).
If it is assumed that the critical depth is less than 2.4 m 0. 02
below the surface. then for d., = 28.75 ,

= 5.0, from Xrr X0.33 'I. 4.1X1.30 + 44.12XrrX
Fig. 3.!0a; that is, Zc = 5.0 X 0.32 1.56 m. Thus, the
assumption is justified. X88
At the critical depth, the effective overburden stress = 2243 kN (252.2t)
26.99kN/m 2 Tne measured ultimate load for pile was 2400 kN
(270 t).
Because the pile has uniform diameter, Fw = I.
For the ultimate base resistance, the value of r/> given
by Eq. (3.17) is 36 .5 . From Fig. 3.11, the value of Nq

is 98.
Substituting into Eq. (3.15),
In examining the behavior of pile groups, it is necessary
+2 99 to distinguish between two types of group:
P,, = ,, X 0.32{[(O ) X1.55 + 26.99 X
(a) A free-standing group, in which the pile cap is not
(2.4 l.56)] Xl.00 +26 .99X( l8.3-2.4) in contact with the underlying soil.
X 1.30 + 26. 99X(20.8 18.3) (b) A "piled foundation," in which the pile cap is in con
tact with the underlying soil.
X l. l 8 +26.99X(24.4 20.8)X1.31

+ 26.99 / rr X O ! 22
X 98
For both types, it is customary to relate the ultimate load
capacity of the group to the load capacity of a single pile
through an efficiency factor T), where
816 + 213
1029 kN (IIS.6 t)
T) == ultimate load capacity of group
sum of ultimate load capacities (3.20)
This com part's with the measured value of I 12 kN ( 125 t). of individual piles
Pile A was a Raymond Standard pile, 10 m long, with
a head diameter d 0.55 m and a tip diameter of 0.20 m.
The pile taper w is 1 . From Fig. 3.12, for W = 1 , Fw

= 3.35 (0 ... 2.4 m), and Fw 4.1 (2.4--18.3 m). The values 3.3.1 Pile Groups in Clay
of tan are as for pile B. Assuming again that the
critical depth is above 2.4 m, z c /d = 5.0 as before, and 3.3.U FREE-STANDING GROUPS
taking an average value of d of 0.51. Z c == 2.55 m, that is,
greater than 2.4 m. However, the difference is negligible For free-standing groups of friction or floating piles in clay,
and hence Z c will be taken as 2.55 m. At this level, the efficiency is unity at relatively large spacings, but
decreases as the spacing decreases. For point-bearing piles,
ac = 2.55X-17.3 == 44.12kN1m2 the efficiency is usually considered to be unity for all
spacings-that is, grouping has no effect on load capacity,
At2.4n;, although in theory the efficiency could be greater than
unity for closely-spaced piles that are point-bearing, f or
a;, = 24 X 17.3 41.52 kN/m 2 example, in dense gravel. For piles that derive their load
capacity from both side-adhesion and end-bearing, Chellis
Since the pile tip is founded in the second stratum,
from Eq. (3.17) is 36 and the corresponding value of
(1962) recommends that the group effect be taken into
is 88. consideration for the side-adhesion component only.
Substituting into Eq. (3.15) and using, for simplicity, Several empirical efficiency fonnulas have been used to
th<' m;;an diameter of the pile in the upper 2.4 m and tiie try and relate group efficiency to pile spacings, among
lower 7.6 m, which are the f ollowing:

(a) Converse-Labarre fonnula, was accompaniecf by the formation of vertical slip planes
joining the perimeter piles, the block of clay enclosed by
, rcn-l)m+(m-l)n] /90 the slip planes sinking with the pile relative to the general
71 - l - ::; tmn
surface of the clay. For wider spacings, the piles penetrated
individually into the clay .The critical spacing was found to
where increase as the number of piles in the group increased_
Although Whitaker's tests confirmed the existence of
m = number of rows the above two types of failure, th transition between the
n = number of piles in a row ultimate group capacity as given by individual pile failure
= arctan d/s, in degrees and that given by block failure was not as abrupt as the
d "' pile diameter Terzaghi and Peck approach suggests. In order to obtain
s == center-to-center spacing of piles a more realistic estimate of the ultimate load capacity of
a group, the following empirical relationship is suggested:
(b) Feld's rule, which reduces the calculated load capacity
of each pile in a group by l/t6 for each adjacent pile, that 5m
is, no account is taken of the pile spacing. l < --1
-: : : J5m
(c) A rule of uncertain origin, in which the calculated load
capacity of each pile is reduced by a proportion I for each
0 0 0 0
adjacent pile where
O____Q _Q_ 0

I == -ci/s (3.22)

A comparison made by Chellis (I 962), between these Clay
50 kPa L = 20 rn d = 0.3 rP
and other empirical formulas shows a considerable variation Cu

in values of 71 for a given group, and since there appears to

be little field evidence to support the consistent use of ai1y J_
empirical formula, an alternative means of estimating
group efficiency is desirable.
One of the most widely used means of estimating
group-load capacity is that given by Terzaghi and Peck r--------
Limit for bock failure

(1948), whereby the group capacity is the lesser of(a) The 30

sum of the ultimate capacities of the individual piles in the I
group; or (b) the bearing capacity for block failure of the /
group, that is, for a rectangular block B, XL,, z single pile /
(3.23) 20 I

where I
c == undrained cohesion at base of group
L = length of piles ::J

Ne = bearing capacity factor corresponding to depth 10

L (see Fig. 3.5)

c "" average cohesion between surface and depth L

Model tests on free-standing groups carried out by Whitaker

( 1957) confirmed the existence of the above two types of
failure. For a given length and number of piles in a group, 10 15

there was a critical value of spacing at which the mechanism

of failure changed from block failure to individual pile FIGURE 3.16 Example of relationship between number of piles
failure. For spacings closer than the critical value, failure and ultimate load capacity of group

obvious 'that virtually no advantage is gained by using more

piles than is required to cause failure of the group as a
block; in the example in Fig. 3.16, increasing the number of
where piles beyond about 80 produces very _little increase in ulti
mate load capacity.
Pu = ultimate load capacity of group A considerable number of model tests have been carried
P1 ultimate load capacity of single pile out to determine group efficiency factors in homogeneous
n = number of piles in group clay-for example, Whitaker (1957), Saffery and Tate
Po = ultimate load capacity of block (Eq. 3.23)- (1961), and Sowers et al. (1961). A summary of some of
these tests has been presented by de Mello (I 969) and is
Eq. (3.24)may be reexpressed as follows: reproduced in Fig. 3.17. From this summary, it may be
seen that higher efficiency factors occur for
n 2Pi
l+-2 (3.25)
Ps (a) Piles having smaller length-to-diameter ratios.
(b) Larger spacings.
where (c) Smaller numbers of piles in the group.

r, = group efficiency For spacings commonly used in practice '(2.Sd to 4d),

1) is on the order of 0.7 to 0.85, and little increase in 71
Figure 3.16 illustrates an example of the relationship occurs beyrnri these spacings, except for large groups of
between the ultimate load capacity of a group of specified relatively long 1 'les.
dimensions and the number of piles in the group, cal Figures 3.18 and 3.19 show comparisons between the
culated using Eq. () .24). This figure shows the transition measured efficie1cy-spacing relationships from the tests
between single-pile failure and block failure as the number ,..,f Whitaker (1957) and those calculated from Eq. (3.25).
of piles increases. In the design of such a group, it is The agreement is generally quite good and the method of

1.0 -------r--------,---------,
3 2 x 30 d(ST)
2 x 12 d(SF)
3 x 12 d(STI I

t o 6 t--------t----c:--t,---r-'


d = diameter

\Iv Whitaker (19571

ST Saffory-Tate (19611
SF= Sowers-Fausold (19611

0.2 '-----------------------
2 3 4
Si:aci ng/diameter
FIGURE 3.17 Relationships for freestanding groups of 2' to 9' 1pi!es oflengths 12d to 48d,frnm model tests (after de Mello, 1969).

1.0 ------- 1.0 .---------- LO.---------.

0 yoo.o
0.6 O 0 0.6 0.6
!] !]

0.2 52 group 0.2 7 2 group 0.2

9 7 group

2 3 4 2 3 4 2 3 4

-'-- s
d d

0 Experimental
(Whitaker, 1957}

_I,_= 48

FIGURE 3.18 Experimental and calculated group efficiency, effect of group size.

calculation appears to predict . with reasonable accuracy has measured the load carried by the piles in model free
the effects of group size, pile spacing, and pile length. standing groups in clay by introducing a small load gauge
It has often been assumed that all piles in a group are at the head of each pile. The results for a 3 2 group of piles
equally loaded. However, if the group supports a rigid at three different spacings are shown in Fig. 3 .20, in which
cap, the load distribution within the group is generally the average percentage of load taken by each pile is plotted
not uniform, the outer piles tending to be more heavily against the group load as a percentage of the group load at
loaded than the piles near the center. Whitaker ( 1957) failure. At spacings of 2d and 4d, the corner piles take the
greatest load (about 13 to 25% more than the average
load) while the center pile takes the least (about 18 to 35%
less than the average), At a spaci.-lg of 8d, virtually no
group action was observed and the load distribution was
uniform. The load distribution for a 5 2 group, at a spacing
of 2d, is shown in Fig. 3 .21. The corner piles reached their
maximum load at about 80% of the ultimate group load,
and carried a constant load thereafter. At failure, the corner
0.5 0.5
0 2 4 (i 8 0 2 4 6 8 piles carried about 28% more than the average load, while
1 s
(a) d (b) d the center and lightest-loaded pile carried about 44%
1 24
less. Therefore, there appears to be a tendency for the load
LO 1.0 distribution to become increasingly nonuniform as the
number of piles in the group increases. A theoretical
method for calculating the load distribution prior to
!] 17 0 ultimate failure is described in Chapter 6, and this method
also confinns the trends displayed by Whitaker's tests. PILED FOUNDATIONS

0.5 0.5
0 2 4 B il 0 2 4 6 8 The ultimate load capacity of a piled foundation (i.e., a pile
d d group having a cap cast on or beneath the surface of the
1 36
1 =48 soil) may be taken as the lesser of the following two values:
d d
o Measured (Wh;taker, i957} (a) The ultimate load capacity of a block containing the
- Calculated
piles (Eq: 3.23) plus the ultimate load capacity of that
FIGURE 3.19 Effect of pile lengh on group efficiency. portion of the cap outside the perimeter of the block.

(a) (:,) (c)
s/d 4
"' s/d = 2 , s/d 8

g' "'
C: 0
"' 0

I ,,
5) I
0. "'
ro -:S

C. ::
I /

0. "'
0 w

/ /
u.c 4
.J 0

40 80 100 0 40 80 100 0 40 80 100

Load on group as a percentage of the load
a\ failure

A B A --- Average of piles A

---- Average of piles B
B C B()
Centre pile
FIGURE 3.20 Load distribution in 3' pile group (Whitaker, 1970).

(b) The sum of the ultimate load capacity of cap and the
piles, acting individually, that is, for group of n pile's of
diameter d and length L. supported by a rectangular cap of
dimensions Be X L e ,

Pu = n(c;As + Ab Cb Ne)+ Ncc Cc (3.26)

(Bc Le - mrd2 /4)


average adhesion along pile

undrained cohesion at level Cl pile tip
undrained cohesion beneath pile cap
bearing capacity factor for pile (see Fig. 3 .5)
cc bearing capacity factor for rectangular cap Be X
Load on a group as a percentage
of the load at failure
Le (L e > B e )"" 5.14 (I+ 0.2 Bc fL c ) (Skempton

co Bl)
Average of piles A
The first value will apply for close pile-spacings while the

Bo Eo oo Bo
..__,, o second will apply at wider spacings when individual action
0 Average of piles can occur.
co Eo 0 \) co
Average of piles C Whitaker (1960) carried out tests on model piled foun
dations in clay and found that at close spacings, block fail
Average of piles D
B "'
0 0 o
E D Average of piles E
ure occurred, and that when the cap did not extend beyond
0 the perimeter of the group, it added nothing to the efficien
Bo o 0 0 C B A
a Pile F
cy of the group. At greater spacings, the efficiency-versus
spacing relationship was found to be an exter.sion of the
FIGURE 3.21 Load dis'tribu tion m 5' pile group at 2d spacing relationship for b.lock failure, with the efficiency exceeding
(Whitaker, 1970). unity because of the effect oft.he cap. Good agreement was

3.3.2 Pile Groups in Sand


There is less information on pile groups in sand than c,n
groups in clay, but it has been fairly well est1b1islted that
0.8 group efficiencies in sand may 0ften be greater than 1.
A summary of s,lme of the available data on piles
is given in Table 3 .4.
;;. 0.7
A summary of some tests on model piles, presented
by Lo (1967), is reproduced in Rg. 3.23. The data shown
w 0.6
in this figure are reasonably consistent with the data in
Table 3.4. Results of tests on somewhat larger model
piles, in groups of four and nine, carried out by Vesic
0.5 (1969), are shown in Fig. 3.24. Vesic measured t!1e point
loacl separately from the shaft resistance, and in tl1 light
of his measurements, he concluded that when the efficiency
0.4 of closely spaced piles was greater than unity, this increase
was in the shaft rather than the point resistance.
The broad conclusion w be drawn from the above
1 2 3 4
data is that unless the sand is very dense or the piles are
Spacing factor, s/d widely spaced, the overall efficiency is likely to be greater
rests on freestandi:1g groups
than 1 .. The maximum efficiency is reached at a spacing of
Tests on oiled foundat10,1s
2 to 3 diameters and generally ra1!ges between 1.3 and 2.
-- Calculated for piled foundations,
assuming block failure U.2.2 INFLUl:fNCE OF PILE CAPS
FIGURE 3.22 Efficiency of piled groups (Whitaker, 1970).
As can be seen in Fig. 3.24, the pile cap can contribute
significantly to the load capacity of the group, particularly
obtained between the model test results and the predicted in the case of the smaller four-pile groups. However, it
efficiency from the block failure equation (Fig. 3.22). seems likely that mobilization of tne bearing capacity of
The load-settlement behavior of piled foundations tlie full area of the cap requires considerably greater move
containing a relatively small number of piles to reduce ment than that required to mobilize the capacity of the
settlement is considered in detail in 2-hapter 10. piles themselves. This is the implication of tests by Vesic.
and for practical purposes, the contribution of the cap
3.3.J.3 ECCENTRIC LOADING can be taken to be the bearing capacity of a strip footing
of half.width equal to the distance from the edge of the
Model tests on groups with small eccentricities of load have cap to the outside of the.pile.
been carried out by Saffery and Tate (1961), who found
that for eccentricities up to two thirds of the spacing, ECCENTRIC LOADING
the group efficiency is not noticeably affected. Meyerhof
(1963) also reported that model tests on piled foundations The influence of eccentric loading on the load capacity
showed that the load eccentricity had no effect on load of pile groups in sand has been studied by Kishida and
ca,iacity for eccentricities up to half the group width. Meyerhof (1965) in a series of model tests. These tests
This behavior is explained by the fact that the reduced showed that small eccentricities of load have no signif
base resistan<::e is offset by mobilization of lateral resis icant influence on the bearing capacity of freestanding
tance. The group capacity can therefore be calculated as groups and piled groups because the applied moment is
for symmetrical vertical loading, except that for groups resisted mainly by the earth pressure moment on the
whose width is on the same order as the pile length, Meyer sides of the group. At larger eccentricities, the load capa
hof (1963) suggests that the shaft resistance can be ignored city decreases rapidly because of smaller point resistance
and the base resistance calculated in a fashion similar to of the group by a reduction of the effective base area.
eccentrically-loaded spread foundations, that is, using a In estimating load capacity, Kishida and dleyerhof
reduced effective base area. suggest that the moment caused by a load Vat eccentricity


Pile Pile Group

Length Diameter Spacing Efficiency
Reference Soil L d L/d ,Group d "I Remarks

Press (1933) Medium-grained 6-10 ft 5 & 6 in. 12-20 2-8 Various >l Driven piles. Max.
moist, dense 11 of l.5 ats/d "',2
sand 23 ft 16 in. 17 2 Various <L Bored piles

Cambefort (1953) Humus/stiff 100 in. 2 in. 50 2--7 2 1.39 Driven piles
clay/sand/ 3 1.64 Average values of 17.
gravel 5 1.17
9 1.07

Kezdi (1957) Moist fine 80 in. 4 in. 20 4 2 2.1 Driven piles.

sand (square) (In line) 3 1.8 Max. 11 ats/d"" 2.
4 1.5 11 greater for square
6 1.05 group.
4 2 2.1
(square) 3 2.0
4 1.75
6 1.1
Bold lines . Data from Bereclugo 3.0 ....------....------....------
Thin lines. Data from others
Solid l<nes ..... Dense sane!
240 Oa$h I Ines Loose sand /
Ra . Rough pile
Sp . Smooth pi!e
220 l. large group, piles I
.Smhll 9rni1r\ shnn pdf;S


''I I "'t'
<S 1



a; 180
Y. t.i'

": 2.0 -- I
# cf
':: 160

"'. r o ra ,
I Ii; 1
"" I . _ __..
' en c y 'vir
140 "'
2 --.....,cap
____,, ii: 1.S I . .,.el'
120 ''
' '-
& Meyerhof,_
40 Point.efficiencv

2 ,ti 4!>
' .0

3;; 3 ,...,,,,,_--=-,
(Average of all P-tests)
- -0.... . I?
80 --..!'s
-- -- -- --
........_ (),"f5

x- --"'*"- - - p, - - - - - - - --
32,s c1, -- 31 5

o.5 L_-1___.L__j___-::-- -';---:;

2 3 4 5 6 8 9 2 3 4 5 6 7
Spacing in pile diameters Prie spacing in diameters

FIGURE 3.23 Measured values of group efficiency in sands-model .ests (Lo, FIGURE 3.24 Pile group efficiencies (Vesic, 1969).
1967). (Reproduced by permission of t1' National Research Council of Canada
from the Canadian Geotechnical Jour.. al, Vol. 4, 1967, pp. 353-354.)


(,;, - 4 3" l
1.25 1.25 Er,1

2 :il
1 p1IE' 0
.'! co 1.00 1.00 "
4 pdes D

u u 9 p,les 6 - ...

C ::, 0.75
::, o_

o_ ::,
0 00 (pile cap
o; J' X J")
0.50 0.50
ro u
u o;
o; C
C c: 0.25 0.25

2 3 4 0 2 3 4

FIGURE 3.25 Bearing capacity of model pile groups under eccentric load in sand: (a) freestanding pile grups; {b) piled foundations
(Kishitla and Mcyerhof, 1965 ). (Copyright Canada, 1965 by University of Toronto Press.)

e is balanced by the moment caused by lateral forces on ing between 20% and 50% more than the average. Thes,:
the sides cf the group until it reaches the maximum value results are in contrast to the load distribution in groups in
corresponding to the coefficient of passive earth pressure. clay, where the center pile carries the least load and the
Within this limit, the eccentricity of load is assumed to corner piles the most.
have no effect on the point resistance. When the moment The influence of the order of driving piles in a group
Ve is greater than can be resisted by side pressure on the on the load distribution has been studied by Beredugo
outer piles., the extra is considered to be taken by an (1966) and Kishida (1967). They found that when the
eccentric base resistance for the case of block failure; load on the group was relatively small, piles that had been
or, for individual pile failure, by the development of uplift installed earlier carried less load than those that have been
resistance of some piles. The total bearing capacity then installed later; but when the failure load of the group was
decreases with further increase in eccentricity. approached, the influence of driving order diminished,
Comparisons between the theoretical and measured and the position of the pile in the group became the domi
effect of load eccentricity on load capacity are shown in nant factor. At this stage, the piles near the center took
Fig. 3.25 for the tests carried out by Kishida and Meyerhof the most load and the corner piles. the least, as in Vesic's
(1965), and there is fair agreement for tests in both loose experiments.
and dense sands. Beredugo also investigated the effects of repeated
loading and found that there was a progressive reduction LOAD DISTRIBUTION IN GROUP
of the influence of driving order, and that for the third
The most detailed data available on load distribution and subsequent loadings, the pile position was the domi
within groups in sand is that reported by Vesic (1969), nant factor at all loads up to the ultimate of the group.
wh0- made axial load measurements in individual piles
during group placement, as well as during loading tests.
For the four-pile groups tested, the measured load distri
bution was JUnost uniform, as expected; the maximum 3.4 PILES TO ROCK
deviation from the average was 3 to 7%. For the nine
pile groups, significant nonuniformit.y of load was mea
sured. The center pile carried about 36% more load than 3.4.1 Point-Bearing Capacity
the average, while the corner piles carried about 12%
less and the edge piles 3% more. Other tests on similar There are a number of possibl approaches to the estima
groups shmved a similar trend, with the center piles carry- tion of point-bearing capacity of piles to rock, including:

(a) The use of bearing-capacity theories to calculate the d..-aws attention to the fact that the load-penetration curve
ultimate point-bearing capacity Pbu. for rocks of medium strength or less ("' 100 MPa) has a
(b) The use of empirical data to determine the allowable large "plastic" component, despite the brittle nature of
point pressure Pba the rock. The curve divides into two portions, with what
(c) The use of in-situ tests to estimate either ultimate appears to be a change of slope associated with the forma
point capacity Pbi, or allowable point pressure Pba. tion of a crushed zone beneath the footing. The displace
ments required to mobilize the full bearing capacity of
Bearing-C:apacity Theories such rocks are very large, and it seems that ; factor of
Pelis (1977) has classified theoretical approaches into three safety of 3 to 4 is required to limit. the displacements to less
categories: than 2% of the footing diameter. Very brittle rocks (qum
> l 50 MPa), do not exhibit this "plastic" load-penetration
1. Methods that essentially assume rock failure to be curve.
"'plastic" and either use soil mechanics-type bearing The presence of jointing in [he rock will tenn to reduce
capacity analyses o; modifications thereof to account for the ultimate bearing capacity. The presence of closely
the curved nature of the peak failure envelope of rock. spaced continuous tight joints may not reduce the bearing
2 . .Methods that idealize the zones of failure beneath a capacity much below that for the intact rock material. If
footing in a form that allows either the brittleness-strength the are open vertical joints with a spacing less than tne
ratio or the brittleness-modular ratio to be taken into width or diameter of the pile point, the point is essentially
account. supported by unconfined rock columns and the bearing
3. Methods based on limiting the maximum stress beneath capacity may be expected to be slightly less than the aver
the loaded area to a value less than required to initiate age uniaxial strength of the rock. If the joint spacing is
fracture. These methods assume essentially tltl.t once the much wider than the footing width, Meyerhof (1953) sug
max.imum strength is exceeded at any point in a brittle gests that the crushed zone beneath the footing splits the
materiaL total collapse occurs. block of rock formed by the joints. Sowers and Sowers
(1970) present a theory for this case that generally indi
For a typical sandstone having an effective frktion cates a bearing capacity slightly greater than the uniaxial
angle ' in excess of 45 , effective cohesion c' of about . strength. Thus, in summary, theoretical considerations sug
one-tenth of the uniaxial strength, qum, and a ratio of gest that the ultimate bearing capacity is unlikely to be ri;;
Young's modulus-to-uniaxial strength of about 200, Pelis duced much below the uniaxial strengtn of the intact rock,
shows that the varios theories predict an ultinrnte point even if open vertical joints are present.
bearing capacity ranging between 4.9qum (incipient failure
theory based on the modified Griffith theory) to 56qum Use ofEmpirical Data
(classical plasticity theory). Various model tests on intact Allowable bearing pressures on rock have often been
rock carried out by Pelis and others indicate ultimate specified by various building codes and authorities, either
capacities ranging between 4 and 11 times qum. Pelis , based on a description of the rock, or in}rms of the
_ :-..
;::, -l. t',. ,, :;t\
_./ j:_:J4 \.. . '\j 1.
,/, . ., {

TABLE 3.5 TYPICAL PROPERTIES OF ROCK (PECK, 1969) , .... , O. _5 v' <:", ,j
Rock Compressive Shear E(IUps1J Poisson's Ratio
ype Strength q 1,1.m Strength
(psi) (psi) Field Lab. Field Lab.

Basalt 28,000-67,000 0.8-3.5 3.6- 5.9 0.30-0.32 0.26-0.28

Granite 10,000-38,700 200()-4260 5.6-11.6 5.4-11.8 0.25-0.27 0.17-0.29
Quartzite 16,00()-44,800 3.1- 8.5 3.6-12.5 0.25-0.30 0.07-0.l 7
Limestone 2450--28,400 1200-2980 3.3-11.9 0.24-0.27
Marble 7900-27,000 1280-6530
San.dslone 4900-20,000 284-2990 1.3- 5 .6 1.0- 9.0 0.28-0,30 0.07-0.17
Slatf 6950-31,000 199()-3550 1.0- 2.5 5.3- 8.4 0.30-0.32 0.24-0.25
Shale 0.20-0.25
___ __ ___
500-6500 0.26-0.27
Concrete 2000-5000 400-1000 2.5- 4.0 2.5- 4.0 0.15 0.15
,.. ,,_, "


---------=-. --.-----,-----,-----,
Y Failed






0 10 20 30 40 50 60
Unconfined strength (MPa)
FIGURE 3.26 Achieved end-bearing pressures in field tests on piles to rock (Thorne, I 977).

uniaxial compressive strength Qum . Some typical values been used but may be expensive if the rock is strong and
of qum and other rock properties are summarized in large loads are required. Freeman et al. ( 1912) have
Table 3 .5 Typically, allowable pressures, Pba , ranging described the use of the Menard Pressuremeter to estimate
between 0.2 and 0.5 times Qum have been stipulated. An the allowable point-bearing capacity, Pba, of piles in rock,
example of stiulated bearing pressures related to rock and suggest that Pba may be taken as the value where the
types is provide by Ordinance No. 70 in New South Wales, pressure-versus-volume relationship starts to become
Australia, in which values of Pba range between 430 kN/m 2 nonlinear. Satisfactory designs of caissons in sound shale
for soft shale to 3210 kN/m 2 for hard sandstone free from bedrock using the above approach have been reported by
defects to a depth of 900 mm. Freeman et al., and design pressures considerably larger
Thorne (1977) has collected data on recorded values than those specified by empirical relationships or building
of bearing capacity, as shown in Fig. 3.26. These values codes have been used.
vary from 0.3qum to about 4q u , ana most cases do not
involve failure. The few recorded failures are in swelling
shales and in fractured rocks, it is clear from these results
that the fracture spacing has an effect on the bearing capa 3.4.2 Pile-Rock Adhesion
city, . although the data is insufficient to quantify this
effect. When piles are socketed or driven into rock, some load
On the basis of the available data, an allowable point transfer to the embedded portion of the shaft will usually
bearing pressure on the order of 0.3qum would appear to occur. Theoretical solutions for load transfer are discussed
be quite conservative for all but swelling shales . Reference in Section 5 .3, and also by Ladanyi (1977). The distribu
to the theoretical solutions shows that such values generally tion of applied load between side-adhesion and end-bearing
imply a factor of safety of at.least 3 in fractured or closely at working loads, as given by theory, has been supported
jointed rocks and 12 or more for intact rocks. by in-situ measurements at a number of sites (Pells, 1977).
There is not a great amount of data on ultimate values of
The Use of In-Situ Tests pile-rock adhesion, but Thorne (1977) has summarized
A number of methods of in-situ testing of rock have been some of the available data, and this summary is reproduced
dmeloped in recent years. Plate-load tests have f requently in 3.27. These results show that a number of failures

5 r----,-----,,-----.------.---------

Canada shale
1-------1-N-ew-ca-stl-e -+---.------+----./----= =-----1
] sandstone


Kings Park
_,.,c;......____ -....1..---=s....
Normal maximum
for 25 MPaconc ret
andesit e--::,,j,-M-el-bo_ u_rn_e Ei+--Ca- li- f-or-n-
i a--l
/ mudstone I shale and
(value appro ximate) sandstone
shale UK siltstone/rn udstone UK shale

0 10 20 30 40 50 60
Unconfined strength (MPa)
FIGURE 3.27 Adhesion attained in field tests on piles in rock (Thorne, 1977).

have occurred, even in relatively unjointed rocks, at values 3.5 USE OF IN-SITU TESTS
on the order of O. l qum . It should. be noted that in many
instances, concrete strength will be the limiting factor,
and in the few instances in which information is available 3 .5 .I Static Cone Penetrometer
on concrete strengths, failure has occurred at an average
shear.stress of between 0.05 to 0.2 times the ultimate com The basis of the test is the measurement of the resistance
pressure strength of concrete, f- However, the tests of to penetration of a 60 cone with a base area of 10 sq
Jaspar and Shtenko (1969) indicated that considerable cm. Two types of cone are commonly used; the standard
caution must be exercised with piles in expansive shales point, with which only point resistance can be measured;
that are likely to be affected by water; an adhesion of and the friction-jacket point, which allows both point
only about 11 psi (75 k.Pa) was measured in these tests. resistance and local skin resistance to be measured (Bege
Freeman et al. (I 972) suggest a design value of allowable mann, 1953 and 1965).
pile-rock adhesion of 100 to 150 psi (700 to 1000 k.Pa), In purely cohesive soils, it is generally accepted that the
depending on the quality of the rock. With such a value, cone-point resistance, Ckd , is related to the undrained cohe
they recommended that the full calculated end-bearing sion, cu , as
capacity be added to obtain the total -design-load capacity.
On the basis of the limited information available, (3.27)
it would appear reasonable to use as a design value an allow
able adhesion of O.OSf or 0.05Qu m, whichever is the As discussed in the previous section, the factor Ne may
lesser value. These values should not be applied to highly vary widely both theoretically and in practice, and values
fractured rocks, for which values of adhesion between 75 of Ne ranging from 10 to 30 have been suggested. The
and 150 kPa may be more appropriate. It must be empha major causes of this variation are the sensitivity of the soil,
sized that care should be exercised to remove all remolded the relative compressibility of the soil, and the occurrence
soil from the socket zone. Furthermore, for uplift loads, of adhesion on the side of the cone. The variation in the
a reduction of the above loads (e.g., by about 30%) appears rate of strain between the cone test and other testing
to be desirable. methods also has an effect on the deduced value of Ne ,

but the use of a constant-penetration rate minimizes cone resistance Ckd within a distance 3.75 db above and
variations from this cause. For design purposes, a value of db below the pile tip, where db is the diameter of the
Ne 15 to 18 appears reasonable (Begemann, 1965; pile tip.
Thomas, 1965; Blight, 1967; Thorne and Burman, 1968). Full-scale tests carried out by Vesic (1967) showed
Van der Veen ( 1957) suggested that the ultimate resis that the point resistance of the piles tested is comparable
tance of a pile point, of diameter db , could be derived from with that of the penetrometer, but the shaft resistance of
the corresponding cone-penetration curve by taking the the piles was approximately double that measured by the
average cone resistance over a distance bdb below the penetrometer. Thus, the ultimate load capacity is given by
pile point and adb above the point. Average values of
a = 3 .75 and b l were suggested by Van der Veen.
The adhesion measured by the friction jacket may
safely be taken as the skin friction for driven piles in clays
(Begemann, 1965). Alternatively, but less desirably, the
cohesion may be estimated from the point resistance and Ckd = measured cone-point resistance at base
an appropriate reduction made to obtain the pile-soil
j average shaft friction along pile, as measured
adhesion (see Section 3.2.1).
on the friction jacket
For piles in sand, various attempts have been made to
relate the cone-poh1t resistance to the angle of friction
and relative density of the sand (Meyerhof, 1956; Shultze
and Mezler, 1965; Plantema, 1957), but it has been found
For driven steel Ii-piles, Meyerhof (1956) suggested that
the above shaft resistance should be halved.
that cone resistance is very sensitive to changes in density. A comparison between the pile and penctrometer
For pactical use, the previously mentioned suggestion of resistances for the tests reported by Ves;c (I 967) is shown
Van der Veen (1957) may be adopted; namely, that the iI_1 Fig. 3.28. The upper and lower limits of the penetro
ultimate point resistance of the pile be taken as the average meter values are shown. Correlation with static cone tests

Pile skin resistance and

Cone and pile point resistance doubled cone shaft resistance
c,d (ton/ft f
fb f,
2fc (ton/ft )


i:l 20

.D 30


- Measured values Shaded area-values calc. from static cone results

'lJRE 3.28 Variation of point and skin resistances with depth (Vesic, 1967).

was found by Vesic to be better than with the results of 3.5.3 Pressuremeter Test
standard penetration tests (see below).
for cases in which separate measurements of friction The use of the pressuremeter in foundation design has been
jacket resistances are not made, Meyerhof (1956) suggested developed extensively in France in recent years. Its appli
that for driven concrete or timber piles, the ultimate skin cation to the estimation of pile load capacity has been
friction fs could be estimated from the cone point resis summarized by Baguelin et al (1978) who present curves
tance Ck d as follows: relating ultimate base capacity to the pressuremeter limit
pressure, for both driven and cast-in-situ piles. Relation
fs 0.005Ck d (3.29) ships are also presented between ultimate skin resistance
and limit pressure for steel or concrete piles in granular and
For driven steel H-piles, Meyerhof suggested that the cohesive soils, and for cast-in-situ piles in weathred rock.
above value be halved. Some comparisons (Mohan et al., The following upper limits on the ultimate skin resistance
1963; Thorne and Burman, 1968) indicate that Eq. (3.29) are suggested by Baguelin et al for pressuremeter limit
underestimates the skin friction by a factor of about pressures in excess of 15 00 k.Pa;
2 if Ckd is less than about 35 kgf/cm 2
In sands, it is necessary to make a distinction between concrete displacement piles in
the skin friction for downward and upward loading. Modi granular soil 122 kPa
fications fo r uplift resistance are discussed in Section 3 .7.
concrete displacement piles in
cohesive soil, or steel displace
ment piles in granular soil 82 kPa
3.5.2 Standard Penetration Test
steel displacement piles in
Meyerhof (1956) has correlated the shaft and base resis cohesive soil 62 kPa
tances of a pile with the results of a standard penetration non-displacement piles in
test. For displacement piles in saturated sand, the ultimate any soil 40 k.Pa
load, in U.S. tons, is given by


3.6.1 Large Bored Piers
f!p standard penetration mimber, N. at pile base
N average value of N along pile shaft Large-diameter bored piles have come into increasing use
in recent years as an alternative to pile groups. They have
For small displacement piles (e.g., steel H-piles), been constructed up to IO ft in diameter and in lengths
exceeding lOO ft, often with an underreamed or belled
base. Such piles have found extensive use in London clay,
and much of the research on large bored piers is based on
their behavior in London clay. Empirical methods of
where design have been developed on the basis of extensive expe
rienc and research. One of the earliest investigations was
Ab net sectional area of toe (sq ft) in model tests on piles with enlarged bases, reported by
A s = gross surface area of shaft (sq ft) (area of all sur- Cooke and Whitaker (1961). These tests evealed that,
faces of flanges and web for H-piles) whereas settlements on the order of 10 to 15% of the base
diameter were required to develop the ultimate base capa
In Eq. (3.30), the recommended upper limit of the unit city, the full shaft resistance was developed at very small
shaft resistance (F//50) is 1 ton/ft 2 and in Eq. (3.31), settlements, on the order of 0.5 to 1.0% of the shaft
0.5 ton/ft2 diameter. (The theory given in Chapter 5 supports these
The above equations have also been used with some findings.) A considerable amount of field-test evidence has
success in stiff clays (Bromham and Styles,1971 ). subsequently been obtained (Whitaker and Cooke, 1966;

Burland et al., 1966), and the behavior of full-scale large where

bored piers has been found to be similar to that of the
model piles. Cb cohesion at pile base
Because of the. different degrees of shaft-and base co cohesion at level of base of bulb
load mobilization at a given pier settlement, it may be Neb = value of Ne at pile-base level
advisable to detennine the working load on a large pier Nco value of Ne at level of base of bulb
by applying separate factors to the ultimate shaft and base ca = average pile-soil adhesion
resistances; for example, Skempton (1966) suggested. a As = surface area of pile shaft
safety factor of 1.5 for shaft resistance and 3.0 for base do = bulb diameter
resistance, for piers with an enlarged base of diameter
6 ft or less .. In many cases, the working load for bored Values of Ca, Neb , and Nco can be obtained from Section
piers, especially those with enlarged bases, will be deter 3.2.
mined by settlement considerations rather than ultimate For double or multiple underreamed piles with the
capacity (Whitaker an4 Cooke, 1966, Burland et al., 1966). bulbs suitably spaced, the soil between the bulbs tends to
Settlement theory is discussed in Chapter 5. act as part of the pile, so that the full resistance of the soil
can be developed on the surface A-A' of a cylinder with a
diameter equal to that of the bulbs and height equal to
3.6.2 Underreamed Bored Piles their spacing. Model tests carried out by Mohan et -al.
(1967) have confinned this behiivior. Mohan et al. (1969)
Underreamed piles have been extensively_ used in India, have suggested two methods_ for estimating the load capa
both . ijS load-bearing and anchor piles in expansive clays. city of multiple underreamed piles:
For anchor piles, a single enlarged bulb is often used,
while for load-bearing, one or more bulbs may be used. 1. Summation of the frictional resistance along the shaft
A single underreamed pile can be treated in a similar man above and below the bulbs, shearing resistance of the
ner to a pile with an enlarged base, except that the bulb cylinder circumscribing the bulbs, and the bearing capacity
may be situated above the base of the pile. Mohan et al. of the bottom bulb and base.
(1967) suggest that the base and shaft resistance be added 2. Summation of the frictional resistance along the shaft
to give the ultate load capacity. Thus, referring to Fig. above the top bulb and below the bottom bulb, and the
3.29, for a pile in clay, bearing capacity of all the bulbs and the base.

It was found that for a typical example of a pile in London

(3.32) clay, these methods give almost identical results. For other
cases, the lesser of the two capacities given by the equations
should be taken.
Mohan et al. (1967) suggest that the optimum spacing
of the bulbs in a multiple underreamed pile lies between
1.25 and 1.5 times the bulb diameter for maximum effi
ciency. As an example of the economy in material that may
be obtained by using underrearned piles, they calculated
d d
that a multiple underreamed pile in London clay can
develop the same load capacity as a uniform pile of about
L four times the volume.

3-.6.3 Screw Piles

Screw piles have been used in several countries for mast

and tower foundations and for underpinning work. Load
tests on model and full-scale screw piles have been reported
(a) Si11gle muferreamed (:) D011i,/e muferreamed by Wilson (I 950) and by Trofimenkov and Mariupolskii
pile pile
(1965). Wilson ( 1950) developed a method of analysis
FIGURE 3.29 Underreamed piles. of the load capacity of screw piles in both sand and clay,

The remolded strength of the soil is used because the clay

adjacent to the shaft is likely to be ahuost fully remolded
by the passage of the screw and by the lateral displacement
caused by the cylinder.
A comparison made by Skempton between measured
and predicted load capacities by the above method showed
that the predicted ultimate loads were within I 5% of the
measured values, although always greater. Trofimenkov
and Mariupolskii ( 1965) employed the same basis of cal
culation as the above and also obtained good agreement
between measured and calculated load capacity.

FIGURE 3.30 Idealized screw pile.


based on the use of elastic theory. In a relatively simple

3. 7.1 Single Piles
analysis for screw piles in clay proposed by Skempton
( 1950), the load capacity is taken to be the sum of the
Piles may be required to resist uplift forces-for example,
bearing capacity of the screw and the side resistance along
in foundations of structures subjected to large overturning
the shaft, assuming no skin friction to be 111obilized for a
moments such as tall chimneys, transmission towers, or
distance above the screw equaJ to its diameter. Thus,
jetty strnctures. Methods of calculating the adhesion to
referring to Fig. 3 .30.
resist uplift are the same as those used for bearing piles.
For a uniform pile in clay, the ultimate uplift resis
(3.33) tance, Puu, is
c, = average remolded shear strength along the shaft
in the length (L -do) where
cb average of undisturbed and remolded shear
strength of soil beneath the screw Wp weight of pile
Ab area of screw ca average adhesion along pile shaft

1.25 ------------------------
Source of data i
Tomlinson ( 1957) Average values
} for pile load tests
Skempton ( 1959)
1.00 f----'><+----+---+-----+-- Mohan and Chandra (1961 ) -j
Turner {19621 Data for p,le
I f:
o F erson and Urie (1964) pulling teSts

0 500 1000 1500 2000 2500 3000 3500 4000

c0 = undrained shear strength (psi)
FIGURE 3.31 ! ,,ationship between ca /cu and undrained shear strength for pulling tests (Sowa, 1970). (Reproduced by permission of the
National Research Council of Canada from the Canadian Geotechnical Journal, Vol. 7, 1970, pp. 482-493.)

Relatively fw pulling tests on piles have been reported (a) The shear resistance of a vertical cylinder above the
in th literature. A summary of some of the available results base, multiplied by a factor k, plus the weight of soil and
is given by Sowa (I 970), who has f ound that the values of pile, W, above the base.
ca lcu agree reasc.nably well with the values for piles sub- (b) The uplift capacity of the base plus W, that is,
jeted to downward loading 3.3l).
For piles of uniform diameter in sand, the ultimate
uplift capacity may be calculated as the sum of the shaft ( 3.34)
resistance plus the weight of the pile. There is, however,
little data available on the skin friction for upward loading, where
and the available data is to some extent conflicting. For
example, tests rported by Ireland ( 1957) on piles driven Nu uplift coefficient
into fine sand suggest that the average skin friction for = Ne for downward load
uplift loading i5 equal to that for downward loading, but
data summarized by Sowa ( 1970) and Downs and Chieurzzi Examination of the results of model and field tests led
(1966) indicates considerable variations in average skin Meyerhof and Adams to suggest the following values of
friction between different tests, although there is a ten k:
dency for the values to be lower than for downward load
ing, especially for cast-in-situ piles. In the absence of other Soft clays k 1-1.25
information, a reduction to two thirds of the calculated Medium clays k 0.7
shaft resistance for downward loading is recommended. S.:iff clays k 0.5
However, a reliable estimate is best deterrr.ined by carrying Stiff fissured clays k 0.25
out a pulling test in-situ.
If static-cor,e-penetration tests are used as a basis for The low values of k in the stiffer clays are partly attributed
estinrnting ultimate uplift skin resistance, Begemann (1965) by Meyerhof and Adams to the influence of tension cracks
suggests that the calculated skin resistan(;e for downward arising from premature tension-failure in the clay.
loading be adjusted by a reduction factor dependent on It has been found that negative pore pressures may
the soil and pile type. He also suggests reduced values of occur in clays during uplift, particularly with shallow
skin resistance be used if the uplift load is oscillating. embedment depihs. The uplift capacity under sustained
Begemann's suggestions, however, should be viewed with loading may therefore be less than the short-term or un
cor,siderable caution, as they are based on limited data. drained capacity, because the clay tends to soften with
Additional uplift resistance may be obtained by under time as the negative pore pressures dissipate. The long-term
reaming or enlarging the base of the pile, and in such uplift capacity can be estimated from the theory f or a
cases, the pile shaft may have little or no influence on the material with both friction and cohesion, using the drained
uplift capacity. Traditional methods of design assume the parameters <Pa and ca of the clay.
resistance of the enlarged base to be the weight of a cone For a soil with both cohesion and friction, the follow

of earth having sides that rise either vert.cally or at 30 ing expressions were obtained by Meyerhof and Adams
f rom the vertical. Neither of these methods has proved for the ultimate load capacity, P uu, of a circular base:

reliable in practice, however. The 30 -cone method is
usually conserv1tive at shallow depths but can give a con (a) Shallow depths(L<db):
siderable overestimate of uplift capacity at large depths
(Turner, 1962). Parr and Varner (1962) showed that the
vertical-failure-surface approach did not apply to piles in (3.35)
clay, although it could apply to backfilled f ootings. Alter
native theories for uplift resistance of enlarged bases have (b) Great depths (L > H):
been proposed by Balla (1961), MacDonald (1963), and
Spence ( 1965)--these theories differing in the assump
tions regarding the shape of the failure surface. (3.36)
Meyerhof and Adams ( 1968) have developed an approx
imate approach based on observations made in laboratory where
model tests. They suggest that the short-term uplift capa
city of a pile in clay ( under undrained conditions) is given 1 soil unit weight*
by the lesser of s = shape factor

= I + mL/db , with a maximum value of values may be appropriate to upward loading. However,
1 t-mH/db the theory for failure of anchor piles with enlarged bases,
K!I ::: earth-pressure coefficient ( approximately 0.9. or of anchor plates more generally, has yet to be fully

0.95 for ,P values between 25 and 40 ) developed.
m coefficient depending on ,p For use in Eqs. (3.35) and (3.36), values of H/db ,
H = limiting height of failure surface above base s, and m, obtained from tests results by Meyerhof and
W weight of soil and pile in cylinder above base* Adams, are shown in Table 3.6. The ultimate uplift capa-.
city should be taken as the lesser value of that given by
The upper lin1it of the uplift capacity is the sum of the net Eq. (3.37) and the appropriate equations 3.35 or 3.36.
bearing-capacity of the base, the side adhesion of the shaft, The results of model tests in clays, reported by Meyer
and the weight of the pile, that is, hof and Adams ( 1968), are shown in Fig. 3.32. !30th the
undrained and long-term pullout loads are shown; and the

20 25 30 35 40 45 48

Ne, Nq bearing-capacity factors

2.5 3 4 5 7 9 11
fs ulti.mate shaft-shear resistance.
Uvb effective vertical stress at level of pile base
m 0.05 0.1 0.15 0.25 0.35 0.5 0.6
Meyerhof and Adams suggest that the values of Ne
1.12 1.30 1.60 2.25 3.45 5.50 7.60
and Nq for downward load can be used in this context,
but theoretically this is incorrect, and somewhat lower
a From Meyerhof and Adams (1968). (Reproduced by permission
of the National Research Council of Canada from the Canadian
Bi,oyant or total, as appropriate. Geo technical journal, Vol. 5, 1968, pp. 225-244.)

Measured undrained capacity


1000 Measured drained capacity


Estim2ted drained capacity

800 160

1;l .I
..Q 600 120 60 30

400 80 40


200 40 20
,-; II
0 L....l--1._.c=--...u.J..LJ......J 0 0
Brick clay Brick clay Brick clay Niagara clay
5"" 5" anchor 1. 125" dia. anchor 1. 125'' dia. anchor 1. 125" dia. anchor
depth 12 in. depth 7.5 in.. depth 2 in. depth 5 in,
D/B 2. 4 D/B 6. 7 D/B = 1.78 D/B = 4 . 4
stiff stiff stiff soft
FIGURE 3.32 Comparisor:. of shrJrt-term and long-term pull-out tests in clay (Meyerhof and Adams, 1968). (Reproduced by permission of the
National Research Council of Canada from the Canadian Geo technical Journal, Vol. 5, I 968, PP: 225-244.)

considerable reduction in load capacity with time can (b) The uplift load of an equivalent pier foundation
clearly be seen. The extent of the load-capacity decrease consisting of the footings and enclosed soil mass.
becomes greater as the soil becomes stiffer. The predicted
long-tenn capacities of the piles show reasonable agree Meyerhof and Adams (1968) have presented some
ment with the measured values. data on the uplift efficiency of groups of two and four
The above theory can also be used to estimate the model circular footings in clay. The results indicate that
uplift capacity of piles in sand. Meyerhof and Adams have the uplift efficiency increases with the spacing of the foot
compared predicted and measured uplift capacities for ings or bases and as the depth of embedment decreases,
buried footings in sand and have found fair agreemet, but decreases as the number of footings or bases in the
although there is a relatively wide scatter of points. group increases. The uplift efficiencies are found to be
in good agreement with those found by Whitaker (1957)
for freestanding groups with downward loads.
For uplift loading on pile groups in sand, there appears
3.7.2 Pile Groups to be little data from full-scale field tests. However, Meyer
hof and Adams (1968) have carried out tests on small
Meyerhof and Adams (1968) suggest that the ultimate groups of circular footings and rough circular shafts, and
uplift load of a group be calculated as the lesser of have analyzed the group efficiencies. For a given sand
density, the uplift efficiencies of the groups increase
(a) The sum of the uplift of the individual footings. roughly linearly with the spacing of the footings or shafts,


Pile Alignment
Reference Pile Type Length Soil Type at Tip Type of Bend

Parsons and Composite: lower 140 ft 20 ft fill, layers of 4.4 ft Gentle sweep over
Wilson 85 ft, I O-in. pipe, organic silt, medium lower length
(1954) top 55 ft, sand, fine sand, silt
corrugated pipe with clay layers,
gravel, bedrock

Bjerrum Steel H-section 30 ft Clay 1.2 ft Gentle sweep

(I 957)

Johnson Composite: lower 40 ft 20 ft silt overlying 8 ft Gentle sweep over

(1962) 40 ft, 1O in. medium sand lower length
upper 50 ft,
corrugated taper pipe

Mohr (1963) 10-in. pipe 85 ft 80 ft soft silt, 10.25 ft Gentle sweep

stiffsand clay,
medium den.se sand

National Precast hexagonal, 60m 50m soft clay, 1 Om l lm Gentle

Swedish Hercules jointed clay, silt, sand,
Council rock at 70m

Hanna (l 96 7) Steel H-section 140 ft 34 ft stiff clay, 50 ft 3.0 ft Triple curvature.

14 BP73 soft clay, 64 ft stiff relatively sharp
clay, shale direction changes
Stee! H-section 138 ft 6.0 ft Double curvature,
14 BP 89 relatively sharp
direction changes

and increase as the depth of embedment becomes smaller. largely be caused by the neglect of the structural strength
The uplift efficiency decreases as the numb of footings or of the pile shell in the design. Long, precast, hexagonal
shafts in the group increases and as the sand density in test piles have also been found to perform satisfactorily,
creases. but Hanna (I 967) has found that for steel H-piles, large
stresses are induced because of bending during driving. Pile
bending is attributed by Hanna to the dvelopment of asym
3.8 LOAD CAPACI.TY OF BENT PILES metrical stresses in the pile as a result of the eccentric pile
tip reaction and eccentric driving inherent in all pile-driving
A number of cases have been reported in which long, work. These eccentric stresses are considered to be suffi
slender piles have become bent during driving. A summary cient to initiate bending, which causes the pile to drive off
of these measurements is shown in Table 3.7. For con vertical. Reverse curvature of the pile may subseq_uently oc
crete-filled steel shell piles, load tests indicated that the cur, and this is believed to result primarily fro the verti
piles could tolerate significant out-of-verticality and still cal-weight component of the inclined pile forcing the pile
carry their design load with safety. This, however, may to bend.

Deflection in inches

- -
80 40 0 40 40 0 40

t --
East North South

a. N-S line throug
0 casmg grooves
inclined 7 degrees
west of N-S pile



- 140
N-S line through
Slope outside range casing grooves
of inscrurnent inclined 15 degrees
west of N-S pile

--- -- Deflection about axis of pile

Deflections rneasured by inclinometer

flGURE 3.33 Measured deflection components of driven pile (Hanna, 1968). (Reproduced by permission of the National Research Council of
Canada from the Canadian Geotechnical Journal, Vol. 5, 1968, pp. 150-172.)

0 2 3 4 5
i 1 I I I I
Scale in ft

FIGURE 3.34 Driven position of pile tips (Hanna. 1968). (Reproduced by permission of the National Research Council of Canada frOJ the
Canadian Geo\eclrnical Journal, Vol. 5, 1968, pp. 150-172.)

Typical deflection profiles, reported by Hanna (1967), Pmax = maximum allowable soil pressure
are shmvn in Fig. 3 .33. These profiles have been obtained Per buckling load of pile
from measurements on an inclinometer installed within the k modulus of subgrade reaction
H-piles. The as-driven positions of the pile tips for every Pmax maximum lateral deflection (deviation of the
20 ft of depth are shown in Fig. 3.34. For the two piles center line of the pile from a straight line con
conside,red, minimum computed radii of curvature were on necting the pile tip and the point at which
the order of 170 ft and I 90 ft at depths of 100 ft and 70 curvature of the pile begins)
ft: these values are about six times less than the suggested
safe minimum value for steel H-piles of 1200 ft (Bjerrum,
For the second criterion to- be satisfied, the allowable load
Methods of estimating the stresses in a pile due to non Pis
verticality have been proposed by Johnson ( l 962), Broms
(1963), Parson and Wilson (1954), and Madhav and Rao (3.39).
(1975). Typical of these methods is that of Brorns, who by
expressing the deflected shape of the pile as a Fourier sine where
series and assuming the soil to be a Winkler medium, was
able to derive a simple approximate equation for the buck b = Pe r + Aamax
ling load on the pile (the subject of buckling is discussed
more fully in Chapter 14). Provided that some information
C (3.39)
of the departure from straightness of the actual piles is
available, the maximum soil pressure along the pile and the
maximum bending moment can then be calculated. As de Per = pile-buckling load
sign criteria, Broms suggested that A area of pile
Orn ax allowable maximum stress in pile
(a) The calculated maximum soil pressure along the pile EP Young's modulus of pile
should not exceed one third of the ultimate value. Ip moment of inertia of pile
(b) The maximum stress (axial plus bending) in the pile Z pile section-modulus
.should be less than the allowable value. Rmin minimum radius of curvature along pile

The first criterion leads to an allowable load P given by

From Eq. (3.39), it may be deduced that the load
carrying capacity will be reduced to zero if c,;;;;; 0, that is, if
p (3.38)
kPmax + Pmax


Section 8P14;89)

Pile lengtn = 50 ft

co. 400

Upper limit for steel stress

<i 200


10--6 10-5 10 3
Maximum curvature (rads/in_)

0L_____ __c_ _ _____ L________J

0 20 40 60
Minimum radius of curvature (ft) lv1axinrnm {!c1lcc:1on (in,}
(a) On basis of steel stre:;s Un On basis of soil pressure

FIGURE 3.35 Allowable loads for bent piles (from Broms' analysis).

For a typical steel H-pile section_ in clay, the allowable pendent of pile length. For the limiting steel-stress criter
loads from Eqs. (3.38) and (3.39) are plotted in Fig. 3.35. ion, an allowable steel stress of 18 kips/sq ft has been
For the limiting-soil-pressure criterion, the allowable load adopted. The allowable load is insensitive to change in soil
increases as the stiffness of the soil increases (K =kd = 33 subgrade-reaction modulus or pile ler:gth.
times the cohesion, has been assumed) buLis almost inde-