This paper summarizes the construction details

and findings of laboratory and field tests of a new

generation of precast composite piles used for the
first time in the construction of the Route 40
highway bridge over the Nottoway River in
Virginia. The piles consisted of 24.6 in. (625 mm)
diameter concrete-filled glass fiber reinforced
polymer (GFRP) circular tubes, with a 0.21 in. (5.3
mm) wall thickness. The composite piles extended
above the ground level and were directly
embedded into the reinforced concrete cap beam
supporting the superstructure. Laboratory tests
included two full-scale composite piles loaded to
failure using four-point bending configuration.
Field testing included a full-scale precast
composite pile and a conventional 20 in. (508
mm) square concrete pile prestressed with
fourteen
1
/
2
in. (12.7 mm) diameter strands. This
paper presents details of the construction and
driving of the piles, comparisons between the
behavior of the composite and prestressed
concrete piles under axial and lateral loading, the
observed failure modes, and the details of the
connection between the piles and the reinforced
concrete cap beam.
P
iles used for bridge foundations are typically pro-
duced using traditional materials such as concrete,
steel, and timber. In recent years, high repair and re-
placement costs have led North American highway agen-
cies and researchers to investigate the feasibility of using
fiber-composite materials for transportation and civil engi-
neering infrastructure, including pile foundations for
Precast Piles for Route 40 Bridge
in Virginia Using Concrete Filled
FRP Tubes
2 PCI JOURNAL
Amir Fam, Ph.D.
Assistant Professor
Department of Civil Engineering
Queens University
Kingston, Ontario, Canada
Miguel Pando, Ph.D.
Assistant Professor
Department of Civil Engineering
University of Puerto Rico
Mayaguez Campus
Mayaguez, Puerto Rico
George Filz, Ph.D., P.E.
Associate Professor
Civil and Environmental Engineering
Department
Virginia Polytechnic Institute and
State University
Blacksburg, Virginia
Sami Rizkalla, Ph.D.
Distinguished Professor of Civil
Engineering and Construction
and
Director of the Constructed Facilities
Laboratory
Civil Engineering Department
North Carolina State University
Raleigh, North Carolina
bridges in corrosive environments.
1,2
For most projects, high quality con-
ventional precast, prestressed concrete
piles will continue to be the preferred
solution for deep bridge foundations;
but for piling exposed to high-corro-
sive environments, the precast FRP
composite piles described in this paper
may provide better durability.
A number of commercial products
known as composite piles have re-
cently become available for structural
applications. The term composite
piles refers to piles composed of fiber
reinforced polymers (FRP), recycled
plastics, or hybrid materials. This
paper focuses on precast concrete-
filled tubular FRP piles, where the
FRP tube serves as permanent
lightweight non-corrosive formwork
and a reinforcement element for con-
crete.
The FRP tube is composed of sev-
eral layers of fibers embedded in a
polymeric resin. The layers of fibers
are oriented in different directions
with respect to the longitudinal axis of
the tube in order to provide strength
and stiffness in both the axial and cir-
cumferential directions. The circum-
ferential stiffness and strength of the
FRP tube provides confinement to the
concrete core under axial stresses,
which increases the strength and duc-
tility of the pile, while the strength and
stiffness of the tube in the axial direc-
tion contributes to the flexural strength
of the pile.
During the relatively brief history of
composite pile use, applications have
been limited mainly to marine fender
piles, loadbearing piles for light struc-
May-June 2003 3
Fig. 2. The old Route 40 Bridge over the Nottoway River in Virginia.
Fig. 1. Location of
the Route 40 Bridge
(Structure No. 1006)
in Sussex County,
Virginia.
tures, and demonstration projects.
2
Composite piles have not yet gained
wide acceptance in practice primarily
because of the lack of a long track
record of performance and higher ini-
tial cost. However, appropriately de-
signed precast FRP composite piles
can result in longer service life, signif-
icantly improved durability in harsh
environments, and substantially re-
duced life-cycle costs.
The short-term structural behavior of
concrete-filled tubular FRP piles, under
axial and flexural loading, has been
studied by several researchers.
3-6
The
geotechnical behavior of these piles has
also been investigated.
7-9
While these
structural and geotechnical studies have
advanced the state of knowledge about
the behavior of composite piles, field
applications of composite piles in
bridges are limited. This paper de-
scribes the application of composite
piles in the new Route 40 Bridge in
Virginia, which is the first bridge
known to the authors to utilize compos-
ite piles in its foundation pier.
BACKGROUND AND
BRIDGE DESCRIPTION
The following sections provide a de-
tailed description of the old Route 40
Bridge in Virginia as well as the new
bridge. Details of the prestressed con-
crete piles as well as the new compos-
ite piles used in the new bridge are
also provided.
The Old Bridge
In 1998, the Virginia Department of
Transportation (VDOT) decided to re-
place the existing Route 40 Bridge
(Structure No. 1006) over the Not-
toway River in Sussex County, Vir-
4 PCI JOURNAL
ture supported by four piers and two
end abutments, as shown in Fig. 4.
Each pier consists of a reinforced con-
crete beam-type elevated pile cap sup-
ported by ten piles. Based on the de-
sign requirements, the VDOT
specified a 20 in. (508 mm) pre-
stressed solid square pile, as shown in
Fig. 5(a). The design axial load of
each pile was 150 kips (667 kN).
As a part of an ongoing study spon-
sored by the Federal Highway Admin-
istration (FHWA), the VDOT took the
initiative to use an alternative compos-
ite pile system, consisting of concrete-
filled FRP tubes, for the entire group of
piles supporting Pier No. 2 (see Fig. 4).
The objectives of the VDOT initia-
tive were to examine the feasibility of
using composite piles for bridge sub-
structures (particularly their drivability
using conventional pile drivers), per-
formance under axial and lateral load-
ing conditions using full-scale field
tests, and comparison to conventional
prestressed concrete piles. The follow-
ing sections provide a brief description
of both the conventional prestressed
piles and the precast composite piles.
Prestressed Concrete Piles
Details of the 20 in. (508 mm)
square prestressed concrete piles spec-
ified for the superstructure supports
are shown in Fig. 5(a). Each pile is
prestressed using a total of fourteen
1
/
2
in. (12.7 mm) diameter seven-wire
strands of 270 ksi (1861 MPa) ulti-
Fig. 4. Schematic of the four piers of the new Route 40 Bridge.
Vertical cracks in columns Spalling and cracking of
bearing seats
Map cracking in beams
Fig. 3. Deterioration of the substructure of the old Route 40 Bridge.
ginia, which was built in the early
1930s. Fig. 1 shows the location of the
bridge. The old bridge, which con-
sisted of a steel truss supported by
concrete piers, as shown in Fig. 2, suf-
fered from excessive deterioration in
both the superstructure and substruc-
ture.
Deterioration included full height
vertical cracks in the concrete
columns of the piers, spalling and
cracking of the concrete bearing seats,
and map cracking, up to
1
/
16
in. (1.6
mm) in width, in the abutment bearing
seats (see Fig. 3).
In addition to the deterioration of
the substructure, excessive corrosion
with large section losses was observed
in the roller bearing devices of the
truss supports. Additionally, the floor
beams suffered from section losses up
to
1
/
8
in. (3.2 mm) in both the web and
bottom flanges. The bridge was func-
tionally obsolete with a roadway
clearance of only 24 ft (7.3 m) curb-
to-curb and was rated as scour critical.
The New Bridge
The new bridge, which is 280 ft
(85.3 m) long and 50 ft (15.2 m) wide,
has a five-span slab-type superstruc-
May-June 2003 5
mate strength, pretensioned to produce
a prestress level of 0.809 ksi (5.6
MPa) in accordance with VDOT stan-
dards. The piles are 43 ft (13.1 m)
long. The specified concrete compres-
sive strength at 28 days is 5.8 ksi (40
MPa). The ties consist of No. 5 gauge
spiral wire.
Precast Composite Piles
The composite piles used in this
project consist of concrete-filled circu-
lar glass FRP (GFRP) tubes. The
GFRP tube is fabricated using the fila-
ment-winding technique, where E-
glass continuous fiber rovings are im-
pregnated with polyester resin and
wound over a rotating steel mandrel,
following a predetermined winding
pattern. The fiber volume fraction of
the tube is 51 percent. Fig. 5(b) pro-
vides details of the composite pile.
The GFRP tube has an outer diame-
ter of 24.6 in. (625 mm) and a total
wall thickness of 0.263 in. (6.68 mm).
The tube has a 0.05 in. (1.27 mm)
thick liner at the inner surface, which
results in a net structural wall thick-
ness of 0.213 in. (5.41 mm).
The wall structure of the tube con-
sists of three layers. The inner and
outer layers are each 0.074 in. (1.88
mm) thick, and they contain layers of
fibers oriented at 34 degrees with re-
spect to the longitudinal axis of the
tube. The middle layer, which is sand-
wiched between the inner and outer
layers is 0.065 in. (1.65 mm) thick and
contains fibers oriented at 85 degrees
with respect to the longitudinal axis.
The GFRP tubes are filled with a
concrete mix that includes an expan-
sive additive to reduce the effect of
shrinkage and therefore improve the
bond between the concrete and the
GFRP tube. It should be noted, how-
ever, that concrete shrinkage in closed
forms is significantly less than conven-
tional concrete members.
10
The speci-
fied compressive strength of the con-
crete fill at 28 days is 6 ksi (41.4 MPa).
The mechanical properties of the
GFRP tubes are provided in Table 1.
LABORATORY TESTING OF
THE COMPOSITE PILES
Prior to construction of the bridge, a
comprehensive experimental study was
Mechanical property Axial direction Hoop direction
Tensile strength (ksi) 31.6* 30.0

36.2

51.2

Compressive strength (ksi) 15.1

N/A
Elastic modulus (ksi) 2196* 2199

2404

2567

Poissons ratio 0.32

0.34

* Coupon test.
13
Manufacturer.
Lamination theory.
Note: 1 ksi = 6.89 MPa.
Table 1. Mechanical properties of composite GFRP tube.
Fig. 5. Details of piling: (a) Prestressed concrete pile; (b) Composite pile.
Fig. 6. Test setup for laboratory bending tests of the composite piles.
conducted to study the structural be-
havior of concrete-filled FRP tubes.
3-6
The behavior of the composite piles
was investigated using full-scale labo-
ratory testing, in which the flexural
stiffness, strength, and failure mode of
the composite piles were determined.
Test Setup and Instrumentation
Two full-scale specimens with
GFRP tubes identical to those pro-
posed for the composite piles of the
Route 40 Bridge were tested in bend-
ing under four-point loads. The tubes
(a) Prestressed concrete pile
(b) Composite pile
6 PCI JOURNAL
about 0.37 in. (9.4 mm). Both labora-
tory test specimens exhibited flexural
tension failure by rupture of the fibers
on the tension side, as shown in Fig.
8(a). Despite the brittle nature of rup-
ture of the FRP tube, the specimen re-
mained intact and was able to support
the dead weight after failure, as shown
in Fig. 6.
Fig. 8(b) shows the cross section of
the specimen at the failure location
after removal of the FRP shell. The
compression zone of the concrete core
could be distinguished from the
cracked zone of the section by exam-
ining the texture of the cross section.
The measured depth of the compres-
sion zone was about 6 in. (152 mm).
Material Testing
To validate the manufacturers data,
the mechanical properties of the com-
posite tubes were measured using
coupon tests and calculated using clas-
sical lamination theory. Strips, 1 in.
wide x 30 in. long (25 x 762 mm),
were cut from the tube in the longitu-
dinal direction.
Since the stress concentration at the
gripping location can severely influ-
ence the strength of FRP coupons, the
two ends of the GFRP strip were in-
serted inside hollow steel tubes, 1 ft
(305 mm) long each, and epoxy resin
was used to fill the steel tubes. The
clear length of the GFRP coupon be-
tween the two steel tubes was 6 in.
(152 mm).
After curing of the epoxy, the
coupons were tested in tension, where
the gripping was applied to the steel
tubes. Fig. 9 shows the coupons after
tension failure and a typical stress-
strain curve for a GFRP coupon com-
pared to the stress-strain curve ob-
tained from classical lamination
theory
11
and the manufacturers data.
Table 1 provides a summary of the me-
chanical properties of the GFRP tube.
Comparison Between Behavior of
Composite and Prestressed
Concrete Piles
The moment-curvature behavior of
the prestressed concrete piles was pre-
dicted analytically using the concept
of equilibrium and strain compatibil-
ity. A modified Ramberg-Osgood
(a) Tension failure by rupture of FRP tube (b) Cross section at failure location
Tension (cracked)
zone
Compression zone
6 in.
Fig. 8. Flexural tension failure of the composite pile.
Fig. 7. Moment-curvature responses of both the composite and prestressed concrete piles.
were filled with 4.8 ksi (33 MPa) con-
crete at Lafarge Canada, in Winnipeg,
Manitoba, using the same procedure
specified for casting the composite
piles for the bridge. The span of the
test specimen was 16.4 ft (5.0 m), and
the distance between the two applied
loads was 4.9 ft (1.5 m).
Fig. 6 shows the test setup of the
specimens in bending. The tests were
conducted using stroke control with a
rate of loading of 0.05 in. per minute
(1.3 mm/min). The specimens were
instrumented to measure the midspan
deflection, the extreme fiber strains at
the tension and compression sides
within the constant moment zone, and
the relative slip between the concrete
core and FRP tube at the ends.
Test Results and Failure Mode
The measured moment-curvature
behavior of the composite piles for the
two identical specimens is shown in
Fig. 7. The average ultimate moment
capacity of the composite piles is 370
kip-ft (502 kN-m). The cracking mo-
ment, 93 kip-ft (126.1 kN-m), is rela-
tively small compared to the ultimate
moment. Since the composite piles
were tested under stroke control, the
behavior shows vertical drops in the
moment when cracks occurred, and
the behavior after first cracking was
almost linear.
At the end of the test, the total slip
measured between the concrete core
and the FRP tube at each end was
May-June 2003 7
function was used to model the stress-
strain relationship of the prestressing
steel strands. A generalized expres-
sion, given by Popovics, was used to
model the nonlinear stress-strain rela-
tionship of concrete.
12
The moment-
curvature response based on the analy-
sis of the prestressed concrete pile is
shown in Fig. 7. The response was ter-
minated when the compressive strain
in the concrete reached 0.003.
For the concrete-filled GFRP tube,
the moment-curvature behavior was
also predicted using an equilibrium
and strain compatibility approach. The
cross section was divided into hori-
zontal layers to account for the circu-
lar geometry of the concrete and the
GFRP tube, which forms continuous
reinforcement around the concrete. A
linear stress-strain relationship of the
GFRP composite tube, based on the
linear regression of the experimental
results shown in Fig. 9, was used. The
forces in both the concrete and the
FRP tube were obtained by numerical
integration.
The forces were then used to calcu-
late the internal moment resistance at
different strain levels, which were
used to generate the predicted mo-
ment-curvature response. Details of
the analytical model can be found
elsewhere.
13
Fig. 7 shows that the pre-
dicted moment-curvature response of
the composite pile agrees well with
the measured response.
Fig. 7 also shows that the flexural
stiffness of the composite pile before
cracking is very similar to that of the
prestressed concrete pile. However,
after cracking, the stiffness of the
composite pile is significantly re-
duced. This is attributed to the pre-
stressing effect and the lower elastic
modulus of GFRP in comparison to
that of steel strands.
The cracking moment of the com-
posite piles is 40 percent lower than
the prestressed concrete pile due to the
absence of prestressing. However,
cracking of composite piles should not
cause concerns regarding durability,
since the internal steel reinforcement
has been eliminated from this system.
The maximum moment capacities of
the two piles are very similar, despite
the differences in geometry, size, and
materials.
(a) Tension
failure
(b) Stress-strain curve in the axial direction
Fig. 9. Failure mode and behavior of longitudinal tension coupons tests of the GFRP tube.
Fig. 10.
Fabrication
and handling
of precast
composite
piles.
(a) Pumping concrete into the upper end
(b) Rear end showing wooden plugs
(c) Handling of the pile using eight-point supports
8 PCI JOURNAL
Fig. 13(a). Driving of composite test pile. Fig. 13(b). Driving of prestressed test pile.
Prestressed
pile
Composite
pile
FIELD TESTING OF
COMPOSITE AND
PRESTRESSED PILES
Prior to installation of the produc-
tion piles for the new Route 40
Bridge, a test pile program was com-
pleted at the bridge site. Axial and lat-
eral loading tests were performed on
two full-scale, 43 ft (13.1 m) long
piles, one a composite pile and the
other a prestressed concrete pile.
Based on the subsurface conditions at
the bridge site, it was anticipated that
the axial load capacity of the pile
would be governed by the geotechni-
cal capacity, i.e., the soil support ca-
pacity rather than the axial structural
capacity of the pile.
Fabrication of the Piles
The GFRP tubes were shipped to the
precast plant and were filled with con-
crete while they were placed in an in-
clined position and supported by a
steel beam along the full length of the
tubes, as shown in Fig. 10. Wooden
Fig. 11. Fabrication of prestressed concrete pile. Fig. 12. Instrumentation of test piles.
The predicted axial load capacities
of the prestressed and composite piles
are 2071 and 2812 kips (9212 and
12508 kN), respectively. These values
are based on the specified concrete
compressive strengths of the pre-
stressed and composite piles, respec-
tively, as well as the properties of the
steel strands and the GFRP tube.
May-June 2003 9
end plugs were used to seal both ends
of the tube. The plugs were secured in
position using metal straps connected
to both the plug and the composite
tube in radial directions, as shown in
Fig. 10(b).
Concrete was pumped into the upper
end through a hole in the wooden
plug. The concrete-filled composite
tubes were handled using eight-point
lifting devices along the length of the
pile, as shown in Fig. 10(c). The pre-
stressed concrete piles were fabricated
at the precast plant in accordance with
VDOT standards, as shown in Fig. 11.
After fabrication and curing, the piles
were shipped to the bridge site.
Instrumentation
Both the composite and the pre-
stressed concrete test piles were in-
strumented with strain gauges in the
axial direction at three levels along the
length of the piles. At each level, two
strain gauges were placed near the two
opposite faces of the pile, as shown in
Fig. 12. The strain gauges were
mounted on 36 in. (914 mm) long No.
4 steel reinforcing bars, which were
embedded inside the concrete core.
Each test pile was also instrumented
with eight lateral motion sensors to
measure the lateral displacement pro-
file of the pile during the lateral load
tests. The sensors were placed inside
PVC tubes embedded inside the test
piles. The locations of the lateral mo-
tion sensors are shown in Fig. 12.
To measure the axial stresses during
driving, pairs of special strain gauges
and accelerometers were installed on
the concrete surface of the piles at a
distance of 4 ft (1.2 m) from the top.
For the composite pile, this was
achieved by cutting 5 x 5 in. (127 x
127 mm) portions of the FRP shell in
order to attach the instrumentation to
the concrete surface.
Pile Driving
Both the composite and prestressed
test piles were driven using a hy-
draulic impact hammer, Type ICE
Model 160S, which was also used to
drive the production piles for the
bridge. The ram weight was 16,000
lbs (71.2 kN). A 7.5 in. (190 mm)
thick plywood pile cushion was used
during pile driving. Both of the test
piles were driven to a depth of about
33.5 ft (10.2 m). The blow count was
six blows per 1 in. (25.4 mm) of pile
penetration for the composite pile and
four blows per 1 in. (25.4 mm) for the
prestressed concrete pile.
Figs. 13(a) and 13(b) show the com-
posite pile during and after driving, as
well as the prestressed pile during
driving. Table 2 provides some of the
measurements obtained during pile
driving. It can be seen that the wave
speeds, maximum compressive
stresses, and maximum tensile stresses
are each similar in the different pile
types. The measured stress levels are
lower than the allowable stresses rec-
ommended for prestressed piles.
14
No
standards are available yet for driving
precast composite piles.
Axial Load Tests
The composite and prestressed piles
were both subjected to axial loads
using the Statnamic Testing System,
as shown in Fig. 14. This system ap-
plies the load to the pile by launching
a heavy reaction mass upwards at
close to 20g (twenty times the gravity
acceleration). When the reaction mass
is launched upwards, an equal and op-
posite reaction force acts on the test
Pile type
Measurement Prestressed Composite
Wave speed 12,150 ft/s 11,840 ft/s
Maximum compressive stress
2.55 ksi 2.78 ksi
measured during driving
Maximum tensile stresses
0.7 ksi 0.42 ksi
measured during driving
Allowable stresses
Tension < 1.02 ksi No standards
Compression < 4.5 ksi available
Note: 1 ft = 0.305 m; 1 ksi = 6.89 MPa.
Table 2. Pile driving measurements for prestressed and composite piles.
Fig. 14. Axial load
test using Statnamic
test setup.
Load cell
Pressure
chamber
Laser beam device
Composite pile
Reaction
mass
10 PCI JOURNAL
pile. The Statnamic load does not have
the impact nature of the hammer blow,
but rather a more gradually increasing
force with a typical duration of 0.2 to
0.3 seconds.
During the Statnamic test, several
measurements are taken, including the
applied load, the pile head displace-
ment using a laser beam device shown
in Fig. 14, and strains. A detailed de-
scription of the Statnamic technique
can be found in Brown.
15
Each test
pile was subjected to three loading cy-
cles with increasing the magnitude of
the applied load. The equivalent axial
static load is determined by subtract-
ing the inertia and damping forces
from the total measured force using
the Unloading Point Method (UPM).
16
The equivalent static load versus
pile head axial displacement for the
prestressed concrete and composite
piles is shown in Fig. 15 for the three
cycles. The behavior during the last
cycle in both piles shows that the
geotechnical capacity of the piles was
fully mobilized, as evident from the
rapid increase of the pile displacement
near the end of the test. The ultimate
equivalent static loads of the pre-
stressed concrete and composite piles
at failure were 942 and 980 kips (4190
and 4359 kN), respectively. Both piles
exhibited similar axial capacities, even
though their shaft and end areas were
slightly different.
Fig. 16 shows the equivalent axial
static load versus axial strain behavior
of both the composite and prestressed
piles based on the peak load and corre-
sponding strain during each of the
three load cycles. The axial strain is
based on the average strain of the two
uppermost strain gauges (closest to the
loading end), since they measured the
maximum axial strain induced in the
pile.
Fig. 16 shows that the behavior of
the two piles is relatively similar, with
the composite pile being slightly stiffer
than the prestressed pile. The figure
also shows the design load of the piles,
150 kips (667 kN), which is signifi-
cantly lower than the ultimate load.
The variation of the axial strains
along the length of both the pre-
stressed and composite piles is shown
in Fig. 17. It can be seen that the mea-
sured strains along the depth are simi-
Fig. 16. Axial load versus axial strain behavior of test piles.
Fig. 17. Variation of axial strain levels along depth of pile.
Fig. 15. Variation of pile head displacement with equivalent axial static load.
May-June 2003 11
lar in both piles. The figure also shows
that the strain level is reduced as the
depth is increased. This is attributed to
the gradual transfer of axial load from
the pile to the soil through skin fric-
tion along the pile shaft.
Lateral Load Tests
Following the axial load tests, lat-
eral load tests were conducted on both
the composite and prestressed piles,
using the Statnamic Testing System
shown in Fig. 18. The loading system
is similar to that used for the axial
load test; however, the setup was
placed horizontally in order to produce
a lateral load. During the Statnamic
test, several measurements were taken,
including the applied load and the lat-
eral displacement of the pile at differ-
ent depths. Each test pile was sub-
jected to four loading cycles by
increasing the magnitude of load.
The equivalent lateral static load ver-
sus the lateral deflection behavior of
the two piles at the loading points is
shown in Fig. 19. The figure shows
that the behavior has a similar moment-
curvature response as shown in Fig. 7,
for both the prestressed concrete and
composite piles. For the composite
pile, the sudden change in stiffness at a
lateral load of 11 kips (48.9 kN) re-
flects the cracking point. For the pre-
stressed pile, the initial stiffness is high
due to the prestressing effect and the
higher elastic modulus of steel.
Composite pile
Reaction
mass
Pressure
chamber
Fig. 18. Lateral load test using Statnamic test setup.
Fig. 20. Lateral deflection distribution along length of pile.
Fig. 19. Lateral load versus displacement response of piles using Statnamic test.
12 PCI JOURNAL
Fig. 20 shows the lateral deflection
distribution along the length of the
pile at the peak load for each of the
four loading cycles, for both the com-
posite and prestressed concrete piles.
The figure indicates that lateral deflec-
tions become insignificant at a certain
depth between 15 and 20 ft (4.5 and
6.1 m) measured from the loading
point. The figure also indicates that,
for the last two load cycles, the deflec-
tion profile along the depth of the
piles is almost bilinear, with a sudden
change in slope occurring at a depth of
about 16 ft (4.8 m), measured from the
loading point. This suggests that, dur-
ing the last two cycles, failure may
have been initiated in both piles at the
location of the change in slope.
CONSTRUCTION OF THE
NEW ROUTE 40 BRIDGE
Based on the results of the field and
laboratory testing of the composite
and prestressed concrete piles, the
Fig. 21. Precast composite piles for Pier No. 2 after driving.
Fig. 22. Details of
pile head
showing bars
used to connect
pile to cap beam.
VDOT decided to use the composite
piles in the construction of Pier No. 2
(see Fig. 4) of the new Route 40
Bridge. The pier consists of a rein-
forced concrete cap beam supported
by ten composite piles.
The composite production piles are
identical to the composite test pile in
terms of size, laminate structure of the
tube, and the methods of construction
and driving. Fig. 10 shows the com-
posite piles during casting and han-
dling, while Fig. 13 shows the pile
driving process. The installed compos-
ite piles of Pier No. 2 are shown in
Fig. 21.
After all the piles were driven, and
prior to casting the cap beams, special
preparation of the pile heads was nec-
essary to facilitate connecting the piles
to the cap beams. Eight, 1 in. (25.4
mm) diameter holes were drilled
through the top flat surface of each
pile, using a regular rock drill. The
holes were 18 in. (457 mm) deep, par-
allel to the longitudinal direction of
the piles.
In the prestressed concrete piles, the
holes were equally spaced in a 13 in.
(330 mm) square pattern, whereas in
the composite pile, the holes were
equally spaced in a 17.6 in. (447 mm)
diameter circular pattern as shown in
Fig. 22. These arrangements allow for
a 3 in. (76 mm) concrete cover for
each hole. Eight, 48 in. (1219 mm)
long, No. 7 steel reinforcing bars were
inserted in the holes, and epoxy resin
was used to secure the bars inside the
concrete. Fig. 22 shows the details of
the pile head and the steel dowels in
both pile types.
The formwork arrangement of the
cap beams were placed such that the
bottom surface of the cap beam is 6 in.
(152 mm) below the upper surfaces of
the piles to allow for embedment of
the piles inside the cap beam. Fig. 23
shows details of Pier No. 2. Fig. 23(b)
shows the No. 7 steel bars, which are
embedded 30 in. (762 mm) inside the
cap beam.
Fig. 24 shows close-up views of Pier
No. 2 and the connection between the
composite pile and cap beam. A gen-
eral view of the new completed bridge
is shown in Fig. 25. The bridge has
been in operation for over two years.
May-June 2003 13
Fig. 23. Connection of composite piles to cap beam at Pier No. 2.
Fig. 24. Pier No. 2 showing composite piles and reinforced concrete cap beam.
Reinforced concrete
cap beam
Cost Comparison
The cost of the composite piles in-
cluding the costs of the GFRP tube,
concrete, casting, and shipping was
$95 per ft. The cost of driving the
composite piles was $20 per ft. This
resulted in a total cost of $115 per ft
for the installed composite pile, in
comparison to a total cost of $65 per ft
for the installed square prestressed
concrete pile. Therefore, the initial
cost comparison indicates about 77
percent higher unit cost for the com-
posite pile. Part of this higher cost
may be due to the fact that composite
piles are not currently produced on a
large-scale basis. Further, current pro-
duction of composite piles is mostly in
Precast composite
piles
Fig. 25. The new Route 40 Bridge over the Nottoway River in Virginia.
14 PCI JOURNAL
smaller sizes for marine fendering ap-
plications.
The higher costs may also be associ-
ated with a lack of contractor experi-
ence in handling and lifting composite
piles, which require special arrange-
ments and equipment. Nevertheless,
precast composite piles may gain
wider acceptance in the construction
industry when life-cycle costs are con-
sidered because annualized mainte-
nance and replacement costs for the
non-corroding composite piles are ex-
pected to be less than for conventional
prestressed concrete piles in corrosive
environments.
CONCLUSIONS
Based on the testing and construc-
tion experience gathered so far, the
following conclusions can be drawn:
1. The use of concrete-filled FRP
tubes as piling for bridge piers is prac-
tical and feasible. The FRP tube
serves as permanent formwork and re-
inforcement at the same time, elimi-
nating the need for internal reinforce-
ment and temporary formwork.
2. The flexural strength of the 24.6
in. (625 mm) diameter composite pile
with a 0.213 in. (5.4 mm) thick GFRP
tube is similar to the flexural strength
of the 20 in. (508 mm) square concrete
pile prestressed with fourteen, 0.5 in.
(12.7 mm) diameter strands. However,
the stiffness of the composite pile after
first cracking was lower.
3. The composite pile failed in
bending by fracture of the GFRP tube
on the tension side, whereas the pre-
stressed concrete pile failed by yield-
ing of the strands in tension, followed
by crushing of the concrete in com-
pression.
4. Both the composite and pre-
stressed concrete piles performed sim-
ilarly during pile driving, as evident
from the measured wave speed gener-
ated by the driving hammer and the
measured compressive and tensile
strains.
5. Both the composite and pre-
stressed concrete piles performed sim-
ilarly under the axial load tests. Full
geotechnical capacity was mobilized
in both cases before structural failure
of the piles. The axial load at failure
was significantly higher than the de-
sign pile load.
6. The lateral load field tests on both
the composite and prestressed piles
showed a similar behavior to that ob-
tained from the laboratory flexural test
and analysis.
7. Similar pile-to-cap beam connec-
tions were used for the composite and
prestressed concrete piles, including
eight No. 7 steel dowels embedded 18
in. (457 mm) inside the piles from one
end and extending 30 in. (762 mm)
into the cap beam. The piles them-
selves were embedded 6 in. (152 mm)
inside the cap beam.
8. An initial cost comparison
showed a unit cost for the installed
composite piles 77 percent higher than
for the installed prestressed piles.
However, as production volume in-
creases, and by considering life-cycle
costs of the low-maintenance compos-
ite piles, the cost comparison may
shift in favor of composite piles in
corrosive environments.
9. To date, no indications of unsatis-
factory performance of the composite
piles of the Route 40 Bridge have
been reported.
Further research is still needed to
examine the following aspects of com-
posite piles:
1. Behavior under combined load-
ing.
2. Methods of improving the bond
between the tube and the concrete
core.
3. Methods for increasing the flexu-
ral stiffness, such as prestressing.
4. Shear behavior.
5. Behavior under cyclic loads.
Additional case histories and long-
term performance data are needed. De-
tailed life-cycle cost analyses are also
needed to determine the cost effective-
ness of these piles.
The VDOT and FHWA are cur-
rently engaged in another composite
pile research project associated with
the new Route 351 Bridge over the
Hampton River in Virginia. This pro-
ject is a continuation of the composite
pile research of VDOT and sponsored
by FHWA to contribute to the scarce
data available on projects with com-
posite piles. This project is different
from the Route 40 Bridge in the
method used for the field test, where
static load tests are performed using
conventional hydraulic jacks instead
of the Statnamic tests.
ACKNOWLEDGMENT
The authors want to acknowledge
the Federal Highway Administration
(FHWA), the Virginia Department of
Transportation (VDOT), the Virginia
Transportation Research Council
(VTRC), and the Network of Centers
of Excellence, ISIS Canada Program,
for supporting and sponsoring the lab-
oratory and field tests.
The authors would also like to ac-
knowledge Lancaster Composite for
providing the composite tubes and
providing valuable guidance during
the construction of the composite
piles.
The authors are also grateful to the
PCI JOURNAL reviewers for their
constructive comments.
May-June 2003 15
1. Lampo, R., Federal Interest Gives Recycled Plastic Lumber a
Leg Up, ASTM Standardization News, 1996, pp. 26-31.
2. Iskander, M., and Hassan, M., State of the Practice Review in
FRP Composite Piling, Journal of Composites for Construc-
tion, American Society of Civil Engineers, V. 2, No. 3, 1998,
pp. 116-120.
3. Fam, A. Z., Flisak, B., and Rizkalla, S., Experimental and
Analytical Investigations of Beam-Column Behavior of Con-
crete-Filled FRP Tubes, ACI Structural Journal, 2002.
4. Fam, A. Z., and Rizkalla, S. H., Confinement Model for Axi-
ally Loaded Concrete Confined by FRP Tubes, ACI Struc-
tural Journal, V. 98, No. 4, July-August 2001, pp. 251-461.
5. Fam, A. Z., and Rizkalla, S. H., Behavior of Axially Loaded
Concrete-Filled Circular Fiber Reinforced Polymer Tubes,
ACI Structural Journal, V. 98, No. 3, May-June 2001, pp. 280-
289.
6. Fam, A. Z., and Rizkalla, S. H., Flexural Behavior of Con-
crete-Filled Fiber-Reinforced Polymer Circular Tubes, Jour-
nal of Composites for Construction, American Society of Civil
Engineers, V. 6, No. 2, May 2002, pp. 123-132.
7. Pando, M., Filz, G., Hoppe, E., Ealy, C., and Muchard, M.,
Performance of a Composite Pile in a Full Scale Statnamic
Load Testing Program, Proceedings, 53rd Canadian Geotech-
nical Conference, V. 1, Montreal, Canada, 2000, pp. 909-916.
8. Pando, M., Filz, G., Dove, J., and Hoppe, E., Interface Shear
Tests on FRP Composite Piles, Proceedings, International
Deep Foundations Congress, American Society of Civil Engi-
neers, Orlando, Florida, February 2002, pp. 1486-1500.
9. Mirmiran, A., Shao, Y., and Shahawy, M., Analysis and Field
Tests on the Performance of Composite Tubes Under Pile
Driving Impact, Composite Structures, V. 55, No. 2, 2002,
pp. 127-135.
10. Naguib, W., and Mirmiran, A., Time-Dependent Behavior of
Fiber-Reinforced Polymer-Confined Concrete Columns Under
Axial Loads, ACI Structural Journal, V. 99, No. 2, March-
April, 2002, pp. 142-148.
11. Daniel, I. M., and Ishai, O., Engineering Mechanics of Com-
posite Materials, Oxford University Press, New York, NY,
1994.
12. Collins, M. P., and Mitchell, D., Prestressed Concrete Struc-
tures, Response Publications, Toronto, Canada, 1997.
13. Fam, A. Z., Concrete-Filled Fiber Reinforced Polymer Tubes
for Axial and Flexural Structural Members, Ph.D. Thesis, The
University of Manitoba, Winnipeg, Manitoba, Canada, 2000,
261 pp.
14. PCI Committee on Prestressed Concrete Piling, Recom-
mended Practice for Design, Manufacture, and Installation of
Prestressed Concrete Piling, PCI JOURNAL, V. 38, No. 2,
March-April 1993, pp. 14-41.
15. Brown, D. A., Evaluation of Static Capacity of Deep Founda-
tions From Statnamic Testing, Geotechnical Testing Journal,
American Society for Testing and Materials, V. 17, No. 4,
April 1994, pp. 403-414.
16. Middendrop, P., Berminghammer, P., and Kuiper, B., Stat-
namic Load Testing of Foundation Piles, Proceedings, Fourth
International Conference of Applications of Stress-Wave The-
ory to Piles, The Hague, Netherlands, 1992, pp. 581-588.
REFERENCES