Construction and Building Materials 69 (2014) 323–334

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier .com/locate /conbui ldmat

Experimental investigation of slender circular RC columns strengthenedwith FRP composites

http://dx.doi.org/10.1016/j.conbuildmat.2014.07.0530950-0618/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Tel.: +966 11 4676962; fax: +966 11 4677008.E-mail address: nadeem@ksu.edu.sa (N.A. Siddiqui).

Nadeem A. Siddiqui ⇑, Saleh H. Alsayed, Yousef A. Al-Salloum, Rizwan A. Iqbal, Husain AbbasDepartment of Civil Engineering, King Saud University, Riyadh 11421, Saudi Arabia

h i g h l i g h t s

� Response of CFRP wrapped circular RC columns under eccentric compression is studied.� Columns of 150 mm diameter having three heights: 600, 900 and 1200 mm, were tested.� Columns were strengthened using three schemes of hoop and longitudinal CFRP wraps.� Hoop wraps provide lateral support to longitudinal fibers and increase column strength.� An expression for the slenderness limit of FRP-strengthened RC columns is proposed.

a r t i c l e i n f o

Article history:Received 31 January 2014Received in revised form 13 June 2014Accepted 18 July 2014Available online 12 August 2014

Keywords:FRPStrengtheningSlender columnsRCCircular columns

a b s t r a c t

The relevant design code provisions for Fiber Reinforced Polymer (FRP) strengthened RC columns arerestricted to the short RC columns strengthened with FRP jackets. These design provisions are thusstrictly not applicable to those long RC columns where second-order/slenderness effect is substantial.In the present study, the effectiveness of hoop and longitudinal Carbon FRP (CFRP) wraps in reducingthe lateral deflections and improving the strength of slender circular RC columns has been studied exper-imentally. A total of 12 small-scale circular RC columns of 150 mm diameter were cast in three groups,each group containing 4 columns of the same height. The columns of the first group belonged to shortcolumns of 600 mm height, whereas the columns of second and third groups of 900 and 1200 mm heightsrespectively represented slender columns. Columns of each group had one control and 3 strengthenedcolumns. The strengthened columns were prepared using three different strengthening schemes. In thefirst strengthening scheme, the columns were wrapped using a single layer of hoop CFRP sheet, whereasother strengthening schemes employed 2 and 4 longitudinal CFRP sheets in addition to one layer of hoopCFRP wrap. The columns were tested under monotonic compression with initial eccentricity of 25 mm. Ingeneral, CFRP-strengthening improves the strength and ductility of slender RC columns substantially. Thetest results indicate that the CFRP hoop wraps provide confinement to concrete and lateral support to thelongitudinal fibers and thus increase the strength of the RC columns. In slender columns, the effect of lon-gitudinal FRP fibers in carrying the load in post-yielding stage is more significant than hoop FRP fibers.The existing ACI expression of slenderness limit for RC columns was extended to propose a simple ana-lytical equation for the slenderness limit of FRP-strengthened RC columns. The proposed expression forslenderness limit is valid for both RC and FRP-confined columns and matches well with the experimen-tally observed slenderness limit.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

All over the world, strengthening and repair of reinforcedconcrete (RC) components using Fiber Reinforced Polymers (FRP)are gaining a wide popularity and acceptance due to its

well-established and promising performance [1–5]. TheFRP-jacketing of short RC circular columns is more effective inimproving its ductility and strength due to the effective lateralconfinement. Although, FRP-strengthening technique is veryeffective for short columns, still the designers are facing difficultiesin using this material for the strengthening of slender/long RCcolumns. Two of the main reasons are the limited researches andunavailability of complete guidelines and design provisions for

324 N.A. Siddiqui et al. / Construction and Building Materials 69 (2014) 323–334

FRP-strengthening of slender RC columns. The design provisions ofmajority of the design Codes such as ACI-440.2R [6,7], CNR-DT200[8]; fib Bulletin No. 14 [9], ISIS Canada [10]; Concrete Society [11];GB-50608 [12] etc. are limited to the FRP-strengthening of shortcolumns in which slenderness or second-order effect is negligible.

The early study in the field of FRP-confined slender column wasperformed by Mirmiran et al. [13]. They studied concrete-filledfiber-reinforced polymer tubes (CFFT) and observed a substantialreduction in the column load-carrying capacity with an increasein the slenderness ratio. The CFFT columns were more sensitiveto slenderness effects because the used GFRP (Glass FRP) materialshave a lower stiffness and higher strength than steel. Pan et al. [14]confirmed that the effectiveness of FRP-strengthening decreaseswith an increase in the slenderness ratio. They concluded thatthe effect of the slenderness ratio on the load-carrying capacityof FRP-strengthened RC columns is more noteworthy than that ofunconfined RC columns because confinement increases thestrength instead of bending stiffness. Tao and Han [15] also con-firmed that increasing load eccentricity and slenderness ratiodecreases the effect of CFRP confinement on the improvement ofcolumn load-carrying capacity. Gajdosova and Bilcik [16] studiedthe two strengthening methods, transverse CFRP sheet wrappingand longitudinal Near Surface Mounted (NSM) CFRP strips bondinginto the grooves in concrete cover. They found that the effect oftransverse CFRP jacketing on the increase in strength is greaterfor short RC columns subjected to predominant compressive load-ing, and longitudinal NSM CFRP strips are more effective inenhancing the flexural load-carrying capacity of slender RC col-umns subjected to eccentric loading. Tao et al. [17], Fitzwilliamand Bisby [18], and Bisby and Ranger [19] studied the long FRP-strengthened RC columns experimentally. All the studied columnswere hinged at the ends and were tested under equal eccentricityconditions. The longest column was 20.4 times its diameter,whereas maximum eccentricity was equal to the diameter of thecolumn. They demonstrated that the longitudinal CFRP not onlydecreases the lateral deflection, but also increases the strength ofthe long RC columns. Jiang and Teng [20,21] proposed analyticalmodels for slender FRP-confined columns and validated the resultswith the test results taken from the literature. Tamuzs et al. [22]studied the stability of plain concrete slender FRP-confined col-umns under concentric compression. They showed that a tangen-tial wrapping increases the load-carrying capacity only forcolumns with slenderness ratio less than 40.

The available researches on FRP-confined slender RC columnshave revealed the two facts: (i) FRP jacketing may convert a shortRC column into a slender column, that is, a reinforced concrete col-umn which was classified as a short column may have to be con-sidered as slender column after FRP confinement; and (ii)effective utilization of enhanced concrete strength due to FRP-con-finement decreases as the column becomes more slender. The

Fig. 1. Idealized axial stress–strain diagram for FRP-confined columns.

prime reason behind above observations is the almost bilinearaxial stress–strain response of FRP-confined columns (Fig. 1) inwhich transition occurs at the level of the concrete unconfinedstrength. Although the two branches are almost linear, the slopeof the second branch is substantially smaller than the first branch.It is due to this reason the elastic modulus of the column decreasesabove failure load value of unconfined column [23]. The flexuralrigidity of the FRP-confined column thus decreases. Consequently,the increase in strength due to FRP-confinement cannot be utilizedunless the slenderness ratio of the confined column is smaller thana limiting value. In other words, if slenderness ratio of FRP-con-fined column is more than the limit value, the column will faildue to buckling much earlier than reaching to its ultimate strength.Consequently, FRP-confinement cannot be exploited fully forincreasing axial load capacity of the column through FRP-confine-ment. In the recent past a few Codes (e.g. ISIS Canada, [10]) and afew investigators [20,21,23] have proposed simple expressions forestimating the slenderness limit of FRP-confined columns. Theirproposed limits however, differ substantially from each other. Thusthere is a need for such a simple and rational expression for slen-derness limit which is valid for both RC and FRP-confined columnsand validated with experimentally observed slenderness limit.

The objective of the present research is to (i) experimentallystudy the effectiveness of hoop and longitudinal CFRP wraps inreducing the lateral deflections and improving the strength of cir-cular RC columns, and (ii) propose a simple and rational expressionfor slenderness limit which is valid for both RC and FRP-confinedcolumns and validated with experimentally observed slendernesslimit.

2. Experimental program

To study the response of CFRP-wrapped columns, 12 small scale circular RC col-umns were cast in three groups, each group having 4 columns of the same height.Columns of the three groups were of 600, 900 and 1200 mm heights respectively.Columns of each group had one unstrengthened (control) column and 3 strength-ened columns. The control columns were steel-reinforced with 4–8 mm diameterlongitudinal rebars (percentage of steel, qg = 1.1%) and 6 mm diameter ties at auniform spacing of 100 mm c/c, as shown in Fig. 2. While the columns are small

Fig. 2. Reinforcement of columns (l = 600, 900 and 1200 mm).

Table 2Details of the test specimens.

Group Specimendesignation

Initial slenderness ratio,klu/r

No. of CFRP layers

HoopCFRP

Long.CFRP

1 CON-600 16 0 0STR1-600 16 1 0STR2-600 16 1 2STR3-600 16 1 4

2 CON-900 24 0 0STR1-900 24 1 0STR2-900 24 1 2STR3-900 24 1 4

3 CON-1200 32 0 0STR1-1200 32 1 0STR2-1200 32 1 2STR3-1200 32 1 4

Table 3Material properties.

Parameter Nominal value

Concrete and steelAverage compressive strength of concrete, f 0c (MPa) 35.1Yield strength of the longitudinal rebars, fy (MPa) 420Yield strength of the transverse rebars, fys (MPa) 275Modulus of elasticity of rebars, Es (GPa) 200

CFRP composite systemType of FRP used Unidirectional CFRP

sheetElastic modulus of CFRP (in primary fibers

direction)77.3 � 103 MPa

Elastic modulus of CFRP (90� to the primary fibers) 40.6 MPaUltimate tensile strength of FRP 846 MPaFracture strain 1.1%Thickness of each layer, tf 1.0 mm

N.A. Siddiqui et al. / Construction and Building Materials 69 (2014) 323–334 325

compared to realistic columns, studies have shown no significant size effect for FRPwrapped RC columns [24]. Table 1 provides the nomenclature and strengtheningschemes used in the present experimental investigation.

By varying the heights of the columns, the slenderness ratio klu/r was varied asshown in Table 2. Here k is the effective length coefficient, taken unity for pinnedends, lu is the unsupported height of the column and r is the radius of gyration ofthe cross section of RC column which is calculated using gross cross-sectional prop-erties of concrete. A slenderness limit of klu/r = 34–12(M1/M2) < 40 is specified byACI 318 [25] for RC columns. With single curvature bending and equal endmoments, M1 = M2, the limit reduces to klu/r = 22. This indicates that 900 and1200 mm size columns are slender.

The strengthened columns in each group were prepared using three differentstrengthening schemes. In the first strengthening scheme, the columns werewrapped using one layer of CFRP sheet with the fibers in the hoop direction. Inthe other two strengthening schemes, the CFRP sheets were first wrapped aroundthe full column circumference keeping their fibers in the longitudinal direction. Asingle layer of hoop FRP sheet was then wrapped over them. In this manner, hoopFRP provided lateral support to the inner longitudinal fibers. The possibility of flex-ural debonding of longitudinal CFRP was thus reduced substantially. The secondand third strengthening schemes differ with each other in number of longitudinalCFRP layers employed. In the second scheme, two layers of longitudinal CFRP sheetswere used whereas in the third scheme, four layers were employed.

2.1. Casting of the test columns

All the columns were cast using single batch of ready mix concrete. Three150 � 300 mm standard cylinders were also cast to obtain mean – compressivestrength of concrete (f 0c ). The steel bars were obtained from the local market touse them as longitudinal and transverse reinforcements. Table 3 shows the averageproperties of the mix and the rebars. After the casting, the columns were curedusing intermittent spraying of water for the two subsequent weeks and then leftas is to dry for the two more weeks.

2.2. Attaching the CFRP-sheets

In order to attach the CFRP sheets to the RC columns, surface of the columnswas grounded using sandpaper and then further smoothened through sand blast-ing. The process was continued until the surface was levelled and no unevennessremained in the surface. The surface was then cleaned with acetone. To make surethat no dust or unwanted particles are adhered to the surface of CFRP sheets, thesheet was also wiped with acetone. A thin layer of the epoxy, consisting of resinand hardener in a ratio of 3:1, was applied to the concrete surface and the CFRPsheets were bonded to the surface of the column. Special attention was given toassure that there was no void between the CFRP sheet and the concrete surface.By pressing the sheets to the concrete surface, any excess epoxy was squeezedout. As mentioned above, in the first strengthening scheme a single layer of CFRPsheet was bonded to the concrete surface with the fibers oriented in the hoop direc-tion. In the second and the third strengthening schemes, fibers were oriented in thelongitudinal direction and the longitudinal wraps were epoxy-bonded around thefull column circumference before bonding the hoop wraps over them to improvetheir anchorage and reduce the likelihood of flexural debonding failures. In the sec-ond scheme, two layers of longitudinal CFRP sheets were used whereas in the thirdscheme four layers were employed. Fig. 3 shows the columns during the attach-ment of sheets on the faces of the columns. There was an overlap of 100 mm inhoop direction for hoop CFRP sheets. However, there was no overlap in the longitu-dinal CFRP sheets. The columns were kept in the laboratory under controlled con-ditions (Temperature = 25 ± 2 �C and relative humidity = 30%) to ensure fullcuring of the epoxy.

Table 1Designation used for the test specimens.

Columndesignation

Number ofcolumns

Details

CON 3 Control RC columnSTR1 3 Strengthened column, obtained after the CFRP-

strengthening of RC columns usingstrengthening scheme #1 (hoop CFRP layer = 1;longitudinal CFRP layer = 0)

STR2 3 Strengthened column, obtained after the CFRP-strengthening of RC columns usingstrengthening scheme #2 (hoop CFRP layer = 1;longitudinal CFRP layers = 2)

STR3 3 Strengthened column, obtained after the CFRP-strengthening of RC columns usingstrengthening scheme #3 (hoop CFRP layer = 1;longitudinal CFRP layers = 4)

Fig. 3. Strengthening of columns.

In the present study, in order to determine the mechanical properties of theCFRP sheets used, coupon samples were cut from the CFRP sheets. Three couponswere tested using INSTRON tensile testing machine with hydraulic grips. Each spec-imen was subjected to a gradually increasing uniaxial load until failure took place.The average mechanical properties obtained are reported in Table 3.

326 N.A. Siddiqui et al. / Construction and Building Materials 69 (2014) 323–334

2.3. Test procedure and instrumentation

All columns were tested using Amsler compression testing machine under pin-ned end conditions and an axial load eccentricity of 25 mm. A displacement controlconfiguration was used for loading. This was chosen based on ACI 318 [25] designrequirements for slender RC columns.

All columns were instrumented with strain gauges and LVDTs attached to theouter surface of columns to measure the strains in the longitudinal and hoop direc-tions at their mid-height. Lateral deflections at one-quarter points along column

Fig. 4. An instrumented column ready for testing.

(a) 600 mm column (b) 900 mm

Fig. 5. Failure pattern obser

height were recorded using linear potentiometers. Fig. 4 shows the test setup usedfor the testing of the columns. Some of the instrumentations used in the presentstudy can also be seen in this picture.

3. Discussion of test results

3.1. General behavior

All columns were tested to failure. Typical failures of columnsare shown in Figs. 5–8. The control columns were generally failedby a sudden loss of concrete cover in compression zone, followedby slight outward buckling of the longitudinal reinforcement. For600 and 900 mm columns, the failure was observed near mid-height whereas for 1200 mm column, it was a little away fromthe mid-height. The failure of the columns of strengtheningscheme 1 was initiated by the appearance of ripples in CFRP sheetin compression zone, followed by rupture of matrix on tension facedue to its longitudinal stretching because of flexural tension. Nofracture of fibers was observed and the ultimate failure was as aresult of crushing of concrete in compression zone. The failure ofcolumns of strengthening schemes 2 and 3 was initiated with theformation of ripples in outer circumferential CFRP sheet in com-pression zone near mid-height which can be attributed to thecrushing of concrete and/or buckling of inner longitudinal fibersafter their debonding from concrete surface. Further increase inload increased the zone of debonding as a result of concrete crush-ing and consequently more buckling of fibers in compression zoneand ultimately rupturing of the outer circumferential CFRP sheet.For strengthening scheme 2, the inner longitudinal fibers on ten-sion face also got ruptured and were exposed due to the ruptureof outer hoop fibers (Fig. 7), however, for strengthening scheme3, this was not observed due to 4 layers of longitudinal fibers(Fig. 8). The magnitude of debonding of longitudinal fibers fromconcrete surface near mid-height of columns increased with theslenderness ratio of columns due to increase in lateral deflectionand secondary bending moment.

3.2. Load–deflection (P–D) and axial load–moment (P–M) response

Figs. 9 and 10 show the variation of load–deformation (P–D)and load–moment (P–M) response for control and the strength-ened columns respectively. The total moment at mid-height, M,in Fig. 10 includes secondary moments determined using theapplied axial load, P, and the measured lateral deflection, D, soM = P(e + D), where, e is the initial eccentricity (=25 mm). There

column (c) 1200 mm column

ved in control columns.

(a) STR1 (b) STR2 (c) STR3

Fig. 6. Failure of 600 mm CFRP-strengthened columns.

(a) STR1 (b) STR2 (c) STR3

Fig. 7. Failure of 900 mm CFRP-strengthened columns.

N.A. Siddiqui et al. / Construction and Building Materials 69 (2014) 323–334 327

are two reasons for nonlinearity in P–D plots – the nonlinear mate-rial response and the lateral deflection due to the buckling of col-umn. Since the effect of nonlinear material behavior is smallcompared to buckling effect, the flat portion in P–D plot shouldbe mainly expected due to instability or buckling of the columns.The presence of nonlinearity in P–D causes higher order nonlinear-ity in P–M plots as P–M plot is quadratic for linear P–D variation.The flat portion in P–D plot is observed for all strengthened col-umns of 900 and 1200 mm height. However the flat portion isnot present in P–D plots of 600 mm column indicating short col-umn behavior which is also evident from the failure of these col-umns (Figs. 5–8). The flat portion in P–D plots for controlcolumns of 900 and 1200 mm height is also evident in Fig. 11where all control columns are plotted together.

Table 4 shows the percent increase in load carrying capacitydue to FRP strengthening. This table clearly illustrates that theincrease in load carrying capacity for strengthening scheme STR-1 varies from 34% to 73% (Table 4) which is mainly due to theincrease in the confining strength of concrete. The effect of hoopwraps on the strength of columns is more significant for short col-umns (73%) than slender columns (61% for 900 mm and 34% for1200 mm columns). The increase in load carrying capacity for

other two strengthening schemes can be attributed to the fibersof longitudinal as well as hoop CFRP sheets. The load is sharedby longitudinal fibers because they are laterally supported in hoopdirection by circumferential CFRP sheet. Moreover, the longitudi-nal fibers also provide resistance to flexure in the slender columnsof 900 and 1200 mm height which consequently increased the loadcarrying capacity. The hoop layer in slender columns also helps inpreventing the debonding of inner longitudinal fibers from con-crete. The increase in load carrying capacity through increase inthe number of longitudinal CFRP layers from two to four is sub-stantially less (27–40%) compared to when it was increased fromzero to two (65–77%). This may be attributed to lesser confinementof longitudinal fibers due to more layers of CFRP (i.e. four) instrengthening scheme STR-3. It is worth mentioning here thatthe role of longitudinal fibers in compression was normally ignoredin past researches (e.g. [18]). However, the present study clearlyillustrates that the longitudinal fibers can take significant load ifthey are adequately supported laterally by hoop fibers.

The unstrengthened column of 600 mm shows a sudden drop inthe load after reaching the peak (Fig. 11). This can be attributed tothe difference in the behavior of concrete inside and outside thecore which resulted in a sudden detachment of concrete cover. It

(a) STR1 (b) STR2 (c) STR3

Fig. 8. Failure of 1200 mm CFRP-strengthened columns.

328 N.A. Siddiqui et al. / Construction and Building Materials 69 (2014) 323–334

is worth mentioning here that the concrete inside the core wasconfined due to the presence of circular ties whereas the concreteoutside the core (i.e. cover) was not confined in unstrengthenedcolumns due to which it got detached at high load. However, instrengthened columns, concrete cover was also confined due tothe presence of CFRP wraps. These wraps prevented detachmentof concrete cover and thus avoided sudden drop in load especiallyin columns of 600 mm size.

3.3. Load-axial strain response

Axial strains at mid-height of columns on the load side (desig-nated as N) and on opposite to the load side (designated as S) areplotted in Fig. 12 for 600, 900 and 1200 mm columns respectively.Examination of the load versus axial strain behavior of 600 mmcolumns (Fig. 12) shows that the magnitude of compressive strainat N is greater than the strain at S up to the failure of column due tothe eccentric loading. The difference in strains at N and S increaseswith the increase in load due to higher level of nonlinearity at N asa result of higher magnitude of strain at N as compared to S.

For 900 and 1200 mm columns (Fig. 12), the magnitude of com-pressive strain at N is greater than the strain at S up to the loadlevel when secondary moment effect was not pronounced. Whencolumns begin to experience significant lateral deflection due tosecondary moment, the difference in axial strains at N and S startsreducing and finally become almost the same. This is because thefirst order moment Pe become insignificant compared to secondorder moment PD which increases sharply near failure. The axialstrain in 1200 mm columns become close to each other much ear-lier than the 900 mm columns owing to the introduction of sec-ondary moment at low axial strains in 1200 mm columns whichis because of their larger slenderness ratio. For STR-1 columns,there is no enhancement in flexural stiffness as there are no longi-tudinal fibers. It is due to this reason that the STR-1 columns afterreaching their peaks experience continuous drop in load with theincrease in secondary moment. However, the columns of the othertwo strengthening schemes (i.e. STR-2 and STR-3) show continu-ous increase in the load but at slower rate due to the contributionof longitudinal fibers in resisting the flexure. In general, the CFRPwrapped columns experienced much larger lateral deflections.

The axial strain versus load curves for the CFRP wrapped col-umns again displayed either an ascending or a descending second-ary branch depending on the slenderness of the columns.Compared with the unwrapped columns, the axial strains through-out the cross-sections of the wrapped columns were considerablyhigher due to the confinement provided by the CFRP wraps, ashas been observed in all previous studies on concentrically loadedFRP confined concrete [26].

3.4. Moment–curvature response and ductility

From the measured compressive strain in concrete and tensilestrain in steel, the relationship between the moment and the cur-vature was tracked up to the failure as shown in Figs. 13(a)–(c).The curvature / was estimated using / ¼ ðec þ esÞ=d. Here, ec andes are the maximum strains in compression concrete and tensionsteel respectively at mid-height section, and d is the distancebetween them. Under increasing moment, the column passesthrough the different stages as shown in Figs. 13(a)–(c). At smallloads when the tensile stresses are less than the modulus of rup-ture the entire cross section of the column resists bending andcompression develops on one side and tension on the other. Upto this point the moment–curvature relationship was linear asobserved in all the specimens (Figs. 13(a)–(c)). When the bendingmoment was sufficiently large to cause the tensile stress in theextreme fibers to be greater than the modulus of rupture, the con-crete on the tension side of the column was cracked and the rein-forcing bars on this side of the column begin to take the tensioncaused by the applied moment. When the moment increasedbeyond the cracking moment, the slope of the curve decreasesbecause of the decrease in the column stiffness due to cracking.In case of control specimens, the diagram follows almost a straightline from cracking moment to the point where the reinforcement isstressed to its yield point. After reaching to this point all the con-trol specimens show failure, with almost similar ductility (Table5). The column strengthened with only single layer of hoop FRP,takes the load further and failure is delayed substantially. The hoopFRP thus not only increases the load carrying capacity but also theductility substantially. For specimens strengthened using singlelayer of hoop FRP (STR-1 specimens), increase in ductility is

N.A. Siddiqui et al. / Construction and Building Materials 69 (2014) 323–334 329

maximum for 1200 mm columns and minimum in 600 mm col-umns whereas the increase in the load is minimum for 1200 mmand maximum for 600 mm columns (Table 4). Those specimenswhich were strengthened with 2 or 4 layers of longitudinal andthen one layer of hoop FRP (STR-2 or STR-3 specimens) the trendof ductility change is same as change in the ultimate load. Thatis, the ductility of 600 mm FRP-strengthened columns, is

0

150

300

450

600

750

900

0 5 10 15 20 25 30 35 40 45 50

Tota

l App

lied

Loa

d (k

N)

Lateral Deflection at Mid-height (mm)

CON-600STR1-600STR2-600STR3-600

(a) 600 mm column

0

150

300

450

600

750

900

0 5 10 15 20 25 30 35 40 45 50

Tota

l App

lied

Loa

d (k

N)

Lateral Deflection at Mid-height (mm)

CON-900STR1-900STR2-900STR3-900

(b) 900 mm column

0

150

300

450

600

750

900

0 5 10 15 20 25 30 35 40 45 50

Tota

l App

lied

Loa

d (k

N)

Lateral Deflection at Mid-height (mm)

CON-1200STR1-1200STR2-1200STR3-1200

(c) 1200 mm column

Fig. 9. Load–deflection response for control and CFRP wrapped columns.

maximum whereas for 1200 mm FRP-confined columns it is mini-mum. It is worth mentioning that ductility is calculated as /u//y,where /u is the ultimate curvature of the mid-height section andit is taken as curvature corresponding to the ultimate moment,and /y is the yield curvature of the same section. It is consideredequal to the curvature at first yield of the longitudinal steel.Although, in the calculation of ductility, ultimate curvature is

0

150

300

450

600

750

900

0 5 10 15 20 25 30 35 40 45 50

Tota

l App

lied

Loa

d (k

N)

Moment (kN.m)

CON-600

STR1-600

STR2-600

STR3-600

M=P.e

(a) 600 mm column

0

150

300

450

600

750

900

0 5 10 15 20 25 30 35 40 45 50

Tota

l App

lied

Loa

d (k

N)

Moment (kN.m)

CON-900

STR1-900

STR2-900

STR3-900

M=P.e

(b) 900 mm column

0

150

300

450

600

750

900

0 5 10 15 20 25 30 35 40 45 50

Tota

l App

lied

Loa

d (k

N)

Moment (kN.m)

CON-1200

STR1-1200

STR2-1200

STR3-1200

M=P.e

(c) 1200 mm column

Fig. 10. Load–Moment (P–M) response for control and CFRP wrapped columns.

;

330 N.A. Siddiqui et al. / Construction and Building Materials 69 (2014) 323–334

generally taken corresponding to 10% or 20% drop of the ultimateload or moment, but in the present study it was taken correspond-ing to ultimate moment itself because in some of the strengthenedspecimens this much drop was not observed.

4. Analytical prediction of peak load and slenderness limit

Consider a circular RC column section of diameter D; reinforcedwith n longitudinal steel bars of diameter / (area of steel = Ast), asshown in Fig. 14. The section is strengthened with externallybonded layers of FRP along the longitudinal and circumferentialdirections. The total thickness of FRP layers in longitudinal and cir-cumferential directions are tfl and tfc respectively. As the objectiveof Fig. 14 is to study the effect of slenderness, FRP layers in hoopdirection are not shown in the figure.

The properties of the gross transformed section can be writtenas [27]:

Ag ¼ gross transformed area of cross-section;

¼pD2

4þðns�1ÞAsþpnf Dtfl¼

pD2

41þðns�1Þqsþ4nf

tfl

D

� � ð1Þ

Ig ¼moment of inertia of gross transformed section;

¼ p64

D4 þ p8ðns � 1ÞD3

s ts þp8

nf D3tfl

¼ p64

D4 1þ 8ðns � 1Þ Ds

D

� �3 ts

D

� �þ 8nf

tfl

D

� �" # ð2Þ

where,

Ds ¼ diameter of equivalent smeared ring for longitudinal steel¼ D� 2dc � / ¼ D� 2de

de ¼ dc þ/2

qs ¼4As

pD2

dc ¼ clear cover to longitudinal reinforcing bars;ts ¼ thickness of equivalent smeared ring for longitudinal steel

¼ As

pDs¼ Dqs

4ðDs=DÞ ¼N/2

4Ds

ts

D¼ qs

4ðDs=DÞ

0

50

100

150

200

250

300

350

400

0 2 4 6 8 10 12

Tota

l App

lied

Loa

d (k

N)

Lateral Deflection at Mid-height (mm)

CON-600

CON-900

CON-1200

Fig. 11. Load–deflection response for control columns.

ns = modular ratio of reinforcing steel ¼ Es

E

Table 4Summa

Specdesi

CONSTR1STR2STR3

CONSTR1STR2STR3

CONSTR1STR2STR3

* Per** Per

cnf = modular ratio of FRP in compression ¼ a

Ef

Eca = a factor that incorporates strength of FRP in compression.The contribution of longitudinal FRP towards compressionwill be considered only if it is laterally supported by hoopFRP layers.

The slenderness ratio, k, of the column with effective length, l,can be calculated using:

k ¼ slenderness ratio ¼ lr¼ klu

rð3Þ

where, k is the effective length factor, lu is the unsupported lengthof the column and r is the radius of gyration of the column sectiongiven by:

r ¼ffiffiffiffiffiIg

Ag

s¼ D

4

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1þ 2ðns � 1Þ Ds

D

� �2

qs þ 8nftfl

D

� �

1þ ðns � 1Þqs þ 4nftfl

D

vuuuuut ð4Þ

The above analysis clearly illustrates that the slendernessratio is not affected by FRP in hoop direction. However, FRP inhoop direction controls the slenderness limit as it confines theconcrete and thus increases the strength of the concrete. TheFRP-confined compressive strength of concrete, f 0cc , can beestimated using [6]:

f 0cc ¼ f 0c 2:25

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1þ 7:9

fl

f 0c

s� 2

fl

f 0c

� �� 1:25

" #ð5Þ

where, f 0c is the unconfined – compressive strength of concrete, fl isthe lateral confining pressure due to the hoop FRP layer. The – con-fining pressure fl can be estimated using:

fl ¼2tflffe

Dð6Þ

where, ffe is the effective stress level in the FRP layer which is takenas 60% of the ultimate strength, ffu. The experimental results of thisstudy support this assumption.

The slenderness limit for change of column behavior from longto short can be estimated by equating the Euler stress to the con-fined compressive strength of concrete, thus giving the theoreticalupper bound value of slenderness ratio as,

kc ¼ p

ffiffiffiffiffiEc

f 0cc

sð7Þ

ry of the test results.

imengnation

Initial slendernessratio, (klu/r)

Peak load(kN)

% Increase with respectto control

-600 16 313.7 –-600 16 541.3 72.6-600 16 745.2 137.6 (65.0*)-600 16 829.9 164.6 (27.0**)

-900 24 244.7 –-900 24 393.7 60.9-900 24 580.9 137.4 (76.5*)-900 24 660.9 170.1 (32.7**)

-1200 32 258 –-1200 32 346.7 34.4-1200 32 545.2 111.3 (76.9*)-1200 32 647.1 150.8 (39.5**)

cent increase with respect to STR1.cent increase with respect to STR2.

N.A. Siddiqui et al. / Construction and Building Materials 69 (2014) 323–334 331

The above expression shows that the value of kc reduces from88.5 for un-strengthened RC column to 69.1 for CFRP confined RCcolumns of this study. As all strengthening schemes have one layerof circumferential CFRP sheet therefore the confined compressive

0

150

300

450

600

750

900

0 1 2 3 4 5 6

Tota

l App

lied

Loa

d (k

N)

Axial strain (%)

Strain at N

Strain at S

STR1-600

STR3-600

STR2-600

CON-600

N S

Loa

d-ax

is

(a) 600 mm column

0

150

300

450

600

750

900

0 1 2 3 4 5 6

Tota

l App

lied

Loa

d (k

N)

Axial strain (%)

Strain at N

Strain at S

STR1-900

STR3-900

STR2-900

CON-900

N S

Loa

d-ax

is

(b) 900 mm column

0

150

300

450

600

750

900

0 1 2 3 4 5 6

Tota

l App

lied

Loa

d (k

N)

Axial strain (%)

Strain at N

Strain at S

STR1-1200

STR3-1200

STR2-1200

CON-1200

N S

Loa

d-ax

is

(c) 1200 mm column

Fig. 12. Applied load versus axial strain for control and CFRP wrapped columns.

strength of concrete is same for all strengthening schemes. Thoughlong column behavior for CFRP confined columns will be for kgreater than or equal to kc but separate formula is required for pre-dicting the behavior of such columns for k smaller than kc. For thetransition from short to long column, the following modified Ran-kine’s formula [28] may be adopted:

0

5

10

15

20

25

30

35

40

45

0.0 0.2 0.4 0.6 0.8 1.0

Mom

ent (

kNm

)

Curvature (/m)

CON-600STR1-600STR2-600STR3-600

(a) 600 mm column

0

5

10

15

20

25

30

35

40

45

0.0 0.2 0.4 0.6 0.8 1.0

Mom

ent (

kNm

)

Curvature (/m)

CON-900

STR1-900

STR2-900

STR3-900

(b) 900 mm column

0

5

10

15

20

25

30

35

40

45

0.0 0.2 0.4 0.6 0.8 1.0

Mom

ent (

kNm

)

Curvature (/m)

CON-1200STR1-1200STR2-1200STR3-1200

(c) 1200 mm column

Fig. 13. Moment–curvature response for control and CFRP wrapped columns.

Table 5Curvature ductility of control and FRP-strengthened columns.

Column ID Slenderness ratio r ¼ffiffiffiffiIg

Ag

qCurvature Ductility

Ultimate Yield /u//y

/u /y

CON-600 16.2 0.058 0.035 1.6STR1-600 16.2 0.415 0.056 7.5STR2-600 15.3 0.514 0.029 17.7STR3-600 14.6 0.460 0.020 22.8

CON-900 24.3 0.027 0.016 1.7STR1-900 24.3 0.320 0.030 10.6STR2-900 22.9 0.478 0.032 14.9STR3-900 22.0 0.519 0.031 16.6

CON-1200 32.4 0.057 0.030 1.9STR1-1200 32.4 0.568 0.050 11.4STR2-1200 30.6 0.542 0.084 6.4STR3-1200 29.3 0.190 0.034 5.5

Table 6Analytical prediction of peak load of control and FRP-strengthened columns.

Column ID Slenderness

ratio, r ¼ffiffiffiffiIg

Ag

q Load carried by column (kN) Predicted/experimental

Experimental Predictedusing Eq. (8)

CON-600 16.2 313.7 275.0 0.88STR1-600 16.2 541.3 520.8 0.96STR2-600 15.3 745.2 637.1 0.85STR3-600 14.6 829.9 750.5 0.90

CON-900 24.3 244.7 268.6 1.10STR1-900 24.3 393.7 487.3 1.24STR2-900 22.9 580.9 600.6 1.03STR3-900 22.0 660.9 710.2 1.07

CON-1200 32.4 258 253.2 0.98STR1-1200 32.4 346.7 446.1 1.29STR2-1200 30.6 545.2 554.3 1.02STR3-1200 29.3 647.1 659.8 1.02

332 N.A. Siddiqui et al. / Construction and Building Materials 69 (2014) 323–334

PAg¼

f cyf 0cc

1þ eyr2

� �f mcy þ f 0cc

mh i1=m ð8Þ

where, y is the distance of extreme fiber from neutral axis; m is amodel parameter; fcy is the Euler stress for FRP confined RC column,given by:

fcy ¼p2Ec

k2 ð9Þ

The value of f 0cc is obviously equal to f 0c for unconfined RC col-umns. The effect of circular steel (ties) confinement is ignored, sothe above formula is applicable only to columns with limited trans-verse steel reinforcement. A comparison of the load carried by col-umn, P, using the above expression with experiments is shown inTable 6. The value of m in this calculation is taken as 1.05. The ratioof the predicted to experimental axial load varies from 0.85 to 1.29(Table 6) which indicates reasonable agreement with experiments.

ACI expression for slenderness limit of RC columns is based on5% axial load reduction [29]. Although MacGregor proposedanother criterion for slenderness limit based on 5% first-ordermoment amplification but the resulting equation remainedunchanged [30]. Keeping the same definition of 5% reduction inaxial load, the value of slenderness limit for FRP confined columnsof all the three strengthening schemes was obtained. Fig. 15 showsthe variation of axial load with slenderness ratio. A vertical line isdrawn in this figure to indicate the drop in axial load by 5%

Fig. 14. Equivalent transformed cross-secti

which is corresponding to a slenderness ratio of 17.4 for all FRPstrengthened concrete columns. Fig. 15 also shows the slendernesslimit for RC columns about 22. This value matches well with ACIproposed limit for RC columns having same end moments (i.e.M1 = M2).

ISIS design manual [10] proposes the following formula for theslenderness limit of FRP confined RC columns:

kc ¼25ffiffiffiffiffi

Nf

Af 0c

q ð10Þ

where Nf = factored axial load carrying capacity of column; andA = gross section area of column. The value of kc obtained fromthe above equation for CFRP confined columns of this study is18.8 which is close to that predicted above. A simplification of theabove equation has been incorporated in ACI slenderness limitand a new equation is proposed:

kc ¼ 34� 12M1

M2

� � ffiffiffiffiffif 0cf 0cc

sð11Þ

where, M1/M2 is the ratio of smaller to larger factored end momentson a compression member, to be taken as positive if member bendsin single curvature, and negative if it bends in double curvature. Thevalue of kc obtained from this equation for the CFRP strengthenedcolumns of the present study is 17.2 which is also very close tothe earlier predicted value of 17.4. The proposed expression being

on of column for analytical prediction.

0

150

300

450

600

750

900

0.0 8.0 16.0 24.0 32.0

Peak

Loa

d (k

N)

Slenderness ratio (kl/r)

Control

Hoop CFRP =1; Long. CFRP = 0

Hoop CFRP = 1; Long. CFRP =2

Hoop CFRP = 1; Long. CFRP = 4

Slenderness limit for FRP-strengthened columns

Slenderness limit for RC control columns

5% of 668 kN

5% of 783 kN

5% of 549 kN

5% of 282 kN

17.4 22

Fig. 15. Effect of slenderness and CFRP wrapping schemes on peak load capacity.

18.8

7.8

13.2

17.4 17.2

22.0

0

5

10

15

20

25

Slen

dern

ess

Lim

it

Fig. 16. A comparison of slenderness limit values.

N.A. Siddiqui et al. / Construction and Building Materials 69 (2014) 323–334 333

simpler and at the same time close to the experiments is recom-mended for calculating slenderness limit for FRP confined columns.

Fig. 16 compares ISIS Canada [10], Jiang and Teng [20], DeLorenzis and Tepfers [23] and proposed slenderness limits withthat of experimentally observed slenderness limit. The chartclearly illustrates that the slenderness limit proposed by all theinvestigators, including present, is less than that of ACI unwrappedcolumn. This is very well expected as FRP wrapping reduces theslenderness limit substantially (due to decrease in flexural rigidityabove failure load of unwrapped column). A comparison of Jiangand Teng [20] and De Lorenzis abd Tepfers [23] predictions withexperimentally observed slenderness limit show that theirpredictions are conservative for the columns of the present study.ISIS Canada [10] slightly overestimates the slenderness limit, but toan extent very close to the proposed and experimentally observed

slenderness limits. The proposed expression, however, yields thelimit which is very close to the experimentally observedslenderness limit. Since this comparison is based on a limitedexperimental data, further experimental verification through awide spectrum of data is required to rely on the validity of thepresent and/or other slenderness limit expressions.

5. Conclusions

The following conclusions can be drawn on the basis of thestudy presented herein:

1. CFRP hoop wraps provide confinement to concrete and lateralsupport to the longitudinal fibers and thus increase the strengthof both short and slender RC columns. However, the effect ofhoop wraps on the strength of columns is more significant forshort columns than slender columns. The outer hoop layer inslender columns also helps in reducing the possibilities offlexural debonding of inner longitudinal FRP fibers from theconcrete surface.

2. The fibers of longitudinal CFRP laminate can also contributesubstantially to the load carrying capacity provided they are lat-erally supported by hoop fibers. In slender columns, the load isprimarily carried by flexural action of longitudinal FRP fibers.

3. The column strengthened with single layer of hoop FRP not onlyincreases the load carrying capacity but also the ductility sub-stantially. For those specimens which were strengthened usingsingle layer of hoop FRP, increase in the ductility is maximumfor 1200 mm columns and minimum in 600 mm columnswhereas the increase in the load is minimum for 1200 mmand maximum for 600 mm columns.

4. For those specimens which were strengthened with 2 or 4 lay-ers of longitudinal and then one layer of hoop FRP, the ductilityand the ultimate load were maximum for 600 mm andminimum for 1200 mm FRP-strengthened columns.

5. The increase in load carrying capacity through increase in thenumber of longitudinal CFRP layers from two to four is substan-tially less (27–40%) compared to when it was increased fromzero to two (65–77%).

334 N.A. Siddiqui et al. / Construction and Building Materials 69 (2014) 323–334

6. The existing ACI expression of slenderness limit for RC columnshas been extended to cover FRP-strengthened columns by intro-ducing the effect of FRP-strengthening in the ACI expression.The proposed expression for slenderness limit is valid for bothRC and FRP-confined columns and matches well with the exper-imentally observed slenderness limit.

Acknowledgements

The Authors would like to extend their sincere appreciation tothe Deanship of Scientific Research at King Saud University for itsfunding of this research through the research group project No.RGP-VPP-104. Authors further extend their acknowledgement tothe MMB Chair at the Department of Civil Engineering, King SaudUniversity for providing technical support.

References

[1] Al-Salloum YA, Almusallam TH, Alsayed SH, Siddiqui NA. Seismic behavior ofas-built ACI-complying and CFRP-repaired exterior RC beam-column joints. JCompos Constr – ASCE 2011;15(4):522–34.

[2] Siddiqui NA. Experimental investigation of RC beams strengthened withexternally bonded FRP composites. Latin Am J Solids Struct (LAJSS)2009;6(4):343–62.

[3] Alsayed SH, Siddiqui NA. Reliability of shear-deficient RC beams strengthenedwith CFRP-strips. Constr Build Mater 2013;42:238–47.

[4] Alsayed SH, Al-Salloum YA, Almusallam TH, Siddiqui NA. Seismic response ofFRP-upgraded exterior RC beam-column joints. J Compos Constr – ASCE2010;14(2):195–208.

[5] Alsayed SH, Almusallam TH, Al-Salloum YA, Siddiqui NA. Seismic rehabilitationof corner RC beam-column joints using CFRP composites. J Compos Constr –ASCE 2010;14(6):681–92.

[6] American Concrete Institute (ACI). Guide for the design and construction ofexternally bonded FRP systems for strengthening concrete structures. ACI440.2R, 2002, Farmington Hills, MI.

[7] American Concrete Institute (ACI). Guide for the design and construction ofexternally bonded FRP systems for strengthening concrete structures. ACI440.2R, 2008, Farmington Hills, MI.

[8] CNR-DT200. Guide for the design and construction of externally bonded FRPsystems for strengthening existing structures. Advisory Committee onTechnical Recommendations for Construction, National Research Council,Rome, Italy; 2004.

[9] Fib Bulletin No. 14. Externally bonded FRP reinforcement for RC structures. TheInternational Federation for Structural Concrete, Lausanne, Switzerland; 2001.

[10] ISIS Canada. Design manual no. 4: Strengthening reinforced concretestructures with externally-bonded fiber reinforced polymers. Intelligentsensing for innovative structures, Winnipeg, Canada; 2001.

[11] Concrete Society. Design guidance for strengthening concrete structures withfiber composite materials, Second Edition, Concrete Society Technical ReportNo. 55, Crow Thorne, Berkshire, UK; 2004.

[12] GB-50608. Technical code for infrastructure application of FRP composites,China Planning Press, China; 2010.

[13] Mirmiran A, Shahawy M, Beitleman T. Slenderness limit for hybrid FRP –concrete columns. J Compos Constr – ASCE 2001;5(1):26–34.

[14] Pan JL, Xu T, Hu ZJ. Experimental investigation of load carrying capacity of theslender reinforced concrete column wrapped with FRP. Constr Build Mater2007;21(11):1991–6.

[15] Tao Z, Han LH. Behavior of fire-exposed concrete-filled steel tubular beamcolumns repaired with CFRP wraps. Thin-Walled Struct 2007;43(5):445–52.

[16] Gajdosova K, Bilcik J. Full scale testing of CFRP-strengthened slenderreinforced concrete columns. J Compos Constr – ASCE 2013;17(2):239–48.

[17] Tao Z, Teng JG, Han LH, Lam L. Experimental behaviour of FRP-confined slenderRC columns under eccentric loading. In: Proc., second int. conf. on advancedpolymer composites for structural applications in construction, Univ. ofSurrey, Guildford, UK; 2004: p. 203–12.

[18] Fitzwilliam J, Bisby LA. Slenderness effects on circular CFRP confinedreinforced concrete columns. J Compos Constr – ASCE 2010;14(3):280–8.

[19] Bisby LA, Ranger M. Axial-flexural interaction in circular FRP-confinedreinforced concrete columns. Constr Build Mater 2010;24(9):1672–81.

[20] Jiang T, Teng JG. Slenderness limit for short FRP-confined circular RC columns. JCompos Constr – ASCE 2012;16(6):650–61.

[21] Jiang T, Teng JG. Theoretical model for slender FRP-confined circular RCcolumns. Constr Build Mater 2012;32:66–76.

[22] Tamuzs V, Tepfers R, Zile E, Valdmanis V. Stability of round concrete columnsconfined by composite wrappings. Mech Compos Mater 2007;43(5):445–52.

[23] De Lorenzis L, Tepfers R. Applicability of FRP confinement to strengthenconcrete columns. Nordic Concrete Research No. 31, The Nordic ConcreteFederation 1/2004, Oslo, June 2004. p. 64–72.

[24] Rocca S, Galati N, Nanni A. Large-size reinforced concrete columnsstrengthened with carbon FRP: validation of existing design guidelines. In:Proc. Composites in Civil Engineering, Florida International University, Miami;2006. p. 231–34.

[25] American Concrete Institute (ACI). Building code requirements for structuralconcrete and commentary. ACI 318–11, 2011, Farmington Hills, MI.

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

[27] Elsanadedy HM, Almusallam TH, Abbas H, Al-Salloum YA, Alsayed SH. Effect ofblast loading on CFRP-Retrofitted RC columns – a numerical study. Latin Am JSolids Struct (LAJSS) 2011;8:55–81.

[28] Ross CTF, Case J, Chilver A. Strength of materials and structures. Butterworth-Heinemann, Aug 27, 1999 – Technology & Engineering.

[29] MacGregor JG, Breen JE, Pfrang EO. Design of slender concrete columns. ACI J1970;67(1):6–28.

[30] MacGregor JG. Design of slender concrete columns-revisited. ACI Struct J1993;90(3):302–9.