Optimized FRP wrapping schemes for circular concrete columns under axial compression

Transcription

1 University o Wollongong Research Online Faculty o Engineering and Inormation Sciences - Papers: Part A Faculty o Engineering and Inormation Sciences 2015 Optimized FRP wrapping schemes or circular concrete columns under axial compression Thong M. Pham University o Wollongong, tmp@uow.edu.au Muhammad N. S Hadi University o Wollongong, mhadi@uow.edu.au Jim Yousse University o Wollongong, jy201@uowmail.edu.au Publication Details Pham, T. M., Hadi, M. N. S. & Yousse, J. (2015). Optimized FRP wrapping schemes or circular concrete columns under axial compression. Journal o Composites or Construction, 19 (6), Research Online is the open access institutional repository or the University o Wollongong. For urther inormation contact the UOW Library: research-pubs@uow.edu.au

2 Optimized FRP wrapping schemes or circular concrete columns under axial compression Abstract This study investigates the behavior and ailure modes o FRP conined concrete wrapped with dierent FRP schemes, including ully wrapped, partially wrapped and non-uniormly wrapped concrete cylinders. By using the same amount o FRP, this study proposes a new wrapping scheme that provides a higher compressive strength and strain or FRP-conined concrete, in comparison with conventional ully wrapping schemes. A total o thirty three specimens were cast and tested, with three o these specimens acting as reerence specimens and the remaining specimens wrapped with dierent types o FRP (CFRP and GFRP) by dierent wrapping schemes. For specimens that belong to the descending branch type, the partially wrapped specimens had a lower compressive strength but a higher axial strain as compared to the corresponding ully wrapped specimens. In addition, the non-uniormly wrapped specimens achieved both a higher compressive strength and axial strain in comparison with the ully wrapped specimens. Furthermore, the partially wrapping scheme changes the ailure modes o the specimens and the angle o the ailure surace. A new equation that can be used to predict the axial strain o concrete cylinders wrapped partially with FRP is proposed. Disciplines Engineering Science and Technology Studies Publication Details Pham, T. M., Hadi, M. N. S. & Yousse, J. (2015). Optimized FRP wrapping schemes or circular concrete columns under axial compression. Journal o Composites or Construction, 19 (6), This journal article is available at Research Online:

3 1 2 Optimized FRP Wrapping Schemes or Circular Concrete Columns under Axial Compression 3 Thong M. Pham A.M.ASCE 1, Muhammad N.S. Hadi, F.ASCE 2 and Jim Yousse 3 4 Abstract This study investigates the behavior and ailure modes o FRP conined concrete wrapped with dierent FRP schemes, including ully wrapped, partially wrapped and non-uniormly wrapped concrete cylinders. By using the same amount o FRP, this study proposes a new wrapping scheme that provides a higher compressive strength and strain or FRP-conined concrete, in comparison with conventional ully wrapping schemes. A total o thirty three specimens were cast and tested, with three o these specimens acting as reerence specimens and the remaining specimens wrapped with dierent types o FRP (CFRP and GFRP) by dierent wrapping schemes. For specimens that belong to the descending branch type, the partially wrapped specimens had a lower compressive strength but a higher axial strain as compared to the corresponding ully wrapped specimens. In addition, the non-uniormly wrapped specimens achieved both a higher compressive strength and axial strain in comparison with the ully wrapped specimens. Furthermore, the partially wrapping scheme changes the ailure modes o the specimens and the angle o the ailure surace. A new equation that can be used to predict the axial strain o concrete cylinders wrapped partially with FRP is proposed CE Database subject headings: Fiber Reinorced Polymer; Coninement; Concrete columns; Strain; Stress-strain relation; Concrete; Cylinders. 1 Postdoctoral Research Associate, School o Civil and Mechanical Engineering, Curtin University, Kent Street, Bentley, WA 6102, Australia; Formerly, PhD Scholar, School o Civil, Mining and Environmental Engineering, University o Wollongong, Wollongong, NSW 2522, Australia. thong.pham@curtin.edu.au 2 Associate Proessor, School o Civil, Mining and Environmental Engineering, University o Wollongong, Wollongong, NSW 2522, Australia (corresponding author). mhadi@uow.edu.au 3 Ph.D. Candidate, School o Civil, Mining and Environmental Engineering, University o Wollongong, Wollongong, NSW 2522, Australia. jy201@uowmail.edu.au 1

4 22 Introduction Fiber Reinorced Polymer (FRP) has been commonly used to strengthen existing reinorced concrete (RC) columns in recent years. In such cases, FRP is a conining material or concrete in which the coninement eect leads to increase in the strength and ductility o columns. In early experimental studies that ocused on retroitting RC columns with FRP, the columns were usually wrapped ully with FRP sheets. This wrapping scheme provides continuous coninement to the columns along their longitudinal axes. Most o the studies in the literature ocus only on columns ully wrapped with FRP (Chaallal et al. 2003; Hadi et al. 2013; Pham et al. 2013; Pham and Hadi 2014a; Smith et al. 2010). In addition, columns wrapped partially with FRP have also been proven to show increases in strength and ductility, as compared to equivalent unconined columns (Colomb et al. 2008; Maaddawy 2009; Turgay et al. 2010). However, there is no study that makes a comparison o the coninement eicacy between partially and ully wrapping schemes in terms o optimization o the FRP amount. In addition, the progressive ailure o those specimens has not been extensively studied. Thereore, it is necessary to investigate the coninement eicacy and ailure mechanisms o columns partially wrapped versus columns ully wrapped with FRP In addition, the available design guidelines or columns wrapped with FRP (ACI 440.2R ; ib 2001; TR ) are utilized to estimate the capacities o partially FRP-wrapped specimens. Among these studies, ACI-440.2R (2008) and technical report TR 55 (2012) do not provide inormation about the coninement eect o concrete columns partially wrapped with FRP. Meanwhile, ib (2001) suggests a reduction actor to take into account the eect o partial wrapping columns. The study by ib (2001) adopts an assumption proposed by Mander et al. (1988) or the coninement eect o steel ties in RC columns to analyze the eicacy o FRP partially wrapped columns. Thereore, there has been a lack o theoretical and 2

5 experimental works about partial FRP-conined concrete. For this reason, an experimental program was developed in this study to compare the coninement eicacy o FRP partially wrapped columns as compared to FRP ully wrapped columns. The same amount o FRP was wrapped onto identical concrete columns by dierent wrapping schemes to achieve an optimized wrapping design. 51 Coninement Mechanism 52 Fully Wrapped Columns In the literature, the term FRP conined concrete is understood automatically as concrete wrapped ully with FRP. When a circular concrete column is horizontally wrapped with FRP around its perimeter, the whole column is conined by the lateral pressure exerted rom the FRP jackets as shown in Fig. 1a. Many studies have been carried out to investigate the behaviors and estimate the capacities o columns wrapped ully with FRP (De Luca and Nanni 2011; Lam and Teng 2003; Pham and Hadi 2014b; Pham and Hadi 2014c; Teng et al. 2009; Toutanji 1999; Wu and Zhou 2010). The conining pressure is assumed to be uniorm in the cross section and along the axial axis o the circular columns. Among the existing studies, the model proposed by Lam and Teng (2003) is adopted in this study to calculate the compressive strength or columns wrapped ully with FRP as ollows: 63 ' cc ' co (1) l ' co 64 where cc and co are respectively the compressive strength o conined concrete and 65 unconined concrete, and l is the eective conining pressure as ollows: 66 l 2E et (2) D 3

6 where E is the elastic modulus o FRP, t is the nominal thickness o FRP jacket, D is the diameter o the column section, and e is the actual rupture strain o FRP in the hoop direction. The model by Lam and Teng (2003) is chosen because it provides a reasonable accuracy with a very simple orm. The simplicity o the model by Lam and Teng (2003) is utilized to establish a new and simple strain model, which is presented in the sections below. The strain model proposed by Pham and Hadi (2013) is adopted to calculate the compressive axial strain o conined concrete as ollows: 2kt e e cc co (3) ' Dco 3. 3t e where cc is the ultimate axial strain o conined concrete, co is the axial strain at the peak stress o unconined concrete, k = 7.6 is the proportion actor, and e is the actual rupture strength o FRP. 78 Partially Wrapped Columns As mentioned above, concrete columns wrapped partially with FRP have been experimentally veriied to increase their strength and ductility. Concrete columns partially wrapped with FRP are less eicient in nature than ully wrapped columns as both conined and unconined zones exist (Fig. 1b). An approach similar to the one proposed by Sheikh and Uzumeri (1980) is adopted to determine the eective conining pressure on the concrete core. The eective conining pressure is assumed to be exerted eectively on the part o the concrete core where the conining pressure has ully developed due to the arching action as shown in Fig. 1b. The arching eect is assumed to be described by a second-degree parabola with initial slope o In such a case, a coninement eective coeicient (k e ) is introduced to take the partial wrapping into account as ollows: 4

7 A e s ke 1 (4) A 2D c where A e and A c are respectively the area o eectively conined concrete core and the crosssectional area, and s is the clear spacing between two FRP bands. Consequently, the compressive strength o concrete columns wrapped partially with FRP could be calculated as: ' cc ' co ke (5) Where k e is estimated based on Eq. 4 and l shown in the ollowing equation is the equivalent conining pressure rom the FRP, assumed to be uniormly distributed along the longitudinal axis o the column. ' l ' co 97 ' l 2E et D w w s (6) where w is the width o FRP bands and s is the clear spacing between FRP bands as shown in Fig. 1b. Experimental Program 101 Design o Experiments A total o thirty three FRP conined concrete cylinders were cast and tested at the High Bay Laboratory o the University o Wollongong. The dimensions o the concrete cylinder specimens were 150 mm in diameter and 300 mm in height. All the specimens were cast rom the same batch o concrete. The 28 day cylinder compressive strength was 52 MPa The experimental program was composed o several groups o cylinders in order to evaluate the coninement eicacy between partially and ully wrapping schemes in terms o optimization o the wrapping schemes. The notation o the specimens consists o three parts: the irst part states the type o conining FRP material, with G and C representing GFRP and CFRP respectively. The second part is either a letter R, F, and P stating the name 5

8 o the sub-group, namely, reerence group (R), ully wrapped group (F) and partially wrapped group (P). The last part o the specimen notation is a number which indicates the number o FRP layers. Details o the specimens are presented in Table The partially wrapped specimens contain FRP bands which are 25 mm in width spaced evenly along the height o the specimen. The optimized partially wrapped specimens include two numbers in the notation, or example GP31. The irst number indicates the number o 25 mm evenly spaced partial FRP layers and the second number depicts the number o FRP layers in between these evenly spaced partial layers. These specimens were designed such that they ollow a non-uniorm wrapping coniguration but ensure the specimen is ully conined at every location. The thicker band is called a tie band and the thinner band is called a cover band. Taking specimen GP31 as an example, the tie bands have three FRP layers which are 25 mm in width, while the cover bands have one FRP layer as shown in Figure 2. Three identical specimens were made or each wrapping scheme In order to analyze the coninement eectiveness between dierent wrapping schemes, the specimens were divided in our groups (as shown in Table 1) such that the specimens in each group incorporate the same amount o FRP but in a dierent wrapping scheme, either ully, partially or optimized non-uniormly wrapped. The specimens in the irst group are reerence specimens which did not include any internal or external reinorcement. The specimens in the second and third groups were conined by GFRP and CFRP respectively, such that the ully, partially and optimized non-uniorm wrapping schemes were equivalent to two layers o ull wrapping. Similarly, the wrapping schemes o the specimens in the ourth group were equivalent to three layers o ull wrapping Ater 28 days, the specimens were wrapped with a number o FRP layers as shown in Table 1. The adhesive used was a mixture o epoxy resin and hardener at 5:1 ratio. Beore the irst 6

9 layer o FRP was attached, the adhesive was spread onto the surace o the specimen and CFRP was attached onto the surace with the main ibers oriented in the hoop direction. Ater the irst layer, the adhesive was spread onto the surace o the irst layer o FRP and the second layer was continuously bonded. The third layer o FRP was applied in a similar manner, ensuring that 100 mm overlap was maintained. The ends o each wrapped specimen were strengthened with additional one layer o FRP strips 25 mm in width. 141 Instrumentation In order to measure the hoop strains o the FRP jacket, three strain gages with a gage length o 5 mm were attached at the mid height o the specimens and evenly distributed away rom the overlap or the ully wrapped specimens. In the partially wrapped specimens, three strain gages were bonded symmetrically on a tie band and other three were bonded on a cover band at midheight o the specimen Furthermore, a longitudinal compressometer as shown in Fig. 3 was used to measure the axial strain o the specimens. A Linear variable dierential transormer (LVDT) was mounted on the upper ring and the tip o the LVDT rests on an anvil. The readability, the accuracy, and the repeatability o the LVDT complies with the Australian standard (Australian Standard ) The compression tests or all the specimens were conducted using the Denison 5000 kn capacity testing machine. The specimens were capped with high strength plaster to ensure ull contact between the loading plate and the specimen. Calibration was carried out to ensure that the specimens were placed at the center o the testing machine. Each specimen was irst loaded to around 30% o its unconined capacity to check the alignment. I required, the specimen was unloaded, realigned, and loaded again. The tests were conducted as delection 7

10 controlled with a rate o 0.5 mm/min. The readings o the load, LVDT and strain gages were taken using a data logging system and were subsequently saved in a control computer. 160 Experimental Results 161 Preliminary tests The actual compressive strength o unconined concrete calculated rom three reerence specimens (R1, R2, and R3) was 54 MPa. The axial strain o unconined concrete at the maximum load was 0.23 %. In this study two types o CFRP were used to conine the concrete, which both had a unidirectional iber density o 340 g/m 2 and a nominal thickness o 0.45 mm, but with varying nominal widths o 75 mm and 25 mm. The GFRP utilized had a unidirectional iber density o 440 g/m 2, a nominal thickness o 0.35 mm and a nominal width o 50 mm Five coupons or each type o FRP were made according to ASTM D7565 (2010) and tested to determine the mechanical properties. The two types o CFRP coupons were made o three layers o FRP with a nominal thickness o 1.35 mm and both types had very similar properties as shown in Table 2. For simplicity the coupons produced rom the 75 mm tape are denoted by CFRP (75) while the coupons rom the 25 mm tape are reerred to as CFRP (25). For GFRP, two-layered coupons containing two overlapping iber sheets were prepared and tested. The nominal thickness o the coupons was 0.7 mm. All coupons had the dimensions 25 mm x 250 mm. The epoxy resin had 54 MPa tensile strength, 2.8 GPa tensile modulus and 3.4% tensile elongation (West System n.d. 2015). 178 Failure Modes All specimens were tested until ailure. The specimens wrapped ully with FRP (CF2, CF3, and GF2) ailed by rupture o FRP at the midheight. The ailure surace o the ully wrapped 8

11 specimens was ound to be approximately 45 degree inclined, as shown in Fig. 4a. Meanwhile, the partially wrapped specimens (CP40, CP60, and GP40) showed many small cracks on the concrete surace at a stress equal to the unconined concrete strength, as shown in Fig. 4b. The concrete between the FRP bands, close to the outer surace o the specimen, started crushing while the concrete core was still conined by the FRP. Cracks on the concrete surace developed as the applied load increased, as shown in Fig. 4c. At the very high stress level, the concrete between the FRP bands spalled o while the concrete under the FRP bands and the core were still conined. These specimens then ailed explosively by FRP rupture at the midheight (Fig. 4d) The angle o the ailure surace with respect to the horizon or the partially wrapped specimens was signiicantly dierent rom the ully wrapping specimens. As shown in Fig. 4d, the ailure surace took place at the spacing between FRP bands. This change o the ailure surace depends on the wrapping schemes and the stiness o the FRP bands. When the axial stress o the conined concrete was higher than the unconined concrete strength, the 45 degree ailure surace may have originally transpired in the concrete cores, but cracks were arrested by FRP bands under the high stress stage. I the stiness o the FRP bands is not strong enough (Specimen GP40) to prevent the development o the cracks, the ailure surace takes place at approximately 45 degrees as shown in Fig. 4e. In contrast, the stiness o the FRP bands in Specimens CP40 and CP60 is great enough so that it changed the ailure surace as depicted in Fig. 4d. It is worth mentioning that the stiness o the FRP bands aects the tangent modulus o FRP-conined concrete. Tamuzs et al. (2008) suggested that the low value o the tangent modulus causes column stability collapse directly as the unconined concrete strength level is surpassed. 9

12 Furthermore, specimens with optimized non-uniorm wrapping schemes showed a dierent ailure mode as compared to the others. At a stress level equal to the unconined concrete strength, the concrete was still conined by the FRP tie bands and cover bands. During the loading process, the lateral strains o the tie bands and the cover bands were almost identical, with the exception o Specimen CP40_3. The ailure modes o these specimens are similar to those o the ull wrapping specimens. The Non-uniorm wrapped specimens ailed by FRP rupture simultaneously at the two bands (tie band and cover band) at the midheight, as shown in Fig. 4. It is worth mentioning that intermittent coninement resulted rom partial coninement (Specimens GP40, CP40, and CP60) makes the concrete to communicate directly with the surroundings, or instance moisture, heat, and evaporation. 214 Stress-Strain Relation Stress-strain relations o the tested specimens were divided into two main types based on the shape o the stress-strain curves. These included specimens in the ascending branch type and descending branch type. A FRP conined concrete column exhibits the ascending type curve as a signiicant improvement o the compressive strength and strain o a FRP conined concrete column could be expected. Otherwise, FRP conined concrete with a stress-strain curve o the descending type illustrates a concrete stress at the ultimate strain below the compressive strength o unconined concrete. Specimens wrapped with glass iber are designed to behave as the descending branch type while specimens wrapped with carbon iber belong to the ascending branch type. Details o all tested specimens are summarized in Table Stress-strain relations o specimens wrapped by equivalent two GFRP layers were plotted in Fig. 5. The specimens which were wrapped with an equivalent o two layers o FRP had identical stress-strain curves at the early stages o loading and experienced slight dierences 10

13 at the latter stage o testing. Specimens GF2 and GP40 had the descending branch type stressstrain curve while the stress-strain curves o Specimens GP31 kept constant ater reaching the unconined concrete strength and then increased again to ailure. The axial stress o Specimens GF2 reached the unconined concrete strength (54 MPa) and then kept constant until the FRP ailed by rupture as shown in Fig. 5a. The average compressive conined concrete strength and strain o Specimens GF2 are 57 MPa and 0.97 %, respectively. Although Specimens GP40 obtained a lower maximum stress (53 MPa) as compared to that o Specimens GF2, they achieved a larger maximum axial strain (1.18%) than the ormer specimens. The axial strain o Specimens GP40 increased by % as compared to that o Specimens GF2 (Fig. 5b). Meanwhile, Specimens GF31 achieved both a higher maximum axial stress (60 MPa) and axial strain (1.02 %), as compared to Specimen GF2, as shown in Fig. 5c Apart rom the specimens above, the specimens which were wrapped with an equivalent o two layers o FRP, had similar stiness during the whole loading process, as shown in Fig. 6. The maximum axial stress o Specimens CF2 was 99 MPa and its corresponding axial strain was 2.13%. Specimens CP40 reached the maximum axial stress at 95 MPa and the corresponding axial strain at 2.08%. Specimen CP40_1 ailed by premature rupture o FRP ( l = 1.18 %) that resulted in very lower maximum axial stress. The average maximum axial stress and axial strain o Specimens CP31 were 98 MPa and 2.12 %, respectively The specimens that were wrapped with an equivalent o three layers o FRP had similar stress-strain curves but experienced a slight dierence in the axial stiness or the whole loading process as shown in Fig. 7. Specimens CF3 obtained average maximum axial stress and strain at 122 MPa and 2.84 %, respectively (Fig. 7a). The partially wrapped Specimens CP60 again had a lower compressive strength but higher axial strain as compared to those o 11

14 Specimens CF3. As shown in Fig. 7b, Specimens CP60 ailed at the average compressive strength o 116 MPa and axial strain o 3.25 %. The axial strain or the specimens CP60 increased by 14.33% in comparison with the Specimens CF3. As compared to Specimens CF3, the non-uniormly wrapped Specimens CP42 had both higher compressive strength and axial strain. Fig. 7d shows that Specimens CP42 ailed at the average compressive strength o 128 MPa and strain o 3.16 %. As a result, the compressive strength and axial strain o these specimens respectively increased by 5.29 % and % as compared to Specimens CF3. In order to compare the eectiveness o dierent wrapping schemes, the stress-strain curves o ive specimens are plotted in Fig. 7e. In reerence to this igure, it can be seen that the partially wrapped Specimens CP60 experienced a lower maximum stress and a higher maximum strain, as compared to Specimens CF3. On the hand, the non-uniormly wrapped specimens CP42 experienced both a higher maximum strain and stress in comparison with Specimens CF3. These indings have also been conirmed by specimens in Group GF2, as shown in Fig. 5d. 266 Analysis and Discussions 267 Lateral Strain The lateral strain o all the specimens are obtained by taking the average o readings rom three strain gages evenly placed along the FRP at locations away rom the overlap. For each specimen, the actual rupture strain o FRP is presented in Table 3. In order to investigate the eectiveness o the iber, the strain eiciency actor k is adopted, which is the ratio o the actual rupture strain o FRP in conined specimens and the rupture strain o the FRP obtained rom the tensile coupon testing. As can be seen rom Table 3, the strain eiciency actors o ully wrapped specimens are approximately 0.83 and 0.87 or glass iber and carbon iber, respectively. For glass iber, the strain eiciency actor o partially wrapped specimens was 12

15 and the corresponding number or non-uniormly wrapped specimens was Meanwhile, the strain eiciency actor o specimens partially wrapped with CFRP was 0.80 and the corresponding number or non-uniormly wrapped specimens was The experimental results have shown that the eectiveness o the iber reduces in the partial wrapping scheme, but increases in the non-uniormly wrapping scheme There is a consensus that the presence o the triaxial stress state in FRP aects the actual rupture strain o the iber (Chen et al. 2013). In this experimental program, it is obvious that the axial stress o the FRP jackets in the ully wrapped specimens is higher than that o the non-uniormly wrapped specimens. The discontinuity o the jacket in the non-uniormly wrapped specimens reduces the axial stress o the FRP jacket, which could be a reason or the increase in the strain eiciency actor in these specimens. Thus, the non-uniormly wrapped specimens had a higher value o k resulting in a higher conined strength and strain. In other words, the discontinuity o the jackets o the partially wrapped specimens did not increase the strain eiciency actor. The partially wrapped specimens experienced a dierent ailure mode as compared with the other wrapping schemes. This dierent ailure mode in partially wrapped specimens may be the reason behind the slight decrease in the strain eiciency actor or these specimens In addition, the lateral strain o the non-uniormly wrapped specimens at both the tie bands and cover bands o the FRP is investigated. For example, the lateral strain axial stress o Specimen CP40_3 (Fig. 8), illustrates that the lateral strain o FRP in a cover band is slightly higher than that o a tie band at any axial stress state. However, there was no dierence in the lateral strain in other specimens. 298 Analytical Veriication 13

16 In order to predict the compressive strength o the tested specimens, the procedure in the section Coninement Mechanism is used. It is noted that the actual lateral strain o each specimen was used in these calculations. The maximum axial strain o the tested specimens is predicted based on the study by Pham and Hadi (2013), in which the relationship between the energies absorbed by the whole column and the FRP was taken into account. Pham and Hadi (2013) assumed that the additional energy in the column core equals the area under the experimental stress-strain curves starting rom the value o unconined concrete strain: 306 U cc cc ( cc cd c co co )( 2 ' co ' cc ) (7) where U cc is the volumetric strain energy o conined concrete, c is the stress o conined concrete, and d c is an increment o the axial strain However, the concrete in the partially wrapped columns is conined in the eective area as shown in Fig. 1. To determine the volumetric strain energy o conined concrete or the whole columns, the value o the conined concrete strength needs to be modiied by the coninement eective coeicient (k e ), which leads to the ollowing equation: 313 U cc cc ( cc co )( cd c 2 co ' co k e ' cc ) (8) 314 Similarly, the energy absorbed by FRP could be calculated as ollows: 315 W 1 Ac ( e e ) (9) where W is the strain energy o FRP, and is the volumetric ratio o FRP as shown in Eq

17 t (10) D The compressive strain o columns partially wrapped with FRP is calculated as ollows: 320 cc co D 2kt ' co e e ' k e cc (11) The predicted results o the compressive strength and strain o the tested specimens are presented in Table 4. This table has shown that the predicted results are quite close to the experimental results. 324 Conclusions This study presented an experimental study on the optimization o concrete cylinders wrapped with FRP. The same amount o FRP was used in each group o specimens but with dierent wrapping schemes, in order to investigate the coninement eicacy between ully, partially and a proposed non-uniorm wrapping scheme or FRP-conined concrete. The indings presented in this study are summarized as ollows: For specimens belonging to the descending branch type, the partially wrapped specimens had a lower compressive strength but a higher strain as compared to the corresponding ully wrapped specimens. On the other hand, the non-uniorm wrapped specimens experienced both a higher compressive strength and axial strain in comparison with the ully wrapped specimens For heavily FRP-conined specimens (CF3, CP60, CP51 and CP42), partial and non- uniorm wrapped specimens provided a higher axial strain as compared to that o ully wrapped specimens. 15

18 The partial wrapping scheme changes the ailure modes o the specimens. I the FRP jackets are strong enough, the angle o the ailure surace signiicantly reduces The actual rupture strain o the FRP jackets is dierent or each wrapping scheme. The strain eiciency actor in the ull wrapping scheme is greater than that o the partial wrapping scheme but is less than that o the non-uniorm wrapping scheme An equation is proposed to estimate the axial strain o partially FRP-conined concrete circular columns Finally, this study proposed a new wrapping scheme that uses the same amount o FRP as compared to the conventional ully wrapping scheme, in order to yield a higher compressive strength and strain. However, urther studies are required to theoretically investigate the behavior o non-uniorm wrapped specimens. 349 Acknowledgement The authors would like to acknowledge the technical assistance o Messrs. Alan Grant, Cameron Neilson, Fernando Escribano, Ritchie McLean and Colin Devenish. The contribution o Mr. Elliot Davison is greatly appreciated. Furthermore, the irst author would like to thank the Vietnamese Government and the University o Wollongong or the support o his ull Ph.D. scholarship. 355 Notations A c A e D d c E = cross-sectional area; = area o eectively conined concrete core; = diameter o the column section; = increment o the axial strain; = elastic modulus o FRP; 16

19 c e cc co l l k k e s t U cc W w e cc co = stress o concrete; = actual rupture strength o FRP; = conined concrete strength; = unconined concrete strength; = eective conining pressure o a column; = equivalent conining pressure rom the FRP; = proportion actor; = coninement eective coeicient; = clear spacing between two FRP bands; = nominal thickness o FRP; = volumetric strain energy o conined concrete; = strain energy o FRP; = width o FRP bands; = actual rupture strain o FRP in hoop direction; = ultimate axial strain o conined concrete; = axial strain o the unconined concrete at the maximum stress; and = volumetric ratio o FRP. 378 Reerences ACI 440.2R-08 (2008). "Guide or the Design and Construction o Externally Bonded FRP Systems or Strengthening Concrete Structures." 440.2R-08, American Concrete Institute, Farmington Hills, MI. ASTM (2010). "Standard test method or tensile properties o iber reinorced polymer matrix composites used or strengthening o civil structures." D7565:2010West Conshohocken, PA. Australian Standard-1545 (1976). "Methods or the Calibration and Grading o Extenometers." Homebush, NSW Chaallal, O., Shahawy, M., and Hassan, M. (2003). "Perormance o axially loaded short rectangular columns strengthened with carbon iber-reinorced polymer wrapping." J Compos Constr, 7(3), Chen, J., Li, S., and Bisby, L. (2013). "Factors Aecting the Ultimate Condition o FRP- Wrapped Concrete Columns." J Compos Constr, 17(1), Colomb, F., Tobbi, H., Ferrier, E., and Hamelin, P. (2008). "Seismic retroit o reinorced concrete short columns by CFRP materials." Compos Struct, 82(4), De Luca, A., and Nanni, A. (2011). "Single-Parameter Methodology or the Prediction o the Stress-Strain Behavior o FRP-Conined RC Square Columns." J Compos Constr, 15(3),

20 ib (2001). "Externally bonded FRP reinorcement or RC structures." Bulletin, 14, 138. Hadi, M. N. S., Pham, T. M., and Lei, X. (2013). "New Method o Strengthening Reinorced Concrete Square Columns by Circularizing and Wrapping with Fiber-Reinorced Polymer or Steel Straps." J Compos Constr, 17(2), Lam, L., and Teng, J. G. (2003). "Design-oriented stress-strain model or FRP-conined concrete." Constr Build Mater, 17(6-7), Maaddawy, T. E. (2009). "Strengthening o Eccentrically Loaded Reinorced Concrete Columns with Fiber-Reinorced Polymer Wrapping System: Experimental Investigation and Analytical Modeling." J Compos Constr, 13(1), Mander, J. B., Park, R., and Priestley, M. J. N. (1988). "Theoretical Stress-Strain Model or Conined Concrete." J Struct Eng, 114(8), Pham, T. M., Doan, L. V., and Hadi, M. N. S. (2013). "Strengthening square reinorced concrete columns by circularisation and FRP coninement." Constr Build Mater, 49(0), Pham, T. M., and Hadi, M. N. S. (2013). "Strain Estimation o CFRP Conined Concrete Columns Using Energy Approach." J Compos Constr, 17(6), Pham, T. M., and Hadi, M. N. S. (2014a). "Coninement Model or FRP Conined Normaland High-Strength Concrete Circular Columns." Constr Build Mater, 69, Pham, T. M., and Hadi, M. N. S. (2014b). "Predicting Stress and Strain o FRP Conined Rectangular/Square Columns Using Artiicial Neural Networks." J Compos Constr, 18(6), Pham, T. M., and Hadi, M. N. S. (2014c). "Stress Prediction Model or FRP Conined Rectangular Concrete Columns with Rounded Corners." J Compos Constr, 18(1), Sheikh, S. A., and Uzumeri, S. M. (1980). "Strength and ductility o tied concrete columns." Journal o the structural division, 106(5), Smith, S. T., Kim, S. J., and Zhang, H. W. (2010). "Behavior and Eectiveness o FRP Wrap in the Coninement o Large Concrete Cylinders." J Compos Constr, 14(5), Tamužs, V., Tepers, R., Zīle, E., and Valdmanis, V. "Mechanical behaviour o FRP-conined concrete columns under axial compressive loading." Proc., 5th International engineering and construction conerence (IECC 5). American Society o Civil Engineers, International Committee, Los Angeles Section, Teng, J. G., Jiang, T., Lam, L., and Luo, Y. Z. (2009). "Reinement o a Design-Oriented Stress-Strain Model or FRP-Conined Concrete." J Compos Constr, 13(4), Toutanji, H. A. (1999). "Stress-strain characteristics o concrete columns externally conined with advanced iber composite sheets." ACI Mater J, 96(3), TR 55 (2012). Design guidance or strengthening concrete structures using ibre composite materials, Concrete Society, Camberley. Turgay, T., Polat, Z., Koksal, H. O., Doran, B., and Karakoç, C. (2010). "Compressive behavior o large-scale square reinorced concrete columns conined with carbon iber reinorced polymer jackets." Materials & Design, 31(1),

21 West System n.d. (2015). "Epoxy resins and hardeners - Physical properties." < (Jan. 31, 2015). Wu, Y. F., and Zhou, Y. W. (2010). "Uniied Strength Model Based on Hoek-Brown Failure Criterion or Circular and Square Concrete Columns Conined by FRP." J Compos Constr, 14(2),

22 List o Figures Figure 1. Coninement mechanism Figure 2. Dierent wrapping schemes Figure 3. Compressometer Figure 4. Failure modes o the tested specimens Figure 5. Stress-strain relation o Group GF2 Figure 6. Stress-strain relation o Group CF2 Figure 7. Stress-strain relation o Group CF3 Figure 8. Lateral strain axial stress relationship o Specimen CP40_3 20

23 List o Table Table 1. Test matrix Table 2. Results o tensile tests on lat coupon tests Table 3. Experimental results o tested specimens Table 4. Veriication o the experimental results 21

24 457 Table 1. Test matrix Group No. o specimens Type o FRP Equivalent FRP layers with ull wrapping Width o each FRP band (w, mm) Clear spacing (s, mm) Type o Wrapping R GF Full GP40 3 GFRP Partial GP Non-uniorm CF Full CP40 3 CFRP Partial CP Non-uniorm CF Full CP Partial CP51 3 CFRP Non-uniorm CP Non-uniorm

25 Table 2. Results o tensile tests on FRP lat coupons Type o Number Width Nominal Average Average Average coupon o FRP (mm) thickness Elastic Tensile Ultimate specimen layers (mm) Modulus Strength Strain (MN/mm) (kn/mm) (mm/mm) CFRP (75) * CFRP (25) ** GFRP * CFRP (75) denotes the coupons made o the FRP sheets that have 75 mm width ** CFRP (25) denotes the coupons made o the FRP sheets that have 25 mm width 23

26 Table 3. Experimental results o tested specimens Specimen Maximum axial stress Maximum axial strain ' cc Average Increase # (MPa) (MPa) (%) cc (%) Average (%) GF2_ Increase # (%) Maximum lateral strain l (%) Average (%) Strain eiciency actor 1.70 GF2_ GF2_ GP40_ GP40_ GP40_ GP31_ GP31_ GP31_ CF2_ CF2_ CF2_ CP40_ * CP40_ CP40_ CP31_ CP31_ CP31_ CF3_ CF3_ CF3_ CP60_ CP60_ CP60_ CP51_ CP51_ CP51_ * CP42_ CP42_ CP42_ * Specimens perormed premature damage # Increase o a specimen compared to the ully wrapping specimens in the same group. k 24

27 Table 4. Veriication o the experimental results Theoretical Experimental (*) Specimen D t s w k l k e (**) cc cc cc cc cc cc (mm) (mm) (mm) (mm) (MPa) (MPa) (%) (MPa) (%) (%) (%) CF CP CF CP cc and cc = dierence between the theoretical values and the corresponding experimental values (*) the values o k e were calculated based on Equation 4 (**) the values o cc were calculated based on Equation 5 25

28 Fig. 1

29 Fig. 2

30 Fig. 3 Compressometer LVDT

31 Fig. 4 4a.. GF2 44b. CP40 (σc = co ) 4d. CP60 4e. GP P40 4 CP40 (σc ~ cu ) 4c. P31 4. GP

32 Fig. 5 c (MPa) c (MPa) GF2_1 GF2_2 GF2_3 c (MPa) GP40_1 GP40_2 GP40_ GP31_ Reerence C GP31_2 ε GF2_3 ε c (%) c (%) c d GP31_3 GP40_2 GP31_ a c (MPa) ε c (%) ε c (%) b

33 Fig CF2_1 CF2_2 CF2_3 a CP40_1 CP40_2 CP40_3 b c (MPa) c (MPa) CP31_ CP31_2 ε c (%) c ε c (%) 100 CP31_3 c (MPa) ε c (%)

34 Fig. 7 c (MPa) c (MPa) c (MPa) CP60_1 CP60_2 CP60_ CP42_ Reerence CP42_2 CP60_2 e ε ε ε c (%) c d (%) c (%) 120 CP42_3 120 CP51_2 CP42_2 CF3_ CF3_1 CF3_2 CF3_3 c (MPa) a c (MPa) b CP51_1 CP51_2 CP51_3 c ε c (%) ε c (%)

35 Fig c (MPa) Tie band Cover band ε l (%)