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

net/publication/326403600

A new perspective on nature of fire-induced spalling in concrete

Article in Construction and Building Materials · September 2018

DOI: 10.1016/j.conbuildmat.2018.06.204

CITATIONS READS

58 1,066

3 authors, including:

Kang-Hai Tan Yao Yao

Nanyang Technological University Northwestern Polytechnical University
339 PUBLICATIONS 6,552 CITATIONS 159 PUBLICATIONS 2,013 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Design Guideline and Analytical Tool to Mitigate Progressive Collapse of Buildings against Explosive Effects (GPC) View project

Construction materials and structure reliability under fire attack View project

All content following this page was uploaded by Jin-Cheng Liu on 24 March 2019.

The user has requested enhancement of the downloaded file.

 1 A new perspective on nature of fire-induced spalling in concrete

2 Jin-Cheng Liua, Kang Hai Tana,*, Yao Yaob

3
a
4 School of Civil and Environmental Engineering, Nanyang Technological
5 University, 639798, Singapore.
6
b
7 School of Mechanics, Civil Engineering and Architecture, Northwestern
8 Polytechnical University, ShaanXi 710072, China.
9
10 * Corresponding author.
11 N1-01c-97, 50 Nanyang Avenue, 639798, Singapore.
12 E-mail: ckhtan@ntu.edu.sg.
13

14 Abstract

15 Spalling of concrete is a great potential threat to fire resistance of concrete structures.

16 Understanding the underlying mechanism is important to predict and mitigate this

17 unfavorable phenomenon. Currently, there are two main mechanisms to explain the

18 fire-induced concrete spalling: viz. spalling due to (a) pore pressure buildup or (b)

19 thermal stress. The relative importance of these two mechanisms has been a subject of

20 intense debate in the research community over the past few decades. This paper

21 presents a critical review of conflicting and concordant points on concrete spalling at

22 high temperature and proposes a unified and coherent fire-induced concrete spalling

23 theory. Therein, the authors propose three types of thermal spalling depending on the

24 governing mechanisms: thermo-hygral spalling, thermo-mechanical spalling and

25 thermo-chemical spalling. The criteria to forecast each type of spalling are established

26 and the spalling temperature range for each of them is analysed. The spalling pattern,

1
27 influencing factors and preventive measures for each type of spalling are also

28 discussed in this paper.

29 Keywords

30 Fire-induced spalling, Concrete, Mechanism, High temperature, Fibres, Protective

31 measures.

32

33 Notation

34 t is the scaling factor

35 c,a is the axial compressive stress

36 c,c is the circumferential compressive stress

37 p,t is the tensile stress caused by peak pore pressure

38 t,ni is the tensile stress normal to interface between the concrete cover and core

39  is the porosity of concrete

40 fc is the compressive strength of concrete

41 ft(T) is the temperature-dependent tensile strength of concrete

42 kc(T) is the compressive strength reduction factor of concrete

43 pp is the peak pore pressure

44  is the temperature of concrete

45 TTCS,t is the threshold temperature for thermo-chemical spalling of concrete

46 TTHS,l is the lower bound temperature for thermo-hygral spalling of concrete

47 TTHS,u is the upper bound temperature for thermo-hygral spalling of concrete

48 TTMS,l is the lower bound temperature for thermo-mechanical spalling of concrete

49 TTMS,u is the upper bound temperature for thermo-mechanical spalling of concrete

2
50 1. Introduction

51 Fire-induced spalling is a phenomenon which occurs in concrete under fire. It is

52 characterised by violent and non-violent dislodgement of pieces or chunks of concrete

53 from heated concrete surface. Spalling results in loss of concrete section, reduction in

54 loading-capacity and consequently fire resistance of concrete members.

55

56 Understanding the physics behind this phenomenon is significant for prediction of

57 spalling and advancement in protective measures against spalling. So far mainly two

58 theories have been proposed to explain thermal instability of concrete. The first theory

59 deems that concrete spalling is mainly caused by the buildup of pore pressure, known

60 as pore pressure spalling [1-8]. The second assumes that concrete spalling is mainly

61 caused by thermal stress, known as thermal stress spalling [9-12]. There is also a third

62 school that combines the effects of thermal stress and pore pressure to account for

63 concrete spalling [13-16]. However, no consensus has been reached on fire-induced

64 concrete spalling mechanism to date.

65 This paper compiles conflicting and common views from previous research works on

66 fire-induced concrete spalling. Based on the analyses of these contradictory and

67 concordant test results, a unified and consistent spalling theory is proposed.

68 According to the driving mechanisms, fire-induced spalling can be categorised into

69 three types: thermo-hygral spalling, thermo-mechanical spalling and thermo-chemical

3
70 spalling. The mechanisms of these three types of spalling are explained and the

71 criteria to forecast each of them are established in this paper. They are found to occur

72 within different temperature ranges. The factors that influence them, solutions to

73 prevent them and interactions between them are also discussed.

74

75 2. Conflicting views regarding current work on fire-induced concrete spalling

76 2.1 Spalling mechanism

77 Fire spalling of concrete is commonly explained by two different mechanisms:

78 (a) Pore pressure spalling mechanism is mainly associated with thermal-hygral

79 process inside the heated concrete. Pore pressure builds up gradually in the

80 micro-pores as a result of heat transfer and moisture migration. When the tensile

81 stress induced by pore pressure is greater than the tensile strength of concrete, spalling

82 occurs [3, 7, 8].

83 (b) Concerning thermal stress spalling mechanism, there are two kinds of views.

84 Some researchers believe that spalling is induced by thermal gradient-induced thermal

85 stress [9, 17], while others insist that spalling is caused by compressive stress

86 resulting from restrained thermal dilation [12].

87

4
 88 2.2 Influence of factors

89 2.2.1 Degree of restraint (DOR)

90 Many experimental results have shown that unrestrained and unloaded concrete

91 specimens experienced explosive spalling under high temperature. The DOR in these

92 tests is zero. According to the theory that concrete spalling is induced by restrained

93 thermal dilation, there should be no spalling in these unrestrained concrete specimens.

94 This inference is obviously contradicted by the experimental results. Therefore, it can

95 be concluded that compressive stress generated from restrained thermal dilation is not

96 a necessary condition for occurrence of spalling in concrete at high temperature.

97

98 2.2.2 Thermal gradient

99 According to thermal gradient-induced thermal stress spalling theory, a high thermal

100 gradient leads to a high spalling risk, and similarly, a low thermal gradient leads to a

101 low spalling probability [18, 19]. Hence, concrete under a higher heating rate should

102 be more prone to spalling that concrete under a lower heating rate.

103 However, Noumowe et al. [18] found severe thermal spalling occurred in concrete at a

104 heating rate as low as 0.5 °C/min. Klingsch [20] studied the influence of heating rate

105 on the occurrence of thermal spalling. Low heating rates were used to minimise the

106 thermal graident-induced thermal stress. The concrete specimens were also found to

107 spall at a heating rate of 0.5 °C/min. These experimental facts were not in favor of

5
108 thermal graident-induced thermal stress spalling mechanism.

109 Furthermore, concrete with certain amounts of polypropylene fibre (PPF) subject to

110 extreme higher heating rate was observed to be free from thermal spalling [2, 21]. In

111 contrast, explosive spalling occurred in concrete without PPF even at 0.5 °C/min.

112 Kanéma et al. [1] found that spalling degree of high strength concrete was greater

113 than ordinary strength concrete, although in their tests, thermal gradient in high

114 strength concrete specimens was lower than that in ordinary strength concrete

115 specimens. These experimental results could not explained by thermal

116 gradient-induced thermal stress spalling mechanism.

117 On the other hand, thermal spalling in concrete was observed to occur in different

118 manners: at low heating rates (0.5-1 °C/min, or insulated with thermal barrier),

119 concrete specimens could explode violently characterised by a loud bang [22, 23],

120 whereas at high heating rates (such as in a hydrocarbon fire), concrete specimens

121 could scale off progressively accompanied by popping sounds [24-26]. It would be

122 far-fetched to rely on thermal gradient-induced thermal stress spalling theory to

123 explain two different spalling patterns exhibited by concrete under different heating

124 rates.

125 These aforementioned experimental results were in conflict with the theory that the

126 thermal spalling in concrete was governed by thermal gradient-induced thermal stress.

127 Therefore, thermal gradient-induced thermal stress is not a critical factor contributing

6
128 to spalling of concrete under heating.

129

130 2.2.3 Steel fibres

131 Different conclusions were made about the effectiveness of steel fibres in mitigating

132 spalling of concrete under fire. Klingsch [22] conducted thermal spalling tests on

133 unrestrained and unloaded concrete cylinders with steel fibres and found that steel

134 fibres had no beneficial effect in minimizing the risk of explosive spalling. However,

135 Kodur et al. [27] conducted fire endurance tests on high strength concrete columns

136 and found that steel fibres reduced spalling in concrete columns and enhanced fire

137 endurance of HSC column. The amounts of steel fibres used in the concrete mixes of

138 [22] and [27] are 195 kg/m3 and 42 kg/m3, respectively.

139

140 2.2.4 Time of spalling

141 Observations by Harmathy [28] indicated that spalling of concrete tended to occur

142 within 10-25 min in the case of ASTM E119 fire. Mindeguia et al. [29] observed that

143 thermal spalling occurred in unrestrained and unloaded concrete slab between 10 and

144 20 minute of ISO fire test. Ko et al. [30] observed that spalling occurred in

145 unrestrained and unloaded concrete slabs within 10 min of ISO 834 (similar to the

146 ASTM E119) fire heating.

7
147 However, Franssen and Dotreppe [31] observed surface spalling in concrete columns

148 between 20 and 60 minutes of fire test. Kodur and Mcgrath [32] observed significant

149 spalling at the corners of concrete column after about 1 hour of ASTM E119 fire

150 heating towards the end of fire endurance test. Han et al. [33] found that spalling in

151 steel reinforced concrete columns occurred at the intermediate and later stage of ISO

152 834 fire that lasted about 2.5 hours. Shah and Sharma [34] also observed significant

153 spalling in reinforced concrete columns at the later stages of ISO 834 fire tests which

154 lasted more than 4 hours.

155 According to the pore pressure spalling mechanism, spalling tends to occur when the

156 tensile stress induced by the vapour pressure exceeds the hot tensile strength of

157 concrete. At 374 °C, the saturated vapour pressure reaches about 22 MPa. It takes a

158 few minutes for the exposed concrete cover to reach this critical temperature under

159 fire, so it is reasonable that concrete spalling is observed to occur in the early stage of

160 a fire. However, occurrences of concrete spalling at intermediate or later stage of a

161 fire cannot be explained by pore pressure spalling mechanism.

162

163 2.2.5 Section size

164 Hertz [35] observed explosive spalling of large concrete cylinders when subjected to a

165 heating rate of 1 °C/min. However, reduction of cylinder size by either 75% or 50%

166 eliminated the risk of explosive spalling. Kanéma et al. [1] examined the size effect

8
167 on thermal spalling risk of concrete. He observed that for the same concrete mixture

168 and heating rate, spalling occurred only in large specimens, although thermal gradient

169 in small specimens was similar to that of large specimens. These experimental

170 evidences were in conflict with thermal gradient-induced spalling theory. Restrained

171 thermal dilation-induced thermal stress spalling mechanism also could not explain this

172 phenomenon, since these tested concrete specimens were restrained and unloaded.

173 However, this could be reasonably explained by the pore pressure spalling theory.

174 Smaller dimensions of specimens facilitate quicker escape of moisture out of the

175 concrete, consequently resulting in a lower pore pressure compared to that of large

176 specimens. Therefore, the risk of spalling is reduced as the specimen size decreases.

177

178 2.2.6 Post-cooling spalling

179 Spalling has been observed in concrete not during the heating phase but after cooling

180 down [20]. This type of spalling is referred to as post-cooling spalling [14]. It was

181 possibly caused by the 44% volume expansion resulting from rehydration of calcium

182 oxide after cooling [36]. This rehydration process takes place when moisture is

183 available for concrete. This post-cooling spalling cannot be explained by either

184 thermal stress spalling theory or pore pressure spalling theory.

185

9
186 2.2.7 Others

187 Arioz [37] observed partial spalling in unloaded and unrestrained concrete specimens

188 exposed to 1200 °C for 2 hours. However, no spalling was observed in concrete

189 specimens exposed to 200, 400, 600, 800, 1000 °C for 2 hours. Both pore pressure

190 spalling theory and thermal spalling theory on their own could not explain this

191 phenomenon. This type of spalling is more likely a result of complete destruction of

192 aggregate-cement bond due to significant decomposition of the cement matrix.

193

194 3. Common views regarding current work on fire-induced concrete spalling

195 3.1 Influence of permeability

196 There is a wide agreement on the effect of permeability on fire-induced concrete

197 spalling. Concrete with a lower permeability is more prone to spall under heating,

198 while concrete with a higher permeability has a decreased propensity for thermal

199 spalling.

200

201 3.1.1 Polypropylene fibre

202 Addition of polypropylene (PP) fibres has been found to decrease spalling risk of

203 concrete under fire [38-42]. Different hypotheses (Table 1) are proposed to explain the

204 mechanism of PP fibres in mitigating spalling. Although no consensus has been

10
205 reached on the mode of action of PP fibres in preventing spalling, all these hypotheses

206 lead to a common consequence, which is a significant increase in permeability of

207 concrete. The significant increase in permeability of PP fibre-reinforced concrete has

208 been confirmed by a number of tests [43-45].

209

210 3.1.2 Silica fume

211 Published data shows that concrete with silica fume had a higher risk of thermal

212 spalling than concrete without silica fume [46-48]. Studies have shown that concrete

213 with silica fume densifies the pore structure and decreases permeability [49, 50]. This

214 lower permeability explains the higher propensity for spalling.

215

216 3.2 Influence of moisture

217 It is widely recognised that moisture content of concrete is one critical factor

218 influencing thermally-induced explosive spalling. The higher the moisture content,

219 the greater is the spalling risk, especially when the moisture content exceeds a

220 threshold limit [30, 51, 52].

221

222 4. Unified and consistent fire-induced concrete spalling theory

223 Analysis of experimental results in Section 2 rules out the possibility that thermal

11
224 gradient-induced thermal stress is a critical factor contributing to concrete spalling.

225 Spalling phenomena of many unloaded and unrestrained concrete specimens, together

226 with the effect of permeability and moisture content, confirm that pore pressure

227 build-up is mainly responsible for these observed spalling.

228 However, spalling of concrete cover in columns, occurred at the intermediate or late

229 stage of fire during the fire endurance tests, can hardly be explained by the pore

230 pressure spalling theory. On the other hand, similar cover spalling has been observed

231 in concrete columns loaded in concentric compression at room temperature before the

232 columns reaching the ultimate capacity [53, 54]. Hence, in these cases, initial applied

233 compressive stress and thermal stress generated from restrained thermal dilation other

234 than pore pressure contributes to this type of fire-induced spalling.

235 Sloughing-off spalling at extreme high temperature and post-cooling spalling are due

236 to complete chemical decomposition of major concrete constituents, and much less

237 related to pore pressure build-up and thermal stress.

238 Therefore, there are three types of fire-induced spalling based on the spalling

239 mechanism, viz., thermo-hygral spalling, thermo-mechanical spalling and

240 thermo-chemical spalling. All three types of spalling are possible to occur in concrete

241 elements under fire but most often at different stages of a fire.

242

12
243 4.1 Thermo-hygral spalling

244 4.1.1 Mechanism

245 Thermo-hygral spalling, also known as pore pressure spalling, is related to

246 thermal-hygral behaviour in concrete. Concrete is a composite building material with

247 porous microstructure. Fig. 1 shows a schematic representation of volumetric

248 proportion of concrete components before and during hydration. The hardened

249 concrete contains aggregate, unhydrated cement, hydrated products, free water, gel

250 water and pores. The hydrated products contain chemically-bound water. This is

251 similar for concrete containing pozzolans (silica fume, fly ash, GGBS, etc.) which

252 involves pozzolanic reaction between portlandite and silicic acid.

253 A schematic description of thermo-hygral spalling in a concrete wall heated from one

254 face is presented in Fig. 2. When concrete is under fire, a temperature gradient is

255 formed in the area near fire-exposed side. Gel water and chemically-bound water will

256 be released into micropores of concrete and added to free water as temperature rises.

257 Pore pressure gradually develops as a consequence of temperature rise and presence

258 of moisture (mixed liquid and vapour). A pressure gradient is formed due to

259 temperature gradient and degree of pore saturation. Moisture is driven by pressure

260 gradient to two opposite directions, one direction towards the heated face while the

261 other towards the deeper, cooler region. As a consequence, three zones will be

262 generated, i.e., dry zone, wet zone and saturated zone (the so-called moisture clog).

13
263 When the peak pressure exceeds tensile strength of concrete at elevated temperature,

264 bursting failure of concrete occurs (Fig. 3). The peak pressure is found to be at the

265 saturation front, i.e. the wet zone in front of saturated zone [55, 56]. In the wet zone,

266 the water exists as a mixture of water vapour and liquid water and the pore pressure is

267 equal to the sum of saturated vapour pressure and partial pressure of dry air.

268 Considering the air pressure is negligible in most cases, so this peak pressure can be

269 approximately taken as saturated vapour pressure, which is consistent with the

270 experimental findings [57, 58].

271

272 4.1.2 Spalling temperature

273 The thermo-hygral spalling of concrete is induced by tensile stress p,t caused by peak

274 pore pressure. When the tensile stress exceeds temperature-dependent tensile strength

275 of concrete (Eq. (1)), spalling occurs.

276 t p p  ft (T ) (1)

277 The peak pore pressure pp at the time of spalling is equal to saturation vapour pressure,

278 which is a function of temperature and can be calculated using Eq. (2) [59]. Eq. (2) is

279 valid within the temperature range between 273.16 K and 647.096 K.

p  T
280 ln  p   c  A1   A 2  1.5  A3  3  A 4  3.5  A5  4  A 6  7.5  (2)
 pc  T

281 with   1  T / Tc , Tc  647.096 K , pc  22.064 MPa , A1=-7.85951783,

14
282 A2=1.84408259, A3=-11.7866497, A4=22.6807411, A5=-15.9618719, A6=1.80122502,

283 and T is temperature (K).

284 The scaling factor t is used to calculate tensile stress induced by pore pressure. If the

285 hollow spherical model is adopted to approximate concrete skeleton, the scaling factor

286 is calculated as a function of porosity  using Eq. (3) [7].

1  2
287 t  (3)
2(1   )

288 However, the micro-pores in concrete are not spherical, and using Equation 2 to

289 calculate t tends to underestimate p,t. According to Tenchev and Purnell [60] and

290 Klingsch [20], t is more reasonably taken as unity. If temperature-dependent tensile

291 strength reduction factor kt(T) and concrete tensile strength are known, the concrete

292 temperature at spalling location can be determined.

293 The compressive strength of concrete in research works lies in the range of 30-200

294 MPa. Assuming that the tensile strength of concrete is 1/10 of its corresponding

295 compressive strength, the lower and upper bound values of tensile strength are 3 MPa

296 and 20 MPa, respectively. The temperature-dependent tensile strength reduction factor

297 of concrete specified in Eurocode 2 [61] is adopted. According to the criterion for

298 thermo-hygral spalling, the concrete temperature at the spalled location (spalling

299 temperature) would be in the range between TTHS,l (220 °C) and TTHS,u (320 °C) as

300 shown in Fig. 4, if thermo-hygral spalling would occur.

15
301 Table 2 shows the temperature information at spalling of unrestrained and unloaded

302 specimens provided by different researchers. Although it is difficult to measure the

303 temperature at the spalled location, the spalling temperature should fall between the

304 recorded surface temperature and core temperature. It can be seen that the theoretical

305 thermo-hygral spalling temperature range generally matches the experimental results.

306

307 4.1.3 Spalling pattern

308 The occurrence of thermal-hygral spalling is accompanied with the release of energy.

309 Hence, the spalled-off concrete pieces break away from concrete elements with a

310 certain initial velocity. In the experiments conducted by Zeiml et al. [62], the

311 maximum observed initial velocity of spalled-off piece reached up to 14 m/s. Hence,

312 the term ―explosive spalling‖ mentioned in previous published works refers in

313 particular to thermal-hygral spalling in the current framework of spalling theory

314 proposed in this paper. When concrete is under slow heating, spalling behaviour tends

315 to be a sudden bursting failure with a loud bang; when concrete is under fast heating,

316 spalling behaviour is dominated by progressive peeling-off with continuous popping

317 sound. Both the one-time explosive spalling and progressive spalling occur in a

318 violent manner.

319

16
320 4.1.4 Influencing factors

321 Permeability: Moisture migration is dependent on permeability of concrete. Low

322 permeability means the moisture trapped in concrete evacuates slowly and facilitates

323 the buildup of pore pressure. Concrete with lower permeability has a higher

324 susceptibility to thermo-hygral spalling, and vice versa. Density and compressive

325 strength of concrete in most cases can serve as a indicator of permeability. High

326 density and high compressive strength usually indicate a low permeability, but this is

327 not always the case.

328 Degree of pore saturation (DPS): Moisture content of concrete has been widely used

329 as one parameter for assessing the thermo-hygral spalling risk. Eurocode 2

330 recommends that explosive spalling is unlikely to occur when the moisture content of

331 concrete is less than 3% by weight. However, moisture-rich concrete may be free

332 from thermo-hygral spalling if the porosity of concrete is high; moreover, concrete

333 with moisture content less than 3% may suffer from thermo-hygral spalling if the

334 porosity is low. Therefore, it is more scientific to use the DPS to assess the explosive

335 spalling risk. A larger DPS tends to increase the thermo-hygral spalling risk. The DPS

336 also influences permeability of concrete. Increased DPS decreases permeability and

337 vice versa. In practice, the DPS of indoor concrete elements decreases as concrete

338 ages, and consequently, explosive spalling risk decreases. However, this is not the

339 case for other situations where concrete elements are exposed to natural environment,

17
340 such as tunnels and offshore structures.

341 Heating rate: A high heating rate tends to result in an early spalling of concrete

342 featured by progressive scaling-off. The extent of thermo-hygral spalling of concrete

343 specimens when subjected to a high heating rate is normally higher than when

344 subjected to a low heating rate. But it is not the case for ultra high performance

345 concrete (UHPC) with extremely low permeability. For UHPC specimens under a low

346 heating rate, one-off explosive spalling may occur which completely disintegrates the

347 specimens.

348 Section dimension: The risk of thermo-hygral spalling increases with the specimen

349 dimension. This is due to the fact that it takes a longer time for trapped moisture to

350 evacuate for a larger specimen. The dimensions of concrete elements used in

351 practice are large. Hence, it is more realistic to use large specimens to evaluate

352 thermo-hygral spalling risk of concrete.

353 Aggregate: Using flint as aggregate of concrete induces aggregate spalling. This

354 aggregate spalling results from thermal instability of flint, which is due to high vapour

355 pressure that builds up inside the laminar microstructure of flint [63]. So this type of

356 spalling is categorised into thermo-hygral spalling.

357

358 4.1.5 Protective measures

359 The protective measures to mitigate thermo-hygral spalling and their disadvantages
18
360 are presented in Table 3. Based on the factors influencing thermo-hygral spalling,

361 thermo-hygral spalling risk of concrete can be minimised by taking adequate

362 measures. There is little flexibility to change DPF and section dimensions. But

363 concrete permeability, heating rate and aggregate type are within the control of

364 engineers. It should be noted that although increasing permeability of concrete

365 reduces explosive spalling risk, it is adverse to durability of concrete. Therefore, a

366 balanced approach is needed between spalling resistance and durability when

367 undertaking concrete mix design.

368

369 4.2 Thermo-mechanical spalling

370 4.2.1 Mechanism

371 A schematic description of the thermo-mechanical spalling in concrete is presented in

372 Fig. 5. When a concrete member is under fire, thermal stress develops inside the

373 concrete due to temperature gradients across the member section and restrained

374 thermal dilation. Therefore, the exterior surface layer of concrete is in a triaxial

375 tension-compression-compression stress state. At first, restrained thermal

376 dilation-induced compressive stress on exterior surface layer is the largest and

377 increases with temperature. When the exterior surface concrete layer reaches its

378 ultimate capacity, the peak of compressive stress profile moves inwards and acts on

379 adjacent interior layer. This process repeats itself until the peak of compressive stress

19
380 profile moves inwards to the interface between the concrete cover and the core. A

381 point at the interface is subjected to axial compressive stress c,a, circumferential

382 compressive stress c,c and tensile stress normal to the interface between the concrete

383 cover and the core t,ni. The tensile stress arises partly from thermal gradient-induced

384 tensile stress, and partly from tensile stress set up at the cover-core interface due to

385 confined effect provided by hoops. When the tensile stress reaches a critical value,

386 approximately vertical cracks will form between concrete cover and concrete core.

387 Interface cracking in itself is not sufficient for concrete cover to break away from the

388 core. Inspection of concrete columns after fire tests conducted by [64] showed that

389 vertical cracks, about 50 mm into the column, had formed. This outer layer of

390 concrete, although not fell off, could be easily removed by hand. Therefore, for

391 thermo-mechanical spalling to occur, a driving force is required. The driving force

392 may be due to the compressive stress exerted on cracked concrete cover, the

393 volumetric expansion of concrete core, or bending in longitudinal reinforcement.

394

395 4.2.2 Spalling temperature

396 Two conditions are needed for the occurrence of thermo-mechanical spalling in a

397 concrete member. The first condition is the formation of longitudinal cracks, and the

398 second is the driving force to separate the concrete cover from the interior core. It is

399 difficult to judge when thermo-mechanical spalling occurs in the finite element model,

20
400 and in fact currently no numerical model is available to predict the occurrence of this

401 type of spalling. A simplified method is proposed here to predict roughly the

402 temperature at the spalled location.

403 The compressive stress in concrete at service condition is typically in the range of

404 0.1fc-0.3fc. The maximum recorded compressive stresses in concrete with an initial

405 stress of 0.1fc and 0.3fc under fully restrained condition are 0.63fc and 0.75fc,

406 respectively, upon heating to 800 °C. Knowing that the compressive stress in heated

407 concrete with no restraint is in the range of 0.1fc-0.3fc, it is reasonable to assume

408 that the peak compressive stress in heated concrete under realistic restraint condition

409 is in the range of 0.5fc and 0.7fc. The compressive strength reduction factor kc(T)

410 specified in Eurocode 2 is adopted to account for the deterioration in compressive

411 strength of concrete with elevated temperature. There are in general two types of

412 concrete based on aggregate types used in concrete: siliceous concrete and calcarous

413 concrete, i.e., concrete with siliceous and calcareous aggregates. As shown in Fig. 6,

414 Eurocode 2 gives different compressive strength reduction factors for the two types of

415 concrete as a function of temperature. It is believed that spalling tends to occur when

416 the concrete cover reaches its instantaneous compressive strength and enters

417 post-peak stage. As shown in Fig. 6, the lower limit of thermo-mechanical spalling

418 temperature TTMS,l is attained when siliceous concrete cover layer under heating

419 reaches a peak compressive stress of 0.7fc; the upper limit of thermo-mechanical

420 spalling temperature TTMS,u is attained when calcareous concrete cover layer under
21
421 heating reaches a peak compressive stress of 0.5fc. Then the concrete temperature at

422 the cover-core interface would be roughly in the range between TTMS,l (430 °C) and

423 TTMS,u (660 °C), if thermo-mechanical spalling were to occur.

424

425 4.2.3 Spalling pattern

426 Thermo-mechanical spalling of concrete occurs in the form of cover spalling or corner

427 spalling. It is not explosive although the popping sound arising from the formation of

428 longitudinal cracks may be alarming.

429

430 4.2.4 Influencing factors

431 Aggregate: Concrete with siliceous aggregate is more susceptible to

432 thermo-mechanical spalling than concrete with calcareous aggregate, since thermal

433 expansion of siliceous concrete is larger than calcareous concrete and strength

434 degradation of siliceous concrete is faster than calcareous concrete.

435 Load level: Increasing load level tends to increase the severity of thermo-mechanical

436 spalling, and vice versa.

437 DOR: Increasing DOR tends to increase the severity of thermo-mechanical spalling,

438 and vice versa.

439 Exposure temperature level: Limit the temperature below potential threshold

22
440 temperature for thermo-mechanical spalling can reduce spalling risk to a minimum.

441 Brittleness: Good ductility of concrete materials at high temperature reduces

442 thermo-mechanical spalling risk, since fibre bridging action can maintain integrity

443 after the formation of longitudinal cracks in concrete elements. Also the existence of

444 fibres can delay the formation of longitudinal cracks. Keeping the concrete cover

445 attached to the concrete core after crack formation is beneficial for fire resistance of

446 structural members. Although the surface concrete layer contributes little to bearing

447 capacity, it protects the longitudinal reinforcement from directly exposing to fire and

448 increases fire resistance.

449

450 4.2.5 Protective measures

451 The protective measures to mitigate thermo-mechanical spalling and their

452 disadvantages are presented in Table 3. Based on the factors influencing

453 thermo-mechanical spalling, using calcareous aggregate instead of siliceous aggregate,

454 decreasing load level, adopting fire-proof coating and adding steel fibres can combat

455 thermo-mechanical spalling.

456

457 4.3 Thermo-chemical spalling

458 4.3.1 Mechanism

459 There are two types of thermo-chemical spalling, viz. one is sloughing-off spalling at
23
460 extreme high temperatures, and the other is post-cooling spalling after exposure to

461 very high temperatures.

462 Sloughing-off spalling occurs at extreme high temperature and results from significant

463 destruction in aggregate-cement bond. One important source for reduction in the

464 aggregate-cement bond is breakdown of calcium hydroxide, calcium-silicate-hydrates,

465 and calcarerous aggreagre, etc. At high temperature cement paste shrinks while

466 aggregates expand. This difference further weakens bond between aggregates and

467 cement paste.

468 Post-cooling spalling occurs after exposure to fire. When concrete is heated to

469 sufficiently high temperatures, calcium carbonate in calcarerous aggreagre

470 decomposes and produces calcium oxide. After cooling down, mositure in the air is

471 absorbed by heated concrete and penerates deep into concrete gradually. The newly

472 generated calcium oxide in the surface concrete layer reacts with the absorbed

473 mositure, which causes expansion (44% increase in volume) in the surface layer. The

474 expansion due to rehydration leads to severe cracking and consequently falling of

475 concrete pieces. Post-cooling spalling may occur in a progressive manner if there are

476 sufficient moisture from the ambient air and sufficient calcium oxide in deeper

477 concrete region. But this type of progressive spalling is much slower than that one in

478 the case of thermo-hygral spalling. It can take a couple of weeks to complete the

479 progressive post-cooling spalling. So far, the knowledge of detailed post-cooling

24
480 spalling mechanism is rather limited. But it reduces service life of concrete buildings,

481 and could even lead to collapse of concrete buildings. Therefore, it is necessary to

482 delve deeper into post-cooling spalling mechanism of concrete.

483

484 4.3.2 Spalling temperature

485 According to the thermo-chemical spalling mechanism, sloughing-off spalling and

486 post-cooling spalling can be linked to the maximum temperature in concrete, i.e.,

487 when the maxmimum temperature in concrete exceeds a threshold, thermo-chemical

488 spalling occurs. The threshold temperature for thermo-chemical spalling, TTCS,t, is

489 related to decarbonation of calcium carbonate and complete dehydration of concrete.

490 Decomposition of calcium carbonate typically occurs between 700 °C and 900 °C

491 [65]. Complete dehyration of concrete usually occurs at a temperature above 800 °C.

492 Thus, it is assumed that the threshold temperature TTCS,t for thermo-chemical spalling

493 is about 700 °C (Fig. 7). When temperature of concrete exceeds TTCS,t,

494 thermo-chemical spalling is likely to occur, especially for calcareous concrete.

495 Table 4 summarises the temperature information at thermo-chemical spalling of

496 unrestrained and unloaded concrete specimens provided by different researchers. It

497 can be seen that the estimated thermo-chemical spalling temperature range generally

498 matches the observed experimental results.

25
499 4.3.3 Spalling pattern

500 Thermo-chemical spalling of concrete occurs in the form of gentle sloughing-off.

501 Although post-cooling spalling is not explosive, if neglected, it could possibly lead to

502 unexpected failure of concrete structures. Therefore, assessment of fire-damaged

503 concrete must include evaluation of post-cooling spalling risk.

504

505 4.3.4 Influencing factors

506 Calcareous aggregate: Decomposition of calcareous aggregate generates calcium

507 oxide. Rehydration of calcium oxide is believed to be the main cause for

508 thermo-chemical spalling.

509 Cement content: Dehydration of calcium hydroxide generates calcium oxide, so it is

510 not recommended to use excessive cement in unit volume of concrete.

511 Exposure temperature level: Thermo-chemical spalling usually occurs at or after

512 temperature higher than 700 °C. Limiting the temperature to below 700 °C can reduce

513 thermo-chemical spalling risk to a minimum.

514

515 4.3.5 Protective measures

516 The protective measures to mitigate thermo-chemical spalling and their disadvantages

517 are presented in Table 3. Based on the factors affecting thermo-chemical spalling,

26
518 avoiding using calcareous aggregate, limiting cement content and using fire-proof

519 coating can mitigate thermo-chemical spalling.

520

521 5. Discussion

522 Behaviour of structural concrete under fire is indeed a coupled

523 thermo-hygro-chemo-mechanical problem. There is interaction between

524 thermo-hygral spalling, thermo-mechanical spalling and thermo-chemical spalling.

525 Under the effect of applied stress, there is a small decrease in permeability up to a

526 certain threshold level. When the threshold stress is exceeded, there will be a

527 significant increase in concrete permeability [66, 67]. Therefore, mechanical damage

528 induced by external load is beneficial to mitigate thermo-hygral spalling.

529 Thermo-hygral spalling reduces cross-section area and increases eccentricity of

530 loading, which tends to induce thermo-mechanical spalling. Damage induced by

531 internal pore pressure and external load increases thermo-chemical spalling risk.

532 Current research work focuses more on spalling of unstressed concrete members, and

533 much less is done on spalling of stressed or restrained concrete. It is common that

534 thermo-hygral spalling and thermo-mechanical spalling are mistaken as the same type

535 of spalling. Future research should be conducted on spalling of stressed or restrained

536 concrete, especially spalling in structural concrete members, since results from

537 unstressed spalling tests can be misleading and impractical.

27
538 Current codes of practice only prescribe some measures against thermo-hygral

539 spalling of concrete. But no measures are given to combat thermo-mechanical spalling

540 and thermo-chemical spalling. Multiple defense lines against fire-induced material

541 deterioration (strength deterioration and three types of spalling) can be used in

542 engineering. For example, hybrid PP and steel fibres can be used in UHPC columns to

543 minimize thermo-mechanical spalling and thermo-chemical spalling; fire-proof

544 coating, hybrid PP and steel fibres can be used simultaneously in UHPC tunnel linings,

545 considering much more intense hydrocarbon fire scenario and the loss as a result of

546 failure of a tunnel.

547

548 6. Conclusions

549 This paper presents a new perspective on fire-induced spalling in concrete. The main

550 conclusions are summarised as follows.

551 (a) There are three types of fire-induced concrete spalling depending on the

552 mechanisms, viz., thermo-hygral, thermo-mechanical and thermo-chemical

553 spalling.

554 (b) Thermo-hygral spalling is induced by moisture clogging and pore pressure

555 buildup inside heated concrete. It typically occurs within 220 °C and 320 °C (i.e.

556 the temperature at the spalled location).

557 (c) Thermo-mechanical spalling is caused by applied stress and restraint-induced

28
558 thermal stress in heated concrete. It typically occurs within 430 °C and 660 °C

559 (i.e. the temperature at the spalled location).

560 (d) Thermo-chemical spalling is related to decomposition of hydrated products and

561 calcite, and rehydration of calcium oxide. It typically occurs at or after

562 temperature greater than 700 °C (i.e. the temperature at the spalled location).

563 (e) There is interaction among these three types of spalling. There is a lack of

564 research work on spalling behaviour of stressed or restrained concrete specimens,

565 especially on full-scale structural members. Test results on stressed or restrained

566 concrete is more useful for designing structural concrete against thermal spalling.

567 (f) Current codes of practice only recommend protective measures against

568 thermo-hygral spalling. Countermeasures against the three types of spalling are

569 presented in this paper. The concept of ‗multiple defense line against fire-induced

570 concrete spalling‘ is introduced for the highest level of protection.

571 (g) Future predictive tools aiming to assess the spalling risk and to determine the

572 spalling extent should consider all three types of spalling and the interactions

573 among them. Furture work can be done to fine-tune the spalling temperature

574 range for each type of fire-induced concrete spalling.

575

29
576 Acknowledgement

577 This material is based on research/work supported by the Land and Liveability

578 National Innovation Challenge (L2NIC) Award No. L2NICCFP1-2013-4.

579

580 Disclaimer

581 Any opinions, findings, and conclusions or recommendations expressed in this

582 material are those of the author(s) and do not necessarily reflect the views of the L2

583 NIC.

584

585 References

586 [1] M. Kanéma, P. Pliya, A. Noumowé, J. Gallias, Spalling, thermal, and hydrous
587 behavior of ordinary and high-strength concrete subjected to elevated temperature,
588 ASCE J Mater Civ Eng 23(7) (2011) 921-930.

589 [2] L.T. Phan, Pore pressure and explosive spalling in concrete, Mater Struct 41(10)
590 (2008) 1623-1632.

591 [3] M.B. Dwaikat, V. Kodur, Hydrothermal model for predicting fire-induced spalling
592 in concrete structural systems, Fire Safety J 44(3) (2009) 425-434.

593 [4] F.L. Monte, F. Lombardi, R. Felicetti, M. Lualdi, Ground-Penetrating Radar

594 monitoring of concrete at high temperature, Constr Build Mater 151 (2017) 881-888.

595 [5] G. Shorter, T. Harmathy, Moisture clog spalling, Proceedings of institution of civil
596 engineers 20 (1965) 75-90.

597 [6] M. Dwaikat, V. Kodur, Response of restrained concrete beams under design fire
598 exposure, Journal of structural engineering 135(11) (2009) 1408-1417.

599 [7] Y. Ichikawa, G. England, Prediction of moisture migration and pore pressure
600 build-up in concrete at high temperatures, Nucl Eng Des 228(1) (2004) 245-259.

30
601 [8] M. Ozawa, H. Morimoto, Effects of various fibres on high-temperature spalling in
602 high-performance concrete, Constr Build Mater 71 (2014) 83-92.

603 [9] J. Zhao, J.-J. Zheng, G.-F. Peng, K. van Breugel, A meso-level investigation into
604 the explosive spalling mechanism of high-performance concrete under fire exposure,
605 Cem Concr Res 65 (2014) 64-75.

606 [10] F.-J. Ulm, O. Coussy, Z.P. Bažant, The ―Chunnel‖ fire. I: Chemoplastic softening
607 in rapidly heated concrete, J Eng Mech 125(3) (1999) 272-282.

608 [11] H. Zhang, C. Davie, A numerical investigation of the influence of pore pressures
609 and thermally induced stresses for spalling of concrete exposed to elevated
610 temperatures, Fire Safety J 59 (2013) 102-110.

611 [12] Y. Msaad, G. Bonnet, Analyses of heated concrete spalling due to restrained
612 thermal dilation: Application to the ―chunnel‖ fire, J Eng Mech 132(10) (2006)
613 1124-1132.

614 [13] R.J. Connolly, The spalling of concrete in fires, Aston University, 1995.

615 [14] G.A. Khoury, Passive fire protection of concrete structures, Proceedings of the
616 Institution of Civil Engineers-Structures and Buildings 161(3) (2008) 135-145.

617 [15] Y. Sertmehmetoglu, On a mechanism of spalling of concrete under fire conditions,

618 University of London, 1977.

619 [16] J.C. Mindeguia, H. Carré, P. Pimienta, C. La Borderie, Experimental discussion

620 on the mechanisms behind the fire spalling of concrete, Fire Mater 39(7) (2015)
621 619-635.

622 [17] Y. Fu, L. Li, Study on mechanism of thermal spalling in concrete exposed to
623 elevated temperatures, Mater Struct 44(1) (2011) 361-376.

624 [18] A. Noumowe, H. Carré, A. Daoud, H. Toutanji, High-strength self-compacting

625 concrete exposed to fire test, ASCE J Mater Civ Eng 18(6) (2006) 754-758.

626 [19] R. Jansson, L. Boström, The influence of pressure in the pore system on fire
627 spalling of concrete, Fire technology 46(1) (2010) 217-230.

628 [20] E. Klingsch, Explosive spalling of concrete in fire, ETH-Zürich, 2014.

629 [21] C.-G. Han, Y.-S. Hwang, S.-H. Yang, N. Gowripalan, Performance of spalling
630 resistance of high performance concrete with polypropylene fiber contents and lateral
631 confinement, Cem Concr Res 35(9) (2005) 1747-1753.

632 [22] E. Klingsch, Explosive spalling of concrete in fire, ETH-Zürich, 2014.

633 [23] L. Phan, J. Lawson, F. Davis, Effects of elevated temperature exposure on

634 heating characteristics, spalling, and residual properties of high performance concrete,

31
635 Mater Struct 34(2) (2001) 83-91.

636 [24] R. Jansson, Fire Spalling of Concrete: Theoretical and Experimental Studies,
637 KTH Royal Institute of Technology, 2013.

638 [25] K.D. Hertz, Limits of spalling of fire-exposed concrete, Fire Safety J 38(2) (2003)
639 103-116.

640 [26] J.-C. Liu, K.H. Tan, Fire resistance of strain hardening cementitious composite
641 with hybrid PVA and steel fibers, Constr Build Mater 135 (2017) 600-611.

642 [27] V. Kodur, F.-P. Cheng, T.-C. Wang, M. Sultan, Effect of strength and fiber
643 reinforcement on fire resistance of high-strength concrete columns, Journal of
644 Structural Engineering 129(2) (2003) 253-259.

645 [28] T. Harmathy, Effect of moisture on the fire endurance of building elements,
646 Moisture in Materials in Relation to Fire Tests, ASTM International1965.

647 [29] J.-C. Mindeguia, P. Pimienta, H. Carré, C. La Borderie, Experimental study on

648 the contribution of pore vapour pressure to the thermal instability risk of concrete, 1st
649 International Workshop on Concrete Spalling due to Fire Exposure, 2009.

650 [30] J. Ko, D. Ryu, T. Noguchi, The spalling mechanism of high-strength concrete
651 under fire, Mag Concrete Res 63(5) (2011) 357-370.

652 [31] J.-M. Franssen, J.-C. Dotreppe, Fire tests and calculation methods for circular
653 concrete columns, Fire technology 39(1) (2003) 89-97.

654 [32] V. Kodur, R. Mcgrath, Fire endurance of high strength concrete columns, Fire
655 technology 39(1) (2003) 73-87.

656 [33] L.-H. Han, Q.-H. Tan, T.-Y. Song, Fire performance of steel reinforced concrete
657 columns, Journal of Structural Engineering 141(4) (2014) 04014128.

658 [34] A.H. Shah, U. Sharma, Fire resistance and spalling performance of confined
659 concrete columns, Constr Build Mater 156 (2017) 161-174.

660 [35] K. Hertz, Heat-induced explosion of dense concretes, Lyngby, Technical

661 University of Denmark, Institute of Building Design Report (166) (1984).

662 [36] E. Annerel, L. Taerwe, Revealing the temperature history in concrete after fire
663 exposure by microscopic analysis, Cem Concr Res 39(12) (2009) 1239-1249.

664 [37] O. Arioz, Retained properties of concrete exposed to high temperatures: size
665 effect, Fire Mater 33(5) (2009) 211-222.

666 [38] G. Khoury, Polypropylene fibres in heated concrete. Part 2: Pressure relief
667 mechanisms and modelling criteria, Mag Concrete Res 60(3) (2008) 189-204.

668 [39] R. Jansson, L. Boström, Factors influencing fire spalling of self compacting
32
669 concrete, Mater Struct 46(10) (2013) 1683-1694.

670 [40] K. Sideris, P. Manita, Residual mechanical characteristics and spalling resistance
671 of fiber reinforced self-compacting concretes exposed to elevated temperatures,
672 Constr Build Mater 41 (2013) 296-302.

673 [41] A. Bilodeau, V. Kodur, G. Hoff, Optimization of the type and amount of
674 polypropylene fibres for preventing the spalling of lightweight concrete subjected to
675 hydrocarbon fire, Cem Concr Compos 26(2) (2004) 163-174.

676 [42] Y. Ding, C. Zhang, M. Cao, Y. Zhang, C. Azevedo, Influence of different fibers
677 on the change of pore pressure of self-consolidating concrete exposed to fire, Constr
678 Build Mater 113 (2016) 456-469.

679 [43] P. Kalifa, G. Chene, C. Galle, High-temperature behaviour of HPC with

680 polypropylene fibres: From spalling to microstructure, Cem Concr Res 31(10) (2001)
681 1487-1499.

682 [44] M. Zeiml, D. Leithner, R. Lackner, H.A. Mang, How do polypropylene fibers
683 improve the spalling behavior of in-situ concrete?, Cem Concr Res 36(5) (2006)
684 929-942.

685 [45] J. Bošnjak, J. Ožbolt, R. Hahn, Permeability measurement on high strength

686 concrete without and with polypropylene fibers at elevated temperatures using a new
687 test setup, Cem Concr Res 53 (2013) 104-111.

688 [46] K. Hertz, Explosion of silica-fume concrete, Fire Safety J 8(1) (1984) 77.

689 [47] A. Noumowe, P. Clastres, G. Debicki, J.-L. Costaz, Transient heating effect on
690 high strength concrete, Nucl Eng Des 166(1) (1996) 99-108.

691 [48] C. Castillo, Effect of transient high temperature on high-strength concrete, Rice
692 University, 1987.

693 [49] M. Khan, C. Lynsdale, Strength, permeability, and carbonation of

694 high-performance concrete, Cem Concr Res 32(1) (2002) 123-131.

695 [50] A. Ramezanianpour, V. Malhotra, Effect of curing on the compressive strength,

696 resistance to chloride-ion penetration and porosity of concretes incorporating slag, fly
697 ash or silica fume, Cem Concr Compos 17(2) (1995) 125-133.

698 [51] S.Y.N. Chan, G.-F. Peng, M. Anson, Fire behavior of high-performance concrete
699 made with silica fume at various moisture contents, ACI Mater J 96(3) (1999)
700 405-409.

701 [52] C. Meyer-Ottens, The question of spalling of concrete structural elements of

702 standard concrete under fire loading, Braunschweig: Technical University of
703 Braunschweig (1972).

33
704 [53] S.J. Foster, On behavior of high-strength concrete columns: cover spalling, steel
705 fibers, and ductility, ACI Structural Journal 98(4) (2001).

706 [54] S.J. Foster, J. Liu, S.A. Sheikh, Cover spalling in HSC columns loaded in
707 concentric compression, Journal of Structural Engineering 124(12) (1998) 1431-1437.

708 [55] M.R. Bangi, T. Horiguchi, Pore pressure development in hybrid fibre-reinforced
709 high strength concrete at elevated temperatures, Cem Concr Res 41(11) (2011)
710 1150-1156.

711 [56] M. Ozawa, S. Uchida, T. Kamada, H. Morimoto, Study of mechanisms of

712 explosive spalling in high-strength concrete at high temperatures using acoustic
713 emission, Constr Build Mater 37 (2012) 621-628.

714 [57] J.-C. Mindeguia, P. Pimienta, A. Noumowé, M. Kanema, Temperature, pore

715 pressure and mass variation of concrete subjected to high temperature—experimental
716 and numerical discussion on spalling risk, Cem Concr Res 40(3) (2010) 477-487.

717 [58] P. Kalifa, F.-D. Menneteau, D. Quenard, Spalling and pore pressure in HPC at
718 high temperatures, Cem Concr Res 30(12) (2000) 1915-1927.

719 [59] W. Wagner, A. Pruss, International equations for the saturation properties of
720 ordinary water substance. Revised according to the international temperature scale of
721 1990. Addendum to J. Phys. Chem. Ref. Data 16, 893 (1987), J Phys Chem Ref Data
722 22(3) (1993) 783-787.

723 [60] R. Tenchev, P. Purnell, An application of a damage constitutive model to concrete

724 at high temperature and prediction of spalling, International Journal of Solids and
725 Structures 42(26) (2005) 6550-6565.

726 [61] B.S. Institution, Eurocode 2: Design of concrete structures: Part 1-2: General
727 rules-Structural fire design, London, 2004.

728 [62] M. Zeiml, R. Lackner, H.A. Mang, Experimental insight into spalling behavior of
729 concrete tunnel linings under fire loading, Acta Geotechnica 3(4) (2008) 295-308.

730 [63] Z. Xing, A.-L. Beaucour, R. Hebert, A. Noumowe, B. Ledesert, Influence of the
731 nature of aggregates on the behaviour of concrete subjected to elevated temperature,
732 Cem Concr Res 41(4) (2011) 392-402.

733 [64] C. Bailey, Holistic behaviour of concrete buildings in fire, Proceedings of the
734 Institution of Civil Engineers-Structures and Buildings 152(3) (2002) 199-212.

735 [65] B.-T. Chen, T.-P. Chang, J.-Y. Shih, J.-J. Wang, Estimation of exposed
736 temperature for fire-damaged concrete using support vector machine, Comp Mater Sci
737 44(3) (2009) 913-920.

738 [66] M. Choinska, A. Khelidj, G. Chatzigeorgiou, G. Pijaudier-Cabot, Effects and

34
739 interactions of temperature and stress-level related damage on permeability of
740 concrete, Cem Concr Res 37(1) (2007) 79-88.

741 [67] M. Hoseini, V. Bindiganavile, N. Banthia, The effect of mechanical stress on

742 permeability of concrete: a review, Cem Concr Compos 31(4) (2009) 213-220.

743 [68] K. Pistol, F. Weise, B. Meng, U. Schneider, The mode of action of polypropylene
744 fibres in high performance concrete at high temperatures, Proceedings of the 2nd
745 International RILEM Workshop on Concrete Spalling due to Fire Exposure, 2011, pp.
746 289-296.

747 [69] H. Fares, A. Noumowe, S. Remond, Self-consolidating concrete subjected to

748 high temperature: mechanical and physicochemical properties, Cem Concr Res 39(12)
749 (2009) 1230-1238.

750 [70] G. Debicki, R. Haniche, F. Delhomme, An experimental method for assessing the
751 spalling sensitivity of concrete mixture submitted to high temperature, Cem Concr
752 Compos 34(8) (2012) 958-963.

753 [71] B. Akturk, N. Yuzer, N. Kabay, Usability of Raw Rice Husk Instead of
754 Polypropylene Fibers in High-Strength Concrete under High Temperature, ASCE J
755 Mater Civ Eng 28(1) (2015) 04015072.

756 [72] J.-C. Liu, K.H. Tan, Fire resistance of ultra-high performance strain hardening
757 cementitious composite: Residual mechanical properties and spalling resistance, Cem
758 Concr Compos 89 (2018) 62-75.

759 [73] J. Xiao, M. Xie, C. Zhang, Residual compressive behaviour of pre-heated

760 high-performance concrete with blast–furnace–slag, Fire Safety J 41(2) (2006) 91-98.

761

762

763

764

765

766

35
767 Tables

768 Table 1 Different modes of action of PP fibres in combating spalling

Hypotheses Description

Vacated micro-channels theory Micro-channels left by the melted fibres

[43] facilitate the escape of the steam.

The ITZ between hydrophobic fibres and

Interfacial Transition Zones (ITZ)
cement matrix facilitate the escape of the
theory [44]
steam.

The PITS between hydrophobic fibres and

Pressure-induced tangential space
cement matrix facilitates the escape of the
(PITS) theory [38]
steam.

The micro-cracks induced by melting of fibres

Micro-cracks theory [68]
facilitate the escape of the steam.

769

770 Table 2 Temperature information at spalling of unrestrained and unloaded specimens

Gas Surface Core Spalling

Source
temperature temperature temperature temperature*

Hertz [35] 300, 400 °C - - <400 °C

Noumowe et al.
- - - 250 - 300 °C
[47]
Phan et al. [23] 300, 450 °C - 240 - 280 °C 240 - 450 °C
Kalifa et al. [43] 190 - 250 °C
Fares et al. [69] 315 °C - - <315 °C
Kanéma et al. [1] - 355 °C - <355 °C
Debicki et al. 150, 223,
- 245, 291, 319 °C 150 - 319 °C
[70] 262 °C
Klingsch [22] - 297 - 433 °C 256 - 304 °C 256 - 433 °C
Akturk et al. [71] 425, 455 °C - - <455 °C
Liu and Tan [72] 400 °C - - <400 °C
771 * Spalling temperature means the temperature of concrete at the spalled location at
772 the time of spalling.

773
36
774 Table 3 Solutions for preventing different types of thermal spalling*

Types of thermal
Methods Protective measures Deficiency
spalling

Add PP fibres, PVA

Compromise in
fibres, jute fibres, raw
durability; disposable
rice husk, air-entraining
Increase and poor resilience.
agent, etc.
permeability
Compromise in
Thermo-hygral Use no or less silica
durability; limited
spalling fume.
application.
Reduce heating Costly; may lead to
Adopt thermal barrier.
rate spalling in deep section.
Selection of No flints or other
aggregate similar aggregates.
Use calcareous
Selection of
aggregate instead of Restrictive.
aggregate
siliceous aggregate.
Decrease initial
Thermo-mechanical Decrease load level. Costly.
applied stress
spalling
Lower the
Adopt thermal barrier. Costly.
temperature
Overcome
Add steel fibres. Costly.
brittleness
Selection of Avoid use of calcareous
Restrictive.
aggregate aggregate.
Limit the
Thermo-chemical Avoid excessive cement
amount
spalling amount.
of cement
Lower the
Adopt thermal barrier. Costly.
temperature
775 * Only passive protective measures are given, active protective measures such as
776 fire suppression system are not covered in this work.

777

778

779

37
780 Table 4 Temperature information of concrete at thermo-chemical spalling

Type of thermo-chemical Spalling

Source Gas temperature
spalling temperature*
Xiao et al. Sloughing-off spalling
800 °C 800 °C
[73]
Sloughing-off spalling
Arioz [37] 1200 °C 1000-1200 °C

Xing et al. Post-cooling spalling

750 °C 750 °C
[63]
Post-cooling spalling 2 hours of ISO 834
Klingsch [20] <1049 °C
fire
781 * Spalling temperature means the maximum recorded temperature at the spalled
782 location of concrete.

783

784 Figures

785

786 Fig. 1. Schematic representation of volumetric proportions of concrete components

787 before and after hydration

788

38
 Temperature
profile

Moisture
content profile
Concrete
wall
exposed
to fire on
Pore pressure one face
profile

Saturated zone

Wet Zone
Wet Zone
Dry zone

Spalling
region

789
790 Fig. 2 Schematic representation of thermo-hygral spalling of a concrete wall
791 exposed to fire on one face

792

793

794 Fig. 3. Pore pressure-induced tensile stress profile at the time of spalling

795

39
796

797 Fig. 4. Thermo-hygral spalling temperature range

798

Load stress

Thermal stress
Cover
Rebar

Hoop

c,a Confined
A action
Core
t,ni
c,c
Stress state of
point A

Spalling
region
Cover-core
799 interface

800 Fig. 5. Schematic representation of thermo-mechanical spalling of concrete

801 members

802

40
 803

804 Fig. 6. Thermo-mechanical spalling temperature range

805

heating
Decomposition of CSH phases Thermo-chemical spalling

20℃ 100℃ 200℃ 300℃ 400℃ 500℃ 600℃ 700℃ 800℃ 900℃

Drying Dissociation of calcium hydroxide Decarbonation

806
807 Fig. 7. Thermo-chemical spalling temperature range

808

41

View publication stats