Interface Characteristics and Laboratory Constructability

Tests of Novel Fiber-Reinforced Polymer/Concrete Piles

M. Sakr1; M. H. El Naggar, M.ASCE2; and M. Nehdi3
Downloaded from ascelibrary.org by University of Southern Queensland on 05/10/20. Copyright ASCE. For personal use only; all rights reserved.

Abstract: Conventional pile materials such as steel, concrete, and timber are prone to deterioration for many reasons. Fiber-reinforced
polymer 共FRP兲 concrete composites represent an alternative construction material for deep foundations that can eliminate many of the
performance disadvantages of traditional piling materials. However, FRP composites present several difficulties related to constructability,
and the lack of design tools for their implementation as a foundation element. This paper describes the results of an experimental study
on frictional FRP/dense sand interface characteristics and the constructability of FRP–concrete composite piles. An innovative toe driving
technique is developed to install the empty FRP shells in the soil and self-consolidating concrete is subsequently cast in them. The
experimental program involves interface shear tests on small FRP samples and uplift load tests on large-scale model piles. Two different
FRP pile materials with different roughness and a reference steel pile are examined. Static uplift load tests are conducted on different piles
installed in soil samples subjected to different confining pressures in the pressure chamber. The results showed that the interface friction
for FRP materials compared favorably with conventional steel material. It was shown that toe driving is suitable for installation of FRP
piles in dense soils.
DOI: 10.1061/共ASCE兲1090-0268共2005兲9:3共274兲
CE Database subject headings: Shear tests; Interface shear; Fiber reinforced materials; Uplift resistance; Skin friction; Sand.

Introduction acceptance in the construction industry due to the lack of efficient

installation techniques and proper design guidelines for predicting
Concrete deterioration, steel corrosion, and marine borer attack their drivability, structural characteristics and load carrying ca-
on timber piles are almost unavoidable and costly problems asso- pacity. Substituting conventional steel shells with FRP materials
ciated with deep foundations installed in harsh environments. For may result in substantial benefits for the piling industry, which is
example, the cost incurred to repair defective foundations due to the underlying motivation for the research reported herein.
corrosion in reinforced concrete and steel shell piles is in excess The frictional resistance at the interface between the pile ma-
of $2 billon per annum in the United States alone 共Iskander et al. terial and surrounding soil constitutes a considerable component
2001兲. A relatively recent trend in the deep foundation industry is of the pile’s capacity in compression and is considered the sole
to use fiber-reinforced polymer 共FRP兲 composite materials in mechanism in the uplift capacity of piles installed in granular
piles because of their light weight, high specific strength, high soils. The interface friction angle can be evaluated using either an
durability, corrosion resistance, chemical and environmental re- interface shear test or a pullout test. Comprehensive studies on the
sistance, and low maintenance cost. Composite piling has been interface behavior between conventional pile materials 共steel,
used in practice in waterfront barriers, fender piles, and bearing concrete, wood兲 and soils using different testing methods such as
piles for light structures 共Iskander and Hassan 1998兲. the simple shear test, ring shear test, and pullout tests have been
Most composite piling products are made of fiberglass or high- reported 共Fukushima and Tatsuoka 1982; Yoshimi and Kishida
density polyethylene with fiberglass reinforcement and additives 1982; and Uesugi and Kishida 1986兲. However, a few studies are
to improve mechanical properties, durability, and ultraviolet pro- available on the interface between FRP and soils. Frost and Han
tection. However, FRP composite piles have not yet gained wide 共1999兲 evaluated the friction between sand and FRP and steel
materials experimentally using the interface shear test. They
1 found that FRP and steel materials exhibit similar relationships
Project Engineer, Thurber Engineering Ltd., Edmonton AB, Canada
between the peak interface friction coefficients and the relative
T6E 6A5; formerly, PhD Candidate, Dept. of Civil and Environmental
Engineering, Univ. of Western Ontario, London ON, Canada N6A 5B9. roughness of granular soils. Pando et al. 共2002兲 conducted inter-
2
Professor, Dept. of Civil and Environmental Engineering, Univ. of face shear tests on two commercially available FRP composite
Western Ontario, London, Ontario, Canada N6A 5B9 共corresponding piles to evaluate their interface friction characteristics. They
author兲. E-mail: [email protected] found that relative roughness, surface hardness and angularity of
3
Associate Professor, Dept. of Civil and Environmental Engineering, the soil grains are the main factors affecting the interface friction
Univ. of Western Ontario, London, Ontario, Canada N6A 5B9. between the FRP composites and granular soils.
Note. Discussion open until November 1, 2005. Separate discussions The installation of cylindrical thin-walled shells in dense sand
must be submitted for individual papers. To extend the closing date by
is difficult, and in many cases hard driving represents a major
one month, a written request must be filed with the ASCE Managing
Editor. The manuscript for this paper was submitted for review and pos- challenge for pile construction. This is especially true when using
sible publication on December 23, 2002; approved on October 10, 2003. low stiffness material such as FRP composites or small gage steel
This paper is part of the Journal of Composites for Construction, Vol. 9, pipes. The conventional driving technique that involves striking a
No. 3, June 1, 2005. ©ASCE, ISSN 1090-0268/2005/3-274–283/$25.00. hammer at the pile head may cause extensive damage to the pile

274 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005

J. Compos. Constr., 2005, 9(3): 274-283

 Table 1. Geometry of Model Piles 共pullout tests兲. The IST tests on small samples are conducted
Diameter Thickness b
Length using two different FRP materials and steel. Uplift pile load tests
d aa ts L 共pullout tests兲 on FRP and steel piles under different confining
Pile ID 共mm兲 共mm兲 共mm兲 Fiber layup pressures that represent different in situ soil depths are conducted.
Two driving techniques are used to install the piles in dense sand
Steel 168.3 6.35 1,524 NAc
in a laboratory setup. The novel toe driving method and the con-
FRP I 162.4 6.00 1,524 关+55° / −55° 兴2
ventional pile driving at head are used to install FRP piles in
FRP II 178.0 7.8 1,524 关CSM/ 0°3 / 90° / CSM兴
a
dense soils. A detailed description of pile materials, SCC proper-
da = average diameter of the model pile shaft. ties, testing methods, and pile installation process and results are
b
ts = thickness of pile wall. presented in the next sections.
c
NA= not available.
Downloaded from ascelibrary.org by University of Southern Queensland on 05/10/20. Copyright ASCE. For personal use only; all rights reserved.

head, or damage at the pile toe if boulders or hard objects are Material Properties
encountered. Mirmiran and his associates 共Mirmiran et al. 2002,
and Mirmiran and Shahawy 2003兲 conducted field tests using Piles
FRP empty tubes, FRP–concrete filled tubes, and prestressed
concrete piles to evaluate their drivability. They observed that Two different FRP composites and steel materials are used in this
the empty FRP tube failed due to buckling during installation at study. The two FRP composite materials are referred as to FRP I
embedment depth of about 3.5 m. Their results also showed that and FRP II. A cylindrical steel open-ended pile is used as a ref-
neither pile head damage nor separation between FRP tube and erence material. It is made of cold drawn steel tubing with a
concrete core were observed when driving FRP concrete filled modulus of elasticity=2.15⫻ 105 MPa. Table 1 shows the geom-
tube. An innovative technique for driving a tubular pile at its toe etry of different piles considered in this study. The dimensions of
using an impact hammer inside the pile, called the toe driving, piles used in this study were chosen to represent a pile segment
was developed and tested in Canada 共Sakr et al. 2003兲. This tech- with total length of 1.524 m and embedment length of 1.2 m.
nique has the potential to be very efficient for the installation of The FRP I and FRP II shells were fabricated using the filament
FRP shells and thin-walled steel piles. winding technique. The FRP I shell, with a wall thickness of
6 mm, consisted of 3.3 mm structural laminate with layup 共i.e.,
number of layers of fiber, fiber orientation, and layer thickness兲 in
Objectives and Scope the form of 关+55° / −55° 兴2, and a 2.7 mm inner liner with glass
content of 30%. Burnout of coupons showed an overall glass
This study is focused on the interface characteristics and con- content of 53.5%. The FRP II with a wall thickness of 7.8 mm
structability of FRP–concrete hybrid piles. Self-consolidating consisted of structural laminate with layup in the form of chopped
concrete 共SCC兲 is used to assure the structural integrity of piles strand mat 共CSM兲 关CSM/ 0°3 / 90° / CSM兴. The exterior CSM was
and to eliminate air pockets. The SCC is cast in FRP tubes that used to increase the surface roughness. The E-glass fiber was
provide durable reinforcement in harsh environments. The prime impregnated with two-part epoxy of laminating resin L135 and
objectives of this study are to evaluate the FRP/sand interface hardener L33-138 manufactured by Martin G. Scheufler, Ger-
behavior for deep foundation applications and to develop an effi- many, at a volumetric ratio of 60/ 40. The FRP II shell was al-
cient installation technique for FRP–SCC hybrid piles. The inter- lowed to cure for 24 h at a temperature of 20– 25° C before re-
face characteristics between FRP and dry dense sand and the con- moving it from the mandrel. The epoxy resin L135 allowed for
structability of FRP–SCC hybrid piles are examined in this paper. the development of the full strength of the FRP composite after
Two different methods are used to evaluate the interface charac- 7 days at a temperature of 20– 25° C. Table 2 shows the mechani-
teristics; the interface shear test 共IST兲 and uplift pile load tests cal properties of resin and FRP pile shells.

Table 2. Properties of Pile Materials

FRP I FRP II

FRP I FRP II
Property 共composite兲 Liner 共composite兲 Resin
Density 共Mg/ m3兲 1.60 1.42 1.65 1.18
Glass content 共%兲 53.5 30 60 —
Compression
Strength 共MPa兲 60 69 380 120
Modulus of elasticity 共GPa兲 8.5 6.9 17 3
Tension
Strength 共MPa兲 193 69 440 70
Modulus of elasticity 共GPa兲 10.5 6.9 19 3
Flexure
Strength 共MPa兲 76 69 480 110
Modulus of elasticity 共GPa兲 4.9 6.9 17 3
Interlaminar shear strength 共MPa兲 — — 38 —

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005 / 275

J. Compos. Constr., 2005, 9(3): 274-283

Downloaded from ascelibrary.org by University of Southern Queensland on 05/10/20. Copyright ASCE. For personal use only; all rights reserved.

Fig. 1. Shear stress versus horizontal displacement 共direct shear test兲

for dense sand Fig. 2. Peak and steady state envelopes for dense Fanshawe brick
sand

Self-Consolidating Concrete
of 2,420 kg/ m3 and slump flow of 550 mm. The 7 and 28 day
The FRP shells were filled with a cost-effective SCC specifically compressive strengths of the SCC placed in the piles were 42 and
developed for deep foundation applications. Its mixture propor- 58 MPa, respectively.
tions 共kilogram/ meter3兲 are as follows: Portland cement ASTM
Type I 共200兲, ground granulated blast-furnace slag 共200兲, water
Soil Sample
共150兲, coarse aggregate with 9.5 mm maximum particle size
共850兲, fine aggregate 共850兲, and water-reducing admixture Fanshawe Bricksand used in the tests consisted of fine subround
共3.0 L / m3兲. Fresh concrete properties are as follows: unit weight to round air-dried sand. The soil used was classified as poorly

Fig. 3. Pile surface roughness profiles for: 共a兲 fiber-reinforced polymer I pile, 共b兲 fiber-reinforced polymer II pile, and 共c兲 steel pile surface

276 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005

J. Compos. Constr., 2005, 9(3): 274-283

 graded sand, with particle sizes in the range of 0.075– 2.00 mm
with an effective diameter D50 = 0.26 mm and uniformity coeffi-
cient Cu = 2.14. A standard density test of the sand showed that it
had a maximum unit weight of 17.72 kN/ m3 and a minimum unit
weight of 14.66 kN/ m3. The maximum void ratio emax = 0.794 and
the minimum void ratio emin = 0.484. The relative density of the
soil samples used for the current testing program was about
90± 2.5%. Soil samples were prepared using a raining technique
共pulveration兲 in order to ensure uniformity of the dense soil
samples.
Fig. 4. Sketch of interface shear test setup
Downloaded from ascelibrary.org by University of Southern Queensland on 05/10/20. Copyright ASCE. For personal use only; all rights reserved.

Methods for Evaluation of Fiber-Reinforced

Polymer/Sand Interface Characteristics Fig. 2 shows that both envelopes are slightly nonlinear and that
the shear stress slightly decreased with an increase in the normal
The internal friction angle of the soil sample used in the tests was stress. The nonlinearity is attributed to soil dilatancy 共Bolton
evaluated using the direct shear test 共DST兲. The surface roughness 1986兲.
of the pile material was evaluated using a Taylor Hobson Talysurf
stylus profilometer. Two different methods were used to evaluate
Surface Roughness Measurement
the characteristics of FRP/dense sand interface in this study: in-
terface shear tests on small FRP composite samples and uplift pile The surface roughness of FRP I, FRP II, and steel pile materials
load testing on model piles. The details of the different tests are was measured using a Taylor–Hobson Talysurf stylus profilome-
described below. ter. Coupons 60⫻ 60 mm were cut from the pile material. A series
of profiles in the longitudinal direction of the piles, each 5 mm in
length and 2.45 mm in width were made on the surface of each
Internal Friction Angle of Sand
coupon. Fig. 3 shows the surface roughness profile for different
The DST was used to find the internal shear frictional angle ␾ for materials considered in this study. The average roughness for FRP
Fanshawe Bricksand. The DST specimen had dimensions of 60 II shown in Fig. 3共b兲 was 1 order of magnitude higher than FRP
⫻ 60 mm and a height of 19 mm. The direct shear tests were I 共with ply angles of ±55°兲 shown in Fig. 3共a兲. However, both
carried out in accordance with ASTM Standard D3080-98 steel 关Fig. 3共c兲兴 and FRP I surfaces have almost similar average
共ASTM 1998兲 using normal stresses that ranged from surface roughness. The values of the average roughness of FRP I,
25 to 200 kPa. Dry samples were tested at a loading rate of FRP II, and the steel pile material were about 1.66, 11.54, and
0.304 mm/ min. The inside of the walls were coated with a thin 1.05 ␮m, respectively. In general, the roughness profile for FRP
film of vacuum grease to minimize the friction between the sand material was more irregular than for the steel material.
and soil box. Fig. 1 shows the shear stress–horizontal displace-
ment behavior for Fanshawe sand over a range of normal stresses.
Interface Shear Test
A peak shear stress followed by reduction to a residual 共postpeak兲
shear stress, characteristic of drained tests on dilatant specimens, The ISTs were conducted on FRP composite specimens using the
was observed for all of the stress–displacement curves. Fig. 2 original direct shear box as shown in Fig. 4. The interface shear
shows the peak and residual shear stress–normal stress envelopes. tests were carried out in accordance with ASTM Standard

Fig. 5. Schematic elevation of axial uplift loading setup

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005 / 277

J. Compos. Constr., 2005, 9(3): 274-283

 D5321-97 共ASTM 2002兲. The FRP composite sections with di- Table 3. Summary of Interface Shear Tests
mensions of 60⫻ 60 mm and a height of 9.5 mm were situated in Normal
the bottom half of a direct shear box so that the top surface was Interface stress ␶ pa d pb ␶ rc d rd
perfectly aligned with the shearing plane. The top part of the 共pile/soil兲 共kPa兲 共kPa兲 共mm兲 共kPa兲 共mm兲
direct shear box was then assembled and positioned underneath
FRP I/sand-A 25 14.8 0.46 13.8 4.86
the raining apparatus to fill the top half of the shear box. The ISTs
75 36.4 0.46 35.4 4.86
were conducted at different values of normal stress ranging from
150 67.5 0.61 68.75 4.86
25 to 200 kPa. Tests were conducted on dry samples at the same
rate as DST 共0.304 mm/ min兲. Frost and Han 共1999兲 conducted a 200 93.8 0.76 92.6 4.86
series of DST and IST on a dry granular soil using a different rate FRP I/sand-B 25 16.0 0.76 16.9 2.43
of shearing ranging from 0.25 to 5.08 mm/ min. They concluded 75 45.7 0.91 49.2 2.74
that the rate of shearing had no significant effect on the interface 200 110.56 1.52 119.3 4.86
Downloaded from ascelibrary.org by University of Southern Queensland on 05/10/20. Copyright ASCE. For personal use only; all rights reserved.

friction angle between the FRP and granular materials. FRP I/sand-C 25 17.1 0.76 15.4 3.95
The FRP I specimens for an off shelf pipe were prepared by
75 45.7 0.76 46.2 3.65
cutting slices 10 mm in width, 60 mm in length from the pipe and
200 120.4 1.06 121.0 4.86
then gluing them on the top of an FRP composite section of 60
⫻ 60 mm and height of 3.15 mm to achieve the required thick- FRP II/sand-A 25 18.6 0.61 14.7 3.0
ness of 9.5 mm for the interface shear test. The FRP II specimens 75 49.0 0.61 39.7 3.65
were prepared during the fabrication of piles using the same pro- 150 92.6 0.61 76.1 1.82
cess. All specimens were tested dry in laboratory temperature 200 121.1 0.76 97.2 3.34
共20° C兲 without conditioning to simulate the interface character- FRP II/sand-B 25 19.9 0.46 16.0 3.34
istics between FRP and dry sand. 75 54.2 0.61 42.9 2.74
200 133.3 0.91 109.3 2.13
Model Piles Uplift Load Test FRP II/sand-C 25 19.4 0.61 15.6 2.12
FRP composite and steel pile segments 1.524 m in length were 75 50.6 0.61 43.6 2.74
installed in dense sand in a pressure chamber to a total penetration 200 131.8 0.91 110.6 3.95
depth of 1.2 m. The pressure chamber consisted of a containment Steel/sand 25 14.7 0.61 14.3 4.86
steel cylinder with 1.34 m inside diameter and 1.52 m in height. 75 40.9 0.91 44.7 4.86
The cylinder was lined with circumferential and vertical air blad- 150 70.3 0.61 74.6 4.86
ders that can apply different pressures to the soil sample in order 200 92.8 1.06 106.1 4.86
to simulate the radial and vertical confinement stresses of various a
␶ p = peak shear stress.
soils at different in situ depths. Ambient effective vertical/radial b
d p = displacement at peak shear stress.
stress combinations in the pressure chamber of 60/ 30, 90/ 45, and c
␶r = residual shear stress.
120/ 60 kPa were maintained during the installation and subse- d
quent uplift loading tests to simulate the vertical and horizontal dr = displacement at residual shear stress.
stresses in normally consolidated dry sand conditions at depths of
4.0, 6.0, and 8.0 m, respectively. used for FRP I and FRP II samples. The three FRP samples were
The piles were instrumented externally with seven full bridge denoted by letters A, B, and C. It is worth mentioning that all
strain gage circuits to measure the skin friction distribution along specimens were situated in the bottom half of direct shear box so
the pile shaft. The strain gages were distributed over the length of that the fiber direction was similar to the orientation of the fibers
the piles such that the first bridge was approximately 206 mm in the pile during pile axial loading 共i.e., shear at the pile/soil
from the pile head above the sand surface and the remaining interface兲. It can be seen from Table 3 that the peak and residual
bridges were disturbed evenly along the pile length. Seven uplift interface shear stress values for both FRP I/sand and steel/sand
tests were conducted on FRP composites and steel piles. Fig. 5 were approximately similar. However, for FRP II/sand interface,
shows a schematic of the axial loading setup. the peak and residual shear stress values were higher than those
The effect of soil particle size is minimized since the ratios of for FRP I/sand interface. The scatter in the results for FRP I
pile circumference to the mean particle size d50 ranged from samples may be attributed to the curved slices that were used in
2,000 to 2,300 and the ratios of da / d50 ranged from 650 to 750. preparing the sample was not flat enough to assure an intimate
Ovesen 共1979兲 showed that the particle size effects become sig- contact between the surface and the soil sample. However, a lin-
nificant for pile circumference to d50 ratios between 20 and 40 for ear fit for FRP I/sand was obtained with a standard deviation of
foundations on quartz sand. Also, Bolton et al. 共1999兲 showed that 0.97 that suggests a reasonable overall consistency of the results
the soil particle size does not affect the cone penetrometer results was achieved.
for da / d50 values higher than 28. Fig. 6 shows the shear stress–horizontal displacement relation-
ship for FRP I/sand, FRP II/sand, and for steel/sand interfaces.
FRP I/sand and steel/sand interfaces 关Figs. 6共a and c兲兴 have ap-
Fiber-Reinforced Polymer/Dense Sand Interface proximately similar plastic postpeak behavior. The curves also
Characteristics show that FRP I/sand and steel/sand have a slightly stress-
hardening behavior, especially at higher normal stress. This may
be attributed to plowing of particles into FRP material surfaces.
Interface Shear Tests
The FRP II/sand interfaces 关Fig. 6共b兲兴 exhibit a postpeak
The results for interface tests performed for FRP I, FRP II, and displacement-softening behavior. The peak in the shear stress–
steel surfaces are summarized in Table 3. Three specimens were horizontal displacement curves is more obvious for cases of high

278 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005

J. Compos. Constr., 2005, 9(3): 274-283

Downloaded from ascelibrary.org by University of Southern Queensland on 05/10/20. Copyright ASCE. For personal use only; all rights reserved.

Fig. 6. Interface shear stress versus horizontal displacement fiber-reinforced polymer for: 共a兲 sand/fiber-reinforced polymer I, 共b兲 sand/fiber-
reinforced polymer II, and 共c兲 sand/steel interface

normal stress. This behavior may be attributed to greater inden- angles were higher than those for the steel material. The FRP
tation of sand particles into the FRP specimen surfaces at higher I/sand and steel/sand peak interface friction angles were almost
normal stresses, promoting particle rearrangement. No dilatancy similar and the residual FRP I/sand friction angle was slightly
was observed in the sand mass during the interface shear tests, lower than that of the steel/sand interface.
even though the sand specimens were dense. This behavior con-
firms that the stress induced during shearing was not large enough
Uplift Pile Load Tests
to disturb the whole sand specimen because the FRP surface was
relatively smooth. Therefore, sliding dominated the shear resis- The FRP/dense sand interface characteristics were also evaluated
tance 共O’Rourke et al. 1990兲. from uplift load tests carried out on model piles in a laboratory
Fig. 7 shows the shear stress envelope for FRP I/sand, FRP setup that closely simulated field conditions. The following sec-
II/sand, and steel/sand interface. The shear stress envelope for tions describe the experimental procedures for the uplift pile load
FRP I/sand, and steel/sand interfaces is approximately linear, tests.
while the FRP II/sand interface shows postpeak displacement-
softening behavior. The peak and residual interface frictional
Uplift Load Tests
angles between sand and FRP and steel materials determined
from IST tests are summarized in Table 4. It can be seen from The model piles were subjected to uplift load tests after the
Table 4 that FRP II/sand peak and residual interface friction completion of driving. Different combinations of radial and ver-

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005 / 279

J. Compos. Constr., 2005, 9(3): 274-283

Downloaded from ascelibrary.org by University of Southern Queensland on 05/10/20. Copyright ASCE. For personal use only; all rights reserved.

Fig. 7. Interface shear stress envelope for steel, fiber-reinforced poly-

mer I and fiber-reinforced polymer II interfaces

tical confinement pressures were applied to the soil sample in the

pressure chamber during the uplift load tests in order to simulate
the in situ state of stress at different depths in normally consoli-
dated sand. One combination of radial and vertical confining pres-
sure was applied to the soil sample 2 h before the start of the
uplift test and was maintained throughout the test. The confining
pressure was monitored using pressure gages mounted on the air
pressure lines and by pressure cells installed in the soil sample
during its preparation. The piles were tested at three different
radial and vertical confinement pressure combinations of 30/ 60, Fig. 8. Shear stress versus uplift displacement for: 共a兲 fiber-
45/ 90, and 60/ 120 kPa to simulate the in situ state of stress at reinforced polymer I pile, 共b兲 fiber-reinforced polymer II pile, and 共c兲
depths equal to 4, 6, and 8 m below ground surface, respectively. steel pile
The model piles were tested in tension up to a total axial
displacement of 25 mm at a rate of 5 mm/ min. The load was
applied at the pile head using a hydraulic jack and both the ap- linear response at the early stages of displacement 共up to 3 mm兲.
plied load and the upward movement were monitored simulta- Unlike the FRP I pile, the FRP II pile exhibited postpeak
neously during the uplift tests. The results for the uplift loading displacement-softening behavior. An obvious peak is observed at
tests under various testing pressures are plotted in Fig. 8 in terms an uplift displacement of about 11 mm. This behavior can be
of the average shear stress versus the upward displacement. The attributed to the high relative roughness of the FRP II interface.
average shear stress was obtained by dividing the measured load The rough surface of the FRP II pile restricted the particle rear-
by the surface area of the pile shaft. rangement at low displacement levels. When the shear strength
Fig. 8 shows that when piles were tested at a high pressure exceeded the internal friction, soil particles at the interface started
combination 共60/ 120 kPa兲 they offered higher shear resistance in to move and softening behavior occurred. The steel pile 关Fig.
the uplift mode than when tested at the low pressure combination 8共c兲兴 exhibited similar behavior to the FRP I pile in terms of an
共30/ 60 kPa兲. It can be seen from Fig. 8共a兲 that the FRP I pile initial linear response 共up to 2.5 mm兲 followed by a nonlinear part
tested at three different confining pressures 共30/ 60, 45/ 90, and and approximately postpeak plastic behavior. The similarity be-
60/ 120 kPa兲 showed similar initial stiffness at the early stages of tween IST results 共Fig. 6兲 and uplift loading tests 共Fig. 8兲 were
loading 共up to 2.5 mm兲. However, piles tested at high confining obvious, which demonstrated that IST tests reasonably captured
pressure combinations 共45/ 90 and 60/ 120 kPa兲 showed a stiffer the interface behavior of FRP materials developed during axial
response at higher displacement levels. The peak shear stresses pile loading tests.
were achieved within a displacement of 16 mm. The FRP II pile
关Fig. 8共b兲兴 tested at pressures of 30/ 60 and 60/ 120 kPa showed a Ultimate Pile Uplift Capacity of Fiber-Reinforced
Polymer Piles
Kulhawy and Hirany 共1989兲 reviewed the existing interpretation
Table 4. Soil Friction Angle and Interface Friction Angle
methods for axial uplift pile load tests and recommended that the
Soil properties FRP II/sand FRP I/sand Steel/sand load corresponding to a displacement equal to 12.7 mm be inter-
␾p ␾r ␦p ␦r ␦p ␦r ␦p ␦r preted as the ⬙failure⬙ load. De Nicola and Randolph 共1999兲 de-
共°兲 共°兲 共°兲 共°兲 共°兲 共°兲 共°兲 共°兲 fined the ultimate tensile pile capacity as the pile head load cor-
responding to a pile head displacement of 5% of the pile diameter.
37.0 31.0 32.3 27.5 27.6 28.2 26.6 26.6
For simplicity, the uplift pile capacity Qup was evaluated based on

280 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005

J. Compos. Constr., 2005, 9(3): 274-283

 Table 5. Uplift Capacity of Tested Piles at Different Confining Pressures Constructability of Fiber-Reinforced Polymer–Self-
Testing pressure Uplift capacity Uplift capacity Consolidating Concrete Piles
共radial/vertical兲 共load test兲 共estimated兲
Pile 共kPa兲 Qup 共kN兲 Qup 共kN兲 Capacity ratio The FRP–SCC hybrid piles examined in the current study have
FRP I 30/ 60 23.7 22.3 0.94
the potential to solve the problems associated with durability and
long-term performance 共Mirmiran and Shahawy 2003兲. The pile
45/ 90 28.2 28.7 1.02
is made of an FRP shell filled with SCC. The FRP shell resists the
60/ 120 32.3 36.7 1.14
bending and tensile stresses while the concrete infill resists the
FRP II 30/ 60 34.4 31.8 0.92 compressive loads. The FRP shell eliminates the need for steel
60/ 120 45.3 50.9 1.12 reinforcement, which is used in the case of conventional concrete
Steel 30/ 60 21.4 20.7 0.97 piles. However, the constructability of these piles is a cause of
concern because of the low stiffness of the FRP shell. Hence,
Downloaded from ascelibrary.org by University of Southern Queensland on 05/10/20. Copyright ASCE. For personal use only; all rights reserved.

60/ 120 33.7 36.6 1.08

techniques that can be used for their installation in dense sand are
explored.
Piles can be installed into the soil using either impact hammers
the uplift load corresponding to a pile head displacement of 5% of or vibratory pile drivers. Conventional pile driving process is very
the pile diameter. The uplift capacity, Qup of different piles used efficient for stiff piles constructed in relatively weak soil deposits
in this study is presented in Table 5. It can be noted from Table 5 such as loose to medium dense sand and soft to stiff clay. How-
that the uplift capacity of the FRP II pile is about 1.3–1.6 times ever, it may cause some problems when driving flexible or thin-
that of the steel pile. The uplift capacity of the FRP I and steel walled piles in relatively strong soils such as dense sand or hard
piles is approximately the same. clay or when driving empty FRP tubes into the ground to large
depth.
Comparison Between Interface Shear Tests and Uplift The vibratory pile driver consists of imparting a dynamic ver-
Loading Tests tical force from a set of two rotating, eccentrically set weights.
These weights are positioned so that their horizontal force com-
For cohesionless soils, the ultimate uplift capacity can be deter- ponents are balanced by each other while their vertical force com-
mined from the following expression: ponents are added. The vibrations generated by these drivers liq-
uefy the surrounding soil and hence reduce the skin friction along

冕
L the pile surface during pile driving. Vibratory hammers are most
Qup = f s p dz 共1兲 effective for pile installation in cohesionless soils 共Prakash and
0
Sharma 1990兲. Vibratory pile drivers can be utilized to install
where f s = ultimate shaft friction, which can be evaluated as f s FRP concrete hybrid piles or empty FRP tubes since there is no
= ␤␴⬘v, where ␤ = combined shaft resistance factor and ␴␯⬘ impact load and therefore bucking associated with low stiffness
= effective overburden stress, L = pile length; and p = its perimeter. FRP tube is considerably reduced. However, this pile installation
The value of ␤ suggested in the Canadian Foundation Engineer- process disturbs the soil surrounding the pile, and reduces the skin
ing Manual 共CFEM 1992兲 for cylindrical piles driven in dense friction and pile capacity. Most design codes recommend design-
sand under uplift loading ranges from 0.4 to 0.6. The values of ␤ ing piles installed using vibratory pile drivers by adopting the
evaluated from the results of the interface shear tests using Eq. 共2兲 same parameters recommended for cast-in-place piles, which lead
are 0.6, 0.49, and 0.5 for the FRP II, FRP I, and steel piles, to a much lower capacity. Moreover, vibration propagation can be
respectively. These values lie within the range suggested in the high, leading to settlement in nearby structures.
CFEM for driven piles in dense sand. Recently, an innovative technique for driving a tubular FRP
The shaft resistance factor ␤ is influenced by the angle of and steel piles at its toe using an impact hammer inside the pile
internal friction of the soil ␾ the frictional angle of pile–soil in- called toe driving was designed and built at the Univ. of Western
terface ␦, the method of installation, and the original state of Ontario, Canada. The principles of toe driving technique and its
stress in the ground, pile material, and size and shape of the pile. components are described in detail by Sakr et al. 共2003兲. The
The shaft resistance factor ␤ can be given by: main idea of this device is to transfer most of the hammer impact
directly to the soil at the pile toe creating a cavity and to utilize
only portion of the impact energy to pull the pile down into the
␤ = Ks tan ␦ 共2兲
created cavity. Toe driving offers several advantages including: it
where Ks = coefficient of lateral earth pressure after pile installa- eliminates the risk of pile damage at the pile head and toe in hard
tion and depends on the method of pile installation and the den- driving conditions; it facilitates driving thin-walled piles and low
sity of the sand. For large displacement driven piles, similar to compressive strength materials such as FRP composites; it re-
those considered in this study, Ks varies between 0.5 and 1. The duces driving energy and time and improves driving efficiency
shaft friction factor ␤ was calculated using Eq. 共2兲 and the values since most of the impact load is transferred directly to the soil,
of interface friction angle ␦ that were determined from IST tests. and it increases the pile load bearing capacity 共Benamar 2000兲.
In this study, the piles were driven in very dense sand 共Dr
= 90% 兲 and thus Ks was assumed to be one. The uplift capacity
Pile Driving Tests
for all tested piles was calculated using Eq. 共1兲 and ␤ values
obtained from Eq. 共2兲, and the results are presented in Table 5. It The FRP–SCC hybrid pile can be constructed in two ways: driv-
can be seen that the calculated values of the uplift capacity agree ing the FRP shell in the ground using a toe driving device first
reasonably well with the measured values, thus confirming the and then pouring the in-fill concrete; or pouring the concrete in
usefulness of the inexpensive IST tests to capture the skin friction the shell above ground then driving the composite section in the
characteristics of FRP piles. ground using conventional head driving. Both techniques are ex-

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005 / 281

J. Compos. Constr., 2005, 9(3): 274-283

Downloaded from ascelibrary.org by University of Southern Queensland on 05/10/20. Copyright ASCE. For personal use only; all rights reserved.

Fig. 9. Driving records for empty fiber-reinforced polymer I pile,

fiber-reinforced polymer I-self-consolidating concrete and steel piles
using toe driving technique and head driving

amined in this study. In addition, for comparison reasons, a steel

pile is also driven using the two different installation methods.
The FRP I empty tube, FRP I-SCC hybrid piles, and steel
piles, considered in this study, were driven in dense sand enclosed
in a pressure chamber using both conventional driving at the pile
head and the innovative toe driving technique. A pile cushion of
51 mm plywood was used when driving the pile at its head to
reduce the peak impact force and to reduce the risk of pile head
damage. Radial and vertical confinement pressures of 30 and
60 kPa, respectively, were applied to the soil sample in the pres-
sure chamber during pile driving. The pile penetration per blow
was maintained in the normal range achieved during in situ pile
driving 共2.5– 5 mm/ blow兲. Fig. 10. Stress wave data for steel cylindrical pile: 共a兲 at pile head,
Fig. 9 shows driving records for FRP and steel piles installed and 共b兲 near pile toe
using toe and head driving techniques in terms of toe penetration
共z兲 normalized by pile diameter 共d兲. Fig. 9 shows that empty FRP
I pile installed using toe driving required about 43% less blow
count than the FRP–SCC pile installed using conventional head the FRP composite material has a significant effect on the FRP/
driving. The easier driving of the empty FRP pile using toe driv- sand behavior during the interface shear test. The rough FRP
ing, despite its low impedance, is attributed to applying the driv- surface coupled with rounded sand results in a brittle interface
ing energy directly to the soil, which increases the overall imped- behavior and strain softening at high displacement levels.
ance of the pile-driving system. A comparison of the steel and Uplift loading tests were conducted on two FRP piles, and a
FRP pile installation using toe driving showed that the steel pile steel pile driven in normally consolidated dense sand 共K0 = 0.5兲
required about 7% less blow count than the FRP pile. It is also and tested at different radial/vertical confining pressures 共ranging
noted that for the steel pile, the driving record was about the same from 30/ 60 to 60/ 120 kPa兲. The uplift loading tests showed simi-
regardless of the installation method. No signs of damage were lar pile interface behavior to the IST. The uplift pile capacity
observed when driving empty FRP tubes even though six driving obtained from uplift loading tests compared reasonably well with
processes were conducted on the same pile. The FRP–SCC hybrid those calculated based on ISTs. Therefore, the inexpensive IST
piles were also installed using head driving without showing any tests can be used efficiently to capture the skin friction character-
visible cracks or damage. istics of FRP piles installed in granular soils.
Fig. 10 shows the stress wave data for the steel pile and Fig. 11 The values of the peak interface friction angle for the two FRP
shows the stress wave data for the FRP I and FRP I–SCC piles, materials/dense sand used in this study 共31 and 26°兲 were similar
both at the pile head and near the pile toe. Figs. 10 and 11 show to, if not higher than, the friction angle for the steel pile/dense
that the stress level was similar for steel and FRP piles. The sand 共26.6°兲. Therefore, the use of FRP materials in deep founda-
figures also show that the peak stresses in the case of toe driving tions should be an added advantage due to their resistance to
were mainly tensile and magnitude of the peak stress was about degradation in addition to increased shaft frictional resistance.
40% of that for the head driving case. Drivability tests conducted as part of this study showed that
FRP–SCC hybrid piles can be constructed either by installing the
composite pile using head driving or installing the FRP shell
Summary and Conclusions
using toe driving and then casting the SCC inside the shell. How-
The IST tests and uplift pile load tests were carried out on two ever, toe driving is shown to be more efficient and potentially
FRP composite materials and on a steel specimen in dry dense more advantageous for the installation of FRP piles in a situation
sand. The IST test results showed that the relative roughness of where hard driving is encountered.

282 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005

J. Compos. Constr., 2005, 9(3): 274-283

 American Society for Testing and Materials 共ASTM兲. 共2002兲. “Standard
test method for determining the coefficient of soil and geosynthetic or
geosynthetic and geosynthetic friction by direct shear method.” D
5321-97, West Conshohocken, Pa.
Benamar, A. 共2000兲. “A new technique to drive piles: ‘Drive-the-hole
piling.” Proc., 6th International Conf. on the Application of Stress
Wave Theory to Piles, Sussumu Niyama and Jorge Beim, eds., Sao
Paulo, Brazil, 117–120.
Bolton, M. D. 共1986兲. “The strength and dilatancy of sands.” Geotech-
nique, 36共1兲, 65–78.
Bolton, M. D., Gui, M. W., Garnier, J., Corte, J. F., Bagge, G., Laue, J.,
and Renzi, R. 共1999兲. “Centrifuge cone penetration tests in sand,”
Geotechnique, 49共4兲, 543–552.
Downloaded from ascelibrary.org by University of Southern Queensland on 05/10/20. Copyright ASCE. For personal use only; all rights reserved.

Canadian foundation engineering manual 共CFEM兲. 共1992兲. Canadian

Geotechnical Society, BiTech Publication, Vancouver, Canada.
De Nicola, A., and Randolph, M. F. 共1999兲. “Centrifuge modelling of
pipe piles in sand under axial loads.” Geotechnique, 49共3兲, 295–318.
Frost, J. D., and Han, J. 共1999兲. “Behavior of interfaces between fiber-
reinforced polymers and sands,” J. Geotech. Geoenviron. Eng.,
125共8兲, 633–640.
Fukushima, S., and Tatsouka, F. 共1982兲. “Deformation and strength of
sand in torsional simple shear.” Proc., IUTAM Conf. on Deformation
and Failure Granular Materials, Delft, The Netherlands, 371–379.
Iskander, M. G., and Hassan, M. 共1998兲. “State of restored practice re-
view in FRP composite piling.” J. Compos. Constr., 2共3兲, 116–120.
Iskander, M. G., Hanna, S., and Stachula, A. 共2001兲. “Drivability of FRP
composite piling.” J. Geotech. Geoenviron. Eng., 127共2兲, 169–176.
Kulhawy, F. H., and Hirany, A. 共1989兲. “Interpretation of load tests on
drilled shafts—Part 2: Axial uplift.” Proc., Foundation Engineering,
ASCE, New York, 1150–1159.
Mirmiran, A., Shao, Y., and Shahawy, M. 共2002兲. “Analysis and field
tests on the performance of composite tubes under pile driving im-
pact.” Compos. Struct., 55, 127–135.
Fig. 11. Stress wave data for fiber-reinforced polymer I and fiber- Mirmiran, A., and Shahawy, M. 共2003兲. “Composite pile: A successful
reinforced polymer I-self-consolidating concrete piles: 共a兲 at pile drive.” Concr. Int., 25共3兲, 89–94.
head and 共b兲 near pile toe O’Rourke, T. D., Druschel, S. J., and Netravali, A. N. 共1990兲. “Shear
strength characteristics of sand-polymer interfaces.” J. Geotech. Eng.,
116共3兲, 451–469.
Acknowledgments Ovesen, N. K. 共1979兲. “Centrifugal testing applied to bearing capacity
problems of footing on sand,” Geotechnique, 25共2兲, 394–401.
The writers would like to thank Mr. Carl Ealy, Head of Deep Pando, M. A., Filz, G. M., Dove, J. E., and Hoppe, E. J. 共2002兲. “Inter-
Foundation Research at the Federal Highway Administration face shear tests on FRP composite piles.” Proc., Deep Foundations
共FHWA兲 for providing the pile driving hammer and Mr. Helge 2002, ASCE, Reston, Va., 1486–1500.
Prakash, S., and Sharma, H. D. 共1990兲. Pile foundations in engineering
Wittholz of Polymarin-Bolwll Composites Inc., Huron Park, On-
practice, Wiley, New York.
tario, for providing the FRP II pipes examined in this study. Sakr, M., El Naggar, M. H., and Nehdi, M. 共2003兲. “Novel toe driving for
pipe pile installation and performance of FRP pile segments.” Can.
Geotech. J., in press.
References Uesugi, M., and Kishida, H. 共1986兲. “Frictional resistance at yield be-
tween dry sand and mild steel.” Soils Found., 26共4兲, 139–149.
American Society for Testing and Materials 共ASTM兲. 共1998兲. “Standard Yoshimi, Y., and Kishida, T. 共1982兲. “A ring torsion apparatus for evalu-
test method for direct shear test of soils under consolidated drained ating friction between soil and metal surfaces.” Geotech. Test. J.,
conditions.” D 3080-98, West Conshohocken, Pa. 4共4兲, 145–152.

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005 / 283

J. Compos. Constr., 2005, 9(3): 274-283