1

1

2

3

Experimental Study of Concrete-Filled CHS Stub Columns 4

with Inner FRP Tubes 5

Yue-Ling Long1,4, Wen-Tao Li2, Jian-Guo Dai3* and Leroy Gardner4 6

7

8

9

10

11

12

13

14

1. Associate Professor, Department of Civil Engineering, Guangdong University of 15 Technology, Guangzhou 510006, China. 16

2. Postgraduate Student, Department of Civil Engineering, Guangdong University 17 of Technology, Guangzhou 510006, China. 18

3. Associate Professor (Corresponding author), Department of Civil and 19 Environmental Engineering, The Hong Kong Polytechnic University, Hong Kong, 20 China, Email: cejgdai@polyu.edu.hk 21

4. Department of Civil and Environmental Engineering, Imperial College London 22 SW7 2AZ, UK 23

24

Long, Y. L., Li, W. T., Dai, J. G. and Gardner, L. (2018). Experimental study of concrete-filled

CHS stub columns with inner FRP tubes. Thin-Walled Structures. 122, 606-621.

2

Abstract: An experimental study into the axial compressive behaviour of 25 concrete-filled circular hollow section (CHS) steel columns with internal fibre 26 reinforced polymer (FRP) tubes is presented in this paper. A total of 17 concrete-filled 27 steel tubular (CFST) columns were tested, 15 with an inner FRP tube and 2 with no 28 inner tube. Complementary material tests and tests on 15 FRP-confined concrete 29 (FCC) columns were also carried out. The varied test parameters included the 30 concrete strength, the ratio of the diameter of the steel tube to that of the FRP tube, the 31 diameter to wall thickness ratio of the inner FRP tube and the type (influencing 32 principally the rupture strain) of the FRP. It was found that the presence of the inner 33 FRP tube led to considerably improved axial compressive behaviour due to the greater 34 levels of confinement afforded to the ‘doubly-confined’ inner concrete core; the 35 load-bearing capacity was increased by between about 10 and 50% and the ductility 36 was also enhanced. Greater benefits arose with (1) increasing diameter of the inner 37 FRP tube due to the increased portion of the cross-section that is doubly-confined and 38 (2) increasing wall thickness of the inner FRP tube due to the increased level of 39 confinement afforded to the inner concrete core. The load-deflection responses of all 40 tested specimens were reported, revealing that failure was generally gradual with no 41 sharp loss in load-bearing capacity, implying that the embedment of the inner FRP 42 tube within the concrete enables it to continue to provide a reasonable degree of 43 confinement even after the initiation of fibre rupture; this is different to the sudden 44 loss of confinement typically observed in FRP externally jacketed concrete columns. 45 46

Keywords: 47

Axial compression; Composite; Concrete-filled steel tube; Ductility; Experiments; 48 Inner FRP tube; Load bearing capacity; Testing 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

3

66

1. Introduction 67 68

Concrete-filled steel tubular (CFST) columns are being extensively used in tall 69 buildings, long-span bridges and other mega structures due to their advantages such as 70 high strength, high ductility and large energy absorption capacity. In addition, the use 71 of CFST columns can bring convenience to construction due to the absence of 72 formwork. In the past few decades, a significant number of experimental and 73 analytical studies into the structural behaviour of CFST columns have been carried 74 out [1-17]. In these studies, it was shown that confinement to the concrete in circular 75 CFST columns brings substantial benefit in terms of load-bearing capacity, but that 76 this effect degrades dramatically upon yielding or local buckling of the steel tube. 77

In recent years, fibre reinforced polymer (FRP) jackets have been used to 78 enhance the structural performance of circular CFST columns by providing additional 79 confinement to the concrete and delaying the occurrence of local buckling of the steel 80 tube (Fig. 1(a)). Following the initial work of Xiao [18], a number of studies have 81 been carried out to investigate the behaviour of circular CFST columns externally 82 confined by FRP jackets [18-25]. Besides FRP-confined circular CFST columns, 83 another structural form of column featuring the combined use of FRP and steel tubes 84 is the FRP-concrete-steel double-skin tubular column (DSTC) as displayed in Fig. 85 1(b), which was originally proposed by Teng et al [26]. This column has an outer FRP 86 tube and an inner steel tube. A number of experimental studies has been carried out on 87 DSTCs by Teng et al. [27-31], Han et al. [32] and Ozbakkaloglu et al. [33-36]. The 88 above-described types of composite column have demonstrated that the combined use 89 of FRP jackets/tubes and steel tubes can offer substantially improved performance 90 over circular concrete columns. However, the external FRP jackets/tubes may not be 91 ideally suited to building construction due to limitations on their fire resistance arising 92 from the rapid degradation of the mechanical properties of FRP and possible smoke or 93 toxic gas generation during a fire. 94

This paper is concerned with a new type of circular CFST column with an inner 95 FRP tube, as shown in Fig. 1 (c). These cross-sections, referred to hereafter as 96 ‘FRP-CFST columns’, are expected to have the following distinct features: (1) the 97 inner FRP tube is expected to provide continuous and additional confinement to the 98 core concrete, further improving the ductility and strength of the columns even after 99 yielding of the steel tube; (2) failure of the FRP tubes will be less brittle due to their 100 embedment in concrete, avoiding a sudden loss of the FRP contribution upon rupture; 101 and (3) the existence of the inner FRP tube will restrict the lateral expansion of the 102 concrete, reducing the hoop strains and hence delaying yielding in the steel tube. 103

A limited number of tests have been performed on square CFST columns with 104 inner FRP tubes by Feng et al. [37-38], demonstrating higher ultimate strengths and 105 better ductility compared to conventional square CFST columns. However, no tests 106 have been conducted on circular FRP-CFST columns, as studied herein. The aim of 107

4

the present paper is therefore to study experimentally the axial compressive 108 behaviour of circular FRP-CFST columns and to advance the understanding of this 109 new structural form. 110

111

FRP jacket

Steel tube

Core concrete

FRP tube

Sandwiched concrete

Steel tube

Core concrete

112 (a) Circular CFST column confined by FRP jacket (b) FRP-concrete-steel double-skin tubular column 113

FRP tube

Steel tube

Core concrete

Sandwiched concrete

114 (c) Circular CFST column with an inner FRP tube 115

Fig. 1. Different forms of concrete-filled steel and FRP tubular columns 116 117 118

2. Experimental Programme 119 120

2.1 Test specimens 121 A total of 32 circular composite stub columns, including 15 FRP-CFST columns, 122

2 conventional CFST columns (without an inner FRP tube) and 15 FRP-confined 123 concrete (FCC) columns were manufactured and tested under axial compression. The 124 measured geometrical and material details of the 15 FRP-CFST specimens (depicted 125 in Fig. 2(a)) and the 2 CFST specimens (depicted in Fig. 2(b)) are listed in Table 1. In 126 the table, the specimens are divided into two groups according to the target concrete 127 strength, with the symbols “L” and “H” representing the lower and higher strength 128 concrete, respectively. The specimen designation system also describes the number of 129 tubes of each type (S = steel and F = FRP), as well as the diameter and thickness of 130 the FRP tube. For instance, specimen 1S1FH-100-2 represents a CFST specimen with 131 an outer steel tube (1S), an inner FRP tube (1F), high strength concrete, Df = 100 mm 132 and tf = 2 mm. Among the 15 FRP-CFST columns, 13 of the inner tubes were made 133 from glass FRP (GFRP) tubes and 2 from high rupture strain (HRS) FRP (i.e. 134 polyethylene terephthalate (PET) FRP and polyethylene naphthalate (PEN) FRP) 135 tubes. High rupture strain FRP composites usually possess a rupture strain greater 136 than 5% [39, 40] and have been recently studied as jacket material for reinforced 137 concrete members [41-44]. The fibres were oriented in the hoop direction resulting in 138 the FRP tubes having high hoop stiffness but low axial stiffness. 139

140

5

  Steel tube

Sandwiched concrete

FRP tube

Core concrete

Ds

Df

ts

tf

Steel tube

Concrete ts

Ds

141

(a) Cross-section of FRP-CFST specimens (b) Cross-section of CFST specimens 142  

FRP tube

Core concrete

tf

Df

143 (c) Cross-section of FCC specimens 144

Fig. 2. Cross-sections of (a) FRP-CFST, (b) CFST and (c) FCC specimens 145 146

All the FRP-CFST specimens and CFST specimens were 273 mm in diameter 147 and 820 mm in height. The height was chosen to be three times the specimen diameter 148 to avoid global buckling and end effects. The steel tubes used in all the specimens had 149 a nominal thickness of 6 mm, leading to diameter to thickness ratio Ds/ts of 45.5. The 150 following parameters were considered in the test programme: (i) concrete cylinder 151 compressive strength (i.e. normal strength f′co=36.5 MPa and high strength f′co=54.7 152 MPa), (ii) diameter of inner FRP tube (100 mm, 150 mm and 200 mm), resulting in 153 three different diameter ratios of outer steel tube to inner FRP tube (i.e. Ds/Df = 2.73, 154 1.82 and 1.37) and (iii) wall thickness of inner FRP tube (i.e. tf = 2 mm, 3 mm and 4 155 mm). As a result, diameter to wall thickness ratios of the inner FRP tube varied 156 between 25 and 100 (i.e. Df /tf =25, 33.3, 37.5, 50, 66.6, 75 and 100) as indicated in 157 Table 1. At the specific diameter of inner FRP tube of 150 mm and in the case of high 158 strength concrete, a range of FRP materials – GFRP, PET FRP and PEN FRP – were 159 employed to assess the influence of rupture strain on the column behaviour. All three 160 types of FRP were designed to have approximately the same tensile stiffness (i.e. Ef tf) 161 in the transverse direction. 162

In parallel to the above-mentioned 15 FRP-CFST column tests, 15 counterpart 163 FCC specimens with the same FRP tube diameter and tensile stiffness were also 164 prepared and tested under axial compression. The purpose of these complementary 165 tests was to investigate how the behaviour of the FRP tubes differs when acting as an 166 outer tube (FCC) and when embedded as an inner tube (FRP-CFST) within the 167 concrete. The FCC specimens are shown in Fig. 2(c). The specimen designation 168 follows the notation “0S1FL-Df-tf” and their key measured properties are given in 169 Table 2. 170

171 2.2 Material properties 172

Two separate batches of commercial self-compacting concrete with different 173 strengths were used to fill the test specimens. Three plain concrete cylinders (152.5 174

6

mm in diameter and 305 mm in height) were prepared and tested in compression 175 following ASTM C39/C39M [45] for each batch to determine the concrete strengths, 176 achieving average measured cylinder strengths of 36.5 MPa and 54.7 MPa at 28 days. 177 The stub column specimens were also tested at 28 days. The outer steel tubes were cut 178 from the same batch of two hot-rolled seamless steel tubes. For each tube, tensile tests 179 on three steel coupons were conducted in accordance with BS18 [46]. The average 180 yield stress from these coupon tests was 300 MPa, with little variation. 181

For the inner FRP tubes made of glass fibres, PET fibres and PEN fibres, hoop 182 tensile tests on 5 FRP rings for each type of FRP material were conducted following 183 ASTM D2290-08 [47]. The average measured ultimate tensile strength, hoop rupture 184 strain and elastic modulus of the FRP tubes obtained from these coupon tests are 185 summarized in Tables 1 and 2. 186

187

2.3 Preparation of FRP-CFST specimens 188 As shown in Fig. 3, a simple timber-steel composite formwork was developed to 189

fix the inner FRP tube and the outer steel tube in position (i.e. concentric) during 190 concrete casting. At the base, a wooden spacer was used to hold the bottom of the 191 steel tube in place and nails were used to maintain the position of the FRP tube 192 relative to the steel tube. At the top, two steel arms were used to maintain the 193 concentric position of the inner and outer tubes. The concrete was poured 194 simultaneously into both the FRP tube and the annular space between the steel tube 195 and the FRP tube to form the FRP-CFST columns. 196

Steel arm

Vertical steel rods

FRP tube

Steel tube

Steel arm

Vertical steel rods

197 Fig. 3. Support system employed to maintain position of tubes during casting of FRP-CFST 198

specimens 199 200

2.4 Instrumentation and testing 201 2.4.1 FRP-CFST & CFST specimens 202 As shown in Fig. 4, for each FRP-CFST or CFST specimen, six hoop strain 203 gauges (i.e. at 60° intervals) and six vertical strain gauges were uniformly placed on 204 the exterior surface of the steel tube at mid-height. For each FRP-CFST specimen, six 205

7

hoop strain gauges and four vertical strain gauges were also attached on the exterior 206 surface of the inner FRP tube. Two linear variable differential transducers (LVDTs) 207 (LVDT1 and LVDT2) placed on opposite sides of the specimen, as shown in Fig. 4, 208 were installed between the top and bottom platens of the testing machine to monitor 209 the overall specimen shortening and to ensure that uniform compression was applied 210 the specimen. Three other LVDTs (LVDT3, LVDT4 and LVDT5) placed at 120° 211 intervals were employed to measure axial shortening overall the central 500 mm 212 height of the test specimens to avoid the influence of local end effects, as shown in 213 Fig. 4. 214

160

500

160

Universal joint

LVDT

End platen

Axial load Strain gauge to measure vertical strainStrain gauge to measure hoop strain

Steel tube

FRP tube60°

LVDT1LVDT2

LVDT3LVDT4

LVDT5 215 Fig. 4. FRP-CFST specimen and instrumentation (dimensions in mm) 216

217 All the FRP-CFST and CFST specimens were tested using a 10000 kN universal 218

loading machine under displacement control. The loading rate was constant at 0.5 219 mm/min. All test data were recorded using a data logger at 1.0 second intervals. The 220 loading process of the FRP-CFST and CFST specimens was terminated manually 221 when the overall axial displacement reached around 85 mm to 100 mm. A photograph 222 of the test setup is shown in Fig. 5. 223

224

Fig. 5. Test setup 225

8

226 2.4.2 FCC specimens 227

For each FCC specimen, six hoop strain gauges and six vertical strain gauges 228 were placed on the exterior surface of the FRP tube wall at mid-height, as indicated in 229 Fig. 6. Three LVDTs placed at 120° intervals were employed to monitor the axial 230 deformation of the specimens within the central 2/3 height. Each FCC and its 231 FRP-CFST counterpart were tested on the same day. 232

233

Universal joint

LVDT

End platen

Axial load Strain gauge to measure vertical strainStrain gauge to measure hoop strain

FRP tube

60°

LVDT1 LVDT2

LVDT3

2L/3

L/6

L/6

234 Fig. 6. FCC specimen and instrumentation 235

236

3. Experimental Results and Discussion 237 In this section, the failure modes, load-deformation responses, ultimate loads and 238

ductility of the tested specimens are reported. The results and the influence of the key 239 varied parameters are also analysed and discussed. 240

241 3.1 Load-axial deformation responses 242

The load-axial deformation responses of the tested FRP-CFST, CFST and FCC 243 specimens are presented in Fig. 7, where the axial deformation is presented in 244 normalised form (i.e. average axial strain). The average axial strain was calculated by 245 normalising the average axial displacement from the central LVDTs (see Figs 4 and 6) 246 by the lengths over which they recorded, equal to 500 mm for the FRP-CFST and 247 CFST specimens and 2/3L for the FCC specimens. It can be seen from Fig. 7 that the 248 load-bearing capacity Nue,FRP-CFST of all the tested FRP-CFST specimens was greater 249 than that of the counterpart CFST specimens Nue,CFST, indicating the positive 250 contribution of the inner FRP tube (see Fig. 7(a) and Fig. 7(b)). Furthermore, it can 251 also be observed that the residual load-bearing capacity Nre (defined as the minimum 252 recorded post-peak resistance) of all FRP-CFSTs except Specimen 1S1FH-150-3 were 253 greater than the peak loads of the corresponding CFSTs. 254

9

The key results from the FRP-CFST and CFST stub column tests are summarised 255 in Table 3, including the ultimate load-bearing capacity Nue (Nue,FRP-CFST for the 256 FRP-CFST specimens and Nue,CFST for the CFST specimens), the average axial strain 257 εau corresponding to Nue, the approximate axial load at which local buckling occurred 258 Nle, which is also illustrated in Fig. 8, the average axial strain εal corresponding to Nle, 259 the residual load-bearing capacity Nre and the Nre/Nue ratio. Note that the load at which 260 local buckling occurred was determined approximately by both monitoring of the 261 axial and hoop strains in the steel tube and by visual observation. The key test results 262 of FCC specimens, including the load-bearing capacity Nue (Nue,FCC specifically) and 263 the axial strain εau and hoop strain εhu corresponding to Nue, are summarised in Table 4. 264 265

0 50000 100000 1500000

2000

4000

6000

8000

10000 1S0FL 1S1FL-150-3 1S1FL-100-2 1S1FL-150-4 1S1FL-100-3 1S1FL-200-2 1S1FL-100-4 1S1FL-200-3 1S1FL-150-2 1S1FL-200-4

Average axial strain ()

Axi

al lo

ad (

kN)

0 50000 100000 150000

0

2000

4000

6000

8000

10000 1S0FH 1S1FH-200-4 1S1FH-150-3 1S1TH-200-30 1S1FH-150-4 1S1NH-200-18 1S1FH-200-3

Average axial strain ()

Axi

al lo

ad (

kN)

266 (a) FRP-CFST & CFST specimens (f'co=36.5 MPa) (b) FRP-CFST & CFST specimens (f'co=54.7 267

MPa) 268

0 20000 40000 600000

2000

4000

6000 0S1FL-100-2 0S1FL-150-4 0S1FL-100-3 0S1FL-200-2 0S1FL-100-4 0S1FL-200-3 0S1FL-150-2 0S1FL-200-4 0S1FL-150-3

Axi

al lo

ad (

kN)

Average axial strain ()

0 20000 40000 600000

2000

4000

6000 0S1FH-150-3 0S1FH-200-4 0S1FH-150-4 0S1TH-200-30 0S1FH-200-3 0S1NH-200-18

Axi

al lo

ad (

kN)

Average axial strain() 269 (c) FCC specimens ( f'co=36.5 MPa) (d) FCC specimens (f'co=54.7 MPa) 270

Fig. 7. Load-average axial strain curves of FRP-CFST, CFST and FCC specimens 271

272 3.2 Influence of key varied parameters on load-axial deformation responses 273 The influence of the key varied parameters on the load-axial deformation response of 274 the tested FRP-CFST and CFST specimens is assessed in Fig. 8. The influence of the 275 thickness of the FRP tube tf on the load-axial strain curves is shown in Figs 8(a) to (e), 276 where it may be seen that increasing tf leads to delayed local buckling and 277 substantially higher ultimate loads and corresponding deformations (Nue and εau, 278 respectively). This is attributed to the greater confinement afforded to the inner 279 concrete core. The influence of the diameter of the FRP tube Df is shown in Figs 280

10

8(f)-(j), where it may be seen that increasing Df results in higher ultimate loads, 281 despite local buckling generally occurring earlier; this is attributed to the increased 282 area that benefits from the ‘double-confinement’ (i.e. from both the outer steel tube 283 and the inner FRP tube) as the diameter of the FRP tube is increased. The effect on the 284 average axial strain at ultimate load εau is marginal. 285

The influence of the steel-to-FRP tube diameter ratio Ds/Df on the load-axial 286 strain curves is shown in Figs 8 (k)-(l), in which the Df/tf ratios are constant. From the 287 figures, it may be seen that decreasing the Ds/Df ratio has little influence on the 288 occurrence of local buckling in the steel tube or the value of εau, but results in 289 significant increases in load-bearing capacity, due to a greater portion of the concrete 290 being doubly-confined. 291

The influence of the type of FRP material employed for the inner tube is assessed 292 in Fig. 8(m). Three FRP-CFST stub columns (1S1FH-200-3, 1S1FH-200-30 and 293 1S1FH-200-18) with the inner FRP tubes made of three different fibre materials (i.e., 294 GFRP, PET and PEN fibres) but with approximately the same tensile stiffness in the 295 hoop direction (i.e.EFRPtf) were compared. It is seen that the ultimate load-bearing 296 capacity Nue of the FRP-CFST stub column with the inner GFRP tube (i.e., 297 1S1FH-200-3) was slightly higher than that of the other two specimens, but the 298 average axial strain at failure εau was significantly lower due to earlier rupture of the 299 FRP tube. These preliminary results indicate that the ductility, but not necessarily the 300 load-bearing capacity, benefits from the use of high rupture strain FRP for the inner 301 tube in the investigated system. 302

303

0 20000 40000 60000 80000 1000000

2000

4000

6000

8000

1S0FL 1S1FL-100-2 (tf=2mm)

1S1FL-100-3 (tf=3mm)

1S1FL-100-4 (tf=4mm)

Average axial strain ()

Axi

al lo

ad (

kN)

Indicates local buckling

0 20000 40000 60000 80000 1000000

2000

4000

6000

8000

1S0FL 1S1FL-150-2 (tf=2mm)

1S1FL-150-3 (tf=3mm)

1S1FL-150-4 (tf=4mm)

Average axial strain ()

Axi

al lo

ad (

kN)

Indicates local buckling

304 (a) Influence of tf (Df=100 mm, f'co=36.5 MPa) (b) Influence of tf (Df=150 mm, f'co=36.5 MPa) 305

306

11

0 20000 40000 60000 80000 1000000

2000

4000

6000

8000

1S0FL 1S1FL-200-2 (tf=2mm)

1S1FL-200-3 (tf=3mm)

1S1FL-200-4 (tf=4mm)

Average axial strain ()

Axi

al lo

ad (

kN)

Indicates local buckling

0 20000 40000 60000 80000 1000000

2000

4000

6000

8000

1S0FH 1S1FH-150-3 (t f=3mm)

1S1FH-150-4 (t f=4mm)

Average axial strain ()

Axi

al lo

ad (

kN)

Indicates local buckling

307 (c) Influence of tf (Df=200 mm, f'co=36.5 MPa) (d) Influence of tf (Df=150 mm, f'co=54.7 MPa) 308

309

0 20000 40000 60000 80000 1000000

2000

4000

6000

8000

1S0FH 1S1FH-200-3 (t f=3mm)

1S1FH-200-4 (t f=4mm)

Average axial strain ()

Axi

al lo

ad (

kN)

Indicates local buckling

0 20000 40000 60000 80000 1000000

2000

4000

6000

8000

1S0FL 1S1FL-100-2 1S1FL-150-2 1S1FL-200-2

Average axial strain ()

Axi

al lo

ad (

kN)

Indicates local buckling

310 (e) Influence of tf (Df=200 mm, f'co=54.7 MPa) (f) Influence of Df (tf=2 mm, f'co=36.5 MPa) 311

312 313

0 20000 40000 60000 80000 1000000

2000

4000

6000

8000

1S0FL 1S1FL-100-3 1S1FL-150-3 1S1FL-200-3

Average axial strain ()

Axi

al lo

ad (

kN)

Indicates local buckling

0 20000 40000 60000 80000 1000000

2000

4000

6000

8000

1S1FL 1S1FL-100-4 1S1FL-150-4 1S1FL-200-4

Average axial strain ()

Axi

al lo

ad (

kN)

Indicates local buckling

314 (g) Influence of Df (tf=3 mm, f'co=36.5 MPa) (h) Influence of Df (tf=4 mm, f'co=36.5 MPa) 315

316

12

0 20000 40000 60000 80000 1000000

2000

4000

6000

8000

1S0FH 1S1FH-150-3 1S1FH-200-3

Average axial strain ()

Axi

al lo

ad (

kN)

Indicates local buckling

0 20000 40000 60000 80000 1000000

2000

4000

6000

8000

1S0FH 1S1FH-150-4 1S1FH-200-4

Average axial strain ()

Axi

al lo

ad (

kN)

Indicates local buckling

317 (i) Influence of Df (tf=3 mm, f'co=54.7 MPa) (j) Influence of Df (tf=4 mm, f'co=54.7 MPa) 318

0 20000 40000 60000 80000 1000000

2000

4000

6000

8000

1S1FL-100-2 (Ds/Df=2.73)

1S1FL-150-3 (Ds/Df=1.82)

1S1FL-200-4 (Ds/Df=1.37)

Axi

al lo

ad (

kN)

Average axial strain ()

Indicates local buckling

0 20000 40000 60000 80000 1000000

2000

4000

6000

8000

1S1FH-150-3 (Ds/Df=1.82)

1S1FH-200-4 (Ds/Df=1.37)

Axi

al lo

ad (

kN)

Average axial strain ()

Indicates local buckling

319 (k) Influence of Ds/Df (Df/tf=50, f'co=36.5 MPa) (l) Influence of Ds/Df (Df/tf=50, f'co=54.7 MPa) 320

321

0 30000 60000 90000 120000 1500000

2000

4000

6000

8000

1S1FH-200-3 (GFRP) 1S1TH-200-30 (PET FRP) 1S1NH-200-18 (PEN FRP)

Axi

al lo

ad (

kN)

Average axial strain ()

Indicates local buckling

322 (m) Influence of FRP type (Df=200 mm, f'co=54.7 MPa) 323

Fig. 8. Influence of key parameters on the load-axial deformation responses of FRP-CFST and CFST 324 specimens 325

326

The influence of concrete strength on the load-axial deformation response of the 327 tested FRP-CFST and CFST stub columns is assessed in Fig. 9, in which the axial 328 load has been normalised by the corresponding peak load of each specimen. The 329 comparisons highlight the earlier loss of stiffness, but more ductile response of the 330 tubes filled with the lower strength concrete. 331

332

13

0 20000 40000 60000 80000 1000000.0

0.2

0.4

0.6

0.8

1.0

1S0FL 1S0FH

Axi

al lo

ad (

N/N

ue)

Average axial strain ()

0 20000 40000 60000 80000 1000000.0

0.2

0.4

0.6

0.8

1.0

1S1FL-150-3 1S1FH-150-3

Axi

al lo

ad (

N/N

ue)

Average axial strain () 333 (a) Without inner FRP tube (b) Df=150 mm, tf=3 mm 334

335

0 20000 40000 60000 80000 1000000.0

0.2

0.4

0.6

0.8

1.0

1S1FL-150-4 1S1FH-150-4

Axi

al lo

ad (

N/N

ue)

Average axial strain ()

0 20000 40000 60000 80000 1000000.0

0.2

0.4

0.6

0.8

1.0

1S1FL-200-3 1S1FH-200-3

Axi

al lo

ad (

N/N

ue)

Average axial strain () 336 (c) Df=150 mm, tf=4 mm (d) Df=200 mm, tf=3 mm 337

338

0 20000 40000 60000 80000 1000000.0

0.2

0.4

0.6

0.8

1.0

1S1FL-200-4 1S1FH-200-4

Axi

al lo

ad (

N/N

ue)

Average axial strain () 339 (e) Df=200 mm, tf=4 mm 340

Fig. 9. Influence of concrete strength on the load-axial strain response of tested FPR-CFST and CFST 341 specimens 342

343 3.3 Contribution of inner FRP tube to axial capacity 344

The contribution of the inner FRP tube to the axial load-bearing resistance of the 345 FRP-CFST stub columns is assessed in this sub-section by examination of the 346 load-axial strain and load-hoop strain responses of the inner FRP tubes in the 347 FRP-CFST specimens and their FCC counterparts. The results are shown in Fig. 10. It 348 may be seen that the peak hoop strain (i.e. corresponding to Nue) in the inner FRP tube 349 of the FRP-CFST specimens was generally less than that in the corresponding FCC 350

14

specimens, indicating that the material strength of the inner FRP tube in the 351 FRP-CFST sections may not have been fully utilized. This is consistent with the 352 failure mode of inner FRP tubes described in Section 3.4, and may also explain the 353 higher residual load-bearing capacity achieved in the FRP-CFST specimens with an 354 inner FRP tube. The lower hoop strains in the inner FRP tubes of the FRP-CFST 355 specimens than the FCC tubes also implies that the load-bearing capacity of the 356 FRP-CFST cross-sections will not be accurately predicted based on the 357 superimposition of the CFST and FCC specimen resistances, as discussed further in 358 Section 3.5. 359

360

0 30000 60000 90000 1200000

2000

4000

6000

8000

10000

Axial strain ()

-15000 -10000 -5000 0

Axi

al lo

ad (

kN)

0S1FL-100-2 Inner tube of 1S1FL-100-2

Hoop strain () 0 30000 60000 90000 120000

0

2000

4000

6000

8000

10000

Axial strain ()

-20000 -15000 -10000 -5000 0

0S1FL-100-3 Inner tube of 1S1FL-100-3

Axi

al lo

ad (

kN)

Hoop strain () 361

(a) 1S1FL-100-2 and 0S1FL-100-2 (b) 1S1FL-100-3 and 0S1FL-100-3 362

0 30000 60000 90000 1200000

2000

4000

6000

8000

10000 0S1FL-100-4 Inner tube of 1S1FL-100-4

Axial strain ()

-20000 -15000 -10000 -5000 0

Axi

al lo

ad (

kN)

Hoop strain ()

0 30000 60000 90000 1200000

2000

4000

6000

8000

10000 0S1FL-150-2 Inner tube of 1S1FL-150-2

Axial strain ()

-15000 -10000 -5000 0

Axi

al lo

ad (

kN)

Hoop strain () 363

(c) 1S1FL-100-4 and 0S1FL-100-4 (d) 1S1FL-150-2 and 0S1FL-150-2 364

0 30000 60000 90000 1200000

2000

4000

6000

8000

10000 0S1FL-150-3 Inner tube of 1S1FL-150-3

Axial strain ()

-15000 -10000 -5000 0

Axi

al lo

ad (

kN)

Hoop strain () 0 30000 60000 90000 120000

0

2000

4000

6000

8000

10000 0S1FL-150-4 Inner tube of 1S1FL-150-4

Axial strain ()

-20000 -15000 -10000 -5000 0

Axi

al lo

ad (

kN)

Hoop strain () 365

(e) 1S1FL-150-3 and 0S1FL-150-3 (f) 1S1FL-150-4 and 0S1FL-150-4 366

15

0 30000 60000 90000 1200000

2000

4000

6000

8000

10000 0S1FL-200-2 Inner tube of 1S1FL-200-2

Axial strain ()-15000 -10000 -5000 0

Axi

al lo

ad (

kN)

Hoop strain () 0 30000 60000 90000 120000

0

2000

4000

6000

8000

10000

Axial strain ()-15000 -10000 -5000 0

0S1FL-200-3 Inner tube of 1S1FL-200-3

Axi

al lo

ad (

kN)

Hoop strain () 367

(g) 1S1FL-200-2 and 0S1FL-200-2 (h) 1S1FL-200-3 and 0S1FL-200-3 368

0 30000 60000 90000 1200000

2000

4000

6000

8000

10000 0S1FL-200-4 Inner tube of 1S1FL-200-4

Axial strain ()

-15000 -10000 -5000 0

Axi

al lo

ad (

kN)

Hoop strain ()

0 30000 60000 900000

2000

4000

6000

8000

10000 0S1FH-150-3 Inner tube of 1S1FH-150-3

Axial strain ()

-15000 -10000 -5000 0

Axi

al lo

ad (

kN)

Hoop strain () 369

(i) 1S1FL-200-4 and 0S1FL-200-4 (j) 1S1FH-150-3 and 0S1FH-150-3 370

0 30000 60000 90000

0

2000

4000

6000

8000

10000 0S1FH-150-4 Inner tube of 1S1FH-150-4

Axial strain ()

-15000 -10000 -5000 0

Axi

al lo

ad (

kN)

Hoop strain ()

0 30000 60000 900000

2000

4000

6000

8000

10000

Axial strain ()

-15000 -10000 -5000 0

Axi

al lo

ad (

kN)

0S1FH-200-3 Inner tube of 1S1FH-200-3

Hoop strain () 371

(k) 1S1FH-150-4 and 0S1FH-150-4 (l) 1S1FH-200-3 and 0S1FH-200-3 372 373

0 30000 60000 900000

2000

4000

6000

8000

10000 0S1FH-200-4 Inner tube of 1S1FH-200-4

Axial strain ()-15000 -10000 -5000 0

Axi

al lo

ad (

kN)

Hoop strain ()

0 50000 100000 1500000

2000

4000

6000

8000

10000

Axial strain ()

-60000 -40000 -20000 0

0S1TH-200-30 Inner tube of 1S1TH-200-30

Axi

al lo

ad (

kN)

Hoop strain () 374

(m) 1S1FH-200-4 and 0S1FH-200-4 (n) 1S1FH-200-30 and 0S1FH-200-30 375 376

16

377

0 30000 60000 90000 1200000

2000

4000

6000

8000

10000

Axial strain ()

-60000 -40000 -20000 0

Axi

al lo

ad (

kN)

0S1NH-200-18 Inner tube of 1S1NH-200-18

Hoop strain () 378

(o) 1S1FH-200-18 and 0S1FH-200-18 379 Fig. 10. Load-axial/hoop strain curves for the FRP tubes in the FRP-CFST and FCC specimens 380

381 The hoop strain-axial strain relationship for the inner FRP tube in each of the 382

tested FRP-CFST specimens is compared with that of the corresponding FCC in Fig. 383 11. It may be that the hoop strain development in the inner FRP tube of the 384 FRP-CFST specimens was consistently slower than that in the corresponding FCC 385 specimen for a given axial strain. This implies that the concrete in the FRP tubes of 386 FRP-CFST and FCC specimens may exhibit different dilation properties, which will 387 influence the load-carrying capacity. This finding is in line with observations from 388 previous studies [41, 48] that the dilation of confined concrete (reflected by the ratio 389 of the hoop strain to the axial strain) is related to the lateral confinement stiffness [41, 390 48]. 391

0 10000 20000 30000 40000 50000 60000 700000

-5000

-10000

-15000

-20000 0S1FL-100-2 Inner tube of 1S1FL-100-2

Axial strain ()

Hoo

p st

rain

(

)

0 10000 20000 30000 40000 50000 60000 700000

-5000

-10000

-15000

-20000 0S1FL-100-3 Inner tube of 1S1FL-100-3

Axial strain ()

Hoo

p st

rain

(

)

392

(a) 1S1FL-100-2 and 0S1FL-100-2 (b) 1S1FL-100-3 and 0S1FL-100-3 393

0 10000 20000 30000 40000 50000 60000 700000

-5000

-10000

-15000

-20000 0S1FL-100-4 Inner tube of 1S1FL-100-4

Axial strain ()

Hoo

p st

rain

(

)

0 10000 20000 30000 40000 50000 60000 700000

-5000

-10000

-15000

-20000 0S1FL-150-2 Inner tube of 1S1FL-150-2

Axial strain ()

Hoo

p st

rain

(

)

394

(c) 1S1FL-100-4 and 0S1FL-100-4 (d) 1S1FL-150-2 and 0S1FL-150-2 395

17

0 10000 20000 30000 40000 50000 60000 700000

-5000

-10000

-15000

-20000 0S1FL-150-3 Inner tube of 1S1FL-150-3

Axial strain ()

Hoo

p st

rain

(

)

0 10000 20000 30000 40000 50000 60000 700000

-5000

-10000

-15000

-20000 0S1FL-150-4 Inner tube of 1S1FL-150-4

Axial strain ()

Hoo

p st

rain

(

)

396

(e) 1S1FL-150-3 and 0S1FL-150-3 (f) 1S1FL-150-4 and 0S1FL-150-4 397

0 10000 20000 30000 40000 50000 60000 700000

-5000

-10000

-15000

-20000 0S1FL-200-2 Inner tube of 1S1FL-200-2

Axial strain ()

Hoo

p st

rain

(

)

0 10000 20000 30000 40000 50000 60000 700000

-5000

-10000

-15000

-20000 0S1FL-200-3 Inner tube of 1S1FL-200-3

Axial strain ()H

oop

stra

in (

) 398

(g) 1S1FL-200-2 and 0S1FL-200-2 (h) 1S1FL-200-3 and 0S1FL-200-3 399

0 10000 20000 30000 40000 50000 60000 700000

-5000

-10000

-15000

-20000 0S1FL-200-4 Inner tube of 1S1FL-200-4

Axial strain ()

Hoo

p st

rain

(

)

0 10000 20000 30000

0

-5000

-10000

-15000

-20000 0S1FH-150-3 Inner tube of 1S1FH-150-3

Axial strain ()

Hoo

p st

rain

(

)

400

(i) 1S1FL-200-4 and 0S1FL-200-4 (j) 1S1FH-150-3 and 0S1FH-150-3 401

0 10000 20000 30000

0

-5000

-10000

-15000

-20000 0S1FH-150-4 Inner tube of 1S1FH-150-4

Axial strain ()

Hoo

p st

rain

(

)

0 10000 20000 30000

0

-5000

-10000

-15000

-20000 0S1FH-200-3 Inner tube of 1S1FH-200-3

Axial strain ()

Hoo

p st

rain

(

)

402

(k) 1S1FH-150-4 and 0S1FH-150-4 (l) 1S1FH-200-3 and 0S1FH-200-3 403

0 10000 20000 30000

0

-5000

-10000

-15000

-20000 0S1FH-200-4 Inner tube of 1S1FH-200-4

Axial strain ()

Hoo

p st

rain

(

)

0 50000 100000 150000

0

-20000

-40000

-60000

-80000 0S1TH-200-30 Inner tube of 1S1TH-200-30

Axial strain ()

Hoo

p st

rain

(

)

404

405 (m) 1S1FH-200-4 and 0S1FH-200-4 (n) 1S1FH-200-30 and 0S1FH-200-30 406

18

0 20000 40000 60000 80000 1000000

-10000

-20000

-30000

-40000

-50000 0S1NH-200-18 Inner tube of 1S1NH-200-18

Axial strain ()

Hoo

p st

rain

(

)

407

(o) 1S1FH-200-18 and 0S1FH-200-18 408 Fig. 11. Hoop strain-axial strain curves for the FRP tube in the FRP-CFST and FCC specimens 409

410 3.4 Failure modes 411

Typical failures modes for the tested CFST and FRP-CFST stub columns are 412 shown in Fig. 12. Both specimen types exhibited outward only local buckling of the 413 steel tube, but this occurred at considerably lower strains (see Fig. 8) for the CFST 414 specimens (Fig. 12(a)) than the FRP-CFST specimens (Figs 12(b) to (d)). Slightly 415 different patterns of local buckling may be seen for differing FRP-CFST cross-section 416 geometries and material properties, as shown in Figs 12(b) to (d). The delayed local 417 buckling in the FRP-CFST specimens is attributed to the reduced transverse 418 expansion of the concrete core due to the additional confining effect of the inner FRP 419 tube. Transverse expansion expediates yielding of the steel tube and cracking of the 420 concrete, and both of these effects promote local buckling of the steel tube. 421

422

423 (a) CFST Specimen 1S0FL (b) FRP-CFST specimen 1S1FL-150-3 424 425

19

426 (c) FRP-CFST specimen 1S1FL-150-4 (d) FRP-CFST specimen 1S1TH-200-30 427

Fig. 12. Typical failure modes of CFST and FRP-CFST specimens 428

429 At the end of the tests, the outer steel tubes of the FRP-CFST specimens were 430

opened using a cutter to expose the concrete and inner FRP tube, as shown in Fig. 431 13(a). The infilled concrete was typically found to be severely crushed, while the FRP 432 tubes were seen to have localised/ partial rupturing, generally adjacent to the point of 433 local buckling of the steel tube – see Fig. 13(a). For the FCC specimens, full rupture 434 of the FRP tubes was observed – see Fig. 13(b). 435

436

437 (a) FRP-CFST specimen (b) FCC specimen 438

Fig. 13. Typical failure modes of FRP-CFST and FCC specimens 439

440 3.5 Ultimate load-carrying capacity of FRP-CFST cross-sections 441

The ultimate load-bearing capacities of the tested FRP-CFST stub columns 442 Nue,FRP-CFST and those of the corresponding CFST stub columns with no inner FRP 443 tube Nue,CFST are reported in Table 3. The Nue,FRP-CFST/Nue,CFST ratios are also presented 444 to demonstrate the strength benefit arising from the addition of the inner FRP tube, 445 which may be seen to have ranged between 12% and 51%. As discussed in Section 3.2, 446 the greatest benefits were achieved when the area of the doubly-confined concrete 447 was increased (i.e. when the Ds/Df ratio was reduced) and when the level of 448 confinement to the inner core was increased (i.e. when the thickness of the FRP tube 449

20

was increased). 450 The cross-section load-bearing performance relative to the full plastic 451

compression resistance of the tested specimens can be assessed through a strength 452 index (SI), as employed by Han [49] and McCann et al.[50]. The strength index is 453 expressed as follows: 454

 

Rdpl,

ueSIN

N

(1) 455

where Npl,Rd is the plastic compressive resistance of the column cross-section, which 456 is defined in EN 1994-1-1 [51], in the absence of steel reinforcement, as: 457

  'cocyaRdpl, fAfAN

(2) 458

in which Aa and Ac are the cross-sectional areas of the steel tube and concrete, 459 respectively, and fy and f’co are the yield strength of steel tube and cylinder strength of 460 the concrete, respectively. 461

The calculated SI values for the tested FRP-CFST specimens are listed in Table 3 462 and plotted against the Df/Ds ratio in Fig.14. It may be seen that the strength index of 463 the FRP-CFST specimens increased with (1) increasing Df/Ds ratio due to the 464 increased portion of the cross-section that is doubly-confined and (2) increasing wall 465 thickness of the inner FRP tube due to the increased level of confinement afforded to 466 the inner concrete core. 467

0.0 0.2 0.4 0.6 0.81.0

1.5

2.0

2.5

Df/D

s

tf=2 mm (f 'co=36.5MPa) No inner tube (f '

co=54.7MPa)

tf=3 mm (f 'co=36.5MPa) tf=3 mm (f '

co=54.7MPa)

tf=4 mm (f 'co=36.5MPa) tf=4 mm (f '

co=54.7MPa)

No inner tube (f 'co=36.5MPa)

SI

468 Fig. 14. Comparison of strength indices for FRP-CFST and CFST specimens 469

470 There are three main contributory components to the axial load-carrying capacity 471 of FRP-CFST cross-sections: (1) the steel tube, (2) the sandwiched concrete between 472 the FRP tube and the steel tube and (3) the doubly-confined inner concrete core, as 473 illustrated in Fig. 15. Similarly, the components contributing to the axial load-bearing 474 capacity of CFST and FCC cross-sections are also illustrated in Fig. 15. 475

It would be anticipated that the ultimate axial load-bearing capacity of the 476 FRP-CFST cross-sections Nue,FRP-CFST could be accurately predicted as the sum of the 477 ultimate load-bearing capacities of the corresponding CFST cross-section (Nue,CFST) 478 and FCC (Nue,FCC) cross-section, minus (to avoid double-counting) the contribution of 479

21

the FRP tube and the inner concrete core i.e. (Accf′co+Aff′co) where Acc is 480 cross-sectional area of the inner core concrete, Af is the cross-sectional area of the 481 FRP tube and f′co is concrete compressive strength. However, while this was found to 482 provide reasonable results, the predicted values were in fact consistently larger than 483 the experimental values Nue,FRP-CFST for all tested FRP-CFST cross-sections, indicating 484 that the material strength of the inner FRP tube was not fully utilized in the 485 FRP-CFST specimens. Hence, a reduction coefficient φ is proposed, as defined by Eq. 486 (3): 487

 '

cof'

coccFCCue,CFSTue,

CFST-FRPue,

fAfANN

N

(3) 488

+ +

+CFST column

FRP-CFST column

FCC column

Steel tube Sandwiched concrete Core concreteConfinement by the inner FRP tubeConfinement by the steel tube

489

Fig. 15. Illustration of contributory components to axial load-carrying capacities of FRP-CFST, CFST 490 and FCC columns 491

492

The calculated values of φ for all the tested FRP-CFST specimens are listed in 493 Table 3, and may be seen to vary between 0.8 and 1.0 depending on the properties of 494 inner FRP tubes; for the GFRP inner tubes, φ varied between 0.92 and 1.0. The values 495 of φ are also plotted against the Df/Ds ratio for the FRP-CFST specimens with GFRP 496 inner tubes in Fig. 16, where it may be seen that there is a trend of increasing φ with 497 increasing Df /Ds ratios, while the influence of tf is less distinct. Further research is 498 required to develop a more in-depth understanding of the above phenomena, but as a 499 preliminary recommendation, a safe-sided value of φ = 0.9 appears appropriate for 500 FRP-CFST cross-sections with GFRP inner tubes, while a value of φ = 0.8 appears 501 suitable for FRP-CFST cross-sections with HRS FRP inner tubes. 502

503

22

0.0 0.2 0.4 0.6 0.8

0.90

0.95

1.00

1.05

D

f/D

s

t f=2 mm (f 'co=36.5MPa) t f=3 mm (f '

co=54.7MPa)

t f=3 mm (f 'co=36.5MPa) t f=4 mm (f '

co=54.7MPa)

t f=4 mm (f 'co=36.5MPa)

504 Fig. 16. Variation of reduction coefficient φ with Df/Ds for tested FRP-CFST specimens 505

506 3.6 Ductility of FRP-CFST cross-sections 507

The ductility of the tested FRP-CFST and CFST specimens can been evaluated 508 through the ductility index DI proposed by Ge and Usami [52] and defined as follows: 509

  yuyu //DI

(4) 510

where δu and u are axial displacement and average axial strain, respectively, 511 corresponding to the ultimate capacity and δy and y are the axial displacement and 512 average axial strain, respectively, at the yield point. In cases where the yield point was 513

not sharply defined, based on the proposals of Schneider [2], the yield strain y was 514 taken as 0.2%. 515

The DI values of all the FRP-CFST specimens and CFST specimens are listed in 516 Table 3. As presented in the table, the DI values of the specimens 1S1TH-200-30 with 517 the PET FRP inner tube and S1NH-200-18 with the PEN FRP inner tube were 64.5 518 and 46.4, respectively, which are the largest and second largest among all the values. 519 This clearly demonstrates that the ductility of FRP-CFST cross-sections can be 520 significantly improved by using an inner FRP tube made of high rupture strain FRP. 521

The DI of the FRP-CFST specimens with inner GFRP tubes are plotted in Fig. 17 522 against the Df/Ds ratio. It may be seen that the DI value exhibits a decreasing trend 523 with increasing Df/Ds ratio and reducing FRP tube thickness. Also, the ductility of the 524 specimens filled with high strength concrete is clearly inferior to those filled with 525 normal strength concrete. 526

527

23

0.0 0.2 0.4 0.6 0.80

10

20

30

40

50 tf=2 mm (f 'co=36.5MPa) tf=3 mm (f '

co=54.7MPa)

tf=3 mm (f 'co=36.5MPa) tf=4 mm (f '

co=54.7MPa)

tf=4 mm (f 'co=36.5MPa)

DI

Df/D

s 528 Fig. 17. Ductility index for FRP-CFST specimens with inner GFRP tubes 529

530 3.7 Residual load-bearing capacity of FRP-CFST cross-sections with GFRP tubes 531

The minimum residual load-bearing capacities Nre of each of the tested 532 FRP-CFST specimens are reported in Table 4, together with the ratios Nre/Nue to 533 assess the post-peak drop-off in capacity with increasing deformations. The residual 534 load-bearing capacities of the FRP-CFST specimens with the high rupture strain inner 535 FRP tubes are not presented in the table since there was no drop in load throughout 536 the load-deformation history of the tests. The Nre/Nue ratio of the FRP-CFST 537 specimens varied between 0.80 and 0.93 depending on the properties of the inner 538 GFRP tube. The Nre/Nue ratios for the tested FRP-CFST cross-sections are plotted 539 against the Df/Ds ratios in Fig. 18. It may be seen that the Nre/Nue ratios show a 540 deceasing trend with increasing FRP tube diameter and thickness. Therefore, although 541 increasing the diameter and thickness of the FRP tube leads to the greatest benefit in 542 terms of ultimate load-carrying capacity, it does also result in the greatest post-peak 543 drop-off in load. 544

545

0.0 0.2 0.4 0.6 0.80.75

0.80

0.85

0.90

0.95

1.00

Df/D

s

Nre/N

ue

t f=2 mm (f 'co=36.5MPa) t f=3 mm (f '

co=54.7MPa)

t f=3 mm (f 'co=36.5MPa) t f=4 mm (f '

co=54.7MPa)

t f=4 mm (f 'co=36.5MPa)

546 Fig. 18. Variation of Nre/Nue ratios for tested FRP-CFST specimens with Df/Ds 547

548

4. Conclusions 549 An experimental study into the axial compressive behaviour of circular CFST 550

columns with an inner FRP tubes (termed FRP-CFST cross-sections herein) has been 551

24

presented. A total of 32 circular composite stub columns including 15 FRP-CFST stub 552 columns, 2 conventional CFST stub columns (without an inner FRP tube) and 15 553 FRP-confined concrete (FCC) stub columns were prepared and tested. Based on the 554 results and analysis presented in this paper, the following conclusions can be drawn: 555

556 (1) All tested FRP-CFST stub columns exhibited superior performance in terms of 557

delayed local buckling, increased load-bearing capacity and improved ductility 558 over their CFST counterparts without inner FRP tubes. 559 560

(2) The axial load-bearing capacity of the FRP-CFST cross-sections increased with 561 increasing thickness of the inner FRP tube due to the greater level of confinement 562 afforded to the inner concrete core and with increasing diameter (and FRP-to-steel 563 tube diameter ratio) of the inner FRP tube due to the greater area of 564 doubly-confined concrete. 565

566 (3) The ultimate load-bearing capacity of the tested FRP-CFST cross-sections was 567

slightly over-predicted by a simple superposition of the capacities of equivalent 568 CFST and FCC specimens minus the over-lapping portions, and some preliminary 569 insights into a possible design treatment were made. 570

571 (4) The material type of the inner FRP tube significantly influenced the ductility of 572

FRP-CFST stub columns, with high rupture strain FRP material showing the best 573 performance. The ductility index also showed a decreasing trend with increasing 574 FRP tube diameter and reducing FRP tube thickness. 575

576 (5) The post-peak residual capacity of the tested FRP-CFST specimens lay between 577

80 % and 93% of the ultimate load-carrying capacity, and showed a deceasing 578 trend with increasing FRP tube diameter and thickness. 579

580 Acknowledgments 581

This research was funded by the National Natural Sciences Foundation of China 582 (Grant No. 51008085 and Grant No. 11472084), Guangzhou Pearl River New Star of 583 Science & Technology Project (Grant No. 2012J2200100), China Scholarship Council 584 (Grant No. 201608440006) and China Postdoctoral Science Foundation (Grant No. 585 2012M511810 and No. 2014T70807). The paper was prepared during the first 586 author’s stay at Imperial College London as a visiting academic. 587

588

References 589 [1] HB Ge, T Usami, Strength of concrete-filled thin-walled steel box column: 590

experiment, J. Struct. Eng. ASCE 118 (11) (1992) 3036-3054. 591 [2] SP Schneider, Axially loaded concrete-filled steel tubes, J. Struct. Eng. ASCE 124 592

(10) (1998) 1125-1138. 593

25

[3] B Uy, Strength of concrete filled steel box columns incorporating local buckling. J. 594 Struct. Eng. ASCE 126 (3) (2000) 341-352. 595

[4] B Uy, Strength of short concrete filled high strength steel box columns, J. Constr. 596 Steel Res. 57 (2) (2001) 114-134. 597

[5] T Fujimoto, A Mukai , I Nishiyama, K Sakino, Behavior of eccentrically loaded 598 concrete-filled steel tubular columns, J. Struct. Eng. ASCE 130 (2) (2004) 203-212. 599

[6] K Sakino, H Nakahara , S Morino, I Nishiyama, Behaviour of centrally loaded 600 concrete-filled steel-tube short columns, J. Struct. Eng. ASCE 130 (2) (2004) 601 180-188. 602

[7] D Liu,WM Gho, Axial load behaviour of high strength rectangular concrete-filled 603 steel tubular stub columns, Thin-Walled Struct. 43(8) (2005) 1131-1142. 604

[8] LH Han, GH Yao, XL Zhao, Tests and calculations for hollow structural steel (HSS) 605 stub columns filled with self-consolidating concrete (SCC), J. Constr. Steel Res. 61 606 (9) (2005) 1241-1269. 607

[9] CW Roeder , DE Lehman, R Thody, Composite action in CFST components and 608 connections, AISC, Eng. J. 47 (4) (2009) 229-242. 609

[10] Z Ou, B Chen, KH Hsieh, MW Halling, PJ Barr, Experimental and analytical 610 investigation of concrete-filled steel tubular columns, J. Struct. Eng. ASCE 137 (6) 611 (2011) 635-645. 612

[11] XH Dai, D Lam, Shape effect on the behaviour of axially loaded concrete filled 613 steel tubular stub columns at elevated temperature, J. Constr. Steel Res. 73 (2012) 614 117-127. 615

[12] Z Tao, ZB Wang, Q Yu, Finite element modelling of concrete-filled steel stub 616 columns under axial compression, J. Constr. Steel Res. 89 (2013) 121-131. 617

[13] YL Long, J. Cai, Stress–strain relationship of concrete confined by rectangular 618 steel tubes with binding bars, J. Constr. Steel Res. 88 (2013) 1-14. 619

[14] JG Nie, YH Wang, JS Fan, Experimental research on concrete filled steel tube 620 columns under combined compression-bending-torsion cyclic load, Thin-Walled 621 Structures 67 (2013) 1-14. 622

[15]Y Yang, Y Wang, F Fu, Effect of reinforcement stiffeners on square 623 concrete-filled steel tubular columns subjected to axial compressive load, 624 Thin-Walled Structures, 82 (2014) 132-144. 625

[16] YL Long, J. Wan, J. Cai, Theoretical study on local buckling of rectangular CFST 626 columns under eccentric compression, J. Constr. Steel Res. 120 (2016) 70-80. 627

[17] Z Lai, AH Varma, Effective stress-strain relationships for analysis of noncompact 628 and slender filled composite (CFST) members, Engineering Structures 124 (2016) 629 457-472. 630

[18] Y Xiao, Applications of FRP composites in concrete columns, Adv Struc Eng. 631 7(4) (2004) 335-343. 632

[19] Y Xiao, W He, KK Choi, Confined concrete-filled tubular columns, J. Struct. Eng. 633 ASCE 131(3) (2005) 488-497. 634

[20] Z Tao, LH Han, JP Zhuang, Axial loading behavior of CFRP strengthened 635

26

concrete-filled steel tubular stub columns, Adv. Struc. Eng. 10(1) (2007) 37-46. 636 [21] KK Choi,Y Xiao, Analytical model of circular CFRP confined concrete-filled 637

steel tubular columns under axial compression, J. Compos. Constr. ASCE 14(1) 638 (2010) 125-133. 639

[22] YM Hu, JG Teng, T Yu, FRP-confined circular concrete-filled thin steel tubes 640 under axial compression, J. Compos. Constr. ASCE 15(5) (2011) 850-860. 641

[23] JG Teng, YM Hu, T Yu, Stress–strain model for concrete in FRP-confined steel 642 tubular columns, Engineering Structures 49 (4) (2013) 156-167. 643

[24] T Yu, YM Hu, JG Teng, FRP-confined circular concrete-filled steel tubular 644 columns under cyclic axial compression, J. Constr. Steel Res. 94 (2014) 33-48. 645

[25] CX Dong, AKH Kwan, JCM Ho, Axial and lateral stress-strain model for 646 concrete-filled steel tubes with FRP jackets, Engineering Structures 126 (2016) 647 365-378. 648

[26] JG Teng, T Yu, YL Wong, Behavior of hybrid FRP–concrete–steel double-skin 649 tubular columns. In: The 2nd international conference on FRP composites in civil 650 engineering-CICE 2004, Adelaide, Australia; 2004. p. 811-8. 651

[27] T Yu, YL Wong, JG Teng, SL Dong, SS Lam, Flexural behavior of hybrid 652 FRP-concrete-steel double skin tubular members, J. Compos. Construct. ASCE 10 653 (5) (2006) 443-452. 654

[28] JG Teng, T Yu, YL Wong, SL Dong, Hybrid FRP concrete steel tubular columns: 655 concept and behavior, Construct. Build. Mater. 21 (2007) 846-854. 656

[29] YL Wong, T Yu, JG Teng, Dong SL, Behavior of FRP-confined concrete in 657 annular section columns, Compos. Part B Eng. 38 (2008) 451-466. 658

[30] T Yu, B Zhang, YB Cao, JG Teng, Behavior of hybrid FRP-concrete-steel double 659 skin tubular columns subjected to cyclic axial compression, Thin Walled Struct. 61 660 (2012) 196-203. 661

[31] B Zhang, JG Teng, T Yu, Experimental behavior of hybrid FRP–concrete–steel 662 double-skin tubular columns under combined axial compression and cyclic lateral 663 loading, Engineering Structures 99 (2015) 214-231. 664

[32] LH Han, Z Tao, FY Liao, Y Xu, Tests on cyclic performance of 665 FRP–concrete–steel double-skin tubular columns, Thin-Walled Struct. 48 (6) 666 (2010) 430-93. 667

[33] B.A.L. Fanggi, T. Ozbakkaloglu, Compressive behavior of aramid 668 FRP-HSC-Steeldouble-skin tubular columns, Constr. Build. Mater. 48 (2013) 669 554-565. 670

[34]T Ozbakkaloglu, BA. Louk Fanggi, Axial compressive behavior of 671 FRP-concrete-steel double-skin tubular columns made of normal-and high-strength 672 concrete, J. Compos. Constr. ASCE 18 (1) (2014) 04013027. 673

[35] B.A.L. Fanggi, T. Ozbakkaloglu, Behavior of hollow and concrete-filled 674 FRP-HSC and FRP-HSC-Steel composites columns subjected to concentric 675 compression, Adv. Struct. Eng 18 (5) (2015)715-738. 676

[36] T Ozbakkaloglu, A novel FRP-dual-grade concrete-steel composite column 677

27

system, Thin Walled Struct. 96 (2015) 259-306. 678 [37] P Feng, S Cheng, Y Bai, L Ye, Mechanical behavior of concrete-filled square 679

steel tube with FRP-confined concrete core subjected to axial compression, 680 Composite Structures 123 (2015) 312-324. 681

[38] S Cheng, P Feng, Y Bai, LP Ye, Load-strain model for 682 steel-concrete-FRP-concrete Columns in Axial Compression, J. Compos. Constr. 683 ASCE, 2016. http://dx.doi.org/10.1061/(ASCE)CC.1943-5614. 0000664. 684

[39] D Anggawidjaja, T Ueda, JG Dai, H Nakai, Deformation capacity of RC piers 685 wrapped by new fiber-reinforced polymer with large fracture strain, Cement Concr. 686 Compos. 28 (10) (2006) 914–927. 687

[40] JG Dai, YL Bai, JG Teng, Behavior and modeling of concrete confined with FRP 688 composites of large deformability, J. Compos. Constr. 15 (6) (2011) 963– 973. 689

[41] JG Dai, L Lam, T Ueda, Seismic retrofit of square RC columns with polyethylene 690 terephthalate (PET) fibre reinforced polymer composites, Constr. Build. Mater. 27 691 (1) (2012) 206–217. 692

[42] YL Bai, JG Dai, JG Teng, Cyclic compressive behavior of concrete confined with 693 large rupture strain FRP composites, J. Compos. Constr. ASCE, 18 (1) (2014) 694 04013025. 695

[43]T Jirawattanasomkul, JG Dai, DW Zhang, M Senda, T Ueda, Experimental study 696 on shear behavior of reinforced-concrete members fully wrapped with large 697 rupture-strain FRP composites, J. Compos. Constr. (2014) 18. Special issue: 10th 698 Anniversary of IIFC, A4013009 699

[44] M Ispir, Monotonic and cyclic compression tests on concrete confined with 700 PET-FRP, J. Compos. Constr. 19 (1) (2014). 04014034 701

[45] ASTM C39/C39M. Standard test method for compressive strength of cylindrical 702 concrete specimens. Philadelphia, USA: American Society for Testing and 703 Materials; 2011. 704

[46] BS 18. Tensile testing of metals (including aerospace materials). London, UK: 705 British Standards Institution; 1987. 706

[47] ASTM D2290-08. Standard test method for apparent hoop tensile strength of 707 plastic or reinforced plastic pipe by split disk method. Philadelphia, USA: 708 American Society for Testing and Materials (ASTM); 1987. 709

[48] JC Lim and T Ozbakkaloglu, Lateral strain-to-axial strain relationship of 710 confined concrete, Journal of Structural Engineering, 141(5) (2014) 04014141. 711

[49] LH Han, Tests on stub columns of concrete-filled RHS sections , J. Constr. Steel 712 Res. 58 (2002) 353-372. 713

[50] F McCann, L Gardner, W Qiu, Experimental study of slender concrete-filled 714 elliptical hollow section beam-columns, J. Constr. Steel Res. 113 (2015) 185-194. 715

[51] Comité Européen de Normalisation, EN 1994-1-1:2004 Eurocode 4: Design of 716 Composite Steel and Concrete Structures-Part 1-1: General Rules and Rules for 717 Buildings, British Standards Institute, 2004. 718

[52] H B Ge, T Usami, Cyclic test of concrete-filled steel box columns, J. Struct. Eng. 719

28

ASCE 122 (10) (1996) 1169-1177. 720 721

722 Nomenclature 723 Ds diameter of steel tube 724 Df diameter of FRP tube 725 ts wall thickness of steel tube 726 tf wall thickness of FRP tube 727 As cross-sectional area of steel tube 728 Af cross-sectional area of FRP tube 729 Ac cross-sectional area of concrete 730 Acc cross-sectional area of core concrete within the FRP tube 731 Es elastic modulus of steel tube 732 EFRP elastic modulus of FRP tube 733 fFRP ultimate tensile stress of FRP tube 734 f′co cylinder strength of concrete 735 fy yield strength of steel tube 736 εau average axial strain at ultimate load 737 εFRP ultimate rupture strain of FRP tube 738 εal average axial strain corresponding to initiation of local buckling 739 Nue ultimate load-bearing capacity of specimens 740 Nue,CFST ultimate load-bearing capacity of CFST specimens 741 Nue,FCC ultimate load-bearing capacity of FCC specimens 742 Nue,FRP-CFST ultimate load-bearing capacity of FRP-CFST specimens 743 744 Nre residual load-bearing capacity of specimens 745 Nle axial load corresponding to initiation of local buckling 746 747 748 749

29

Table 1. Properties of FRP-CFST and CFST columns

Specimen L

(mm)

Ds

(mm)

Df

(mm)

ts

(mm)

tf

(mm) Ds/Df Df/tf

Concrete Steel tube FRP properties

f′co

(MPa)

fy

(MPa）

Es

(GPa) Type

EFRP

(GPa)

fFRP

(MPa) FRP

1S0FL 820 273 - 6.1 - - - 36.5 300.4 193 - - - -

1S1FL-100-2 820 273 100 6.1 2.0 2.73 50 36.5 300.4 193 GFRP 87 2714 0.0312

1S1FL-100-3 820 273 100 6.1 3.0 2.73 33.3 36.5 300.4 193 GFRP 87 2714 0.0312

1S1FL-100-4 820 273 100 6.1 4.0 2.73 25 36.5 300.4 193 GFRP 87 2714 0.0312

1S1FL-150-2 820 273 150 6.1 2.0 1.82 75 36.5 300.4 193 GFRP 87 2714 0.0312

1S1FL-150-3 820 273 150 6.1 3.0 1.82 50 36.5 300.4 193 GFRP 87 2714 0.0312

1S1FL-150-4 820 273 150 6.1 4.0 1.82 37.5 36.5 300.4 193 GFRP 87 2714 0.0312

1S1FL-200-2 820 273 200 6.1 2.0 1.37 100 36.5 300.4 193 GFRP 87 2714 0.0312

1S1FL-200-3 820 273 200 6.1 3.0 1.37 66.7 36.5 300.4 193 GFRP 87 2714 0.0312

1S1FL-200-4 820 273 200 6.1 4.0 1.37 50 36.5 300.4 193 GFRP 87 2714 0.0312

1S0FH 820 273 - 6.1 - - - 54.7 300.4 193 - - - -

1S1FH-150-3 820 273 150 6.1 3.0 1.82 50 54.7 300.4 193 GFRP 87 2714 0.0312

1S1FH-150-4 820 273 150 6.1 4.0 1.37 37.5 54.7 300.4 193 GFRP 87 2714 0.0312

1S1FH-200-3 820 273 200 6.1 3.0 1.82 66.7 54.7 300.4 193 GFRP 87 2714 0.0312

1S1FH-200-4 820 273 200 6.1 4.0 1.37 50 54.7 300.4 193 GFRP 87 2714 0.0312

1S1TH-200-30 820 273 200 6.1 3.0 1.37 6.7 54.7 300.4 193 PET FRP 8.7 760 0.0873

1S1NH-200-18 820 273 200 6.1 18.0 1.37 11.1 54.7 300.4 193 PEN FRP 13.9 856 0.0616

30

Table 2. Properties of FCC specimens

Note: ‘-‘ indicates value not available

Specimen L

(mm)

Df

(mm)

tf

(mm) Df/tf

Concrete FRP properties

f′co

(MPa) Type

EFRP

(GPa)

fFRP

(MPa) FRP

0S1FL-100-2 300 100 2.0 50 36.5 - - - -

0S1FL-100-3 300 100 3.0 33.3 36.5 GFRP 87 2714 0.0312

0S1FL-100-4 300 100 4.0 25 36.5 GFRP 87 2714 0.0312

0S1FL-150-2 450 150 2.0 75 36.5 GFRP 87 2714 0.0312

0S1FL-150-3 450 150 3.0 50 36.5 GFRP 87 2714 0.0312

0S1FL-150-4 450 150 4.0 37.5 36.5 GFRP 87 2714 0.0312

0S1FL-200-2 600 200 2.0 100 36.5 GFRP 87 2714 0.0312

0S1FL-200-3 600 200 3.0 66.7 36.5 GFRP 87 2714 0.0312

0S1FL-200-4 600 200 4.0 50 36.5 GFRP 87 2714 0.0312

0S1FH-150-3 450 150 3.0 50 54.7 GFRP 87 2714 0.0312

0S1FH-150-4 450 150 4.0 37.5 54.7 GFRP 87 2714 0.0312

0S1FH-200-3 600 200 3.0 66.7 54.7 GFRP 87 2714 0.0312

0S1FH-200-4 600 200 4.0 50 54.7 GFRP 87 2714 0.0312

0S1TH-200-30 600 200 30.0 6.7 54.7 PET-FRP 8.7 760 0.0873

0S1NH-200-18 600 200 18.0 11.1 54.7 PEN-FRP 13.9 856 0.0616

31

Table 3. Key test results of FRP-CFST and CFST specimens

Note: ‘-‘ indicates value not available

Specimen εal au Nle

(kN)

Nue

(kN)

Nue,FRP-CFST

/ Nue,CFST

Nre

(kN) Nre/Nue

Nre /

Nue,CFST φ SI DI

1S0FL 0.0310 0.0439 4167 4238 - 4167 0.98 - 1.22 21.9 1S1FL-100-2 0.0417 0.0501 4609 4734 1.12 4404 0.93 1.04 0.96 1.37 25.1

1S1FL-100-3 0.0520 0.0607 4729 4877 1.15 4527 0.93 1.07 0.92 1.41 30.3

1S1FL-100-4 0.0568 0.0684 4970 5154 1.22 4747 0.92 1.12 0.92 1.49 34.2

1S1FL-150-2 0.0360 0.0426 4973 5156 1.22 4669 0.91 1.1 0.97 1.49 21.3

1S1FL-150-3 0.0405 0.0499 5183 5422 1.28 4831 0.89 1.14 0.96 1.57 24.9

1S1FL-150-4 0.0435 0.0581 5490 5849 1.38 4889 0.84 1.15 0.94 1.69 29.1

1S1FL-200-2 0.0336 0.0376 5327 5519 1.30 4885 0.88 1.15 1.00 1.60 18.8

1S1FL-200-3 0.0382 0.0450 5597 5811 1.37 4892 0.84 1.15 0.98 1.68 22.5

1S1FL-200-4 0.0395 0.0492 6147 6392 1.51 5151 0.81 1.21 0.94 1.85 24.6

1S0FH 0.0038 0.0038 5573 5573 - 4594 0.82 - - 1.26 1.9

1S1FH-150-3 0.0210 0.0259 6107 6252 1.12 5304 0.85 0.95 0.93 1.41 13.0

1S1FH-150-4 0.0260 0.0295 6676 6721 1.21 5577 0.83 1.00 0.93 1.52 14.8

1S1FH-200-3 0.0197 0.0228 6507 6624 1.19 5601 0.85 1.00 0.97 1.50 11.4

1S1FH-200-4 0.0202 0.0259 6863 7190 1.29 5731 0.80 1.03 0.97 1.62 12.9

1S1TH-200-30 0.0475 0.1291 5373 6601 1.18 - - - 0.80 1.49 64.5

1S1NH-200-18 0.0419 0.0928 5451 6467 1.16 - - - 1.00 1.46 46.4

32

Table 4. Key test results of FCC specimens

Specimen Df

(mm) tf

(mm) Df/tf FRP type

Nue (kN) au εhu

0S1FL-100-2 100 2.0 50 GFRP 999 0.0472 0.0146

0S1FL-100-3 100 3.0 33.3 GFRP 1323 0.0604 0.0160

0S1FL-100-4 100 4.0 25 GFRP 1671 0.0636 0.0181

0S1FL-150-2 150 2.0 75 GFRP 1718 0.0362 0.0125

0S1FL-150-3 150 3.0 50 GFRP 2080 0.0453 0.0139

0S1FL-150-4 150 4.0 37.5 GFRP 2657 0.0563 0.0167

0S1FL-200-2 200 2.0 100 GFRP 2370 0.0286 0.0110

0S1FL-200-3 200 3.0 66.7 GFRP 2866 0.0354 0.0116

0S1FL-200-4 200 4.0 50 GFRP 3676 0.0449 0.0133

0S1FH-150-3 150 3.0 50 GFRP 2141 0.0185 0.0130

0S1FH-150-4 150 4.0 37.5 GFRP 2650 0.0203 0.0138

0S1FH-200-3 200 3.0 66.7 GFRP 2955 0.0135 0.0122

0S1FH-200-4 200 4.0 50 GFRP 3546 0.0186 0.0125

0S1TH-200-30 200 30.0 50 PET FRP 4360 0.0917 0.0527

0S1NL-200-18 200 18.0 50 PEN FRP 2480 0.0422 0.0407