You are on page 1of 40

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/353605627

Behaviour and design of stainless steel-concrete composite beams

Article in Journal of Constructional Steel Research · October 2021


DOI: 10.1016/j.jcsr.2021.106863

CITATIONS READS
13 401

6 authors, including:

Yifan Zhou Brian Uy


The University of Sydney UNSW Sydney
3 PUBLICATIONS 17 CITATIONS 106 PUBLICATIONS 1,530 CITATIONS

SEE PROFILE SEE PROFILE

Jia Wang Dongxu Li


The University of Sydney The University of Sydney
29 PUBLICATIONS 385 CITATIONS 52 PUBLICATIONS 854 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Dongxu Li on 31 July 2021.

The user has requested enhancement of the downloaded file.


1 Behaviour and design of stainless steel-concrete composite beams
2 Yifan Zhou*, Brian Uya, Jia Wanga, Dongxu Lia, Zhichao Huanga and Xinpei Liua
a
3 School of Civil Engineering, The University of Sydney, Sydney, NSW 2006, Australia
4

5 Abstract

6 Stainless steel-concrete composite beam has been regarded as a new application that is widely applied

7 in offshore bridges and high-rise buildings and is attracting increasing interest from engineering

8 community. The benefits brought by stainless steel, such as superior corrosion and fire resistance, high

9 strength and ductility as well as improved durability, could offer optimal and sustainable solutions for

10 infrastructures and highways. However, the current design provisions for composite beams are based on

11 the assumption of an elastic-perfectly-plastic steel model which is different to stainless steel material with

12 a rounded stress-strain curve and significant strain hardening. To develop rational design guidance, this

13 paper carried out a comprehensive experimental, numerical and analytical study on the behaviour of

14 stainless steel-concrete composite beams. A series of tests were conducted on eighteen full-scale stainless

15 steel composite beams under bending, shear as well as combined bending and shear. Three dimensional

16 finite element (FE) model was developed in ABAQUS and FE simulations were carried out to complement

17 the experimental program. New analytical models were proposed for predicting the moment capacity of

18 stainless steel-concrete composite beam. Additionally, a modified design approach for shear capacity was

19 recommended based on the design guidance in AS/NZS 2327. The applicability of current moment-shear

20 interaction defined in AS/NZS 2327 and EN 1994-1-1 was assessed by the obtained results.

21

22 Keywords: stainless steel composite beam, analytical model, finite element modelling, design guidance,

23 bending behaviour, shear behaviour

24

1
25 1 Introduction

26 Stainless steel has become a competitive alternative to carbon steel in structural applications, which is

27 well-known for its aesthetic appearance, superior corrosion resistance and improved durability [1-3].

28 Stainless steel-concrete composite beam has been regarded as a relatively new application and is attracting

29 increasing interest from engineering community. It is noted that, the traditional carbon steel-concrete

30 composite beams, which are widely applied in offshore bridges and high-rise buildings [4, 5] and exposed

31 to harsh environment with high salinity and humidity, have encountered the issue of corrosion. In this

32 light, the utilisation of stainless steel in composite beam provides an optimal solution for the offshore

33 infrastructures. The durability of structure is significantly improved and the maintenance costs over the

34 lifetime can be reduced [6]. Moreover, stainless steel has significant fire resistance that can maintain

35 satisfactory strength in very high temperature. Thus, the stainless steel composite structure is considered

36 as a fire protection solution for high-rise buildings [7]. Additionally, excellent mechanical performance in

37 terms of improved fatigue resistance and high ductility is also well known as the advantages of stainless

38 steel, which can particularly benefit the buildings and bridges [8].

39 To promote the application of stainless steel in composite structures, extensive experimental and

40 numerical investigations have been carried out on the behaviour of concrete-filled stainless steel tubular

41 (CFSST) columns. Young and Ellobody [9] and Lam and Gardner [10] demonstrated the satisfactory

42 behaviour of CFSST by tests on 14 and 12 stub columns, respectively. More recently, Uy et al. [11] and

43 Kazemzadeh Azad et al. [12, 13] conducted a series of experiments on short and slender CFSST columns

44 with austenitic, duplex and lean duplex stainless steels being used. On the other hand, very limited research

45 was found on stainless steel-concrete composite beams, although significant studies on carbon steel-

46 concrete composite beams were carried out in terms of the flexural behaviour, shear behaviour, as well as

47 combined flexure and shear behaviour. Back to the early age, Chapman [14] tested seventeen simply

2
48 supported composite beams with welded studs and proposed a new design method for composite section

49 and shear connectors. Recently, the benefits of demountable shear connectors have been recognised by

50 researchers. For example, Kwon et al. [15] constructed five large-scale composite beams retrofitted with

51 post-installed shear connectors and a substantial increase in stiffness and strength was achieved with a

52 small number of bolts. Lam et al. [16] compared the beams with bolted and welded shear connectors and

53 the results showed the former type had similar stiffness and superior ductility. Tan and Uy [17] and

54 Vasdravellis et al. [18, 19] conducted a series of experimental tests on composite beams subjected to

55 combined loading conditions and nonlinear FE analysis was also performed in these studies. In addition

56 to the flexural behaviour, the shear strength enhancement of composite beam has been compared with

57 bare steel beams by Johnson et al. [20] and Ansourian [21]. Nie et al. [22, 23] carried out an experimental

58 test on sixteen composite beams and quantified the strength contributed by concrete slab which was 33-

59 56%. Moreover, Vasdravellis and Uy [24] and Liang et al. [25, 26] tested composite beams with various

60 lengths and found out the moment resistance reduced when an acting shear force exceeded 60% of the

61 shear strength. Up to now, Shamass and Cashell [27] was the only one that proposed analytical models

62 for stainless steel composite beams through continuous strength method, with which the strain hardening

63 effects were involved. Therefore, the experimental and numerical investigation of the behaviour of

64 stainless steel-concrete composite beams is urgently required.

65 This paper presents a comprehensive experimental, numerical and analytical study on the behaviour of

66 stainless steel-concrete composite beams under bending, shear as well as combined bending and shear. In

67 particular, a total of eighteen full-scale composite beams fabricated with austenitic, duplex and lean duplex

68 stainless steel were tested under different loading conditions. To provide the interaction between concrete

69 slab and stainless steel beam, two types of stainless steel shear connectors (bolted connector and welded

70 stud) were employed to avoid the potential galvanic corrosion issues. Composite beams with full and

3
71 partial shear connection were both considered. Besides the experimental tests, parallel three-dimensional

72 FE models were developed and validated against experimental results, which were further used to extend

73 the available test database. The results obtained from experimental tests and numerical simulations were

74 used to assess the applicability of current codes of practice, such as AS/NZS 2327 [28] and EN 1994-1-1

75 [29].

76

77 2 Experimental programme

78 2.1 Material test

79 The standard coupons were extracted from the spare stainless steel plates, bolted connectors, welded

80 studs and reinforcing bars to assess their material properties. A testing procedure considering stress

81 relaxation was followed [30]. The main properties of stainless steel, including 0.01% and 0.2% proof, as

82 well as ultimate stress and ultimate strain are summarised in Table 1, where E denotes the Young’s modulus,

83 n and m denote the exponents from Ramberg-Osgood model. The typical stress-strain curves of stainless

84 steel plates and shear connectors are illustrated in Fig. 1 (a) and (b), respectively.

85 Twelve concrete cylinders were tested to assess the concrete compressive strength. The concrete

86 cylinders were cast together with concrete slabs and the mean values of the concrete compressive strength

87 at the age of 7th, 28th and testing days were 32, 37 and 42MPa, respectively.

88

89 2.2 Pushout test

90 2.2.1 Pushout specimens and test setup

91 To assess the shear resistance (Fsc) of duplex and austenitic stainless steel bolted connectors and welded

92 studs, three stainless steel pushout specimens with identical composite cross-section and material

93 properties were constructed. Fig. 2 (a) shows the geometry and instrument setup of pushout test, which

4
94 were designed according to AS/NZS 2327 [28]. As shown in the figure, each specimen consisted of one

95 stainless steel beam, two concrete slabs (600 mm × 650 mm × 100 mm) and eight stainless steel shear

96 connectors. Stainless steel shear connectors were installed in the beam flanges and embedded into the

97 concrete slabs. A 2 mm bolt-to-hole clearance was employed for specimens with bolted connectors, and

98 the connectors were tightened into the stainless steel beams by using a torque wrench with a torque load

99 of 120 N.m. Stiffeners were welded to the beam tip to prevent the premature failure due to the concentrated

100 stress.

101 The pushout specimens were equipped with eight LVDTs, as shown in Fig. 2(a). LVDT 1 was used to

102 acquire the vertical displacement of stainless steel beam. The horizontal separation of two concrete slabs

103 was obtained by LVDT 2, whereas the separations between stainless steel beam and concrete slab were

104 measured by LVDTs 5 and 6. Moreover, the relative slips between stainless steel beam and concrete slab

105 at two rows of shear connectors were recorded by LVDTs 3, 4 and LVDTs 7, 8, respectively. The pushout

106 tests were performed using a testing machine with a maximum capacity of 5000kN. The loading procedure

107 followed the guidance stated in AS/NZS 2327 [28]. An axial load was applied on the top of the stainless

108 steel beam up to 40% of the predicted ultimate load resistance (Fu). A cyclic loading protocol ranging

109 from 0.05Fu to 0.4Fu was repeated 25 times to examine the elastic behaviour of the shear connectors. The

110 specimens were thereafter loaded at a constant displacement-controlled rate of 0.5 mm/min and terminated

111 once the load dropped to 0.8Fu.

112

113 2.2.2 Bearing resistance of stainless steel shear connectors

114 Load-slip curves obtained from the stainless steel pushout tests are illustrated in Fig. 2 (b), where the

115 total load recorded by the loading cell is plotted against the average slip of shear connectors. The failure

116 mode of tested specimens was concrete crushing along with bending of embedded connector shanks.

5
117 Bearing resistance (PRK) of each shear connector is summarised in Table 2, which was determined by

118 dividing the total load by eight shear connectors. The bearing resistance of duplex stainless steel bolted

119 connector and welded stud were 111kN and 96kN, respectively, whilst that of austenitic stainless steel

120 bolted connector was 102kN. According to AS/NZS 2327 [28], a reduction factor of 0.9 on shear connector

121 bearing resistance was employed for the calculation of shear connection degree. In addition, the maximum

122 slip su at failure was evaluated, which satisfied the required minimum slip stated in AS/NZS 2327 [28]

123 (6.6 mm) and EN1994-1-1 [29] (6 mm) for ductile shear connectors. In this light, the duplex and austenitic

124 stainless steel shear connectors can be properly used in composite beams with partial shear connection

125 (PSC).

126

127 2.3 Stainless steel-concrete composite beams tests

128 2.3.1 Test preparation

129 A total of eighteen full-scale stainless steel-concrete composite beams were designed and fabricated

130 according to AS/NZS 2327 [28]. The typical cross-section of a composite beam is presented in Fig. 3 (a).

131 The details of specimens are summarised in Table 3, where Le denotes the effective length, a denotes the

132 shear span and β denotes the shear connection degree. All the specimens were labelled in a form of

133 “specimen number-connector type and the level of shear connection degree”. For instance, Specimen

134 CSSB1-BF was the composite stainless steel beam CSSB1 with full shear connection (FSC) utilising

135 bolted connectors. Each specimen consisted of a concrete slab, a welded stainless steel beam and a number

136 of shear connectors. The concrete slab was designed to be 600 mm wide and 100 mm thick, and the

137 stainless steel beam was designed as per a compact universal beam 180UB18.1. The stainless steel bolted

138 connectors and welded studs were 16 mm in diameter and 80mm in embedded height. The fully threaded

139 bolted connectors were placed through an 18 mm clearance hole and fastened with one embedded nut. All

6
140 the bolted connectors had a same preload to guarantee consistent performance. By adjusting the spacing

141 between the shear connectors, FSC or PSC was achieved for different stainless steel composite beams.

142 Moreover, two layers of stainless steel reinforcement meshes were placed inside the concrete slab to avoid

143 any concrete splitting failure. Both longitudinal and transverse reinforcement is 10 mm in diameter. The

144 spacing between longitudinal and transverse reinforcement is 180 mm and 150 mm, respectively. In order

145 to prevent the premature local buckling caused by concentrated stress, the stiffeners were welded to the

146 web at the positions of the loading points and supports. In addition, different types of stainless steel,

147 namely lean duplex, duplex and austenitic stainless steels, were used for the fabrication of beams, shear

148 connectors and reinforcing meshes.

149 The composite beams having various shear spans were tested under bending, shear and combined

150 bending and shear. These tests were conducted by a hydraulic actuator with a capacity of 1000kN. Each

151 specimen was initially loaded at a constant rate of 1 mm/min until the first yield was reached, that the first

152 yield was defined when the stainless steel bottom flange first reached its 0.2% proof stress. Followed by

153 an unloading procedure to 10kN and the specimen was reloaded up to failure with an increased speed of

154 1.5 mm/min. A small amount of 10kN was applied for unloading procedure was to maintain the contact

155 surfaces in position. The test was paused at 20%, 40%, 60% and 80% of expected ultimate load to record

156 the sequence of failure modes.

157 The tested stainless steel composite beams were equipped with a series of linear variable displacement

158 transducers (LVDTs), strain gauges and inclinometers as illustrated in Figs. 3 (b) and (c). In particular,

159 four and three LVDTs placed at the bottom flange were used to measure the vertical displacement of

160 composite beams under four- and three-point loading tests, respectively. Two additional LVDTs were

161 installed at the ends of beam to measure the interface slip between the concrete slab and stainless steel

162 beam. Strain gauges were located in sets of nine at the critical sections (Section A, B and C) of the

7
163 specimens as shown in Fig. 3 (a), where the strain developments at the concrete top surface, stainless steel

164 reinforcement, stainless steel beam flanges and webs were recorded. Additionally, inclinometers were

165 attached to the concrete top surface at supports, with which the rotations of the specimen were monitored

166 during the tests.

167

168 2.3.2 Bending test

169 Eight stainless steel composite beams were tested in bending under a simply support condition, in which

170 Specimens CSSB1-6 were subjected to four-point loading and Specimens CSSB7-8 were tested under

171 three-point loading. In four-point bending test, a pure bending state was observed in the mid span that

172 stress and strain distributions under pure moment can be produced. On the other hand, a complicated stress

173 state was noticed in three-point bending test that the influence on bending behaviour by shear stress could

174 be observed. The specimens under four-point loading were loaded at one- and two-thirds of the effective

175 length through a distributed beam, which was placed on a pair of roller and pin. As for the three-point

176 loading test, the load was applied through a pin at the mid span of the specimens. To avoid the stress

177 concentration at the loading and support points, 80mm wide rigid steel plates were utilised to spread the

178 load. The detailed configurations of loading and instrument setup for four-point and three-point loading

179 are illustrated in Figs. 3 (b) and (c). Different from Specimens CSSB1-2 and CSSB 5-6 having FSC,

180 Specimens CSSB3-4 were designed as PSC composite beams by increasing the spacing of shear connector

181 from 200 mm to 400 mm.

182

183 2.3.3 Shear test

184 A group of six short composite beams (Specimens CSSB9-14) were cut with various lengths (Type A

185 and B) from the tested four-point bending specimens (as shown in Fig.3 (c)) to achieve different shear

8
186 connection degrees for the following three-point loading shear tests. It is noteworthy that both the steel

187 and concrete of these cut segments remained in elastic during the previous bending tests, which could be

188 demonstrated by the recorded strain development in critical cross section C.

189

190 2.3.4 Combined bending-shear test

191 The remaining four Specimens CSSB15-18 were tested under combined bending and shear. Different

192 levels of moment and shear were achieved by adjusting the effective length (Le) of composite beams. In

193 these tests, all specimens were designed to be FSC composite beams and tested under three-point loading.

194

195 3. Experimental results and discussions

196 3.1 Stainless steel-concrete composite beams under bending

197 3.1.1 Failure modes

198 The failure modes for stainless steel-concrete composite beams under four-point and three-point

199 bending are illustrated in Fig. 4. Specimens CSSB1-6 experienced typical flexural failure in the four-point

200 bending test. During the test, flexural cracks initiated from the bottom of concrete slab at the loading

201 points, which was observed when the load reached around 60% of the ultimate load. Afterwards,

202 significant plastic strain was found at the bottom flange of stainless steel beams. The concrete cracks

203 gradually extended through the depth of concrete slab across the pure bending region, followed by

204 horizontal cracks appeared at the mid layer of the concrete slab. The specimens failed in concrete crushing

205 on the top surface, which randomly appeared in pure bending span of the composite beam. Additionally,

206 for PSC composite beams, the initiation of concrete cracks was observed at 75% of the ultimate load.

207 Noticeable end slips were found in PSC specimens, indicating that the weak shear connectors had effect

208 on the composite action between stainless steel beam and concrete slab.

9
209 Typical flexural failure was also found in Specimens CSSB7-8 that were subjected to three-point

210 bending, where concrete cracks were observed at 25% of ultimate load. Compared to Specimens CSSB1-

211 6, the premature cracks initiation of Specimens CSSB7-8 were mainly attributed to a higher stress

212 concentration. The concrete crushing was finally observed near the loading point together with significant

213 plastic strain at bottom flange.

214

215 3.1.2 Mid span moment-displacement curve

216 The mid span moment-displacement curves of stainless steel composite beams obtained from the four-

217 point and three-point bending tests are illustrated in Figs. 5 (a) and (b), where the comparison of different

218 shear connector types, shear connection degrees, beam materials and loading strategies are demonstrated.

219 Results of the bending tests on stainless steel-concrete composite beams are summarised in Table 4, where

220 Vu and Mu denote the maximum shear and moment of the tested specimens, respectively. In addition, the

221 initial stiffness (Ki.M) of each specimen is also summarised, which is defined as the tangent to the

222 unloading branch of the moment-displacement curve [31]. From the observations, the moment increased

223 linearly at early stage. Soon after, the curves exhibited non-linear behaviour before reaching the yield

224 point, which was defined as the point when the bottom flange of stainless steel beam started yielding. The

225 moment gradually developed until severe concrete crushing occurred at concrete surface, which led to the

226 final failure.

227 Specimens CSSB1-BF and CSSB2-SF with different stainless steel shear connector types exhibited

228 similar initial stiffness. However, the stiffness started to diverge once the curves behaved non-linearly,

229 where a weaker stiffness was observed for Specimen CSSB1-BF with stainless steel bolted connectors. At

230 this stage, the friction force induced by intended pretension had been overcome at the interface. The slip

231 between concrete slab and stainless steel beam was due to the 2 mm bolt-to-hole clearance and led to a

10
232 reduction in stiffness. This slip was recorded and further evidenced in Fig. 6 (a). Furthermore, the

233 Specimen CSSB1-BF with bolted connectors reached the similar ultimate moment resistance as Specimen

234 CSSB2-SF which employed the stainless steel welded studs.

235 The flexural behaviour of stainless steel composite beams with various levels of shear connection was

236 also compared in Fig. 5 (a). Regardless of the connector type, the PSC composite beams (CSSB3-BP and

237 CSSB4-SP) can still reach a high level of the load resistance, compared to their counterparts with FSC. In

238 addition, discrepancy in the non-linear stage was found between Specimens CSSB1-BF and CSSB3-BP.

239 This was due to the fact that more bolted connectors were assembled in Specimen CSSB1-BF and higher

240 total pretension force was applied. Thus, a higher interface friction delayed the commencement of

241 nonlinear stage, which is also reflected in the moment-end slip curves shown in Fig. 6 (b). On the other

242 hand, for composite beams with stainless steel welded studs, both specimens with FSC (CSSB2-SF) and

243 PSC (CSSB4-SP) exhibited similar initial stiffness and non-linear behaviour until the different ultimate

244 load resistance was reached.

245 Fig. 5 (a) also illustrates the mid span moment-displacement curves of composite beams utilising lean

246 duplex (Specimen CSSB1-BF), austenitic (Specimen CSSB5-BF) and duplex (Specimen CSSB6-BF)

247 stainless steels. As can be seen in the figure, duplex and lean duplex stainless steel composite beams

248 exhibited great similarity in terms of stiffness and ultimate moment resistance. Although the initial

249 stiffness of austenitic stainless steel specimen was similar to the former two, the sustained ultimate

250 moment resistance of Specimen CSSB5-BF was only around 68% of Specimen CSSB1-BF and CSSB6-

251 BF. The lower moment resistance was due to the lower material strength of austenitic stainless steel, which

252 can be demonstrated from the coupon test results. Furthermore, composite beams made with various

253 stainless steel grades presented ductile flexural performance characterised by significant vertical

254 displacement during the bending test.

11
255 Figs. 5 (a) and (b) compare the behaviour of stainless steel composite beams under four-point and three-

256 point loading. For the composite beams under four-point bending, the maximum moment was reached

257 right before the failure of concrete slab. On the other hand, the maximum moment resistance of specimens

258 subjected to three-point bending maintained for certain displacement before the failure occurred. As seen

259 from Table 4, the ultimate moment sustained by composite beams under three-point bending were 5%

260 higher than those from four-point bending test. The difference was because the strain hardening of bottom

261 flange contributed to an increase in the moment resistance once the stainless steel beam under three-point

262 bending reached the yielding near the loading point. However, the strain hardening only took effect until

263 the yielding was spread through the uniform moment region for the beams under four-point bending. To

264 the end, the moment resistance improved by the strain hardening effect under the former loading condition

265 was always more significant than that under the latter loading condition until failure. This phenomenon

266 was originally specified by Reck et al. [32], which was known as inelastic reserve strength.

267

268 3.1.3 Strain distributions

269 Strain evolution of stainless steel-concrete composite beams was monitored by sets of strain gauges,

270 which were arranged at the concrete slab surface, upper layer of reinforcement, stainless steel flanges and

271 web. Accordingly, the strain distributions of mid-span section for FSC and PSC composite beams are

272 illustrated in Fig. 7.

273 Take Specimen CSSB1-BF with FSC as an example, the strain distribution showed continuity across

274 the interface, indicating that the shear force was fully transferred between concrete slab and stainless steel

275 beam through the stainless steel bolted connectors. At yield point, the bottom flange firstly achieved its

276 yielding while the other region remained elastic. The elastic neutral axis determined from the strain

277 distribution was 68 mm from top surface. As shown in the figure, when the specimen reached the ultimate

12
278 moment resistance, the extreme fibre compressive strain was 2700 με, which denoted the concrete crushing

279 emerged in the concrete slab. Moreover, the development of plasticity exhibited in the most region of

280 stainless steel beam, where the maximum tensile strain was over 12,000 με in the bottom flange. This large

281 tensile strain demonstrated the significant strain hardening developed in stainless steel section prior to the

282 final failure, which was also reflected in the discrepancy of the plastic neutral axis determined from the

283 test and rigid plastic analysis (RPA) method. The inaccuracy in estimating the neutral axis with RPA

284 method was due to the ignorance of strain hardening effect in stainless steel elements and consequently

285 led to the underestimation of moment capacity.

286 Fig. 7 also presents the strain distribution of Specimen CSSB4-SP with PSC (β = 0.6). The recorded

287 strains showed discontinuity at the interface between concrete slab and stainless steel beam, as the stainless

288 steel welded studs were unable to provide enough shear force to fully transfer the composite action. When

289 the bottom flange of this specimen reached its yielding strain, the strain in concrete slab was lower than

290 that of FSC composite beam (CSSB1-BF), which explained the fact that the initiation of concrete cracks

291 in PSC composite beam occurred at a relatively higher moment as discussed in Section 3.1.2. Two elastic

292 neutral axes were determined accordingly, with the first one extending 45mm from the concrete slab and

293 second one 5 mm below the interface. Moreover, the extreme fibre compressive and tensile strain of PSC

294 composite beam at ultimate load level was much lower than that of FSC counterparts, which was the

295 reason that a lower moment resistance was achieved for PSC specimens.

296

297 3.2 Stainless steel-concrete composite beam under shear

298 3.2.1 Failure modes

299 Six stainless steel-concrete composite beams CSSB9-14 experienced shear failure in this loading

300 condition. The typical shear failure modes are presented in Fig. 8. During the test, minor vertical cracks

13
301 at the bottom of concrete slab initiated at a loading level of 30% ultimate shear resistance. With the

302 increasing applied load, the vertical cracks extended through the depth of concrete slab and horizontal

303 cracks were observed along the length of the slab. The drop of load resistance was triggered by the

304 appearance of large diagonal concrete cracks. Afterwards, significant diagonal buckling occurred in

305 stainless steel beam web, which was a result of the intended web stiffeners. At the end of the shear tests,

306 the composite beam exhibited significant ductile performance with large vertical displacements.

307 Compared with the FSC specimens, apparent end slip was observed in the PSC composite beams.

308

309 3.2.2 Load-displacement curve

310 The shear test results of stainless steel-concrete composite beams are summarised in Table 5 and the

311 load-displacement response at mid span is presented in Fig. 9. Generally, the load-displacement curves

312 exhibited a linear increment at the initial stage followed by a nonlinear transition stage. The shear load

313 gradually reduced with the diagonal cracks propagating towards the loading point, which highlighted the

314 loss of concrete strength. Meanwhile, the shear load was redistributed to the stainless steel beam until

315 significant web buckling occurred.

316 As shown in Fig. 9 (a), limited influence on the initial shear stiffness (Ki.V) was found by changing the

317 types of shear connectors, although a smaller shear stiffness was observed in Specimen CSSB9-BF at the

318 nonlinear stage owing to the bolt-to-hole clearance. Moreover, Specimens CSSB9-BF and CSSB10-SF

319 with different stainless steel shear connectors attained similar shear capacity indicating that both bolted

320 connectors and welded studs can provide excellent composite action for composite beams under shear.

321 Fig. 9 (a) also highlights the influence of shear connection degree on the shear behaviour of stainless

322 steel-concrete composite beams. Take Specimen CSSB11-BP as an example, the specimen having shear

323 connection degree of 0.7 achieved 95% ultimate shear resistance of the FSC counterparts (Specimen

14
324 CSSB9-BF). The sustained load decreased when the specimen had lower shear connection degree. It is

325 also worth noting that, unlike FSC composite beams, PSC specimens did not exhibit a sharp peak at the

326 ultimate load. This was due to that weak shear resistance of stainless steel shear connectors led to interface

327 slip and consequently diminished the shear strength of stainless steel composite beams. Moreover, the

328 shear connection degree had limited influence on the initial stiffness of composite beams.

329 The effect of different stainless steel grades was examined as illustrated in Figs. 9 (a) and (b). As can

330 be seen, the shear capacities sustained by lean duplex and duplex stainless steel composite beams were

331 approximately identical, which was 25% higher than that of austenitic stainless steel composite beams.

332 Besides, the initial stiffness of austenitic stainless steel composite beam was lower than that fabricated

333 with duplex stainless steel. In addition, remarkable plastic deformations were observed for all of the tested

334 specimens after concrete failure, which was contributed by the high ductility of stainless steel material.

335

336 3.3 Stainless steel-concrete composite beam under combined bending and shear

337 3.3.1 Failure modes

338 Typical failure modes of tested specimens CSSB15-18 are presented in Fig. 10, which were tested under

339 combined bending and shear. The critical failure mode was highly dependent on the shear span of

340 composite beams. The failure of Specimens CSSB15-BF and CSSB17-SF was dominated by flexural

341 failure, which was characterised by crushing of concrete slab. Apparent diagonal cracks on concrete slab

342 were also observed during the test, indicating that the specimens sustained high shear stress in the

343 meantime. The failure of Specimens CSSB16-BF and CSSB18-SF were governed by shear failure, which

344 were represented by significant concrete diagonal cracks and web buckling in stainless steel beams.

345

346 3.3.2 Load-displacement curve

15
347 The mid span load-displacement curves of selected specimens are illustrated in Fig. 11, which consist

348 of the lean duplex stainless steel composite beams under bending, shear and combined bending and shear.

349 These specimens were divided into two groups according to the type of stainless steel shear connector,

350 and in each group the shear span of the specimens varied to achieve different levels of moment and shear.

351 The ultimate moment and shear resistances of the tested specimens are summarised in Table 6 and are

352 normalised by the maximum shear and moment resistances. The maximum moment resistance (Mu.m) and

353 shear resistance (Vu.m) were defined by the Specimens CSSB9-BF/CSSB10-SF and Specimens CSSB7-

354 BF/CSSB8-SF for composite beams utilising bolted and welded shear connectors, respectively. As

355 evidenced in the table, the moment capacity of composite beams has been limitedly decreased when

356 applied shear force was less than 60% Vu.m. On the other hand, an intended moment of 75% Mu.m can

357 reduce the shear capacity of the composite beam to 93% Vu.m.

358

359 4. Finite element modelling

360 Three-dimensional FE models of stainless steel-concrete composite beams were developed by using

361 commercial program ABAQUS [33], which was used to achieve two purposes, 1) to explore the

362 mechanical performance of composite beams; 2) to extend the test database and facilitate the establishment

363 of design solutions.

364 4.1 Geometry, boundary and loading conditions

365 The concrete slabs and shear connectors were modelled by eight-node linear brick elements (C3D8R),

366 while the stainless steel beams utilised four-node shell elements with reduced integration (S4R).

367 Additionally, stainless steel reinforcement meshes were modelled by truss elements (T3D2) and were

368 embedded in the concrete slab. The dimensions of numerical model were referred to the geometries of

369 tested specimens. The threaded bolt shank was simulated with a reduced area equalled to 78% of bolt

16
370 gross sectional area [34]. Fig. 12 shows the detailed geometry of FE models.

371 The boundary and loading conditions were carefully prescribed in the numerical model. The roller-

372 pinned supports and axial load were defined by reference points via “MPC” constraint. To simulate the

373 interaction between different composite components, a general contact was defined with “Hard contact”

374 in normal direction and a friction coefficient of 0.4 in the tangential direction [35]. A “tie” constraint was

375 used for the welded elements, such as welded studs and web stiffeners. The pretension effect on bolted

376 connectors was reproduced by applying a thermal expansion coefficient defined in the temperature field

377 [36]. Moreover, the dynamic explicit solver was adopted to analyse the behaviour of stainless steel-

378 concrete composite beams under bending, shear as well as combined bending and shear. The explicit

379 solver could easily overcome the convergence problem involving complicated geometries and material

380 nonlinearity [37]. To achieve accurate outcomes with efficient computational costs, a mesh sensitive

381 analysis was carried out with a mesh of 4 mm being selected for critical elements.

382

383 4.2 Stainless steel material model

384 The stress-strain relationships of stainless steel beams, shear connectors and reinforcements obtained

385 from coupon tests were used as inputs of FE model. A von Mises plasticity constitutive formulation

386 together with the isotropic hardening was employed to capture the behaviour of lean duplex, duplex and

387 austenitic stainless steels [38]. As per ABAQUS manual [33], true stress-strain relationships of stainless

388 steel elements were adopted in FE model which were derived from the engineering stress and strain

389 obtained from the coupon results.

390

391 4.3 Concrete material model

392 The ascending branch of the concrete compressive stress-strain relationship was proposed by Carreira

17
393 and Chu [39] up to the maximum compressive stress (f’c). The descending branch was followed by an

394 inverse function suggested by Pathirana et al. [40], for the purpose of mirroring the post-peak behaviour

395 of composite beam, where the ultimate compressive strain was taken as 0.03 to eliminate the unrealistic

396 overestimation of concrete strength [41]. In addition, a concrete damage evolution law was defined, with

397 which the concrete damage performance can be well captured [42].

398 The tensile behaviour of concrete was defined to be linear up to the ultimate tensile stress (f’t = f’c/10).

399 For the descending branch of tensile stress-strain curve, the tensile stress linearly reduced to zero with the

400 ultimate tensile strain recommended to be 0.1 [43].

401

402 4.4 Validation of finite element models

403 Results from the developed FE models were compared with the corresponding experimental data.

404 Failure modes of the composite beams during the FE simulation was identified by one of the following

405 situations: 1) concrete crushing; 2) stainless steel web shear buckling. For composite beams under bending

406 test, concrete crushing was observed when the principal compressive strain at concrete top fibre reached

407 the ultimate strain. Moreover, in the simulation of the shear test, the diagonal buckling on stainless steel

408 web can be straightforwardly observed in ABAQUS. The comparisons in terms of the mid span load-

409 displacement curve for selected specimens with different loading conditions are presented in Fig. 13. As

410 observed, the results obtained from FE model show a good agreement with the test results. The general

411 trend and initial stiffness of the tested specimens were well predicted by the numerical models, whilst the

412 discrepancy regarding the ultimate load between test and FE results was less than 10%. Therefore, the

413 developed FE model herein is reliable and can well predict the behaviour of stainless steel composite

414 beams subjected to bending, shear, as well as combined bending and shear.

415

18
416 5. Evaluation on codes of practice

417 5.1 Analytical model for moment capacity

418 5.1.1 FSC stainless steel-concrete composite beams

419 The current design provisions, such as AS/NZS 2327 [28] and EN 1994-1-1 [29], employed RPA

420 method for prediction of the bending moment capacity (Mpl) for the traditional composite beam. This

421 design method assumed an elastic-perfectly-plastic material behaviour for steel section and neglected the

422 strain hardening effect. Fig. 14 (a) presents the typical strain and stress distributions of composite beam

423 predicted by RPA method. The bending moment capacity is determined based on the equilibrium of forces

424 where the concrete slab reaches its maximum compressive strength and steel sections are stressed to the

425 yield strength. Unlike the carbon steel that has a clear yielding point and minor strain hardening, stainless

426 steel exhibits a rounded stress-strain curve and significant strain hardening in post yielding stage. As

427 evidenced in the bending test conducted herein, a considerable level of strain hardening developed in

428 stainless steel beams prior to failure. The moment capacities obtained in four-point and three-point

429 bending tests were 15% and 23% higher than that predicted by RPA method, respectively, owing to the

430 strength enhancement brought by strain hardening. In this case, a new analytical model is required to

431 accurately predict the moment capacity (Mpl) of stainless steel-concrete composite beams involving the

432 strain hardening effect.

433 To account for the strength contribution from strain hardening effect, a bi-linear stainless steel material

434 model was modified based on the coupon results and the methods proposed by [44-47], which is presented

435 in Fig. 15. The modified stainless steel stress-strain curve exhibits elastic behaviour up to the yielding

436 point defined as (ε0.2, f0.2). The strain hardening slope Esh can be determined by a line passing through the

437 yielding point and the specified maximum point (C2εu, fu). The intersection of proposed line and original

438 stress-strain curve is defined to be the point (C1εu, fc1εu). The parameters C1 and C2 and strain hardening

19
439 slope Esh can be determined from the following expressions:

𝜀0.2 + 0.1(𝜀𝑢 − 𝜀0.2 )


𝐶1 = (1)
𝜀𝑢
𝜀0.2 + 0.3(𝜀𝑢 − 𝜀0.2 )
𝐶2 = (2)
𝜀𝑢
𝑓𝑢 − 𝑓0.2
𝐸𝑠ℎ = (3)
𝐶2 𝜀𝑢 − 𝜀0.2
440 In addition to the modified material model, a strain hardening analysis (SHA) method was proposed in

441 this paper. The typical strain and stress distributions developed by SHA method are illustrated in Fig. 14

442 (b). The SHA model assumed the stainless steel sections were stressed to post yielding strength (f0.2) region

443 and the strain linearly distributed through the depth of stainless steel beam section for FSC composite

444 beams. Accordingly, the stress can be determined based on strain distribution by applying the modified

445 bi-linear stainless steel material model. Thus, the increment of stress (∆fh) brought by the strain hardening

446 effect at a depth (h) can be written as a function of the corresponding strain (εh).

∆fℎ = 𝐸𝑠ℎ (𝜀ℎ − 𝜀0.2 ) (4)


447 The plastic neutral axis (dNA) and bending moment capacity (Mpl) that account for the strain hardening

448 effects can be therefore determined by the equilibrium of strain distribution and internal forces. It should

449 be noted that the outer fibre concrete strain (εc) was assumed to be 0.003 [28] and the strain (εs) at the

450 bottom outer fibre of stainless steel section should be less than C1εu (the stress was assumed to be fC1εu if

451 the strain (εs) was greater than C1εu). Moreover, the reinforcement in concrete slab under compression was

452 neglected during the calculation, while the reinforcement under tension was assumed to be stressed to its

453 yielding strength. The comparisons regarding the moment capacities (Mpl) obtained by the proposed

454 design model and experimental tests are summarised in Table 4. It can be concluded that the accuracy of

455 prediction has been well improved by SHA method compared with RPA method. The improvement in

456 prediction accuracy by SHA method will be more significant for composite beams consisting of stainless

457 steel grade with higher strain hardening effect or large-dimension of concrete slab.

20
458 Despite the improved prediction accuracy, the computational complexity of detailed SHA method might

459 limit its wide application, owing to the demand of strain and stress calculation for each stainless steel

460 element. The stress distribution for composite section was therefore modified with a simplified SHA

461 method, as illustrated in Fig. 14 (c). A linear strain distribution was still adopted in this model, while the

462 stress distribution of stainless steel beam section was modified as a rectangular with a constant value of fs

463 [27], which was the sum of the yielding stress (fy) and the stress increment (∆f). The increment of stress

464 (∆f) was defined as half of the strength brought by strain hardening effect at the bottom outer fibre of

465 stainless steel section. The function of stress increment (∆f) is expressed by Eq. (5)

∆f = 1/2𝐸𝑠ℎ (𝜀𝑠 − 𝜀0.2 ) (5)


466 As the distribution of strain is continuous throughout the depth of composite section, the plastic neutral

467 axis (dNA) could therefore be determined based on the equilibrium of internal forces and strain gradient,

468 which are given as:

F𝑐 − F𝑟𝑡 − F𝑠 = 0.85f𝑐 𝑑𝑁𝐴 𝑏𝑠𝑙𝑎𝑏 − f𝑦 𝐴𝑟𝑡 − (∆f + f𝑦 )𝐴𝑠 = 0 (6)


𝜀𝑐 𝜀𝑠
= (7)
𝑑𝑁𝐴 𝐷𝑐𝑜𝑚𝑝 − 𝑑𝑁𝐴
469 Thus, the bending moment capacity (Mpl) can be obtained by the sum of internal forces (Fc, Frt, Fs)

470 times their corresponding lever arms. It is noteworthy that the plastic neutral axis (dNA) and total internal

471 force of stainless steel beam (Fs) generated from detailed and simplified SHA methods was same, but the

472 reduction in lever arms of Fs lead to a conservative prediction on moment capacity by simplified SHA

473 method. The obtained results (MSHA.S) of tested specimens from simplified SHA method are summarised

474 in Table 4. It was evidenced that the predictions from simplified SHA model could acquire similar

475 accuracy as the detailed SHA model for stainless steel composite beams.

476

477 5.1.2 PSC stainless steel-concrete composite beams

478 The simplified SHA method was also applicable to stainless steel composite beam with PSC. The typical

21
479 strain and stress distributions of a composite section are presented in Fig. 14 (d), where two neutral axes

480 lie within concrete slab and stainless steel beam. With the known load bearing capacity (Fsc) provided by

481 shear connectors at the interface, the neutral axis (dNA1) in concrete slab could be easily determined by the

482 equilibrium of forces (Fc, Frt and Fsc). The curvature (κ) herein could be determined as a ratio of outer

483 fibre concrete strain (εc = 0.003) over the depth of first plastic neutral axis (dNA1). Meanwhile, the curvature

484 (κ) could also be expressed as the ratio of strain at outer fibre of stainless steel beam over the distance

485 from second neutral axis (dNA2) to bottom surface, as it assumed that no separation between concrete slab

486 and beam section. The expression of curvature (κ) is given as,

ε𝑐 ε𝑠
κ= = (8)
dNA1 𝐷𝑠𝑡𝑒𝑒𝑙 − dNA2
487 The second neutral axis (dNA2) could therefore be obtained based on the equilibrium of forces generated

488 in stainless steel beam and shear connectors. Finally, the bending moment capacity (Mpl) of stainless steel-

489 concrete composite beam with PSC could be calculated by internal forces times their corresponding level

490 arms. The maximum moment resistance of tested specimens with partial shear connection obtained by the

491 simplified SHA method are summarised in Table 4.

492 A series of FE simulation was also conducted on stainless steel composite beams with various degrees

493 of shear connection. The obtained experimental and numerical results are compared with those calculated

494 by the simplified SHA model as illustrated in Fig. 16. As per required in AS/NZS 2327 [28], the moment

495 resistance only considered the contribution from steel beam when the shear connection degree was zero.

496 As evidenced in the plot, the obtained results from the simplified SHA method showed reasonable and

497 conservative predictions on moment capacity of stainless steel composite beams with shear connection

498 degree larger than 0.2. Owning to the fact that the composite beam with low shear connection degree is

499 not allowed in engineering practice and the minimum shear connection degree is 0.4 as per AS/NZS 2327

500 [28], the proposed simplified SHA method was applicable for PSC stainless steel composite beams.

22
501 5.2. Design approaches for shear capacity

502 5.2.1 FSC stainless steel-concrete composite beams

503 The design approach of shear capacity (Vcomp) for stainless steel composite beams has not been outlined

504 in existing codes of practice. The current design approaches adopted in international standards AS/NZS

505 2327 [28] and EN 1994-1-1 [29] were developed for carbon steel composite beams. As stated in EN 1994-

506 1-1 [29], to obtain a conservative prediction, only the shear strength contributed by the steel web (Aw) was

507 considered, which is given as:

Vw = 0.6fy Aw (9)
508 To take the advantage of composite action, a design approach including the shear strength of both steel

509 web and concrete slab was outlined in AS/NZS 2327 [28]. In this design approach, the ultimate shear

510 capacity (Vcomp) of composite beam was expressed as:

Vcomp = Vw + Vslab (10)


Vslab = f(λ𝑠𝑑 )(𝑏𝑓 𝐷𝑠𝑙𝑎𝑏 )0.7 √𝑓′𝑐 (11)
f(λ𝑠𝑑 ) = 110λ𝑠𝑑 + 13 (12)
λ𝑠𝑑 = 𝐷𝑠𝑙𝑎𝑏 /𝐷𝑐𝑜𝑚𝑝 (13)
511 The shear capacities (Vcomp) of tested specimens determined from two design approaches were

512 compared with the experimental results in Table 5. In particular, EN 1994-1-1 [29] provided an

513 overconservative estimation on the shear capacity (Vcomp) as it neglected the contribution from concrete

514 slab. On the contrary, the results obtained from AS/NZS 2327 [28] exhibited overestimated predictions on

515 shear capacity (Vcomp) of stainless steel composite beams. The reason for this overestimation was that the

516 early failure of concrete slab limited the shear strength development of stainless steel beam. Owing to this

517 fact, it was unsafe to assume the stainless steel web section stressed to shear yielding when determining

518 the shear capacity (Vcomp) of stainless steel composite beam. In this view, the authors herein proposed a

519 new design solution by regression analysis method. Reduction shear strength of stainless steel was

520 employed and expressed as Eq. (14).

23
𝑛
Vw = 0.6 √0.15f0.2 Aw (14)
521 where, n denotes the exponent from Ramberg-Osgood model based on 0.01% and 0.2% proof stress [48].

522 The results of tested specimens calculated by new expression (Vnew) are listed in Table 5 and

523 demonstrate conservative estimations on shear capacities of stainless steel composite beams.

524

525 5.2.2 PSC stainless steel-concrete composite beams

526 The effect of shear connection degree on the shear capacity (Vcomp.β) of carbon steel composite beams

527 has been well evaluated by previous studies [24, 26, 49] and such influences were also reflected in the

528 current code of practice AS/NZS 2327 [28]. It is believed that the shear strength of composite beam is

529 negatively influenced when shear connectors are unable to provide full shear connection between concrete

530 slab and steel beam. As stated in AS/NZS 2327 [29], the influence on shear capacity (Vcomp.β) of composite

531 beam with PSC can be expressed by a reduction factor (α):

Vcomp.β = 𝛼Vcomp (15)


1−β
𝛼= +β (16)
0.76λ𝑠𝑑 + 0.92
532 To evaluate if this design solution was still applicable for stainless steel composite beam with PSC, a

533 series of FE simulations was carried out on lean duplex and austenitic stainless steel composite beams.

534 The geometry of FE model was identical to the Specimen CSSB9-BF in present test programme. Different

535 shear connection degrees were achieved by modifying the diameters of the stainless steel shear connectors,

536 whilst the other conditions remain unchanged. As shown in Fig. 17, the obtained results from the

537 experimental tests, FE simulations and design standard AS/NZS 2327 [28] are normalised by dividing

538 Vcomp, which is derived with Eqs. (10) and (15). As observed, the shear capacity of stainless steel composite

539 beams with PSC is improved with an increase of shear connection degree, which can be conservatively

540 predicted by AS/NZS 2327 [28].

541

24
542 5.3. Comparison moment-shear interaction with the existing code of practice

543 The moment-shear interaction behaviour of stainless steel composite beam has been evaluated in this

544 section. A series of FE simulations was conducted to assess the influence of shear span on the moment

545 and shear resistances. The developed models had identical composite section but varied in shear span. The

546 moment-shear interaction diagrams predicted by AS/NZS 2327 [28] and EN 1994-1-1 [29] are plotted in

547 Fig. 18, in which the moment (Mpl) and shear capacities (Vcomp) were calculated based on the proposed

548 methods in the Section 5.1.1 and 5.2.1, respectively. In addition, the obtained results from experimental

549 tests and FE simulations were also included. As shown in the figure, the outcomes from the experimental

550 tests and FE simulations follow the same trend of the interaction diagrams generated from codes of

551 practice. The reduction in ultimate moment resistance of composite beams emerges when the applied shear

552 force is greater than 60% Vcomp. Meanwhile, a reduction in the ultimate shear resistance can be observed

553 when the applied moment exceeds 50% Mpl.

554

555 6. Conclusions
556 This paper presented an experimental, numerical and analytical study on the behaviour of stainless

557 steel-concrete composite beams subjected to bending, shear as well as combined bending and shear. A

558 total of eighteen full-scale composite beams made with lean duplex, duplex and austenitic stainless steels

559 were examined. In parallel with the experimental tests, three-dimensional FE models were developed in

560 commercial programme ABAQUS and the verified FE models were further employed for extensive

561 numerical analysis. In addition, the international codes of practice were evaluated by obtained

562 experimental and numerical results. Herein, new analytical models and design approaches were proposed

563 for stainless steel-concrete composite beam. Based on the study, the following conclusions can be made:

564 • All tested specimens under bending, shear as well as combined bending and shear exhibited

565 very high ductility owing to the advantage of stainless steel members.

25
566 • The stainless steel-concrete composite beams with bolted connectors exhibited similar ultimate

567 moment resistance as those with welded studs. However, the former possessed smaller flexural

568 stiffness due to the bolt-to-hole clearance. Moreover, the strain distribution of composite section

569 in the test was evidently different from the prediction by RPA method owing to the significant

570 strain hardening, which can be explicitly found in stainless steel composite beams.

571 • The ultimate shear resistance of stainless steel composite beams with bolted connectors and

572 welded studs was approximately identical, but the shear capacity could be reduced with the

573 lower shear connection degree.

574 • By comparing with experimental results, the developed FE model was proved to be reliable to

575 predict the behaviour of stainless steel-concrete composite beams.

576 • Compared with the existing codes of practice (AS/NZS 2327 [31] and EN 1994-1-1 [32])

577 designed for carbon steel composite beam, the stainless steel-concrete composite beam could

578 sustain higher flexural capacity owing to its remarkable strain hardening effects.

579 • New simplified SHA model incorporating the strain hardening effects were proposed for FSC

580 and PSC stainless steel-concrete composite beams, which could accurately predict the moment

581 capacity, and were suggested as a straightforward design tool for engineer practice.

582 • The design approach stated in AS/NZS 2327 [31] overestimated the shear capacity of the

583 stainless steel composite beams since the early concrete failure limited the development of

584 stainless steel strength. A new design method was proposed by employing a reduced strength of

585 stainless steel, which showed conservative predictions on the shear capacities.

586 • The design approaches stated in AS/NZS 2327 [28] and EN 1994-1-1 [29] could provide a

587 reliable prediction on the moment-shear interaction behaviour of stainless steel-concrete

588 composite beams, given that modified design methods for Mpl and Vcomp are used.

26
589 Acknowledgements

590 This research work has been supported by Australian Research Council under its Discovery Project

591 Scheme (Project ID: DP180100418). The authors would like to acknowledge Outokumpu and Stirlings

592 performance steel for the generous donations of stainless steel plates. The generous donations of stainless

593 steel bolted connectors from Bumax are also gratefully acknowledged. A special thanks to assistance and

594 advices from Dr. Mohanad Mursi, Mr. Garry Towell and the technicians from J.W.Roderick Laboratory.

595

596 References
597
598 [1] E. Mirambell, E. Real, On the calculation of deflections in structural stainless steel beams: an
599 experimental and numerical investigation, J. Constr. Steel Res. 54(1) (2000) 109-133.
600 [2] L. Gardner, M. Theofanous, Discrete and continuous treatment of local buckling in stainless steel
601 elements, J. Constr. Steel Res. 64(11) (2008) 1207-1216.
602 [3] D. Li, B. Uy, F. Aslani, C. Hou, Behaviour and design of spiral-welded stainless steel tubes subjected
603 to axial compression, J. Constr. Steel Res. 154 (2019) 67-83.
604 [4] B. Uy, M. Bradford, Elastic local buckling of steel plates in composite steel-concrete members, Eng.
605 Struct. 18(3) (1996) 193-200.
606 [5] B. Uy, Applications, behaviour and design of composite steel-concrete structures, Adv. Struct. Eng.
607 15(9) (2012) 1559-1571.
608 [6] A. Roufegarinejad, B. Uy, M. Bradford, A.J. Deeks, H. Hao, Behaviour and design of concrete filled
609 steel columns utilising stainless steel cross sections under combined actions, Proceedings of the 18th
610 Australasian Conference on the Mechanics of Structures and Materials, Perth, Australia, 2004, pp. 159-
611 65.
612 [7] L.H. Han, F. Chen, F.Y. Liao, Z. Tao, B. Uy, Fire performance of concrete filled stainless steel tubular
613 columns, Eng. Struct. 56 (2013) 165-181.
614 [8] J. Wang, B. Uy, D. Li, Y. Song, Fatigue behaviour of stainless steel bolts in tension and shear under
615 constant-amplitude loading, Int. J. Fatigue 133 (2020) 105401.
616 [9] B. Young, E. Ellobody, Experimental investigation of concrete-filled cold-formed high strength
617 stainless steel tube columns, J. Constr. Steel Res. 62(5) (2006) 484-492.
618 [10] D. Lam, L. Gardner, Structural design of stainless steel concrete filled columns, J. Constr. Steel Res.
619 64(11) (2008) 1275-1282.
620 [11] B. Uy, Z. Tao, L.H. Han, Behaviour of short and slender concrete-filled stainless steel tubular columns,
621 J. Constr. Steel Res. 67(3) (2011) 360-378.
622 [12] S. Kazemzadeh Azad, D. Li, B. Uy, Axial slenderness limits for austenitic stainless steel-concrete
623 composite columns, J. Constr. Steel Res. 166 (2020) 105856.
624 [13] S. Kazemzadeh Azad, D. Li, B. Uy, Axial slenderness limits for duplex and lean duplex stainless
625 steel-concrete composite columns, J. Constr. Steel Res. 172 (2020) 106175.
626 [14] J. Chapman, The behaviour of composite beams in steel and concrete, Struct. Engr. 42 (1964) 115-
627 125.
628 [15] G. Kwon, M.D. Engelhardt, R.E. Klingner, Experimental behavior of bridge beams retrofitted with
27
629 postinstalled shear connectors, J. Bridg. Eng. 16(4) (2011) 536-545.
630 [16] D. Lam, X. Dai, A. Ashour, N. Rehman, Recent research on composite beams with demountable shear
631 connectors, Steel Constr. 10(2) (2017) 125-134.
632 [17] E.L. Tan, B. Uy, Experimental study on straight composite beams subjected to combined flexure and
633 torsion, J. Constr. Steel Res. 65(4) (2009) 784-793.
634 [18] G. Vasdravellis, B. Uy, E.L. Tan, B. Kirkland, The effects of axial tension on the sagging-moment
635 regions of composite beams, J. Constr. Steel Res. 72 (2012) 240-253.
636 [19] G. Vasdravellis, B. Uy, E.L. Tan, B. Kirkland, Behaviour and design of composite beams subjected
637 to sagging bending and axial compression, J. Constr. Steel Res. 110 (2015) 29-39.
638 [20] R.P. Johnson, R.T. Willmington, Vertical shear in continuous composite beams, Proc. Inst. Civ. Eng.
639 53(2) (1972) 189-205.
640 [21] P. Ansourian, Experiments on continuous composite beams, Proc. Inst. Civ. Eng. 73(1) (1982) 26-51.
641 [22] J. Nie, Y. Xiao, L. Chen, Experimental studies on shear strength of steel–concrete composite beams,
642 J. Struct. Eng. 130(8) (2004) 1206-1213.
643 [23] J. Nie, L. Tang, Vertical shear strength of composite beams with compact steel sections. Part Ⅰ:
644 Composite beams subjected to sagging moment, China Civ. Eng. 3 (2008).
645 [24] G. Vasdravellis, B. Uy, Shear strength and moment-shear interaction in steel-concrete composite
646 beams, J. Struct. Eng. 140(11) (2014) 04014084.
647 [25] Q.Q. Liang, B. Uy, A. Bradford, H.R. Ronagh, Ultimate strength of continuous composite beams in
648 combined bending and shear, J. Constr. Steel Res. 60(8) (2004) 1109-1128.
649 [26] Q.Q. Liang, B. Uy, A. Bradford, H.R. Ronagh, Strength analysis of steel–concrete composite beams
650 in combined bending and shear, J. Constr. Steel Res. 131(10) (2005) 1593-1600.
651 [27] R. Shamass, K. Cashell, Analysis of stainless steel-concrete composite beams, J. Constr. Steel Res.
652 152 (2019) 132-142.
653 [28] Australian/New Zealand standard AS/NZS 2327:2017 Composite Steel-Concrete Construction in
654 Buildings, 2017.
655 [29] EN 1994-1-1. BS EN Eurocode 4: Design of Composite Steel and Concrete Structures, 2004
656 [30] Y. Huang, B. Young, The art of coupon tests, J. Constr. Steel Res. 96 (2014) 159-175.
657 [31] D. Lam, K. Elliott, D. Nethercot, Experiments on composite steel beams with precast concrete hollow
658 core floor slabs, Proc. Inst. Civ. Eng. Struct. Build. 140(2) (2000) 127-138.
659 [32] H.P. Reck, T. Peköz, G. Winter, Inelastic strength of cold-formed steel beams, J. Struct. Div. 101(11)
660 (1975) 2193-2203.
661 [33] ABAQUS, 2016 Documentation, Dassault Systemes, Simulia, Providence, RI, USA, 2016.
662 [34] Y. Song, J. Wang, B. Uy, D. Li, Stainless steel bolts subjected to combined tension and shear:
663 behaviour and design, J. Constr. Steel Res. 170 (2020) 106122.
664 [35] X. Liu, A. Bradford, M.S. Lee, Behavior of high-strength friction-grip bolted shear connectors in
665 sustainable composite beams, J. Struct. Eng. 141(6) (2015) 04014149.
666 [36] Y. Zhou, B. Uy, X. Liu, D. Li, Finite element modelling of stainless steel shear connectors in
667 composite beams, Proceedings of International Association of Structural Engineering and Mechanics,
668 ASEM19 Jeju Island, 2019.
669 [37] H.T. Thai, T.P. Vo, T.K. Nguyen, C.H. Pham, Explicit simulation of bolted endplate composite beam-
670 to-CFST column connections, Thin Wall Struct. 119 (2017) 749-759.
671 [38] J. Wang, B. Uy, D. Li, Behaviour of large fabricated stainless steel beam-to-tubular column joints
672 with extended endplates, Steel Compos. Struct. 32(1) (2019) 141-156.
673 [39] D.J. Carreira, K.-H. Chu, Stress-strain relationship for plain concrete in compression, ACI J. Proc
674 82(6) (1985) 797-804.

28
675 [40] S.W. Pathirana, B. Uy, O. Mirza, X. Zhu, Strengthening of existing composite steel-concrete beams
676 utilising bolted shear connectors and welded studs, J. Constr. Steel Res. 114 (2015) 417-430.
677 [41] S.W. Pathirana, B. Uy, O. Mirza, X. Zhu, Bolted and welded connectors for the rehabilitation of
678 composite beams, J. Const. Steel Res. 125 (2016) 61-73.
679 [42] U. Katwal, Z. Tao, M.K. Hassan, B. Uy, D. Lam, Load sharing mechanism between shear studs and
680 profiled steel sheeting in push tests, J. Constr. Steel Res. 174 (2020) 106279.
681 [43] S.W. Pathirana, B. Uy, O. Mirza, X. Zhu, Flexural behaviour of composite steel–concrete beams
682 utilising blind bolt shear connectors, Eng. Struct. 114 (2016) 181-194.
683 [44] S. Afshan, L. Gardner, The continuous strength method for structural stainless steel design, Thin
684 Walled Struct. 68 (2013) 42-49.
685 [45] A. Liew, L. Gardner, Ultimate capacity of structural steel cross-sections under compression, bending
686 and combined loading, Structure 1 (2015) 2-11.
687 [46] M.N. Su, B. Young, L. Gardner, Testing and design of aluminum alloy cross sections in compression,
688 J. Struct. Eng. 140(9) (2014) 04014047.
689 [47] L. Gardner, X. Yun, L. Macorini, M. Kucukler, Hot-rolled steel and steel-concrete composite design
690 incorporating strain hardening, Structure 9 (2017) 21-28.
691 [48] EN 1993-1-4, Eurocode 3: Design of Steel Structures - Part 1.4: General Rules-Supplementary Rules
692 for Stainless Steel, 2015
693 [49] R.C. Donahey, D. Darwin, Web openings in composite beams with ribbed slabs, J. Struct. Eng. 114(3)
694 (1988) 518-534.
695
696

29
(a) Stainless steel plates (b) Stainless steel shear connectors

Fig. 1 Stress-strain relationship of stainless steel sections


697
698
699

(a) Geometry and test setup of pushout test (b) Load-slip curves of pushout specimens

Fig. 2 Pushout test setup and results


700
701

30
(a) Composite beam cross section and strain gauges arrangement (b) Three-point loading test setup

(c) Four-point loading test setup and cut segments

Fig. 3 Configuration of geometry and instrument setup for stainless steel-concrete composite beam tests
702

CSSB1-BF CSSB2-SF CSSB3-BP CSSB4-SF

CSSB5-BF CSSB6-BF CSSB7-BF CSSB8-SF

Fig. 4 Photos of typical flexural failure modes observed in four-point and three-point bending tests
703

31
(a) Four-point bending test (b) Three-point bending test

Fig. 5 Moment vs. mid-span displacement curves for bending tests


704

(a) Comparison of shear connector types (b) Comparison of shear connection degrees

Fig. 6 Moment vs. end slip curves from bending test


705
706
707

Fig. 7 Strain distribution of stainless steel-concrete composite beams under bending test
708
709
710

32
711
712
713

CSSB9-BF CSSB10-SF CSSB11-BP

CSSB12-SP CSSB13-BF CSSB14-BF

Fig. 8 Photos of typical shear failure modes observed in shear test


714

(a) LD2101 composite beams (b) A304 and D2205 composite beams

Fig. 9 Load vs. mid-span displacement curves from shear test


715
716
717

CSSB15-BF CSSB16-BF CSSB17-BF CSSB18-BF

Fig. 10 Photos of typical failure modes observed in combined bending and shear test
718

33
(a) Composite beams with bolted connectors (b) Composite beams with welded studs

Fig. 11 load vs. mid-span displacement curves from combined bending and shear test
719

Fig. 12 Finite element model of stainless steel-concrete composite beam


720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735

34
(a) CSSB1 (b) CSSB7

(c) CSSB9 (d) CSSB16

Fig. 13 Validation of FE model with experimental results


736
737

35
(a) Rigid plastic analysis (RPA) model

(b) Strain hardening analysis (SHA) model

(c) Simplified SHA model for FSC stainless steel composite beams

(d) Simplified SHA model for PSC stainless steel composite beams

Fig. 14 Analytical models for moment behaviour of stainless steel-concrete composite beams

36
738
739

Fig. 15 Modified bi-linear stainless steel material model Fig. 16 Effect of shear connection degree on moment capacity
740

Fig. 17 Effect of shear connection degree on shear capacity Fig. 18 Moment-shear interaction diagram of stainless steel-concrete composite
beam
741
742

37
743 Table 1. Material test results for stainless steel sections
E f0.01 f0.2 fu
Material εu n m
(GPa) (MPa) (MPa) (MPa)
LD2101-5mm 191 344 500 718 0.36 8.01 3.44
LD2101-8mm 201 321 490 673 0.29 7.08 3.55
A304-5mm 193 178 295 626 0.55 5.93 2.65
Plates
A304-8mm 190 152 265 599 0.54 5.39 2.55
D2205-5mm 182 309 470 683 0.27 7.14 3.41
D2205-8mm 191 344 525 719 0.28 7.09 3.55
D2205bolt 196 157 690 1102 0.035 2.02 3.19
Shear
A316Lbolt 197 158 590 929 0.037 2.27 3.22
connectors
D2205stud 187 281 380 608 0.28 9.93 3.19
Reinforcement D2205 201 503 938 1116 0.024 4.81 3.94
bars A316L 192 288 750 962 0.19 3.13 3.73
744
745 Table 2. Specimen details and obtained results of pushout test
Specimens Materials Connector types PRk (kN) 0.9PRk (kN) su (mm)
S1 D2205 Bolted connector 111 100 10.6
S2 D2205 Welded stud 96 86 9.2
S3 A316L Bolted connector 102 92 9.8
746
747 Table 3. Specimen details of stainless steel-concrete composite beams
Beam Le a Loading Connector Connector Connector
Specimens spacing (mm) β
materials (mm) (mm) method* types materials
CSSB1-BF LD2101 5800 2000 FPL Bolt D2205 200 1
CSSB2-SF LD2101 5800 2000 FPL Stud D2205 200 1
CSSB3-BP LD2101 5800 2000 FPL Bolt D2205 400 0.7
Bending CSSB4-SP LD2101 5800 2000 FPL Stud D2205 400 0.6
test CSSB5-BF A304 5800 2000 FPL Bolt A316L 200 1
CSSB6-BF D2205 5800 2000 FPL Bolt D2205 200 1
CSSB7-BF LD2101 2000 1000 TPL Bolt D2205 200 1
CSSB8-SF LD2101 2000 1000 TPL Stud D2205 200 1
CSSB9-BF LD2101 550 275 TPL Bolt D2205 100 1
CSSB10-SF LD2101 550 275 TPL Stud D2205 100 1
Shear CSSB11-BP LD2101 550 275 TPL Bolt D2205 100 0.7
test ** CSSB12-SP LD2101 550 275 TPL Stud D2205 100 0.6
CSSB13-BF A304 550 275 TPL Bolt A316L 100 1
CSSB14-BF D2205 550 275 TPL Bolt D2205 100 1
CSSB15-BF LD2101 1600 800 TPL Bolt D2205 100 1
Combined
bending CSSB16-BF LD2101 800 400 TPL Bolt D2205 100 1
and shear CSSB17-SF LD2101 1600 800 TPL Stud D2205 100 1
test TPL D2205 100
CSSB18-SF LD2101 800 400 Stud 1

38
748 * FPL: four-point loading; TPL: three-point loading
749 ** CSSB9-BF (Type A) and CSSB11-BP (Type B) were cut from CSSB1-BF; CSSB10-SF (Type A) and CSSB12-SP (Type
750 B) were cut from CSSB2-SF; CSSB13-BF (Type B) was cut from CSSB5-BF; CSSB14-BF (Type A) was cut from CSSB6-B
751 F;
752
753 Table 4. Results of stainless steel composite beams from bending tests and analytical models
Bending Vu Mu Ki.M MRPA MSHA MSHA.S
Specimens M/MRPA M/MSHA M/ MSHA.S
test (kN) (kNm) (kNm/mm) (kNm) (kNm) (kNm)
CSSB1-BF 107 204 2.86 177 1.15 190 1.07 190 1.08
CSSB2-SF 108 206 2.89 177 1.16 190 1.09 190 1.09
CSSB3-BP 102 194 2.45 170 1.14 175 1.11 175 1.11
FPL
CSSB4-SP 101 192 2.73 165 1.16 170 1.13 170 1.13
CSSB5-BF 73 139 2.55 118 1.18 126 1.10 125 1.12
CSSB6-BF 108 205 2.62 177 1.15 190 1.08 190 1.08
CSSB7-BF 213 213 21.74 175 1.22 190 1.12 190 1.12
TPL
CSSB8-SF 218 218 22.35 175 1.24 190 1.15 190 1.15
754
755 Table 5. Results of stainless steel composite beams from shear tests and design approaches
Vu Ki.S VAS/NZS2327 Vu/ VEN 1994-1-1 Vu/ Vnew Vu/
Specimens
(kN) (kN/mm) (kN) VAS/NZS2327 (kN) VEN 1994-1-1 (kN) Vnew
CSSB9-BF 422 552 464 0.91 263 2.10 408 1.03
CSSB10-SF 444 535 464 0.96 263 2.04 408 1.09
CSSB11-BP 411 494 442 0.93 263 1.88 401 1.03
CSSB12-SP 402 572 434 0.93 263 2.22 393 1.02
CSSB13-BF 316 495 348 0.82 147 3.37 313 1.01
CSSB14-BF 450 542 464 0.97 263 2.06 390 1.15
756
757 Table 6. Results of stainless steel composite beams under combined bending and shear tests
Specimens Vu (kN) Vu / Vu.m Mu (kNm) Mu / Mu.m
CSSB7-BF 213 50% 213 100%
CSSB15-BF 262 62% 210 99%
CSSB16-BF 388 92% 155 73%
CSSB9-BF 422 100% 116 54%
CSSB8-SF 218 49% 218 100%
CSSB17-SF 270 61% 216 99%
CSSB18-SF 413 93% 165 76%
CSSB10-SF 444 100% 122 56%
758

39

View publication stats

You might also like