Effect of FRP wrapping in seismic performance of RC buildings with and without special detailing -A case study

Composites: Part B 45 (2013) 12651274

Contents lists available at SciVerse ScienceDirect

Composites: Part B
journal homepage: www.elsevier.com/locate/compositesb

Effect of FRP wrapping in seismic performance of RC buildings with and without special detailing A case study
A. Eslami, H.R. Ronagh
School of Civil Engineering, The University of Queensland, Brisbane, QLD 4072, Australia

a r t i c l e

i n f o

a b s t r a c t
The results of a numerical investigation into the efciency of glass bre reinforced polymers (GFRPs) in improving the seismic performance of an 8-storey moment resisting reinforced concrete building are presented. In order to assess the effect of the transverse reinforcement, the building is detailed with different levels of transverse reinforcement representing well-conned and poorly-conned conditions. Although GFRP wrapping of columns at critical regions is the main retrotting technique considered in this study, the effect of increasing the beam ductility on the seismic performance of a structure is also evaluated for the code-compliant building. The retrotting strategy aims to provide both columns and beams with more ductility and energy dissipation instead of increasing the lateral strength. The loaddisplacement curves obtained from pushover analysis of the frames are then used in the seismic assessment using a capacity spectrum approach (the N2 method). The results conrm that GFRP wraps is capable of improving the seismic performance and ductility of the poorly-conned structure substantially, compared to the original structure. However, it was found that using FRP composites in order to increase the ductility of code-compliant building only, was not that effective. 2012 Elsevier Ltd. All rights reserved.

Article history: Received 22 April 2012 Received in revised form 14 September 2012 Accepted 14 September 2012 Available online 23 September 2012 Keywords: A. Glass bres B. Plastic deformation B. Strength C. Analytical modelling Pushover analysis

1. Introduction After the 1994 Northridge and the 1995 Kobe earthquakes, many questions arose concerning the distinct characteristics of earthquakes in the areas near the fault rupture. Failure of modern engineered structures during these earthquakes revealed the inadequacy of current design methods and code regulations for nearfault pulse-type ground motions. Despite several studies on the phenomenon, researchers could not agree on an appropriate and accurate, yet simple design method, for the design of buildings located in the proximity of fault rupture. The scarcity of reliable data is considered one of the major contributors to this issue. In addition to near-fault earthquakes, changes in seismic hazard levels, design methods, and serviceability requirements, are amongst other reasons for retrotting a code-compliant structure subjected to an ordinary earthquake. Thus, many existing buildings have yet to be retrotted in order to remain reasonably intact and safe during pulse-type ground motions or more severe earthquakes than those they have been designed for. During the last couple of decades, application of composite materials for retrotting/upgrading of reinforced concrete (RC) buildings has experienced a sharp increase. Compared to other building materials, FRPs offer several advantages; such as, possessing high tensile
Corresponding author. Tel.: +61 7 3365 9117; fax: +61 7 3365 4599.
E-mail address: h.ronagh@uq.edu.au (H.R. Ronagh). 1359-8368/$ - see front matter 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.compositesb.2012.09.031

strength, low specic weight, high resistance to corrosion, and ease of application. These advantages make FRP a more suitable candidate for retrotting RC buildings in comparison to other traditional methods, such as bracing and shear walls. Seismic strengthening/repairing of structural elements with FRP materials has been studied by many researchers worldwide. Numerous past studied have been carried out on the application of FRP in strengthening of beams [13], columns [48], and beam to column connections [913]. Researchers have also investigated related problems such as de-bonding failure and interfacial interaction of FRP laminates and concrete substrate [1416]. Ascione et al. [17] have also utilised GFRP bars as internal reinforcement in concrete beams. Comparing the experimental results and theoretical predictions, both at service and ultimate conditions, their large experimental program was conducted to provide a contribution to the design procedure of FRP reinforced beams. Despite the large volume of data available on FRP strengthening of RC components (beams, columns, and beamcolumn joints), only a handful of researchers have investigated the overall behaviour of FRP-rehabilitated RC structures. In an experimental study, Balsamo et al. [18] assessed the seismic performance of a full-scale RC structure repaired using carbon bre reinforced (CFRP) laminates and wraps. Their results proved that a large displacement capacity exists in the repaired structure while no reduction of strength is seen after the application of FRP at beamcolumn joints and walls. In addition, the energy dissipation remains almost identical to the original structure. On the contrary, a reduction in

1266

A. Eslami, H.R. Ronagh / Composites: Part B 45 (2013) 12651274

the deformability of shear walls was observed during the experimental activity due to the presence of CFRP laminates over the entire height. In another experimental study, Di Ludovico et al. [19] investigated seismic retrotting of an under-designed, full-scale RC structure with FRP wrapping. In their study, a bi-directional test with peak ground acceleration (PGA) equal to 0.2 g was performed for which the structure was found decient and after which the gravity-designed structure was retrotted to withstand 0.3 g PGA. Their research conrmed the effectiveness of FRP in conning effects to improving the global performance of a structure in terms of ductility and energy dissipating capacity. Improving the seismic behaviour of decient RC structures with FRP composites has also been approved by Garcia et al. [20] as well, through experimental tests and numerical modelling. They found that an FRP retrot resulted in substantial improvement of seismic performance of the investigated damaged RC frames. Recently, Niroomandi et al. [21] investigated the seismic performance of an ordinary RC frame. Their pushover analysis showed that relocating plastic hinges away from the column faces through web-bonded FRP retrotting of joints in an 8-storey frame increased the lateral load carrying capacity and seismic behaviour factor by 40% and 100%, respectively. However, it is worth mentioning that web-bonded technique comes with certain limitations in practical applications. This paper aims to evaluate the efciency of FRP composites applied at the critical regions of RC members used to enhance the seismic performance of RC structures. The method is to increase the ductility of plastic hinges at beam and column ends without varying their strength levels. Technically speaking, in the latter case, exural strengthening of columns and beams with proper anchorage of composite sheets are necessary. An 8-storey RC building is used as the case study structure. The effects of torsion have been neglected in this study and two-dimensional (2-D) frames analysed as being representative of regular RC buildings. The frame was detailed based on two different reinforcement spacings resulting in two RC frames with different levels of transverse steel reinforcement. Hereafter, these two structures are called intermediate and poorly-conned frames. The nonlinear static (pushover) analysis was performed in order to estimate the seismic response of the structures. In addition, the concept of lumped plasticity was used in the characterisation of nonlinear properties of members. A nite element analysis program, SAP 2000 [22], commonly used by structural engineering professionals, was utilised to run the nonlinear static analysis. The N2 method [23] was employed in order to evaluate the seismic demand and capacity of the retrotted and original frames. 2. Research signicance Despite the vast amount of research carried out on improving the load carrying capacity and ductility of RC members using composite materials, a few studies have focused on the behaviour of FRP retrotted RC structures in particular the structures that are code-compliant. Based on the literature reviewed in this study, almost all researchers have focused on the rehabilitation of decient RC structures built according to the provisions of older codes. However, many modern engineered buildings are supposed to be designed for the higher seismic demands suggested by the more recent codes. The results of this paper provides a basis for further study on different retrotting congurations in order to increase strength and/or ductility of RC structures built according to the provisions of current building codes. 3. Design and description of the frames The structure considered was an 8-storey moment resisting RC frame representing a mid-rise building. The frame considered to be

part of the lateral resisting system of a residential building with three bays (each equal to 5 m). The height was assumed to be equal to 3 m for all stories. The seismic loads were considered according to the provisions of the Iranian seismic code [24]. In the design of the moment resisting frame, the design dead load and live load were assumed to be equal to 30 kN/m and 10 kN/m, respectively which were applied to the beams in addition to the self-weight of the structure. In addition, the compressive strength of concrete was taken as 25 MPa and deformed bars of Grade 60 (fy = 420 MPa) were considered as steel reinforcement. Design base shear was determined considering a peak ground acceleration of 0.3 g representing a high seismic hazard and soil type-III which is similar to class D of FEMA-356 [25]. In order to investigate the connement effect of steel, the frame was reinforced at two different transverse reinforcement levels creating two distinct frames that are called intermediate and poorly-conned as mentioned previously. The former was detailed based on the intermediate provisions of the ACI 318-02 [26], whereas in the latter only the shear design was considered for transverse reinforcing. The poorly-conned frame represents the situations in which the current code provisions are not satised. The longitudinal reinforcements in both frames were similar as they were designed for similar gravity and seismic load levels. For both frames, deformed steel bar, 10 mm in diameter, was selected as transverse reinforcement. Analysis and design of the selected RC frame was carried out using SAP 2000 [22]. The fundamental period of the structure was calculated to be around 1.28 s. As shown in Fig. 1, for a typical beam and column section, the column longitudinal reinforcement was distributed around the section, while the beam longitudinal bars were positioned at the top and bottom of the section in all frames. Fig. 2 provides a schematic illustration along with the dimensions and exural/ transverse reinforcement of the members in both considered frames.

4. Nonlinear analysis of the frames Increasing evidence from past earthquakes has proven the inelastic behaviour of structures under seismic loading. Recourse to nonlinear analysis is unavoidable when it comes to inelastic analysis of a structure subjected to ground shaking. These nonlinear procedures consist of nonlinear static or pushover analysis and nonlinear dynamic or time history analysis. Pushover analysis of a structure includes increasing monotonic lateral loading up to a predened lateral displacement or the onset of plastic collapse mechanism. It is well accepted that nonlinear time history analysis results in a more accurate prediction of the seismic response of a structure subjected to elaborate modelling, precise denition of nonlinear properties of components and selection of a suitable set of ground motion records. However, some past studies have suggested that

b d d

b
As

A' s

d'

Typical column section

Typical beam section

Fig. 1. Distribution of longitudinal reinforcement in a typical beam and column section.

d'

A. Eslami, H.R. Ronagh / Composites: Part B 45 (2013) 12651274

1267

CL

F F C C F F E E E E D D D D D D D D C C

F F F F E E E E D D D D D D D D

8@3m

3@5m

600

Fig. 2. Reinforcement details of the 8-storey intermediate frame and poorly-conned frame.

within the boundary of its limitation, pushover analysis could provide a good estimate of seismic response [2729] in lieu of a time history analysis. The simplicity of pushover analysis compared to inelastic dynamic analysis and the implementation of performance based concepts in the design and rehabilitation of structures have given pushover analysis a signicant boost in recent years which is well deserved. In this section, pushover analysis of unreinforced original frames is presented. The plastic behaviour of each beam and column member has been considered using the lumped plasticity approach. After verication of the adopted assumptions and nonlinear analysis results, the nonlinear static analysis of frames was carried out in SAP 2000 [22] and a characteristic force displacement curve obtained for each frame. Generally, any reference force and reference displacement can be chosen. However, in this study, the base shear and roof displacement were used as the references for force and displacement, respectively.

4.1. Lumped plasticity modelling of the frames Nonlinear analysis of RC buildings using concentrated plastic hinges requires the determination of plastic hinge properties in order to represent the inelastic behaviour of the structure. For every degree of freedom, a rigid plastic hinge could be dened. In a comprehensive nonlinear analysis, consideration of the shear failure in members is necessary. However, some past studies [19,30,31] have reported that even for under-designed RC buildings, detailed with inadequate transverse reinforcement, the shear demand is signicantly lower than the shear capacity in both beams and columns and that no shear failure is observed. Therefore, only the exural inelastic behaviour of the elements was considered in this study using lumped plasticity at both ends. As shown in Fig. 3, a simplied bilinear momentrotation curve was used for each plastic hinge. This approach has also been implemented by Park and Paulay [32] for the nonlinear analysis of RC structures.

1268

A. Eslami, H.R. Ronagh / Composites: Part B 45 (2013) 12651274

The momentrotation properties of the plastic hinges were determined using section analysis. For this purpose, XTRACT software [33] was employed to calculate the momentcurvature relationship based on bre analysis of the end sections in the beams and columns considering section properties, reinforcement details and a constant axial load. Axial loads on the columns were assumed to be equal to the resultant loads calculated from dead loads plus 20% of live loads (as recommended in the seismic design code of the selected structures [24]). On the beams, the axial forces due to gravity loads were assumed to be equal to zero. The commonly used conned concrete model proposed by Mander et al. [34], was implemented and an elastic perfectly plastic model with parabolic strain hardening was considered for steel. The properties recommended in ASTM [35] were used for the steel reinforcement. The design properties of steel reinforcement and concrete were also used during the nonlinear analysis. In Fig. 3, point B corresponds to the tensile yield strain in steel indicating yield moment, My, and yield rotation, hy, while point C belongs to the ultimate conditions; namely ultimate moment, Mu, and corresponding ultimate rotation, hu. The ultimate condition is considered to be the attainment of one of the following conditions; whichever happened earlier [30]. (1) A 20% drop in the moment capacity of member. (2) When the tensile strain in the longitudinal steel reaches the ultimate tensile strain. (3) The attainment of the ultimate compression strain in concrete using the following equation proposed by Scott et al. [36].

should be noted that according to Paulay and Priestley [37], Eq. (2) results in accurate values for conventional beam and column sections. The axial-moment and exural moment hinges were introduced at the end of columns and beams, respectively and the calculated nonlinear properties based on the section analysis were then imported to each hinge. For the columns, the yield moment changes according to the axial load. Thus, a yield moment-axial load interaction curve needs to be dened for each column. During the modelling of the structure, rigid plastic hinges were positioned taking into consideration the beam/column dimensions and the plastic hinge length. In this way, rigid plastic hinges were modelled at a distance of 0.5lP from the beam to column joint faces so that the plastic hinge length spreads equally to the sides of the position considered for the concentrated plastic hinges [30]. 4.2. Verication of the modelling assumptions In order to validate the aforementioned assumptions for the quantication of the plastic hinge properties and the pushover results, two 2-D reinforced concrete frames which were studied by Filiatrault et al. [39,40] were selected. Filiatrault et al. studied the behaviour of these frames using shaker table and numerical analyses using a computer program called RUAUMOKO [41] which was capable of performing nonlinear dynamic and static analysis using the concept of spread plasticity. The two test structures were designed, at their reduced scale, according to the provisions of the National Building Code of Canada [42] and the Canadian concrete standard [43]. Each structure was assumed to be part of the lateral load resisting system of a building, with two stories (each 1.5 m high) and two bays (each 2.5 m wide). Herein, the pushover results of the ductile frame is selected in order to be compared with the dened hinge model characterised based on the assumptions of the previous section. The ductile frame incorporated full seismic details, composed of rectangular hoops, with 135 hooks, spaced at 30 mm on centre in critical locations of the beams, columns, and joints. In this study, the pushover analysis of the ductile structure was carried out using the dened hinge model by SAP 2000 and the outcomes in terms of loaddisplacement curve and sequence of hinging were compared with those reported by Filiatrault et al. [39]. Similar to their analysis, the distribution of lateral loads in pushover analysis was identical to the one used for the design of the structures. Also, the full gravity load was applied to the structure. The plastic properties of dened hinges were calculated using member reinforcement and the assumed models for concrete and steel described in the previous section. For this quantication, the actual material properties obtained by Filiatrault et al. [40] from tensile test on reinforcing steel and compressive test on concrete cylinders were considered. The base shear versus roof lateral displacement (so-called pushover) curve obtained from SAP 2000 was compared with that of Filiatrault et al. [39] in Fig. 4. Considering the differences between the lumped and spread plasticity approach, the two loaddisplacement curves agree well, proving the accuracy of the above mentioned assumptions for the denition of plastic hinges. In particular, the failure points predicted by the two methods correspond to each other reasonably well. The sequence of plastic hinging in the static nonlinear analysis of a ductile structure using the adopted lumped plasticity conforms to the capacity design philosophy in the Canadian concrete standard [43] to which the structure was designed. Similar to the static nonlinear results observed in the experimental test and numerical analysis by Filiatrault et al. [39,40], the rst four hinges occur in the beams followed by three others at the base of the columns. It is worth mentioning that there was no inelastic behaviour at the top of the columns up to the point of failure.

ecu 0:004 0:9qs

 fyh 300

In the above equation, ecu is the ultimate compressive strain of concrete, qs is the volumetric ratio of conning reinforcement and fyh represents the yield strength of transverse steel. Plastic rotation is dened as the difference between the ultimate and the yield curvature (curvature ductility) multiplied by the plastic hinge length. Many equations have been proposed by researchers for the plastic hinge length [32,37]. In this study, the plastic hinge was calculated as:

lP H=2

where Lp and H are the plastic hinge length and the height of section, respectively. This simple, yet adequately accurate relation for the plastic hinge length is also recommended by ATC-40 [38]. It

C B

Fig. 3. Idealised momentrotation (or momentcurvature) curve for a exural plastic hinge.

A. Eslami, H.R. Ronagh / Composites: Part B 45 (2013) 12651274

1269

80 70 60

Base shear, V(kN)

50 40 30 20 10 0

RUAUMOKO [39] SAP 2000 with lumped plasticity

20

40

60

80

100

120

140

160

180

200

220

Roof lateral displacement (mm)

Fig. 4. Comparison of pushover curves of ductile RC frame.

4.3. Pushover results of the original frames Pushover analysis consists of an incrementally increasing lateral load applied to the structure up to the failure point in the presence of a constant gravity load. In this study, the total dead load plus 20% of the live load based on the Iranian seismic code [24], is applied to the frame studied. For the seismic evaluation of a building, the lateral force prole applied to the building should represent, albeit approximately, the likely distribution of inertial forces induced during an earthquake. In the current paper, an inverted triangular distribution over the height is used as the lateral load pattern. This lateral load pattern provides better estimates of the capacity curve and seismic responses when compared with a uniform distribution. In addition, while inverted triangular distribution is more practical than multi-modal distribution, it would yield similar results [29]. It should be mentioned that the selected load pattern is similar to the lateral load distribution used for the seismic design of considered structures and has been suggested in the Iranian seismic code. Also, the effect of P D has been considered in all nonlinear analyses. Due to the exural cracking of the RC members, the stiffness of the members is reduced during the seismic loads. The reduction in the exural stiffness was considered in the nonlinear analysis by introducing the equivalent moment of inertia, Ieq, given by:

FRP connement. In the latter, the structure would attract higher seismic forces and hence the effectiveness of the retrot is decreased. 5. Design of composite wraps From the retrotting point of view using composite material, it was decided to increase the displacement capacity of the structure and provide more energy dissipation under seismic loads. Following this objective, two retrotting strategies were selected: (1) Fully exploiting the rotational capacity of beams and columns through GFRP wrapping of potential plastic hinge regions of each column in both code-compliant and poorly-conned frames, as demonstrated in Fig. 6. The primary length of connement, Lc, was considered to be equal to the plastic hinge region suggested in ACI 318-02 [26]. In practical implementation, a secondary region should be conned adjacent to the primary connement region. The thickness of FRP for conning the secondary connement region was suggested to be half of that in primary connement region [44]. (2) Increasing the ductility of beams in code-compliant structure in addition to column connement. This could be achieved using web bonded FRP with bres oriented in the direction of transverse reinforcement regions prone of inelastic behaviour. To provide connement in the columns, glass bres are more attractive than others. While they have the highest ultimate strain of any high modulus bre, their low fatigue and creep rupture resistance are not a crucial factor in this type of application [8]. A comparison between GFRP and CFRP material showed that the former provides more displacement ductility for the conned concrete with lower amount of composite material. In addition glass bre materials are more cost effective than the carbon bres. Thus the GFRP composite materials were considered for the connement of columns in this study. The design properties of unidirectional glass bre sheets as provided by the manufacturer are as follows: bre thickness tf = 0.589 mm per layer, tensile modulus Ef = 72397 MPa, ultimate tensile strain efu = 0.045, and ultimate tensile strength ffu = 3241 MPa [8]. The stressstrain model proposed by Lam and Teng [45,46] was selected for FRP-conned concrete. A comprehensive investigation

Ieq

My Ec /y

where My and /y are yield moment and yield curvature, respectively and Ec is Youngs modulus of concrete. Table 1 presents stiffness properties of all members together with the ratios of the equivalent moment of inertia (Ieq) to the gross moment of inertia (Ig). On account of negligible effect of connement on the yielding point, the stiffness ratio for both frames was calculated to be almost identical. The lateral loaddisplacement curves obtained from the nonlinear pushover analyses of code-compliant and poorly-conned frames are compared in Fig. 5. While the lateral-load carrying capacities of two structures are identical, lower connement resulted in a signicant decrease (about 34%) in the displacement ductility in the poorly-conned frame. In addition, considering the pre-yielding and post-yielding stiffness, both frames behaved in a similar manner. The fact that the connement of the members only increases the displacement capacity of the structures instead of stiffness, demonstrates the distinct advantage of retrotting by

1270

A. Eslami, H.R. Ronagh / Composites: Part B 45 (2013) 12651274

Table 1 Stiffness properties for nonlinear static analysis of the frames. Member type Columns Section AA Story no. 1 2 BB 3 4 CC 5 6 7 8 Mean value for columns Beams DD EE FF 14 56 78 6.22 E4 5.18 E4 3.81 E4 Pos. Internal External Internal External Internal External Internal External Internal External Internal External Internal External Internal External EcIeq (kN m2) 1.27 1.21 1.26 1.19 9.22 8.27 8.89 8.01 4.45 4.06 4.25 3.89 4.06 3.72 3.72 3.54 E5 E5 E5 E5 E4 E4 E4 E4 E4 E4 E4 E4 E4 E4 E4 E4 Ieq/Ig 0.50 0.48 0.49 0.47 0.36 0.33 0.35 0.32 0.36 0.33 0.35 0.32 0.33 0.3 0.3 0.29 0.37 0.51 0.42 0.31 0.41

Mean value for beams

1600 1400 1200

Base shear (kN)

1000 800 600 400 200 0

Intermediate frame Poorly-confined frame

100

200

300

400

500

600

700

Roof lateral displacement (mm)

Fig. 5. Lateral loaddisplacement curves of original frames.

GFRP confinement
Lc

RC column GFRP confinement

Lc

by the authors and other researchers [47] proved that this model is most suitable for predicting the maximum conned compressive strength and strain of circular and rectangular RC columns. This model was adopted by design guidelines [48,49] for strengthening of RC structures using externally bonded FRPs. Due to the effect of non-uniform stress distribution and curvature in the FRP jacket, the rupture strain of the FRP is lower than the ultimate tensile strain determined from direct coupon tests. Based on the evaluation of experimental data, Lam and Teng [45,46] suggested a value of 62.4% ultimate tensile strength for GFRP composites. 5.1. Intermediate frame The rst retrotting conguration consisted of wrapping the column at the high seismic demand regions in order to provide more ductility capacity without changing the location of plastic hinges. During the nonlinear static analysis of the original frame,

Fig. 6. FRP jacketed regions of columns for seismic retrotting.

A. Eslami, H.R. Ronagh / Composites: Part B 45 (2013) 12651274

1271

only plastic hinges in the columns of the bottom stories exhibited severe plastic behaviour. Thus the columns of the rst four stories were wrapped with GFRP in the potential regions of plastic hinge formation. In addition, a sensitivity analysis conrmed that the GFRP connement of the top columns would not affect the nonlinear response of the frame signicantly. To quantify the amount of GFRP to be installed in the columns, the internal column at the rst storey was selected since it carried the maximum axial force due to gravity loads and thus had the minimum rotational capacity. The composite thickness was calculated based on changing the column failure mode in the section analysis from concrete crushing to steel rupture. For the selected column, it could be achieved with four layers of GFRP sheets which increased the ultimate strain of column concrete to ecu = 0.0235. The effect of FRP connement was taken into account by modifying the exural inelastic behaviour of the elements at the member ends, where it was assumed that the nonlinear plastic hinges would be placed. Fig. 7 compares the pushover curves of the retrotted and the original frame. As expected, due to the adequate connement provided through the transverse steel reinforcement, the additional connement provided by composite material could not enhance the lateral displacement capacity signicantly (roughly 7%). More exact evaluation of hinge damage state illustrated that a higher displacement capacity for the structure is achievable by increasing the rotational capacity of the plastic hinges at the lower beams. In order to pursue the second retrotting technique; the plastic rotational capacity of hinges at the beams was increased by 24%. This value was selected from the average ductility enhancement in the code-compliant RC connections reported in the past studies [9,50]. In the calculation of this mean value, those CFRP congurations that increase both the strength and stiffness have been neglected. The number and thickness of CFRP wrap to reach the targeted displacement ductility level could be calculated using the nite element analysis of the retrotted joints. As observed in Fig. 7, more ductility capacity was provided for the intermediate frame with retrotting both beams and columns of the rst four stories. Retrotting congurations that increase the load carrying capacity of this type of building might result in a higher seismic resistance. Discussion on this type of strengthening conguration is beyond the scope of this paper. 5.2. Poorly-conned frame Compared to the code-compliant building, the poorly-conned frame suffered from the inadequate transverse steel reinforcement
1600 1400 1200

resulting in low ductility capacity of the structure during the lateral loading. This represents the situation of many RC structures built based on the pre-seismic code provisions. In order to overcome this deciency, using FRP wraps to provide additional connement is an efcient rehabilitation method since it increases the ductility capacity without considerable strength increment. In this study, the efciency of GFRP wraps to improve a displacement capacity of a poorly-conned frame to the levels seen in an intermediate frames is studied. For this purpose, the plastic rotation of columns is enhanced using FRP connement of concrete. The columns of the rst four stories were wrapped with four layers of GFRP, while three layers of GFRP wraps was applied on the columns of the other stories. Composite thickness in the bottom column was selected to be similar to the intermediate frame, while three layers of GFRP was calculated to be adequate to provide stress strain model for conned concrete similar to the bottom columns. In both column categories, the ultimate concrete strength 0 and strain were fcc 46 MPa and ecu = 0.0235, respectively. The lateral loaddisplacement curves obtained from the nonlinear static analyses of FRP-conned and original frames are compared in Fig. 8. The displacement capacity of frame was enhanced by approximately 38% using GFRP wrapping of columns. It is worth mentioning that FRP-connement has no effect on the slope of the pushover curve before yielding. The nonlinear outcomes emphasised the capability of column connement in improving the seismic performance of decient RC building with inadequate transverse reinforcement.

6. Seismic assessment of the retrotted structure The capacity diagrams, obtained using the pushover analysis of the original and retrotted frames, could be compared with the demand spectrum to assess the seismic performance of structures at different levels of ground motion. In this study, amongst different methods developed, the N2 method [23] was employed for the seismic analysis of the poorly-conned frame. The method is formulated in accelerationdisplacement (AD) format with a visual representation of the procedure. Seismic demand was dened with the elastic response spectrum (soil type III, 5% damping) suggested for the design of structures in the Iranian seismic code [24] (Fig. 9a). In the capacity spectrum approach, the elastic acceleration (Sae) and displacement spectrum (Sde) are plotted in AD format in order to calculate the

Base shear (kN)

1000 800 600 400 200 0

Original frame FRP wrapping of columns FRP wrapping of columns and beams

100

200

300

400

500

600

700

800

Roof lateral displacement (mm)

Fig. 7. Comparison of pushover curves of intermediate frame retrotted using different FRP application.

1272

A. Eslami, H.R. Ronagh / Composites: Part B 45 (2013) 12651274

1600 1400 1200

Base shear (kN)

1000 800

FRP retrofitted frame

600 400 200 0

Original frame

100

200

300

400

500

600

Roof lateral displacement (mm)

Fig. 8. Comparison of pushover curves for poorly-conned frame.

seismic demand of the equivalent SDOF system, as shown in Fig. 9b. Neglecting the performance level restrictions, the maximum sustainable ground motion for the equivalent SDOF systems of the original and retrotted frame was calculated using the N2 method to be 0.32 g and 0.52 g, respectively. These peak ground accelerations were calculated using a feedback procedure to achieve the same elastic displacement demand and elastic dis-

placement capacity for equivalent SDOF. The capacity curve obtained from the pushover analysis was transformed from the MDOF system to an equivalent SDOF. The transformation factor is given by:

CP

mi /i mi /2 i

(a)

(b)

where mi and /i are the mass and normalised displacement of the ith story. Assuming an equal mass for all stories, the displacement distribution pattern is similar to the lateral load distribution and the transformation factor was calculated to be 1.41. The seismic demand for the equivalent SDOF system could be determined by plotting both demand spectra and capacity diagram in the same graph. The intersection of radial line corresponding to the elastic period of the idealised bilinear system, T, with the elastic demand spectrum determines the acceleration demand and corresponding displacement demand required for the elastic behaviour. The amount of ductility factor, l depends on whether T is larger or smaller than the characteristic period of ground motion, Tc. The inelastic demand of SDOF system in terms of acceleration and displacement is dened by the intersection of the idealised capacity curve and the inelastic demand spectrum corresponding to l. The displacement demand of the MDOF system is obtained by multiplying the SDOF demand by the transformation factor. Fig. 10 provides a graphical demonstration of the N2 method for the poorly-conned frame. The requested ductility for the PGA of 0.52 g and 0.32 g was equal to 3 and 1.97, respectively. As observed from the idealised bilinear capacity curves, the available structural ductility of the original frame increased by 52% after FRP application. This resulted in an increase in the seismic load capacity to the PGA of 0.52 g instead of 0.32 g. 7. Conclusions This study reported on the results of a numerical investigation on the seismic enhancement of code-compliant and poorly-conned RC buildings. Retrotting strategy was focused on increasing the ductility instead of strengthening using FRP wrapping. Nonlinear static analyses of the frames were carried out in order to simulate the seismic performance of buildings during ground motion. Plastic behaviour of the elements was characterised using two nonlinear hinges at the end sections of the members based on the lumped plasticity concept. The nonlinear properties of plastic hinges were

Fig. 9. (a) Acceleration response spectrum normalised to 1.0 g peak ground acceleration (soil type III, 5% damping); and (b) spectrum in AD format.

A. Eslami, H.R. Ronagh / Composites: Part B 45 (2013) 12651274

1273

Fig. 10. Demand spectra versus capacity diagrams of poorly-conned frame.

calculated based on the well-established models for concrete and steel reinforcement. After verication of the adopted assumptions for nonlinear characterisation of members, the results of pushover analysis were implemented in seismic assessment of frames. Nonlinear analyses results outlined the inefciency of column wrapping in improving the ductility of code-compliant structures. This is particularly due to the adequate lateral connement provided by transverse steel reinforcements. For the intermediate frame, increasing the beam ductility appears to be more effective. However, FRP wraps could enable the poorly conned structure to resist much higher ground motion (PGA = 0.52 g) through increasing the ductility and energy dissipation capacity. More studies need to be conducted on the FRP-retrotting of pre-existing RC buildings in order to not only scrutinise their seismic response but also provide a more economical retrotting scheme for this type of buildings. Considering the more stringent demands set by UBC97 [51] and more recent seismic codes such as IBC2006 [52], which basically increase the lateral design forces in order to combat the effect of near-fault pulse type ground motions; strengthening of buildings designed based on the older versions of seismic codes might be necessary. It should be pointed out that seismic evaluation of building could be carried out considering the rational nonlinear analysis methods and criteria adopted in well-known rehabilitation codes such as ATC-40 [53] or FEMA356 [25]. In addition, FRP retrotting/reinforcing applications should comply with the regulations of international codes [5457]. Acknowledgment The authors gratefully thank Imbsen & Associate, Inc. for providing a free license for the XTRACT software [33]. References
[1] Saadatmanesh H. Extending service life of concrete and masonry structures with bre composites. Constr Build Mater 1997;11:32735. [2] Shahawy MA, Arockiasamy M, Beitelman T, Sowrirajan R. Reinforced concrete rectangular beams strengthened with CFRP laminates. Compos Part B: Eng 1996;27:22533. [3] Rabinovitch O, Frostig Y. Experiments and analytical comparison of RC beams strengthened with CFRP composites. Compos Part B: Eng 2003;34:66377. [4] Saadatmanesh H, Ehsani MR, Jin L. Repair of earthquake-damaged RC columns with FRP wraps. ACI Struct J 1997;94:20615.

[5] Chaallal O, Shahawy M. Performance of bre-reinforced polymer-wrapped reinforced concrete column under combined axialexural loading. ACI Struct J 2000;97:65968. [6] Harajli MH, Rteil AA. Effect of connement using bre-reinforced polymer or bre-reinforced concrete on seismic performance of gravity load-designed columns. ACI Struct J 2004;101:4756. [7] Toutanji H, Han M, Gilbert J, Matthys S. Behaviour of large-scale rectangular columns conned with FRP composites. J Compos Constr 2010;14:6271. [8] De Luca A, Nardone F, Matta F, Nanni A, Lignola GP, Prota A. Structural evaluation of full-scale FRP-conned reinforced concrete columns. J Compos Constr 2011;15:11223. [9] Mostonejad D, Talaeitaba SB. Finite element modeling of RC connections strengthened with FRP laminates. Iran J Sci Technol Trans B: Eng 2006;30:2130. [10] Granata PJ, Parvin A. An experimental study on Kevlar strengthening of beam column connections. Compos Struct 2001;53:16371. [11] Attari N, Amziane S, Chemrouk M. Efciency of beamcolumn joint strengthened by FRP laminates. Adv Compos Mater 2010;19:17183. [12] Mahini SS, Ronagh HR. Web-bonded FRPs for relocation of plastic hinges away from the column face in exterior RC joints. Compos Struct 2011;93:246072. [13] Mukherjee A, Joshi M. FRPC reinforced concrete beamcolumn joints under cyclic excitation. Compos Struct 2005;70:18599. [14] Antonopoulos CP, Triantallou TC. Analysis of FRP-strengthened RC beam column joints. J Compos Constr 2002;6:4151. [15] Mosallam AS, Banerjee S. Shear enhancement of reinforced concrete beams strengthened with FRP composite laminates. Compos Part B: Eng 2007;38:78193. [16] Ascione L, Berardi VP, Feo L, Mancusi G. A numerical evaluation of the interlaminar stress state in externally FRP plated RC beams. Compos Part B: Eng 2005;36:8390. [17] Ascione L, Mancusi G, Spadea S. Flexural behaviour of concrete beams reinforced with GFRP bars. Strain 2010;46:4609. [18] Balsamo A, Colombo A, Manfredi G, Negro P, Prota A. Seismic behaviour of a full-scale RC frame repaired using CFRP laminates. Eng Struct 2005;27:76980. [19] Di Ludovico M, Prota A, Manfredi G, Cosenza E. Seismic strengthening of an under-designed RC structure with FRP. Earthquake Eng Struct Dynam 2008;37:14162. [20] Garcia R, Hajirasouliha I, Pilakoutas K. Seismic behaviour of decient RC frames strengthened with CFRP composites. Eng Struct 2010;32:307585. [21] Niroomandi A, Maheri A, Maheri MR, Mahini SS. Seismic performance of ordinary RC frames retrotted at joints by FRP sheets. Eng Struct 2010;32:232636. [22] Computers and Structures Inc. Static and dynamic nite element analysis of structures. SAP 2000. 14.1.0 ed. Berkeley, CA; 2009. [23] Fajfar P. A nonlinear analysis method for performance-based seismic design. Earthquake Spectra 2000;16:57392. [24] Permanent committee for revising the Iranian code for seismic resistant design of buildings. Iranian code of practice for seismic resistant design of buildings (Standard No. 280005). 3rd ed. Tehran (Iran): Building and Housing Research Center (BHRC); 2005. [25] American Society of Civil Engineering (ASCE). Prestandard and commentary for the seismic rehabilitation of buildings (FEMA-356). Washington (DC): Federal Emergency Management Agency; 2000.

1274

A. Eslami, H.R. Ronagh / Composites: Part B 45 (2013) 12651274 [42] NBCC. National building code of Canada. Ottawa (Ont): Associate Committee on the National Building Code, National Research Council of Canada; 1995. [43] CSA. Design of concrete structures for building. Standard CAN-A-233-94. Rexdale (Ont).: Canadian Standards Association; 1994. [44] Zou XK, Teng JG, De Lorenzis L, Xia SH. Optimal performance-based design of FRP jackets for seismic retrot of reinforced concrete frames. Compos Part B: Eng 2007;38:58497. [45] Lam L, Teng JG. Design-oriented stressstrain model for FRP-conned concrete. Constr Build Mater 2003;17:47189. [46] Lam L, Teng JG. Design-oriented stressstrain model for FRP-conned concrete in rectangular columns. J Reinf Plast Compos 2003;22:114986. [47] Rocca S, Galati N, Nanni A. Interaction diagram methodology for design of FRPconned reinforced concrete columns. Constr Build Mater 2009;23:150820. [48] ACI Committee 440.2. Guide for the design and construction of externally bonded FRP systems for strengthening concrete structures (ACI 440.2R-08). Farmington Hills (MI): American Concrete Institute; 2008. [49] Concrete society. Design guidance for strengthening concrete structures with bre composite materials. Technical Rep. No. 55, 2nd Ed., Crowthorne, Berkshire, UK; 2004. [50] Talaeitaba SB. Ductility enhancement of RC joints with FRP laminates. Isafahn (Iran): Isfahan University of Technology (IUT); 2003 [n Farsi]. [51] International conference of building ofcials. Whittier (CA): Uniform Building Code (UBC); 1997. [52] IBC. International building code falls church. Virginia: International Code Council; 2006. [53] ATC. Seismic evaluation and retrot of concrete buildings (ATC-40). California: Applied Technology Council; 1996. [54] ACI. Guide for the design and construction of structural concrete reinforced with FRP bars (ACI 440.1R-06). Farmington Hills (MI): American Concrete Institute; 2006. [55] CSA. Specication for bre reinforced polymers (CSA-S807-10). Toronto (Canada): Canadian Standards Association (CSA) International; 2010. [56] Fib bulletin no. 40. FRP reinforcement in RC structures. 2007 160pp. [57] JSCE. Recommendation for upgrading of concrete structures with use of continuous ber sheets. Concrete Engineering Series 41. Tokyo (Japan): Japan Society of Civil Engineers; 2001.

[26] ACI Committee 318. Building code requirements for structural concrete (ACI 31802) and commentary (ACI 318R-02). Farmington Hills, MI: American Concrete Institute; 2002. [27] Krawinkler H, Seneviratna GDPK. Pros and cons of a pushover analysis of seismic performance evaluation. Eng Struct 1998;20:45264. [28] Sadjadi R, Kianoush MR, Talebi S. Seismic performance of reinforced concrete moment resisting frames. Eng Struct 2007;29:236580. [29] Mwafy AM, Elnashai AS. Static pushover versus dynamic collapse analysis of RC buildings. Eng Struct 2001;23:40724. [30] Inel M, Ozmen HB. Effects of plastic hinge properties in nonlinear analysis of reinforced concrete buildings. Eng Struct 2006;28:1494502. [31] Jeong SH, Elnashai AS. Analytical assessment of an irregular RC frame for fullscale 3D pseudo-dynamic testing Part I: Analytical model verication. J Earthquake Eng 2005;9:95128. [32] Park R, Paulay T. Reinforced concrete structures. New York: Wiley; 1975. [33] Imbseon and Associates Inc. Cross section analysis program for structural engineers. XTRACT. 3.0.8 ed. California; 2011. [34] Mander JB, Priestley MJN, Park R. Theoretical stressstrain model for conned concrete. J Struct Eng 1988;114:180426. [35] ASTM A615M. Standard specication for deformed and plain carbon-steel bars for concrete reinforcement. Philadelphia (PA): American Society for Testing and Materials; 2009. [36] Scott BD, Park R, Priestley MJN. Stressstrain behaviour of concrete conned by overlapping hoops at low and high strain rates. J Am Concr Inst 1982;79:1327. [37] Paulay T, Priestley MJN. Seismic design of reinforced concrete and masonry buildings. New York: Wiley; 1992. [38] ATC. Seismic evaluation and retrot of concrete buildings (ATC-40). California: Applied Technology Council; 1996. [39] Filiatrault A, Lachapelle E, Lamontagne P. Seismic performance of ductile and nominally ductile reinforced concrete moment resisting frames. II. Analytical study. Can J Civil Eng 1998;25:34252. [40] Filiatrault A, Lachapelle E, Lamontagne P. Seismic performance of ductile and nominally ductile reinforced concrete moment resisting frames. I. Experimental study. Can J Civil Eng 1998;25:33141. [41] Carr AJ. RUAUMOKO the Maori god of volcano and earthquake. Computer Program Library. 27 January 1996 ed. Christchurch (New Zealand): Department of Civil Engineering, University of Canterbury; 1996.