1-s2.0-S0143974X12001423-main

g of thin-walled steel structures against local buckling, and strengthening of hollow

. . . .ngtheniteel .. . . .. . . .. . . .. . . .

3.3.2. Bond-slip relationship .

Journal of Constructional Steel Research 78 (2012) 131143

Contents lists available at SciVerse ScienceDirect

Journal of Constructional Steel Research4.2. Intermediate debonding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1374.3. Other issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137

5. Fatigue strengthening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1376. Strengthening of steel structures against local buckling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

6.1. Buckling induced by high local stresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1386.2. Buckling induced by other loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

7. FRP connement of hollow steel tubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1388. FRP connement of concrete-lled steel tubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1399. Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140

9.1. Steel surface treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1419.2. Selection and formulation of adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

9.3. Bond behavior and debonding failu9.4. Fatigue strengthening . . . . . .9.5. FRP connement of tubular structu9.6. Other issues . . . . . . . . . .

Corresponding author. Tel.: +852 2766 6012.E-mail address: [email protected] (J.G. Teng).

0143-974X/$ see front matter 2012 Elsevier Ltd. Aldoi:10.1016/j.jcsr.2012.06.011. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

4. Flexural strengthening of steel beams . .

4.1. Plate end debonding . . . . . .Contents

1. Introduction . . . . . . . . .2. Appropriate use of FRP in the stre3. Bond behavior between FRP and s

3.1. General . . . . . . . .3.2. Adhesion failure . . . .3.3. Bond behavior . . . . .

3.3.1. Bond strength .research needs. 2012 Elsevier Ltd. All rights reserved.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132ng of steel structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135or concrete-lled steel tubes through external FRP connement. The paper concludes with comments on future

Composite materials steel structures, strengthenin

Retrot tween FRP and steel and its aReview

Strengthening of steel structures with ber-reinforced polymer composites

J.G. Teng a,, T. Yu b, D. Fernando c

a Department of Civil and Structural Engineering, The Hong Kong Polytechnic University, Hong Kong, Chinab School of Civil, Mining & Environmental Engineering, Faculty of Engineering, University of Wollongong, Northelds Avenue, Wollongong, NSW 2522, Australiac Institute of Construction and Infrastructure Management (IBI), Department of Structural, Environmental and Geomatic Engineering (D-BAUG), ETH Zrich, Zrich, Switzerland

a b s t r a c ta r t i c l e i n f o

Article history:Received 27 February 2012Accepted 29 June 2012Available online 30 July 2012

Keywords:Steel structuresFRP compositesStrengthening

Over the past two decades, ber-reinforced polymer (FRP) composites have gradually gainedwide acceptance incivil engineering applications due to their unique advantages including their high strength-to-weight ratio andexcellent corrosion resistance. In particular,manypossibilities of using FRP in the strengthening and constructionof concrete structures have been explored. More recently, the use of FRP to strengthen existing steel structureshas received much attention. This paper starts with a critical discussion of the use of FRP in the strengtheningof steel structures where the advantages of FRP are appropriately exploited. The paper then provides a criticalreview and interpretation of existing research on FRP-strengthened steel structures. Topics covered by the re-view include steel surface preparation for adhesive bonding, selection of a suitable adhesive, bond behavior be-

ppropriatemodeling,exural strengthening of steel beams, fatigue strengthening ofres . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1res . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

l rights reserved.41414141

Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

1. Introduction

Fiber-reinforced polymer (FRP) composites are formed by embed-ding continuous bers in a polymeric resinmatrixwhich binds the berstogether. Common bers used in FRP composites include carbon, glass,aramid and basalt bers while common resins are epoxy, polyester,and vinyl ester resins. The most widely used FRP composites are glassber-reinforced polymer (GFRP) composites and carbonber-reinforcedpolymer (CFRP) composites, while aramid ber-reinforced polymer

reducing disturbance to services and trafc. Another signicant advan-tage of FRP, which applies only to FRP laminates formed via the wetlay-up process, is the ability of such FRP laminates to follow curved andirregular surfaces of a structure. This is difcult to achieve using steelplates. A third advantage of FRP is that its material properties in different

500 Mild Steel Sika CarboDur S 165 2800 1.70

132 J.G. Teng et al. / Journal of Constructional Steel Research 78 (2012) 13114300 0.5 1 1.5 2 2.5 3

Strain (%)

(high strength CFRP)Sika CarboDur M(intermediate modulus CFRP)

210 2400 1.20

Sika CarboDur H(high modulus CFRP)

300 1300 0.45

a(AFRP) composites and basalt ber-reinforced polymer (BFRP) compos-ites are less frequently used. A useful general background to the compo-sition of these materials and their mechanical properties can be found inRefs. [14]. Fig. 1 shows typical stressstrain responses of FRP compos-ites in contrast with that of mild steel, where it is clearly seen that FRPcomposites exhibit a linear elastic stressstrain behavior before brittlefailure by rupture. This linearelasticbrittle stressstrain behavior hasimportant implications for the structural use of FRP composites in civilengineering applications.

FRP composites possess several advantages over steel, the most sa-lient of which are their high strength-to-weight ratio and excellent cor-rosion resistance. The structural use of FRP in civil infrastructure isgenerally based on the exploitation of these advantages. In particular,FRP, being a material of high tensile strength, can generally be used toits greatest advantages, when combined with concrete which is strongin compression but poor in tension. Therefore, the use of FRP in concretestructures has been amajor focus of existing research [2,46]. Suchappli-cations include the external bonding of FRP to concrete structures forstrengthening purposes, concrete structures reinforced or prestressedwith FRP, concrete-lled FRP tubes as columns and piles, as well asFRP-concrete hybrid beams/bridge decks. More recently, the use of FRPcomposites in combination with steel, particularly in the strengtheningof steel structures, has received much attention. This paper rst exam-ines applicationswhere the use of FRP in the strengthening of steel struc-tures presents signicant advantages and then provides a critical reviewand interpretation of existing research on FRP-strengthened steelstructures.

2. Appropriate use of FRP in the strengthening of steel structures

Since steel is also amaterial of high elastic modulus and strength, theuse of FRP in strengthening steel structures calls for innovative exploita-tions of the advantages of FRP. The main advantage of FRP over steel inthe strengthening of steel structures is its high strength-to-weightratio, leading to ease and speed of transportation and installation, thus

1000

1500

2000

2500

3000

Stre

ss (M

Pa)

High modulus CFRP

Intermediate modulus CFRP

High strength CFRP

GFRPFig. 1. Typical FRP and mild steel stressstrain curves.directions can be tailored for a particular application. As a result of thesecond and third advantages, FRP jackets with bers oriented only orpredominantly in the circumferential direction can be used to connesteel tubes/shells or concrete-lled steel tubes to delay or eliminatelocal buckling problems in steel tubes/shells, thereby enhancing thestrength and/or seismic resistance of such structures (e.g. [712]). Themethod of FRP connement is attractive not only in the strengtheningof steel tubular structures, but also in the construction of new tubularcolumns.

The combination of adhesive bonding with shape exibility makesbonded wet lay-up FRP laminates an attractive strengthening methodin a number of applications. Needless to say, steel plates can also beadhesively-bonded but bonding is less attractive for steel plates dueto their heavy weight and inexibility in shape. Furthermore, for thesame tensile capacity, a steel plate has a much larger bending stiffnessthan an FRP laminate so a steel plate leads to higher peeling stressesat the interface between the steel plate and the steel substrate. It isalso easier to anchor FRP laminates to a steel member by wrappingFRP jackets around the steel member.

Steel plates can also be attached by welding to strengthen existingsteel structures, but the bonding of FRP laminates is superior to thewelding of steel plates in the following situations:

(1) Bonding of FRP laminates for enhanced fatigue resistance has theadvantage that the strengthening process does not introducenew residual stresses;

(2) In certain applications (e.g. oil storage tanks and chemical plants)where re risks must be minimized, welding needs to be avoidedwhen strengthening a structure; bonding of FRP laminates is thena very attractive alternative;

(3) High-strength steels suffer signicant local strength reductionsin heat-affected zones of welds, so bonded FRP laminates offeran ideal strength compensation method [13].

The use of both CFRP and GFRP to strengthen steel structures hasbeen explored. For the strength enhancement of steel structures, CFRPis preferred over GFRP due to the much higher elastic modulus of theformer. In particular, when the enhancement of buckling resistance isthe aim, the use of high or ultra-high modulus CFRP is very attractive.Table 1 shows the properties of pultruded CFRP plates supplied bySIKA; these three types of CFRP plates are referred to herein as highstrength, intermediate modulus and high modulus plates respectivelyand their stressstrain curves are illustrated in Fig. 1. By contrast,for the connement of steel tubes, particularly when ductility

Table 1Properties of SIKA CFRP platesa.

Product Elastic modulus(GPa)

Tensile strength(MPa)

Ultimatestrain (%)Extracted from the manufacturer's product data sheet.

enhancement is the main aim, GFRP is more attractive as it is cheaperand offers a greater strain capacity (>2%). An issue to note is that ofgalvanic corrosion when steel is in direct contact with CFRP [14,15], soa layer of GFRP has been advised to be sandwiched between them bysome researchers (e.g. [15]). A detailed discussion of the issue ofgalvanic corrosion is given in Ref. [16].

Since FRP composites, particularly CFRP composites are an expen-sive material, in all applications, the amount of FRP material requiredshould be minimized. For this reason, where the amount of FRP mate-rial required is small by nature of the problem (e.g. local strengthen-ing under a concentrated force), FRP strengthening is more likely to

on in some recent research [17]. The authors thus strongly believethat in FRP-strengthened steel structures, interfacial failure shouldoccur within the adhesive layer in the form of cohesion failure(Fig. 3), and a proper surface treatment procedure together with anappropriate adhesive should be used to ensure that such cohesionfailure is critical.

3.2. Adhesion failure

In an FRP-to-steel bonded joint, adhesion failure may occur at thesteel/adhesive interface or at the FRP/adhesive interface. However,adhesion failure at the FRP/adhesive interface seldom occurs whenthe FRP is formed and applied to the structure via a wet lay-up pro-cess on site; when a pultruded FRP plate/strip is used, such failurecan generally be avoided through the use of a peel-ply which is re-

133J.G. Teng et al. / Journal of Constructional Steel Research 78 (2012) 131143be attractive.

3. Bond behavior between FRP and steel

3.1. General

Similar to the structural use of FRP in concrete structures, thestructural use of FRP with steel can be classied into two categories:(a) bond-critical applications where the interfacial shear stress transferfunction of the adhesive layer that bonds the steel and the FRP togetheris crucial to the performance of the structure; and (b) contact-criticalapplications where the FRP and the steel need to remain in contact foreffective interfacial normal stress transfer which is crucial to ensurethe effectiveness of the FRP reinforcement. The use of FRP in thestrengthening of steel structures provides good examples for bothcategories: externally bonded FRP reinforcement for the exuralstrengthening of steel beams falls into the rst category, while conne-ment of concrete-lled steel tubular members with FRP jackets belongsto the second category.

In all bond-critical applications, the interfacial behavior betweenFRP and steel is of critical importance in determining when failure oc-curs and how effectively the FRP is utilized. An important difference inbond behavior between FRP-strengthened concrete structures andFRP-strengthened steel structures is the exact location of interfacialfailure: for the former interfacial failure generally occurs in the sub-strate concrete and the design theory has been developed with this na-ture of interfacial failure implicitly or explicitly assumed; for the latterinterfacial failure cannot possibly occur in the substrate steel due tothemuch higher tensile strength of steel than that of adhesives. As a re-sult, for the latter, interfacial failure can only occur within the adhesivelayer (i.e. cohesion failure) or at the material interfaces (adhesion fail-ure) between the steel and the adhesive (referred to as the steel/adhe-sive interface hereafter) or between the adhesive and the FRP (referredto as the FRP/adhesive interface hereafter). A summary of possiblefailure modes is shown in Fig. 2.

If adhesion failure controls the strength of FRP-strengthened steelstructures, then the interfacial bond strength depends on how thesteel surface and the FRP surface are treated as well as the bond capa-bility of the adhesive. As adhesion failure depends on the method anddegree of surface treatment, especially to the steel substrate, which isdifcult to control on site, the development of a design theory be-comes much more involved. This important issue has not beengiven adequate attention in previous studies, but has been focused

Interlaminar failure of FRP

Steel

Adhesive

CFRP

Adhesion failure at FRP/adhesive interface

Cohesion failure in adhesive

Adhesion failure at steel/adhesive interface

FRP RuptureFig. 2. Possible failure modes of FRP-to-concrete bonded joints.moved prior to bonding to ensure a clean and rough FRP surface forbonding [15] or by abrading and cleaning the FRP surface beforebonding. By contrast, failure at the steel/adhesive interface is muchmore likely to happen. For various reasons, the treatment and charac-terization of steel surfaces for adhesive bonding has received muchresearch attention [1822].

The adhesion strength of a steel/adhesive interface results from bothchemical bonding and mechanical bonding between the two adherends[18,21,23]. It is evident that a strong steel/adhesive interface requiresthe adhesive to be in intimate contact with the steel surface. This gener-ally means that the adhesive should have a sufciently low viscosity sothat it can ow easily over the surface and ll the pores [24], and thatthe steel surface should be clean and should have a sufciently large sur-face energy so that it can be easily wetted [20,21]. When the twoadherends are in intimate contact, the strength of chemical bonding de-pends mainly on the chemical composition of the steel surface and thatof the adhesive and whether they are chemically compatible [21]. Bycontrast, apart from the properties of the adhesive, the strength of me-chanical bonding depends mainly on the roughness and topography ofthe steel surface; roughening the surface can signicantly enhance thestrength of mechanical bonding [23,25], but it may also reduce thelevel of contact between the two adherends [26,27]. Therefore, thethree main properties of a steel surface, namely, surface energy, surfacechemical composition and surface roughness and topography, are oftenused to characterize the capacity of a surface for bonding [20,2830].

Existing approaches of steel surface treatment generally aim toenhance the two bonding mechanisms (i.e. chemical bonding and me-chanical bonding) by: (1) cleaning the surface; (2) changing the prop-erties of the surface. The most popular approaches include solventcleaning and mechanical abrasion through grit blasting or using othertools (e.g. wire brushes, abrasive pads and wheels, and needle guns)[15,21]. Solvent cleaning removes the contaminants from the surface(e.g. grease, oil and water) but does not change the surface properties,so it alone only has a limited effect on the adhesion strength [20]. It ishowever a necessary step of any surface treatment process and shouldFig. 3. Surface of the FRP plate after cohesion failure.

initiates at an FRP plate end due to a combination of high interfacialshear and peeling (normal) stresses. Intermediate debonding hasbeen observed in laboratory tests on FRP-strengthened steel beamswith or without an initial defect (e.g. [3739]) and steel sectionsstrengthened with FRP against local buckling (e.g. [40]), while plateend debonding has been observed in laboratory tests onexurally-strengthened steel beams (e.g. [41]) and on steel sectionsstrengthened against end bearing loads (e.g. [42,43]) or other loads in-ducing local buckling (e.g. [44]).

It has been widely recognized [4550] that in order to understandand model debonding failures, the bond behavior between the sub-strate material and the bonded FRP reinforcement needs to be studied,commonly through pull tests on simple bonded joints (Fig. 4(a))[48,5153]. In a pull test, the adhesive layer is primarily subjected to in-terfacial shear stresses and debonding is caused by Mode II fracture infracturemechanics terms. The interfacial behavior of such simple bond-ed joints is similar to that of an FRP-to-steel interface in a beam whereintermediate debonding is critical, as interfacial shear stresses dominatethe debonding process in both cases. This interfacial shear behavior isalso an important basis for understanding the behavior of FRP-to-steelinterfaces subjected to combined shear stresses and peeling stresses.

Different from FRP-to-concrete bonded joints where the concrete isusually the weak link, the adhesive is the weak link in FRP-to-steel

134 J.G. Teng et al. / Journal of Constructional Steel Research 78 (2012) 131143normally be conducted at the beginning of the process [15,22]. It is im-portant to use a volatile solvent (e.g. acetone) so that the contaminantson the surface (and hence their negative effects on the adhesionstrength) are minimized [4,18]. Mechanical abrasion roughens the sur-face and removes the weak surface layer (e.g. oxide layer) which ischemically inactive [20,21], so that the surface in contactwith the adhe-sive is sufciently rough, clean and chemically active. Among variousmechanical abrasion approaches, grit blasting appears to be themost ef-fective [15,20,31,32] and is recommended by some existing guidelineson the FRP strengthening of metallic structures [22,33]. Tests recentlyconducted by Teng et al. [17] showed that with the four types of differ-ent adhesives used in their study, adhesion failure was avoided whenthe steel surface was grit-blasted prior to bonding.

The grit used in grit blasting may be made of different materialsand have different particle sizes. Existing studies [17,20,34] haveshown that grit blasting can modify the chemical composition of thesurface by introducing grit residues to the surface, so it is importantto choose a grit material which is chemically compatible with the ad-hesive. The particle size of grit may have a pronounced effect on sur-face energy and surface roughness, but the limited existing studies[17,20] have revealed that within the range of grit particle sizes ex-amined in these studies (i.e. from 0.125 mm to 0.5 mm), the effectof particle size on adhesion strength is limited.

During the grit blasting process, ne abrasive dust is produced andbecomes additional surface contaminants [15]. Therefore, it is importantto clean the surface again after grit blasting. Hollaway and Cadei [15]suggested to remove the ne dust using dry-wiping or using a vacuumhead instead of solventwiping as they believed that solventwiping is ca-pable of only partial removal of the dust and is likely to redistribute theremaining dust on the surface. El Damatty and Abushagur [35] howevershowed that with the use of an excessive amount of solvent, the dust canbe completely removed and a clean surface can be produced.

After surface treatment, an adhesive/primer should be applied assoon as possible to avoid any contamination of the surface or formationof weak oxide layers on the surface [36]. Cadei et al. [33] recommendedthat the period between grit blasting and adhesive/primer applicationshould not exceed 2 h,while Schnerch et al. [22] suggested amore prac-tical maximum period of 24 h for the application of adhesive.

Apart from the adoption of an appropriate surface treatment proce-dure, it is also important to characterize the surface to determinewhether a sufcient adhesion strength can be developed. The followingmethods are available for surface characterization: (a) a VCA (videocontact angle) device can be employed to obtain contact angle mea-surements from which the surface energy can be evaluated; (b) anSEM/EDX (scanning electron microscopy/energy dispersive x-ray) sys-tem can be used to measure the surface chemical composition; and(c) a prolometer can be used tomeasure the surface roughness and to-pography [17]. By using these devices, Teng et al. [17] showed that thecharacteristics of surfaces are consistent after being grit-blasted usingthe same grit, which suggest the possibility of developing a standardpreparation process to ensure a good surface with a sufcient adhesionstrength.

3.3. Bond behavior

Similar to reinforced concrete (RC) structures strengthenedwith ex-ternally bonded FRP reinforcement, interfacial debonding failures alsocontrol the load-carrying capacity of steel structures strengthenedwith externally bonded FRP reinforcement in many cases. A simply-supported steel beam strengthened in exure using a bonded softFRP plate is a typical bond-critical case where the following two distinctdebonding failure modes can occur: (1) intermediate debonding; and(2) plate end debonding. In the former mode, debonding initiatesaway from the FRP plate ends and at a location where high interfacialshear stresses arise from either the presence of a defect (e.g. crack) or

local yielding of the steel substrate. In the latter mode, debondingbonded joints, provided that adhesion failure at the steel/adhesive in-terface and the FRP/adhesive interface is avoided by careful selectionof the adhesive and appropriate surface preparation of the steel andthe FRP. As a result, the behavior of FRP-to-steel bonded joints is similarto that of steel-to-steel bonded joints, so available tests on the latter arealso included in the discussion below to supplement the limited avail-able studies on FRP-to-steel bonded joints [35,51,5365]. In addition,existing studies on FRP-to-concrete bonded joints are referred to wher-ever appropriate, as the generic concepts (e.g. the interfacial fractureenergy and the effective bond length) well established for these jointsare also applicable to FRP-to-steel bonded joints.

Different test methods for bonded joints have been used by differentresearchers [52], including single-lap pull tests (Fig. 4(a)) [51,60],double-lap pull tests (Fig. 4(b)) [54,58,65], double-lap shear tests undercompression [35], and beam tests [56]. Despite the variations in the testmethod, most of the existing studies were focused on the two importantcharacteristics of the interface: the ultimate load of the joint (i.e. the bondstrength) and the relationship between the interfacial shear stress andthe interfacial relative displacement between the two adherends at a

a

b

FRP plate

Adhesive

Steel substrate

AdhesiveSteel plates FRP plateGap Fig. 4. Pull tests of bonded joints. (a) Single-lap pull test. (b) Double-lap pull test.

energy under shear (Mode II) loading, and (L) is a function of thebond length.

Fernando [60] and Xia and Teng [51] recently conducted two se-ries of single-lap pull tests aiming to understand the full-range behav-ior of FRP-to-steel bonded joints. Their test results claried the effectsof adhesive properties, adhesive layer thickness, and the plate axial ri-gidity of FRP on the bond strength, and veried the applicability ofEqs. (1) and (2) to FRP-to-steel bonded joints. Fernando [60] also pro-posed an equation to predict the Mode II interfacial fracture energy Gfbased on the thickness and tensile strain energy (i.e. the area underthe uniaxial tensile stressstrain curve) of the adhesive.

3.3.2. Bond-slip relationshipAn accurate bond-slip model for FRP-to-steel interfaces is of funda-

mental importance to the understanding and modeling of the behaviorof FRP-strengthened steel structures. A bond-slip model depicts the rela-tionship between the local interfacial shear stress and the relative slipbetween the two adherends and can be experimentally obtained throughbonded joint tests. To study the bond-slip behavior of FRP-to-concretebonded joints, the single-lap pull test with the steel block supported atthe loaded end (Fig. 4(a)) is probably the most suitable [48] and wasalso used in the recent studies on the full-range behavior of FRP-to-steelbonded joints [51,60,84,86].

For FRP-to-concrete bonded joints, Lu et al. [87] conducted a thor-ough review of bond-slip models and proposed three two-branch (anascending branch and a descending branch) bond-slip models of differ-ent levels of sophistication. The simplest of the bond-slip models pro-posed by Lu et al. [87] is a bi-linear model with sufcient accuracyfor practical use (Fig. 5(a)). The key parameters of the bilinear bond-slip model are the maximum local bond shear stress max and the

a

Elastic Softening region Debonding

135J.G. Teng et al. / Journal of Constructional Steel Research 78 (2012) 131143specic location on the interface (i.e. the local bond-slip relationship). Inthe following discussion, a single-lap pull test is assumed for simplicity ofdescription and a double-lap pull test can be seen as two single-lap pulltests being conducted simultaneously.

3.3.1. Bond strengthThe bond strength is the ultimate tensile force that can be resisted

by the FRP plate in a bonded joint test before the FRP plate debondsfrom the substrate [4]. Existing studies [51,53,57,59,60] have shownthat the bond strength of an FRP-to-steel bonded joint initially in-creases with the bond length, but when the bond length reaches athreshold value, any further increase in the bond length does notlead to a further increase in the bond strength. This observation issimilar to that found in tests on FRP-to-concrete bonded joints[48,50,66,67], and the threshold bond length value is commonly re-ferred to as the effective bond length (Le) [66].

Two main approaches have been developed to predict the bondstrength of FRP-to-steel bonded joints: (1) strength-based approach[22,57,68] which assumes that the bond strength is reached whenthe maximum stress/strain in the adhesive reaches its correspondingultimate value; and (2) fracture mechanics-based approach [60,69]which is similar to that employed to predict the bond strength ofFRP-to-concrete bonded joints [66,70] where the bond strength is re-lated to the interfacial fracture energy.

Apart from studies on FRP-to-steel joints, the strength-based ap-proach has also been adopted in some studies on steel-to-steel bondedjoints [7173]. The failure criteria for the adhesive used in these studiesinclude themaximumshear stress criterion [71], themaximumprincipalstress criterion [72] and the maximum shear strain criterion [73]. Thestrength-based approach generally implies that the ultimate load of thebonded joint is reachedwhen therst crack occurs in the adhesive. How-ever, Fernando [60] found from single-lap pull tests that the tensile forceresisted by the FRP plate can still increase signicantly after the initiationof the rst crack in the adhesive, provided that the bond length is suf-ciently long. In addition, the existence of an effective bond length is notcompatible and cannot be explained with the strength-based approach.Therefore, it can be concluded that the strength-based approach doesnot reect the debonding failure mechanism of an FRP-to-steel bondedjoint; however, it may provide reasonable predictions when thebond length is small so that debonding failure of the bonded jointfollows immediately the occurrence of the rst crack in the adhesive.In applying the strength-based approach, an accurate analysis of in-terfacial stresses and/or strains in the adhesive is needed. Both ana-lytical studies [71,7479] and nite element (FE) studies [45,80,81]have been conducted to predict interfacial stresses in bonded joints,but many of them suffer from various limitations [60], including theomission of interfacial peeling stresses (e.g. [71]), the assumption of aconstant stress state over the thickness of the adhesive (e.g. [76,78]),and the inaccurate simulation of the edge shape of the FRP plate end(e.g. [82]). A thorough review of interfacial stress analysis can befound in Ref. [81].

The fracture mechanics-based approach has been successfullyemployed to predict the bond strength of FRP-to-concrete bonded jointsand steel-to-concrete bonded joints [66,70,83]. This approach providesthe theoretical basis for the existence of an effective bond length whichhas also been observed in FRP-to-steel bonded joint tests [51,53,56,84].In this approach, the bond strength depends on the interfacial fractureenergy as given below [51,67,85] instead of the strength of the adhesive:

Pu bp2EptpGf

qwhen L Le 1

Pu L bp2EptpGf

qwhen L Le 2

where Pu is the bond strength, bp is the plate width, Ep is the elastic mod-

ulus of the plate, tp is the plate thickness, Gf is the interfacial fractureSlip

Inte

rfaci

al sh

ear s

tress

Area under the curve= Gf

b

ElasticSoftening region Debonding

Constant stress region

Area under the curve= GfInt

erfa

cial

shea

r stre

ss

Slip

Fig. 5. Bond-slip curves for linear and nonlinear adhesives. (a) Linear adhesives.

(b) Nonlinear adhesives.

corresponding slip 1, the ultimate slip f when the local bond shearstress rst reaches zero, and the interfacial fracture energy Gf which isequal to the area enclosed by the bond-slip curve and the horizontalaxis. For FRP-to-concrete bonded joints, these parameters are generallyrelated to the tensile strength of concrete as the concrete is usually theweak link of the joint.

A two-branch bond-slip model without a plateau at the peak stresshas been shown to perform well for almost all FRP-to-concrete bondjoints because of the brittle nature of concrete. However, such a modelmay not work well for FRP-to-steel bonded joints where the weak linkis the adhesive whose behavior may be brittle or ductile. As a result,the bond-slip response of FRP-to-steel interfaces may also be brittle orductile as it depends on the material properties of the adhesive.Fernando [60] recently conducted a series of single-lap pull tests onFRP-to-steel bonded joints formed using four different adhesives. Resultsfrom Fernandos study [60] showed that while a two-branch bond-slipmodel is suitable for bonded joints with a brittle linear adhesive, it isnot suitable for joints with a more ductile nonlinear adhesive having ahigh strain capacity (up to 2.9%). The shape of the bond-slip curve forhis joints with a nonlinear adhesive was shown to be trapezoidal(Fig. 5(b)). Based on these test results, Fernando [60] proposed threebond-slip models, two for linear adhesives and one for nonlinear adhe-sives respectively, where the parameters of both types of models are re-lated to the material properties of the adhesive.

vent plate end debonding failure [98].As plate end debonding in FRP-plated beams depends strongly on the

localized interfacial stresses, many studies have been conducted on theprediction of these interfacial stresses, including both analytical solutions[76,78,79,90,104] and numerical investigations [45,81,90,105]. Thesestudies have been based on different simplifying assumptions and thuspossess different levels of sophistication [81]. Despite such differences,these existing studies generally assumed that the adhesive layer is line-arly elastic. A comparison of different modeling approacheswas recentlypresented by Zhang and Teng [81], which illustrates clearly how each as-sumption affects the predicted interfacial stresses. Stress singularityarises at the bi-material interfaces when a sharp square edge is assumed[45,106] but this issue cannot be properly dealt with by the existing an-alytical solutions. In real applications, the edge shape can be quite differ-

136 J.G. Teng et al. / Journal of Constructional Steel Research 78 (2012) 131143a

Steel I beam

FRP plateAdhesive layer

FRP U-jackets

b

FRP plateAdhesive layer

Steel I beam

Fig. 6. Strengthening of steel beams with a bonded FRP plate. (a) Side view.4. Flexural strengthening of steel beams

Similar to an RC beam, a steel beam (or a composite steel-concretebeam) can be strengthened by bonding an FRP (generally CFRP) plateto its tension face (i.e. the soft if a beam in positive bending is assumed,see Fig. 6) [37,54,57,8897]. The bonded FRP plate can enhance not onlythe ultimate load but also the stiffness of the beam (especially when ahigh modulus CFRP is used) [90,93,98,99]; the latter means that thestrains in the beam are reduced under the same load and the rst yield-ing of the beam is delayed. A number of failure modes (Fig. 7) are pos-sible for such FRP-plated steel beams, including: (a) in-plane bendingfailure [96]; (b) lateral buckling [37]; (c) plate-end debonding [41,97];and (d) intermediate debonding due to local cracking or yielding(b) Cross-sectional view.away from the plate ends [37]. Additional failure modes include:(e) local buckling of the compression ange; and (f) local buckling ofthe web. It should be noted that even in a beam for which these localbuckling modes are not critical before FRP strengthening, they can be-come critical after strengthening, particularly when the strengtheninginvolves only the bonding of FRP to the tension ange only. This isbecause the compression ange and the web now need to sustain ahigher load level before the beam fails in one of the other modes, buttheir local buckling resistance does not benet from the bonded FRPreinforcement.

The in-plane bending capacity of an FRP-plated steel beam can beeasily determined, provided that debonding does not become criticaland hence the plane section assumption can still be used [33,100,101].Many existing analytical studies [33,90,95,96,100,101] on FRP-platedsteel beams adopted this simple assumption,whichmeans that the pre-diction of debonding failures was beyond their scope. Nevertheless, re-search on debonding failures has attracted considerable attentionworldwide (e.g. [16,33,91,100]) as discussed below.

4.1. Plate end debonding

As described earlier, plate end debonding in an FRP-plated steelbeam is due to high localized interfacial shear stresses and peelingstresses in the vicinity of the plate end. The magnitudes of these local-ized interfacial stresses depend on a number of factors [78,81], includ-ing the bending moment and the shear force in the beam at the plateend location. In a simply-supported beam in three- or four-point bend-ing, plate end debonding is more likely to occur when the plate end isfarther away from the adjacent support (i.e. when the plate end mo-ment is larger) but can be delayed or even avoided when the plateend is very close to the adjacent support [41]. Besides the plate end lo-cation, the localized interfacial stresses can also be reduced using othermeasures. Examples include the use of a spew llet of excess adhesiveat the plate end [73], the use of a softer adhesive near the plate end[102], tapering the thickness of the plate near the plate end [22,103],and a combination of these measures [22]. Obviously, clamps or othertypes of mechanical anchors should be used wherever possible to pre-

Intermediatedebonding

Beam

AdhesivePlate enddebonding

Flange buckling

Web buckling

FRP rupture

FRP Plate

Fig. 7. Some of the failure modes of steel beams bonded with an FRP plate.ent from a sharp square edge because of the existence of a llet of excess

137J.G. Teng et al. / Journal of Constructional Steel Research 78 (2012) 131143adhesivewhich is introduced during the installation process; this changein the edge shape may signicantly reduce the interfacial stresses, but ithas seldom been appropriately considered.

While existing solutions for interfacial stresses in FRP-plated beamsbased on the assumption of linear elastic material behavior are helpfulfor understanding the occurrence of plate end debonding, they cannotbe used directly to predict debonding failure as debonding is controlledby the interfacial fracture energy rather than by stress values. In someexisting studies (e.g. [22,91]), it was simply assumed that plate enddebonding occurs when the maximum interfacial stresses found fromanelastic analysis reach their correspondingmaterial strengths; this ap-proachmay signicantly underestimate the plate end debonding failureload for reasons similar to those already discussed for bonded joints.

To accurately predict plate end debonding, the nonlinear and dam-age behavior of the interface in both the normal (i.e. peeling) direction(i.e. under Mode I loading) and the shear direction (i.e. under Mode IIloading) and their interaction should be appropriately simulated.Fernando [60] made the rst attempt to model plate end debonding ofFRP-plated steel beams using a so-called mixed-mode cohesive law tosimulate this complex behavior of the FRP-to-steel interface. Fernandosmixed-mode cohesive law [60] was based on a bond-slip model forMode II behavior developed from pull tests and certain assumptionsfor Mode I behavior and for interaction between the two modes [60].Itwas shown that byusing thismixed-mode cohesive law, both the pro-cess of and the ultimate load at plate end debonding can be closelypredicted [60].

More recently, Chiew et al. [107] proposed an approach similar to themixed-mode fracture criterion, where the dilatational and distortionalstrain energy densities are used as variables instead of the Mode I andMode II interfacial fracture energy. Chiew et al. [107] also veried theirapproach using their own test results [108]. However, in Chiew et al.sstudy [107], the critical values for the dilatational and distortional strainenergy densities and the failure envelope accounting for the interactionbetween the two energy density components were both based on theirown bonded joint tests where only one single adhesive was used. Thewide applicability of their approach thus remains uncertain.

4.2. Intermediate debonding

Intermediate debonding generally initiates at a defect (e.g. crack)[38,39] or a location of concentrated plasticity of the steel substrate[37] where the FRP plate is highly stressed; it then propagates towardsa plate end. Although both plate end debonding and intermediatedebonding are brittle failuremodes, the latter, involving amore gradualprocess of debonding, is generally less brittle than the former [60].

Comparedwith plate end debonding, much less research is availableon intermediate debonding in FRP-plated steel beams [60]. Intermedi-ate debonding in FRP-plated steel beams is similar in nature tointermediate-crack debonding (IC debonding) in FRP-plated RC beams[47]: both initiate from a location where the FRP is highly stressedand both are dominated by interfacial shear stresses. Therefore, it canbe expected that the intermediate debonding strength depends strong-ly on the interfacial shear fracture energy obtained from pull tests onbonded joint tests [60]. For the accurate prediction of intermediatedebonding failure in an FRP-plated steel beam, an accurate bond-slipmodel that captures the nonlinear behavior of the FRP-to-steel interfaceis needed. Fernando [60] showed that with the use of a cohesive lawbased on a bond-slip model for Mode II behavior, both the processof and the ultimate load at intermediate debonding can be closelypredicted.

4.3. Other issues

Although steel beams are often prevented from lateral bucklingfailure by slabs and other adjacent structural members, this mode of

failure is still possible in some situations. The elastic lateral bucklingproblem has been studied by Zhang and Teng [109], but much morework is needed before a design method can be established.

In the strengthening of steel or steel-concrete composite bridges,the speed of strengthening operations is of great importance whenclosure of trafc needs to be avoided to reduce economic losses.Hollaway et al. [110] and Zhang et al. [111] investigated the rapidstrengthening of steel bridges using prepregs and lm adhesive.Using this new method, a bridge may be strengthened in as short as4 h. They also examined the effect of trafc-induced vibration duringthe curing of the FRP system on the performance of the strengthenedstructure. The effectiveness and reliability of this rapid strengtheningmethod for steel structures were demonstrated by their study [110].

5. Fatigue strengthening

One of the most important aspects of FRP strengthening of steelstructures is its capability to improve their fatigue life [112118]. Fatiguestrengthening studies have been carried out on beams [92,94,119121],steel plates [116,117,122126], steel rods [127] and steel connections[128130].

Similar to the behavior of FRP-to-steel joints under static loading, Liuet al. [116,117] found that the fatigue life of FRP-strengthened steelplates initially increased with the bond length until the effective bondlength Le was reached, after which any further increase in the bondlength did not further increase the fatigue life. In the strengthening ofsteel members (e.g. plates, beams and rods), a bond length longerthan Le is easy to achieve, but this may be difcult in the strengtheningof steel connections where the bond length of FRP is limited. In suchcases, the adhesive should be carefully selected to minimize the effec-tive bond length.

Stress intensity factors (SIFs) are commonly used in fracture me-chanics to describe the stress state at a crack tip due to appliedloads and/or residual stresses [122,131]. The fatigue strengtheningof steel structures generally aims to reduce the SIF at a (potential)crack tip and thus increase their post-crack fatigue life. As may beexpected, the use of a stiffer FRP plate (i.e. a thicker plate or a platewith a higher elastic modulus) or a stiffer adhesive (i.e. with a higherelastic modulus) can reduce the SIF [116,117,128]. One exception tothis statement is that when a relative thin steel plate is strengthenedon one side only, an excessively stiff plate can induce out-of-planebending of the steel plate which can lead to premature debondingof FRP [132]. Debonding near the crack tip can lead to a signicant in-crease in the SIF, which is detrimental to the fatigue life of thestrengthened structure [114]. In addition to experimental work, anumber of analytical studies [114,116,132] have been conducted onthe prediction of SIFs at crack tips in FRP-strengthened steel struc-tures. Such analysis is necessary and useful in the design of FRP sys-tems for the fatigue strengthening of steel structures.

Debonding along the CFRP-to-steel interface is also a key issue ofconcern in the fatigue strengthening of steel beams with CFRP,where both plate end debonding and intermediate debonding arepossible. While plate end debonding may be prevented using variousmeasures (see Section 4.1) and is often not a concern, intermediatedebonding of the FRP can have a signicant effect on the crack growthrate in the steel [94,114] in fatigue-strengthened steel beams. How-ever, in most of the existing literature, debonding between the FRPand the substrate is either not considered at all or is modeled basedonly on a prescribed debonding shape and size as a function of thesubstrate crack width when the SIF is evaluated [133,134]. More re-search is therefore necessary to gain a better understanding of the cy-clic behavior of CFRP-to-steel bonded interfaces and the interactionbetween intermediate debonding and fatigue crack growth in steelbeams so that the detrimental effect of debonding on the fatigue lifeof the CFRP-strengthened steel beam can be predicted.

Pre-stressing the bonded FRP reinforcement can signicantly en-

hance the effectiveness of fatigue strengthening. By pre-tensioning

the FRP plate, compressive stresses are induced in the steel substrate toachieve crack closure, resulting in improved fatigue performance. Theeffect of the pre-tensioning level on the fatigue crack growth rate has

structure. Such local buckling failure may be prevented by bonding

responses of a rectangular hollow section (RHS) tube subjected to anend-bearing load when ve different adhesives were used to bond theCFRP. Depending on the adhesive used, the failure mode varied fromthe debonding initiating at a plate end to FRP rupture failure; theamount of strength enhancement achieved also varied signicantly. Itwas shown in this study that debonding was less likely to occur whenan adhesive with a larger ultimate tensile strain was used, which ledto a greater load-carrying capacity of the strengthened tube [43,60].

6.2. Buckling induced by other loads

FRP, especially CFRP, has also been used in the strengthening ofother steel structures against local buckling, including steel square col-umns [135], lipped channel steel columns [136], and steelWT compres-sion members [44,137,138] subjected to axial compression. The FRPstrengthening has been shown to be very effective [44,60] in delayinglocal buckling and thus enhancing the strength of the steel structure,especially when a slender section is used. While crushing of the FRP

Fig. 8. Debonding failure of CFRP-strengthened rectangular steel tube subjected to anend bearing load.

b

138 J.G. Teng et al. / Journal of Constructional Steel Research 78 (2012) 131143FRP patches. Local high tensile stresses may also be addressed in thesame way.

A practically important problem is the web crippling failure ofthin-walled sections under a bearing force [42]. Zhao et al. [42] foundfrom their experimental study that bonded CFRP can be an effective so-lution to this problem. Fernando et al. [43] further investigated the ef-fect of adhesive properties on the effectiveness of this strengtheningtechnique. Fig. 8 which is extracted from Ref. [43] shows the different

abeen studied both experimentally and numerically [114,122,131]. Byevaluating the SIF at the crack tip of the strengthened system, thepre-tensioning force needed to stop the growth of a fatigue crackcan be predicted [131]. The level of pre-tensioning that can be im-posed on an FRP strengthening system depends on the static andfatigue strength of the bonded joint, where a good understandingof the behavior of bonded interfaces under fatigue cyclic loading isagain required.

6. Strengthening of steel structures against local buckling

6.1. Buckling induced by high local stresses

In practice, high stresses in a local zone often arise, due to concen-trated loads and the need to introduce discrete supports, openingsand other local features. Under local high compressive stresses, localbuckling failure is likely to control the thickness of a thin-walled steelFig. 9. Elephants foot buckling in a steel tube or shell. (a) Failure near thCourtesy of Dr. H.B. Ge, Nagoya University and Prof. J.M. Rotter, Edinburgplate was observed in some experiments [135], debonding has beenfound to be the most likely failure mode in the strengthened struc-ture [44,135,136].More research is therefore needed ondebonding pro-cesses in buckling failures of FRP-strengthened steel structures wherethe FRP is commonly loaded in compression.

7. FRP connement of hollow steel tubes

Hollow steel tubes are used in many structures. Local buckling canoccur in these tubular members when they are subjected to axial com-pression alone or in combinationwith monotonic/cyclic lateral loading.For example, hollow steel tubes are often used as bridge piers and suchbridge piers suffered extensive damage and even collapse during the1995 Hyogoken-Nanbu earthquake [139]. A typical local bucklingmode of circular hollow steel tubes involves the appearance of an out-ward bulge near the base and is often referred to as elephants footbuckling (Fig. 9). In typical circular tubular structures, elephants footbuckling appears after yielding and the appearance of this inelasticlocal bucklingmode normally signies the exhaustion of the load carry-ing capacity and the end of the ductile response. The latter is of partic-ular importance in seismic design, as the ductility and energyabsorption capacity of the column dictate its seismic resistance. In rect-angular (including square) steel tubes, a similar failuremode can occur.Here, the buckling deformation is normally outwards on theanges andinwards on the webs.

The enhancement of ductility and hence seismic resistance of hollowtubular columns through connement by an FRP jacket has been

e base of a steel tube. (b) Failure at the base of a liquid storage tank.

h University.

explored by the authors' group [8,140,141] as an extension of Xiaosidea of conning concrete-lled steel tubes with FRP [9]. The techniquewas shown to be highly effective. The failure modes of hollow steeltubes with and without FRP connement are shown in Fig. 10(a) and(b), while the axial stress-nominal axial strain (axial shortening/tubeheight) curves are shown in Fig. 10(c). It is clear that through FRPconnement, the elephants foot mode of buckling failure is preventedand the ductility of the tube is greatly enhanced. Nishino and Furukawa[142] also explored the same technique for hollow steel tubesindependently. More recent work on FRP-strengthened hollow steeltubes/cylindrical shells can be found in [143145].

These results also show that when the jacket thickness reaches athreshold value for which inward buckling deformations dominate thebehavior, further increases in the jacket thickness do not lead to signi-cant additional benets as the jacket provides little resistance to inwardbuckling deformations. It is signicant to note that FRP connement ofsteel tubes leads to large increases in ductility but limited increases inthe ultimate load, which is often desirable in seismic retrot of columnswhich are part of a larger structure, so that the retrotted tube will notattract forces which are so high that adjacent members may be put indanger.

The elephants foot buckling mode is not only the critical failuremode in commonly used circular steel tubular columns under axialcompression and/or bending, it also occurs in much thinner cylindricalshells in steel storage silos and tanks under combined axial compressionand internal pressure. This failure mode has been commonly observedin earthquakes [146] and under static loading [147]. The use of FRP

jackets to strengthen thin steel cylindrical shells against local elephantsfoot buckling failure at the base has also been explored through niteelement analyses by Teng and Hu [141]. The limited numerical resultsfor a thin cylindrical shell with a radius-to-thickness ratio of 1000and subjected to axial compression in combination with internalpressure indicate that the method leads to signicant increases ofthe ultimate load. The FRP jacketing of steel cylindrical shells can alsobe used in the construction of new tanks and silos to enhance their per-formance. A similar and related study on the strengthening of such cy-lindrical shells has been conducted by Chen et al. [148] where anoptimally-located ring stiffener is proposed as the strengtheningmeth-od. This ring stiffener may well be a CFRP cable that provides the samecircumferential stiffness and the needed strength. More recent work onthe local connement of cylindrical shells against elephants foot buck-ling and on the strengthening of cylindrical shells against bucklingusingbonded FRP reinforcement can be found in Refs. [11,12].

8. FRP connement of concrete-lled steel tubes

Concrete-lled steel tubes (CFSTs) are widely used as columns inmany structural systems. In CFSTs, inward buckling deformations ofthe steel tube are prevented by the concrete core, but degradationin steel connement, strength and ductility can result from inelasticoutward local buckling. When used as columns subjected to com-bined axial and lateral loads, the critical regions are the ends of thecolumn where the moments are the largest. Under seismic loading,plastic hinges form at the column ends and large plastic rotations

a b

s

0Axi

eel Tly F

FRP

ly FR

139J.G. Teng et al. / Journal of Constructional Steel Research 78 (2012) 131143c

Elephantfoot buckling

0

50

100

150

200

250

300

350

400

0 0.01Nominal

Axi

al S

tress

(N/m

m2)

Bare StSingle-p

Two-plyThree-pFig. 10. Suppression of local buckling in hollow circular steel tubes. (a) Bare steel tube af.02 0.03al Strain

ubeRP Jacket

JacketP Jacketter test. (b) FRP-conned tubes after test. (c) Axial stress-nominal axial strain curve.

without signicant degradation in stiffness and strength are demandedhere. Against this background, Xiao [9] proposed a novel form of con-ned concrete-lled steel tubular columns, in which the end portionsare conned with steel tube segments or FRP wraps. In these columns,due to the additional connement from an FRP or steel segment, boththe inward and the outward buckling deformations of the steel tubeare constrained, so the ductility and strength of the column can be sub-stantially enhanced in the end regions. In addition, the concrete is betterconned with the additional connement from the FRP or steel seg-ment. Although Xiao's work [9] was directed at new construction, thesame concept can be applied in the strengthening/retrot of CFSTs:FRP wrapping provides a simple and effective method to enhance theload-carrying capacity and/or ductility of CFSTs, which is similar to theFRP wrapping for strengthening RC columns [6,149]. Following Xiao'sinitial work [9], a number of studies have been conducted by Xiaoand associates [10,150,151] as well as other researchers [152160]on the effectiveness of FRP wrapping in improving the structural be-havior of both circular [10,151155] and square/rectangular CFSTs[150,153,154,156].

The structural behavior of FRP-conned CFSTs has recently been in-vestigated systematically by the authors' group [161163]. Within thisstudy, several series of laboratory tests were conducted to examine

the behavior of FRP-conned CFSTs under monotonic axial compres-sion, cyclic axial compression and the combined action of constantaxial compression and cyclic lateral loading. In addition, theoreticalmodels were developed to predict the experimental observations.

Existing research has indicated that FRP jacketing is highly effec-tive in delaying or even preventing the outward local buckling andin enhancing the performance of CFSTs subjected to various loadingschemes (i.e. monotonic and cyclic axial compression, and combinedaxial compression and cyclic lateral loading), in terms of both thestrength and ductility of the column [161,162]. Fig. 11 shows the en-hancement of the load-carrying capacity of CFSTs under axial com-pression by FRP jacketing.

9. Concluding remarks

External bonding of FRP reinforcement has been clearly establishedas a promising alternative strengthening technique for steel structuresby existing research. As more research is conducted and more reliabledesign guidelines become available, the technique is also expected toreceive increasing acceptance in practice. Based on the discussionspresented in this paper, it is recommended that future research shouldaddress the following issues with priority.

a b

orte

t. (

140 J.G. Teng et al. / Journal of Constructional Steel Research 78 (2012) 1311430 50

500

1000

1500

2000

2500

3000

Axi

al lo

ad (k

N)

Axial sh

c

Fig. 11. Strengthening of axially-loaded concrete-lled steel tubes with FRP connemen

shortening curves.10 15ning (mm)

CFSTConfined CFST

a) CFST specimen after test. (b) FRP-conned CFST specimen after test. (c) Axial load

[19] McKnight SH, Bourban PE, Gillespie JWJ, Karbhari VM. Surface preparation of

surface-roughness. Rubber Chem Technol 1995;68(1):1325.

141J.G. Teng et al. / Journal of Constructional Steel Research 78 (2012) 1311439.1. Steel surface treatment

More work should be conducted on the treatment of steel surfacepreparation and characterization to develop a widely accepted proce-dure for use in practice that can avoid adhesion failure at theadhesive/steel interface.

9.2. Selection and formulation of adhesives

In FRP-strengthened steel structures, the weak link is the adhesivelayer, provided adhesion failure at the adhesive/steel interface andthe FRP/adhesive interface can be avoided through appropriate sur-face preparation. As a result, at least for bond-critical applications,the material properties of the bonding adhesive play a key role in de-termining the load-carrying capacity of the strengthened structure;design theory needs to reect the mechanical properties of the adhe-sive. To maximize the effectiveness of FRP strengthening, the selec-tion of an appropriate adhesive is very important. The selectionprocess needs to consider not only the short-term mechanical perfor-mance but also its long-term durability and ease for handling on theconstruction site. It may also become necessary and benecial for ma-terial researchers to explore the development of better adhesiveswith properties tailored to the needs of FRP strengthening of steelstructures.

9.3. Bond behavior and debonding failures

Debonding failures are the most challenging issue in the exuralstrengthening of steel beams and the strengthening of thin-walledsteel structures against local buckling. As the adhesive is the weaklink, debonding failures in FRP-strengthened steel structures dependon the properties of the adhesive.Morework is needed to develop accu-rate bond-slip models for FRP-to-steel interfaces under Mode II loadingand under mixed mode loading, with parameters of bond-slip modelsbeing given in generic adhesive properties. Particular attention needsto be paid to debonding of bonded FRP plates loaded in compression,which has received little attention so far; such debonding arises oftenin the strengthening of steel structures against buckling failures.

9.4. Fatigue strengthening

Bonded CFRP patches provide a highly effective method for fatiguestrengthening of steel structures. In such applications, the elasticmodulus of the FRP rather than its ultimate tensile strength/strain isthe key parameter. Pre-tensioning of the FRP patch is highly desirable,but simple methods to pre-tension and anchor such FRP plates havenot yet been developed. In terms of theoretical modeling, a keyissue is the interaction between debonding and crack propagation. Abond-slip model for FRP-to-steel interfaces subjected to cyclic loadingis expected to be the key element in fatigue life prediction ofFRP-strengthened steel structures.

9.5. FRP connement of tubular structures

External FRP connement has been found to be an effectivestrengthening method for circular steel tubes with or without a con-crete inll, but not so effective for square or rectangular columns. Effec-tive methods for the strengthening of the latter columns need to bedeveloped.

9.6. Other issues

In addition to the topicsmentioned above, several other topics do notappear to have been explored and should be given due attention in thefuture, including: (1) durability of the bonding adhesive; (2) re resis-

tance of FRP-strengthened steel structures; (3) strengthening of steel[26] Tamai Y, Aratani K. Experimental study of relation between contact angle andsurface-roughness. J Phys Chem 1972;76(22):326771.

[27] Hitchcock SJ, Carroll NT, Nicholas MG. Some effects of substrate roughness onsteel for adhesive bonding in rehabilitation applications. ASCE Materials Engi-neering Conference. San Diego, CA; 1994. p. 114855.

[20] Harris AF, Beevers A. The effects of grit-blasting on surface properties for adhe-sion. Int J Adhes Adhes 1999;19(6):44552.

[21] Baldan A. Adhesively-bonded joints and repairs in metallic alloys, polymers andcomposite materials: adhesives, adhesion theories and surface pretreatment.J Mater Sci 2004;39(1):149.

[22] Schnerch D, Dawood M, Rizkalla S, Sumner E. Proposed design guidelines forstrengthening of steel bridges with FRP materials. Constr Build Mater2007;21(5):100110.

[23] Packham DE. Surface energy, surface topography and adhesion. Int J AdhesAdhes 2003;23(6):43748.

[24] Rosen SL. Fundamental principles of polymeric materials. John Wiley and Sons;1993.

[25] Gent AN, Lai SM. Adhesion and autohesion of rubber compounds effect ofstructures against blast and impact loading; (4) the use of external FRPreinforcement for combined strengthening and corrosion protection.

Acknowledgment

The authors are grateful for the nancial support from The HongKong Polytechnic University provided through its Niche Area FundingScheme, through a Postdoctoral Fellowship to the second author andan International Scholarship for PhD Studies to the third author. Inpreparing this paper, they have beneted from the list of references com-piled on the topic by Prof. X.L. Zhao of Monash University which wasmade available to members of the Working Group on FRP-StrengthenedMetallic Structures of the International Institute for FRP in Construction.

References

[1] Holloway LC, Head PR. Advanced polymer composites and polymers in the civilinfrastructure. Elsevier; 2001.

[2] Bank LC. Composites for construction: structural design with FRP materials. JohnWiley & Sons; 2006.

[3] ACI-440. Report on ber-reinforced polymer (FRP) reinforcement for concretestructures. ACI440; 2007.

[4] Hollaway LC, Teng JG. Strengthening and rehabilitation of civil infrastructuresusing bre-reinforced polymer (FRP) composites. England: Woodhead Publish-ing and Maney Publishing; 2008.

[5] Hollaway LC, Leeming MB. Strengthening of reinforced concrete structures usingexternally-bonded FRP composites in structural and civil engineering. Cambridge,UK: Woodhead Publishing Limited; 1999.

[6] Teng JG, Chen JF, Smith ST, Lam L. FRP strengthened RC structures. UK: JohnWiley & Sons; 2002.

[7] Tani K, Matsumura M, Kitada T, Hayashi H. Experimental study on seismicretrotting method of steel bridge piers by using carbon ber sheets. Proceed-ings of the Sixth Korea-Japan Joint Seminar on Steel Bridges. Tokyo, Japan;2000. p. 43745.

[8] Teng JG, Hu YM. Suppression of local buckling in steel tubes by FRP jacketing.Proceedings of the Second International Conference on FRP Composites in CivilEngineering. Adelaide, Australia; December 810 2004. p. 74953.

[9] Xiao Y. Applications of FRP composites in concrete columns. Adv Struct Eng2004;7(4):33543.

[10] Xiao Y, HeW, Choi K. Conned concrete-lled tubular columns. J Struct Eng ASCE2005;131(3):48897.

[11] Batikha M. Strengthening of thin metallic cylindrical shells using bre reinforcedpolymers. UK: The University of Edinburgh; 2008.

[12] Batikha M, Chen J, Rotter J, Teng J. Strengthening metallic cylindrical shellsagainst elephant's foot buckling with FRP. Thin-Walled Struct 2009;47(10):107891.

[13] Jiao H, Zhao X. CFRP strengthened butt-welded very high strength (VHS) circularsteel tubes. Thin-Walled Struct 2004;42(7):96378.

[14] Tavakkolizadeh M, Saadatmanesh H. Galvanic corrosion of carbon and steel inaggressive environments. J Compos Constr 2001;5(3):20010.

[15] Hollaway LC, Cadei J. Progress in the technique of upgrading metallic structureswith advanced polymer composites. Prog Struct Eng Mater 2002;4(2):13148.

[16] Schnerch D, Dawood M, Rizkalla S, Sumner E, Stanford K. Bond behavior of CFRPstrengthened steel structures. Adv Struct Eng 2006;9(6):80517.

[17] Teng JG, Fernando D, Yu T, Zhao XL. Treatment of steel surfaces for effective ad-hesive bonding. Proceedings of the 5th International Conference on FRP Com-posites in Civil Engineering, CICE-2010. Beijing, China; September 2729 2010.

[18] Mays GC, Huthcinson AR. Adhesives in civil engineering. New York: CambridgeUniversity Press; 1992.wettability. J Mater Sci 1981;16(3):71432.

142 J.G. Teng et al. / Journal of Constructional Steel Research 78 (2012) 131143[28] Gent AN, Lin CW. Model studies of the effect of surface-roughness and mechan-ical interlocking on adhesion. J Adhes 1990;32(23):11325.

[29] Amada S, Satoh A. Fractal analysis of surfaces roughened by grit blasting. J AdhesSci Technol 2000;14(1):2741.

[30] Lavaste V, Watts JF, Chehimi MM, Lowe C. Surface characterisation of compo-nents used in coil coating primers. Int J Adhes Adhes 2000;20(1):110.

[31] Sykes JM. Surface analysis and pretreatment of plastics and metals. Essex, England:Applied Science Publishers Ltd; 1982.

[32] Schnerch D. Strengthening of steel structures with high modulus carbon berreinforced polymer (CFRP) materials. Raleigh (NC): North Carolina State University;2005.

[33] Cadei JMC, Stratford TJ, Hollaway LC, Duckett WG. Strengthening metallic struc-tures using externally bonded bre-reinforced polymers- C595. London: CIRIA;2004.

[34] Gettings M, Kinloch AJ. Surface analysis of polysiloxane-metal oxide interfaces.J Mater Sci 1977;12:25118.

[35] El Damatty AA, Abushagur M. Testing and modeling of shear and peel behaviorfor bonded steel/FRP connections. Thin-Walled Struct 2003;41(11):9871003.

[36] Allan RC, Bird J, Clarke JD. Use of adhesives in repair of cracks in ship structures.Mater Sci Technol 1988;4(10):8539.

[37] Sallam HEM, Ahmad SSE, Badawy AAM, Mamdouh W. Evaluation of steel I-beamsstrengthened by various plating methods. Adv Struct Eng 2006;9(4):53544.

[38] Kim YJ, Garrett B. Interaction between CFRP-repair and initial damage ofwide-ange steel beams subjected to three-point bending. Compos Struct2011;93(8):198696.

[39] Kim YJ, Kent AH. Predictive response of notched steel beams repaired with CFRPstrips including bond-slip behavior. Int J Struct Stab Dyn 2012;12(1):121.

[40] Silvestre N, Young B, Camotim D. Non-linear behaviour and load-carrying capac-ity of CFRP-strengthened lipped channel steel columns. Eng Struct 2008;30(10):261330.

[41] Deng J, Lee MMK. Behaviour under static loading of metallic beams reinforcedwith a bonded CFRP plate. Compos Struct 2007;78:23242.

[42] Zhao XL, Fernando D, Al-Mahaidi R. CFRP strengthened RHS subjected to trans-verse end bearing force. Eng Struct 2006;28(11):155565.

[43] Fernando D, Yu T, Teng JG, Zhao XL. CFRP strengthening of rectangular steeltubes subjected to end bearing loads: effect of adhesive properties and nite el-ement modelling. Thin-Walled Struct 2009;47(10):10208.

[44] Harries KA, Peck AJ, Abraham EJ. Enhancing stability of structural steel sectionsusing FRP. Thin-Walled Struct 2009;47(10):1092101.

[45] Teng JG, Zhang JW, Smith ST. Interfacial stresses in reinforced concrete beamsbonded with a soft plate: a nite element study. Constr Build Mater2002;16(1):114.

[46] Teng JG, Chen JF, Smith ST, Lam L. Behaviour and strength of FRP-strengthenedRC structures: a state-of-the-art review. Proc Inst Civil Eng 2003;156(1):5162.

[47] Teng JG, Smith ST, Yao J, Chen JF. Intermediate crack-induced debonding in RCbeams and slabs. Constr Build Mater 2003;17(67):44762.

[48] Yao J, Teng JG, Chen JF. Experimental study on FRP-to-concrete bonded joints.Compos B 2005;36(2):99113.

[49] Yao J, Teng JG, Lam L. Experimental study on intermediate crack debonding inFRP-strengthened RC exural members. Adv Struct Eng 2005;8(4):36596.

[50] Yao J, Teng JG. Plate end debonding in FRP-plated RC beams-I: experiments. EngStruct 2007;29(10):245771.

[51] Xia SH, Teng JG. Behavior of FRP-to-steel bond joints. International Symposiumon Bond Behaviour of FRP in Structures (BBFS 2005). Hong Kong, China; 2005.

[52] Zhao XL, Zhang L. State-of-the-art review on FRP strengthened steel structures.Eng Struct 2007;29(8):180823.

[53] Yu T, Fernando D, Teng JG, Zhao XL. Experimental study on CFRP-to-steel bondedinterfaces. Compos B 2012;43(5):227989.

[54] Miller TC, Chajes MJ, Mertz DR, Hastings JN. Strengthening of a steel bridge gird-er using CFRP plates. J Bridge Eng, ASCE 2001;6(6):5238.

[55] Sebastian WM. Nonlinear proportionality of shear-bond stress to shear force inpartially plastic regions of asymmetric FRC-laminated steel members. Int J SolidsStruct 2003;40(1):2546.

[56] Nozaka K, Shield CK, Hajjar JF. Design of a test specimen to assess the effectivebond length of carbon ber-reinforced polymer strips bonded to fatigued steelbridge girders. J Compos Constr 2005;9(4):30412.

[57] Nozaka K, Shield CK, Hajjar JF. Effective bond length of carbon-ber-reinforcedpolymer strips bonded to fatigued steel bridge I-girders. J Bridge Eng, ASCE2005;10(2):195205.

[58] Colombi P, Poggi C. Strengthening of tensile steel members and bolted jointsusing adhesively bonded CFRP plates. Constr Build Mater 2006;20(12):2233.

[59] Fawzia S, Al-Mahaidi R, Zhao XL, Rizkalla S. Strengthening of circular hollowsteel tubular sections using high modulus CFRP sheets. Constr Build Mater2007;21(4):83945.

[60] Fernando D. Bond behaviour and debonding failures in CFRP-strengthened steelmembers. PhD, Hong Kong Polytechnic University, Kowloon; 2010.

[61] Haghani R. Analysis of adhesive joints used to bond FRP laminates to steel mem-bers - a numerical and experimental study. Constr Build Mater 2010;24(11):224351.

[62] Al-Zubaidy H, Al-Mahaidi R, Zhao XL. Experimental investigation of bond charac-teristics between CFRP fabrics and steel plate joints under impact tensile loads.Compos Struct 2012;94(2):5108.

[63] Hidekuma Y, Kobayashi A, Okuyama Y, Miyashita T, Nagai M. Experimental studyon debonding behavior of CFRP for axial tensile reinforced steel plate by CFRPstrand sheets. Proceedings of the Third Asia-Pacic Conference on FRP in Struc-tures (APFIS2012), Sapporo, Japan; 2012.[64] Matsumoto Y, Long ND, Yamada S, Matsui T, Matsumoto Y, Long ND. Mechanicalcharacteristics of CFRP reinforcement for corroded steel under axial tension. Pro-ceedings of the Third Asia-Pacic Conference on FRP in Structures (APFIS2012),Sapporo, Japan; 2012.

[65] Wu C, Zhao XL, Duan WH, Al-Mahaidi R. Bond characteristics between ultra highmodulus CFRP laminates and steel. Thin-Walled Struct 2012;51:47157.

[66] Chen JF, Teng JG. Anchorage strength models for FRP and steel plates bonded toconcrete. J Struct Eng, ASCE 2001;127(7):78491.

[67] Yuan H, Teng JG, Seracino R, Wu ZS, Yao J. Full-range behavior of FRP-to-concretebonded joints. Eng Struct 2004;26(5):55365.

[68] Bocciarelli M. Response of statically determined steel beams reinforced by CFRPplates in the elasticplastic regime. Eng Struct 2009;31(4):95667.

[69] Bocciarelli M, Colombi P. Elasto-plastic debonding strength of tensile steel/CFRPjoints. Eng Fract Mech 2012;85:5972.

[70] Yuan H, Wu Z. Theoretical solutions on interfacial stress transfer of externallybonded steel/composite laminates. Proceedings of the Symposium of Chinaand Japan: Science and Technology of 21st Century, Tokyo; 1999.

[71] Volkersen O. Die nietkraftverteilung in zugbeanspruchten Nietverbindungen mitkonstanten laschenquerschnitten. Luftfahrtforschung 1938;15:417.

[72] Adams RD, Peppiatt NA. Stress analysis of adhesively bonded lap joints. J StrainAnal Eng Des 1974;9(3):18596.

[73] Hart-Smith LJ. Development in adhesives-2. London: Applied Science Publishing;1981.

[74] Goland M, Reissner E. The stresses in cemented joints. J Appl Mech 1944;11:A1727.

[75] Adams RD, Peppiatt NA. Effect of Poisson's ratio strains in adherends on stressesof an idealized lap joint. J Strain Anal Eng Des 1973;8(2):1349.

[76] Rabinovitch O, Frostig Y. Closed-form high-order analysis of RC beams strength-ened with FRP strips. J Compos Constr 2000;4(2):6574.

[77] Shen HS, Teng JG, Yang J. Interfacial stresses in beams and slabs bonded withthin plate. J Eng Mech, ASCE 2001;127(4):399406.

[78] Smith ST, Teng JG. Interfacial stresses in plated beams. Eng Struct 2001;23(7):85771.

[79] Yang J, Ye JQ. An improved closed-form solution to interfacial stresses in platedbeams using a two-stage approach. Int J Mech Sci 2010;52(1):1330.

[80] Adams RD. Stress analysis: a nite-element analysis approach. In: Kinloch AJ, ed-itor. Developments in adhesives-2. London: Applied Science Publishers; 1981.

[81] Zhang L, Teng JG. Finite element prediction of interfacial stresses in structuralmembers bonded with a thin plate. Eng Struct 2010;32:45971.

[82] Harrison NL, Harrison WJ. Stresses in an adhesive layer. J Adhes 1972;3(3):195212.

[83] Taljsten B. Dening anchor lengths of steel and CFRP plates bonded to concrete.Int J Adhes Adhes 1997;17(4):31927.

[84] Akbar I, Oehlers DJ, Mohamed-Ali MS. Derivation of the bond-slip characteristicsfor FRP plated steel members. J Constr Steel Res 2010;66(89):104756.

[85] Wu ZS, Yuan H, Niu H. Stress transfer and fracture propagation in different kindsof adhesive joints. J Eng Mech, ASCE 2002;128(5):56273.

[86] Ibrisam A, Oehlers DJ, Ali MSM. Derivation of the bondslip characteristics forFRP plated steel members. J Constr Steel Res 2010;66(89):104756.

[87] Lu XZ, Teng JG, Ye LP, Jiang JJ. Bond-slip models for FRP sheets/plates bonded toconcrete. Eng Struct 2005;27(6):92037.

[88] Tavakkolizadeh M, Saadatmanesh H. Strengthening of steel-concrete compositegirders using carbon ber reinforced polymers sheets. J Struct Eng, ASCE2003;129(1):3040.

[89] Al-Saidy AH, Klaiber FW, Wipf TJ. Repair of steel composite beams with carbonber-reinforced polymer plates. J Compos Constr 2004;8(2):16372.

[90] Colombi P, Poggi C. An experimental, analytical and numerical study of the staticbehavior of steel beams reinforced by pultruded CFRP strips. Compos B2006;37(1):6473.

[91] Lenwari A, Thepchatri T, Albrecht P. Debonding strength of steel beamsstrengthened with CFRP plates. J Compos Constr 2006;10(1):6978.

[92] Deng J, Lee MMK. Fatigue performance of metallic beam strengthened with abonded CFRP plate. Compos Struct 2007;78(2):22231.

[93] Schnerch D, Rizkalla S. Flexural strengthening of steel bridges with high modu-lus CFRP strips. J Bridge Eng 2008;13(2):192201.

[94] Shaat A, Fam A. Repair of cracked steel girders connected to concrete slabs usingcarbon-ber-reinforced polymer sheets. J Compos Constr 2008;12(6):6509.

[95] Fam A, MacDougall C, Shaat A. Upgrading steel-concrete composite girders andrepair of damaged steel beams using bonded CFRP laminates. Thin-Walled Struct2009;47(10):112235.

[96] Linghoff D, Haghani R, Al-Emrani M. Carbon-bre composites for strengtheningsteel structures. Thin-Walled Struct 2009;47(10):104858.

[97] Haghani R, Al-Emrani M. A new design model for adhesive joints used to bondFRP laminates to steel beams: part B: experimental verication. Constr BuildMater 2012;30:68694.

[98] Sen R, Liby L, Mullins G. Strengthening steel bridge sections using CFRP lami-nates. Compos B:Eng 2001;32(4):30922.

[99] Tavakkolizadeh M, Saadatmanesh H. Repair of damaged steel-concrete compos-ite girders using carbon ber-reinforced polymer sheets. J Compos Constr2003;7(4):31122.

[100] Moy SSJ. FRP composites: life extension and strengthening of metallic structures.Institution of Civil Engineers Design and Practice Guide. London: ThomasTelford; 2001.

[101] Deng J, Lee MMK, Li S. Flexural strength of steelconcrete composite beamsreinforced with a prestressed CFRP plate. Constr Build Mater 2011;25(1):37984.

[102] Fitton M, Broughton I. Variable modulus adhesives: an approach to optimisedjoint performance. Int J Adhes Adhes 2005;25(4):32936.

[132] Tsouvalis NG, Mirisiotis LS, Dimou DN. Experimental and numerical study of thefatigue behaviour of composite patch reinforced cracked steel plates. Int JFatigue 2009;31(10):161327.

143J.G. Teng et al. / Journal of Constructional Steel Research 78 (2012) 1311432010;44(3):287302.[104] Stratford T, Cadei J. Elastic analysis of adhesion stresses for the design of a

strengthening plate bonded to a beam. Constr Build Mater 2006;20(12):3445.[105] Linghoff D, Al-Emrani M. Performance of steel beams strengthened with CFRP

laminate - part 2: FE analysis. Compos B 2010;41(7):51622.[106] Rabinovitch O, Frostig Y. On edge stresses control in strengthened RC beams

with FRP strips: adhesive layer prole effect. J Eng Mech 2001;127(4):31725.[107] Chiew SP, Yu Y, Lee CK. Bond failure of steel beams with FRP laminates - part 1:

model development. Compos B 2011;42(5):111421.[108] Yu Y, Chiew SP, Lee CK. Bond failure of steel beams with FRP laminates - part 2:

verication. Compos B 2011;42(5):112234.[109] Zhang L, Teng JG. Elastic lateral buckling of Isection beams bonded with a thin

plate. Proceedings of the Third International Conference on Advanced Compos-ites in Construction (ACIC 2007), University of Bath, UK; April 24 2007.p. 4416.

[110] Hollaway LC, Zhang L, Photiou NK, Teng JG, Zhang SS. Advances in adhesivejoining of carbon bre/polymer composites to steel members for repair and re-habilitation of bridge structures. Adv Struct Eng 2006;9(6):791803.

[111] Zhang L, Hollaway LC, Teng JG, Zhang SS. Strengthening of steel bridges underlow frequency vibrations. Proceedings of the Third International Conference onFRP Composites in Civil Engineering. Miami, Florida, USA; December 13152006.

[112] Bassetti A, Nussbaumer A, Hirt MA. Crack repair and fatigue life extension ofriveted bridge members using composite materials. Bridge Engineering Confer-ence 2000: Past Achievements, Current Practices, Future Technologies. Egypt:Sharm El-Sheikh; 2000.

[113] Bassetti A, Nussbaumer A, Hirt MA. Fatigue life extension of riveted bridge mem-bers using prestressed carbon ber composites. International Conference onSteel Structures of the 2000's, Istanbul, Turkey; 2000. p. 37580.

[114] Colombi P, Bassetti A, Nussbaumer A. Crack growth induced delamination onsteel members reinforced by prestressed composite patch. Fatigue Fract EngMater Struct 2003;26(5):42937.

[115] DawoodM, Rizkalla S, Sumner E. Fatigue and overloading behavior of steel-concretecomposite exural members strengthened with high modulus CFRP materials.J Compos Constr 2007;11(6):65969.

[116] Liu H, Al-Mahaidi R, Zhao XL. Experimental study of fatigue crack growth behaviorin adhesively reinforced steel structures. Compos Struct 2009;90:1220.

[117] Liu H, Xiao Z, Zhao XL, Al-Mahaidi R. Prediction of fatigue life forCFRP-strengthened steel plates. Thin-Walled Struct 2009;47:106977.

[118] Jiao H, Zhao XL, Mashiri F. Improving fatigue performance of CFRP strengthenedsteel beams by applying vacuum pressure in the wet layup of CFRP wovensheets. Proceedings of the Third Asia-Pacic Conference on FRP in Structures(APFIS2012), Sapporo, Japan; 2012.

[119] TavakkolizadehM, Saadatmanesh H. Fatigue strength of steel girders strengthenedwith carbon ber reinforced polymer patch. J Struct Eng, ASCE 2003;129(2):18696.

[120] Egilmez OO, Yormaz D. Cyclic testing of steel I-beams reinforced with GFRP.Steel Compos Struct 2011;11(2):93114.

[121] Ghafoori E, Motavalli M, Botsis J, Herwig A, Galli M. Fatigue strengthening of dam-agedmetallic beams using prestressed unbonded and bonded CFRP plates, Int J Fa-tigue (in press).

[122] Colombi P, Bassetti A, Nussbaumer A. Analysis of cracked steel membersreinforced by pre-stress composite patch. Fatigue Fract Eng Mater Struct2003;26(1):5966.

[123] Nakamura H, Maeda K, Suzuki H, Irube T, Nakamura H, Maeda K, Suzuki H, IrubeT. Monitoring for fatigue crack propagation of steel plate repaired by CFRP strips.Proceedings of the 4th International Conference on Bridge Maintenance, Safetyand Management, IABMAS'08; 2008. p. 294350.

[124] Bocciarelli M, Colombi P, Fava G, Poggi C. Fatigue performance of tensile steelmembers strengthened with CFRP plates. Compos Struct 2009;87(4):33443.

[125] Colombi P, Giulia F. Fatigue behaviour of tensile steel/CFRP joints. Compos Struct2012;94(8):240717.

[126] Wu C, Zhao XL, Al-Mahaidi R, Emdad M, Duan WH. Fatigue tests of cracked steelplates strengthened with UHM CFRP plates, Adv Struct Eng (in press).

[127] Jones SC, Civjan SA. Application of ber reinforced polymer overlays to extendsteel fatigue life. J Compos Constr 2003;7(4):3318.

[128] Nakamura H, Jiang W, Suzuki H, Maeda K, Irube T. Experimental study on repairof fatigue cracks at welded web gusset joint using CFRP strips. Thin-WalledStruct 2009;47(10):105968.

[129] Chen T, Yu QQ, Gu XL, Zhao XL. Study on fatigue behavior of strengthenednon-load-carrying cruciform welded joints using carbon ber sheets. Int J StructStab Dyn 2012;12(1):17994.

[130] Xiao ZG, Zhao XL. CFRP repaired welded thin-walled cross-beam connectionssubject to in-plane fatigue loading. Int J Struct Stab Dyn 2012;12(1):195211.

[131] Taljsten B, Hansen CS, Schmidt JW. Strengthening of old metallic structures in fa-tigue with prestressed and non-prestressed CFRP laminates. Constr Build Mater2009;23(4):166577.[133] Megueni A, Bouiadjra BB, Belhouari M. Disbond effect on the stress intensity fac-tor for repairing cracks with bonded composite patch. Comput Mater Sci2004;29:40713.

[134] Ouinas D, Serier B, Bouiadjra BB. The effects of disbonds on the pure mode IIstress intensity factor of aluminum plate reinforced with bonded composite ma-terials. Comput Mater Sci 2007;39:7827.

[135] Shaat A, Fam AZ. Slender steel columns strengthened using high-modulus CFRPplates for buckling control. J Compos Constr 2009;13(1):212.

[136] Silvestre N, Camotim D, Young B. On the use of the EC3 and AISI specications toestimate the ultimate load of CFRP-strengthened cold-formed steel lipped chan-nel columns. Thin-Walled Struct 2009;47(10):110211.

[137] Kim YJ, Harries KA. Behaviour of tee-section bracing members retrotted withCFRP strips subjected to axial compression. Compos B:Eng 2011;42(4):789800.

[138] El-Tawil S, Ekiz E, Goel S, Chao S. Retraining local and global buckling behavior ofsteel plastic hinges using CFRP. J Constr Steel Res 2011;67(3):2619.

[139] Kitada T, Yamaguchi T, Matsumura M, Okada J, Ono K, Ochi N. New technologiesof steel bridges in Japan. J Constr Steel Res 2002;58(1):2170.

[140] Hu YM. Behaviour of FRP-conned steel tubes under axial compression. BEngThesis, Department of Civil and Structural Engineering, The Hong KongPolytechnic University, April; 2004.

[141] Teng J, Hu YM. Behaviour of FRP-jacketed circular steel tubes and cylindricalshells under axial compression. Constr Build Mater 2007;21(4):82738.

[142] Nishino T, Furukawa T. Strength and deformation capacities of circular hollowsection steel member reinforced with carbon ber. Proceedings of the SeventhPacic Structural Steel Conference. American Institute of Steel Construction;March 2004.

[143] Haedir J, Zhao XL, Grzebieta RH, Bambach MR. Non-linear analysis to predict themoment-curvature response of CFRP-strengthened steel CHS tubular beams.Thin-Walled Struct 2011;49:9971006.

[144] Haedir J, Zhao XL. Design of short CFRP-reinforced steel tubular columns.J Constr Steel Res 2011;67:497509.

[145] Bhetwal KK, Yamada S. Effects of CFRP reinforcements on the buckling behaviorof thin-walled steel cylinders under compression. Int J Struct Stab Dyn2012;12(1):13151.

[146] Manos GC, Clough RW. Tank damage during the Coalinga earthquake. Earth-quake Eng Struct Dyn 1985;13:44966.

[147] Rotter JM. Local collapse of axially compressed pressurized thin steel cylinders.J Struct Eng, ASCE 1990;116(7):195570.

[148] Chen JF, Rotter JM, Teng JG. A simple remedy for elephants foot buckling in cy-lindrical silos and tanks. Adv Struct Eng 2006;9(3):40920.

[149] Teng JG, Lam L. Behavior and modeling of ber reinforced polymer-connedconcrete. J Struct Eng, ASCE 2004;130(11):171323.

[150] Mao XY, Xiao Y. Seismic behavior of conned square CFT columns. Eng Struct2006;28(10):137886.

[151] Shan JH, Chen R, ZhangWX, Xiao Y, Yi WJ, Lu FY. Behavior of concrete lled tubesand conned concrete lled tubes under high speed impact. Adv Struct Eng2007;10(2):20918.

[152] Gu W, Guan CW, Zhao YH, Cao H. Experimental study on concentrically-compressed circular concrete lled CFRP-steel composite tubular short columns.J Shenyang Archit Civil Eng Univ 2004;20(2):11820 (in Chinese).

[153] Tao Z, Han LH, Zhuang JP. Axial loading behavior of CFRP strengthenedconcrete-lled steel tubular stub columns. Adv Struct Eng 2007;10(1):3746.

[154] Wang ZB, Tao Z, Han LH. Performance of CFRP-strengthened concrete lled steeltubes under axial compression. Proceedings, 10th International Symposium onStructural Engineering for Young Experts, Hunan University, Changsha, China;2008. p. 68994.

[155] Liu L, Lu YY. Axial bearing capacity of short FRP conned concrete-lled steel