Ž

.

Construction and Building Materials 14 2000 73

]78

Behaviour of precast reinforced concrete pile caps

Toong Khuan Chan

a,

U

, Chee Keong Poh

b

a

School of Ci

¨

il and Structural Engineering, Nanyang Technological Uni

¨

ersity, Nanyang A

¨

enue, Singapore 639798, Singapore

b

Land Transport Authority, Singapore, Singapore

Received 5 July 1999; received in revised form 1 November 1999; accepted 6 January 2000

Abstract

The objective of this investigation is to study the behaviour of precast reinforced concrete pile caps and the ultimate

load-carrying capacity. Three pile cap units were cast and tested to failure. One unit was a control pile cap cast in situ and the
other two were precast reinforced units with in situ concrete infill. The experimental results showed that the precast pile cap
behaved in a similar manner as compared with the conventional cast in situ pile cap. Furthermore, all the three units failed at
loads exceeding the failure loads predicted using conventional design methods and exhibited predicted failure modes. In addition,
there was a substantial increase in productivity as the precast pile caps could be constructed quickly and thus reducing the risk of
exposing the excavated pit to rain and possible failure of the unsupported sides.

Q 2000 Elsevier Science Ltd. All rights reserved.

Keywords: Precast; Pile caps; Ultimate load

1. Introduction

The current trend of increasing efficiency and pro-

ductivity in the management of construction activities
has placed considerable emphasis on the use of precast
members where off-site manufacture, under controlled
conditions, and uncoupled from site processes and de-
lays, can provide a constant supply of precast elements.
The use of precast elements is more crucial at locations
where heavy rains can cause serious delays due to a
difficult working environment. This is particularly evi-
dent for foundation works in soft or slimy soils where
heavy rainfall can cause the sides of the excavation to
fail and thus requires further time and effort to rectify
the excavation.

The construction of conventional cast in situ pile

Ž

.

caps see Fig. 1 requires an excavation for the pile cap,

U

Corresponding author. Tel.:

q65-790-5283; fax: q65-791-0676.

Ž

.

E-mail address: [e-mail] T.K. Chan

base preparation with a layer of lean concrete, con-
struction of forms, installation of a steel reinforcement
cage and placing of fresh concrete. This sequence of
work may easily take up to 2 days for a small pile cap
of 1

]2 m width. The steel cage may be pre-constructed

and lifted into the pit to speed up this process. This
current practice is vulnerable to heavy rains especially
when the surrounding soil is weak. Flooding followed
by failure of the sides of the pit is not uncommon.

An innovative system of precast pile caps is proposed

where no extensive ground preparation or external
forms are required. The steel cage can be constructed
separately and cast with a thin layer of concrete on the
sides to form a precast reinforced concrete shell as
illustrated in Figs. 2 and 3. This shell serves as a
permanent form for the pile cap and rests directly on
the cut-off piles. The precast shell is then infilled with
in situ concrete to complete the construction of the pile
cap. A lean concrete layer, which is normally required
to provide a firm base, may not be necessary with this
system.

0950-0618

r00r$ - see front matter Q 2000 Elsevier Science Ltd. All rights reserved.

Ž

.

PII: S 0 9 5 0 - 0 6 1 8 0 0 0 0 0 0 6 - 4

(

)

T.K. Chan, C.K. Poh

rConstruction and Building Materials 14 2000 73]78

74

Fig. 1. Unit A: conventional cast in situ pile cap.

The objectives of this project are to compare the

ultimate load-carrying capacity of precast reinforced
concrete pile caps with conventional cast in situ pile
caps and to study the behaviour of these precast units.
No previous experimental work on precast pile caps
was reported in the literature.

2. Design concept

The concept of this precast pile cap is to cast a thin

concrete shell together with the steel reinforcement
cage to provide a permanent form to hold the fresh
concrete. The sides of the steel cage are cast with a
thin layer of concrete of approximately 70 mm to
provide an outer cover of at least 50 mm to the steel
bars. Inner cover to the steel is not required, as the in
situ concrete will protect the bars. The bottom of the
steel cage is left open to enable it to rest on the top of
the piles, leaving a small clearance between the precast
shell and the ground. After proper alignment of the
precast element and the addition of the column starter
bars, the pile cap can be infilled with fresh concrete.

The bottom steel bars provide all necessary anchorage
between the precast shell and the cast in situ concrete.

Two approaches are available for the analysis of pile

Ž .

Ž .

caps: i the beam method; and ii the strut-and-tie
analogy method. The British design code for the struc-

w x

tural use of concrete 1,2 allows both the beam ap-
proach and the strut-and-tie method to be used for the
design of pile caps. When the pile cap is designed by
beam theory, it is assumed to act as a beam spanning
between the piles and is designed for usual conditions
of bending and shear. The bending moment is taken as
the sum of the moments acting from the centre of the
pile to the column face. Consequently, the reinforce-
ment in the pile cap for bending is placed uniformly
across the full width of the cap.

The pile cap can also be idealised as a strut-and-tie

w

x

model 3

]6 , with compression struts transferring load

from the column to the top of the piles, and tension
ties equilibrating the outward components of the com-
pression thrusts. The tension ties have constant forces
in them and must be anchored for the full horizontal
tie force outside the intersection of the pile and the
compression strut. No curtailment of reinforcement
within the cap is allowed and full anchorage must be
provided beyond the piles. The reinforcement bars are

Fig. 2. Unit B: precast reinforced concrete shell with cast in situ
concrete infill.

(

)

T.K. Chan, C.K. Poh

rConstruction and Building Materials 14 2000 73]78

75

Fig. 3. Unit C: precast reinforced concrete shell with cast in situ
concrete infill.

placed in concentrated bands in the direction of the tie
forces to resist the tensile forces.

The precast pile cap conforms to the assumptions of

both these methods of analysis as the embedding of the
steel bars into the precast shell provides the necessary
anchorage. The concrete in the compression zones is
confined by links, which are provided in the precast
shell. The interface between the shell and the in situ
concrete is subjected to only compressive forces and
needs no further ties.

3. Methods

Three pile cap units for a four-pile group were

fabricated. The first unit is a conventional cast in situ
pile cap of 1000

=1000=400 mm designed in accor-

dance with BS8110 and to fail in flexure. Four 150-mm
concrete cubes were utilised to represent the piles. The
second unit was of similar dimensions and steel rein-
forcements, but precast with a shell thickness of 70
mm. The third unit of 1000

=1000=300 mm was pre-

cast and was constructed with a larger amount of
reinforcement to investigate failure in shear. These
units will be labelled as A, B and C, respectively.

Table 1
Results of material tests

Ž

.

Material

Type

Strength MPa

Concrete

Pile cap A, cast in situ

39.7

Pile cap B, precast shell

33.4

Pile cap B, cast in situ infill

38.3

Pile cap C, precast shell

35.8

Pile cap C, cast in situ infill

36.4

Steel

10-mm-diameter deformed bars

480.7

The details of the specimens are shown in Figs. 1

]3.

The precast shells were left to cure for at least 28 days
before the in situ concrete infill. Three 100-mm cubes
were cast for each batch of concrete made and were
tested on the same day as the load test on the pile cap
units. Each of the test results tabulated in Table 1 is
the average of three cube specimens. Tensile tests on
the reinforcing bars were also carried out to determine
the yield strength.

Ž

A total of 20 electrical-resistance strain gauges TML

.

FLA-5-11 were installed at various locations on se-
lected reinforcing bars in the three units as shown in
Figs. 2 and 3. Displacement transducers were used to
monitor deflections at various positions on the units
during the test.

A 2000-kN testing frame was used to apply compres-

sive load onto the units. A load cell was placed on top
of a 20-mm steel plate, which was used to transfer the
load onto the column stump. The four concrete pile
supports were supported on 12-mm-thick steel plates,
which in turn were supported on rocker bearing sup-
ports as shown in Fig. 4. The load was increased at
intervals of 20 kN until failure. Care was taken to
check the unit for visible cracks. The vertical displace-
ments and strain gauge readings were automatically
recorded during the tests using a computerised data
acquisition system.

Fig. 4. Test load arrangement.

(

)

T.K. Chan, C.K. Poh

rConstruction and Building Materials 14 2000 73]78

76

Table 2
Crack width observations at first crack

Ž

.

Unit

Load

Disp.

Crack width mm

Ž

.

Ž

.

kN

mm

North

West

South

East

A

840

1.87

0.16

0.20

0.12

0.24

B

900

1.92

0.30

0.24

0.18

0.26

C

450

1.72

0.20

0.14

0.12

0.08

4. Results

The observed load-deflection relationships at the pile

cap centre, for the three units, are shown in Fig. 5. The
observed crack widths and ultimate loads of the three
pile cap units are tabulated in Tables 2 and 3, respec-
tively.

Pile cap A was predicted to fail at a total load of 890

kN by the design equations of BS8110, with flexure
being critical. However, the unit failed at 38% higher
load of 1230 kN. The load-deflection curve was linear
up to a point of more than 800 kN, where a definite
softening occurred. This point coincided with the ap-
pearance of the first crack in the unit. The load contin-
ued to increase at a lower rate up to a maximum of
1230 kN where the unit continued to deflect with no
further increase in load. The load carrying capacity
began to reduce noticeably after a deflection of more
than 6 mm. The strains in the reinforcing bars exhib-
ited a sudden increase after the appearance of the first
cracks. The strains continued to increase as the load
was increased and were all beyond the yield stress at
the maximum load.

Pile cap B, which has similar reinforcing steel ratio

and layout, has the same predicted failure load. The
ultimate load was also very similar at 1250 kN; an
increase of 41% over the BS8110 predictions. The
load-displacement behaviour was very similar to speci-
men A with the first crack at a load of 900 kN and a
0.05-mm difference in maximum centre displacement
at ultimate load compared with pile cap A. There was a
similar increase in the strains in the reinforcing bars
after the appearance of the first cracks. At the maxi-
mum load, the strains in the reinforcing bars have
exceeded yield stress.

Pile cap C, which was designed with a shallower

depth and larger amount of steel reinforcement failed
at a maximum load of 870 kN. The failure load was 7%
higher than the prediction of BS8110. There was a
significant drop in the stiffness of the pile cap after the
first crack at 450 kN. When the load reached a maxi-
mum of 870 kN, a shear failure occurred with a punch-
ing cone extending from the outside faces of the column
to the inside edges of the piles. The strains in the
reinforcing bars exhibited a sudden increase after the
appearance of the first cracks at 450 kN. At the maxi-

Table 3
Comparison of experimental and predicted failure loads

Unit

Failure

Load ratio

BS8110 predicted loads

Ž

.

load kN

Exp.

rPred.

Ž

.

Ž

.

Flexure kN

Shear kN

A

1230

890

1240

1.38

B

1250

890

1240

1.41

C

870

904

811

1.07

Fig. 5. Load-displacement behaviour of the pile cap units.

(

)

T.K. Chan, C.K. Poh

rConstruction and Building Materials 14 2000 73]78

77

mum load, all the measured strains in the reinforcing
bars were observed to have exceeded the yield stress
indicating that the pile cap was also close to its flexural
capacity.

5. Crack behaviour

The pile caps typically had very few cracks prior to

failure. Units A and B failed in flexure with flexural
cracks extending diagonally between the piles. Failure
of unit C was with a square crack pattern within the
four piles indicative of punching shear failure.

Fig. 6 shows the deformation pattern at the soffit of

the pile caps at failure and Fig. 7 shows the crack
patterns at the sides of the pile caps. The flexural
cracks originated from the centre of the soffit of pile
cap A, extending diagonally towards the piles and prop-
agating outwards. Cracks were first observed on the
vertical faces of the unit when the loading was approxi-
mately 840 kN. At this loading, the largest crack width
measured was 0.24 mm. Many new cracks developed on
the four vertical faces of the unit just before the failure
load was reached.

Cracks first appeared on the vertical faces of unit B

at approximately 900 kN and the largest recorded crack
width measured 0.30 mm. As in pile cap A, the cracks
originated from the bottom centre of the unit. Simi-
larly, the flexural cracks extended diagonally towards
the piles and failure was characterised by the rapid
development of many new cracks on the vertical faces.
The crack patterns for both units A and B were similar.

Cracks were first observed on the vertical faces of

unit C at a load of 450 kN with the largest crack width
being 0.20 mm. In contrast to pile caps A and B, there
were cracks at the bottom of pile cap C that ran

Fig. 6. Crack patterns at the soffit of the pile cap units.

parallel to the sides of the unit, indicating a ‘drop’ of
the concrete mass due to the punching shear failure.
Thus, a punching cone had extended from the loaded
area to the inside of the piles.

6. Discussion

The comparison of crack widths indicates that the

precast unit B has slightly larger crack widths com-
pared to the conventional cast in situ unit A. However,

Fig. 7. Crack patterns at the sides of the pile cap units.

(

)

T.K. Chan, C.K. Poh

rConstruction and Building Materials 14 2000 73]78

78

the first crack occurred at a marginally higher load in
the precast unit. The loads at which these cracks oc-
curred in units A and B were higher than the estimated
working loads that the pile caps were designed for. It is
evident from the crack widths, loads at which these
cracks occurred and the crack patterns that the precast
pile cap exhibits similar behaviour as a conventional
cast in situ pile cap. The first crack unit C occurred at
approximately the designed working load of the pile
cap.

The failure loads of the precast pile cap can be

predicted using conventional design equations as re-
ported in these tests. The results suggest that the
precast shell does not reduce the load-carrying capacity
or cause a weak joint in the function of the pile cap.
The interface between the in situ concrete and the
precast shell is not subjected to large stresses based on
the beam approach as the moment of resistance is
assumed to fall off according to the bending moment
diagram and therefore only nominal steel is required
beyond the pile.

w

x

According to the strut-and-tie model 4

]6 , the flow

of forces is within the concrete unit and the concrete in
the shell does not contribute to the areas which com-
prise the compression struts. The nodal zones of high
compressive stresses are entirely within the in situ
concrete, which is effectively confined by the precast
shell. Extending the steel bars into the precast segment
provides full anchorage of the tension tie. The inter-
face is therefore under compressive confining stresses
and not expected to fail.

The durability of these precast units should not

differ significantly from conventional pile caps as the
interface is not under tensile stresses. Pile cap B
cracked at a slightly higher load compared to the
conventionally cast pile cap A, confirming that the
precast shell did not induce cracking at a lower load or
at the interface. There were no cracks at the soffit of
pile caps B and C until the loads exceeded 0.7 and 0.5
of the ultimate load, respectively. It should be further
noted that more durable concrete could be provided
for the precast shell to provide additional resistance to
chemical attack although it has been reported that
small cracks of less than 0.5 mm very rarely pose any
particular corrosion risk, whatever the nature of the

w x

environment 7,8 . No shrinkage cracks were observed
at the bottom face of the precast units as the pile caps
were provided with a reinforcement ratio of 0.0016.
This was more than the recommended reinforcement
ratio of 0.0013 to be provided in two orthogonal direc-

w x

tions on the top and bottom faces of pile caps 9 .
However, a faint shrinkage crack was observed at the

top surface between the precast shell and the concrete
infill as no top steel was provided for these units. The
provision of minimum steel at the top face would
eliminate the shrinkage cracks.

7. Summary and conclusions

A comparison of the observed failure loads of the

two precast test units with predictions from the British
code indicates that the failure load of precast pile caps
was approximately 40% and 7% higher when the units
failed in flexure and shear, respectively. The behaviour
of the precast unit is similar to the corresponding cast
in situ unit with only a slight increase in crack widths.
It is therefore expected that current design equations
for conventional cast in situ construction can be used
to predict the failure loads of the pre cast units al-
though the predictions may be conservative in certain
cases. These findings lead to the conclusion that the
precast pile cap is a feasible method of construction.

In addition, there was a substantial increase in pro-

ductivity as the precast shells could be placed over the
cut piles, aligned, levelled and infilled with concrete in
a short time. The risk of exposing the excavated pit to
rain and possible failure of the unsupported sides was
also reduced.

References

w x

1

British Standards Institution. BS 8110 Part 1: 1985 structural
use of concrete. British Standards Institution, 1985.

w x

2

Taylor HPJ, Clarke JL. Some detailing problems in concrete

Ž .

frame structures. Struct Eng 1976;54 1 :19

]32.

w x

3

Canadian Standards Association. CAN3-A23.3-M94 Design of
concrete structures for buildings. Canadian Standards Associa-
tion, 1994

w x

4

Adebar P, Kuchma D, Collins MP. Strut-and-tie models for the
design of pile caps: an experimental study. ACI Struct J

Ž .

1990;87 1 :81

]92.

w x

5

Adebar P, Zhou LZ. Design of deep pile caps by strut-and-tie

Ž .

models. ACI Struct J 1996;93 4 :437

]448.

w x

6

Siao WB. Strut-and-tie model for shear behaviour in deep
beams and pile caps failing in diagonal splitting. ACI Struct J

Ž .

1993;90 4 :356

]363.

w x

7

Rowe RE, Sommerville G, Beeby AW et al. Handbook to
British Standard BS8110, 1985: structural use of concrete.
Palladian Publications Ltd, 1987.

w x

8

Beeby AW. Cracking and corrosion. Concrete in the oceans
report No. 1 CIRIA

rUEG. London: Cement and Concrete

Association, Department of Energy, 1978.

w x

9

The Institution of Structural Engineers, The Institution of Civil
Engineers. Manual for the design of reinforced concrete build-
ing structures. London, UK: The Institution of Structural Engi-
neers, October 1985.