Concrete or FRP Jacketing of Columns with Lap Splices for

Journal of Advanced Concrete Technology Vol. 4, No. 3, 431-444, October 2006 / Copyright © 2006 Japan Concrete Institute 431

Scientific paper

Concrete or FRP Jacketing of Columns with Lap Splices for Seismic Rehabilitation Stathis Bousias1, Alexis-Loukas Spathis2 and Michael N. Fardis3

Received 14 January 2006, accepted 18 June 2006

Abstract Seismic rehabilitation of rectangular columns with corner deformed bars lap-spliced at floor level through RC jacketing or CFRP wrapping, is experimentally studied. In three unretrofitted columns, deformation capacity and energy dissipa-tion drop fast with lap length below 45-bar diameter. Three columns were cyclically tested after concrete jacketing of their full length and another nine after CFRP wrapping of the lap splice and plastic hinge region. Five CFRP layers were found to be more effective than two, but the improvement was not major. The positive effect of FRP wrapping on flex-ural resistance, ultimate deformation and energy dissipation declined with decreasing lap length. There is a limit to the improvement that can be obtained through FRP wrapping: if lapping is as short as 15-bar diameter, its adverse effects on force capacity and energy dissipation cannot be sufficiently removed by FRP wrapping. Overall, RC jacketing is more effective than FRP wrapping for the improvement of the deformation capacity of columns with 15-bar diameter lapping.

1. Introduction

Most concrete buildings in earthquake-prone regions of the world were constructed before the enforcement of modern seismic design codes and thus are inherently vulnerable to earthquakes. Short lap splices in columns are a typical deficiency in such buildings, especially when they are located at floor level, i.e. in the plastic hinge region. When subjected to deformation reversals well into the post-elastic region, such columns exhibit a rapid loss of strength and low deformation capacity. To enhance the ductility of old-type columns, conventional as well as innovative materials are currently employed. The first category comprises jackets of cast-in-situ or sprayed concrete. Externally bonded FRP wraps, en-hancing the limited confining action provided by exist-ing sparse stirrups, belong to the latter category.

Concrete jacketing of columns has traditionally been the method of choice in seismic rehabilitation of con-crete buildings, especially after damage due to a strong earthquake. The method is cost-effective, due to: (1) The familiarity of engineers and of the construction industry as a whole with the field application of struc-tural concrete. Rehabilitation, especially through modi-fication of existing components, does not lend itself to prefabrication in shop and structural concrete is the prime structural material for field fabrication and appli-cation. (2) The suitability of concrete jacketing for simultane-

ous repair of serious seismic damage, including local or more extensive concrete crushing, or even buckling of bars and fracture of stirrups: replacement of the crushed concrete is done in the same operation as casting or shotcreting of the jacket, while full restoration of buck-led or fractured bars may not be absolutely necessary, if such reinforcement is replaced by the new reinforce-ment of the jacket. This factor is of greater importance when seismic rehabilitation is triggered by seismic dam-age (as is often the case). (3) The versatility and shape-adaptability of reinforced concrete to fully encapsulate existing members and joints and provide structural continuity between differ-ent components (e.g., between a joint and the adjoining members, or between members in adjacent stories). (4) The ability of concrete jackets to have, through appropriate reinforcement, multiple effects, enhancing at the same time stiffness, flexural and shear strength, deformation capacity and anchorage and continuity of reinforcement in development or splicing zones. The enhancement of stiffness and flexural strength is due to the increased cross-sectional dimensions and the added longitudinal reinforcement, which—most importantly and unlike other retrofit techniques of individual mem-bers—can easily extend beyond the member end into and through joint regions. Although the added concrete is also a factor, the increase in shear strength, deforma-tion capacity and reinforcement development and splic-ing is mainly due to the added transverse reinforcement, which works for shear, confinement and restraint of longitudinal bars against buckling.

The multiple effectiveness of concrete jackets is what mainly differentiates them from steel or FRP jackets that do not normally extend beyond the end of the existing member and into the joints. With the increase in stiff-ness, global seismic displacements in general decrease and deformation demands on all members may be re-

1Assistant Professor, Structures Laboratory, Civil Engineering Dept., University of Patras, Greece. E-mail: [email protected] 2Civil Engineer, PhD 3Professor, Structures Laboratory, Civil Engineering Dept., University of Patras, Greece.

432 S. Bousias, A.-L. Spathis and M. N. Fardis / Journal of Advanced Concrete Technology Vol. 4, No. 3, 431-444, 2006

duced below the corresponding capacities. Moreover, in concrete frames jacketing can increase the flexural strength of weak columns above that of stronger beams.

Concrete jackets have certain drawbacks in compari-son to FRP wrapping for seismic rehabilitation of con-crete buildings. They considerably increase the cross sectional dimensions of members, which may be a seri-ous drawback when floor area is at a premium; they disrupt occupancy and produce dust and debris (espe-cially if shotcrete is used), and they create noise pollu-tion and health or safety hazards for workers to a greater degree than other rehabilitation techniques. As the im-portance of the factors for which concrete jackets are at a disadvantage relative to competing techniques tends to increase, the balance is turning against concrete jackets, despite their present advantage in direct rehabilitation costs.

Experimental work on the cyclic behavior of RC jacketed rectangular columns, beams or walls is very limited (Alcocer 1992; Alcocer et al. 1993; Bett et al. 1988; Choudhuri et al. 1992; Ersoy et al. 1993; Gomez et al. 1998; Ilki et al. 1998; Julio et al. 2005; Rodriguez et al. 1994; Stoppenhagen et al. 1995) and does not in-clude members with lap splices in the plastic hinge zone. By contrast, FRP jackets have been extensively studied experimentally, including rectangular columns with de-formed bars lap-spliced in the plastic hinge zone (Chang et al. 2001; Elnabelsy et al. 2004; Haroun et al. 2001; Harries et al. 2003; Saadatmanesh et al. 1997a; Saadat-manesh et al. 1997b).

FRP jackets are primarily used to enhance shear strength, deformation capacity and reinforcement devel-opment and splicing, but normally not flexural strength and stiffness. Although their cost-effectiveness has not been fully established yet, encasing RC members in FRP jackets is rapidly becoming the method of choice in seismic rehabilitation.

This paper presents a comparative experimental pro-gram on the rehabilitation of rectangular RC columns through concrete or FRP jackets. The vertical rein-forcement of the original column consists of deformed

bars at the corners of the section, lap-spliced at floor level. The length of lapping is the main parameter stud-ied. An additional parameter in the tests of FRP jacketed columns is the number of FRP layers.

2. Test specimens, materials, test setup and testing program

The experimental program comprised a total of 16 col-umns having dimensions, materials and detailing and cover of reinforcement typical of old RC buildings with-out proper detailing for earthquake resistance. Columns were tested as simple cantilevers fixed into a heavily reinforced 0.6 m deep footing, within which the vertical bars were anchored with 90 degree hooks at the bottom. The shear span, measured from the top of the footing to the point of application of the lateral load, was equal to 1.6 m, meant to represent approx. half a typical story height. The original column section was 250 by 500 mm (Fig. 1(a)) and was reinforced with a 18 mm deformed bar at each corner, having a yield stress equal to 514 MPa and tensile strength equal to 659 MPa (mean value from three coupons). It had 8 mm plain (smooth) stir-rups positioned at 200 mm intervals, anchored by 135 degree hooks at one end and by 90 degree hooks at the other. The tie yield stress and tensile strength were equal to 425 and 596 MPa, respectively (mean from three coupons). The strength of the concrete of the original columns at the time of testing, measured on 150 by 300 mm cylinders, fc, and the effective depth of the section are listed in Tables 1 to 3.

In one of the specimens, R-0 in Table 1, which was used as an unretrofitted control, the vertical bars of the original column continued into the footing for anchor-age, while in all other columns they started at the top of the footing and were lapped over a 15-, 30- or 45-bar diameter length (specimens designated as L1, L3 or L4, respectively) with straight starter bars coming from the footing (Fig. 2(a)) and anchored with a 90 degree hook at the bottom.

The concrete jacket applied to three of the columns

Direction of testing

250m

m

500 mm

Φ8/200 Φ18

~440 mm

~185

mm

Φ8/200Φ18

Φ18

650 mm

400

mm

250 mm

500 mm

Φ8/200

Φ18

~440 mm

~185

mm

R = 30 mm

(a) (b) (c) Fig. 1 Cross-section of (a) original column, (b) RC jacketed column, and (c) FRP-retrofitted column (dimensions in mm).

S. Bousias, A.-L. Spathis and M. N. Fardis / Journal of Advanced Concrete Technology Vol. 4, No. 3, 431-444, 2006 433

was 75 mm thick and consisted of shotcrete. This thick-ness is the minimum that provides proper cover to the perimeter tie and allows 135 degree hooks at its ends. The jacket was reinforced with six 18 mm vertical de-formed bars with the same properties as the vertical bars of the original column. These bars were embedded in the footing at the time of casting of the original column (Fig. 2(b)). The jacket transverse reinforcement con-

sisted of a single 10 mm perimeter tie at 100 mm cen-ters, with yield stress of 599 MPa. After the addition of the RC jacket, the total cross-sectional dimensions of the specimen became 650 by 400 mm (Fig. 1(b)). The jacket was not applied over the full height of the column, but stopped approx. 250 mm below the point where the lateral load was applied. No special measures were taken for the connection of the jacket to the old concrete,

Table 1 Unretrofitted columns: test parameters and key results.

Specimen Lap length

Effective depth (mm)

Concrete cylindrical strength, fc (MPa)

ν=N/bhfc

Yield moment,

My (kNm)

Drift at yielding, Ry (rad)

Yield curvature, φy (1/m)

Ultimate drift, Ru

(rad)

Ultimate curvature, φu (1/m)

Main features of behavior and failure mode

R-0 - 455 31.0 0.26 298 0.0075 0.010 0.025 0.04 Diagonal tension failure after yield-ing; buckling of all corner bars

R-0L1 15dbL 470 27.4 0.23 225 0.005 0.005 0.019 0.048

Heavy spalling of corners all along lapping; major crack formed at level of 1st tie above base, near end of lap-splice

R-0L3 30dbL 465 27.4 0.28 279 0.0075 0.005 0.019 0.044 Spalling of all four corners all along lapping

R-0L4 45dbL 470 27.4 0.28 283 0.011 0.009 0.025 0.048Spalling of two corners all along lapping; tendency of bars to buckle along lap splice

US-0 (Bousias

et al. 2004)

- 470 18.3 0.38 283 0.0105 0.0125 0.028 0.045 Concrete crushing; bar buckling

(a) (b) Fig. 2 Laps at column base in specimen R-0L1 (a); jacket bars embedded in column base (b).


as Bousias et al. (2006) have found, through a separate experimental campaign and by comparing test results of others in the literature, that in RC jacketed columns hav-ing continuous vertical smooth bars in the original col-umn, positive measures of connection (such as roughen-ing the interface, steel dowels, or connection of the new corner bars to the old ones through welded U-bars) are not essential for the full composite action of the old and the new concrete under cyclic loading. The same con-clusion was reached in a recent comparative monotonic testing program of six jacketed columns with different levels and means of connection at the interface (Julio et al. 2005).

The measured cylindrical concrete strength of the jacket and of the original column at the time of testing and the effective depth of the jacketed section are listed in Table 2.

Six columns were tested after retrofitting with wraps of resin-impregnated practically unidirectional carbon-fiber-reinforced-polymer (CFRP) sheets. Each layer of CFRP had a nominal thickness 0.13 mm, elastic modulus of 230 GPa, and tensile strength of 3450 MPa (failure strain 1.5%) in the (main) direction of the fibers. The CFRP sheets were attached to the concrete surface via epoxy resin, after thorough cleaning of the surface from loose material and rounding of the corners of the section to a 30 mm radius (Fig. 1(c)), to avoid stress concentrations that may have lead to premature FRP rupture and to enhance confinement of the corners of the section. The CFRP wrapping started at a distance of approx. 10 mm from the base and was applied over the lowest 600 mm of the column. A CFRP lap length of one full side of the cross-section was provided at the end of the wrap. Two different numbers of CFRP plies were used: two or five, with specimens denoted as P2 and P5, respectively. The measured cylindrical concrete strength at the time of testing and the effective depth of the section are listed in Table 3.

Unidirectional deflection cycles were applied with amplitude increasing by 5 mm from one cycle to the next. A load history with closely spaced single cycles was chosen in the present study for better identification of possibly small effects of single testing parameters on the deformation capacity of the column, which could possibly be missed if the usual protocol of three cycles at a few fixed displacement ductility levels were applied. In general, testing continues (with increasing deflection amplitude) until and beyond the ultimate column de-formation, following its now established conventional definition as that where peak resistance in a cycle drops below 80% of the maximum recorded lateral resistance of the column (French et al. 1991; Saatcioglou 1991). The conventionally defined ultimate deformation often takes place when one or more vertical bars break, or when the FRP wrapping ruptures.

A jack at the top of the column maintained the axial load at approx. 800 kN. This jack acted against vertical rods connected to the strong laboratory floor through a hinge. With this setup, the P-∆ moment at the base of the column was equal to the axial load times the ratio of the distance of the hinge from the column base to that to the point of application of the lateral load (i.e. times 0.5/1.6=0.3125). Bending moments reported in Tables 1 to 3 include the P-∆ contribution. Tables 1 to 3 give the mean value of the axial load during the test, normalized to the product of the external dimensions b and h of the jacketed section and to the concrete strength. For the RC jacketed columns, the fc value of the jacket is used in the normalization of axial load as ν=N/bhfc, because it is the concrete of the jacket that mainly controls the flexural or shear resistance of the column and its ultimate de-formation, as these are controlled by the extreme com-pression fibers.

Measurements of LVDTs placed at opposite sides of the column in the direction of loading provided the rela-

Table 2 RC jacketed columns: test parameters and key results. Concrete

strength, fc (MPa) Specimen Lap

length

Effective depth (mm) original

column jacket

ν=N/bhfc jacketed column

Yield moment,

My (kNm)

Drift at yielding,Ry (rad)

Yield curvature, φy (1/m)

Ultimate drift, Ru

(rad)


Main features of behavior and of failure mode

R-RCL1 15dbL 600 36.7 55.3 0.066 533 0.01 0.014 0.042 0.04 Heavy bond splitting/spalling all along corner bars; tendency of bars to buckle

R-RCL3 30dbL 605 36.8 55.3 0.066 563 0.01 0.015 0.038 0.125

Significant bond split-ting/spalling all along corner bars; diffused inclined cracks; tendency of bars to buckle

R-RCL4 45dbL 600 36.3 55.3 0.052 522.5 0.01 0.018 0.047 0.15

Partial height bond cracking along one corner bar; lower-most tie opened; buckling of jacket bars and one interior bar (in original column)


Table 3 FRP-jacketed columns: test parameters and key results.

Specimen Lap length

No. of

FRP plies

Effective depth (mm)

Concrete strength, fc

(MPa) ν=N/bhfc

Yield moment,

My (kNm)

Drift at yielding, Ry (rad)

Yield cur-vature, φy

(1/m)

Ultimate drift, Ru

(rad)


Main features of behavior and

failure mode

R-P5 - five 460 32.9 0.23 323.2 0.01 0.00625 0.053 0.075

Over lower 20 mm FRP frac-ture and bar buckling

R-P2 - two 462 32.9 0.23 323.2 0.01 0.085 0.044 0.058

Wide crack at column base; FRP fracture over lower 120 mm; concrete disintegration and bar buck-ling inside

R-P5L1 15dbL five 465 27.0 0.28 284.0 0.0095 0.016 0.05 0.18

Wide crack at column base; spalling at top of footing around bar starting there

R-P5L3 30dbL five 470 27.0 0.29 323.1 0.0095 0.01 θu>0.056 0.1

Wide crack at column base; end of test be-fore reaching ultimate drift

R-P5L4 45dbL five 470 27.0 0.29 322.6 0.008 0.00925 θu>0.056 0.0925

Over lower 20 mm FRP frac-ture and bar buckling; test ended before ultimate drift

R-P2L1 15dbL two 460 26.9 0.30 258.1 0.0065 0.01 0.034 0.062

R-P2L3 30dbL two 470 26.9 0.28 303.2 0.01 0.013 0.047 0.085

Concrete spalling at top of footing around end of bar starting from there

R-P2L4 45dbL two 470 26.2 0.28 303.6 0.0115 0.008 0.056 0.085

FRP fracture over lower 120 mm; concrete disintegration inside and bar buckling; di-agonal cracking above FRP

R-P2L4a 45dbL two 480 26.9 0.265 312.8 0.0095 0.01 - - Problem in footing ended test early

US-C5 (Bousias et al. 2004)

- five 470 17.9 0.39 324.0 0.011 0.03 θu>0.069 ϕu>0.22

Ultimate de-formation not reached, but by end of test the FRP bulged at column base

US-C2 (Bousias et al. 2004)

- two 480 18.1 0.37 323.0 0.0105 0.028 0.056 0.165 FRP fracture; bar buckling


tive rotation and the mean axial strain at two levels of the column above the base, at 250 mm and 500 mm (i.e. at 50% or 100% of the section depth, h, of the original column in the testing direction). LVDT measurements include the column section at the face of the footing and hence are affected by vertical bar pull-out from the base.

3. Experimental results and discussion

The five columns in Tables 1 to 3 before the last one give the moment, the drift ratio (chord rotation) and the curvature at the base at yielding of the column, My, Ry and φy, respectively, and the ultimate drift ratio (chord rotation) and ultimate curvature at the base, Ru and φu, respectively. For the purposes of these tables, yielding is identified with a distinct reduction in the slope of the force-drift curve and is close to the corner point of a bilinear envelope that can be fitted to the ascending part of the force-drift response, i.e. up to peak resistance. Admittedly, this definition may lead to subjective quan-tification of yielding; however, as the experimental force-drift results are given in the figures, the reader may choose to identify yielding in the experiments adopting alternative, e.g. mechanistic, criteria.

The value of the curvature listed is the one obtained

from the rotation of the upper of the two instrumented sections relative to the base, and as such, it is less af-fected by bar pull-out from the footing. Ultimate col-umn deformation is conventionally identified with a post-ultimate strength drop of the lateral force resistance by more than 20% with respect to the peak resistance (ultimate strength) of the column.

3.1 Unretrofitted columns Force-displacement loops of the unretrofitted specimens are shown in Fig. 3. Specimen R-0 with continuous lon-gitudinal reinforcement (bottom right in Fig. 3), served as the unretrofitted control column of the group of specimens with lap splicing. It yielded in flexure but exhibited a rapid drop in resistance with cyclic loading after ultimate strength, accompanied by inclined crack-ing—suggesting effects of shear—and leading to bar buckling and disintegration of the concrete core above the base. The ultimate drift was low (2.5%). The initial stiffness of columns with lap splices was similar to that of the control with continuous reinforcement, R-0. The effect of lap length on the force-deflection loops became clear at and after yielding. Specimen R-0L1, with the 15 bar diameter splices, had the lowest strength and exhib-ited early spalling at the corners and loss of bond. The

-250

-200

-150

-100

-50

0

50

100

150

200

250

Forc

e (k

N)

-6 -4 -2 0 2 4 6-250

-200

-150

-100

-50

0

50

100

150

200

250

Drift (%)

Forc

e (k

N)

-4 -2 0 2 4 6Drift (%)

R-0L1 R-0L3

R-0L4 R-0

Fig. 3 Effect of lap length in unretrofitted columns (dashed line: force at yielding according to calculation).


maximum moment reached by columns R-0L1 and R-0L3 was approx. 80% and 95%, respectively, of the theoretical yield moment of the column with continuous bars, whilst column R-0L4, as well as the control, R-0, reached a moment capacity almost 110% of the theoreti-cal yield moment. Thus, lap length affected peak resis-tance, which dropped by 30% in R-0L1 with the 15 bar diameter lap splice, or by more than 10% in R-0L3 with the 30 bar diameter lap splice, compared to R-0 or R-0L4. (Figure 4 shows also the force corresponding to the theoretical yield moment of the column with continuous bars, after its reduction due to P-∆).

The width of hysteresis loops and energy dissipation decreased with decreasing lapping length. Although column R-0L4 with the 45-bar diameter lapping had similar strength and deformation capacity as the control column R-0, its hysteresis loops exhibited to a certain degree the effect of lap splicing, albeit much less than in columns R-0L1 and R-0L3.

Damage in specimens R-0L1 and R-0L3 appeared first as splitting of the cover all along the lapping and progressed with crushing of concrete just ahead of the

starter bars, due to bearing stresses (Fig. 4(a), (b)). In the absence of dense stirrups, concrete crushing ahead of a starter bar was promoted by the location of the starter bars at the corner of the ties (due to the sequence of construction) and hence closer to the surface. Al-though in column R-0L4, with the 15-bar diameter lap splice, splitting cracks appeared also along the lap length at two corners (Fig. 4(c)), the strength and de-formation capacity was not conditioned by the splice.

Table 1 summarizes also (at the bottom line) the re-sults for a column tested by Bousias et al. (2004), which was similar to the control column R-0 in every respect other than the concrete strength, fc, which was lower (hence the normalized axial load was proportionally higher) and the effective depth of the section, which was closer to that of the unretrofitted columns with lap splic-ing.

3.2 RC jacketed columns The force-deformation loops of the three RC jacketed columns with lap splices are shown in Fig. 5. All jack-eted columns exceeded the theoretical flexural capacity

(a) (b) (c)

Fig. 4 Unretrofitted columns with lap-spliced bars after failure: (a) R-0L1; (b) R-0L3; (c) R-0L4.

-6 -4 -2 0 2 4 6-400

-300

-200

-100

0

100

200

300

400

Drift (%)

Forc

e (k

N)

-4 -2 0 2 4 6Drift (%)

-4 -2 0 2 4 6Drift (%)

R-RCL1 R-RCL3 R-RCL4

Fig. 5 Effect of lap length of bars in original column on load-deflection loops of RC jacketed columns (dashed line: force at yielding according to calculation).


of a similar monolithic column with the lap-spliced bars considered as continuous and with the strength of the jacket concrete taken to apply over the entire section (witness in Fig. 5 the force corresponding to this theo-retical flexural capacity, after its reduction due to P-∆). Thus, RC jackets seem to be quite effective regarding flexural resistance and mobilization of the strength of insufficiently spliced bars. Strength was roughly equal in all three columns. The jacket concrete cracked and spalled along corner bars due to bond distress (see Fig. 6). The extent and severity of bond splitting/spalling decreased with lap length increases. It is noted in this respect that, due to the small jacket thickness (75 mm) and the space needed for the 135 degree hook of ties, cover to the vertical bars was, on average, only 35-40 mm and cover to the stirrup only 25-30 mm.

Columns R-RCL1 (15-bar diameter lapping) and R-RCL3 (30-bar diameter lapping) yielded in flexure, reaching peak resistance at 1.5% drift (2.5 times that of the unretrofitted column R-0L1 and twice that of R-0L3). Some inclined cracking developed near peak re-sistance, but diagonal cracks did not open up with dis-placement cycling thereafter. These two columns ulti-mately failed (see 3.1 for definition of failure) at a drift of approx. 4.2%. Their rather narrow force-deflection loops in Fig. 5 suggest bond- or shear-dominated cyclic behavior after ultimate strength. (The shear span ratio of the jacketed columns was equal to 2.46, and shear may have affected cyclic behavior; moreover, these two jacketed columns had extensive bond splitting/cracking along corner bars). The force-deformation response of column R-RCL4 with the 45-bar diameter lapping (see Fig. 5, right) did not differ much from that of the other two retrofitted columns. It yielded in flexure and, owing to its longer lap length, it maintained its peak resistance over more cycles. Diagonal cracking developed in this column as well, but there was little bond cracking along corner bars of the jacket and damage concentrated near

the base. Crushing of the concrete there extended over the full width of the jacket, accompanied by disintegra-tion of the concrete in the old column inside and buck-ling of some of its bars as well (Fig. 6(c)). Hysteresis loops were appreciably wider than in the retrofitted col-umns with the shorter lapping.

Overall, the hysteresis loops suggest that, although the adverse effects of short lap splices in the original column are very much reduced in the RC jacketed col-umn, they do not disappear completely.

Figure 7 compares the envelope curves of the four unretrofitted and the three RC jacketed columns. The improvement effected by the RC jacket not only in strength and stiffness, but in deformation capacity as well, is evident.

3.3 FRP retrofitted columns Four specimens similar to the unretrofitted ones (i.e. specimens without lap splices and three with 15-, 30- and 45-bar diameter splicing) were tested after wrap-ping with five layers of CFRP. The force-deflection loops of all four columns are shown in the middle col-umn of Fig. 8 (Figs. 8 (b), (e), (h) and (k)). In column R-P5L1 (with 15-bar diameter lapping) FRP wrapping restored member strength to above 90% of that of the control specimen with continuous bars in Fig. 8(a), while pre-yield stiffness remained unaffected. In the two other columns with lap splices, strength increased well above that of the unretrofitted column. The deformation capacity of all three retrofitted columns with lap splices was much higher than that of the unretrofitted column without lap splices. Notably, no crushing of concrete ahead of the starter bars occurred in the three retrofitted columns, even in specimen R-P5L4, where the length of application of CFRP (0.6 m) was shorter than the lap length (0.81 m) and did not reach the end of the starter bar: improvement in bond conditions along the 0.6 m-long confined part of the lap splice reduced the force to

(a) (b) (c) Fig 6 RC jacketed columns with lap-spliced vertical bars in original column after failure: (a) R-RCL1; (b) R-RCL3; and (c) R-RCL4.


be transferred by direct bearing at the head of the starter bar and, with it, the possibility of local crushing at that location.

Regardless of lap-splice length, all columns tested af-ter retrofitting sustained drift ratios of at least 3.4% at conventionally defined failure (see Table 3). The im-provement in performance compared to the unretrofitted columns is due to the confinement of concrete, espe-cially in the splice region. Despite the dilatation of con-crete inside the FRP jacket, the jacket itself did not rup-ture in any specimen of the 5-ply group.

Figures 8(c), (f), (i) and (l) show test results from columns with two layers of CFRP, without lapping (R-P2) and with short to medium lap length (R-P2L1, R-P2L3, R-P2L4). The retrofitting effect of two layers (the practical minimum) is less than that of five layers. How-ever, the reduction is less than proportional to that of the number of layers. For the specimen with 30-bar diame-ter laps, which shows very satisfactory deformation ca-pacity and energy dissipation when wrapped with five CFRP layers, wrapping with just two FRP layers (R-P2L3) is clearly insufficient. Another notable observa-tion is that, no matter the number of layers, wrapping with FRP cannot fully remove the deficiency of very short lapping of straight bar ends: although columns R-P5L1 and R-P2L1 had larger deformation capacity than the unretrofitted column without lap splices, they had lower force resistance and low energy dissipation—as evidenced by their narrow loops. It seems that the ad-verse effects of a lap length of at least 30-bar diameter can be almost fully removed by FRP wrapping, but, by contrast, if the lapping is as short as 15-bar diameter, some of its adverse effects cannot be removed by FRP wrapping.

The last two lines of Table 3 summarize the results for two columns tested by Bousias et al. (2004), which

were similar to columns R-P5, R-P2 in every respect other than the concrete strength, fc, which was lower (hence the normalized axial load was proportionally higher) and the effective depth of the section, which was closer to that of the FRP wrapped columns with lap splicing.

3.4 Mean axial strain at bottom of columns The evolution with deflection of the top of the column of the mean (over the bottom 500 mm from the base) axial strain at the center of the section is presented in Fig. 9 for the unretrofitted columns, in Fig. 10 for the RC-jacketed ones and in Fig. 11 for the FRP wrapped columns. These strains are derived from the average of the LVDT measurements on opposite sides of the col-umn. In the RC jacketed specimens, they essentially reflect the axial displacement of the lowermost 500 mm of the jacket with respect to the footing. The change in column length in each cycle of loading developing approx. in proportion to the lateral deflection is due to bending according to the plane-sections hypothesis.

The most interesting result in Figs. 9 to 11 is the ratcheting axial shortening or extension with the cyclic deflection. The four unretrofitted columns in Fig. 9 ex-hibit moderate ratcheting axial shortening, as cyclic loading leads them beyond yielding and towards failure, due to accumulation of permanent compressive strains in the concrete. The large ratcheting axial elongation of the three RC jacketed columns in Fig. 10, is partly due to their very low normalized axial load (owing to the very high compressive strength of their jacket shotcrete), but mainly due to the lack of any positive measures to connect the old concrete to the jacket. In a separate ex-perimental campaign on jacketed columns with continu-ous vertical smooth bars in the original column, Bousias et al. (2006) have found that ratcheting axial elongation

-6 -4 -2 0 2 4 6-400

-300

-200

-100

0

100

200

300

400

Drift (%)

Forc

e (k

N)

R-0L0R-0L1R-0L3R-0L4R-RCL1R-RCL3R-RCL4

Fig. 7 Comparison of envelope curves of unretrofitted and of RC jacketed columns.


-200-150-100-50

050

10050

200250

Forc

e (k

N)

-200-150-100-50

0 5010050

200250

Forc

e (k

N)

-200-150-100-50

0 5010050

200250

Forc

e (k

N)

-6 -4 -2 0 2 4 6-250-200-150-100-50

0 5010050

200250

Drift (%)

Forc

e (k

N)

-4 -2 0 2 4 6Drift (%)

-4 -2 0 2 4 6Drift (%)

(a) R-0L0 (b) R-P5L0 (c) R-P2L0

(d) R-0L1 (e) R-P5L1 (f) R-P2L1

(g) R-0L3 (h) R-P5L3 (i) R-P2L3

(j) R-0L4 (k) R-P5L4 (l) R-P2L4

Fig. 8 Effect of lap length of vertical bars in original column on load-deflection loops of FRP wrapped columns - Top row: columns with continuous bars; 2nd row: 15-bar diameter laps; 3rd row: 30-bar diameter laps; Bottom row: 45-bar diameter laps – Left column: unretrofitted specimens; Center: 5-ply FRP wraps; Right column: 2-ply FRP wraps.


of the type shown in Fig. 10 is significantly reduced, if roughening of the surface of the old column, or U-bars welded to its corner bars are used to connect the jacket

to the old column, while dowels alone at the interface are only partly effective in this respect. In that same investigation, no significant effect on the global cyclic

0

0.01

0.02

0.03

Drift (%)

Mea

n ax

ial s

train

-5 -4 -3 -2 -1 0 1 2 3 4 5-0.01

0

0.01

0.02

0.03

Drift (%)

Mea

n ax

ial s

train

-4 -3 -2 -1 0 1 2 3 4 5Drift (%)

R-0 R-0L1

R-0L3 R-0L4

Fig. 9 Evolution of axial strain at center of section (mean value over lower 500 mm from base) with lateral displacement of unretrofitted columns.

-5 -4 -3 -2 -1 0 1 2 3 4 5

0

0.01

0.02

0.03

Drift (%)

Mea

n ax

ial s

train

-4 -3 -2 -1 0 1 2 3 4 5Drift (%)

-4 -3 -2 -1 0 1 2 3 4 5Drift (%)

R-RCL1 R-RCL3 R-RCL4

Fig. 10 Evolution of axial strain at center of section (mean value over lower 500 mm from base) with lateral displacement of RC jacketed columns.


behavior of the columns could be attributed to the appli-cation or non-application of the above-mentioned posi-tive measures of connection between the jacket and the old column. Moreover, the analysis of a large database of test results on jacketed members in the literature in Bousias et al. (2006) has shown that, if there are no positive measures of connection at the interface with the old column, or if they are limited to dowels or welded U-bars, only the secant stiffness at incipient yielding is reduced (on average by 15%) compared to than of jack-eted columns with a roughened interface.

The FRP-retrofitted columns in Fig. 11 show very stable axial behaviour, with neither ratcheting net short-ening nor net extension to the end of the test, except column R-P2L1, which exhibits ratcheting shortening towards its premature failure. This is consistent with the lower My and ultimate drift (Ru) values of this column in Table 3.

4. Conclusions

The paper presents cyclic test results on rectangular columns with corner deformed bars lap-spliced in the plastic hinge region, including retrofitting through RC jacketing or CFRP wrapping. The conclusions apply only to rectangular or square columns with reinforce-ment amount and detailing fairly similar to that of the

present specimens and for low to moderate axial load ratio, ν=N/bhfc.

It has been confirmed that old-type columns with de-formed bars lap-spliced in the plastic hinge region have significantly reduced cyclic deformation capacity and energy dissipation. It was found, however, that in the case of lapping of at least a 45-bar diameter, energy dissipation is acceptable and the cyclic deformation capacity is not significantly reduced compared to a col-umn with continuous bars. Lapping of straight bar ends by as little as 15-bar diameter reduces appreciably flex-ural resistance and gives rapid post-peak strength and stiffness degradation, as well as low energy dissipation capacity.

Concrete jackets were found to be quite effective for remedying the adverse effect of lap-splicing straight deformed bars on strength and deformation capacity, even for very short lap lengths. Nonetheless, some ad-verse effect on hysteretic energy dissipation due to very short lapping in the original column is retained in the jacketed column. Large ratcheting axial elongations at the base of the RC jacketed columns are attributed mainly to the lack of positive measures to connect the old concrete and the jacket: in separate tests of RC jack-eted columns with continuous vertical smooth bars in the original column, Bousias et al. (2006) found much less ratcheting when the surface of the old column was

-0.01

0

0.01

0.02

0.03

Mea

n ax

ial s

train

-5 -4 -3 -2 -1 0 1 2 3 4 5-0.01

0

0.01

0.02

0.03

Drift (%)

Mea

n ax

ial s

train

-4 -3 -2 -1 0 1 2 3 4 5Drift (%)

-4 -3 -2 -1 0 1 2 3 4 5Drift (%)

R-P2 R-P2L1 R-P2L3

R-P5 R-P5L1 R-P5L3

Fig. 11 Evolution of axial strain at center of section (mean value over lower 500 mm from base) with lateral displacement of FRP wrapped columns.


roughened, or when the connection of the jacket to the old column was improved by welding U-bars between the corner bars of the jacket and the old column (but not so if only dowels at the interface were used). However, Bousias et al. (2006) found no significant effect on the global cyclic behavior of the columns due to such posi-tive connection measures (other than a small reduction in the secant stiffness at incipient yielding, identified from a large database of tests on jacketed members in the literature).

Regarding FRP retrofitting, the present investigation was limited to columns with only corner bars, which are very effectively confined by the FRP wrapping. As the bond along lap-spliced intermediate bars between the corners of the section would not benefit at all from FRP wrapping, the conclusions of the present study cannot be extrapolated to bigger columns with several interme-diate lap-spliced bars. The improvement due to FRP wrapping was found to be less than proportional to the number of CFRP layers. Despite the significant positive effect of FRP wrapping on flexural resistance, ultimate deformation and energy dissipation, all these perform-ance parameters decrease with a reduction in lap length, even in the FRP-retrofitted column. It seems that there is a limit to the improvement that FRP wrapping may bring about in columns with lap-spliced straight de-formed bars: if the lap length is as short as 15-bar-diameters, its adverse effects on force capacity and en-ergy dissipation cannot be sufficiently removed by FRP wrapping. Overall, for columns with just 15-bar-diameter laps, concrete jacketing seems to improve more the deformation capacity than FRP wrapping. In view of the reduction in deformation demands that would result from the enhanced stiffness of the RC-jacketed column, concrete jackets seem to represent a better option for rehabilitation of columns with lap length as short as 15-bar-diameters. This would be even more so for big columns having several intermediate bars with short lap-splicing (cf. the word of caution at the beginning of this paragraph).

Acknowledgments This study was funded in part by the European Com-mission (research project SPEAR of the GROWTH pro-gram, contract no. G6RD-2001-00525) and, in part by NATO (via its Science for Peace Program). References Alcocer, S. M. (1992). “Rehabilitation of RC frame

connections using jacketing.” 10th World Conference of Earthquake Engineering, Acapulco, Mexico, 5235-5240.

Alcocer, S. M. and Jirsa, J. O. (1993). “Strength of reinforced concrete frame connections rehabilitated by jacketing.” ACI Structural J., 90 (3), 249-261.

Bett, J., Klingner, R. and Jirsa, J. (1988). “Lateral load response of strengthened and repaired reinforced concrete columns.” ACI Structural J., 85 (5), 499-507.

Bousias, S. N, Triantafillou, T. C., Fardis, M. N., Spathis, L.-A. and O’Regan, B. (2004). “FRP retrofitting of rectangular RC columns with or without corrosion.” ACI Structural J., 101 (4), 512-520.

Bousias, S. N., Biskinis, D., Fardis, M. N. and Spathis, A.-L. (2006). “Strength, stiffness and cyclic deformation capacity of concrete jacketed columns.” ACI Structural J. (submitted).

Chang, K. C., Lin, K. Y. and Cheng, S. B. (2001). “Seismic retrofit study of RC rectangular bridge columns lap-spliced at the plastic hinge zone.” International Conference of FRP composites in Civil Engineering, Hong Kong, 869-875.

Choudhuri, D., Mander, J. B. and Reinhorn, A. M. (1992). “Evaluation of seismic retrofit of reinforced concrete frame structures: Part I Experimental performance of retrofitted subassemblages.” Report NCEER-92-0030, National Center for Earthquake Engineering Research, State University of New York, Buffalo, NY.

Elnabelsy, G. and Saatcioglu, M. (2004). “Seismic Retrofit of Circular and Square Bridge Columns with CFRP Jackets.” 4th International Conference on Advanced Composite Materials in Bridges and Structures, Calgary, Canada.

Ersoy, U., Tankut, T. and Suleiman, R. (1993). “Behavior of jacketed columns.” ACI Structural J., 90 (3), 288-293.

French, C. W. and Schultz, A. E. (1991). “Minimum Available Deformation Capacity of Reinforced Concrete Beams.” S. K. Ghosh, Ed., SP127 “Earthquake Resistant Concrete Structures. Inelastic Response and Design.” American Concrete Institute, Detroit, 363-410.

Gomes, A. M. and Appleton, J. (1998). “Repair and strengthening of reinforced concrete elements under cyclic loading.” 11th European Conference of Earthquake Engineering, Paris.

Harries, K. A., Ricles, J. R., Pessiki, S. and Sause, R. (2003). “Rehabilitation of lap-splices in non-ductile reinforced concrete columns using CFRP jackets.” M. Forde, Ed. 10th International Conference Structural Faults and Repair, Eng. Technic Press.

Haroun, M. A., Mosallam, A. S., Feng, M. Q. and Elsanadedy, H. M. (2001). “Experimental investigation of seismic repair and retrofit of bridge columns by composite jackets.” International Conference of FRP composites in Civil Engineering, Hong Kong, 839-848.

Ilki A., Darilmaz, K., Bakan, I., Zorbozan, M., Yuksel, E., Saruhan, H. and Karadogan, F. (1998). “Jacketing of prefabricated columns.” 2nd Japan-Turkey Workshop on Earthquake Engineering, Istanbul, 329-336.

Iliya, R. and Bertero, V. V. (1980). “Effects of amount and arrangement on wall-panel reinforcement on hysteretic behavior of reinforced concrete walls.”


Report No. UCB/EERC-80/04, University of California, Berkeley, CA.

Julio, E. N. B. S., Branco, F. A. B. and Silva, V. D. (2005). “Reinforced concrete jacketing-Interface influence on monotonic loading response.” ACI Structural J., 102 (2), 252-257.

Rodriguez, M. and Park, R. (1994). “Seismic load tests on reinforced concrete columns strengthened by jacketing.” ACI Structural J., 91 (2), 150-159.

Saadatmanesh, H., Ehsani, M. R. and Jin, L. (1997a). “Repair of earthquake-damaged RC columns with FRP wraps. ” ACI Structural J., 94 (2), 206-215.

Saadatmanesh, H., Ehsani, M. R., M. and Jin, L.

(1997b). “Seismic Retrofitting of Rectangular Bridge Columns with Composite Straps.” Earthquake Spectra, 13 (2). 281-304.

Saatcioglu, M. (1991). “Deformability of Reinforced Concrete Columns.” S. K. Ghosh, Ed., SP-127 “Earthquake Resistant Concrete Structures. Inelastic Response and Design.” ACI, Detroit, Mich., pp. 421-452.

Stoppenhagen, D. R., Jirsa, J. O. and Wyllie, L. A. (1995). “Seismic repair and strengthening of a severely damaged concrete frame.” ACI Structural J., 92 (2), 177-187.

Documents

Concrete or FRP Jacketing of Columns with Lap Splices for