See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/304749283

Probabilistic assessment of FRP-confined reinforced concrete columns

Article  in  Composite Structures · July 2016

DOI: 10.1016/j.compstruct.2016.07.003

CITATIONS READS

7 328

3 authors:

Hassan Baji Hamid Reza Ronagh

Central Queensland University Western Sydney University
49 PUBLICATIONS   112 CITATIONS    215 PUBLICATIONS   2,038 CITATIONS   

SEE PROFILE SEE PROFILE

Chun Qing Li
RMIT University
175 PUBLICATIONS   1,622 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Corrosion-Induced Deterioration of Reinforced Concrete Structures View project

Seismic Performance of Innovative Structural Fuse for Concentric Braced Frames. View project

All content following this page was uploaded by Hassan Baji on 18 December 2017.

The user has requested enhancement of the downloaded file.

 Probabilistic assessment of FRP-confined reinforced concrete columns

Hassan Baji1*, Hamid Reza Ronagh2 and Chun Qing Li3

1
Research Fellow, School of Engineering, Civil, Environmental and Chemical
Engineering, RMIT University, 376 Swanston Street, Melbourne, Victoria, Australia
Email: [email protected]

2
Professor, Institute for Infrastructure Engineering, Western Sydney University, Locked
bag 1797, Penrith, NSW, 2751, Australia
Email: [email protected]

3
Professor and Head of School, School of Engineering, Civil, Environmental and
Chemical Engineering, RMIT University, 376 Swanston Street, Melbourne, Victoria,
Australia Email: [email protected]

Corresponding author: Hassan Baji

School of Engineering, Civil, Environmental and Chemical Engineering, RMIT University,
Melbourne, Victoria, 3001, Australia
Email: [email protected]
Phone: +61 3 99251800

1
Abstract: This paper presents a reliability study on reinforced concrete (RC) columns retrofitted by

Fibre Reinforced Polymers (FRP) confinement. First, variability of the compressive resistance of

FRP-confined RC columns considering uncertainty in the model error, material properties, section

dimensions, FRP strain efficiency factor and minimum load eccentricity as well as any existing

correlation among random variables was studied. Then, reliability of FRP-confined RC columns

designed based on the ACI 440 and the fib bulletin 14 guidelines was investigated. Results showed

that in order to achieve the target reliability used in design of conventional RC columns, resistance

reduction factors of these guidelines require modification. For the design of FRP-confined RC

columns, a resistance reduction factor of 0.85 for the ACI 440 guideline and a partial safety factor

of 1.45 for the fib bulletin 14 are suggested.

Keywords: Fibre reinforced polymers (FRP); Reinforced concrete columns; Reliability; Design

guidelines; Resistance reduction factor

1 Introduction

Vast asset of reinforced concrete (RC) infrastructure and structures such as buildings, road and

railway bridges and airports are aging. Most of these structures are exposed to long-term damage

from harsh environment such as exposure to chloride ions, carbon dioxide, sulphate dioxide, etc.,

and as such, they are subject to continuous deterioration resulting in reduction of load carrying

capacity. On the other hand, many of these structures were designed before the current design

provisions had been adopted. The substandard older designs results in structures with lower strength

than needed. Furthermore, these structures might become subjected to increasing demand during

their operating lifetime. One apparent solution to deal with deterioration and increasing demand is

to replace the infrastructure but this is very costly and more importantly is not sustainable due to

increasing scarcity of resources. The most viable and practical solution for maintaining safe

operation of these structures is to restore and/or increase their capacity through

retrofitting/strengthening. Traditional techniques of retrofitting RC structures include bonding of

2
steel plates and post-tensioning. Due to their excellent mechanical and chemical properties, ease

and efficiency of application and relatively low cost, strengthening using Fibre Reinforced

Polymers (FRP) is emerging as an alternative to the traditional methods and is becoming an

increasingly popular retrofit technique among researchers and engineers worldwide. In addition to

strengthening, external FRP retrofitting offers the advantage of protection against external attack of

damaging agents such as chloride.

One major application of FRP materials to structural retrofit is in strengthening RC columns. In the

last decade or so, the retrofit of RC columns by wrapping them with FRP sheets has gained

increasing popularity in the engineering community [1-5]. A considerable number of researchers

have investigated the behaviour of FRP-confined concrete and proposed constitutive models [6-16].

A comprehensive review of models developed to predict the axial stress-strain behaviour of FRP-

confined concrete in circular sections can be found in Ozbakkaloglu et al. [17]. On the other hand, a

number of design guidelines [18-23] have been developed for external strengthening of RC

structures including the RC columns. Nonetheless, very few reliability-based studies in this area can

be found in the current literature. With rapid developments in the research and field applications of

FRP in strengthening RC columns, there is a growing need for establishing a probability-based

design methodology for the design of external strengthening of RC structures.

Val [24] studied the reliability of short columns with circular cross-sections under static loads. The

ultimate tensile strength of FRP was treated as a random variable. However, other properties of FRP

material, such as the thickness and the modulus of elasticity, were treated as deterministic variables.

Experimental data from different sources were used to calibrate the model error in predicting the

strength and strain at the ultimate state. For a target reliability of 3.8, a confinement ratio-dependent

resistance reduction factor smaller than 0.75 was proposed. Bisby et al. [25] made a comparison

among performance of confinement models for FRP-confined columns. Without providing details

of the reliability analysis and including the effects of steel reinforcement and model error, they

reported that the reliability of strengthened columns designed based on the ACI guidelines is low,
3
when compared to the target reliabilities available in the literature. Zou and Hong [26] assessed the

reliability level of FRP-confined RC columns that had been designed according to the Canadian

design codes. They considered the uncertainty in the concrete, steel and FRP materials, as well as

the uncertainty in the model error. Based on a target reliability index of 3.5, they suggested an FRP

resistance reduction factor of 0.75. Casas and Chambi [27] investigated the reliability of CFRP-

confined bridge piers designed based on the European codes. Partial safety factors dependent on the

thickness of the FRP wrap were proposed. They also concluded that partial safety factors proposed

by the European codes were conservative. In a recent study, Wang and Ellingwood [28] studied

reliability of FRP-confined bridge piers designed based on the AASHTO requirements and the ACI

440 guideline [18]. It was reported that in order to achieve a target reliability of 4.5 in design of

FRP strengthened RC bridge columns, a resistance reduction factor of 0.65 is required.

The aim of the present study is to evaluate the reliability of FRP-confined RC columns subjected to

concentric axial loading. There are several reasons to conduct further reliability studies on FRP-

retrofitted RC columns. Firstly, previous studies on retrofitted RC columns did not consider the

model error or assumed it similar to that of ordinary RC columns. In this study, based on the

available experimental data on FRP-confined RC columns, the uncertainty associated with the

strength model is estimated. Secondly, although variability in composite material properties has

been considered in the abovementioned reliability studies, variability in the FRP strain efficiency

factor has not been fully incorporated in the reliability analysis. Detailed statistical analysis leading

to an appropriate probability density function for this factor is performed in this research. Thirdly,

comparison between reliability levels provided by the current design guidelines has not been

investigated. Provisions of both the ACI 440 [18] and fib bulletin 14 [19] guidelines are subject of

investigation in this paper. Lastly, no study on the reliability of FRP-confined RC columns has

considered variability in the axial load eccentricity, and this study considers this variability. The

uncertainty in the model error as well as other uncertainties associated with material properties,

loads and correlation among random variables is taken into account in the reliability analysis.

4
Furthermore, standard code calibration procedure is employed for calibrating the resistance

reduction factors of the considered design guidelines.

2 Strength model for FRP-confined concrete

For a circular concrete cylinder of diameter D, as shown in Fig. 1, as the axial stress increases, the

corresponding lateral strain increases and the confining jacket develops a tensile hoop stress,

balanced by a uniform radial pressure, which reacts against the concrete lateral expansion. The FRP

wraps effectively curtail the lateral expansion of the concrete as it reaches the unconfined strength

level, and then reverse the direction of the volumetric response as the concrete responds via large

stable volume contraction. Using the force equilibrium and linearity of FRP materials, the lateral

(radial) confining pressure acting on the concrete core at the ultimate state (rupture of FRP jacket),

fr, is given by:

2 fu t f 2 E f  fu t f
fr    E j fu (1)
D D

where,  fu and t f are the ultimate strength and thickness of the FRP jacket, respectively. E f and

 fu denote the Young’s modulus and the ultimate strain of the FRP jacket. Ej in Eq. (1) represents

the lateral stiffness of the FRP jacket, which controls the relation between stress and strain in the

FRP jacket.

Fig. 1.

Assuming a perfect bond between the FRP jacket and the concrete core, the hoop strain in the FRP

shell is equal to that of the concrete core. The ultimate state is characterised by the tensile rupture of

the wrap. However, it has been observed that, at failure, tensile strength of the FRP jacket is

generally lower than the uniaxial tensile strength of the FRP obtained from coupon tests [8, 29]. To

address the difference between the actual rupture strain and the ultimate strain of FRP material,

Lam and Teng [8] proposed the use of the term “actual confining pressure” (at failure state), here

5
denoted by f r ,rup , by replacing the ultimate strain of the FRP jacket,  fu , with the actual strain of

the FRP jacket (at failure state),  f ,rup , as shown in Eq. (2).

2 E f  f ,rupt f 2 E f k  fu t f
f r ,rup   (2)
D D

Results from different tests show that the so called ‘‘strain efficiency factor’’, k , given by the ratio

 f ,rup /  fu varies for different types of FRP materials. More discussion on this factor and its

probabilistic evaluation will be presented in Section 4.4. It is also worth noting that at failure the

ultimate strain of the FRP jacket, εfu, provided by the manufacturer is less than that obtained from

the coupon tensile tests.

Linear as well as nonlinear models have been proposed for predicting the ultimate strength of FRP-

confined concrete. However, due to their simplicity and their adequate accuracy, linear models are

more convenient. These linear models generally follow the original form suggested by Richart et al.

[30]. Eq. (3) shows the general format of the linear model.

2k  fu t f
f cc'  f c'  kf r ,rup  f c'  k (3)
D

In Eq. (3), k is the confinement effectiveness coefficient. The ACI 440 guideline [18] proposes a k

factor equal to 3.30. A value of 0.55 is also suggested for the FRP strain efficiency factor, kε. In

order to assess the confinement model shown in Eq. (3) and calibrate the k factor, a comprehensive

set of test results for the axial compressive resistance of FRP-confined circular plain concrete

specimens with Carbon Fibre Reinforcement Polymer (CFRP) and Glass Fibre Reinforcement

Polymer (GFRP) confinement is collected in this study [3, 12, 16, 29, 31-87]. Only experimental

data that had a complete set of information for the ultimate strength and the FRP strain efficiency

factor were considered. The selected database contains 658 test results, including significant test

programs performed in recent years. Only specimens with strength enhancement are included in the

database. Fig. 2 illustrates the distribution of concrete compressive strengths (unconfined

6
specimens), diameter of specimens, normalised lateral pressure and the normalised jacket lateral

stiffness for the collected database. It should be noted that the repetition of concrete compressive

strengths for specimens with different numbers of FRP layers is considered in the frequency graph

shown in Fig. 2. The considered database covers an appropriate range, with a majority of the cases

in the practical range.

Fig. 2.

If a regression analysis is used to fit the data to a line relating the axial and the lateral stresses, the

slope of this line represents the k coefficient. In the regression analysis that leads to the

determination of k, the fitting line is required to pass through the point representing the uniaxial test

results of plain unconfined concrete. Fig. 3 illustrates all of the experimental data and the fitted line

for the k coefficient as well as the line representing the ACI 440 model. The resultant best-fit model

with k = 3.25 is very close to the model proposed by the ACI 440 guideline [18]. Therefore, for

consistency with provisions of the ACI 440 guideline, k coefficient of 3.30 will be adopted in the

reliability analysis. It should be noted that, as Fig. 3 shows, the accuracy of the linear model for

predicting the ultimate strength of the FRP-confined concrete is adequate. In terms of the

coefficient of determination, R2, there is very slight difference between the linear and nonlinear

models.

Fig. 3.

In the subsequent sections, using the model error, performance of this linear strength model in

predicting the ultimate resistance of FRP-confined RC columns will be assessed.

3 Design of FRP-confined RC columns

With the rapid advancement in the area of strengthening RC members using FRP material, different

design guidelines have been developed worldwide. Some of the widely used international design

guidelines with the corresponding confinement models for the FRP-confined concrete are shown in

Table 1. For the sake of consistency, the confinement models shown in Table 1 follow the notation
7
used in this study and it may be different from the notation used in the original design guideline

document. A review on performance of confinement models proposed by these design guidelines

can be found in Guler and Ashour’s study [88].

Table 1.

ACI 440 [18] and the fib bulletin 14 [19] guidelines have different design philosophy in applying

the resistance reduction factors. In terms of method of applying the safety factors, other design

guidelines, some of which shown in Table 1, follow the safety format of either of these guidelines.

In this paper, structural reliability of concentrically loaded FRP-confined RC columns design based

on these two design guidelines is investigated.

The ACI 440 guideline [18] applies the resistance reduction factor to the behavioural effect (here,

compression). The ultimate load case considering the dead and live loads specified by ACI 318

design code [89] is determined by,

c  PRn   1.20PDLn  1.60PLLn (4)

where PDLn and PLLn are the nominal dead and live load effects and PRn is the nominal compressive

resistance. The c factor is the resistance reduction factor for compression, which is equal to 0.75

for RC columns with spiral steel confinement and 0.65 for columns with closed steel ties. The factor

α accounts for minimum load eccentricity due to geometric imperfections and is 0.85 for columns

with spiral steel confinement and 0.80 for columns with closed steel ties. An alternative to applying

this factor is to apply the minimum load eccentricity directly. According to the ACI 318

commentary, a reduction factor of 0.85 in spirally confined columns is equivalent to normalised

minimum load eccentricity, emin / D, of 0.05, and 0.10 for that of columns with closed ties. In this

study, the latter method is adopted for calculating the compressive resistance of RC columns.

Concrete, steel and FRP materials contribute to the nominal compressive resistance, PRn , which is

calculated as follows,

8
PRn  0.85 f cc'  Ag  As   As f y (5a)
2k  fu t f
f cc'  f c'  f 3.30 (5b)
D

where Ag and As are the gross area and reinforcement area of column cross-section, respectively.

The reduction factor 0.85 in Eq. (5a) is a concrete stress block parameter accounting for the

difference between the standard cylindrical test and in-situ concrete strength. Contribution of FRP

material in the compressive resistance shown in Eq. (5a) is hidden inside the ultimate strength of

confined concrete, f cc' . Eq. (5b) shows the formula for calculating this ultimate resistance of FRP-

confined concrete according to the ACI 440 guideline [18]. All the variables in Eq. (5b) are

previously defined. The factor ψf accounts for variability in the FRP material, and its probabilistic

calibration is the main goal of this research. Value of 0.95 is currently recommended by the ACI

440 guideline.

Design provisions of fib bulletin 14 [19] are in line with the general design procedure of the fib

Model Code 2010 [90]. It should be noted that the fib Model Code 2010 uses different notations

than the ACI 318 code. However, for consistency, the ACI notation is used throughout this paper.

Eq. (6) shows the design of concentrically loaded RC column based on the basic gravity load

combination.

 PRn   1.35PDLn  1.50PLLn (6)

Safety factors of the fib Model Code 2010 are directly applied onto the material properties. The

minimum load eccentricity due to geometric imperfections, as shown in Eq. (7), has to be taken into

account.

 
0.00167l0  emin  max 0.005 l0 , D / 30  0.00250l0 (7)

where, l0 is the effective length of RC column in metres, which can be conservatively considered as

the column height for braced frames under gravity loads; D is the cross-section diameter of column

and emin is the minimum load eccentricity.

9
It should be noted that the process of calculating the ultimate strength of FRP-confined concrete,

f cc' , in the fib bulletin 14 [19] is based on an iterative procedure using Mander’s confinement

model [91]. For consistency in comparing this guideline with the ACI 440 guideline [18], same

linear strength model, shown in Eq. (5b), is adopted for calculating the ultimate compressive

strength of FRP-confined concrete. Thus, the nominal compressive resistance of FRP-confined RC

column based on fib bulletin 14 is calculated as follows,

PRn  f cc'  Ag  As  
As f y
(8a)
s
f c' 2k  fu t f
f cc'   3.30 (8b)
c fD

In Eqs. (8a) and (8b), γs, γc and γf are the material partial safety factors for steel, concrete and FRP

materials, respectively. According to the fib Model Code 2010 [90], factors γs and γc are 1.15 and

1.50, respectively. The fib bulletin 14 relates the γf factor to FRP material type, e.g. Carbon, Glass,

or Aramid polymers and method of application. For instance, for application of wet lay-up systems

under normal quality control, the value of 1.35 is suggested for the CFRP. This factor plays a role

similar to that of ψf in the ACI 440 guideline [18].

4 The probabilistic procedure

4.1 Case study

The case study considered in this paper is a deficient concentrically loaded circular RC column of a

building that only carries dead and live loads. This is a common case in buildings with shear walls

as the solo lateral resisting system, in which columns are only designed to withstand the gravity

loads. It is assumed that due to deterioration or substandard design, the column does not have the

adequate capacity to resist the applied load and therefore is in need of strengthening. The FRP

confinement strengthening increases the capacity, restoring the reliability to an acceptable level. For

each main random variable, different cases are considered. The cases include three cross section

diameters (300 mm, 400 mm and 500 mm), three concrete compressive strengths (25 MPa, 35 MPa

10
and 45 MPa) and three reinforcement percentages (0.01, 0.02 and 0.03). For all cases, cover of

concrete to rebar centre is 50 mm. The axial load eccentricity due to geometric imperfection is set to

0.05 of the cross section diameter. Furthermore, a wide range of live to dead load ratios and FRP

confinement levels is used in the reliability analysis. The general shape of the considered FRP-

confined RC column is shown in Fig. 4.

4.2 Reliability analysis methodology

The performance of a structure can be represented by a limit-state function, also known as a

performance function, g. In the current study, preventing the ultimate failure of FRP-confined RC

columns due to concentric axial load is the main goal of establishing the limit state. In Eq. (9), the

limit state, g, containing the effect of dead and live loads is shown.

g   PR  PDL  PLL (9)

where PR is the actual compressive resistance of RC column and PDL and PLL are dead and live

load effects, respectively. ξ in Eq. (9) represents the model error. The axial compressive resistance

of the FRP-confined RC column depends on the geometric dimensions and material properties. Fig.

4 illustrates derivation of the axial compressive resistance of the column with geometric

imperfections schematically. In this figure, PR0 is the maximum compressive resistance of the RC

column without effect of the minimum load eccentricity.

Fig. 4.

As is shown in Fig. 4, the minimum normalised load eccentricity is taken as 0.05 in this study, and

in the absence of the probability density function for the eccentricity due to geometric

imperfections, the actual eccentricity is modelled as a uniformly distributed variable with minimum

and maximum values of 0.0 and 0.05, respectively. Using the Uniform distribution for probabilistic

modelling of geometric imperfection has been adopted by other researchers [92] as well. In deriving

the capacity curve for the column section, the standard procedure based on the mechanics of RC

structures considering force equilibrium and strain compatibility is utilised [93]. Elastic-perfectly-

11
plastic behaviour is assumed for the reinforcement steel material and the code-specified (here ACI

318 [89] or fib Model Code 2010 [90]) concrete stress block parameters are used for calculating

concrete force in the cross section.

The reliability index β can be defined as a measure of the probability that g (as a function of basic

random variables x) is less than zero,

p f  P  g (x)  0       (10)

In Eq. (10),  function is the cumulative standard normal distribution function. The reliability

index can be calculated using different methods. In this study, the Advanced First Order Second

Moment (AFOSM) method [94] is employed. AFOSM method takes the first two statistical

moments of a linear approximation of the limit state function, which is a function of vector of

random variables (x). The limit state used in this study, shown in Eq. (9), contains three basic

random variables that are the resistance and the dead and live load effects. The reliability index, β,

is the minimal distance from the origin to the tangent hyper-plane representing the linearized limit

state function in the standardised space in which all the variables are transformed into standard

normal space. Mathematical representation of the procedure is shown in Eq. (11). The output of this

optimisation problem is x*, and the reliability index is β = (x*x*T)1/2.

E  g (x) E  g (x* )  ( x  x * )T G* 
Min 
x
VAR  g (x) G*T Cx G* (11)
subject to g (x* )  0

where the point x*, called the design point, is the point at which the limit state function is

linearized. E and VAR show the mean and standard deviation operators. Cx is the covariance matrix

of the basic random variables and μx is the vector of mean of basic random variables. G* is the

gradient vector of the limit state function evaluated at the design point. Several algorithms have

been proposed to solve this problem. The reader is referred to classic references e.g. [94] for more

details on solution of this problem.

12
In summary, the procedure of calculating the reliability index (β), shown as a flowchart in Fig. 5,

can be summarised as follows:

(i) Prepare the input such as cross-sectional diameter and material properties and live to dead load

ratio;

(ii) Set a value for the f r / f c' ratio;

(iii) Using the fr/f’c ratio, design the required FRP jacket thickness based on Eqs. (4), (5a) and (5b)

for the ACI 440 guideline [18], and (6), (8a) and (8b) for the fib bulletin 14 [19];

(iv) Find the best-fit Lognormal distribution for the actual compressive resistance of FRP-confined

RC columns, PR (see Fig. 4), using probabilistic models taken from the literature and those

obtained in this study and Monte Carlo simulation technique [94];

(v) Calculate the reliability index, shown in Eq. (10), using the Advanced First Order Second

Moment (AFOSM) method [94];

Fig. 5.

The resulted reliability indices are compared against the acceptable reliability index known as the

target reliability index. The target reliability index depends on the consequences of failure and cost

of increasing/decreasing the safety margin by a unit (marginal cost of safety). The range in which

target reliability indices vary, changes due to many factors such as the importance of the structural

member, the expected mode of failure, the ratio of live loads to dead loads, and so on. It has been

suggested that the target reliability index for members that fail suddenly with little warning, such as

FRP-confined RC columns, should be increased [95]. According to the fib Model Code 2010 [90],

based on moderate cost of safety measure and great consequence of failure, the appropriate target

reliability index is 3.8. In this study, the reliability indices corresponding to the current design

practice of ordinary RC columns are used as the reference in the calibration process, under the

assumption that rehabilitated buildings should have comparable reliability to new buildings.

Szerszen and Nowak [96] showed that, based on the gravity load combinations and resistance

13
reduction factor of 0.75, cast in-place RC columns with spiral, designed based on the ACI 318 code

[89] have a reliability index of above 4.0. Therefore, in the reliability analysis of this study, a target

reliability index of 4.0 will be adopted.

4.3 Uncertainty modelling

Structural members in need of strengthening, due to either deterioration or substandard design, are

in general relatively old. Therefore, in any reliability analysis that deals with safety of these

members, using old statistical data is more appropriate. In this research, statistical models for

concrete and reinforcement steel materials and sectional dimension are taken from the old literature,

while those of the FRP material are taken from recent studies. Probabilistic models for the model

error and the FRP strain efficiency factor are derived in this study and will be discussed in the

subsequent sections.

The statistical model for column cross-section diameter, D, and the effective depth, d, are taken

from Israel et al. [97] study. The Normal distribution is proposed for representing the probabilistic

distribution of these variables.

A general probabilistic model for the in-situ concrete compressive strength i.e. actual strength of in-

place concrete cast in structural members consists of two parts: the first part accounts for variability

in the standard cylinder compressive strength, while the second part deals with variability in

concrete strength due to age, consolidation and curing of in-place concrete [98-100]. The model

calibrated by Bartlett and MacGregor [100] was used for probabilistically representing the in-situ

concrete compressive strength.

Probabilistic models for yield strength and Young’s modulus of steel reinforcement are taken from

Mirza and MacGregor study [101]. The nominal yield strength of steel reinforcement is set to 424

MPa. This type of steel corresponds to G60 steel in the ACI 318 code [89] or S400 steel in the fib

Model Code 2010 [90].

14
Limited statistical data for the properties of wet lay-up FRP are available in the literature. One of

the most comprehensive probabilistic studies on the mechanical properties of wet lay-up FRP

composites was conducted by Atadero [102]. Statistical models from that study, as shown in Table

2, are used in the reliability analysis. Atadero and Karbhari [103] also reported that the correlation

between thickness and strength, strength and modulus of elasticity, and thickness and modulus of

elasticity were -0.517, 0.313 and -0.362, respectively. These correlation coefficients are used in the

reliability analysis.

A summary of all of the main random variables used in the reliability analysis is given in Table 2.

Dn and dn in this table show the nominal values of the cross-section diameter and effective depth,

respectively.

Table 2.

The statistical models for dead and live loads were taken from Ellingwood et al. [104] study. As

previously discussed, the normalised load eccentricity of column is modelled as a uniformly

distributed variable within the range of 0.00 to 0.05.

In what follows, with the aid of large experimental databases, probabilistic models for the FRP

strain efficiency factor and the model error, kε and ξ, are derived.

4.4 Probabilistic model for the FRP strain efficiency factor

Reliable strength model for the FRP-confined concrete is highly influenced by the FRP strain

efficiency factor. Therefore, in order to evaluate effective confining pressure exerted by the jacket,

appropriate model for this factor is required. Although, many researchers have observed that the

actual rupture strain of FRP is less than the strain measured in standard coupon tests [8, 14, 16, 29,

43, 51, 105], causes of variation in this factor have not been fully understood. Pessiki et al. [29]

related this phenomenon to the localisation in the cracked concrete, which leads to non-uniform

distribution of stress in the FRP jacket. In an experimental investigation, Lam and Teng [55]

identified the following three factors to be causing the low rupture strain in FRP cylinders: the non-
15
uniform deformation of cracked concrete; the existence of an overlapping zone in which the

measured strain are much lower than the strains obtained elsewhere; and the curvature of the FRP

jacket. In a contradictory conclusion, Wu and Jiang [81] claimed that the curvature of FRP jacket

does not significantly reduce the efficiency factor. They stated that small FRP strain efficiency

values were attributed to differences between the strain measurement systems used for the FRP

jacket and the flat coupon, and differences in the workmanship. Although, inefficiency of the strain

measurement systems is definitely an issue as was highlighted in Bisby and Take’s study [69], who

used digital image correlation technique for investigating variation of hoop strain within the FRP

jacket, the workmanship factor is an inherent part of uncertainty in any material application and it

should be dealt with through a probabilistic-based approach. Shahawy et al. [43] suggested that the

issue of the FRP strain efficiency factor is a probabilistic problem and that, for design purposes,

proper confidence levels must be set by the reliability analysis of the effective hoop rupture strain of

the jacket. In this study, the focus is on statistically evaluating the randomness in this factor.

Mechanical evaluation of this factor is beyond the scope of this paper.

A full probabilistic study on variability of the FRP strain efficiency factor is carried out in this

research. The statistical analysis of kε is performed by using the collected database. As is shown in

Figs. 6a and 6b, the general trend is that the FRP strain efficiency factor is smaller than 1.0.

However, in 49 tests, comprising less than 7.5% of the database, values larger than 1.0 are reported.

These values seem to be unacceptable. De Lorenzis and Tepfers [105] attributed these high values

to the scatter in the FRP tensile strength and in the strain measurement. Close look at the data shows

that only 14 out of 60 sources contain data with FRP strain efficiency factor greater than 1.0

(Mathyas et al. [39], 2 of 3; Saafi and Toutanji [12], 6 of 6; Kshirsagar et al. [42], 2 of 3; Ilki et al.

[49], 2 of 6; Harries and Kharel [52], 2 of 8; Carey and Harries [57], 1 of 2; Ciupala et al. [62], 1 of

4; Cui [70], 18 of 48; Xiao et al. [76], 1 of 9; Elsanadedy et al. [77], 1 of 6; Liang et al. [78], 1 of

12; Wu and Jiang [81] 9 of 34; Lim and Ozbakkaloglu [86], 1 of 18; Wu et al. [87], 2 of 7). Four

references [12, 42, 70, 81] contain 71% of these data. This interesting statistics shows that high

16
values of the FRP strain efficiency factor have systemically appeared in some of the collected

sources and confirms the effect of the measurement system (used in each experimental program) on

the evaluation of the FRP strain efficiency factor. Investigation into the effect of the measurement

system and the possibility of having values greater than 1.0 for the FRP efficiency factor needs

more research and is beyond the scope of this paper. Therefore, in what follows FRP strain

efficiency factors greater than 1.0 will not be used in the probability density function fitting.

Exclusion of values greater than 1.0 has been adopted by other researchers [9, 106] and it brings

more conservatism in the design procedure.

Fig. 6.

Level of FRP confinement, represented by the fr/f’c ratio as shown in Fig. 6b, is another influential

factor on the FRP strain efficiency factor, which is not extensively researched in the current

literature. Most of the data with high FRP strain efficiency factor are in the region with low

confinement ratio. About 70% of the data with FRP strain efficiency factor greater than 1.0 lay in

the range of 0 to 0.50 of the fr/f’c ratio and nearly all the data with fr/f’c greater than 1.0 have FRP

strain efficiency factor in the range of 0.20 and 1.0. This might be due to cluster of experimental

data in the low range of fr/f’c ratio. Visual inspection of the scatter of data in Figs. 6a and 6b also

indicates that the FRP strain efficiency factor is slightly sensitive to the type of FRP material. Based

on the entire database (without removing data with FRP strain efficiency factor greater than 1.0) the

mean values of the FRP strain efficiency factor for CFRP- and GFRP-confined specimens are 0.67

and 0.59, and, the coefficients of variation are about 0.35 and 0.39, respectively, indicating slightly

higher variability in the GFRP-confined type.

In the absence of strong cause and effect relationship between the FRP strain efficiency factor and

the mechanical properties of concrete and FRP jacket, correlation coefficient would be a useful

measure of relating these variables. Nonetheless, other researchers have proposed some empirical

expressions relating the FRP strain efficiency factor to the concrete compressive strength and

mechanical properties of FRP jacket such as Young’s modulus and ultimate tensile strength [10, 70,
17
85, 107]. Due to weak correlation between these variables, as shown in Fig. 7, these types of

expressions would not be reliable, and the strategy of using correlation coefficients, followed in this

research, would be more appropriate.

Fig. 7.

Considering the upper and lower bounds of the FRP strain efficiency factor, bounded or truncated

probability density functions are the most appropriate distribution functions to fit to the

experimental data. On the other hand, as Fig. 8 shows, the histogram of test data for FRP strain

efficiency factor has clear left skewness. Therefore, probability density functions bounded in the [0,

1] region with negative skewness would be more appropriate for fitting to the FRP strain efficiency

factor data. The Beta distribution is the most common bounded distribution used for representing

the bounded data, and it has been used for fitting yield and ultimate strength of mild steel [108].

This distribution is also capable of handling the negative skewness. The cumulative function for this

distribution is given below.

B  x; a, b  x
FX  x   ; a, b  0; 0.0  x  1.0; B  x; a, b    t a1 (1  t )b1 dt (12)
B 1; a, b  0

where a and b are the shape parameters of the Beta distribution. Fx denotes the cumulative

distribution function of variable x and B is the incomplete beta function.

The upper truncated Weibull distribution is on the other hand suitable for modelling the data with

left skewness and is truncated in the upper tail side. The cumulative distribution function for this

distribution is given below.

   x a  
FX  x   c 1  exp       ; a, b  0 ; 0.0  x  1.0 (13)
   b  
 

where a and b are the shape parameter and scale parameter, respectively. The c parameter is an

adjustment required for truncating the distribution from the upper tail i.e. to ensure that the area

under the probability density function is 1.0.

18
The fitted distributions, as shown in Fig. 8, are very close. Nonetheless, the truncated Weibull

distribution better fits distribution of the data.

Fig. 8.

In order to determine which distribution best fits the data, three different goodness-of-fit measures

that are Pearson’s Chi-square, Kolmogorov–Smirnov and Anderson-Darling [109] are used. The

Pearson’s Chi-square test is the simplest measure. The Anderson-Darling is the enhanced version of

the Kolmogorov–Smirnov that gives more weight to the tails. As Table 3 shows, both of the tested

distributions can be considered as acceptable fit to the data based on 95% confidence level. To

determine the best fit, the different goodness-of-fit measures are compared to each other, with the

lowest test statistic indicating the best fitting distribution. As is evident from the results shown in

Table 3, the truncated Weibull distribution is superior to the Beta distribution.

Table 3.

In summary, the truncated Weibull distribution with the mean and coefficient of variation shown in

Table 3 and the cumulative distribution function shown in Fig. 8 will be used in the reliability

analysis. It is worth noting that although Realfonzo and Napoli [10] considered only FRP strain

efficiency factors smaller than 1.0 in their statistical analysis, they did not use the truncated Weibull

distribution.

4.5 Calibration of the model error

Model error covers the uncertainties in the mathematical modelling of a structure, where those

uncertainties arise from the necessary idealisation of different parts of the structure. In what

follows, a probabilistic model for resistance is sought. As for the problem under consideration, the

mathematical definition of the model error can be defined as.

PR,test
 (14)
PR,model

19
In Eq. (14),  shows the model error. PR,model is the compressive resistance predicted by model and

PR,test is the compressive resistance obtained from test. In order to calibrate the model error,

experimental results on FRP-confined RC columns are required. The existing literature was

extensively surveyed and a total of 91 FRP-confined RC columns (82 confined with CFRP material

and 9 confined GFRP material) were collected to evaluate statistics of the model error. Table 4

shows the collected database. All these columns are tested to failure under concentric compression

and the rupture of FRP jacket controls the ultimate compression resistance.

Table 4.

The compressive resistance of FRP-confined RC columns is calculated as shown in Eq. (15). This

expression is similar to expressions used for calculating the nominal capacity but it does not contain

any resistance reduction factor.

 2k  fu t f    D 2 
PR ,model   f c'  3.30   As   As f y (15)
 D  4 

As Table 4 shows, the statistical mean of ξ is 1.12, and the coefficient of variation is 0.15. Scatter of

the model versus test values for compressive resistance of columns, shown in Fig. 9, indicates that

the model prediction reasonably correlates with the experimental results. A mean value of 1.12 for

the model error indicates that the model underestimates the actual value by 12%.

Fig. 9.

Distribution fitting based on the Anderson-Darling goodness-of-fit measure shows that within the

probability density functions with positive domain, the Lognormal distribution, with mean and

coefficient of variation of 1.12 and 0.15 respectively, provides the best fit to the model error data,

and as such this distribution will be used in the probabilistic evaluation of compression resistance of

FRP-confined RC columns.

20
5 Statistics of the compressive resistance

The actual compressive resistance of FRP-confined RC column with circular section can be

expressed as shown in Eq. (16), in which the factor  accounts for the load eccentricity due to

geometric imperfection and  is the model error. The rest of the variables in Eq. (16) that are

related to material properties and sectional dimensions of the section are previously defined.


 2k  fu t f    D 2  

 PR     f c'  3.30   As   As f y  (16)

 D  4  


The Monte Carlo simulation technique can be used for deriving the probabilistic distribution for the

compressive resistance. In each simulation run, 10000 samples are generated. As is shown in Fig.

10, by having the lowest Anderson-Darling goodness-of-fit measure, the Lognormal distribution

provides the best-fit distribution for the compressive resistance of FRP-confined RC columns.

Therefore, in all reliability analyses, the best-fit Lognormal distribution is used for the compressive

resistance. Results in Fig. 10 are based on concrete compressive strength of 25 MPa, reinforcement

percentage of 0.01, steel yield strength of 424 MPa and section diameter of 300 mm. The fr /f’c ratio

also equals 0.50. Furthermore, by dividing the actual compressive resistance resulted from Eq. (16)

to the nominal compressive resistance (calculated based on nominal values of the random

variables), the compressive resistance is normalised.

Fig. 10.

In total, ten random variables are involved in calculating the compressive resistance of FRP-

confined RC columns. The variable α in Eq. (16) depends on the minimum load eccentricity, emin,

and as Fig. 4 shows, using the bending moment and axial force interaction diagram it can be

calculated. Therefore, the minimum load eccentricity influences the compressive resistance of RC

column through the variable α. A sensitivity analysis based on the Monte Carlo simulation results

and the Pearson’s correlation coefficient is performed to find out sensitivity of the compressive

resistance to each of the random variables. As is seen in Fig. 11, the FRP strain efficiency factor, kε,

21
is the most influential variable. The results in Fig. 11 are based on the same case used for finding

the best-fit distribution of the compressive resistance in Fig. 10.

Fig. 11.

Effects of FRP strain efficiency factor and ultimate strength of FRP jacket are more pronounced for

higher fr /f’c ratios. On the other hand, sensitivity of the compressive resistance to all other random

variables, especially concrete strength, decreases with an increase in the fr/f’c ratio. This is attributed

to the increasing contribution of the FRP jacket in carrying the axial load. Compressive resistance

of RC columns with low eccentricity is less dependent on mechanical properties of longitudinal

reinforcement. On the other hand, variability of mechanical properties of reinforcement (variables

As, Es and fy) is relatively low. Therefore, as is expected, sensitivity of compressive resistance to

these variables is not high.

The mean and coefficient of variation of the resulted normalised strength are used for assessing

variability of this variable. In total, 27 cases (three section dimensions, three concrete compressive

strengths and three longitudinal reinforcement percentages, refer to the case study in Section 4.1)

are considered. Fig. 12 shows the range of the mean and coefficient of variation for the considered

cases. Both the coefficient of variation and the mean decrease as the fr /f’c ratio increases. This is

attributed to the growing contribution of the FRP material. Due to high uncertainty in the

mechanical properties of FRP jacket, increasing contribution of the FRP confinement leads to

higher variability in the compressive resistance.

Fig. 12.

Addition of the FRP jacket with fr/f’c, of 3.0 increases the mean and the coefficient of variation by

about 12% and 68%, respectively. It will be seen in the next section that higher mean and

coefficient of variation for the normalised compressive resistance will result in lower reliability

indices for the strength ultimate limit state. Therefore, it is expected that for higher fr/f’c ratios,

lower resistance reduction factors or higher partial safety factors are required. According to the

22
results, the mean of strength is at least 50% higher than the nominal value (see Fig. 12a). This is

attributed to bias factors of greater than 1.0 for some input variables such as the yield strength of

steel reinforcement and ultimate strength of FRP jacket (see Table 2). The nominal value of

resistance (corresponding to normalised resistance of 1.0) is generally located in the lower tail of

the probability density function as shown in Fig. 13. Results in this figure are based on diameter of

300 mm, reinforcement percentage of 1% and concrete compressive strength of 25 MPa.

Fig. 13.

As is seen in Fig. 13, the chance of having strength values higher than the nominal value is more

than 90%.

6 Reliability assessment and calibration

The strengthening design of the deficient or deteriorated RC columns requires calibrating the

resistance reduction or partial safety factors. As discussed previously, the main assumption is that

because of the deterioration process, the RC column has lost a certain amount of axial strength.

Therefore, the FRP jacket is designed (the thickness of the jacket) to restore the RC column to an

acceptable condition i.e. the reliability index is restored to an acceptable level. It is assumed that the

acceptable reliability level corresponds to the reliability of an ordinary RC column without

strengthening. Different FRP confinement levels designated with different fr/f’c ratios are

considered for the reliability analysis. The case with zero fr/f’c ratio is identical with the design of a

conventional RC column, and its reliability can be set as a reference value in the calibration

procedure. Design provisions of both the ACI 440 guideline [18] and fib bulletin 14 [19] are subject

of investigation. Nonetheless, the general parametric study results are only presented for the ACI

440 guideline.

First, results of the reliability analysis for RC columns designed based on the ACI 440 guideline

without the FRP resistance reduction factor, i.e. ψf = 1.00, are presented. As is shown in Eq. (16),

23
compressive resistance of FRP-confined RC columns consists of three parts: the strength of the

unconfined concrete, the strength added due to the confinement provided by the FRP jacket and

strength of steel reinforcement. In comparison with capacities contributed by the steel and the

concrete portions, the contribution of FRP confinement to the capacity is highly uncertain. The

results presented in Figs. 14a to 14d, clearly show that when the contribution of FRP confinement

increases (i.e. fr/f'c increases), the reliability indices decrease. Please note that the reliability index

can be calculated based on steps of the procedure outlined in section 4.2 (i.e. steps i to v) and

illustrated as a flowchart in Fig. 5. This suggests that the reliability of the FRP-confined RC column

depends on the level of confinement. This is in line with findings of other researchers [24]. Due to

higher variability in the probabilistic model of the live load, as is shown in Fig. 14a, live-to-dead

load ratio has a converse effect on the reliability indices. Nevertheless, the effect of the live-to-dead

load ratio is less pronounced for high fr/f'c ratios. On the other hand, with low reinforcement

percentage; contribution of steel reinforcement in carrying the axial compression is low, and as the

steel material in comparison with FRP and concrete materials has lower variability, it is expected

that the reinforcement percentage shows an adverse effect on the reliability index. The results

depicted in Fig. 14b confirm such expectation. On the other hand, concrete compressive strength

and cross-sectional dimension have slight converse effect on the reliability (see Figs. 14c and 14d).

Fig. 14.

Following the parametric study on the reliability index and identifying the most influential

variables, reliability of FRP-confined RC columns designed based on the considered design

guidelines is now investigated. For the purpose of comparison between reliability indices of these

guidelines [18, 19], a design case with cross section diameter of 400 mm, live-to-dead load ratio of

1.0, reinforcement percentage of 1% and concrete compressive strength of 35 MPa is considered. It

should be noted that, the current FRP resistance reduction factor in the ACI 440 guideline [18] is

0.95, while that of the fib bulletin 14 [19] depends on the FRP application method and FRP material

type. For wet lay-up application, the partial safety factors for CFRP, AFRP and GFRP materials are
24
1.35, 1.45 and 1.50, respectively. Due to its high strength and stiffness, CFRP material type is a

more appropriate and practical choice for strengthening deteriorated RC column. Fig. 15 shows

reliability of FRP-confined RC columns based on different resistance reduction factors.

Fig. 15.

For ordinary RC columns, corresponding to fr/f'c = 0.0 in Fig. 15, both design guidelines provide

reliability indices above 4.0. However, as the fr/f'c ratio increases, the reliability indices decrease to

as low as 3.5 level. If fr/f'c ratio is less than 1.0, for both design codes the reliability index is still in

the range of reliability of ordinary RC columns. Aiming at the reliability index of ordinary RC

columns as the target reliability, for high fr/f'c ratios, the FRP resistance reduction factor or the

partial safety factor requires modification. The general conclusion from reliability indices of the

considered case (see Fig. 15) is that for fr/f'c ratios less than 1.00, the current safety factors of both

design guidelines are appropriate. However, for ratios larger than 1.00, resistance reduction factor

of about 0.75 is recommended for ACI 440 guideline [18]. In case of the fib bulletin 14 [19], partial

safety factor of 1.65 provides a reliability index above 4.0. This conclusion is based on a specific

design case. Detailed calibration involving different design cases will be conducted in what follows.

An iterative calibration procedure that minimizes the differences between the evaluated reliability

indices and the target reliability can be used for estimating an optimum for the resistance reduction

or partial safety factor. The basic intent of the procedure is to define the range of cases to which the

calibrated factors will be applicable, create many trial designs spanning the range of applicability.

In the common code calibration procedures, the load factors are generally available from the

loading codes. However, the resistance reduction factor/partial safety factor is to be determined.

Calculations are carried out for several possible values of partial safety factor. Seventy-two (72)

design cases with reasonable combinations of influential design variables are selected for the

calibration as follows:

 The live-to-dead load ratio, L/D: 0.50, 1.00, 1.50 and 2.0

25
 The reinforcement percentage: 1.0%, 2% and 3.0%

 The fr/f'c ratio: 0.50, 1.00, 1.50, 2.00, 2.50 and 3.00

Parametric study showed that the section dimension and concrete compressive strength have

marginal effect on the reliability index (see Fig. 14). Therefore, average diameter of 400 mm and

concrete compressive strength of 35 MPa are set in the calibration procedure.

A least-squares-analysis of the differences between the reliability achieved in each case and the

target reliability of 4.0, illustrated in Figs. 16a and 16b, shows that when each of these 72 cases are

equally weighted, a common ψf of about 0.87 (for ACI 440 guideline [18]) and γf of about 1.46 (for

fib bulletin 14 [19]) would lead to the desired target reliability of 4.0.

Fig. 16.

Rounding the resulted resistance reduction factors to the nearest 0.05, the ψf and γf of 0.85 and 1.45

are recommended for design of FRP-confined RC columns.

7 Conclusion

A reliability assessment of FRP-confined columns with circular cross-section under concentric

gravity loads was presented in this paper. Initially, an extensive experimental database was used for

assessing the FRP confinement model and finding the best-fit probability density function for the

FRP strain efficiency factor. It was found that the strength model for FRP-confined circular

columns proposed by the ACI 440 guideline is very close to the best-fit model obtained from the

regression analysis, and the best-fit probability density function for the FRP strain efficiency factor

is the Truncated Weibull distribution, with mean of 0.62 and coefficient of variation of 0.33.

Another experimental database with more than 90 FRP-confined RC columns was used to find

statistical measures and the best-fit distribution for the model error. Then, considering uncertainty

in the mechanical properties, section dimensions, model error and the correlation amongst random

variables, variability in the compressive resistance of FRP-confined RC columns was assessed.

26
Finally, based on an ultimate strength limit state, reliability of the current ACI 440 and fib bulletin

14 guidelines with respect to the design of FRP-confined columns was investigated. It was found

that the reliability of strength limit state depends on the confinement level. Furthermore, results of

code-calibration procedure showed that a resistance reduction factor of 0.85 for the ACI 440

guideline and a partial safety factor of 1.45 for the fib bulletin 14 would lead to minimum target

reliability of 4.0 in design of FRP-confined RC columns.

8 Notations

Ag = Gross area of column cross-section

As = Reinforcement area of column cross-section
a, b = Parameters of the Beta and Weibull distributions
Cx = Covariance matrix of the basic random variables
D, Dn = Diameter and nominal diameter of column cross-section or concrete cylinder
d, dn = Effective depth and nominal effective depth of column cross-section
Ef = Young’s modulus of the FRP jacket
Ej = Lateral stiffness of the FRP jacket
Es = Young’s modulus of steel reinforcement
emin = Minimum load eccentricity
Fx = Cumulative distribution function of x variable
f’c = Concrete compressive strength
f’cc = Ultimate strength of FRP-confined concrete
fr = Lateral (radial) confining pressure exerted by the FRP jacket (at the ultimate state)
fr,rup = Actual confining pressure (at failure state)
fy = Yield strength of the steel reinforcement
G* = Gradient vector of the limit state function evaluated at the design point
k = Confinement effectiveness coefficient
kε = FRP strain efficient factor
l0 = Effective length of column
PDL, PDLn = Actual and nominal dead load effects
PLL, PLLn = Actual and nominal live load effects
PR, PRn = Actual and nominal compressive resistances of FRP-confined RC column
PR,test = Compressive resistance obtained from test
PR,model = Compressive resistance predicted by model

27
pf = Probability of failure
tf = Thickness of the FRP jacket
x = Vector of basic random variables
x* = Design point
α = Factor accounts for minimum eccentricity due to geometric imperfections
β = Reliability index
εf,rup = Actual strain of the FRP jacket (at failure state)
εfu = Ultimate strain of the FRP jacket
γc, γf, γs = Material partial safety factors for concrete, FRP and steel materials
μx = Vector of mean of basic random variables
ξ = Model error
σfu = Ultimate strength of the FRP jacket
ϕc = Resistance reduction factor for compression
Ф = Cumulative standard normal distribution function
ψf = FRP material resistance reduction factor

28
9 References

[1] Saadatmanesh H, Ehsani MR, Li MW. Strength and ductility of concrete columns
externally reinforced with fiber composite straps. ACI Structural Journal. 1994;91(4):434-47.

[2] Ozcan O, Binici B, Ozcebe G. Seismic strengthening of rectangular reinforced concrete

columns using fiber reinforced polymers. Engineering Structures. 2010;32(4):964-73.

[3] Demers M, Neale K. Fourth International Conference on Short and Medium Span Bridges.
Developments in short and medium span bridge engineering. Montreal, Canada: Canadian
Society for Civil Engineering; 1994. p. 895-905.

[4] Binici B. Design of FRPs in circular bridge column retrofits for ductility enhancement.
Engineering Structures. 2008;30(3):766-76.

[5] Mirmiran A, Shahawy M. Behavior of concrete columns confined by fiber composites.

Journal of Structural Engineering. 1997;123(5):583-90.

[6] Cui C, Sheikh S. Experimental study of normal- and high-strength concrete confined with
fiber-reinforced polymers. Journal of Composites for Construction. 2010;14(5):553-61.

[7] Fardis MN, Khalili HH. FRP-encased concrete as a structural material. Magazine of
Concrete Research. 1982;34(121):191-202.

[8] Lam L, Teng J. Design-oriented stress-strain model for FRP-confined concrete.

Construction and Building Materials. 2003;17(6):471-89.

[9] Ozbakkaloglu T, Lim JC. Axial compressive behavior of FRP-confined concrete:

Experimental test database and a new design-oriented model. Composites Part B:
Engineering. 2013;55:607-34.

[10] Realfonzo R, Napoli A. Concrete confined by FRP systems: Confinement efficiency and
design strength models. Composites Part B: Engineering. 2011;42(4):736-55.

[11] Rousakis TC, Rakitzis TD, Karabinis AI. Design-oriented strength model for FRP-
confined concrete members. Journal of Composites for Construction. 2012;16(6):615-25.

[12] Saafi M, Toutanji H, Li Z. Behavior of concrete columns confined with fiber reinforced
polymer tubes. ACI Materials Journal. 1999;96(4):500-9.

[13] Samaan M, Mirmiran A, Shahawy M. Model of concrete confined by fiber composites.

Journal of Structural Engineering. 1998;124(9):1025-31.

[14] Spoelstra MR, Monti G. FRP-confined concrete model. Journal of Composites for
Construction. 1999;3(3):143-50.

[15] Tamuzs V, Tepfers R, Sparnins E. Behavior of concrete cylinders confined by carbon

composite 2. Prediction of strength. Mechanics of Composite Materials. 2006;42(2):109-18.

[16] Xiao Y, Wu H. Compressive behavior of concrete confined by carbon fiber composite

jackets. Journal of Materials in Civil Engineering. 2000;12(2):139-46.

29
[17] Ozbakkaloglu T, Lim JC, Vincent T. FRP-confined concrete in circular sections: Review
and assessment of stress-strain models. Engineering Structures. 2013;49:1068–88.

[18] ACI 440. Guide for the design and construction of externally bonded FRP systems for
strengthening concrete structures. Farmington Hills, MI, USA American Concrete Institute;
2008.

[19] fib bulletin 14. Externally bonded FRP reinforcement for RC structures. Lausanne,
Switzerland: fédération internationale du béton (fib); 2001. p. 165.

[20] CSA S-806. Design and construction of building components with fibre-reinforced
polymers. Mississauga, Ontario, Canada: Canadian Standards Association; 2012.

[21] CNR-DT. Guide for the Design and Construction of Externally Bonded FRP Systems for
Strengthening Existing Structures. Rome, Italy: National Research Council, Advisory
Committee on Technical Regulations for Constructions; 2004.

[22] ISIS Canada. Strengthening reinforced concrete structures with externally-bonded ribre
reinforced polymers. Manitoba, Canada: University of Manitoba; 2001.

[23] TR55. Design guidance for strengthening concrete structures using fibre composite
materials. Century House, Berkshire, UK: Concrete Society; 2004.

[24] Val DV. Reliability of fiber-reinforced polymer-confined reinforced concrete columns.

Journal of Structural Engineering. 2003;129(8):1122-30.

[25] Bisby LA, Dent AJS, Green MF. Comparison of confinement models for fiber-reinforced
polymer-wrapped concrete. ACI Structural Journal. 2005;102(1):62-72.

[26] Zou Y, Hong HP. Reliability assessment of FRP-confined concrete columns designed for
buildings. Structure and Infrastructure Engineering. 2011;7(3):243-58.

[27] Casas JR, Chambi JL. Partial safety factors for CFRP-wrapped bridge piers: Model
assessment and calibration. Composite Structures. 2014;118:267-83.

[28] Wang N, Ellingwood BR. Limit state design criteria for FRP strengthening of RC bridge
components. Structural Safety. 2015;56:1-8.

[29] Pessiki S, Harries KA, Kestner JT, Sause R, Ricles JM. Axial behavior of reinforced
concrete columns confined with FRP jackets. Journal of Composites for Construction.
2001;5(4):237-45.

[30] Richart FE, Brandtzaeg A, Brown RL. A study of the failure of concrete under combined
compressive stresses. Engineering Experimental Station. Champaign, Illinois: University of
Illinois; 1928. p. 110.

[31] Picher F, Rochette P, Labossiére P. Confinement of concrete cylinders with CFRP. In: H.
Saadatmanesh MRE, editor. Proceedings of the First International Conference on Composite
Infrastructures. Tucson, Arizona: University of Arizona; 1996. p. 829-41.

[32] Rochette P, Labossiere P. A plasticity approach for concrete columns confined with
composite materials. 2nd International Conference on Advanced Composite Materials in
Bridges and Structures. Montreal, Québec, Canada: Canadian Society of Civil Engineering;
1996. p. 359-66.
30
[33] Mastrapa J. Effect of construction bond on confinement with fiber composites [Master
Thesis]. Orlando, Florida: University of Central Florida; 1997.

[34] Watanabe K, Nakamura H, Honda Y, Toyoshima M, Iso M, Fujimaki T, et al.

Confinement effect of FRP sheet on strength and ductility of concrete cylinders under uniaxial
compression. Non-metallic (FRP) reinforcement for concrete structures. Sapporo, Japan:
Japan Concrete Institute; 1997. p. 233-40.

[35] Kono S, Inazumi M, Kaku T. Evaluation of confining effects of CFRP sheets on

reinforced concrete members. Second International Conference on Composites in
Infrastructure. Tucson, Arizona,USA: University of Arizona; 1998. p. 343-55.

[36] Mirmiran A, Shahawy M, Samaan M, El Echary H, Mastrapa JC, Pico O. Effect of

column parameters on FRP-confined concrete. Journal of Composites for Construction.
1998;2(4):175-85.

[37] Owen LM. Stress-strain behavior of concrete confined by carbon fiber jacketing [Master
Thesis]. Seattle: University of Washington; 1998.

[38] Demers M, Neale K. Confinement of reinforced concrete columns with fibre-reinforced

composite sheets - An experimental study. Canadian Journal of Civil Engineering.
1999;26(2):226-41.

[39] Matthys S, Taerwe L, Audenaert K. Tests on axially loaded concrete columns confined
by fiber reinforced polymer sheet wrapping. ACI Special Publication. 1999;SP-188:217-28.

[40] Purba BK, Mufti AA. Investigation of the behavior of circular concrete columns
reinforced with carbon fiber reinforced polymer (CFRP) jackets. Canadian Journal of Civil
Engineering. 1999;26(5):590-6.

[41] Toutanji H. Stress-strain characteristics of concrete columns externally confined with

advanced fiber composite sheets. ACI Materials Journal. 1999;96(3):397-404.

[42] Kshirsagar S, Lopez-Anido RA, Gupta RK. Environmental aging of fiber-reinforced

polymer-wrapped concrete cylinders. ACI Materials Journal. 2000;97(6):703-12.

[43] Shahawy M, Mirmiran A, Beitelman T. Tests and modeling of carbon-wrapped concrete

columns. Composites Part B: Engineering. 2000;31(6):471-80.

[44] Dias da Silva V, Santos J. Strengthening of axially loaded concrete cylinders by surface
composites. In: Figueiras J, Juvandes L, Faria R, Marques AT, Ferreira A, Barros J, et al.,
editors. Proceedings of the International Conference: Composites in Constructions. Lisse, The
Netherlands: Balkema Publishers; 2001. p. 257-62.

[45] Micelli F, Myers J, Murthy S. Effect of environmental cycles on concrete cylinders

confined with FRP. In: Figueiras J, Juvandes L, Faria R, Marques AT, Ferreira A, Barros J, et
al., editors. Proceedings of the International Conference: Composites in Constructions. Lisse,
The Netherlands: Balkema Publishers; 2001. p. 317-22.

[46] Rousakis T, Tepfers R. Experimental investigation of concrete cylinders confined by

carbon FRP sheets, under monotonic and cyclic axial compressive load. Work 44. Goteborg,
Sweden: Division of Building Technology, Chalmers University of Technology; 2001. p. 87.

31
[47] Wang F, Cheong K. RC columns strengthened by FRP under uniaxial compression. In:
Teng J, editor. Proceedings of the International Conference on FRP Composites in Civil
Engineering. Oxford, UK: Elsevier Science Ltd; 2001. p. 327-34.

[48] De Lorenzis L, Micelli F, La Tegola A. Influence of specimen size and resin type on the
behaviour of FRP-confined concrete cylinders. Advanced Polymer Composites for Structural
Applications in Construction: Proceedings of the First International Conference. Southampton
University, UK: Thomas Telford; 2002. p. 231-40.

[49] Ilki A, Kumbasar N, Koc V. Strength and deformability of low strength concrete
confined by carbon fiber composite sheets. 15th ASCE Engineering Mechanics Conference.
New York, NY: Columbia University; 2002.

[50] Bullo S. Experimental study of the effects of the ultimate strain of fiber reinforced plastic
jackets on the behavior of confined concrete. In: Cosenza E, editor. Second International
Conference on Composite in Construction. Rende, Italy: University of Calabria; 2003. p. 465-
70.

[51] Harries KA, Carey SA. Shape and “gap” effects on the behavior of variably confined
concrete. Cement and Concrete Research. 2003;33(6):881-90.

[52] Harries KA, Kharel G. Experimental investigation of the behavior of variably confined
concrete. Cement and Concrete Research. 2003;33(6):873-80.

[53] Rousakis T, You C-S, De Lorenzis L, Tamužs V, Tepfers R. Concrete cylinders confined
by CFRP sheets subjected to cyclic axial compressive load. In: Tan K, editor. Proceedings of
the 6th International Symposium on FRP Reinforcement for Concrete Structures. Singapore:
World Scientific; 2003. p. 571-80.

[54] Faella C, Realfonzo R, Salerno N. Sulla resistenza e deformazione di elementi in ca

confinati con tessuti in FRP. Proceedings of the XI National Congress “L’ingegneria sismica
in Italia”. Genova, Italy: ANIDIS; 2004.

[55] Lam L, Teng J. Ultimate condition of fiber reinforced polymer-confined concrete.

Journal of Composites for Construction. 2004;8(6):539-48.

[56] Berthet J, Ferrier E, Hamelin P. Compressive behavior of concrete externally confined by

composite jackets. Part A: Experimental study. Construction and Building Materials.
2005;19(3):223-32.

[57] Carey SA, Harries KA. Axial behavior and modeling of confined small-, medium-, and
large-scale circular sections with carbon fiber-reinforced polymer jackets. ACI Structural
Journal. 2005;102(4):596-604.

[58] Lam L, Teng JG, Cheung CH, Xiao Y. FRP-confined concrete under axial cyclic
compression. Cement and Concrete Composites. 2006;28(10):949-58.

[59] Silva MA, Rodrigues CC. Size and relative stiffness effects on compressive failure of
concrete columns wrapped with glass FRP. Journal of Materials in Civil Engineering.
2006;18(3):334-42.

[60] Yan Z, Pantelides CP, Reaveley LD. Fiber-reinforced polymer jacketed and shape-
modified compression members: I- Experimental behavior. ACI Structural Journal.
2006;103(6):885-93.
32
[61] Almusallam TH. Behavior of normal and high-strength concrete cylinders confined with
E-glass/epoxy composite laminates. Composites Part B: Engineering. 2007;38(5–6):629-39.

[62] Ciupala M, Pilakoutas K, Mortazavi A. Effectiveness of FRP composites in confined

concrete. Proceedings of the 8th International Symposium on Fiber Reinforced Polymer
Reinforcement for Concrete Structures (FRPRCS-8). Patras, Greece: University of Patras;
2007. p. 1-10.

[63] Jiang T, Teng J. Analysis-oriented stress-strain models for FRP-confined concrete.

Engineering Structures. 2007;29(11):2968-86.

[64] Shehata I, Carneiro L, Shehata L. Strength of confined short concrete columns.

Proceedings of the 8th International Symposium on Fiber Reinforced Polymer Reinforcement
for Concrete Structures. Patras, Greece: University of Patras; 2007. p. 1-10.

[65] Teng J, Yu T, Wong Y, Dong S. Hybrid FRP-concrete-steel tubular columns: Concept

and behavior. Construction and Building Materials. 2007;21(4):846-54.

[66] Valdmanis V, De Lorenzis L, Rousakis T, Tepfers R. Behaviour and capacity of CFRP-

confined concrete cylinders subjected to monotonic and cyclic axial compressive load.
Structural Concrete. 2007;8(4):187-200.

[67] Tamuzs V, Valdmanis V, Tepfers R, Gylltoft K. Stability analysis of CFRP-wrapped

concrete columns strengthened with external longitudinal CFRP sheets. Mechanics of
Composite Materials. 2008;44(3):199-208.

[68] Wang L-M, Wu Y-F. Effect of corner radius on the performance of CFRP-confined
square concrete columns: Test. Engineering Structures. 2008;30(2):493-505.

[69] Bisby LA, Take WA. Strain localisations in FRP-confined concrete: New insights.
Proceedings of the ICE-Structures and Buildings. 2009;162(5):301-9.

[70] Cui C. Behaviour of normal and high strength concrete confined with fibre reinforced
polymers (FRP) [PhD Thesis]: University of Toronto; 2009.

[71] Aire C, Gettu R, Casas J, Marques S, Marques D. Concrete laterally confined with fibre-
reinforced polymers (FRP): Experimental study and theoretical model. Materiales de
Construcción. 2010;60(297):19-31.

[72] Eid R, Roy N, Paultre P. Normal-and high-strength concrete circular elements wrapped
with FRP composites. Journal of composites for construction. 2009;13(2):113-24.

[73] Benzaid R, Mesbah H, Chikh NE. FRP-confined concrete cylinders: Axial compression
experiments and strength model. Journal of Reinforced Plastics and Composites.
2010;29(16):2469-88.

[74] Chastre C, Silva MAG. Monotonic axial behavior and modelling of RC circular columns
confined with CFRP. Engineering Structures. 2010;32(8):2268-77.

[75] Mohamed HM, Masmoudi R. Axial load capacity of concrete-filled FRP tube columns:
Experimental versus theoretical predictions. Journal of Composites for Construction.
2010;14(2):231-43.

33
[76] Xiao Q, Teng J, Yu T. Behavior and modeling of confined high-strength concrete.
Journal of Composites for Construction. 2010;14(3):249-59.

[77] Elsanadedy HM, Al-Salloum YA, Alsayed SH, Iqbal RA. Experimental and numerical
investigation of size effects in FRP-wrapped concrete columns. Construction and Building
Materials. 2012;29:56-72.

[78] Liang M, Wu ZM, Ueda T, Zheng JJ, Akogbe R. Experiment and modeling on axial
behavior of carbon fiber reinforced polymer confined concrete cylinders with different sizes.
Journal of Reinforced Plastics and Composites. 2012;31(6):389-403.

[79] Song X, Gu X, Li Y, Chen T, Zhang W. Mechanical behavior of FRP-strengthened

concrete columns subjected to concentric and eccentric compression loading. Journal of
Composites for Construction. 2012;17(3):336-46.

[80] Wang Z, Wang D, Smith ST, Lu D. Experimental testing and analytical modeling of
CFRP-confined large circular RC columns subjected to cyclic axial compression. Engineering
Structures. 2012;40:64-74.

[81] Wu YF, Jiang JF. Effective strain of FRP for confined circular concrete columns.
Composite Structures. 2012;95(1):479-91.

[82] Micelli F, Modarelli R. Experimental and analytical study on properties affecting the
behaviour of FRP-confined concrete. Composites Part B: Engineering. 2013;45(1):1420-31.

[83] Ozbakkaloglu T, Vincent T. Axial compressive behavior of circular high-strength

concrete-filled FRP tubes. Journal of Composites for Construction. 2013;18(2):1-11.

[84] Vincent T, Ozbakkaloglu T. Influence of concrete strength and confinement method on

axial compressive behavior of FRP confined high- and ultra high-strength concrete.
Composites Part B: Engineering. 2013;50:413-28.

[85] Lim JC, Ozbakkaloglu T. Hoop strains in FRP-confined concrete columns: Experimental
observations. Materials and Structures. 2015;48(9):2839-54.

[86] Lim JC, Ozbakkaloglu T. Investigation of the influence of the application path of
confining pressure: Tests on actively confined and FRP-confined concretes. Journal of
Structural Engineering. 2015;141(8):1-18.

[87] Wu Y-F, Yun Y, Wei Y, Zhou Y. Effect of predamage on the stress-strain relationship of
confined concrete under monotonic loading. Journal of Structural Engineering.
2014;140(12):04014093.

[88] Guler S, Ashour A. Review of Current Design Guidelines for Circular FRP-Wrapped
Plain Concrete Cylinders. Journal of Composites for Construction. 2015;20(2):04015057.

[89] ACI 318. Building Code Requirements for Structural Concrete and Commentary.
Farmington Hills, MI, USA: American Concrete Institute; 2011.

[90] Comité Euro-International du Béton (fib). fib Model Code for Concrete Structures 2010.
Lausanne, Switzerland: Ernst & Sohn; 2010. p. 434.

[91] Mander JB, Priestley MJ, Park R. Theoretical stress-strain model for confined concrete.
Journal of Structural Engineering. 1988;114(8):1804-26.
34
[92] Buonopane S. Strength and reliability of steel frames with random properties. Journal of
Structural Engineering. 2008;134(2):337-44.

[93] Park R, Paulay T. Reinforced concrete structures. New York: Wiley; 1975.

[94] Melchers RE. Structural reliability analysis and prediction. New York: John Wiley &
Sons; 1999.

[95] MacGregor JG. Load and resistance factors for concrete design. ACI Journal
Proceedings. 1983;80(4): 279-87.

[96] Szerszen MM, Nowak AS. Calibration of design code for buildings (ACI 318): Part 2-
Reliability analysis and resistance factors. ACI Structural Journal. 2003;100(3):383-91.

[97] Israel M, Ellingwood B, Corotis R. Reliability-based code formulations for reinforced

concrete buildings. Journal of Structural Engineering. 1987;113(10):2235-52.

[98] Mirza SA, MacGregor JG, Hatzinikolas M. Statistical descriptions of strength of

concrete. Journal of the Structural Division. 1979;105(6):1021-37.

[99] Bartlett F. Canadian Standards Association Standard A23.3-04 resistance factor for
concrete in compression. Canadian Journal of Civil Engineering. 2007;34(9):1029-37.

[100] Bartlett FM, Macgregor JG. Statistical analysis of the compressive strength of concrete
in structures. ACI Materials Journal. 1996;93(2):158-68.

[101] Mirza SA, MacGregor JG. Variability of mechanical properties of reinforcing bars.
Journal of the Structural Division. 1979;105(5):921-37.

[102] Atadero RA. Development of load and resistance factor design for FRP strengthening of
reinforced concrete structures [PhD Thesis]. San Diego, USA: University of California; 2006.

[103] Atadero RA, Karbhari VM. Sources of uncertainty and design values for field-
manufactured FRP. Composite Structures. 2009;89(1):83-93.

[104] Ellingwood B, Galambos T, McGregor J, Cornell C. Development of a probability-

based load criteria for American Standard A 58. NBS Special Publication. Washington, D.C.,
USA: National Bureau of Standards; 1980.

[105] De Lorenzis L, Tepfers R. Comparative study of models on confinement of concrete

cylinders with fiber-reinforced polymer composites. Journal of Composites for Construction.
2003;7(3):219-37.

[106] Realfonzo R, Napoli A. Confining concrete members with FRP systems: Predictive vs
design strain models. Composite Structures. 2013;104(0):304-19.

[107] Lim JC, Ozbakkaloglu T. Confinement model for FRP-confined high-strength concrete.
Journal of Composites for Construction. 2013;17(5):1-19.

[108] Bournonville M, Dahnke J, Darwin D. Statistical analysis of the mechanical properties

and weight of reinforcing bars. Lawrence, KS: University of Kansas; 2004. p. 194.

[109] Ross SM. Introduction to probability and statistics for engineers and scientists.
Burlington, USA: Academic Press; 2014.
35
[110] Atadero R, Lee L, Karbhari VM. Consideration of material variability in reliability
analysis of FRP strengthened bridge decks. Composite Structures. 2005;70(4):430-43.

[111] Zhang Y-x. Behavior of large-size FRP-jacketed circular and rectangular reinforced
concrete columns [Master Thesis]. Hong Kong, China: The Hong Kong Polytechnic
University; 2013.

[112] Lee J-Y, Yi C-K, Jeong H-S, Kim S-W, Kim J-K. Compressive response of concrete
confined with steel spirals and FRP composites. Journal of Composite Materials.
2010;44(4):481-504.

[113] Cairns SW. Circular concrete columns externally reinforced with pre-fabricated carbon
polymer shells [Master Thesis]. Toronto, Canada: University of Toronto; 2001.

[114] Jaffry SAD. Concrete filled glass fibre reinforced polymer (GFRP) shells under
concentric compression [Master Thesis]. Toronto, Canada: University of Toronto; 2001.

36
List of Tables

Table 1. FRP confinement models based on some of the international design guidelines

Table 2. Probabilistic models for the main random variables

Table 3. Collected database for calibrating the model error

Table 4. Goodness-of-fit and statistical measures of the FRP strain efficiency factor

37
Table 1. FRP confinement models based on some of the international design guidelines

Design guideline Confinement model

 f  f
ACI 440 [18] f cc'  f c' 1  3.3 r'  , r'  0.08
 fc  fc
Bases on analysis-oriented model (theoretical) or
fib bulletin 14 [19]  f 
f cc'  f c'  0.2  3.0 r'  (practical)
 fc 
 f  
0.17

CSA S-806 [20] f cc'  f c'  0.85  6.7 r ' 

 f c 

 f 
ISIS Canada [22] f cc'  f c' 1  2.0 r'  , 4 MPa  f r  0.29 f c'
 fc 
 1 2E f t f  2E f t f mm2
TR55 [23] f  f 
'
cc c
'
 0.05( ) ,  0.183
 0.8 Df c'  D( f c' )2 N
 f  f
CNR-DT [21] f cc'  f c'  2.6  ( r' )2/3  , r'  0.05
 fc  fc

38
 Table 2. Probabilistic models for the main random variables

Variable Nominal Bias COV Distribution Reference

D (mm) 300/400/500 1.00 0.60/Dn Normal
Israel et al. [97]
d (mm) 250/350/450 1.00 0.40/dn Normal
f’c (MPa) 25/35/45 1.20 0.18 Lognormal Bartlett and MacGregor [100]
fy (MPa) 424 1.16 0.093 Beta
Es (MPa) 200000 1.00 0.033 Lognormal Mirza and MacGregor [101]
As/Ag 0.01/0.02/0.03 0.99 0.024 Normal
tf (mm) Eq. (3) 1.00 0.05 Lognormal
Atadero et al.[110]
σfu (MPa) 887.0 1.20 0.15 Weibull
kε 0.55 1.12 0.33 Truncated Weibull
This study
 1.00 1.12 0.15 Lognormal
PDL - 1.05 0.10 Normal
Ellingwood et al. [104]
PLL - 1.00 0.25 Extreme Type I
emin/D 0.05 0.50 0.58 Uniform Assumed
Bias = Mean/Nominal
COV = Coefficient of Variation

39
 Table 3. Collected database for calibrating the model error

Reference No HxD f c' Model Error

(mm) (MPa) Mean COV
Zhang [111] 5 1300 x 650 25.0-46.4 1.02 0.063
915 x 605
Wang et al. [80] 12 24.5 1.07 0.075
612 x 204
750 x 150
Chastre and Silva [74] 13 35.2-38.0 1.25 0.139
750 x 250
1200 x 303
Eid and Paultre [72] 16 29.4-50.8 1.13 0.106
1200 x 253
Lee et al. [112] 19 300 x 150 36.2 1.11 0.165
Cairns [113] 8 1525 x 356 32.9-33.3 1.37 0.092
Jaffry [114] 8 1525 x 356 29.8 1.11 0.176
Pessiki [29] 2 1830 x 508 32.8 0.94 0.075
Demers and Neale [38] 8 1200 x 300 25.0-40.0 0.98 0.084
Total 91 - 24.5-50.8 1.12 0.150

40
 Table 4. Goodness-of-fit and statistical measures of the FRP strain efficiency factor

Goodness-of-fit Statistics
Pearson’s Kolmogorov- Anderson- 5th
Distribution Mean COV
Chi-square Smirnov Darling percentile
Truncated Weibull 6.082 0.031 0.660 0.61 0.33 0.248
Beta 10.374 0.047 1.109 0.62 0.33 0.263

41
List of Figures

Fig. 1. Confining action of the FRP jacket on concrete core

Fig. 2. Histogram of different variables in the collected database

(a) Diameter of concrete cylinder
(b) Concrete compressive strength
(c) Normalised actual confining pressure
(d) Normalised lateral stiffness of the FRP jacket

Fig. 3. Experimental data versus the best-fit confinement models

Fig. 4. Deriving compressive resistance of FRP-confined RC column

Fig. 5. Reliability analysis procedure

Fig. 6. Scatter of experimental data for the FRP strain efficiency factor
(a) Relation with concrete compressive strength
(b) Relation with the normalised confining pressure

Fig. 7. Correlation of the FRP strain efficiency factor with f’c, Ef and σfu

Fig. 8. The best-fit probability density functions for the FRP strain efficiency factor

Fig. 9. Model versus test results for compressive resistance of FRP-confined RC columns

Fig. 10. The best-fit probability density functions for compressive resistance of FRP-confined RC columns
(a) Normal distribution
(b) Lognormal distribution
(c) Weibull distribution
(d) Gamma distribution

Fig. 11. Sensitivity of compressive resistance of FRP-confined RC columns to the main random variables

Fig. 12. Statistics of the normalised compressive resistance of FRP-confined RC columns

(a) Mean
(b) Coefficient of variation

Fig. 13. Cumulative distribution function of the simulated normalised compressive resistance
Fig. 14. Parametric study on reliability indices of FRP-confined RC columns
(a) Live-to-dead load ratio, L/D
(b) Reinforcement percentage, ρ
(c) Concrete compressive strength, f’c
(d) Cross-section dimension, D

Fig. 15. Reliability of FRP-confined RC columns with different FRP resistance reduction factors
(a) ACI 440 guideline [18]

42
 (b) fib bulletin 14 [19]

Fig. 16. Calibration of FRP resistance reduction factors

(a) ACI 440 guideline [18]
(b) fib bulletin 14 [19]

43
Fig. 1. Confining action of the FRP jacket on concrete core

44
 (a) Diameter of concrete cylinder (b) Concrete compressive strength

(c) Normalised actual confining pressure (d) Normalised lateral stiffness of the FRP jacket

Fig. 2. Histogram of different variables in the collected database

45
Fig. 3. Experimental data versus the best-fit confinement models

46
Fig. 4. Deriving compressive resistance of FRP-confined RC column

47
Fig. 5. Reliability analysis procedure

48
(a) Relation with concrete compressive strength (b) Relation with the normalised confining pressure

Fig. 6. Scatter of experimental data for the FRP strain efficiency factor

49
Fig. 7. Correlation of the FRP strain efficiency factor with f’c, Ef and σfu

50
Fig. 8. The best-fit probability density functions for the FRP strain efficiency factor

51
Fig. 9. Model versus test results for compressive resistance of FRP-confined RC columns

52
 (a) Normal distribution (b) Lognormal distribution

(c) Weibull distribution (d) Gamma distribution

Fig. 10. The best-fit probability density functions for compressive resistance of FRP-confined RC

columns

53
Fig. 11. Sensitivity of compressive resistance of FRP-confined RC columns to the main random

variables

54
 (a) Mean (b) Coefficient of variation

Fig. 12. Statistics of the normalised compressive resistance of FRP-confined RC columns

55
Fig. 13. Cumulative distribution function of the simulated normalised compressive resistance

56
 (a) Live-to-dead load ratio, L/D (b) Reinforcement percentage, ρ

(c) Concrete compressive strength, f’c (d) Cross-section dimension, D

Fig. 14. Parametric study on reliability indices of FRP-confined RC columns

57
 (a) ACI 440 guideline [18] (b) fib bulletin 14 [19]

Fig. 15. Reliability of FRP-confined RC columns with different FRP resistance reduction factors

58
 (a) ACI 440 guideline [18] (b) fib bulletin 14 [19]

Fig. 16. Calibration of FRP resistance reduction factors

59

View publication stats