4e Confrence spcialise en gnie des structures

de la Socit canadienne de gnie civil

4th Strutural Specialty Conference
of the Canadian Society for Civil Engineering
Montral, Qubec, Canada
5-8 juin 2002 / June 5-8, 2002

INVESTIGATION OF PRESTRESSED CONCRETE GIRDERS

STRENGTHENED WITH EXTERNALLY BONDED CARBON FIBRE
SHEETS
A. HarraqA, K. W. NealeA and P. PaultreA
A Department of Civil Engineering, Universit de Sherbrooke, Quebec, Canada
ABSTRACT: Prestressed concrete beams may be under-strength because of deficiencies in
design, increases in applied loads, loss of prestress, damage due to the effect of collision,
corrosion or military operations and salt damage. There has been limited research on the repair of
PC bridge girders. Carbon laminates can solve a wide variety of structural problems due to
environmental exposure, inadequate design and heavy traffic. Because they are impermeable, the
laminates resist moisture and the resulting corrosion. An added advantage of carbon laminates is
their ease of installation due to their extremely light weight compared to steel or concrete. This
paper describes the strength enhancement provided to the flexural capacity of prestressed
concrete girders by the external bonding of CFRP laminates through a large scale beam study.
Seven half scale standard CPCI 900 mm pretensioned bridge girders comprising two nonreinforced and five reinforced by CFRP were considered in this study.
1.

INTRODUCTION

During the last two decades engineers responsible for the maintenance of highway bridges have focused
their attention on the deteriorating condition of existing bridges. Thousands of bridges across Canada built
prior to and after the second world war are approaching the end of their design lives. These structures
which are usually designed for a 50-year life are also facing in many cases increasing vehicular wheel
load and traffic volume. Thus a large number of bridges must be either replaced or extensively
rehabilitated to carry loads which exceed those for which they were designed. Salt damage is also a
serious problem for concrete structures containing steel as a reinforcing material, since it may sometimes
lead to corrosive expansion of the steel which may in turn cause the concrete cross section to break and
fall-off. When the salt damage is so considerable as to cause partial corrosion and damage of PC steel
wires, a new method that can provide both salt damage prevention and reinforcing effect is required.
The performance of beams and girders depends greatly on the quantity and condition of the tensile
reinforcement in these members. Additional tensile reinforcement may be required to replace corroded or
damaged steel reinforcement to restore the strength and stiffness of the member. Several rehabilitation
methods have been developed which strengthen flexural concrete members by adding tensile
reinforcement. A method that has recently been gathering attention to counter those situations is to use
carbon fibre, which has superior properties such as light weight, high strength and high corrosion
resistance as reinforcing materials for existing concrete structures.
1

This paper presents the strengthening effects of externally bonded carbon fibre sheets on PC girders.
The objectives of this research are to develop design and analysis guidelines for strengthened systems
and to assess and demonstrate the strength enhancement achievable using FRP materials.
2.

ANALYTICAL STUDY

2.1

Determination of the nominal moment

A comprehensive design approach for prestressed concrete flexural beams strengthened with
externally bonded fibre-reinforced-polymer (FRP) plates is presented herein. This approach
complies with the Canadian Concrete Standard. The models presented are based on the common
principles of compatibility of deformations and equilibrium of forces. They can be used to predict
the ultimate strength in flexure that can be achieved by such elements, given the FRP crosssectional area, or the required FRP cross-sectional area to achieve the desired resisting moment
for rehabilitated flexural elements.
The following assumptions are considered:
1. A perfect bond between the FRP and the bonded substrate
2. Plane sections remain plane
The general equation for the nominal moment capacity of a prestressed concrete section
strengthened with FRP flexural reinforcement is given by:
a
a
a
a
M n = As ' f s ' ( d s ' ) + f s As (d ) + A ps f ps (d p ) + A frp f frp (h )
2
2
2
2

Where:

As : Area of the tension steel (mm2)

As' : Area of the compression steel (mm2)

A frp : Total area of the fiber contained in the FRP laminate (mm2)
f s : Stress in the tension steel (MPa)
f s' : Stress in the compression steel (MPa)
f ps : Stress in prestressing tendon (MPa)
f frp : Stress in the FRP laminate (MPa)
An approach estimating the depth to the neutral axis is required to determine the stress levels in
each of the materials. The estimate on the neutral axis depth must be checked by finding the strain
and stress levels in all the materials and substituting them into the following equation:

c=

A ps f ps + As f s As' f s' + A frp f frpu

1 f c' 1b

f ps is given by : f ps = f pu 1

f pu
d
'
p ' +
d
1
fc
p

where:

f pu : Ultimate strength of prestressing tendon (MPa)

p:

Factor accounting for the shape of stress-strain relationship of the prestressing steel

( p = 0.28 for
low relaxation strand)
p : Ratio of prestressed reinforcement ( = A ps bd p )

d : Distance from extreme compression fiber to centroid of non-prestressed reinforcement (mm)

d p : Distance from extreme compression fiber to centroid of prestressed reinforcement (mm)
=

fy

' = '

f c'

fy

, where

f c'

, where

is the ratio of non-prestressed reinforcement ( = As bd )

' is the ratio of compression reinforcement ( '

= As bd )

If failure is governed by concrete crushing, the following equations must be used to determine the
strain levels in the FRP and the reinforcing steel:

c = cu , s = c

s '= c

(c d s ')
c

(d s c )
c

, frp = c

(h c )
c

If failure is governed by FRP rupture, the following equations must be used:

frp = c

(h c ) ,
c

hc

c = frp

d c
c ds '
, s ' = frp

hc

s = frp
Where :

c : Maximum compressive strain in the concrete

cu : Maximum compressive strain in the concrete at ultimate
s : Strain in the tension steel (MPa)
s' : Strain in the compression steel (MPa)
frp : Strain in the FRP laminate (MPa)
h : Total height of the section and depth to the flexural reinforcement (mm)
c : Depth of the neutral axis (mm)

d s : Depth to the tension steel reinforcement centroid (mm)

d s' : Depth to the compression steel reinforcement centroid (mm)
d p : Depth to the prestressing steel centroid (mm)
2.2

Moment curvature response

The determination of the moment - curvature relation of the section of a beam permits us to
understand its behaviour during all phases of loading, to identify the critical points where this
behaviour changes. At a section of the beam subjected to an external moment, the member is
divided into a number of layers. An initial linear distribution of strain is assumed. Using the
uniaxial stress-strain relation, the normal stress distribution associated with the average strain in
a layer is calculated. The forces and the moments caused by the stresses are integrated over the
height of the section to determine the moment acting on the beam. The calculations are repeated
until the entire moment-curvature response is determined. This procedure has been implemented
in the WMNPhi software. This software is used to evaluate the moment-curvature response of the
prestressed girders strengthened with CFRP sheets.
2.3

Materials behaviour

The stress-strain curves of the materials modelled with the WMNPhi software are shown in Figure
1. The stress-strain curve for the concrete is considered as a parabolic function. A perfectly elastic
behaviour is assumed for the composite material, the stress-strain relation of the reinforcing
steel is assumed to be elastoplastic whereas the stress-strain response of prestressing strand is
represented by the Ramberg-Osgood function.
4.

EXPERIMENTAL PROGRAM

This section describes the ongoing experimental program for investigating the behaviour and the FRP
strengthening in bending of prestressed concrete I - girders. The experimental program is conducted at
the Universit de Sherbrooke. The objective of the experimental investigation is to examine the
effectiveness and feasibility of using CFRP sheets to increase the flexural capacity, and simultaneously
improve the serviceability of a structurally damaged concrete bridge girder model. The properties of the
materials, the test set-up and the instrumentation are presented.
4.1

Specimens

The pretensioned concrete specimens are half scale configurations standard CPCI 900 mm deep bridge
girders. The length of the specimens are 4.4 m with a depth of 530 mm. The typical geometry with the
other dimensions are shown in Figure 2. The girders are pretensioned with 9.53 mm diameter steel strand
to an initial stress of 1200 MPa. The strands are released at a concrete age of 7 days. Steel stirrups
spaced at 200 mm are used as shear reinforcement. Seven girders are considered in this experimental
program; two of them are not strengthened and served as control beams and five strengthened with
CFRP. The details of the reinforcement are given in the Figures 3 to 5.
Three FRP strengthening systems are used in this research project. These systems consist of two types
of CFRP sheets (Systems R and C) and one type of GFRP sheets (System T). The carbon fibre sheets
are used for the flexural strengthening and the GFRP sheets (U-strips) are used as anchors to provide
mechanical connection between the CFRP and the concrete to avoid delamination. The features of these
specimens are itemized in Table1.
4.2

Materials used

In the following experimental program three different bonding composite materials have been used,
namely carbon plates, carbon sheets and glass sheets with the mechanical properties as reported in
Table 2. All the products have been fabricated with unidirectional carbon or glass pattern. The properties
of concrete are given in Table 3, and the properties of the reinforcing bars are summarized in Table 4.

4.3

Instrumentation

Displacements along the length of the beams is monitored by 6 linear voltage displacement transducers
(LVDTs). Strains in the beam are monitored by electrical strain gauges mounted on reinforcement in the
beams. Strain gauges are placed on each stirrup in the shear span and on the tensile reinforcement at
mid-span and the quarter span, a strain gauge is placed on the compression reinforcement at the midspan and two strain gauges are placed on the prestressing strand at mid-span and at the quarter span.
Furthermore strain gauges are placed to the surface of the concrete and to the surface of CFRP sheets. In
addition, demec points are glued on four rows along the depth of the beam mid-span to measure crack
widths. The details of the instrumentation are presented in Figure 6.
4.4

Test set-up

All the beams are subjected to bending moment. The beams are subjected to a static monotonic load, and
are tested until the ultimate strength state is attained, and their behaviour is monitored.
5.

ANALYTICAL RESULTS AND DISCUSSION

A 25 % loss in load carrying capacity is assumed due to corrosion related problems. Two control beams
were investigated; PCB1 prestressed with four 9.53 mm strands before losses and PCB2 prestressed with
only two 9.53 mm strands after losses. To compensate for these losses, beam PCB2 was strengthened
using different types and configurations of CFRP laminates. Figure 8 shows the predicted momentcurvature relationships for the beams considered. The application of externally bonded CFRP laminates
increased the capacity of beams PCB2R2 ( 2 layers of Replark), PCB2R3 ( 3 layers of Replark), and
PCB2C1 (Carbodur) by approximately 20 %, 30 % and 50 %, respectively, compared to that of the control
beam PCB2. It should be noted that all strengthened beams failed by rupture of FRP laminates. This may
be expected due to the section of the prestressed beam which is under-reinforced.
6.

CONCLUSION

When assessing the performance of concrete structures such as bridges, there are comparatively few
cases of deterioration due to acting loads; there are more cases of corrosion of reinforcing bars and
prestressing steels. Considering the excellent corrosion resistance properties of FRP products, it may be
expected that FRP products will see increased use in concrete structures. The effects of reinforcement by
bonding carbon fibre sheets is satisfactory, and the reinforcement effect can be obtained by the
conventional RC calculation method. In addition, durability can be improved by increasing the quantity of
carbon fibre sheets for reinforcement. By bonding the CFRP sheets we are able to compensate for the
loss of prestressing.
7.

FUTURE WORK

The experimental results will be compared to the theoretical analyses. Also the testing program will be
followed by numerical simulations (Finite Element Analysis) aiming to reproduce the real behaviour of the
tested beams. The model will account for the non-linear behaviour of the concrete material and the
interface between concrete and composite will be modelled.

8.

AKNOWLEDGEMENTS

This research project was supported by the Natural Sciences and Engineering Research Council of
Canada (NSERC) through the Network of Centres of Excellence (NCE) ISIS Canada. The technical
assistance of Claude Aub and Marc Demers was essential in the completion of the laboratory set-up.
9.

REFERENCES

Collins, M.P. and Mitchell, D. (1997) Prestressed Concrete Structures, Response Publications, Canada.
Manfroni, O., Di Tommaso, A. and Bergmeister, K. (2000) Full Scale Bending Tests up to Collapse of PC
Beams Strengthened with Bonding FRP, Advanced Composite Materials in Bridges and Structures,
Editors., Humar. J., Razaqpur, A.G., Ottawa, Ontario, Canada.
Nanni, A., Huang, P.C. and Tumialan, J.G. (2001) Strengthening of Impact-Damaged Bridge Girder Using
FRP Laminates, Structural Faults and Repairs, London, UK,
Picard, A. (2001) Bton Prcontraint, 2 dition, Gatan Morin, Qubec, Canada.
Sakamoto et al., (1995). A test on PC Girders Reinforced by using Carbon Fibre Sheets, The Japan
Prestressed Concrete Engineering Association, Proceeding of the 5th Symposium on Developments in
Prestressed Concrete.
Vrana, J., (1983). The Behaviour and Repair of Prestressed Bridge Girders in Shear, Masters Thesis,
Department of Civil Engineering, McGill University, Montreal, Quebec, Canada.
WMNPhi (2001). WMNPhi Version. P & P Engineering, Qubec, Canada.
Yagi K., Tanaka T. and Jinnai T., (1996). Experimental Studies on Strengthening of Prestressed Concrete
Beam with Carbon Fiber Sheet. Proceedings of ICCI96, Saadatmanesh H., Ehsani M.R., editors, Tucson
US.

40

frp,max

35

30

Contrainte, MPa

25

20

Efrp

15

10

frp,max

-5
-0.15

-0.10

-0.05

0.00

0.05
0.10
0.15
Dformation[e-2]

0.20

0.25

0.30

0.35

(a) Parabolic stress-strain curve for concrete

(b) Stress-strain curve for composite material

s
fy

Es

y
(c) Stress-strain response of prestressing strand
(Ramberg-Osgood function)

s,max

(d) Stress-strain curve for reinforcing steel

Figure 1. Stress-strain curves of the constituent materials

Figure 2. Geometry of bridge girders

Figure 3. Reinforcement details

Figure 4. Girders with two pretensioned strands

Figure 5. Girders with four pretensioned strands

LVDT # 1
LVDT # 3

LVDT # 2

LVDT
5 et # 6

LVDT # 4
1000 mm

1000 mm
J2 , J3 & J4
J1

J5
J6 & J7

J' 3 J'5& J'6

Demec points

J'1

J' 4

J'2

1500 mm

1500 mm

Figure 6. Instrumentation Lay-out

600

PCB1R3

PCB2R3
PCB1

PCB2C1

500

Moment ( kN . m )

400

300

PCB2R2
PCB2

200

PCB2R2

100

0
0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

Curvature ( rad / m)

Figure 7. Moment-curvature response

0.08

0.09

Table 1. List of Specimens

Specimen

Prestressing area Aps (mm2)

PCB1 (Control)
PCB2 (Control)
PCB2R2
PCB2R3
PCB2C1
PCB1R3
PCB1R3D

Number of layers

Remarks

220
110
110
110
110
220

Type of
fibres
R* + T
R+T
C + T
R+T

Double-layer
Triple-layer
Single-layer
Triple-layer

220

R+T

Triple-layer

Reference specimen
Reference specimen
Reinforced beam
Reinforced beam
Reinforced beam
Beam loaded until cracking
and reinforced
Beam with 2 rods cut at 400
mm from the centre in both
sides and reinforced

* Replark-20 (Mitsubishi)
Carbodur-S (Sika)

Tyfo-S (Hexcel Fyfe)

Table 2 Physical Properties of the composite materials
Type of FRP
FRP name
Thickness
Tensile strength
Ultimate strain
Elastic modulus

Type R
Replark-20 (Mitsubishi)
0.11 mm/ layer
3400 MPa
0.0148
230 GPa

Type C
Carbodur-S (Sika)
1.2 mm/ layer
2400 MPa
0.00155
155 GPa

Type T
Tyfo-S (Hexcel Fyfe)
1.3 mm/ layer
454 MPa
0.02
22.7 GPa

Table 3 - Properties of concrete

fc (MPa)
40

c (%)
0.0021

fr (MPa)
3.795

E (MPa)
28460.5

cmax (%)
0.0035

Table 4 Properties of reinforcement

Size
10 M
D6.5
D8.5
9.53 mm strand

A (mm2)
100
42.1
54.9
54.8

E (MPa)
200000
200000
200000
185000

10

fy (MPa)
400
400
400
1715

Function
Longitudinal
Stirrups
Shear Connectors
Prestressing