Composite Structures 204 (2018) 333–341

Contents lists available at ScienceDirect

Composite Structures
journal homepage: www.elsevier.com/locate/compstruct

Design approach for flexural capacity of concrete T-beams with bonded T

prestressed and nonprestressed FRP reinforcements
⁎
Fei Peng, Weichen Xue
Department of Structural Engineering, Tongji University, Shanghai 200092, China

A R T I C LE I N FO A B S T R A C T

Keywords: Concrete beams with prestressed and nonprestressed fiber-reinforced polymer (FRP) reinforcements are com-
FRP reinforcement monly employed in field applications. However, available flexural strength design approaches mainly focus on
Prestressed concrete beam rectangular concrete beams exclusively prestressed with FRP tendons. This paper, therefore, presents a simpli-
Failure mode fied yet rational design approach for flexural capacity of concrete T-beams with bonded prestressed and non-
Transition region
prestressed FRP reinforcements. Firstly, a new transition region between tension- and compression- controlled
Flexural capacity
Design equation
sections was proposed in terms of ratio of provided-to-balanced reinforcement (ρe,b < ρe ≤ 1.5ρe,b) based on a
statistical analysis of an experimental database of 83 beams. Afterwards, numerical sectional analysis procedure
of tension-controlled sections was developed by using an accurate stress block to approximate the nonlinear
compressive stress distribution in concrete. Based on a detailed parametric study of over 160,000 sections,
simplifed design equations for flexural capacity of tension-controlled section is derived from multiple regression
analyses. Then, design equations were presented for flexural capacity of compression-controlled sections.
Finally, the performance of the proposed approach was evaluated by comparing their predictions with experi-
mental results of the 83 beams.

1. Introduction nonprestressed FRP reinforcement has been proven an efficient way to

The deterioration in concrete structures due to corrosion of steel
strands and bars is one of the major challenges facing the construction
industry. When exposed to deicing salts and marine environments,
concrete structures are susceptible to corrosion. A promising solution to
eliminate the corrosion-related problems associated with conventional
prestressed and reinforced concrete structures is the application of fiber
reinforced polymer (FRP) reinforcements [1,2]. FRP reinforcements
exhibit several properties that include high resistance to corrosion, high
strength-weight ratio, outstanding fatigue resistance, lower elastic
modulus compared to steel, and a linear stress-strain relationship.
Among FRPs, aramid FRP (AFRP) and carbon FRP (CFRP) are suited for
prestressing tendons because of their higher strength and lower sensi-
tivity to creep rupture. Glass FRP (GFRP) and Blast FRP (BFRP) are
more widely used as nonprestressed bars because of their lower cost.
Although the use of FRP as structural reinforcement shows great
promise in terms of durability, issues of its low elastic modulus and
linear elastic to failure without yielding must be addressed in a prac-
tical manner [3,4]. In practice, the concrete beams exclusively pre-
stressed with FRP tendons are rarely employed because of their lack of
ductility and control of crack distribution. Combining prestressed and
improve cracking behavior and minimize deflections at service loads
while providing additional deformability [4,5]. To date, remarkable
progresses have been made in field applications of prestressed concrete
bridges with FRP reinforcements in North America, Europe, Japan and
China [6–8].
Nowadays, considerable experimental studies have been devoted to
investigate flexural response of concrete beams with bonded pre-
stressed and nonprestressed FRP reinforcements [9–14]. Three flexural
failure modes, including rupture of prestressed FRP, rupture of non-
prestressed FRP, and concrete crushing, were observed in these ex-
periments. Generally, the first two modes are regarded as tension
failure, and the last is regarded as compression failure. Although
compression failure is the preferred failure mechanism since it is more
progressive and less catastrophic, in practice, where the beam and the
topping slab work together in a composite action, tension failure is a
common practice [15,16].
Theoretically, the flexural failure mode can be determined by
comparing the reinforcement ratio with the balanced reinforcement
ratio. However, available studies have shown that the actual flexural
failure mode may not coincide with the predicted one [17]. That is,
there is a transition region where concrete crushing and FRP rupture

⁎
Corresponding author at: Department of Structural Engineering, Tongji University, Siping Road 1239, Shanghai 200092, China.
E-mail address: xuewc@tongji.edu.cn (W. Xue).

https://doi.org/10.1016/j.compstruct.2018.07.091
Received 10 May 2018; Accepted 26 July 2018
Available online 31 July 2018
0263-8223/ © 2018 Elsevier Ltd. All rights reserved.
F. Peng, W. Xue
are possible. For concrete members with steel reinforcements, the de-
finition of the transition region between tension- and compression-
controlled section was proposed to achieve sufficient ductility and give
adequate warning prior to failure. For concrete members with FRP re-
inforcement, however, a brittle flexural failure is unavoidable and this
may lead to a change in the definition of the transition region. Cur-
rently, ACI 440.4R-04 [18], in accordance with ACI 318-14 [19], de-
fines the FRP-PC sections in transition region as those that have net
tensile strain in the extreme tension reinforcement at nominal strength
greater than 0.002 and less than 0.005. It should be mentioned that the
strain limits were originally proposed for flexural concrete members
with conventional Grade 60 (fy = 414 MPa) steel bars or Grade 270
(fpy = 1860 MPa) steel strands [20]. Considering that high-strength
steel bars or strands that have different stress-strain relationships, the
strain limits defining the transition region should be modified [21–24].
Since FRP is linear elastic to failure, the traditional strain limits de-
veloped for conventional steel-PC members may not be necessarily
applicable to FRP-PC members.
So far, valuable design approaches have been developed for flexural
capacity of PC members with boned FRP reinforcements [18,25–27].
Previous approaches mainly focused on concrete beams exclusively
prestressed with FRP tendons [18,25,26]. However, very limited re-
search was carried out to predict the flexural capacity of concrete beam
with a combination of prestressed and nonprestressed FRP reinforce-
ments, especially for the beam with T-section. Grace et al. [27] pre-
sented a design approach for concrete beams with prestressed and
nonprestressed CFRP. However, the failure due to nonprestressed CFRP
rupture, which was observed by Yonekura et al. [10], was not discussed
in their approach. Moreover, it is important to recognize that the state
of the under-reinforced section at ultimate in an FRP-PC beam is not
analogous to that of a steel-PC beam. In a steel-PC beam, the failure of
under-reinforced section is yielding of the tension steel, followed by
eventual crushing of the concrete. This means that the concrete in
compression experiences stress redistribution after the steel yields. In
this case, Whitney’s rectangular stress block can be used to determine
the flexural capacity. In the case of an FRP-PC beam, however, the
failure of the under-reinforced section is rupture of FRP. This means
that there is no redistribution of stresses in the section and a cata-
strophic collapse will occur when the FRP fails. Consequently, an
iterative procedure is required to determine the concrete stress dis-
tribution in compression. This iterative procedure is overly complex for
normal calculations in design offices [28]. To avoid this iteration, ACI
440.4R-04 [18] and Grace et al. [27] approximated the concrete com-
pressive stress distribution by using Whitney’s rectangular stress block.
This simplification, however, overestimates the compressive stress of
concrete and consequently the lever arm, which may result in a non-
conservative design.
From the above discussions, it can be concluded that there is very
limited research on flexural strength design approaches of concrete
beams with prestressed and nonprestressed FRP reinforcements, espe-
cially for the beams with T-sections. Besides, the brittle behavior of FRP
may lead to a change in definition of the transition region, which hasn’t
been investigated. Moreover, an iterative procedure is required when
FRP rupture governs the design. In this paper, a simplified yet rational
flexural strength design approach is developed for concrete T-beams
with bonded prestressed and nonprestressed FRP reinforcements.
Firstly, sectional analyses under balanced failure are firstly carried out
to distinguish possible flexural failure modes. Based on a statistical
analysis of an experimental database of 83 beams, a new transition
region is proposed. Then, rigorous sectional analyses and multiple
linear regression analyses are carried out to obtain simplified design
equations for predicting the flexural capacity of under-reinforced sec-
tions. Subsequently, design equations are presented to predict the
flexural capacity of over-reinforced sections. Finally, the accuracy of
the proposed approach is verified by experimental results of the 83
beams. It should be mentioned that the proposed design approach has
334
Composite Structures 204 (2018) 333–341
been incorporated in Shanghai’s Construction Standard “Design and
Construction of Concrete Structures with FRP Reinforcements”.
2. Distinction of flexural failure modes
2.1. Basic assumptions
Referring to ACI 440.1R-15 [29] and ACI 440.4R-04 [18], the fol-
lowing assumptions are made in predicting ultimate flexural strength
behaviors of concrete beams with bonded prestressed and non-
prestressed FRP reinforcements:
1. There is a perfect bond between the FRP reinforcement and the
concrete;
2. A plane-cross section remains plane after bending deformation;
3. The maximum usable compressive strain in the concrete, εcu, is as-
sumed to be 0.003;
4. The stress-strain curve of FRP reinforcement is idealized as linear
elastic to failure;
5. The stresses in the concrete can be computed from the strains by
using stress-strain curves for concrete;
6. The tensile strength of concrete is ignored; and
7. The compression contribution of the FRP reinforcements is ne-
glected.
2.2. Balanced failure mode
Theoretically, the distinction among different flexural failure modes
of a concrete beam can be achieved through the predicted balanced
failure. Based on sectional analyses, two balanced failure conditions
and three flexural failure modes are identified for concrete beams with
prestressed and nonprestressed FRP reinforcements. The three failure
modes include rupture of prestressed FRP (Mode I), rupture of non-
prestressed FRP (Mode II) and concrete crushing (Mode III), which
coincides with the observations in available experiments [10–14]. The
sectional analyses indicate that the flexural failure modes depend on
the arrangement of the flexural reinforcements. If the parameter χ
defined by Eq. (1) is not less than 1.0, the beam will fail due to pre-
stressed FRP rupture (Mode I) or concrete crushing (Mode III). Other-
wise, the failure mode will be governed by nonprestressed FRP rupture
(Mode II) or concrete crushing (Mode III).
(εfu + εcu ) dp
χ =
(εpu−εpe + εcu ) df (1)
where εpe and εpu are the effective prestressing strain and ultimate
tensile strain in prestressed FRP, respectively, εfu is the ultimate tensile
strain in nonprestressed FRP, and dp and df are the distance from ex-
treme concrete compression fiber to centroid of prestressed and non-
prestressed FRP, respectively.
Fig. 1(b) and (c) show the strain and stress distributions across the
depth of a section of concrete beam, respectively. The balanced failure
condition between Mode I and Mode III is obtained when the strains in
the concrete and prestressed FRP simultaneously reach their maximum
values. However, the stresses in nonprestressed FRP at failure, ff,
Fig. 1. Strain and stress conditions for εcf = εcu: (a) cross section (b) strain (c)
stress (d) stress (equivalent).
F. Peng, W. Xue Composite Structures 204 (2018) 333–341

doesn’t reach the ultimate tensile strength ffu. In this case, the concrete The corresponding ratio of depth of neutral axis to that of non-
stress distribution in compression zone can be approximated by Whit- prestressed FRP reinforcement under balanced condition kfu,b can be
ney’s rectangular stress block [see Fig. 1(d)]. Generally, the decom- expressed as
pression strain and the strain resulting from sustained loads are negli- cb εcu
gible [18,27], and therefore these two strain values are assumed zero. kfu, b = =
df εfu + εcu (9)
Based on the equilibrium of internal forces and strain compatibility, the
balanced failure condition can be expressed as:
2.3. Determination of transition region
Ap Af Ef ⎡ df 0.85f ′c (b−bw ) hf
+ (εpu−εpe + εcu )−εcu⎤−
b w dp ⎢
bw dp fpu ⎣ dp ⎥ fpu b w dp
⎦ As FRP possesses mechanical properties different from steel, the
f ′c
εcu design philosophy of concrete beams with FRP reinforcements differs
= 0.85β1
fpu εcu + εpu−εpe (2) from that of traditional concrete beams with steel reinforcements
[28,29]. Traditional steel reinforced/prestressed concrete design phi-
where fpu is the ultimate tensile strength of prestressed FRP, fc′ the is losophy strongly encourages the use of under-reinforced sections to
cylinder compressive strength of concrete, Ap and Af are the area of ensure a ductile failure mode. For FRP-PC beams, however, a brittle
prestressed and nonprestressed FRP, respectively, Ef is the modulus of flexural failure is unavoidable. This is fundamentally different from
elasticity of nonprestressed FRP, bw is the width of web (width of a steel reinforced/prestressed concrete. Although compression failure
rectangular cross section), b is the width of compression face of mode is considered as the preferred failure mechanism, current design
member, hf is the depth of the flange in T-section, and β1 is the ratio of guidelines permit tension failure as well [5,18].
depth of equivalent rectangular stress block to depth of neutral axis. Theoretically, if the effective reinforcement ratio ρe [given by Eq.
Define the effective prestressed FRP reinforcement ratio, ρep, as (3) or Eq. (7)], is less than the corresponding balanced ratio ρe,b [given
by Eq. (4) or Eq. (8)], tension failure mode (FRP rupture) governs.
Ap Af Ef ⎡ df 0.85f ′c (b−bw ) hf
ρep = + (εpu−εpe + εcu )−εcu⎤− Otherwise, compression failure mode (concrete crushing) governs. Al-
b w dp bw dp fpu ⎢ d ⎥ fpu b w dp (3)
⎣ p ⎦ though the flexural failure mode can be predicted based on calcula-
Then, the corresponding balanced reinforcement ratio ρep, b between tions, the actual member may not coincide with the predicted one. This
Mode I and Mode III can be expressed as mainly attributes to the uncertainties of material strengths, assumptions
made in analysis, and variations in locations of reinforcements and
f ′c εcu dimensions of concrete sections [17]. In this study, a total of 83 PC
ρep, b = 0.85β1
fpu εcu + εpu−εpe (4) beams with FRP reinforcements tested in flexure were collected from
the available experiments. Balanced reinforcement ratios provisions
Alternatively, the balanced condition between Mode I and Mode III
were performed to predict the flexural failure modes. It was found that
can be expressed in terms of the ratio of depth of neutral axis to that of
five beams did not fail in the predicted failure mode, and that there was
prestressed FRP kpu. Based on a sectional analysis, Eq. (5) can be ap-
a transition region where tension and compression failure mode are
plicable for different geometry and cross sections of beams. For Mode I
possible. Similar findings in FRP reinforced concrete beams were re-
kpu is less than kpu,b; and for Mode III kpu is greater than kpu,b, where kpu,b
ported by other researchers [17,30].
is ratio of depth of neutral axis to that of prestressed FRP under ba-
For FRP-PC members, ACI 440.4R-04 [18], in accordance with ACI
lanced failure conditions.
318-14 [19], defines the sections in transition region in terms of the net
cb εcu tensile strain limits. Note that, the strain limits were originally pro-
kpu, b = =
dp εcu + εpu−εpe (5) posed to achieve sufficient ductility and give adequate warning prior to
failure for flexural members with conventional steel bars and strands
where cb is the depth of neutral axis under balanced condition. Note
[20]. Since FRP is linear elastic to failure, a ductile failure does not
that if the cb is less than hf, the section can be treated as a rectangular
occur in FRP-PC concrete. Therefore, the strain provision for defining
section with a width of b.
the transition region may not be applicable for FRP-PC beams [21–24].
The balanced failure condition between Mode II and Mode III is
Alternatively, the transition region can be stated in terms of the ratio of
obtained when the strains in the concrete and nonprestressed FRP si-
provided-to-balanced reinforcement [17,29], which not only reflects
multaneously reach their maximum values. However, the stresses in
the flexural behavior of the member but also considers the effect of
prestressed FRP at failure, fp, doesn’t reach the ultimate tensile strength
mechanical properties of the materials. For example, current ACI
fpu. Base on a sectional analysis, the balanced failure condition can be
440.1R-15 [29] defines the sections in transition region as those that
expressed as:
have FRP reinforcement ratio greater than the balanced reinforcement
Af Ap Ep ⎡ dp 0.85f ′c (b−bw ) hf ratio ρb and less than 1.4ρb. In this study, therefore, balanced re-
+ (εfu + εcu ) + εpe−εcu⎤−
b w df bw df f fu ⎢⎣ d f
⎥
⎦ f fu b w df inforcement provision was adopted to define the transition region of
FRP-PC beams. The upper bound of the transition region can be de-
f ′c εcu
= 0.85β1 termined by Eq. (10) [17,30].
f fu εfu + εcu (6)
ρe, b
ρ*=
where f fu is the ultimate tensile strength of nonprestressed FRP re- 1−3σC (10)
inforcement.
Define the effective nonprestressed FRP reinforcement ratio, ρef, as: where ρe,b is the balanced reinforcement ratio, which can be determined
from Eq. (4) and Eq. (8) for χ ⩾ 1 and χ < 1, respectively, and σC is the
Af Ap Ep ⎡ dp 0.85f ′c (b−bw ) hf standard deviation of the ratio between the calculated and experi-
ρef = + (εfu + εcu ) + εpe−εcu⎤−
b w df bw df f fu ⎢
⎣ d f
⎥
⎦ f fu b w df (7) mental flexural capacity of the beams failing in compression failure.
The standard deviation σC in Eq. (10) is equal to 10.8% based on a
Then the balanced reinforcement ratio between Mode II and Mode comparison between the predicted and experimental flexural capacity
III is given by Eq. (8). of 22 beams failing in compression with ρe > ρe,b. Therefore, ρ* should
f ′c εcu be taken as 1.5ρe,b. Note that the suggested upper limit of 1.5ρe,b is in
ρef , b = 0.85β1 accordance with the limit of FRP reinforced concrete beams suggested
f fu εfu + εcu (8)
by Xue et al. [17].

335
F. Peng, W. Xue Composite Structures 204 (2018) 333–341

Table 1 Todeshini et al. [33] is adopted, which is given as:

Reinforcement ratio versus failure mode.
2f ′c (εc / ε0 )
Condition Reinforcement ratio Failure mode fc =
1 + (εc / ε0 )2 (13a)
χ⩾1 ρep ⩽ ρep, b Prestressed FRP rupture
ρep, b < ρep ⩽ 1.5ρep, b Transition region f ′c
ε0 = 1.71
ρep > 1.5ρep, b Concrete crushing Ec (13b)
χ<1 ρef ⩽ ρef , b Nonprestressed FRP rupture where εc is the compressive concrete strain, ε0 is the compressive
ρef , b < ρef ⩽ 1.5ρef , b Transition region concrete strain at maximum strength, and Ec is the concrete modulus.
ρef > 1.5ρef , b Concrete crushing Referring to Fig. 2(b), the compressive concrete strain along the
depth of the neutral axis can be expressed as:

At the ultimate limit state, most of design provisions recommend the x

εc (x ) = εcf
compression-controlled seciton design since it is less catastrophic and c (14)
more progressive. Besides, available reliability studies on concrete The strain compatibility from Fig. 2(b) allows determination of the
members with FRP reinforcements recommended that the strength re- kpu in terms of the available strains.
duction factor of tension-controlled sections was less than that of
compression-controlled sections [31,32]. In design practice, the tran- c εcf
kpu = =
sition region can be taken as ρe, b < ρe ⩽ 1.5ρe, b . Table 1 lists the flexural dp εpu−εpe + εcf (15)
reinforcement ratios and corresponding flexural failure modes.
Based on the equilibrium of internal forces and linear strain dis-
tribution assumption (as shown in Fig. 2), the following equations are
3. Prediction of ultimate flexural capacity
obtained
3.1. Mode I: Prestressed FRP rupture Ap fpu + Af f f = αβkpu fc′ bdp + αfc′ (b−bw ) hf (16)

When rupture of prestressed FRP governs the design, the ultimate
tensile strength of prestressed FRP, fpu, is reached. However, the com-
pressive strain at the extreme compression fiber of concrete, εcf, doesn’t
reach the maximum concrete strain εcu. Therefore, Whitney’s rectan-
gular stress block cannot be assumed to represent the nonlinear stress
distribution in the concrete. In addition, the stress in nonprestressed
FRP, ff, and the depth of the neutral axis, c, are unknown. Consequently,
calculation of the flexural capacity of the section requires use of the
nonlinear stress–strain curve of the concrete, and this necessitates a
numerical solution procedure. In this study, an equivalent rectangular
stress block with two strain-dependent and stress-dependent para-
meters α and β is used to approximate the nonlinear stress distribution
in the concrete [as shown in Fig. 2(d)]. For the resultant compressive
forces of the actual and equivalent stress blocks to have the same
magnitude and line of action, the values for the parameters α and β
must be
c c
αfc′ bw βc + αfc′ (b−bw ) hf = bw · ∫0 fc (x ) dx + (b−bw )·
βc hf c step 5.
αfc′ bw βc ⎛c− ⎞ + αfc′ (b−bw ) hf ⎛c− ⎞ = bw
⎜ ⎟ ∫0
⎝ 2⎠ ⎝ 2⎠
c Repeating the iterative procedure for increasing values of the re-
+ (b−bw ) ∫c−h f
fc (x )·xdx
where fc (x ) is the compressive stress in the compression zone, and x is
the distance from the neutral axis to any location through the depth.
To determine the parameters α and β, the stress-strain relationship
in concrete is required. In this study, the concrete model proposed by
ψf −kpu
f f = Ef (εpu−εpe ) ⩽ f fu
1−kpu (17)
where ψf = df / dpf .
In order to determine the stress distribution in the concrete, the
stress in nonprestressed FRP and the location of neutral axis, a nu-
merical procedure is required. The procedure can be implemented
through the following steps:
1. Select a small reinforcement ratio ρe at a given Af/Ap ratio;
2. Assume a value for the concrete strain εcf at extreme compression
fiber of concrete;
3. According to the linear strain distribution, determine the value of
kpu using Eq. (15);
4. Solve the block parameters α and β by using Eqs. (11)–(14);
5. According to Eqs. (16) and (17), determine the value of kpu and ff;
and
∫c−h f
fc (x ) dx
6. If kpu in step 3 equals to the value in step 5, the coefficient kpu can be
(11) calculated from Eq. (15). Otherwise, εc should be changed and steps
2 to 5 should be repeated until kpu in step 4 equals to the value in
fc (x )·xdx
(12) inforcement ratio ρep until it reaches the balanced reinforcement ratio
ρep,b, the corresponding non-dimensional equivalent neutral axis depth
βkpu can be computed and, thus, tables of ρep/ρep,b and βkpu can be
generated.
Based on aforementioned procedure, parametric studies were per-
formed using a wide range of values for design parameters. A total of
160,380 isolated tension-controlled sections were simulated, each
having a different combination of specified properties. Table 2 shows
the design parameters and their values, which represents wide ranges of
these variables used in construction industry. A correlation analysis of
the numerical βkpu data indicated that the ρep/ρep,b and b/bw had the
most significant effect on the βkpu. Base on a multiple linear regression
analysis of the numerical βkpu data, Eq. (18) for the non-dimensional
equivalent neutral axis depth βkpu was proposed

βkpu = ζβ1 kpu, b (18a)

Fig. 2. Strain and stress conditions for εcf < εcu: (a) cross section (b) strain (c)
stress (d) stress (equivalent). where

336
F. Peng, W. Xue Composite Structures 204 (2018) 333–341

Table 2
Specified parameters of tension-controlled section studied.
Parameters Values No. of case

Compressive concrete strength fc′ 40 MPa; 50 MPa; 60 MPa 3

Modulus of elasticity of prestressed 75 GPa (AFRP); 145 GPa 2
FRP Ep (CFRP)
Modulus of elasticity of 75 GPa (AFRP); 145 GPa 3
nonprestressed FRP Ef (CFRP); 45 GPa (GFRP)
Ratio of area of nonprestressed FRP 0; 1.0; 2.0 3
to that of prestressed FRP Af/Ap
Ultimate tensile strain of FRP εpu 0.015; 0.02; 0.025 3
Effective reinforcement ratio ρe 0.2ρe,b–1.0ρp,ba or 33
0.35ρe,b–1.0ρe,bb
Effective prestressing stress fpe 0.4fpu; 0.5fpu; 0.6fpu 3
Depth of the flange in T-section hf 0.1 h; 0.15 h; 0.2 h 3
Width of compression face of bw; 2bw; 3bw; 4bw 4
member b
Total cases 160,380

a
For rectangular section. Fig. 4. Stress in nonprestressed FRP ff for Mode I.
b
For T-section.
ψf −ζkpu, b
f f = Ef (εpu−εpe ) ⩽ f fu
⎡ b b ρep ⎤ 1−ζkpu, b (19)
ζ = ⎢ ⎛0.15 + 0.1 ⎞ + ⎛0.85−0.1 ⎞
bw ⎠ ρep, b ⎥
⎜ ⎟ ⎜ ⎟

⎝ bw ⎠ ⎝
⎣ ⎦
A sensitive analysis was carried out to assess the consistency of Eq.
(18). Fig. 3 compares the βkpu data from numerical analysis and those
from Eq. (18). As shown, the proposed approach provided consistent
and accurate predictions of the βkpu over a wide region of numerical
data. The coefficient of determination R2 of Eq. (18) and the corre-
sponding standard error Se were found to be 0.982 and 0.0041, re-
spectively, representing adequate and fair correlation [34].
For simplify, the stress in nonprestressed FRP at failure is approx-
imate by
(18b)
Fig. 4 compares the ff obtained from the numerical procedure and
that obtained from Eq.(19). It is found that Eq. (19) can accurately
predict the ff. Once the βku and f f are obtained from Eqs. (18) and (19),
respectively, the strain-dependent parameter α can be calculated from
Eq.(16). Fig. 5 compares the α obtained from the proposed simplified
procedure against the numerical data. Comparisons demonstrate that
the proposed simplified procedure can provide good predictions of the
α. Note that the strain-dependent parameter α should not be greater
than 0.85.
From the moment equilibrium, the ultimate flexural capacity of the

Fig. 3. Variation in non-dimensional equivalent neutral axis depth βkpu for Mode I.

337
F. Peng, W. Xue Composite Structures 204 (2018) 333–341

Fig. 7. Comparison of stress increase in prestressed FRP for Mode II.

Fig. 5. Variation in strain-dependent parameter α for Mode I.

nonprestressed GFRP. It is found that the βkfu obtained from Eq. (21) is
section then can be computed by Eq. (20)
in good agreement with the numerical results.
βkpu dp ⎞ βkpu dp ⎞ The stress in prestressed FRP at failure fp can be approximate by
Mu = Ap fpu ⎛dp−
⎜ + Af f f ⎛df −
⎟ ⎜ ⎟

⎝ 2 ⎠ ⎝ 2 ⎠ 1/ ψf −ζkfu, b ⎞
βkpu dp hf ⎞ fp = Δfp + fpe = Ep ⎜⎛ ⎟ εfu + f pe ⩽ f pu
+ αfc′ (b−bw ) hf ⎛ ⎜ − ⎟
⎝ 1−ζkfu, b ⎠ (22)
⎝ 2 2⎠ (20)
where Δfp is the stress increase in prestressed FRP at failure. Fig. 7
compares the Δfp obtained from Eq. (22) against these from numerical
3.2. Mode II: Nonprestressed FRP rupture procedure. Comparisons demonstrate that Eq. (22) can provide good
predictions of the Δfp .
When nonprestressed FRP rupture governs the design, the ultimate Once the βkfu and fp are determined from Eq. (21) and (22), re-
tensile strength of nonprestressed FRP, ffu, is reached. Howbeit, the spectively, the strain-dependent parameter α can be calculated from Eq.
stresses in prestressed FRP, fp, and the concrete compressive strain, εcf, (23).
at failure are unknown. Based on aforementioned procedure, a detailed
parametric study in conjunction with a multiple regression analysis Ap fp + Af f fu
α =
were conducted to statistically derive the expression of the on-dimen- βkfu f ′c bdf + f ′c (b−bw ) hf (23)
sional equivalent neutral axis depth βkfu:
From the moment equilibrium, the ultimate flexural capacity of the
βkfu = ζβ1 kfu, b (21a) section can be computed by Eq. (24)

where βkfu df ⎞ βkfu df ⎞

Mn = Ap fp ⎛dp−⎜ + Af f fu ⎛df −
⎟ ⎜ ⎟

⎝ 2⎠ ⎝ 2 ⎠
⎡ b b ρef ⎤
ζ = ⎢ ⎛0.15 + 0.1 ⎞ + ⎛0.85−0.1 ⎞
⎜ ⎟ ⎜
⎥⎟ βk fu d f h f
⎣⎝
b w⎠ ⎝ b w ⎠ ρef , b
⎦ (21b) + αfc′ (b−bw ) hf ⎛ ⎜ − ⎞ ⎟

⎝ 2 2⎠ (24)
Fig. 6 shows (as an example) the βkfu coefficient graphically in terms
of ρef / ρef , b and b/ bw , for the case of a section with prestressed AFRP and
3.3. Mode III: Concrete crushing

When concrete crushing governs the design, the compressive strain

at the extreme compression fiber of concrete, εcf, reaches the ultimate
compressive strain εcu. The stresses in prestressed and nonprestressed
FRP reinforcement at failure, however, are unknown. In such a case, the
stress distribution in the concrete can be approximated by Whitney’s
rectangular stress block. As shown in Fig. 1, the force equilibrium
condition and compatibility condition on the cross section can be ex-
pressed as

0.85β1 fc′ bc + 0.85fc′ (b−bw ) hf = Ap fp + Af f f (25)

dp εcu
fp = Ep ⎛⎜ + εpe−εcu⎞ ⩽ fpu ⎟

⎝ c ⎠ (26)

df −c
f f = Ef εcu ⩽ f fu
c (27)
Fig. 6. Variation in non-dimensional equivalent neutral axis depth βkfu for Summing moments about the compression centroid defines the ul-
Mode II. timate moment capacity as

338
F. Peng, W. Xue Composite Structures 204 (2018) 333–341

Table 3
Comparisons between the predicted and experimental flexural capacity.
Ref. Specimen Geometries fc′ Prestressed FRP Pea Nonprestressed FRP ρe ρe/ρe,b Actual modeb Mexp Mpre/Mexp
MPa kN % kN·m
b h Ap Ep fpu Af Ef ffu
mm mm mm2 GPa MPa mm2 GPa MPa

[12] CFRP_S1 152 305 41 50 107 1607 48 108 107 1607 0.25 0.41 TR 45.3 0.84
CFRP_S2 152 305 41 79 107 1607 76 108 107 1607 0.39 0.64 TR 63.1 0.92

[10] CI-C0-S0-2 150 220 56 76 145 2110 55 25 54 1900 0.42 0.95 RR 28.2 1.03
AI-A0-S0-2 150 220 57 85 54 1900 102 152 145 2110 0.40 1.05 RRc 33.0 0.98
CI-CI–S0-2 150 220 60 76 145 2110 77 57 54 1900 0.90 1.59 CC 55.9 0.98
AI-A-S0-2 150 220 53 85 54 1900 82 152 145 2110 0.50 1.70 CC 34.6 0.93
CII-CI–S0-2 150 220 55 155 139 1970 82 152 145 2110 1.25 2.82 CC 54.5 1.01
CII-CI–S0-4 150 220 65 155 139 1970 163 57 54 1900 1.44 2.00 CC 60.2 1.05
AI-A-A0-2 150 220 64 85 54 1900 82 61 140 2200 0.50 1.43 CC 41.5 0.85
CI-C-C0-3 150 220 64 76 145 2110 95 25 54 1900 0.45 0.63 TR 49.4 0.72

[4] PT2-ST 600 300 64 245 144 1765 217 1187 55 934 0.40 0.48 TR 249.4 0.92
PT2-G 600 300 59 245 144 1765 242 1187 55 934 0.38 0.45 TR 217.9 0.97
PT2-S 600 300 65 245 144 1765 245 1187 55 934 0.37 0.41 TR 217.0 0.97
PT2-S45 600 300 65 245 144 1765 246 1187 55 934 0.37 0.40 TR 207.4 0.96
PT4 600 300 56 491 144 1765 489 1187 55 934 0.56 0.70 TR 258.1 1.09
PT4-G 600 300 59 491 144 1765 474 1187 55 934 0.56 0.68 TR 273.0 1.05

[11] II-3 150 250 45 123 143 2400 162 157 567 42 700 1.92 CC 66.5 0.93

[35] 3 100 150 35 56 140 1734 68 – – – 0.56 0.85 TR 9.0 0.89

4 100 150 35 56 140 1724 57 – – – 0.56 1.01 CC 9.0 0.89
5 100 150 35 30 140 1875 23 – – – 0.30 0.83 TR 5.5 0.93
7 100 150 35 30 140 1875 23 30 140 1875 0.55 1.49 CC 9.3 0.94
8 100 150 35 30 140 1875 23 39 130 1578 0.60 1.63 CC 9.5 0.97
10 100 150 35 30 140 1875 37 30 140 1875 0.45 0.86 TR 8.6 0.99
11 100 150 35 30 140 1875 37 39 130 1578 0.48 0.92 TR 8.9 1.03

[13] CFCC beam 457 356 64 462 159 2590 443 347 159 2590 0.45 1.14 CC 381.8 1.14
CFRP beam 457 356 50 288 157 2344 385 216 157 2344 0.25 0.52 TR 315.2 0.91

[14] C-S-F-U 457 406 65 462 159 2590 587 – – – 0.27 0.59 TR 376.4 1.09
C-S-F-B 457 406 53 462 159 2590 587 347 159 2590 0.66 1.18 TRc 462.0 1.03
C-S-F-O 457 406 65 462 159 2590 587 578 159 2590 0.79 1.43 CC 495.0 1.16

[36] T-4-5.H 600 330 47 201 147 2250 156 – – – 0.12 0.35 TR 127.3 0.95
R-4-5.H 200 330 47 201 147 2250 139 – – – 0.36 1.10 CC 107.2 1.02
T-4-5.V 600 330 61 201 147 2250 154 – – – 0.13 0.30 TR 117.5 0.94
R-4-5.V 200 330 61 201 147 2250 161 – – – 0.39 0.89 TR 108.2 0.97
T-4-7.V 600 330 64 201 147 2250 237 – – – 0.13 0.23 TR 122.6 0.89
R-4-7.V 200 330 64 201 147 2550 225 – – – 0.39 0.95 TR 117.7 1.01
T-2-5.V 600 330 70 108 147 2950 81 – – – 0.07 0.23 TR 67.6 1.09
R-2-5.V 200 330 70 108 147 2950 84 – – – 0.20 0.70 TR 68.2 1.05

[25] CFCC1 241 305 50 76 137 2118 81 – – – 0.13 0.28 TR 34.9 1.05
Strawman1 241 305 31 108 146 1912 107 – – – 0.17 0.44 TR 39.2 1.08
Strawman2 241 305 31 108 146 1912 107 – – – 0.17 0.44 TR 39.3 1.08
Strawman3 181 241 31 50 146 1912 53 – – – 0.14 0.34 TR 19.1 0.95

[37] B1 150 300 45 120 50 1330 80 – – – 0.32 0.72 TR 35.7 1.05

[38] B1 150 300 41 58 53 2500 65 – – – 0.15 1.22 TRc 34.8 0.95

B2 150 300 40 58 53 2500 66 – – – 0.15 1.24 TRc 33.2 0.93
B3 150 300 40 58 53 2500 66 – – – 0.15 1.22 TRc 40.0 0.89

[39] TB(1) 100 200 49 25 54 1800 31 – – – 0.19 0.46 TR 6.4 0.87

TB(2) 100 200 49 25 54 1800 31 – – – 0.19 0.46 TR 6.4 0.87
FB 100 200 57 33 69 1430 27 – – – 0.25 0.35 TR 6.2 0.93

[40] Oc-Ar-UR-L 106 200 57 25 54 1802 41 – – – 0.18 0.55 TR 7.2 0.85

Oc-Ar-OR-L 106 200 49 123 54 1846 106 – – – 0.87 3.32 CC 22.9 0.83
Cc-Ar-OR-L 106 200 52 123 54 1846 108 – – – 0.87 3.08 CC 17.6 1.02
FRc-Ar-OR-L 106 200 49 123 54 1846 106 – – – 0.87 3.25 CC 22.1 0.80

[26] C1-H 1220 460 43 264 146 1822 267 – – – 0.06 0.10 TR 166.0 0.99
C2-H 1220 460 43 264 146 1822 267 – – – 0.06 0.10 TR 157.0 1.05

[1] A2-S4 120 210 37 84 76 1639 55 – – – 0.50 1.62 CC 17.9 0.81

B1-S3 120 210 38 90 68 1461 25 – – – 0.54 1.98 CC 14.5 0.83
C1-S6 120 210 32 180 63 1392 116 – – – 1.07 3.24 CC 22.6 0.83
C1-S7 120 210 32 180 63 1392 116 – – – 1.07 3.23 CC 22.6 0.83
D1-S3 120 210 39 90 68 1461 56 – – – 0.54 1.43 CC 14.0 1.02

[41] Beam1 305 914 86 560 147 2600 1120 – – – 0.23 0.23 TR 1006.3 1.09
Beam2 305 914 85 560 147 2600 683 – – – 0.23 0.42 TR 1284.6 0.87

(continued on next page)

339
F. Peng, W. Xue Composite Structures 204 (2018) 333–341

Table 3 (continued)

Ref. Specimen Geometries fc′ Prestressed FRP Pea Nonprestressed FRP ρe ρe/ρe,b Actual modeb Mexp Mpre/Mexp
MPa kN % kN·m
b h Ap Ep fpu Af Ef ffu
mm mm mm2 GPa MPa mm2 GPa MPa

[42] B1 108 203 95 25 161 2200 30 – – – 0.14 0.14 TR 8.5 1.06

B2 108 203 85 25 161 2200 30 – – – 0.14 0.16 TR 11.2 0.83
B3 108 203 93 25 161 2200 30 – – – 0.14 0.14 TR 11.2 0.83
B4 108 203 100 25 161 2200 30 – – – 0.14 0.13 TR 10.9 0.85
B5 108 203 108 25 161 2200 30 – – – 0.14 0.13 TR 11.2 0.89
B6 108 203 82 25 161 2200 30 – – – 0.14 0.16 TR 10.5 0.90
B8 108 203 83 25 161 2200 30 – – – 0.14 0.16 TR 10.6 0.91
B9 108 203 77 25 161 2200 30 – – – 0.14 0.17 TR 10.0 0.93

[43] TR-1-7.5/7 500 550 51 568 137 2150 575 – – – 0.25 0.56 TR 515.2 1.01
TR-2-5/1 500 550 60 568 137 2150 575 – – – 0.25 0.47 TR 510.3 1.03
TR-3-4/7 500 550 61 568 137 2150 575 – – – 0.25 0.47 TR 490.9 1.07
LL-4-2B 500 550 61 473 147 2950 575 – – – 0.21 0.70 TR 581.6 1.03
LL-5-1B 500 550 66 473 147 2950 575 – – – 0.21 0.65 TR 617.2 0.98

[44] C80-120-S 150 300 77 100 147 2250 115 – – – 0.29 0.42 TR 47.8 1.04
C40-80-L-C 150 300 37 100 147 2250 79 – – – 0.29 1.04 CC 43.75 1.07
C40-120-L-C 150 300 37 100 147 2250 122 – – – 0.29 0.81 TR 46.25 1.02

[45] S30-2 150 250 62 127 144 1765 68 – – – 0.39 0.61 TR 42.4 1.05
S45-4 150 250 50 127 144 1765 94 – – – 0.39 0.66 TR 43.5 1.00
N30-1 150 250 65 127 144 1765 71 – – – 0.39 0.58 TR 42.3 1.05
N60-3 150 250 65 127 144 1765 130 – – – 0.39 0.42 TR 44.6 0.99
N60-4 150 250 65 127 144 1765 133 – – – 0.39 0.41 TR 43.7 1.00
N60-4 150 250 65 127 144 1765 133 – – – 0.39 0.41 TR 43.7 1.00

a
Pe = Effective prestressing force.
b
Failure modes: CC = concrete crushing; TR = Prestressed FRP rupture; RR = Nonprestressed FRP rupture.
c
Indicates disagreement between predicted and experimentally observed flexural failure modes.

βc βc β c hf predictions regardless of the flexural reinforcement ratio. For over-re-

Mn = Ap fp ⎛dp− 1 ⎞ + Af f f ⎛df − 1 ⎞ + 0.85fc′ (b−bw ) hf ⎛ 1 − ⎞
⎜ ⎟ ⎜ ⎟ ⎜ ⎟

⎝ 2 ⎠ ⎝ 2 ⎠ ⎝ 2 2 ⎠ (28) inforced beams, the proposed approach slightly underestimates the load
capacity. This may be attributed to the fact that the approach ignores
reinforcement in the compression zone and the maximum concrete
4. Validation of design approach crushing strain can reach higher values than what is assumed in ana-
lysis [46]. For the beams in the transition region where the flexural
Test results of 83 FRP-PC beams collected from available literature failure mode is uncertain, it is found that the predictions of the beams
were used to validate the proposed design approach. The database obtained from equations for compression failure modes [Eqs.
contains 32 beams with T-shaped or I-shaped cross sections, and 61 (25)–(28)] are good agreement with the experimental flexural capa-
beams with rectangular cross sections. All 83 beams were prestressed cities.
with bonded FRP reinforcements and reported to fail in flexure. Table 3
lists the geometrical and material properties of all beams. Note that all 5. Conclusions
material reduction and safety factors in the design equations were set
equal to 1.0 for the sake of comparison. Table 3 compares the predic- This paper presents a theoretical investigation on the flexural ca-
tions from the proposed approach against the experimental flexural pacity of prestressed concrete T-beams with bonded prestressed and
capacities. Generally, the predictions are in good agreement with the nonprestressed FRP reinforcements. The following conclusions can be
experimental results. The average and standard deviation of the ratio of drawn from the study:
computed to experimental flexural capacities (Mpre/Mexp) are 0.96 and
10%, respectively. 1. Three flexural failure modes of concrete T-beams with prestressed
Particularly, the influence of reinforcement ratio on the accuracy of and nonprestressed FRP reinforcement were identified. The three
the design approach was investigated. Table 4 presents the means and failure modes include rupture of prestressed FRP, rupture of non-
the standard deviations of the ratio of Mpre/Mexp of under-reinforced prestressed FRP and concrete crushing.
beams (ρe ≤ ρe,b), beams in the transition region (ρe,b < ρe ≤ 1.5ρe,b) 2. An experimental database including 83 flexural tests on FRP-PC
and over-reinforced beams (ρe > 1.5ρe,b), respectively. It can be found beams available in the literature has been established. Based on a
that the proposed approach provides consistent and accurate statistical analysis of the experimental database, a new transition
region is proposed in terms of ratio of provided-to-balanced re-
Table 4 inforcement (ρe, b < ρe ⩽ 1.5ρe, b ) instead of the net tensile strain limits
Performance of flexural design approach. in current ACI 440.4R-04.
3. Based on rigorous sectional analyses and parameter studies of over
Reinforcement ratio Mpre/Mexp
160,000 tension-controlled sections, simplified yet rational design
Mean Standard deviation equations for predicting flexural capacity of the sections is derived
from multiple regression analyses.
0 < ρe < ρe, b (56 beams) 0.97 0.09
4. Comparisons between the predicted flexural capacities of FRP-PC
ρe, b ≤ ρe ≤ 1.5 ρe, b (14 beams) 0.98 0.11
beams and experimental results showed that the proposed approach
ρe > 1.5 ρe, b (13 beams) 0.91 0.10
provides consistent and accurate predictions. The predicted ultimate
ρe > 0 (83 beams) 0.96 0.10
flexural capacity is on average 0.96 of the experimental value, with

340
F. Peng, W. Xue
a standard deviation of 10%.
Acknowledgements
The authors gratefully acknowledge the financial support provided
by the Project of National Key R&D Plan of China (No.
2017YFC0703000), the National Natural Science Foundation (No.
51678433), Fundamental Research Funds for the Central Universities
(No. 0200219151) and Project of Shanghai Science Technology
Commission (No. 16DZ1201802).
References
[1] Nanni A, Tanigaki M. Pretensionded prestressed concrete members with bonded
fiber reinforced plastic tendons: development and flexural bond lengths (static). ACI
Struct J 1992;89(4):433–41.
[2] Lou T, Liu M, Lopes SMR, Lopes AV. Effect of bond on flexure of concrete beams
prestressed with FRP tendons. Compos Struct 2017;173:168–76.
[3] Barros JAO, Taheri M, Salehian H, Mendes PJD. A design model for fibre reinforced
concrete beams pre-stressed with steel and FRP bars. Compos Struct
2012;94(8):2494–512.
[4] Noël M, Soudki K. Effect of prestressing on the performance of GFRP-reinforced
concrete slab bridge strips. J Compos Constr 2013;17(2):188–96.
[5] CAN/CSA S806-12. Design and construction of buildings components with fiber-
reinforced polymers. Rexdale (ON, Canada): Canadian Standards Association; 2012.
[6] Grace N, Bebawy M, Kasabasic M. Evaluation and analysis of decked bulb T beam
bridge Report NoRC-1620 Southfield (MI): Lawrence Technological University;
2015.
[7] Karbhari VM. Use of composite materials in civil infrastructure in Japan WTEC
report Baltimore (Md): International Technology Research Institute, World
Technology (WTEC) Division; 1998
[8] Xue W, Peng F, Tan Y, Zhang S, Fu K, Yang Y. A review of studies on concrete
structures with FRP reinforcements in China. In: 7th int. conf. on advanced com-
posite materials in bridges and structures (ACMBS-VII), Vancouver, BC, Canada;
2016.
[9] Atutis M, Valivonis J, Atutis E. Experimental study of concrete beams prestressed
with basalt fiber reinforced polymers. Part I: flexural behavior and serviceability.
Compos Struct 2018;183:114–23.
[10] Yonekura A, Tazawa E, Nakayama H. Flexural and shear behavior of prestressed
concrete beams using FRP rods as prestressing tendons. Proc, int symp fiber re-
inforced polymer reinforcement for reinforced concrete structures, ACI special
publication 138. Farmington Hills (MI): American Concrete Institute; 1993. p.
525–48.
[11] Xue W, Wang X. Experiment and nonlinear analysis of concrete beams with bonded
prestressing CFRP tendons. China J Highway Transp 2007;20(4):41–7. (in Chinese).
[12] Liang Y, Sun C, Ansari F. Damage assessment and ductility evaluation of post ten-
sioned beams with hybrid FRP tendons. J Compos Constr 2011;15(3):274–83.
[13] Grace NF, Enomoto T, Baah P, Bebawy M. Flexural behavior of CFRP precast pre-
stressed decked bulb T-beams. J Compos Constr 2012;20(3):225–34.
[14] Grace NF, Ushijima K, Matsagar V, Wu C. Performance of AASHTO-type bridge
model prestressed with carbon fiber-reinforced polymer reinforcement. ACI Struct J
2013;110(3):491–501.
[15] Gar SP, Mander JB, Hurlebaus S. Deflection of FRP prestressed concrete beams. J
Compos Constr 2018;22(2):04017049.
[16] Forouzannia F, Gencturk B, Dawood M, Belarbi A. Calibration of flexural resistance
factors for load and resistance factor design of concrete bridge girders prestressed
with carbon fiber-reinforced polymers. J Compos Constr 2016;20(5):04015050.
[17] Xue W, Peng F, Zheng Q. Design equations for flexural capacity of concrete beams
reinforced with glass fiber–reinforced polymer bars. J Compos Constr
2016;20(3):04015069.
[18] ACI 440.4R-04. Prestressing concrete structures with FRP tendons. Farmington Hills
(MI): American Concrete Institute; 2004.
341
Composite Structures 204 (2018) 333–341
[19] ACI Committee 318. Building code requirements for structural concrete (ACI
318–14) and commentary. Farmington Hills (MI): American Concrete Institute;
2014.
[20] Mast RF. Unified design provisions for reinforced and prestressed concrete flexural
and compression members. ACI Struct J 1992;89(2):185–99.
[21] Mast RF, Dawood M, Rizkalla SH, Zia P. Flexural strength design of concrete beams
reinforced with high-strength steel bars. ACI Struct J 2008;105(4):570–7.
[22] Park H, Cho J. Ductility analysis of prestressed concrete members with high-
strength strands and code implications. ACI Struct J 2017;114(2):407–16.
[23] Shahrooz BM, Reis JM, Wells EL, Miller RA, Harries KA, Russell HG. Flexural
members with high-strength reinforcement: behavior and code implications. J
Bridge Eng 2014;19(5):04014003.
[24] AASHTO. (American Association of State Highway and Transportation Officials).
Bridge design specifications. Washington, DC; 2017.
[25] Burke CR, Dolan CW. Flexural design of prestressed concrete beams using FRP
tendons. PCI J 2001;46(2):76–87.
[26] Dolan CW, Swanson D. Development of flexural capacity of a FRP prestressed beam
with vertically distributed tendons. Compos Part B: Eng 2002;33(1):1–6.
[27] Grace NF, Singh SB. Design approach for carbon fiber-reinforced polymer pre-
stressed concrete bridge beams. ACI Struct J 2003;100(3):365–76.
[28] Bank LC. Composites for construction: Structural design with FRP materials. New
York: Wiley; 2006.
[29] ACI 440.1R-15. Guide for the design and construction of structural concrete re-
inforced with FRP bars. Farmington Hills (MI): American Concrete Institute; 2015.
[30] Vijay PV, GangaRao HVS. Bending behavior and deformability of glass fiber-re-
inforced polymer reinforced concrete members. ACI Struct J 2001;98(6):834–42.
[31] Kim YJ, Nickle RW. Strength reduction factors for fiber-reinforced polymer-pre-
stressed concrete bridges in flexure. ACI Struct J 2016;113(5):1043–52.
[32] Zadeh HJ, Nanni A. Reliability analysis of concrete beams internally reinforced with
fiber-reinforced polymer bars. ACI Struct J 2013;110(6):1023–32.
[33] Todeschini CE, Bianchini AC, Kesler CE. Behavior of concrete columns reinforced
with high strength steels. ACI J Proc 1964;61(6):701–16.
[34] Haldar A, Mahadevan S. Probability, reliability and statistical method in en-
gineering design. New York: John Wiley & Sons; 2000.
[35] Kakizawa T, Ohno S, Yonezawa T. Flexural behavior and energy absorption of
carbon FRP-reinforced concrete beams. Fiber reinforced plastic reinforcement for
concrete structures, ACI special publication 138. Farmington Hills (MI): American
Concrete Institute; 1993. p. 585–98.
[36] Abdelrahman AA, Rizkalla SH. Serviceability of concrete beams prestressed by
carbon-fiber-reinforced-plastic bars. ACI Struct J 1997;94(4):447–54.
[37] Saafi M, Toutanji H. Flexural capacity of prestressed concrete beams reinforced
with aramid rectangular tendons. Constr Build Mater 1998;12(5):245–9.
[38] McKay KS, Erki MA. Flexural behaviour of concrete beams pretensioned with
aramid fibre reinforced plastic tendons. Can J Civil Eng 1993;20(4):688–95.
[39] Lees JM, Burgoyne CJ. Experimental study of influence of bond on flexural behavior
of concrete beams pretensioned with aramid fiber reinforced plastics. ACI Struct J
1999;96(3):377–85.
[40] Morais MM, Burgoyne CJ. Experimental investigation of the ductility of beams
prestressed with FRP (June). 6th int conf on fiber reinforced polymers for reinforced
concrete structuresSingapore: World Scientific; 2003. p. 1013–22.
[41] Stoll F, Saliba JE, Casper LE. Experimental study of CFRP-prestressed high-strength
concrete bridge beams. Compos Struct 2000;49(2):191–200.
[42] Mertol HC, Rizkalla S, Scott P, Lees JM, El-Hacha R. Durability and fatigue behavior
of high-strength concrete beams prestressed with CFRP bars. ACI special publica-
tion SP245-1: case histories and use of FRP for prestressing applications.
Farmington Hills (MI): American Concrete Institute; 2006. p. 1–20.
[43] Fam A, Rizkalla S, Tadros G. Behavior of CFRP for prestressing and shear re-
inforcements of concrete highway bridges. ACI Struct J 1997;94(1):77–86.
[44] Zou P. Load-deflection response of high strength concrete beams pretensioned by
carbon fibre reinforced polymers. In: Seracino R, editor. FRP composites in civil
engineering—CICE 2004. London: Taylor & Francis; 2004. p. 781–91.
[45] Krem S. Bond and flexural behaviour of self consolidating concrete beams re-
inforced and prestressed with FRP bars (PhD thesis) Canada: University of
Waterloo, Waterloo; 2013.
[46] Kara IF, Ashour AF. Flexural performance of FRP reinforced concrete beams.
Compos Struct 2012;94(5):1616–25.