Reliability Analysis of RC Columns and Frame with FRP

Strengthening Subjected to Explosive Loads

Hong Hao, F.ASCE 1; Zhong-Xian Li, M.ASCE 2; and Yanchao Shi 3
Downloaded from ascelibrary.org by Guru Nanak Dev Engineering College on 09/07/22. Copyright ASCE. For personal use only; all rights reserved.

Abstract: Some structures, both military and civilian, might experience explosive loads during their service life. Owing to high uncertainties
in blast load predictions and structural parameters, accurate assessment of the performances of structures under explosion loads is a chal-
lenging task. For example, a number of experimental studies using fiber-reinforced plastic (FRP) strengthening of RC structures have been
reported in the literature. Most of these studies demonstrate that FRP strengthening is effective in increasing the blast load–carrying capacities
of RC structures. However, significant variations in the effectiveness of FRP strengthening have also been observed owing to the large
uncertainties in blast loading, RC and FRP material properties, and workmanship in preparing the test specimens and conducting exper-
imental tests. Very few studies that take into consideration these uncertainties in analyzing the effectiveness of FRP strengthening of RC
structures on blast loading resistance can be found in the literature. This study performs a reliability analysis to assess the performance of RC
columns with or without FRP strengthening in resisting blast loads. Statistical variations of blast loading predictions derived in a previous
study are adopted in this study. To define structural performance, pressure-impulse (P-I) curves with a damage criterion on the basis of axial
load-carrying capacity that were developed in previous studies for RC columns without strengthening or with FRP strip, FRP wrap, or both
FRP strip and wrap strengthening are used. Considering the uncertainties in blast loading predictions and RC column and FRP material
properties and dimensions, limit-state functions corresponding to different damage levels of RC columns with or without FRP strengthening
are formulated. The statistical variations of blast loading, RC column dimension, longitudinal and transverse reinforcement ratio, and con-
crete, steel, and FRP material strength and FRP thickness are considered. The failure probabilities of RC columns corresponding to different
damage levels with or without FRP strengthening are calculated. The effectiveness of FRP strengthening on the RC column’s blast loading
resistance capacities is discussed. The importance of considering the random fluctuations on blast loading and RC column parameters in
assessing the blast loading effect on RC columns is demonstrated. A structural system reliability analysis is also carried out to examine the
probability of structural collapse as a result of blast loading applied to the front of an example two-span multistory RC frame. The results
obtained in this study demonstrate the effectiveness of FRP strengthening on structure protection and can also be used to assess RC structure
performance under blast loadings. DOI: 10.1061/(ASCE)CF.1943-5509.0000748. © 2015 American Society of Civil Engineers.
Author keywords: Fiber-reinforced plastic (FRP); RC columns; Strengthening; Uncertainty; Reliability analysis; Blast loading.

Introduction the necessity for an effective structure and human protection

against explosion loads (DOS 2011). Unlike military structures, it
Some civilian structures in their service life might experience blast is not realistic to prevent public access to civilian structures for
loadings from terrorist attacks and/or accidental explosions. For ex- blast protection. Therefore, blast-resistant design and structural
ample, the U.S. Department of State (DOS) released a list of 44 strengthening are critical to the safety of civilian structures, where
terrorist organizations, and the annual number of recorded attacks there is also a high risk of casualties.
has been more than 11,000 since 2005 (DOS 2009). In 2010 alone, Many structures are constructed of RC material. Intensive re-
these attacks caused 13,186 fatalities and 30,665 injuries, and search has been carried out to assess the capacities of RC structures
among the 11,604 attacks, 13% of them involved explosions target- in resisting blast loads (Beckmann et al. 2011; Shi et al. 2008; Zhou
ing civilian or commercial structures, indicating great threats and et al. 2008). To increase the load-carrying capacities of RC struc-
tures, a number of strengthening techniques that are able to adapt to
1 a variety of existing conditions have been investigated. These in-
Professor, Joint Research Center of Structural Monitoring and Protec-
tion, Tianjin Univ., Tianjin, China; and School of Civil and Mechanical
clude the FRP composites that have been widely used to strengthen
Engineering, Curtin Univ., Bentley, GPO Box U1987, Perth, WA 6845, existing concrete and other structures. Crawford et al. (1996) inves-
Australia. E-mail: [email protected] tigated the effect of retrofitting RC columns with FRP wraps to
2 improve their resistance to blast loads. It is found that shear failure
Professor, Key Laboratory of Coast Civil Structure Safety, Ministry
of Education, Tianjin Univ., Tianjin 300072, China; and School of Civil is a potential major collapse mechanism, and FRP wrap is a means
Engineering, Tianjin Univ., Tianjin 300072, China. E-mail: [email protected] of improving this. Crawford et al. (2001) looked into a procedure
3
Associate Professor, Key Laboratory of Coast Civil Structure Safety, for increasing the blast resistance of RC columns by incorporating
Ministry of Education, Tianjin Univ., Tianjin 300072, China; and School of FRP wrap and strips. The result verified that FRP strengthening
Civil Engineering, Tianjin Univ., Tianjin 300072, China (corresponding
improves the flexural and shear capacities of RC columns. Because
author). E-mail: [email protected]
Note. This manuscript was submitted on July 23, 2014; approved on
a number of tests show that columns retrofitted with transverse FRP
December 26, 2014; published online on March 12, 2015. Discussion per- display much higher levels of ductility and deform flexurally rather
iod open until August 12, 2015; separate discussions must be submitted for than failing in shear, Isaac et al. (2011) developed a plasticity-based
individual papers. This paper is part of the Journal of Performance of approach to evaluate the responses of RC columns with FRP
Constructed Facilities, © ASCE, ISSN 0887-3828/04015017(15)/$25.00. strengthening.

© ASCE 04015017-1 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2016, 30(2): 04015017

 To take into consideration the statistical variations in RC struc- compared the peak reflected pressures and positive blast loading
tural and FRP material parameters, Atadero and Karbhari (2005) duration predicted from the preceding empirical relations and com-
examined the feasibility of a probabilistic-based design approach puter code ConWep (Hyde 1992), and found that these empirical
for FRP strengthening of RC. They looked to determine appropriate relations may give different predictions of blast loads from the
resistance factors for use in design on the basis of the coefficient of same blasting conditions, especially in the range with the explosion
variation (COV) of the FRP strength. Owing to significant varia- scaled distance being smaller than 1.0 m=kg1=3 and larger than
tions caused by inconsistency in quality control, manufacturing, 10 m=kg1=3 , with an average coefficient of variation of 0.3227
and installation of FRP, the authors concluded that it is difficult for peak reflected pressure and 0.130 for positive blast loading du-
to generate a general guide for FRP-strengthening design and as- ration (Low and Hao 2001). By analyzing 325 peak pressure and
sessment. Wang et al. (2010) attempted to create reliability-based 320 impulse data obtained from testing at a scaled distance varying
guidelines for externally bonded FRP composites. Workmanship from 0.5 to 2.6 m=kg1=3 , Twisdale et al. (1994) found significant
Downloaded from ascelibrary.org by Guru Nanak Dev Engineering College on 09/07/22. Copyright ASCE. For personal use only; all rights reserved.

was found to be a significant factor in the characteristics of the scattering of the recorded data. The ratios of the individual record to
FRP. Buchan and Chen (2007) performed a comprehensive review the mean value of all the data vary from 0.4 to 3 for both peak
of experimental and numerical research available for retrofitting pressure and impulse. A similar study that compared the peak in-
structures with FRP and polymers against blast loading. They cident and reflected positive pressure, incident and reflected pos-
identified FRP as a suitable candidate for improving the existing itive impulse, incident and reflected peak negative pressure, and
physical properties of structures so that they could withstand blast impulse estimated by using three computer codes, ConWep (Hyde
loading, either as a wrap, in strips, or as a spray. Although FRP 1992), SHOCK (Wager 2005), and BlastX (Britt et al. 2001),
strengthening has been found to be highly effective at increasing and approximately 300 measured data in field tests in the scaled
the strength and ductility of concrete structures in resisting blast distance range of 1.19 m=kg1=3 (3 ft=lb1=3 ) to 23.79 m=kg1=3
and impact loads (Buchan and Chen 2007; Crawford et al. 2001; (60 ft=lb1=3 ) concluded that the average coefficient of variation
Teng et al. 2003; Triantafillou 1998), there are some contradicting of the reflected peak positive pressure is approximately 0.24 and
observations regarding its effectiveness owing primarily to prema- that of the incident positive impulse is 0.18 (Bogosian et al. 2002).
ture debonding failure of FRP (Razaqpur et al. 2007; Khalifa et al. Netherton and Stewart (2010) performed a statistical analysis of
1999; Lawver et al. 2003; Mosalam and Mosallam 2001; Muszynski blast loading data and revealed a high level of variability in peak
and Purcell 2003a, b; Wu et al. 2009). This is because of high in- reflected pressure, impulse, and time of positive phase duration. It
terfacial stress near the boundary between the concrete structure and was found that the TM5-1300 (U.S. Department of the Army 1990)
the FRP sheet that is generated by the blasting load–induced stress design values for peak reflected pressure and time of positive phase
wave or in the flexural-shear cracks formed in the concrete system. duration adequately represent median values of the probability
Besides premature debonding failure and workmanship that affect distribution of blast loads. The TM5-1300 design values for peak
the quality of the FRP retrofitting of structures and hence the struc- reflected impulse were 40% higher than the median values, with
ture performance, fluctuations of the RC structure and FRP material probabilities of exceedance of 4–23%. These studies demonstrated
properties also contribute to the observed variations of performances the significant variations of blast loads estimated from various em-
of structures strengthened with FRP. As a result, despite most studies pirical relations or computer code commonly used in practice.
that found that FRP strengthening is effective, there is, in general, a Compared with variations in blast-loading estimations, the
lack of quantitative predictions for the effectiveness of FRP strength- structural parameters, such as material properties and dimensions,
ening on concrete structure blast load–carrying capacities. also inevitably vary but in a relatively less extent. This is because
A pressure-impulse (P-I) curve is a graphical representation of of construction quality control and material deterioration. For ex-
the damage threshold of a structure when a dynamic load is applied ample, it was reported that the average coefficient of variation of
to it. The damage threshold is usually defined according to the re- RC structure dimensions under normal construction quality control
sponse levels of the structure, such as the maximum ductility ratio is 3% and that of concrete strength is 11% (Low and Hao 2001).
and maximum displacement. The P-I curves are commonly used to The coefficient of variation of FRP strength was found to range
correlate the blast load to the corresponding damage, which can be from 4 to 24% (Wang et al. 2010). These variations affect the struc-
readily used for quick damage assessment of structures. To quantify tural parameters, such as stiffness, material strength, and dimen-
the effectiveness of FRP strengthening on blast load–carrying sions, which in turn affect the structural responses (Connolly 1976;
capacities of concrete columns, intensive laboratory drop weight Ellingwood 1978; MacGregor et al. 1983; Mirza et al. 1979; Val
tests and numerical simulations were carried out to construct P-I et al. 1997). Therefore, a more reliable prediction of structural re-
curves of concrete structures with different dimensions, reinforce- sponse and assessment of the effectiveness of FRP strengthening of
ment ratios, and material properties, which were strengthened with RC structures against blast loading should take into consideration
FRP of different configurations, thickness, and epoxy and FRP these variations.
strengths (Mutalib and Hao 2011). These P-I curves can be used Considering random fluctuations of structural and loading
for easy assessment of blast load–carrying capacities of RC col- parameters, reliability and failure probability analyses of structural
umns with or without FRP strengthening. components and structural systems to static and other types of dy-
However, the previously discussed P-I curves were developed namic loads, such as earthquake, wind, and wave loads, have been
with the assumption of deterministic blast loading and structural widely reported in the literature. However, reliability analysis of the
parameters. In reality, it is very difficult to predict the exact blast structural response to blast loads is rather limited. Low and Hao
loads acting on a structure. This is because many factors affect (2001, 2002) performed reliability analyses of simply supported
the explosion wave, including the effectiveness of the explosive RC beams and slabs subjected to blast loads with randomly varying
material, ignition points, shape, and location of the explosive and parameters. In these studies, the structure is simplified to one (Low
interaction of blast waves with the surrounding structures. A num- and Hao 2001) or two equivalent single degree of freedom (SDOF)
ber of empirical relations and computer packages are available to systems (Low and Hao 2002), respectively, representing the flexu-
estimate blast loads (Baker and Cox 1983; Brode 1959; Henrych ral and shear failure mode. The failure probabilities and boundary
1979; Hyde 1992; Kingery et al. 1984; U.S. Department of the between direct shear failure and flexural failure of the beam and
Army 1969, 1990; Sadovskii 1952; DoD 2008). A previous study slab as a function of the blast loading duration were determined.

© ASCE 04015017-2 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2016, 30(2): 04015017

 Stewart and Netherton (2005) calculated the failure probabilities of
glazing subjected to blast loads, and subsequently extended the
work to perform a risk assessment and analysis of the optimization
strategies for blast mitigations of built infrastructure (Stewart
2006). Considering various uncertainties, Stewart also discussed
the risk acceptability and cost-effectiveness of infrastructure protec-
tion (Stewart 2010). Using the developed P-I curve for RC col-
umns (Shi et al. 2008), Hao et al. (2010) conducted reliability
analyses of RC columns to blast loads. It was concluded that ran-
dom variations of blast loading govern the variation in failure prob-
abilities, whereas those of RC column parameters have a relatively
Downloaded from ascelibrary.org by Guru Nanak Dev Engineering College on 09/07/22. Copyright ASCE. For personal use only; all rights reserved.

lower effect on the variation of failure probabilities owing to their

relatively insignificant fluctuations. A study of the failure probabil-
Fig. 1. FRP strengthened RC columns: (a) FRP strip; (b) FRP wrap;
ities of RC columns strengthened with FRP and structural collapse
(c) FRP strip and wrap
to blast loads cannot be found in the open literature yet.
This paper presents the numerical results of the failure proba-
bilities of RC columns with or without FRP strengthening to blast
represents low damage, D ¼ 0.5 represents medium damage,
loads. This work is an extension of that reported by Hao et al.
and D ¼ 0.8 represents a high level of damage or almost collapse
(2010), which studied the failure probabilities of RC columns with-
of the column. These damage levels are defined according to the
out FRP strengthening. In the present study, the structural compo-
remaining axial load-carrying capacity of the column with respect
nent reliability analysis method is used to estimate the failure
to its design axial load-carrying capacity (Mutalib and Hao 2011).
probabilities of RC columns corresponding to different damage lev-
As can be noted, FRP strengthening clearly enhances the blast
els subjected to blast loads. The P-I curves developed in a previous
load–carrying capacities of RC columns.
study (Mutalib and Hao 2011) are used to estimate the column
The P-I curves might change curvature in the impulsive loading
damage. The analyses consider random fluctuations of RC column
region owing to negative-phase blast loading and/or changing of
dimensions, reinforcement ratios, material properties, FRP configu-
the primary structural response mode from flexural to shear mode.
rations, thickness, and strength and the random fluctuations of blast
In the previous study (Mutalib and Hao 2011), negative-phase blast
loadings corresponding to different scaled distances. The computer
loading was not considered. Therefore, P-I curves can be repre-
code CALREL is used to calculate the failure probabilities (Liu et al.
sented by Eq. (19) without curve changing.
1989). Numerical results of failure probabilities corresponding to
different damage levels of three generic RC columns with or with-
out FRP strengthening that were subjected to blast loadings of dif-
ferent scaled distances are presented. Discussions on the influences Random Fluctuations of RC Column and FRP
of the random fluctuations of structural properties and blast load- Properties
ings on RC column failure probabilities are given. The ranges of The previously discussed P-I diagrams developed by Mutalib and
the blast loadings that are very unlikely to cause any damage to sure Hao (2011) are constructed by assuming deterministic RC column
collapse of the RC columns are derived. A system reliability analy- and FRP parameters. In reality, depending on construction quality
sis is also carried out to predict the probability of structural collapse control and environmental conditions, RC structural parameters,
as a result of blast loading applied to the front of an example two- such as dimensions, reinforcement ratios, and material strength,
span multistory building frame. The results obtained in this study inevitably fluctuate. Intensive studies of statistical variations of
can be used to assess the failure probabilities of RC columns and RC structural parameters have been reported in the literature. In
effectiveness of FRP strengthening of RC structures on improving these studies, statistical analyses of available data were carried
their blast loading resistance capacities. out to determine the COV of each RC structure parameter. It was
found that the standard deviation of concrete strength in the United
Kingdom varies between 4.5 and 7.0 MPa, which results in a COV
Pressure-Impulse Diagrams of RC Columns with or between 7.5 and 23.0% (Teychenne 1966) for 30–60 MPa concrete.
without FRP Strengthening Ellingwood (1978) reported a COV of 20.7% for in situ concrete
compressive strength. Wang et al. (2011) found that the COV of
A computer model was developed to calculate the responses of RC concrete strength is 18% when fcu ¼ 19.2 MPa, 14% when f cu ¼
columns with or without FRP strengthening to blast loads. The val- 26 MPa, and 12% when f cu ¼ 33.6 MPa. Val et al. (1997) sug-
idity of the model in predicting the structural column responses gested the COV of concrete compressive strength be estimated by
were verified with drop weight impact tests and field blasting tests.
Using the model, intensive numerical simulations were carried out COV ¼ 8=½1.64ðf cu Þ ð1Þ
to construct P-I diagrams of RC columns with or without FRP
strengthening under blast loading. These P-I diagrams will be used This gives a COV of 8% for 60-MPa concrete and 16% for
to model the RC column capacities to resist blast loads in this study. 30-MPa concrete. A previous study (Low and Hao 2001) conducted
For brevity, the Appendix provides formulae for constructing the a review of the available data and found that the average COV for
P-I diagrams. Mutalib and Hao (2011) provide detailed descrip- concrete compressive strength under normal construction quality
tions of the numerical model, its validity, and the derived P-I dia- control is 11%.
grams for RC columns without FRP strengthening, with different The previous study (Low and Hao 2001) reviewed the COV of
types of FRP strengthening, i.e., strip, wrap, and both strip and reinforcement strength, concrete and steel elastic modulus, and
wrap, as illustrated in Fig. 1, and different FRP thicknesses and structural dimensions. It was found that the average COV of steel
strengths. Fig. 2 shows the constructed P-I curves of RC columns yield stress is 0.08 and that of structural dimension is 0.03. Not
with and without FRP strengthening. The damage level D ¼ 0.2 many studies on the statistical variations of longitudinal and

© ASCE 04015017-3 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2016, 30(2): 04015017

 100000 Table 1. Statistical Variation (COV) of RC Column and FRP Parameters
D=0.8, FRP strip
RC column and FRP parameters COV
D=0.5, FRP strip
Width b (mm) 0.03
D=0.2, FRP strip Depth h (mm) 0.03
Pressure (kPa)

10000 D=0.8, unstrengthened Height H (m) 0.03

Concrete strength f cu (MPa) 0.11
D=0.5, unstrengthened Longitudinal reinforcement ratio ρ 0.10
D=0.2, unstrengthened Transverse reinforcement ratio ρs 0.15
FRP strip strength fstrip (MPa) 0.05
FRP wrap strength f wrap (MPa) 0.125
1000 FRP wrap thickness twrap (mm) 0.07
Downloaded from ascelibrary.org by Guru Nanak Dev Engineering College on 09/07/22. Copyright ASCE. For personal use only; all rights reserved.

200
1000 10000 100000 field-manufactured composites with average quality control for
(a) Impulse (kPa*ms) FRP tensile strength. Atadero and Karbhari (2005) assumed a range
of COV from 0.05 to 0.20 for FRP strength in their study. In gen-
100000 eral, FRP strips are more likely to be manufactured and applied
D=0.8, FRP wrap under higher quality controls when compared with FRP wrap.
D=0.5, FRP wrap The preceding review indicates significant variations of RC struc-
ture and FRP material properties. These are only the variations of the
D=0.2, FRP wrap
RC structure and FRP parameters under normal construction quality
D=0.8, unstrengthened
Pressure (kPa)

10000 control. Possible variations induced by deterioration and overstress-

D=0.5, unstrengthened ing for in-service structures are not considered. If these variations are
D=0.2, unstrengthened included, the COV values might be higher. However, the influences
of these variations are not considered in the present study.
For a reliability analysis, besides the mean value and standard
1000 deviation, information regarding the probability distribution of
each parameter and cross correlation among all of the parameters
is needed. Unfortunately, most of such information is not available.
200 Because of the lack of data, most previous studies assumed that
1000 10000 100000
each of the structural parameters has a normal distribution and is
(b) Impulse (kPa*ms)
statistically independent of one other. The present study also adopts
100000
these assumptions, i.e., all of the structural parameters are assumed
statistically independent of one other and normally distributed.
D=0.8, FRP wrap & strip
Table 1 lists the random variations of the RC column and FRP prop-
D=0.5, FRP wrap & strip erties used in this study. They are on the basis of the available stat-
D=0.2, FRP wrap & strip istical data in the literature reviewed previously. The yield stress of
the reinforcement bars is assumed to be 550 MPa as a deterministic
Pressure (kPa)

D=0.8, unstrengthened
10000
D=0.5, unstrengthened
parameter in this study. The effects of random variations of steel
reinforcement are implicitly accounted for by assuming a relatively
D=0.2, unstrengthened
large COV for reinforcement ratios, as given in Table 1.

1000 Statistical Variations of Blast Loadings

A number of computer tools, design charts, empirical formulae, and

200 computational fluid dynamics models can be used to estimate blast
1000 10000 100000
(c) loads for structural response analysis and design. Among them, the
Impulse (kPa*ms)
design charts in TM5-1300 (U.S. Department of the Army 1990) or
Fig. 2. Comparison of P-I curves of unstrengthened and FRP- UFC (DoD 2008) are most commonly used to estimate the reflected
strengthened RC column: (a) column strengthened with 2,280-MPa peak pressure Pr and reflected impulse I r corresponding to a 2,4,6-
1-mm-thick FRP strips; (b) column strengthened with 2,280-MPa trinitrotoluene (TNT) equivalent charge weight and standoff dis-
3-mm-thick FRP wrap; (c) column strengthened with 2,280-MPa 1-mm- tance. These Pr and I r values can then be used together with a
thick strips and 3-mm wrap P-I curve, as shown in Fig. 2, to predict the expected damage levels
of the column with or without FRP strengthening. The approach is
straightforward, but as reviewed previously, the actual blast loadings
might vary from what are estimated from a chosen design chart or
transverse reinforcement ratios can be found in the open literature. empirical relation owing to many uncertainties involved in modeling
In this study, they are assumed to have a normal distribution, with or measuring an explosion process, shock wave propagations, and
their design value being the corresponding mean value, and an interaction with structures.
assumed coefficient of variation of 0.10 for the longitudinal rein- In a previous study (Hao et al. 2010), the mean and coefficient of
forcement ratio and 0.15 for the transverse reinforcement ratio. variation of the peak reflected pressure and duration estimated from
Similarly, only limited studies on the variability in FRP properties nine empirical relations and design charts, including those given in
exist. Wang et al. (2010) reported a range of COVs from 0.04 for UFC and ConWep as a function of scaled distance, are derived. Fig. 3
shop-manufactured composites with high quality control to 0.24 for shows the mean and standard deviation and their corresponding

© ASCE 04015017-4 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2016, 30(2): 04015017

 500000
log10 σPr ¼ 3.03 − 3.533log10 ðZÞ þ 0.4534½log10 ðZÞ2
100000 Mean value þ 0.3248½log10 ðZÞ3 þ 0.07896½log10 ðZÞ4 ;
Standard deviation
10000 0.3 m=kg1=3 ≤ Z ≤ 40 m=kg1=3 ð3Þ
Pr(kPa)

1000 where Pr is in kPa; and Z = scaled distance in m=kg1=3 . The

coefficient of variation of the peak reflected pressure could be
100 derived by
σ Pr
COVPr ¼ ð4Þ
10 μ Pr
Downloaded from ascelibrary.org by Guru Nanak Dev Engineering College on 09/07/22. Copyright ASCE. For personal use only; all rights reserved.

1 Those of the normalized positive blast loading duration are

0.2 0.5 1 5 10 50 (Hao et al. 2010)
1/3
Scaled Distance (m/kg )
log10 μtd =W 1=3 ¼ −0.00307 þ 1.2186log10 ðZÞ − 0.5207½log10 ðZÞ2
10
5 − 0.2835½log10 ðZÞ3 þ 0.2132½log10 ðZÞ4 ;
0.3 m=kg1=3 ≤ Z ≤ 40 m=kg1=3 ð5Þ
td/W1/3 (ms/kg1/3 )

1
0.5 log10 σtd =W 1=3 ¼ −0.8433 þ 1.0982log10 ðZÞ − 0.8127½log10 ðZÞ2
þ 0.4214½log10 ðZÞ3 − 0.1046½log10 ðZÞ4 ;
0.1 0.3 m=kg1=3 ≤ Z ≤ 40 m=kg1=3 ð6Þ
0.05 Mean value
Standard deviation where td is in milliseconds; and W = TNT equivalent explosive
weight in kilograms. Again, the coefficient of variation of the nor-
0.01 malized positive blast loading duration is given by
0.2 0.5 1 5 10 50
Scaled Distance (m/kg1/3 )
σtd =W 1=3
COVtd =W 1=3 ¼ ð7Þ
μtd =W 1=3
Fig. 3. Mean and standard deviation of peak reflected pressure and
normalized positive loading duration and their best fitted functions Because only limited data are available, it is not likely to get an
unbiased prediction of the distribution types of the peak reflected
pressure and duration. In this study, the peak reflected pressure and
normalized duration are assumed to be statistically independent and
fitted functions of the peak reflected pressures and normalized have a normal distribution at each scaled distance, with the mean
positive blast loading duration. They are obtained as the ensemble and standard deviation defined previously. Assuming the triangular
means and standard deviations according to a simple statistics blast loading shape, the impulse is calculated by
analysis. As shown, when the scaled distance is smaller than
1.0 m=kg1=3 or larger than 10 m=kg1=3 , the variations of the peak 1
I r ¼ P r td ð8Þ
reflected pressure are very significant, with the standard deviations 2
comparable to the mean values. The variation of the normalized
positive loading duration is not as significant as that of the peak These reflected pressure and impulse values are used in the
reflected pressure, with the standard variation being approximately present study to calculate the failure probabilities of RC columns
10% of the mean value, i.e., the coefficient of variation is approx- with or without FRP strengthening. Negative-phase blast loading is
imately 0.1 throughout the scaled distance range considered in the neglected in the present study.
study. These observations indicate the significant fluctuations of the
blast loads estimated from the available empirical relations. These
values are valid only for an open field explosion and large enough Reliability Analysis
flat reflection surface. For a complex explosion scenario, such as an
explosion in a complex city environment, more significant varia- To assess the effectiveness of FRP strengthening in enhancing RC
tions are expected because of blast wave interaction with surround- column capacities in resisting blast loading, four types of columns
ing structures. However, without loss of generality, these blast are considered in this study. They are (1) unstrengthened RC col-
loading models are used in the present study to estimate blast load- umns as references for comparison, (2) an RC column strengthened
ing and its statistical variations. Prediction of blast loading and the with FRP strips, (3) an RC column strengthened with FRP wrap,
associated statistical variation in a complex environment is beyond and (4) an RC column strengthened with FRP strips and then FRP
the scope of the present study. wrap. Fig. 1 shows the configurations of RC columns with FRP
The corresponding best fitted functions of the peak reflected strips and wrap.
pressure and its standard deviation are given as (Hao et al. 2010) The computer program CALREL (Liu et al. 1989) is used to cal-
culate the failure probabilities of the RC columns at different dam-
age levels defined by the P-I diagrams, as given in the Appendix.
log10 μPr ¼ 3.651 − 3.018log10 ðZÞ þ 0.1967½log10 ðZÞ2 The program has the capability of calculating the various reliability
þ 0.8873½log10 ðZÞ3 − 0.3795½log10 ðZÞ4 ; indices, including the first-order second moment (FOSM) reliabil-
ity index, Hasofer-Lind (HL) reliability index, generalized reliabil-
0.3 m=kg1=3 ≤ Z ≤ 40 m=kg1=3 ð2Þ ity index, and full distribution reliability index if the cumulative

© ASCE 04015017-5 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2016, 30(2): 04015017

 probability density function is known. It also performs Monte Carlo Although each random parameter is assumed to have a normal
simulations. The definitions of these reliability indices can be found distribution and all of the parameters are assumed statistically in-
in any standard reliability textbook and the CALREL user manual dependent of one other, integration of Eq. (13) is not straightfor-
(Liu et al. 1989), which are therefore not introduced here. ward. In this study, the computer program CALREL (Liu et al.
Users need to define the limit state function and program and 1989) is used to calculate the failure probability. The preceding
link it to the CALREL main program as a user-provided subroutine limit state function, with P-I curves defined in Eqs. (19)–(31) and
for analysis. Considering the random variations of the RC column blast loadings defined in Eqs. (2)–(8), is programmed and linked to
properties listed in Table 1, the P-I curves defined by Eqs. (19)– CALREL as a user-defined subroutine. The CALREL program cal-
(31) in the Appendix vary accordingly. Similarly, the estimated culates failure probabilities according to the four reliability indices
reflected pressure Pr and impulse I r at each scaled distance vary. and Monte Carlo simulations as described previously. It was found
Taking all of these variations into account, the limit state function that the failure probabilities from all of these approaches yield very
Downloaded from ascelibrary.org by Guru Nanak Dev Engineering College on 09/07/22. Copyright ASCE. For personal use only; all rights reserved.

gðXÞ, corresponding to a given damage level defined by the respec- similar results because of the statistical independence and normal
tive P-I curve, is given as distribution assumption of the random variables. For this reason,
hereafter, only the failure probabilities for different damage levels
gðXÞ ¼ Uðρs ;ρ;f cu ;h;b;H;fstrip ;f wrap ;twrap ;Pr ;td Þ − U  ðPr ;td Þ ¼ 0 of RC columns are presented without differentiating the associated
ð9Þ reliability index.

where U depends on the random parameters of the RC column;

FRP = distance from the origin of the P-I diagram to the intersec- Failure Probabilities of RC Columns with and
tion point of the P-I curve with a line linking the origin of the P-I without FRP Strengthening
diagram to the random loading point (Pr , I r ), as shown in Fig. 4;
In this study, eight example columns are considered. They are three
and U  = distance from the P-I diagram origin to the point in the
RC columns without FRP strengthening designated as C1, C2, and
P-I diagram determined by the peak reflected pressure and impulse
C3 and five FRP-strengthened RC columns. To study the effective-
of blast loading. Because of random variations of the column
ness of FRP strengthening, Concrete Column C2 is chosen as the
parameters and blast loading, as discussed previously, both U and
reference column, i.e., different FRP strengthening is applied to
U  vary randomly. They are defined as
RC Column C2 and designated as C2FRP1–C2FRP5, in which
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi C2FRP1, C2FRP2, and C2FRP3 represent Column C2 being
U  ðPr ; I r Þ ¼ P2r þ I 2r ð10Þ strengthened with FRP strips and wrap of different thicknesses and
C2FRP4 and C2FRP5 represent Column C2 being strengthened
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi by FRP strips and wrap of the same thickness but different FRP
UðPI ; I I Þ ¼ P2I þ I 2I ð11Þ strength. Table 2 gives the mean values of these column parameters.
Substituting these mean values into Eqs. (20)–(31), the pressure
and impulse asymptotes corresponding to different damage levels
The intersection point (PI , I I ) is straightforwardly derived from can be calculated, which are given in Table 3. These asymptotes can
Eq. (19) as be used to construct P-I diagrams according to Eq. (19). As can be
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi noticed, the pressure and impulse asymptotes of FRP-strengthened
P ðI 0 þ PIrr P0 Þ þ ðI 0 − PIrr P0 Þ2 þ 48 PIrr ðP20 þ I20 Þ1.5 columns are larger than those of the unstrengthened column and
PI ¼ r I I ; II ¼ increase with FRP strength and thickness, indicating the effective-
Ir 2
ness of using FRP to strengthen RC columns for blast-loading
ð12Þ resistance. In this study, the TNT equivalent charge weight of
1,000 kg is assumed in all of the simulations, representing a van
where P0 and I 0 are defined in Eqs. (20)–(31) for different damage or a light truck bomb (FEMA 2003). The explosion effect on struc-
levels. They depend on the random variables of the RC column tures depends on both explosion charge weight and stand-off
parameters. distance. At the same scaled distance, a smaller charge weight cor-
With the limit state function, the failure probability can be cal- responds to a short distance and hence short blast loading duration
culated by and smaller impulse, as compared with those from an explosion of a
Z larger charge at a longer distance, although the peak reflected pres-
Pf ¼ P½gðXÞ ≤ 0 ¼ f x ðxÞdx ð13Þ sure is the same. As a result, the failure probability of the structure
gðXÞ≤0 at the same scaled distance from a smaller charge weight is smaller

Fig. 4. Formation of limit-state function

© ASCE 04015017-6 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2016, 30(2): 04015017

 Table 2. Mean Values of the Eight Example Columns 1
0.9
f cu d b H f strip f wrap twrap
Column ρs P (MPa) (mm) (mm) (mm) (MPa) (MPa) (mm) 0.8

Failure Probability
0.7
C1 0.006 0.01 30 400 400 3,600 — — —
C2 0.016 0.02 40 600 600 4,600 — — — 0.6
C3 0.032 0.03 50 800 800 5,500 — — — 0.5
C2FRP1 0.016 0.02 40 600 600 4,600 1,700 1,700 2
0.4
C2FRP2 0.016 0.02 40 600 600 4,600 1,700 1,700 4 Damage Level D=0.2
C2FRP3 0.016 0.02 40 600 600 4,600 1,700 1,700 6 0.3
Column 1
C2FRP4 0.016 0.02 40 600 600 4,600 1,000 1,000 4 0.2 Column 2
C2FRP5 0.016 0.02 40 600 600 4,600 2,400 2,400 4 Column 3
0.1
Downloaded from ascelibrary.org by Guru Nanak Dev Engineering College on 09/07/22. Copyright ASCE. For personal use only; all rights reserved.

0
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
1/3
Scaled Distance (m/kg )
Table 3. Mean P0 and I 0 Values for the Eight Example Columns
1
Damage D ¼ 0.2 Damage D ¼ 0.5 Damage D ¼ 0.8
0.9
P0 I0 P0 I0 P0 I0 0.8
Column (kPa) (kPa · ms) (kPa) (kPa · ms) (kPa) (kPa · ms)

Failure Probability
0.7
C1 605 1,096 847 1,869 1,180 2,094
0.6
C2 1,095 3,256 1,535 3,997 1,639 4,849
C3 1,661 5,700 2,354 6,318 2,179 7,651 0.5
C2FRP1 1,197 3,572 1,753 4,216 1,797 5,061 0.4
C2FRP2 1,264 3,840 1,862 4,468 1,857 5,483 Damage Level D=0.5
0.3
C2FRP3 1,375 4,336 1,971 5,008 1,939 6,749 Column 1
C2FRP4 1,207 3,828 1,830 4,195 1,820 5,107 0.2 Column 2
Column 3
C2FRP5 1,351 3,853 1,893 5,117 1,902 6,409 0.1
0
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
Scaled Distance (m/kg1/3 )
than that from a larger charge. In this study, however, only a
1,000-kg explosion charge weight is considered. 1
Fig. 5 shows the probability of failure of unstrengthened RC 0.9
columns (C1–C3) for the three damage levels defined previously. 0.8
As shown, increasing the column size and reinforcement ratio
Failure Probability

0.7
greatly enhances its capacity to resist blast loads. Taking Column
0.6
C1 as the example, its failure probability, which corresponds to
damage level D ¼ 0.8, indicating sure collapse of this column 0.5

at a scaled distance of 0.9 m=kg1=3 , is 1.0, whereas the larger 0.4

Damage Level D=0.8
Column C3 experiences the same failure probability only at a 0.3
Column 1
scaled distance of 0.4 m=kg1=3 , implying significant enhancement 0.2 Column 2
of the blast loading resistance capacity. The results also show that 0.1
Column 3
Column C1 is not likely to suffer any damage, i.e., D < 0.2, when
0
the scaled distance is larger than 2.3 m=kg1=3 , whereas Column C3 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
will not suffer any damage when the scaled distance is larger than Scaled Distance (m/kg1/3 )
1.8 m=kg1=3 . The results indicate that the three unstrengthened RC
columns are likely to experience a certain level of damage and even Fig. 5. Probability of unstrengthened RC column failure at different
collapse, depending on the column size and reinforcement ratio damage levels
when the scaled distance is less than 1.8 m=kg1=3 . When the scaled
distance is larger than 2.3 m=kg1=3 , the three example RC columns
are not likely to suffer any damage from blast loading.
Fig. 6 shows the probabilities for the three damage levels of the of FRP strengthening. By comparing the results in Figs. 5 and 6, it
unstrengthened RC Column C2 and that of the strengthened with is obvious that similar effects can be achieved by either increasing
FRP columns of different thicknesses at different scaled distances. the column size and reinforcement ratios or by using FRP strength-
As expected, FRP strengthening increases the column capacity ening. As shown, the effect of increasing the column dimension
to resist blast loads. Increasing the FRP thickness results in a fur- from 600 × 600 to 800 × 800 mm and increasing the reinforcement
ther increase in the blast load resistance capacity of the column. ratio from 0.016 to 0.032 is similar as strengthening the column
As shown, the sure collapse (D ¼ 0.8) for Column C2 without with 6-mm FRP. Increasing the column size and reinforcement ratio
strengthening occurs when the scaled distance is less than or equal is slightly more effective than using 6-mm FRP strengthening. This
to approximately 0.7 m=kg1=3. This is reduced to 0.5 m=kg1=3 is because increasing the column size greatly increases its mass and
when the column is strengthened with 6-mm-thick FRP wrap. hence inertial resistance. Increasing the size and reinforcement ratio
When the column is not strengthened, it is safe and not likely also greatly enhances its flexural and shear resistance capacities.
to suffer any damage when the scaled distance is larger than The drawback of increasing the column size is that it will take more
1.6 m=kg1=3 or larger than 1.5 m=kg1=3 when it is strengthened space, and bulky columns are not necessarily aesthetically appeal-
with 6-mm FRP. These observations demonstrate the effectiveness ing and architecturally acceptable. Strengthening with FRP avoids

© ASCE 04015017-7 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2016, 30(2): 04015017

 1 1
0.9 0.9
0.8 0.8

Failure Probability
Failure Probability

0.7 0.7
0.6 0.6

0.5 0.5

0.4 0.4
Damage Level D=0.2 Damage Level D=0.2
0.3 0.3
Column 2 No FRP Column 2 No FRP
0.2 Column 2 FRP1 (2mm) 0.2 Column 2 FRP4
Column 2 FRP2 (4mm) Column 2 FRP2
0.1 0.1 Column 2 FRP5
Column 2 FRP3 (6mm)
Downloaded from ascelibrary.org by Guru Nanak Dev Engineering College on 09/07/22. Copyright ASCE. For personal use only; all rights reserved.

0 0
0.8 1 1.2 1.4 1.6 1.8 2 0.8 1 1.2 1.4 1.6 1.8
Scaled Distance (m/kg ) 1/3 Scaled Distance (m/kg 1/3)

1
1
0.9
0.9 Damage Level D=0.5
Column 2 No FRP 0.8
0.8
Column 2 FRP1 (2mm)

Failure Probability
Failure Probability

Column 2 FRP2 (4mm) 0.7

0.7
Column 2 FRP3 (6mm) 0.6
0.6
0.5
0.5
0.4
0.4
Damage Level D=0.5
0.3 0.3
Column 2 No FRP
0.2 Column 2 FRP4
0.2
Column 2 FRP2
0.1 0.1 Column 2 FRP5
0 0
0.8 1 1.2 1.4 1.6 1.8 2 0.8 1 1.2 1.4 1.6
Scaled Distance (m/kg1/3 ) Scaled Distance (m/kg 1/3)

1 1
0.9 Damage Level D=0.8 0.9
Column 2 No FRP
0.8 0.8
Column 2 FRP1 (2mm)
Failure Probability

Failure Probability

0.7 Column 2 FRP2 (4mm) 0.7

Column 2 FRP3 (6mm)
0.6 0.6
0.5 0.5
0.4 0.4
0.3 Damage Level D=0.8
0.3
Column 2 No FRP
0.2 0.2 Column 2 FRP4
0.1 Column 2 FRP2
0.1 Column 2 FRP5
0
0.8 1 1.2 1.4 1.6 1.8 2 0
0.8 1 1.2 1.4 1.6
Scaled Distance (m/kg1/3)
Scaled Distance (m/kg 1/3)

Fig. 6. Failure probability of unstrengthened and FRP-strengthened

Fig. 7. Failure probabilities of unstrengthened and FRP-strengthened
RC columns with different FRP thickness at different damage levels
RC columns with different FRP strength at different damage levels

these shortcomings and it is also likely to increase the service life of on column threshold failure probabilities. This is because at this
the column because FRP is more corrosion resistant than reinforce- level, only concrete material experiences some damage and no
ment bars. Moreover, the construction cost of FRP strengthening is damage to FRP sheets occurs. The FRP wrap provides confinement
relatively low, although the FRP material cost is high. These ob- to concrete and therefore the blast loading resistance capacity in-
servations demonstrate the advantages of using FRP strengthening creases. At D ¼ 0.5 and 0.8, i.e., the intermediate and severe dam-
of RC columns to resist blast loads. age levels, both concrete material damage and FRP rupture occur,
Fig. 7 shows the probabilities of damage of the unstrengthened and increasing FRP strength and thickness therefore has a pro-
and FRP-strengthened RC Column C2 with different FRP strength nounced effect on column failure probabilities. Increasing the
at three damage levels. As shown, FRP strengthening is effective in FRP strength from 1,000 to 2,400 MPa reduces the corresponding
increasing the column blast loading resistance capacity, but increas- scaled distances for the failure probability of 1.0 of damage level
ing the FRP strength has different influences on columns at differ- 0.8 from 0.6 to 0.5 m=kg1=3 .
ent failure probabilities corresponding to the different damage As discussed previously, blast loading prediction displays more
levels, as also indicated in Table 3. At D ¼ 0.2, the minor damage significant statistical variations than RC column and FRP material
level, increasing the FRP strength has only an insignificant effect parameters. It was found in the previous study of RC column failure

© ASCE 04015017-8 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2016, 30(2): 04015017

 1
probabilities that the statistical variations in blast loading predic-
tions contribute most to the variations in failure probability estima- 0.9
tions, i.e., the failure probabilities of the RC columns obtained by 0.8
assuming deterministic RC column parameters are close to those

Failure Probability
0.7
obtained by considering random variations in both the blast loading
0.6
and column parameters, whereas assuming deterministic blast load-
ings results in very different predictions of column damage prob- 0.5
abilities (Hao et al. 2010). The observation made in the latter study 0.4
Column 2 No FRP D=0.8
considered the statistical variations of RC column parameters. In 0.3 Random
the present study, with the involvement of statistical variations 0.2 Load Deterministic
of FRP material properties, a greater number of column parameters Column Deterministic
0.1
Downloaded from ascelibrary.org by Guru Nanak Dev Engineering College on 09/07/22. Copyright ASCE. For personal use only; all rights reserved.

experiences statistical variations. To investigate the influence of

statistical variations in blast loading predictions (Pr and I r ) and 0
0.8 1 1.2 1.4 1.6
column and FRP parameters (ρs ; ρ; f cu ; h; b; H; f strip ; fwrap , and
Scaled Distance (m/kg1/3)
twrap ) on failure probabilities, similar analyses, i.e., calculations
of failure probabilities by considering deterministic blast loading 1
or deterministic structural parameters are also carried out. Fig. 8 0.9
shows the failure probabilities of the three columns corresponding
0.8
to a damage level of 0.8. As shown, variations in blast loadings

Failure Probability
dominate the influences of random variations on failure probabil- 0.7
ities. Failure probabilities obtained by considering only the statis- 0.6
tical blast loading variations are very similar to those obtained by 0.5
considering statistical fluctuations in both the blast loading and 0.4
column parameters, indicating that the statistical variations in col- Column 2 FRP2 D=0.8
0.3 Random
umn parameters can be neglected. However, these results are ob-
0.2 Load Deterministic
tained because of the more significant statistical variations in blast Column Deterministic
loading than in column parameters. In certain cases, for example, 0.1
when considering structural property deterioration and/or poor con- 0
struction quality control, more intensive fluctuations of structural 0.8 1 1.2 1.4 1.6
parameters are expected and different results might be observed. Scaled Distance (m/kg1/3)

1
0.9
Collapse Probabilities of an Example RC Frame
0.8
Structure
Failure Probability

0.7
The preceding component reliability analysis is extended to a 0.6
system reliability analysis to estimate the collapse probability of
0.5
an example frame structure subjected to the blast attack shown
in Fig. 9. The two-bay 6-story RC frame structure is designed ac- 0.4
Column 2 FRP5 D=0.8
cording to the Australian Standards AS3600 (2009) and AS1170 0.3 Random
(2002). Beams and columns of the same cross sections and rein- 0.2 Load Deterministic
forcement ratios are assumed for all stories. Fig. 10 shows the cross Column Deterministic
0.1
section of the column and beam. Table 4 lists the column param-
0
eters. The three ground floor columns are designated as FC1–FC3, 0.8 1 1.2 1.4 1.6
as shown in the figure. The yield strength of steel reinforcement Scaled Distance (m/kg1/3)
is f y ¼ 550 MPa.
According to the Australian Standards AS3600 (2009) and Fig. 8. Failure probabilities of unstrengthened and FRP-strengthened
AS1170 (2002), static loading on the example structure includes columns obtained by considering either blast loading or column para-
the weight of the concrete columns, beams, and a 160-mm concrete meters deterministically
slab for each floor (all with an assumed weight of 25 kg=m3 ), a live
load of 4 kPa, and a uniform dead load (UDL) of 4.5 kN=m rep-
resenting the walls. All loads on the building are unfactored in this
study. The live load and slab weight are distributed to the surround- The capacity of the concrete column is represented by a strength
ing beams assuming two-way slab design, which is appropriate be- interaction diagram in accordance with AS3600 (2009), in which
cause the dimensions are 8 × 10 m. four points in the diagram bound the capacity region. They are the
For the system reliability analysis of progressive collapse, only squash load point (considers only axial loads), decompression point
the bottom floor of the structure is considered, as shown in Fig. 9; (zero tension in concrete), balanced point (concrete crushing in
thus, the loading on this system consists of a UDL of 21.5 kN=m compression and steel yielding in tension), and pure bending point
along the entire beam, and point loads of 1,255, 2,066, and (considers only bending). For the column section defined in Fig. 10,
1,255 kN over the three columns. Because structural collapse usu- these four points can be easily calculated. In this study, without loss
ally occurs after the action of the blast loading owing to a very short of generality, the concrete strength fcu is assumed to deteriorate
loading duration, collapse of the structure is assumed to take place with column damage
if the blast-damaged columns are not strong enough to resist these
static loads. fcu−d ¼ ð1 − DÞf cu ð14Þ

© ASCE 04015017-9 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2016, 30(2): 04015017

 Table 4. Statistical Parameters of the RC Column
Statistical fcu
1255kN 2066kN 1255kN parameters ρs P (MPa) d (mm) b (mm) H (mm)
21.5kN/m μ 0.006 0.011 32 400 400 4,000
COV 0.15 0.1 0.11 0.03 0.03 0.03
6x4=24m Distribution Normal Normal Normal Normal Normal Normal
FC1 FC2 FC3

yielding moment shown in Fig. 11 needs be scaled down according

to the moment magnifier value, i.e., the yielding moment shown in
Downloaded from ascelibrary.org by Guru Nanak Dev Engineering College on 09/07/22. Copyright ASCE. For personal use only; all rights reserved.

8m 8m 10m
the figure should be divided by the moment magnifier. For the f cu
corresponding to the three damage cases given in Table 5, the
Fig. 9. Example RC frame building (dimension not to scale)
moment magnifiers are calculated to be 1.27, 1.52, and 2.64,
respectively.
Using the column capacity corresponding to the three damage
The column load-carrying capacities then decrease with the levels determined previously, failure of the frame structure can be
damage. For the three damage levels considered in this study, evaluated. In this study, elements in the example building are as-
Table 5 gives the corresponding parameters for the column strength sumed to be ductile; thus, failure of the columns will not neces-
interaction diagram, and Fig. 11 shows the interaction diagrams. sarily result in failure of the system. This is a fair assumption
Similarly, the flexural capacity of the beam shown in Fig. 10 is and is in fact specified in many design manuals, such as FEMA
calculated to be 221 kN · m. This is used to assess the collapse of (2003). Thus, in the analysis, if static loading from the structure
the frame. If the bending moment in the beam exceeds its capacity, is greater than the capacity of the failed column, the system is
collapse of the frame is assumed. remodeled such that the column is only providing its yielding
For a concrete column, a moment magnifier (δb ) also needs to be force/moment determined from Fig. 11 and the moment magnifier
considered when assessing its capacity to account for the secondary calculated previously. Moment redistribution in the frame occurs. If
effect. According to AS3600 (2009), δ b is calculated by the bending moment in the beam exceeds its capacity, collapse of
the frame occurs. In this study, direct blast loading on the structure
km is assumed to only influence the columns and have no effect on the
δb ¼  ≥ 1.0 ð15Þ
1 − NN c capacity of the beam. Although this is not an accurate assumption
because the beams might also suffer structural damage because of
where blast loading, damage to the beam from direct blast loading is ex-
pected to be sufficiently less severe than damage to the column
M1
km ¼ 0.6 − 0.4 ð16Þ owing to a significantly larger inertia resistance from the floor slab.
M2 Therefore, it is an acceptable assumption and is commonly used in
analyzing blast-induced progressive collapse of structures.
and
Damage to the column will reduce its load-carrying capacity, as
π2 182d0 ΦM c shown in Fig. 11. Complete damage (D ¼ 0.8) of a column does
Nc ¼ ð17Þ not necessarily lead to collapse of the frame, whereas low and/or
l2e 1 þ β d
medium damage to two or three columns might lead to structural
where N  = design axial load (kN); N c = buckling load (kN); M 2 collapse. To determine the possible structural failure cases, static
and M 1 = design moments at the top and bottom of the column, analyses are carried out according to the procedure outlined previ-
respectively; le = column effective length (m); d0 = distance from ously, with all the possible combinations of column damages,
extreme compressive fiber of the concrete to the centroid of the i.e., three columns each with damage equal to 0.2, 0.5, and 0.8,
outermost layer of tensile reinforcement (m); M c = buckling mo- to derive the moment distributions in the frame. A detailed static
ment (kN · m); coefficient Φ ¼ 0.6; and β d = factor, which equals analysis is not described here because it is straightforward and
zero in this case. When assessing the capacity of the column, the can be performed with the standard structural mechanics method.

Fig. 10. Cross section view of (a) beam; (b) column of example RC frame

© ASCE 04015017-10 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2016, 30(2): 04015017

 Table 5. Data for Axial-Moment Diagram of the RC Column
fcu ¼ 32 MPa (D ¼ 0) f cu ¼ 25.6 MPa (D ¼ 0.2) f cu ¼ 16 MPa (D ¼ 0.5)
Data point Mu (kN · m) Nu (kN) Mu (kN · m) Nu (kN) Mu (kN · m) Nu (kN)
Squash load point 0.00 5,298.04 0.00 4,437.48 0.00 3,146.65
Decompression point 235.93 3,544.50 200.98 2,924.26 148.55 1,993.90
Balanced point 322.72 1,541.49 281.85 1,205.70 220.55 702.01
Pure bending point 157.36 0.00 153.85 0.00 147.67 0.00

5400 FC1 is 0.5 and that of the other two columns is 0.1. Unfortunately,
Downloaded from ascelibrary.org by Guru Nanak Dev Engineering College on 09/07/22. Copyright ASCE. For personal use only; all rights reserved.

4800 fcu=32MPa (D=0.0) Mutalib and Hao (2011) define only three column damage levels.
fcu=25.6MPa (D=0.5) Therefore, the collapse analysis can only be performed on the basis
Axial Load Capacity (kN)

4200 fcu=16MPa (D=0.8) of the combinations of these three damage levels of the three col-
3600 umns. In the present study, the column damage combinations that
lead to frame collapse are listed, in which any value indicates frame
3000
collapses irrespective of the three damage levels of the particular
2400 column. In the collapse probability calculation, component damage
1800 probability corresponding to the damage level D ¼ 0.2 is taken be-
cause it gives a larger collapse probability among the three possible
1200
component failure probabilities, corresponding to damage levels at
600 0.2, 0.5, and 0.8.
0 Combining the three collapse combinations results in a typical
0 40 80 120 160 200 240 280 320 360 series system in the system reliability analysis (Stewart and
Moment Capacity (kNm)
Melchers 1997). Denoting the limit state functions of the three
Fig. 11. Axial force-bending moment interaction diagram of column columns as g1 ðXÞ, g2 ðXÞ, and g3 ðXÞ, respectively, each of them
is defined by Eq. (9), and the structural collapse probability is

Y
3

Table 6 lists the combinations of column damage that will lead to Pf ¼ 1 − ð1 − Pfi Þ ¼ 1 − ð1 − Pf1 Þð1 − Pf2 Þð1 − Pf3 Þ
i¼1
structural failure.
The failure combinations listed in Table 6 are the possible dam- ð18Þ
age conditions that will lead to frame collapse. They indicate, for
example, if the damage level in Column FC1 is 0.5 or above, irre- where Pfi = damage probability of the ith column, which is sim-
spective of the damage level of the other two columns, the structure ilarly calculated using the preceding component reliability method
will collapse because the redistributed bending moment in the beam by specifying the damage level for each column given in Table 6.
under static loading is larger than the beam capacity. Conversely, In this study, without loss of generality, two explosion cases, as
even all three columns experience damage, but if the damage level indicated in Fig. 12, are considered. Case 1 assumes the explosion
in Column FC2 is less than 0.8 and less than 0.5 in the two corner occurs directly in front of the center column, and Case 2 occurs in
columns, structural collapse will not occur. In this study, structural front of Column FC1. The scaled distance is measured as the short-
collapse is assumed when the bending moment in the beam exceeds est distance from the explosion center to the frame. Blast loading
the beam capacity. Contribution from the slab to resist structural and its variation are calculated according to Eqs. (2)–(7) with modi-
collapse is not considered, which may lead to overprediction of fication because of different angles of incidence, as specified in the
the structural collapse probability. In the current study, only three UFC guidelines (DoD 2008). Possible further variation in blast
component damage levels are considered in deriving the damage loading prediction owing to an oblique incident blast wave along
combination. This may also lead to some errors in estimating the column height is not considered. Fig. 13 shows the collapse
the collapse probability. For example, through a static moment re- probability of the structure under two explosion scenarios at differ-
distribution analysis, it is found that when the damage level of FC1 ent scaled distances obtained by considering all parameters ran-
is 0.5, the frame collapses irrespective of the damage levels of the domly or assuming either blast loading or structural parameters
other two columns, i.e., 0.2, 0.5, or 0.8. In a real case, collapse deterministically. As shown again, neglecting random variations
might also occur when the damage level of FC1 is 0.4 and that of structural parameters in the analysis gives only slightly different
of the other two columns is 0.25 or when the damage level of collapse probabilities, but neglecting random variations in blast
loading may greatly underestimate the collapse probability of

Table 6. Combinations of Column Failure to Cause Structural Collapse

Failure
combination Column FC1 Column FC2 Column FC3
FC3
1 D ≥ 0.5 Any value Any value FC1 FC2
2 Any value D ≥ 0.8 Any value d2 d1
3 Any value Any value D ≥ 0.5
Case 2 Case 1
Note: Any value indicates that system collapse will occur irrespective of the
damage level of the column. In this study, D ¼ 0.2 is used in the
Fig. 12. Two explosion scenarios
calculations.

© ASCE 04015017-11 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2016, 30(2): 04015017

 1 1
0.9 0.9
0.8 0.8

Failure Probability
Failure Probability

0.7 0.7
0.6 0.6
0.5 0.5
0.4 0.4
Case 1 Column 2 Case 1
0.3 0.3
Random Column 2 Case 2
0.2 Load Deterministic 0.2 Column C2FRP2 Case 1
Column Deterministic Column C2FRP2 Case 2
0.1 0.1
Downloaded from ascelibrary.org by Guru Nanak Dev Engineering College on 09/07/22. Copyright ASCE. For personal use only; all rights reserved.

0 0
1.2 1.4 1.6 1.8 2 0.8 1 1.2 1.4 1.6
Scaled Distance (m/kg1/3 ) Scaled Distance (m/kg1/3)

1 Fig. 15. Collapse probability of example frame with unstrengthened

0.9 and FRP-strengthened RC column
0.8
Failure Probability

0.7
center imposes a lower risk to the structure than that at the corner
0.6 of the structure.
0.5 To assess the effectiveness of FRP strengthening in reducing
0.4 the structural collapse probability, the preceding analysis can be
0.3
Case 2 repeated to consider the cases of columns with or without FRP
Random strengthening. For brevity, in the present study, the columns of
0.2 Load Deterministic
Column Deterministic
the example frame are replaced by Column C2 without FRP
0.1 strengthening and C2FRP2 with FRP strengthening, as defined pre-
0 viously. The combinations of column failure to cause structural
1.2 1.4 1.6 1.8 2 collapse are assumed to be the same, as given in Table 6. This
Scaled Distance (m/kg1/3 ) is a reasonable assumption because, as discussed previously, the
collapse of the frame is induced owing to an insufficient beam
Fig. 13. Collapse probabilities of example structure under two explo-
flexural capacity owing to static moment redistributions after col-
sion scenarios
umn damage. Because the beams of the example frame are the same
and Eq. (14) defines the capacity reduction of the column after
damage in the present study, which is directly proportional to
the structure owing to significant blast loading variations, as dis- Damage Level D, the combinations of column failure to induce
cussed previously. Fig. 14 compares the collapse probabilities of frame collapse therefore remain the same.
the structure under two explosion scenarios. As shown, at a scaled Fig. 15 shows the collapse probabilities of the building frame
distance larger than 1.5 m=kg1=3 , with a collapse probability of with C2 and C2FRP2 columns. As shown, strengthening the col-
less than 95%, both explosion cases give almost identical collapse umns reduces the collapse probability of the example frame. When
probabilities. However, when the scaled distance is larger than the scaled distance is small, i.e., the standoff distance is relatively
1.5 m=kg1=3 , explosion scenario Case 1 induces a relatively lower short, the collapse probability corresponding to the Case 1 explo-
collapse probability, indicating that the explosion occurring at the sion scenario is slightly larger. When the scaled distance is large,
the collapse probability of explosion scenario Case 2 is higher. This
is because the frame structure is more sensitive to the damage of the
corner column. When the corner column suffers medium level dam-
1 age, moment redistribution results in frame collapse, but the
0.9 medium level damage to the central column alone would not lead
0.8 to frame collapse, as indicated in Table 6. Compared with the re-
sults in Fig. 14, it is obvious that increasing the column size and
Failure Probability

0.7
reinforcement ratio is very effective in protecting the frame struc-
0.6
ture against blast loads.
0.5
0.4
0.3
Case 1 Discussion and Comparison
Case 2
0.2 On the basis of the preceding results, the ranges of scaled distances
0.1 that cause minor (D ¼ 0.2) to severe (D ¼ 0.8) damage to the ex-
0 ample columns with and without FRP strengthening are extracted
1.2 1.4 1.6 1.8 2 and listed in Table 7. As shown, with consideration of random var-
Scaled Distance (m/kg 1/3 ) iations in blast loading predictions and column parameters, there is
a rather large scaled distance range for each column at each damage
Fig. 14. Comparison of collapse probabilities of two explosion
level. For the small Column C1 considered in the study, minor dam-
scenarios
age might occur when the scaled distance is less than 2.3 m=kg1=3

© ASCE 04015017-12 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2016, 30(2): 04015017

 Table 7. Range of Scaled Distances to Cause Minor to Severe Damages to the possible random variations in blast loading and RC column
Example Columns parameters is also demonstrated. The results presented in this study
D ¼ 0.2 D ¼ 0.5 D ¼ 0.8 can be used to better assess the performance of RC columns and RC
frame structures under different explosion scenarios than the range
Column Pf ¼ 0 Pf ¼ 1 Pf ¼ 0 Pf ¼ 1 Pf ¼ 0 Pf ¼ 1
of the effects chart given by FEMA.
C1 2.3 1.2 2.0 1.0 1.8 0.9
C2 1.8 1.0 1.7 0.8 1.6 0.7
C3 1.6 0.6 1.5 0.5 1.4 0.4
C2FRP1 1.8 0.8 1.6 0.7 1.5 0.6 Appendix. P-I Diagrams of RC Columns with or
C2FRP2 1.7 0.8 1.6 0.6 1.5 0.5 without FRP Strengthening
C2FRP3 1.7 0.7 1.5 0.6 1.5 0.5
C2FRP4 1.8 0.8 1.6 0.7 1.5 0.6 The P-I diagrams of unstrengthened and FRP-strengthened RC
Downloaded from ascelibrary.org by Guru Nanak Dev Engineering College on 09/07/22. Copyright ASCE. For personal use only; all rights reserved.

C2FRP5 1.7 0.7 1.6 0.6 1.5 0.5 columns can be expressed as (Mutalib and Hao 2011; Shi et al.
2008)
 
P I 1.5
and surely occurs when the scaled distance is less than ðP − PO ÞðI − I O Þ ¼ 12 O þ O ð19Þ
2 2
1.2 m=kg1=3 , whereas severe damage might occur when the scaled
distance is less than 1.8 m=kg1=3 and surely occurs when it is less where PO and I O = pressure and impulse asymptotes of the P-I
than 0.9 m=kg1=3 . Increasing the column size and reinforcement curves, respectively. The analytical formulae to estimate the pressure
ratio greatly enhances its blast loading resistance capacity. The cor- and impulse asymptotes corresponding to low damage (D ¼ 0.2),
responding scaled distances for the large Column C3 considered intermediate damage (D ¼ 0.5), and high damage (D ¼ 0.8) are
in the study reduce to 1.6 and 0.6 m=kg1=3 for minor damage derived as (Mutalib and Hao 2011)
and 1.4 and 0.4 m=kg1=3 for severe damage. Strengthening with
FRP also enhances the column blast loading resistance capacity PO ð0.2Þ ¼ 7.25fcu þ 2.37d − 0.147H − 0.414b þ 7,342.47ρ
of the column. The scaled distance range for severe damage of þ 10,073.44ρs þ α1 ð20Þ
the unstrengthened Column C2 is between 1.6 and 0.7 m=kg1=3 ,
which is reduced to 1.5 and 0.5 m=kg1=3 for the FRP-strengthened
Column C2FRP5. I O ð0.2Þ ¼ 25fcu þ 7.289d − 0.158H − 0.168b þ 19,261.3ρ
For quick determination of the required stand-off distance in a þ 44,864.881ρs − 2,398.62 þ α2 ð21Þ
given explosion scenario, FEMA (2003) recommends that building-
specific range-to-effect charts be developed. For various explosion
sources, FEMA (2003) gives the chart of blast range and effects PO ð0.5Þ ¼ 2fcu þ 3.174d − 0.217H − 0.445b þ 15,786.72ρ
ranging from 4.536 to 4,536 kg (10 to 10,000 lb). From the chart þ 18,137.95ρs þ 210 þ α3 ð22Þ
(FEMA 2003), it is easy to obtain that for the TNT-equivalent
charge weight of 1,000 kg (2,200 lb), the charge weight considered
in this study, the threshold stand-off distance for the concrete col- I O ð0.5Þ ¼ 27.5fcu þ 9.75d − 0.168H − 1.776b þ 13,121.77ρ
umn to fail is approximately 12.19 m (40 ft), which gives a scaled þ 29,433.94ρs − 1,848.178 þ α4 ð23Þ
distance of 1.22 m=kg1=3 . Although FEMA (2003) does not quan-
titatively define the RC column and failure of the column, this value
is in the scaled distance range obtained in the present study, indi- PO ð0.8Þ ¼ 11fcu þ 3.456d − 0.268H − 1.552b þ 14,753.44ρ
cating the consistency of the present results with that which FEMA þ 8,924.068ρs þ 851.90 þ α5 ð24Þ
derived primarily from observations in explosion events. With a
reliability analysis that takes into consideration the possible varia-
tions in blast loadings and structural parameters, the ranges of fail- I O ð0.8Þ ¼ 59fcu þ 13.16d − 0.43H − 0.26b þ 1,091.78ρ
ure probabilities at specific damage levels of small to large example þ 489.97ρs − 3,302.33 þ α6 ð25Þ
RC columns with and without FRP strengthening are generated
in this study. This gives a more accurate assessment of column where α1 ; α2 ; α3 ; α4 ; α5 ; α6 ¼ 0 for nonretrofitted RC columns. For
performance under explosion scenarios than the chart given by FRP-strengthened RC columns
FEMA (2003).
α1 ¼ expð0.000169fstrip þ 0.000423fwrap þ 0.252twrap þ 3.114Þ
ð26Þ
Conclusions

This paper presents both component and system reliability analyses α2 ¼ expð0.000163f strip − 0.000132fwrap þ 0.307twrap þ 5.09Þ
of RC structures subjected to blast loading. The component reli- ð27Þ
ability analyses are performed to calculate the failure probabilities
of columns with or without FRP strengthening. System reliability α3 ¼ 0.0539f strip − 0.00909f wrap þ 54.53twrap þ 32.302 ð28Þ
analyses are performed to calculate the collapse probabilities of an
example RC frame under two explosion scenarios. Random varia-
tions in blast loadings and RC column parameters are considered. α4 ¼ expð−0.00000295f strip þ 0.00124fwrap þ 0.382twrap þ 2.524Þ
The scaled distance ranges from zero probability to sure probabil- ð29Þ
ity, corresponding to the different damage levels of the RC columns
with and without FRP strengthening, are determined. The effective-
α5 ¼ expð0.000189fstrip þ 0.0000795f wrap 0.16twrap þ 4.286Þ
ness of FRP strengthening of the RC columns on their blast load-
carrying capacities is demonstrated. The importance of considering ð30Þ

© ASCE 04015017-13 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2016, 30(2): 04015017

 α6 ¼ expð0.0000868fstrip þ 0.0012f wrap þ 0.549twrap þ 2.068Þ Hyde, D. (1992). ConWep—Application of TM5-855-1, Structural Mechan-
ics Division, Structures Laboratory, U.S. Army Corps of Engineers
ð31Þ Waterway Experimental Station, Vicksburg, MS.
Isaac, P., Darby, A., Ibell, T., and Evernden, M. (2011). “Plasticity based
where the column width b, depth d, height H, and FRP wrap thickness approach for evaluating the blast response of RC columns retrofitted
twrap are all in millimeters; concrete strength f cu , FRP strip strength with FRP.” Int. J. Protective Struct., 2(3), 367–380.
f strip , and FRP wrap strength f wrap are all in MPa; and ρ and ρs = Khalifa, A., Alkhrdaji, T., Nanni, A., and Lansburg, S. (1999). “Anchorage
longitudinal and transverse reinforcement ratio of the column. of surface mounted FRP reinforcement.” Concr. Int., 21(10), 49–54.
These analytical formulae are valid for a reinforcement steel Kingery, C. N., and Bulmash, G. (1984). “Air blast parameters from TNT
strength of 550 MPa. For reinforcements with other strengths, the spherical air burst and hemispherical surface burst.” Ballistic Research
equivalent longitudinal and transverse steel area Ase should be used Laboratories, Aberdeen Proving Ground, MD.
when calculating the respective reinforcement ratio Lawver, D., Daddazio, R., Oh, G. J., Lee, C. K. B., Pifko, A. B., and
Downloaded from ascelibrary.org by Guru Nanak Dev Engineering College on 09/07/22. Copyright ASCE. For personal use only; all rights reserved.

Stanley, M. (2003). “Simulating the response of composite reinforced

fy floor slabs subjected to blast loading.” Proc., ASME 2003 Int. Mechani-
Ase ¼ A ð32Þ
550 s cal Engineering Congress and Exposition: Applied Mechanics and
Biomedical Technology, ASME, New York.
Liu, P. L., Der Kiureghian, A., and Lin, H. Z. (1989). CALREL user
Acknowledgments manual, Dept. of Civil Engineering, Univ. of California, Berkeley,
CA.
The authors gratefully acknowledge support from the Australian Low, H. Y., and Hao, H. (2001). “Reliability analysis of reinforced concrete
Research Council and University of Western Australia final-year slabs under explosive loading.” Struct. Saf., 23(2), 157–178.
student Thomas Nyman. The third author also acknowledges the fi- Low, H. Y., and Hao, H. (2002). “Reliability analysis of direct shear and
nancial support from the China National Nature Science Foundation flexural failure modes of RC slabs under explosive loading.” Eng.
(Grant Number 51008209) and Tianjin Nature Science Foundation Struct., 24(2), 189–198.
(Grant Number 12JCQNJC04800) for supporting this research. MacGregor, J. G., Mirza, S., and Ellingwood, B. (1983). “Statistical analy-
sis of resistance of reinforced and prestressed concrete members.” ACI
J. Proc., 80(3), 167–176.
Mirza, S. A., MacGregor, J. G., and Hatzinikolas, M. (1979). “Statistical
References descriptions of strength of concrete.” J. Struct. Div., 105(6), 1021–1037.
Mosalam, K. M., and Mosallam, A. S. (2001). “Nonlinear transient analysis
Atadero, R., and Karbhari, V. (2005). “Probabilistic based design for FRP
of reinforced concrete slabs subjected to blast loading and retrofitted
strengthening of reinforced concrete.” Proc., 7th Int. Symp. Fiber-
with CFRP composites.” Composites Part B, 32(8), 623–636.
Reinforced Polymer Reinforcement for Concrete Structures, Special
Publication SP-230, American Concrete Institute, Farmington Hills, MI. Muszynski, L., and Purcell, M. (2003a). “Composite reinforcement to
Baker, W. E., and Cox, P. (1983). Explosion hazards and evaluation, strengthen existing concrete structures against air blast.” J. Compos.
Elsevier Scientific, Amsterdam, Netherlands. Constr., 10.1061/(ASCE)1090-0268(2003)7:2(93), 93–97.
Beckmann, B., Hummeltenberg, A., Weber, T., and Curbach, M. (2011). Muszynski, L., and Purcell, M. (2003b). “Use of composite reinforcement
“Concrete under high strain rates: Local material and global structure to strengthen concrete and air-entrained concrete masonry walls against
response to impact loading.” Int. J. Protective Struct., 2(3), 283–294. air blast.” J. Compos. Constr., 10.1061/(ASCE)1090-0268(2003)7:2(98),
Bogosian, D., Ferritto, J., and Shi, Y. (2002). “Measuring uncertainty and 98–108.
conservatism in simplified blast models.” 30th Explosive Safety Seminar, Mutalib, A. A., and Hao, H. (2011). “Development of P-I diagrams for FRP
Atlanta. strengthened RC columns.” Int. J. Impact Eng., 38(5), 290–304.
Britt, J. R., Ranta, D. E., and Joachim, C. E. (2001). “BlastX code, version Netherton, M., and Stewart, M. (2010). “Blast load variability and accuracy
4.2, user’s manual.” ERDC/GSL TR-01-2, U.S. Army Engineer of blast load prediction models.” Int. J. Protective Struct., 1(4),
Research and Development Center, Vicksburg, MS. 543–570.
Brode, H. L. (1959). “Blast wave from a spherical charge.” Phys. Fluids, Razaqpur, G. A., Tolba, A., and Contestabile, E. (2007). “Blast loading
2(2), 217–229. response of reinforced concrete panels reinforced with externally
Buchan, P. A., and Chen, J. F. (2007). “Blast resistance of FRP composites bonded GFRP laminates.” Composites Part B, 38(5–6), 535–546.
and polymer strengthened concrete and masonry structures—A state-of- Sadovskii, M. A. (1952). Mechanical effects of air shock waves
the-art review.” Composites Part B, 38(5–6), 509–522. from explosions according to experiments, Izd Akaud Nauk SSSR,
Connolly, J. P. (1976). “Construction tolerances in reinforced concrete Moscow.
beams/joists.” ACI J. Proc., 73(11), 613–617. Shi, Y., Hao, H., and Li, Z.-X. (2008). “Numerical derivation of pressure–
Crawford, J. E., Malvar, L. J., Dunn, B. W., and Gee, D. J. (1996). “Retrofit impulse diagrams for prediction of RC column damage to blast loads.”
of reinforced concrete columns using composite wraps to resist blast Int. J. Impact Eng., 35(11), 1213–1227.
effects.” Karagozian and Case, Glendale, CA.
Standards Association of Australia. (2002). “Structural design actions.”
Crawford, J. E., Malvar, L. J., Morrill, K. B., and Ferritto, J. M. (2001).
AS1170, Sydney, NSW, Australia.
“Composite retrofits to increase the blast resistance of reinforced
Standards Association of Australia. (2009). “Concrete structures.” AS3600,
concrete buildings.” Proc., 10th Int. Symp. on Interaction of the Ef-
Sydney, NSW, Australia.
fects of Munitions with Structures, Woodhead Publishing, Sawston,
Cambridge. Stewart, M. G. (2006). “Risk assessment and optimization of blast mitiga-
Ellingwood, B. (1978). “Reliability basis of load and resistance factors for tion strategies for design and strengthening of built infrastructure.”
reinforced concrete design.” NBS building science series 110, U.S. Trans. Tianjin Univ., 12(z1), 8–15.
Dept. of Commerce, National Bureau of Standards, Washington, DC. Stewart, M. G. (2010). “Acceptable risk criteria for infrastructure
FEMA. (2003). “Reference manual to mitigate potential terrorist attacks protection.” Int. J. Protective Struct., 1(1), 23–40.
against buildings.” Risk management series, Washington, DC. Stewart, M. G., and Melchers, R. E. (1997). Probabilistic risk assessment of
Hao, H., Stewart, M., Li, Z.-X., and Shi, Y. (2010). “RC column engineering systems, Chapman & Hall, London.
failure probabilities to blast loads.” Int. J. Protective Struct., 1(4), Stewart, M. G., and Netherton, M. D. (2005). “Blast reliability curves and
571–591. uncertainty modelling for glazing subject to explosive blast loading.”
Henrych, J. (1979). Dynamics of explosion and its use, Elsevier Scientific, Proc., 6th Asian-Pacific Conf. on Shock and Impact Loads on Struc-
Amsterdam, Netherlands. tures, CI-Premier, Singapore.

© ASCE 04015017-14 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2016, 30(2): 04015017

 Teng, J., Chen, J., Smith, S., and Lam, L. (2003). “Behaviour and strength Val, D., Bljuger, F., and Yankelevsky, D. (1997). “Reliability assessment of
of FRP-strengthened RC structures: A state-of-the-art review.” Proc. damaged RC framed structures.” J. Struct. Eng., 10.1061/(ASCE)0733-
ICE Struct. Build., 156(1), 51–62. 9445(1997)123:7(889), 889–895.
Teychenne, D. C. (1966). “The variability of the concrete and its treatment Wager, P. (2005). SHOCK user’s manual version 1.0, Engineering Service
in codes of practice.” Struct. Concr., 3(1), 33–47. Center, Port Hueneme, CA.
Triantafillou, T. C. (1998). “Strengthening of structures with advanced Wang, N., Ellingwood, B., and Zureick, A. (2010). “Reliability-based
FRPs.” Prog. Struct. Eng. Mater., 1(2), 126–134. evaluation of flexural members strengthened with externally bonded
Twisdale, L. A., Sues, R. H., and Lavelle, F. M. (1994). “Reliability-based fiber-reinforced polymer composites.” J. Struct. Eng., 10.1061/(ASCE)
design methods for protective structures.” Struct. Saf., 15(1–2), 17–33.
ST.1943-541X.0000199, 1151–1160.
U.S. Department of Defence (DoD). (2008). “Structures to resist the effects
Wang, Y., Lin, F., and Gu, X. (2011). “Load and resistance factors for
of accidental explosion.” UFC 3-340-02, Washington, DC.
progressive collapse resistance design of reinforced concrete building
U.S. Department of State (DOS). (2009). Country reports on terrorism
2008, Washington, DC. structures.” Adv. Mater. Res., 338–344.
Downloaded from ascelibrary.org by Guru Nanak Dev Engineering College on 09/07/22. Copyright ASCE. For personal use only; all rights reserved.

U.S. Department of State (DOS). (2011). Country reports on terrorism Wu, C., Oehlers, D. J., Rebentrost, M., Leach, J., and Whittaker, A. S.
2010, Washington, DC. (2009). “Blast testing of ultra-high performance fibre and FRP-
U.S. Department of the Army. (1969). “Structures to resist the effects of retrofitted concrete slabs.” Eng. Struct., 31(9), 2060–2069.
accidental explosion.” AFR 88-22, Arlington, VA. Zhou, X. Q., Kuznetsov, V. A., Hao, H., and Waschl, J. (2008). “Numerical
U.S. Department of the Army. (1990). “Design of structures to resist the prediction of concrete slab response to blast loading.” Int. J. Impact
effects of accidental explosions.” TM5-1300, Arlington, VA. Eng., 35(10), 1186–1200.

© ASCE 04015017-15 J. Perform. Constr. Facil.

J. Perform. Constr. Facil., 2016, 30(2): 04015017