Construction and Building Materials 304 (2021) 124608

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Ultra-High performance concrete (UHPC) with polypropylene (Pp) and

steel Fibres: Investigation on the high temperature behaviour
Francesca Sciarretta a, Stefano Fava b, Marco Francini b, Luca Ponticelli c, Mauro Caciolai c,
Bruno Briseghella d, Camillo Nuti e, *
a
Dipartimento di Culture del Progetto, Universita’ IUAV di Venezia, Venice, Italy
b
Buzzi Unicem S. p. A., via Luigi Buzzi, 6 - 15033 Casale Monferrato (AL), Italy
c
Corpo Nazionale dei Vigili del Fuoco, Rome, Italy
d
College of Civil Engineering, Fuzhou University, Fujian, China
e
Department of Architecture, University Roma Tre, Rome, Italy

A R T I C L E I N F O A B S T R A C T

Keywords: Ultra-high performance concretes (UHPC) are advanced cement-based materials characterised by superior me­
Ultra-High-Performance Concrete chanical properties with respect to normal and high-strength concretes; however, their dense and compact matrix
Polypropylene fibres can facilitate the onset of spalling at high temperatures. This problem is often coped up by adding polypropylene
Fire behaviour
(PP) fibres to the mix design, alone or with other types of fibres; steel fibres enhance the material’s tensile ca­
Experimental tests
Spalling
pacity. The paper presents a series of tests on two UHPC types (150 and 180 N/mm2) with PP fibres (0.27% of
volume) and variable content of steel fibres (0% to 1.92%), aimed at investigating the residual mechanical
properties of the material after high temperature exposure. The experimental results are compared to available
research on small UHPC specimens exposed to high temperatures, with dosages in PP fibres from 0.03% to 2%,
and in steel fibres from 0 to 3%. The results of this research demonstrate that UHPCs need hybrid fibre rein­
forcement (PP + steel) to withstand high temperatures, and that the residual strength increases after 200 ◦ C
exposure, at all steel fibre dosages; this is in line with literature. Available research also shows that strength loss
is possible in hot conditions, as found in the present research, while PP fibres alone do not always prevent the
occurrence of spalling in small UHPC samples.

1. Introduction used for joints between precast bridge decks, shear connector pockets,
and retrofitting of damaged elements [5–7]. Other prominent applica­
In the last 25 years, ultra-high-performance concrete (UHPC) has tions are high-rise building columns, precast bridge girders and foot­
been gaining attention in the field of construction industry and research, bridges [8–11].
due to its very high mechanical strength, high energy absorption ca­ There is a need of experimental research on UHPC mechanical
pacity in tension and durability [1,2]. To define this class of advanced properties under fire exposure (or ‘hot’ properties), as well as after
cement-based materials, the minimum strength of 120 N/mm2, i. e. the heating and subsequent cooling (or ‘residual’). This is important for
upper limit of high-strength concrete (HSC), is generally accepted [3]. performance-based fire design for any construction material [12], but it
The superior properties of UHPC are due to its peculiar mix design, is a quite an open issue in UHPC. As for the testing methods, no specific
which includes only fine aggregate (maximum grain diameter < 1 mm), standards for mechanical tests on UHPC under/after high temperatures
and low water to cement ratio (generally not higher than 0,25). Ac­ exist yet [13]. The compressive strength of UHPCs exposed to high
cording to the study by Stengel and Schiessl [4], the mean quantities of temperatures has been investigated since the 2000s especially in the
the different components in UHPC mixes are outlined in Table 1, where residual condition, with tests on small samples subjected to pre-
the volume fraction is the ratio of quantity to density. established target temperatures and slowly cooled down, to avoid
As far as structural applications are concerned, UHPCs are frequently thermal shock. The material strength of UHPC is inevitably investigated

* Corresponding author.
E-mail addresses: [email protected] (F. Sciarretta), [email protected] (S. Fava), [email protected] (M. Francini), [email protected]
(L. Ponticelli), [email protected] (M. Caciolai), [email protected] (B. Briseghella), [email protected] (C. Nuti).

https://doi.org/10.1016/j.conbuildmat.2021.124608
Received 9 February 2021; Received in revised form 14 August 2021; Accepted 15 August 2021
Available online 2 September 2021
0950-0618/© 2021 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
F. Sciarretta et al. Construction and Building Materials 304 (2021) 124608

Table 1
Mean quantities of the different components in UHPC mixes, according to [4].

fibres
Steel
Components Quantity Density Weight ratio Volume

100

150

150
50
(kg/m3) (kg/m3)

0
to to fraction
cement binder

fibres
Cement 752 3150 1.00 1.00 0.24

2.5

2.5

2.5
PP
Reactive powders 173 2200 0.23 0.08
(fly ash, silica
fume)

Water
Inert powders 169 variable 0.22 0.18

156

165

165
–
Silica sand and 887 1500 1.18 0.96 0.60
gravel
Superplasticiser 31 1100 0.04 0.03 0.03
(Polycarboxylate

Gravel Dolomia
ether)

Genova (4–8)
Steel fibres 242 7800 0.32 0.26 0.03
PP fibres 2 910 0.003 0.002 0.002
Water 184 1000 0.24 0.20 0.184

550

730

850
on small samples; the material properties must be considered along with
the structure’s size, loading and environmental conditions.
The paper presents experimental research on the hot and residual

Sand Dolomia
Genova (0–4)
compressive strength of two UHPC mixes with polypropylene (PP) and
steel fibres at 200 ◦ C. Tests at higher temperature will be carried out in
the future on the basis of the present experience. In fact, explosive

750

310

210
spalling occurred at 230–290 ◦ C, indicating the need of changes in
hybrid fibre composition with respect to what here investigated. The

aerator
experimental results here presented are discussed in comparison to
available literature data on high temperature heating of UHPC samples

De-

0.2

0.5

1.0
with reinforcement of PP and PP and steel (i. e. hybrid) fibres. In­
dications for the future developments are deduced.
The use of PP and steel fibres is frequent in UHPC mixes [14–18]. The

plasticiser
addition of PP fibres [19,20] is usual in concrete to withstand high

Super-
temperatures. PP fibres melt at 165 ◦ C, leaving a network of micro-

40

40

40
channels which increases the material permeability, relieving the pore
pressure and avoiding the development of high thermal gradients Metakaolin
[21,22]. As UHPC releases a high amount of energy at cracking, the
addition of steel fibres improves the tensile behaviour of the concrete 30

30

30
and can reduce the risk of explosive spalling [23].
Eurocode 2 [24] recommends a minimum PP fibre content of 2 kg/
Blast furnace

m3 (circa 0.2% of volume) for the highest grades of HSC; such quantity
can be effective for most types of PP fibres in UHPC [25]. According to
the American Concrete Institute, a 0.1% of volume in PP fibres is able to
slag
100

100

100

significantly reduce spalling [26]; for hydrocarbon fire exposure, the

Canadian standard [27] specifies a minimum of 0.3% in volume. On the
other hand, research suggests that the quantity of PP fibres required to
Silica
fume

avoid explosive spalling depends on the UHPC mix design [28]; values
80

80

80

from 0.03% to 2% of volume are reported (see Section 3 below).

The evaluation of UHPC compressive strength require tests on small
specimens; spalling investigations on UHPC can be performed by heat­
Concrete I

ing small specimens in a furnace. While the use of low heating rates can
52.5 R

be convenient for strength assessment, high heating rates are usual for
300

spalling investigations measuring the mass loss after the target tem­
–

–
Composition of the tested UHPC mixes (kg/m3).

perature is attained [21,28–37]. On the other hand, the spalling

behaviour of concretes can be investigated by blowtorch exposure of a
Content (kg/m3)

Compound 5941

specified area on a block or wall specimen [38]. As well, research reports

the occurrence of explosive spalling during heating cycles on UHPC
Nanodur

samples to be mechanically tested under/or after high temperatures

1000

[14,15,21,39]. Thus, spalling can impair the mechanical investigations

800

700

on UHPC.
In the present paper, the tested UHPC mixes behaved in a satisfactory
CLS-A(fc = 150 N/mm2)

Bt
B-

B-
Lt

manner at 200 ◦ C; to withstand higher temperatures, the content in PP

fibres needs to be increased, while maintaining the highest percentages
CLS-B(fc = 180

of steel fibres here adopted.

N/mm2)
Table 2

ID

2
F. Sciarretta et al. Construction and Building Materials 304 (2021) 124608

Fig. 1. Temperature recordings of performed thermal cycles: a) target 200 ◦ C; b) target 400 ◦ C.

Table 3
Test results.
Material age at PP fibres (% steel fibres cold tests hot tests – 200 ◦ C residual tests – 200 ◦ C
test (d) vol.) (% vol.) n. of fcm (N/ n. of fcm,θ (N/ %f n. of fcm,θ,res (N/ %fres
tests mm2)/CoV tests mm2)/CoV tests mm2)/CoV

1 2 3 4 5 6 7 8 9 10 11 12
CLS-A(150 N/ 28 0.27 0 3 141.9/0.04 3 135.2/0.10 − 5% 3 166.8/0.15 +18%
mm2) 0.62 3 144.3/0.02 3 135.5/0.17 − 6% 3 183.2/0.01 +27%
1.25 3 146.5/0.03 3 162.2/0.02 +11% 3 199.0/0.01 +36%
90 0.27 0 3 135.1/0.14 3 123.4/0.15 − 9% 3 183.0/0.05 +36%
0.62 3 149.2/0.04 3 152.4/0.03 +2% 3 186.1/0.05 +25%
1.25 3 167.4/0.02 3 166.9/0.04 − 2% 3 198.1/0.02 +18%
CLS-B-Bt(180 28 0.27 0 3 169.8/0.06 1 133.7 − 21% 3 190.0/0.01 +12%
N/mm2) 1.9 3 185.1/0.01 3 160.8/0.02 − 13% 3 199.9/0.03 +8%
CLS-B-Lt(180 N/ 90 0.27 0 3 171.6/0.08 0 – – 3 198.3/0.03 +16%
mm2) 1.9 3 177.0/0.02 3 178.1/0.03 +1% 3 224.6/0.03 +27%

1.1. Studies about the effects of PP, steel and hybrid fibre content matrix to the fibres before cracking, and thus reduces the confinement
effect. For the same reasons, the effectiveness of steel fibres in pre­
Some studies are available to understand the behaviour of UHPC venting explosive spalling is reduced in UHPCs [15,31,50,51]. Fibres
subjected to high temperatures, with the contribution of PP and steel have a bridge effect across cracks, but they may have insufficient stiff­
fibres [40–44]. ness to ensure the integrity of the UHPC specimen; moreover, their
The UHPC matrix has a very compact structure; high temperatures strength reduces with temperature [50–52]. Way and Wille [43] observe
bring on a further increase in compactness, through chemical reactions embrittlement and weakening of steel fibres after exposure to 500 ◦ C
(decomposition and hydration) [43]. Moreover, the very dense micro­ and beyond, and complete melting after 800 ◦ C.
structure of UHPC leads to an “internal autoclave” effect, i. e. hydration Hybrid reinforcement appears to be a need for UHPC exposed at high
is entrapped during the heating process, creating a high-temperature temperatures. PP and steel fibres together have shown effectiveness in
and high-pressure environment [15,45,46]. This leads to strength improving the residual strength of UHPC and reducing the explosive
reduction, in absence of the confinement exerted by steel fibre spalling hazard [53,30]. In [49], a UHPC with 2% in volume of PP fibres
[20,22,36]. – a very high fibre content, which may impair the material workability –
The available studies agree about the positive contribution of PP fi­ and 1% in volume of steel fibres retains the 70% of the compressive
bres in reducing explosive spalling phenomena, e. g. [18,29,47]; this is strength after exposure to 1000 ◦ C.
in line with experience on normal concrete and HSC. To be effective, The available tests in literature (reported in Tables A1 and A2 in the
fibres must have sufficient length, maintaining a minimal diameter not Appendix) do not allow to draw exhaustive conclusions about the effects
to affect too much the concrete volume [47]. The escape of steam can be of different dosages and geometry of PP fibres (Section 3); further
facilitated by the micro-channels and/or micro-cracks generated by the research seems useful in this field.
melting of PP fibres, around 165 ◦ C [32]. [38] and [48] enlighten the
important role of micro-cracks in connecting the channels network left 2. Experimental investigation
by melted out fibres, and the good performance of PP fibres 6 mm long,
with a diameter of with 0.018 mm (aspect ratio 330). At 360–400 ◦ C, 2.1. Materials and testing procedures
other phenomena can affect the pore pressure (the melted PP releases
various volatiles, e. g. pentane, propylene) [49], but such cases have not The UHCP mixes CLS-A and CLS-B (Table 2) were designed to attain
been experienced in the present research. compressive strength of 150 and 180 N/mm2 respectively, after 28 days.
The effect of steel fibres on the compressive strength of UHPCs at The materials were developed at the R&D Laboratory of Unical s.p.a.
normal temperature is not as substantial as for NSC and HSC [23]; in (Buzzi Unicem group) in Calenzano (Florence), Italy, where the UHPC
fact, the higher stiffness of UHPC impairs the transfer of energy from the specimens were made and left to harden before testing. Then, they were

3
F. Sciarretta et al. Construction and Building Materials 304 (2021) 124608

Fig. 2. Results of the single tests (3 samples per age per S) at 20 ◦ C, 200 ◦ C hot and 200 ◦ C residual of CLS-A (a, c, e) and of CLS-B (b, d, f).

4
F. Sciarretta et al. Construction and Building Materials 304 (2021) 124608

Fig. 3. Trends of temperature-dependent compressive strength of a) CLS-A in hot conditions, b) CLS-A in residual conditions, c) CLS-B in hot conditions, d) CLS-B in
residual conditions (d = days).

brought to the Laboratory of Strength of Materials of the National Fire Three samples could be tested for each target. Concrete cubes of
Brigade in Capannelle (Rome), where the heating cycles and mechanical 100x100x100 mm were used as samples, according to the capacity of the
tests in hot conditions were performed. testing equipment and the very high material strength.
A single percentage in volume of PP fibres was adopted for all mixes: To perform the hot tests, the samples were heated inside a muffle
0,27%. This dosage was determined according to available recommen­ oven (Linn High Therm, model KK-120) and immediately tested ac­
dations [24–26,54]; the PP fibres have diameter of 0.05 mm and length cording to EN 12390–3 protocol [56], with a RMU C190 hydraulic press.
of 12 mm (aspect ratio: 240). The length is higher than the average This procedure had been previously adopted at the Fire Brigade Labo­
values in literature, and the aspect ratio is at the lower bound of liter­ ratory; the oven and testing machine were moved close to each other
ature average (see Table A1 in the Appendix), to affect as less as possible and an insulating box was used to keep the sample at the attained
the workability of the fresh material. temperature, once out of oven, before and during the mechanical test.
CLS-A (150 N/mm2) has a single mix design and three different steel The box was made of four refractory bricks and rockwool layers, and it
fibre contents, i. e. 0%, 0.62% and 1.25%. CLS-B, which was designed to was pre-heated in another oven before use, to the same temperature of
achieve 180 N/mm2, has two different mixes. The mix indicated with B- the sample.
Bt (where ‘Bt’ is for short term) reaches the desired performance in 28 The cold and residual tests were performed at the Material and
days, while B-Lt (where ‘Lt’ is for long term) could reach the expected Structures Research Laboratory of the University of Roma Tre, with a
compressive strength only after 90 days. For both mixes, two steel fibre MetroCom Engineering hydraulic press.
dosages were adopted, i. e. 0% and 1.9% of volume. Attention was paid to the measurement of the temperature reached
For both concretes (A and B), the steel fibres were of two types, i. e. by the concrete; literature shows that for Ø100 mm cylindric samples,
straight fibres of diameter 0.4 mm and length 20 mm (aspect ratio: 50), a<2-hour hold does not allow reaching the target temperature at the
and hooked fibres of diameter 0.7 mm and length 30 mm (aspect ratio: core of the sample [14–16,31]. Thermocouples were inserted in some
43). samples to record the core temperature, through a hole of 4 mm diam­
The settings of the thermal cycles were established on the grounds of eter and 50 mm depth. The recordings verified that the samples reached
the provisions published by the Italian Council for Research [55]; target the target temperature (Fig. 1a).
temperatures were selected as multiples of 200 ◦ C, with a very low The research programme was interrupted after the cycles at target
heating rate (0.5 ◦ C/min). temperature 200 ◦ C, due to explosive spalling of all the samples

5
F. Sciarretta et al. Construction and Building Materials 304 (2021) 124608

Fig. 4. a) Trends of experimental hot compressive strength, compared to literature data (PP = 0.33%, different S dosages, [17]) b) zoom out.

Fig. 5. a) Trends of experimental residual compressive strength, compared to literature datum (PP = 0.25%, S = 0% [21]), b) zoom out.

(Fig. 1b). In detail, the 28-days and the 90-days old samples exploded and fcm,θ,res respectively for cold, hot and residual tests, with the coef­
respectively at temperatures in the range 230–260 ◦ C (CLS-A) and ficient of variation (CoV, evaluated on the three samples only) and the
around 290 ◦ C (CLS-B). per cent difference of the property in hot or residual conditions versus
the respective cold value (%f and %fres). The strength values of each
2.2. Results sample are depicted in histograms in Fig. 2.
Observing the original values (mean values in Table 3, column 6), the
The hot and residual compressive test results, for samples exposed to 28-days compressive strength of CLS-A shows no relevant variation with
200 ◦ C, are commented below in comparison to cold tests (at 20 ◦ C). the fibre content. On the other hand, at 90-days it increases with steel
Henceforth, the volume percentages in PP and steel fibres are abbrevi­ fibre content, i. e. from 135 to 149, and to 167 N/mm2, meaning + 3%
ated into PP and S respectively. and + 14% respectively for S = 0.62% and S = 1.25%. The single
Table 3 reports the values of mean compressive strength fcm, fcm,θ, specimen results in Fig. 2a show a good uniformity at 28 days for each

6
F. Sciarretta et al. Construction and Building Materials 304 (2021) 124608

Table 4
Mechanical tests, from research on UHPCs with PP fibres.
Reference ID fibres (%vol) sample age at hardening pre-drying heating hold
steel PP (cm) test means ◦
C duration ◦
C duration (◦ C/min) (h)
(d) (d) (d)

Chen, Yu & Tang UHPC-1 0.50 0.03 cylinder 56 water N/A 55 – – 2 0.5
2020 [14] ∅10x20
Yang et al. 2019 HF-CA 1.00 0.15 cube 10 56 water 20 56 105 to desired 2 2
[16] HF-noCA water
content
Liang et al. 2018 QS (quartz 01.002.00 02.00 cube 5 28 water 90 1 90 1 4 2
[15] sand)
SS (steel
slag)
Xiong & Liew Ducorit®- – 00.100.250.50 cylinder 28 air (RH 25–30 28 – – 530 4
2015 [21] D4 ∅10x20 85%)
Pimienta et al. BSI®-fire 2.50 0.33 cylinder N/A N/A N/A N/A N/ N/A 1 N/A
2012 (review ∅10x30 A
paper) [17] Ductal®-AF 1.87 0.46 cylinder N/A N/A N/A N/A N/ N/A 2 N/A
∅7x14 A
Italcementi – 0.57 cylinder N/A N/A N/A N/A N/ N/A 0.5 N/A
∅3.6x11 A
CERIB_I 1.00 0.33 cylinder N/A N/A N/A N/A N/ N/A 1 N/A
∅10x30 A
CERIB_II 1.71 0.33 cylinder N/A N/A N/A N/A N/ N/A 1 N/A
∅10x30 A
BCV® 2.00 – cylinder N/A N/A N/A N/A N/ N/A 3.3 N/A
∅4x6 A
Burke 2011 [18] Ductal®-AF 1.94 0.55 cylinder N/A steam 90 2 N/ N/A N/A 01236
∅7.5x15 A
Diederichs & FIR/LA/PF/ – 0.06 cylinder >180 steam 90 2 N/ N/A 3 1
Mertsch 2008 H ∅7x19.5 A
[41]
Felicetti et al. RPC 2.00 2.00 cubes N/A N/A N/A N/A N/ N/A 1 N/A
2000 [42] (mortar) A

steel dosage, and at 90 days except for S = 0%. It may be assumed that of a relevant content in steel fibres.
the longer curing time increases the bond at the matrix-fibres interfaces, The residual tests after exposure to 200 ◦ C (Table 3, columns 11 and
giving a positive contribution to strength. 12) show increased mean compressive strength, with respect to the
For CLS-B, both mixes have the same basic strength. They show a original values, for all the samples of both mixes. For CLS-A, the mean
moderate strength increase with S = 1.92% for the 28-day old samples strength values at each steel fibre dosage are almost equal for 28- and
(from 170 to 185 N/mm2, +9%), while for the 90-day old the increase is 90-days old samples. Fig. 2e indicates dispersion in the S = 0% tests at
very slight (from 172 to 177 N/mm2, +4%). In the latter case, the single 28 days, while the data are quite compact for the hybrid fibre rein­
specimens indicate some dispersion in the results for S = 0% (Fig. 2b); forcement and show a significant strength increase for both steel fibre
thus, the limited performance improvement may be disregarded. dosages (+27% and + 36% in the mean values for S = 0.62% and S =
As far as it concerns the 200 ◦ C hot tests (Table 3, columns 8 and 9), 1.25% respectively).
the steel fibre dosage affects the strength of CLS-A. At 28 days of The mean residual strength of CLS-B with S = 1.92% is moderately
maturation, the samples with S = 0% and S = 0.62% undergo a very higher than the original for the 28-days old samples (+8%) and signif­
slight reduction in strength (about − 5% of mean original strength), icantly higher (+27%) for the 90-days old samples. This mix with the
while the 1.25% content of steel fibres brings on a moderate increase hybrid reinforcement shows uniform results (Fig. 2f).
(+11%). For the former two S percentages, Fig. 2c shows some disper­ On the grounds of the above presented results, trends of the factor of
sion in the results of single samples, while data are quite compact in the original strength can be plotted in the range 0–200 ◦ C, for the tested
latter case. At 90 days, the mean strength reduces by − 9% for S = 0%, UHPC mixes (Fig. 3).
while for both dosages of steel fibres the strength is almost equal to the Comparing the results of the present research to literature, Figs. 4
original (+2% and − 2%). Fig. 2c points out a good uniformity in the and 5 show the plots of Fig. 3a-d, together with the data of CERIB and
results of S = 0.62% and S = 1.25%. Thus, at hot conditions (200 ◦ C), the BSI®-fire concretes [17], both having PP = 0.33% (for the hot condition)
effect of curing time on the strength variation of CLS-A is very limited. and Ducorit®D4 concrete [21], of PP = 0.25% (for the residual
The CLS-B material with PP fibres only (S = 0%) has exhibited a condition).
fragile behaviour both in cold and hot conditions. In particular, very CERIB UHPC-I and UHPC-II have respectively a compressive strength
violent failures have occurred to the loaded specimens of both CLS-B of 170 N/mm2 with S = 1%, and 200 N/mm2 with S = 1.7%. Both have
mixes in hot conditions, so that the machine plates were damaged. the same types of fibres, i. e. PP fibres 12 mm long, aspect ratio 667, and
The only sample successfully tested, at 28 days (Fig. 2d), resulted in a steel fibres 13 mm long, aspect ratio 81. In hot conditions, the latter has
21% reduction in strength. It was then decided not to complete the hot a higher strength gain than the former. BSI®-fire has a strength of
tests on the CLS-B samples without steel fibre reinforcement. Such 148–165 N/mm2 with S = 2.5%, with the same PP fibres as the previous
fragile behaviour was assumed to impair any applicability of the mate­ two, and steel fibres 20 mm long with aspect ratio 67. This material
rial without steel fibre addition. undergoes a slight reduction in strength in hot conditions. Finally,
The hybrid fibre reinforced samples of CLS-B, S = 1.92%, have Ducorit®D4 has a strength of 167 N/mm2 with S = 0%, and PP fibres 13
encountered more ductile failure modes; the mean strength decreases by mm long with aspect ratio 433; it shows a drop (-50%) in the residual
− 13% for the 28-day old samples and undergoes no relevant variation strength beyond 200 ◦ C exposure. The graphs’ captions report the ma­
for the 90-day old. Again, Fig. 2d enlightens uniform results in presence terials’ IDs (see Tables A1 and A2 for other details) and the fibre content

7
F. Sciarretta et al. Construction and Building Materials 304 (2021) 124608

Fig. 6. Linear regression of single fc,θ/fc-θ series and polynomial regression of all the data set, at variable PP, a) 0%, b) 0.03–0.17%, c) 0.20–0.33%, d) 0.44–0.66%,
and e) 2.00%

8
F. Sciarretta et al. Construction and Building Materials 304 (2021) 124608

Table 5
θ0.5fc values according to single linear and collective polynomial regressions.
PP content range ID PP (%) S (%) single linear regressions collective polynomial regression
n. data R2 θ0.5fc (◦ C) mean CoV R2 θ0.5fc (◦ C)

1 2 3 4 5 6 7 8 9 10 11
NULL QS 0 0 2 1.00 330 418 0.34 0.61 385
QS 0 2.00 2 1.00 310
SS 0 0 2 1.00 350
Ducorit®D4, 5 ◦ C/min 0 0 4 0.88 445
BCV® 0 2.00 4 0.65 655
LOW UHPC-1 0.03 0.50 8 0.95 530 632 0.25 0.65 630
FIR/LA/PF/H 0.06 0 4 0.62 470
Ducorit®D4, 5 ◦ C/min 0.10 0 4 0.99 600
Ducorit®D4, 30 ◦ C/min 0.10 0 4 1.00 560
HF-CA 0.15 1.00 5 0.77 725
HF-noCA 0.15 1.00 5 0.52 905
MEDIUM Ducorit®D4, 5 ◦ C/min 0.25 0 4 0.96 570 687 0.18 0.58 675
Ducorit®D4, 30 ◦ C/min 0.25 0 4 0.98 570
BSI®-fire 0.33 2.50 5 0,77 695
CERIB-I 0.33 1.00 4 0,50 725
CERIB-II 0.33 1.70 4 0,32 875
HIGH Ductal®-AF 0.46 1.90 2 1.00 900 744 0.24 0.58 675
Ducorit®D4, 5 ◦ C/min 0.5 0 5 0.96 550
Ducorit®D4, 30 ◦ C/min 0.5 0 3 1.00 540
Ductal®-AF, no hold 0.55 1.90 5 0.56 –
Ductal®-AF, 1 h hold 0.55 1.90 3 0.99 920
Ductal®-AF, 2 h hold 0.55 1.90 4 0.96 750
Ductal®-AF, 3 h hold 0.55 1.90 4 0.97 995
Ductal®-AF, 6 h hold 0.55 1.90 3 0.71 725
Italcementi 0.57 0 4 0.82 570
MAX QS 2.00 0 5 0.77 810 765 0.10 0.65 918
QS 2.00 1.00 5 0.74 805
SS 2.00 1.00 5 0.57 >1000
RPC 2.00 2.00 2 1.00 680

(% in volume). samples are taken into consideration. Such processes refer to tests of hot
In hot conditions at 200 ◦ C (Fig. 4a), the behaviour of CLS-A is in­ and residual compressive strength tests as well as of spalling.
termediate between the two CERIB concretes. As above noted, CLS-A
shows a moderate strength increase for the highest dosage in steel fi­ 3.1. Effect of fibres on UHPC residual compressive strength
bres (S = 1.25%) at 28 days of age; this is in line with the behaviour of
CERIB UHPC II, which displays a strength increase of the same entity at A collection of literature data allows making considerations about
150 ◦ C. the conditions of effectiveness of PP fibres (with and without steel fibres)
The decreasing strength of CLS-B at 28 days of age is in line with the in compressive tests on UHPC exposed to high temperatures. Table 4 lists
behaviour of CERIB UHPC I (S = 1%) and BSI®-fire (S = 2.5%); as noted the fibre content, sample types, hardening, pre-drying and testing con­
above, the different CLS-B mix with longer curing time and steel rein­ ditions (i. e. heating rate and hold of the target temperature); N/A de­
forcement S = 1.92% withstands the heating condition without losing notes unavailable data. The heating rates are generally low, to minimise
strength, which is an intermediate behaviour between CERIB UHPC-I the risk of spalling (except one study which compares the effects of slow
and -II. and fast heating [21]). The samples are generally cooled down in air, by
In residual conditions (Fig. 5a), CLS-A exhibits a stronger increase in turning off the heat and leaving the samples inside the oven; this is due
strength than the literature case Ducorit®D4 after 200 ◦ C exposure; the to the need of excluding thermal shock effects. For all the references
effect of the steel fibre reinforcement (which is absent in the literature listed in Table 4, the Appendix provides the concrete mixes (Table A1)
case) is evident. and fibres’ geometry (Table A2).
As well (Fig. 5a), the performance of the CLS-B mix at 28 days of age On the grounds of the collected references, graphs of the
is very similar to Ducorit®D4, regardless of the steel reinforcement. The temperature-dependent compressive strength fc,θ (Fig. 6) allow to indi­
other mix, which attains the desired strength at 90 days, shows larger viduate possible relationships between residual strength and PP fibre
strength increases than the literature case, especially with the presence content, and to evaluate the dispersion of results. Based on the PP vol­
of steel fibres. ume percentages listed in Table 4, the data are subdivided into five in­
Generally, all the cases tested in the present research lie within the tervals, i. e. null (PP = 0%), low (0.03–0.17%), medium (0.20–0.33%),
envelope of the literature cases. Beyond 200 ◦ C, in hot conditions, the high (0.44–0.66%) and maximum (2%), whose data are plotted in the
presented reference cases show a stable behaviour up to 600 ◦ C. In re­ graphs of Fig. 6. Regression lines are determined for the five groups of
sidual conditions, the cited case undergoes mechanical decay (Fig. 4b fc,θ/fc - θ data, following the least square method for each single series;
and 5b). the amount of data for each series and the determination coefficient R2
for each respective linear regression are listed in Table 5, columns 5 and
3. Discussion 6. Then, all the points for each group (null, low, medium, high and max)
are collected together; second order polynomial expressions, also shown
Literature information and the data of the present research are dis­ in Fig. 6, are determined for the collective series (Equations (1) to (5)).
cussed here below, to provide understanding of the effect of PP fibre The R2 coefficients are indicated in each subfigure of Fig. 6 and listed in
percentages (alone and combined with steel fibres into hybrid rein­ Table 5, column 10.
forcement) on the behaviour and properties of UHPCs exposed to high
temperatures. In detail, processes of furnace heating on small UHPC fcθ /fc = 7E − 06θ2 − 0.009θ + 2.9438for PP = 0%vol. (1)

9
F. Sciarretta et al. Construction and Building Materials 304 (2021) 124608

Fig. 7. Mean values and intervals of confidence of 65% and 95% for the considered data sets, a) null PP fibre content, b) low, c) medium, d) high and e) maximum.

Fig. 8. Mean values and intervals of confidence of 65% and 95% for the considered data sets.

of the present research (240, see Section 2 above); five values are higher
fcθ /fc = − 9E − 07θ2 − 0.0005θ + 1.148 forPP = 0.03 − 0.17% vol. (2)
(667 to 1000) [16,17,30], while in 9 out of 23 cases the aspect ratio of
PP fibres is not declared. Concerning the steel fibres, the aspect ratio is
fcθ /fc = − 3E − 06θ2 + 0.0015θ + 0.7954 forPP = 0.20 − 0.33% vol. (3)
between 65 and 83; only two are lower (40, 47) and one is higher (1 0 0).
In the present research, the aspect ratios of steel fibres (43 and 50) are
fcθ /fc = − 3E − 06θ2 + 0.0015θ + 0.8197 for PP = 0.44 − 0.66% vol. (4)
thus lower than the majority of references.
fcθ /fc = − 1E − 06θ2 + 0.0005θ + 1.1261 forPP = 2% vol. (5) The temperature after which the residual strength is reduced to the
50% of original value (θ0.5fc) is selected to evaluate the influence of PP
In the considered literature, the aspect ratio of PP fibres (reported in parameter. The values of θ0.5fc are determined by the intersection of the
Table A2 in the Appendix) is mainly low (167 to 433), as it is in the case regression lines with fc,θ/fc = 0.5, as shown in the graphs of Fig. 4. The

10
F. Sciarretta et al. Construction and Building Materials 304 (2021) 124608

linear and nonlinear trends have increasing abscissae of intersection

hold (θ0.5fc) at increasing PP range.
(h) Table 5 lists the values of θ0.5fc determined in two different ways, i. e.

0.5

3
1

4
–

–
1) by taking the mean of all the θ0.5fc values from single linear re­
gressions intersecting fc,θ/fc = 0.5, and 2) considering the value θ0.5fc
determined by the intersection of each collective polynomial regression

300 to
target

1000

1300

1000
(◦ C)

600

945

842

945

700
800

900

400
line with fc,θ/fc = 0.5. The values of θ0.5fc provided by the collective
polynomial regressions are mostly on the safe side with respect to the
mean value given the single linear regressions, except for the maximum
~ 60(ISO 834)

~ 60(ISO 834)

~ 60(ISO 834)

~ 60(ISO 834)

~ 60(ISO 834)

~ 60(ISO 834)
heating (◦ C/

~ 500(HCM
range of PP fibre content. As from Fig. 6C and 6D, the trends for the
0.2 to 2.45 medium and high range are almost equal. The corresponding θ0.5fc are
min)

180)
N/A
also equal, i. e. 675 ◦ C as per Table 5 column 11. This could indicate that
510

5
the medium range (PP = 0.2 to 0.33%) could be sufficient for substantial
improvement in θ0.5fc, but the large dispersion of data should induce
days

20 (50% RH)
≥90

N/A

N/A

150
pre-heating
pre-drying

20 (sealed)
some caution. Further performance increase seems obtainable with PP
1
–

= 2%, which might compromise the workability of the fresh concrete.

105

105

Then, the dispersion of the raised data is analysed by determining the

N/

N/

20
C

A
–

–
◦

intervals of confidence of the 65% and 95% of the values of residual

20–30

compressive strength (Fig. 7). The mean values and intervals of confi­
days

N/A

N/A

N/A
27

28

2–
dence are calculated for each PP range, at the temperatures for which
2

1
there are at least three measurements of the residual strength (Table 5,
25–30

column 5).
N/A

N/A

N/A

N/A

N/A

N/A

80–
90

90

90
C

It is worth noting that, at any PP range, the values of residual

◦

strength have very narrow intervals of confidence at the lowest tem­

peratures (150–200 ◦ C). The graph of examples with null PP (Fig. 7A)
lime-saturated

lime-saturated
airair + water

air (RH 85%)

shows that a considerable strength increment (a little<40% of the

original strength) after exposure at 200 ◦ C could be assumed. However,
curing
means

steam

steam

steam

water
water

water

N/A

N/A

although data in between are scarce, at 400 ◦ C the 65% interval lies well
air–

under the 50% of original strength, being affected by the number of

spalling cases before such temperature (zero values).
>90

≥90
N/A

N/A

N/A

N/A
300

150
age

The low range of PP (0.03–0.17%, Fig. 7B) displays a quite uniform

(d)

28

28

28

confidence between 200 and 800 ◦ C, showing decrease in residual

compressive strength at increasing temperature beyond 200 ◦ C.
cylinder ∅11x22cube

It must be noticed that, while in the medium and high ranges

cylinder ∅7.5x15

cylinder ∅7.5x15

cylinder ∅10x20
cylinder ∅10x20

cylinder ∅10x20

cylinder ∅10x20

prism 20x20x60
cylinder ∅5x10

cylinder ∅5x10

(0.20–0.33% and 0.44–0.66%) the regression lines are quite close to

sample (cm)

each other (Fig. 6C-D), their respective confidence intervals are quite
different (Fig. 7C-D). In fact, the dispersion at 450 ◦ C and 600 ◦ C is much
N/A

larger for the medium than for the high range of PP, underlying the need
10

for further experimental research. Finally, the available data on UHPC

0.0750.150.25

with PP = 2.00% (Fig. 7E) show a peculiar increase in residual

compressive strength after exposure to 400 ◦ C, with very narrow con­
fidence interval.
nylon

The numeric values defining all the lines plotted in Fig. 7 (i. e. mean
–

values, as well as the minimum and maximum values defining the in­
0.0750.150.25

tervals of 95% and 65% confidence) are listed in Table A3 in the Ap­
0.220.440.66

0.080.160.25

pendix. 2D and 3D graphs are also given in the Appendix (Fig. A1).
0.10.2505

0.330.66

Finally, to get a better focus on the effect of hybrid fibre reinforce­

0.30.5

0.33
0.22

0.33

0,22
fibres (%vol)

ment, the temperatures after which the material retains only the 50% of
PP

12

original strength (θ0.5fc) are put into relationship with the PP fibre
2.53.5

content, at variable steel fibre content (Fig. 8). Three ranges are selected,
steel
Spalling tests, from research on UHPCs with PP fibres.

2.58
123

1.5

2.5

i. e. S = 0%, 0.5–1% and 1.5–2.5%. Table A4 in the Appendix lists the

2
–

values plotted with the circle markers in Fig. 8, where second order
polynomial regressions are plotted and chosen as a basis for the confi­
UHPCPP-1
UHPCPP-2

Ducorit®-

200 MPa
150 MPa
UHPC/O

UHPC F

dence intervals. The regressions are expressed by Equations (6) to (8),

BSI®
B5Q

where vPP is the percent value of PP. The determination coefficients R2

D4
ID

IC
S

–
–

are also indicated.

Heinz, Dehn & Urbonas 2004 [30]

[
Hosser, Kampmeier & Hollmann

θ0.5fc = − 65.734(vPP )2 + 311.08vPP + 445.41 R2

Zhang, Dasari & Tan 2018 [34]

]
(6)
Mindeguia et al. 2007 [37]

= 0.69 for steel fibres = 0% volume

Li, Tan & Yang 2019 [33]

Xiong & Liew 2015 [21]

Liu & Zhang 2020 [32]

Ozawa et al. 2019 [31]

Choe at al. 2015 [35]

[ ]
θ0.5fc = − 292.4(vPP )2 + 675.98vPP + 621.35 R2 = 0.27 for steel fibres
Ye at al. 2012 [29]
Sarwar 2017 [28]

= 0.5 − 1% volume (7)

2012 [36]
Reference
Table 6

11
F. Sciarretta et al. Construction and Building Materials 304 (2021) 124608

Fig. 9. Spalling occurrence at increasing PP fibre percent in volume, per temperature ranges a) 200–300 ◦ C, b) 300–400 ◦ C, c) 400–500 ◦ C, d) 500–600 ◦ C,
e) 600–700 ◦ C.

12
F. Sciarretta et al. Construction and Building Materials 304 (2021) 124608

Table 7 Fig. 9 shows the occurrence of spalling within five temperature

Spalling cases, from research on UHPCs with PP fibres. ranges, for the cases of Table 6. The graphs exclude the cases which do
Temperature Spalling cases No spalling cases not mention the spalling temperatures; no spalling is reported beyond
range (◦ C) 650 ◦ C. The steel fibre dosage is subdivided into three ranges; the
200–300 [28] medium PP + high [14–16,21–23,28–31,33,34,36,37] different markers denote the heating rate. For every temperature range,
steel fibre content + the related references, fibre content and heating rates of spalling cases
low heating ratepresent are reported in Table 7.
research: medium PP It can be noticed that spalling can occur in the 200–300 ◦ C and
+ any steel fibre
content + low heating
300–400 ◦ C ranges even at low heating rates. Low and very high con­
rate tents in PP fibres, as well as steel fibres, can be involved. Within the high
300–400 [15,36] no fibres + low [14–16,18,21–23,28,31,33,34,36,37] PP range (0.44–0.66%), there is only one spalling case, however the
heating rate[15] total cases are significantly less than at PP < 0.4%.
maximum PP + high
steel fibre content +
low heating rate 3.2.1. PP fibre content.
400–500 [15] maximum PP + [14–16,18,21–23,28,31,33,34,36,37] Fig. 10 shows the spalling/no spalling occurrence for all the cases in
null steel fibre content the considered references, subdivided into hybrid reinforcement (A),
+ low heating rate[15] and only PP fibres (B), per each of the PP ranges considered above. In
maximum PP + high
steel fibre content +
detail, for hybrid reinforcement with PP = 0.44–0.66% (Fig. 10A), the
low heating rate[16] data are quite uniform in pointing out an increased effectiveness of the
low PP + medium steel hybrid reinforcement (1 spalling case of 8 total); while for the range of
fibre content + 0.20–0.33% of volume, the good numerosity of data indicates increased
medium heating rate
effectiveness, but with higher scatter (9 spalling cases in 24 total). On
500–600 [14] low PP + medium [15,16,18,21,28,31,33,34,36,37]
steel fibre content + the other hand, for the PP fibres alone, it is difficult to establish an op­
low heating rate[16] timum range of fibre volume percentage, since very few data refer to PP
low PP + medium steel > 0.33% (Fig. 10B). In both cases, increasing PP to 2% does not seem to
fibre content + produce commensurate benefits; however, information on PP = 2% is
medium heating rate
from one research only.
600–700 [31] zero PP + high [15,16,18,21,28,33,34,36,37]
steel fibre content + A number of studies ascribe the spalling cases to the effect of vapour
high heating rate[31] pressure inside the pores of the UHPC matrix [14–16,28,29,31], up to
low PP + high steel PP = 0.66%. On the other hand, for PP = 2%, explosive spalling
fibre content + high
occurred without vapour emission and was rather due to thermal tension
heating rate[31]
medium PP + high and strain incompatibility between the cement paste and aggregates
steel fibre content + [15]. It is thus likely that, for very high percentages of PP fibres, a switch
high heating rate[31] occurs from pressure- to stress-driven spalling mechanism, but in-depth
high PP + high steel investigation should confirm this induction.
fibre content + high
Comparing Fig. 10A to 10B, the percentage of spalling cases (histo­
heating rate
grams on the right) decreases at increasing PP (up to 0.66%) if in
combination with steel fibres. At the opposite, when PP fibres are used
[ ]
θ0.5fc = − 410.18(vPP )2 + 910.3vPP + 497.54 R2 = 0.59 for steel fibres alone, their benefit – as far as the available data can show – is much
reduced beyond the 0.03–0.17% volume range. This might be due to the
= 1.5 − 2.5% volume
possible aggravation of pore pressure due to polypropylene decompo­
(8) sition into volatile compounds [49]. Thus, the steel fibres, by playing a
The highest range of S, i. e. 1.5–2.5%, brings on a much higher in­ complementary role – i. e. bridging the incipient cracks and increasing
crease in θ0.5fc at growing PP fibre content, as it is clear from the three the material’s ductility – can act as a guarantee for the beneficial effect
graphs in Fig. 6. The highest content of PP fibres (vPP = 2.00), brings the of PP fibres on spalling probability, up to medium and high PP fibre
highest benefit to UHPC reinforced with PP only, by taking θ0.5fc to content. Li, Tan and Yang have enlightened this synergetic effect of the
about 800 ◦ C (Fig. 6A); vice versa, vPP = 2.00 appears to give no increase fibres [33].
when the reinforcement is hybrid (Fig. 6B-C). However, these observa­ The spalling cases observed in the present research (PP only and
tions are based only on a few data related to vPP = 2.00, and the lack of hybrid fibre reinforcement, with 0.27% in volume of PP fibres) fit in the
data for 0.66 < vPP < 2.00 impairs any conclusion. Data are scattered, not negligible spalling occurrence (i. e. 38%) at PP = 0.20–0.33% re­
with high R2 coefficients, therefore further experimental information is ported in the graphs of Fig. 10A-B.
needed.
3.2.2. Curing, pre-drying conditions and heating rates.
Some information is available about the effect of different curing
3.2. Effect of fibre content and other parameters on the spalling of UHPC conditions for UHPCs subjected to high temperatures. Table 8 resumes
samples the characteristics of the research carried on by Liu and Zhang [32] and
Xiong and Liew [21]. The two studies took into account only PP fibre
A collection of studies on UHPC spalling, carried on with furnace reinforcement, samples of the same type and size, same heating rates
heating on small samples, is listed in Table 6. Such information is treated and no pre-drying; both of them accounted for wet and dry curing
jointly with the previously cited studies on temperature-dependent conditions.
compressive strength (Table 4), to analyse the effects of furnace heat­ Each of the two investigations showed uniform results – all the
ing on small samples in terms of spalling occurrence. As a difference samples underwent spalling in the former study and no spalling in the
from the heating processes related to mechanical tests, in most cases the latter – across the selected parameters, demonstrating that the curing
spalling tests entail much higher heating rates. The target temperatures conditions had negligible effect. While the higher original strength of
are very high and the durations at maximum temperature are mostly the concrete in Xiong and Liew’s research could have played some role,
short (Table 6). the relevant differences between the two series of tests could rather be in

13
F. Sciarretta et al. Construction and Building Materials 304 (2021) 124608

Fig. 10. Spalling occurrence per PP fibre content range, with (a) and without steel fibres (b).

Table 8
Comparison between different experimental results at partially similar conditions.
Ref. % % PP sample age at material mix proportions MPa curing pre- Heating result
steel (cm) testing drying rate target hold
(d) (◦ C/ (◦ C) (h)
min)

[32] 0 0.22 cylinder >90 UHPCPP- cement 1silica fume 100.1 air (water none 510 600 1 explosive
10x20 1 0.2silica sand 0.6 content spalling in
superplast. 0.02–0.03);air all cases
0.049fine aggreg. + water 3–4
0.6w/c: 0.288 weeks (water
UHPCPP- cement 1silica fume 116.7 content 0.05)
2 0,2silica powder 0.6
superplast. 0.049w/
c: 0.24
[21] 0 00.10.250.5 cylinder 28 Ducorit®- cement, Densit® 163 air 25–30 ◦ C, none 5 800 4 no spalling
10x20 D4 binder, 85% RH;steam in any case
superplasticiser, 28 ◦ C, 100%
fine bauxite RH;sealed
aggregatew/mix: specimens +
0.076 air 25–30 ◦ C,
85%RH

the materials’ mix design. It can be inferred that the high content in development of excessive thermal gradients. A high heating rate (30 ◦ C/
silica and ordinary cement could have favoured the onset of spalling min) could bring on a faster decrease of the residual strength of PP-fibre
[28], while the presence of bauxite aggregate could have significantly reinforced UHPC at increasing temperatures [21]. Though a relatively
improved the fire performance of the basic material [21]. high moisture content has a significant role in facilitating the onset of
Contrasting points of view are present in literature about the de­ explosive spalling [16], nonetheless spalling can be experienced at any
pendency of spalling on the heating rate and initial moisture content of heating rate and moisture content, e. g. [22,25]. It can be said that the
UHPC. Spalling tests mostly envisage very fast heating (Table 6); slow heating rate is of importance as far as it contributes to the pore pressure,
rates (<1 to 5 ◦ C/min) are usually applied in thermal cycles before which is the predominant trigger of spalling [28,42].
mechanical tests of residual properties (Table 4), just to avoid the Finally, pre-drying procedures are usually applied to minimise the

14
F. Sciarretta et al. Construction and Building Materials 304 (2021) 124608

Fig. 11. Total spalling cases vs pre-drying conditions of samples in the considered references.

effect of vapour pressure inside the concrete pores; this allows avoiding • Available research involving the furnace heating of small UHPC
the spalling of specimens. However, some research suggests that the pre- samples encompasses PP = 0–0.66% and 2.00%, and S = 0–3.00%.
drying in oven at 60 to 105 ◦ C could impact the material’s microstruc­ The results of the present research are confirmed by the available
ture, increasing the concrete’s porosity due to capillary stress, cement data about mechanical tests, whose dispersion is low at 200 ◦ C.
hydrates desiccation and potential micro-crack generation in relation­ • Available research shows a considerable dispersion of data about
ship with internal thermo-hydric stress [57]. In absence of specific residual compressive strength at PP = 0.20–0.33%, while there is a
studies, it is worth noting that, in the considered researches, spalling and clear trend for PP = 0.44–0.66%. This is a good indication for the
pre-drying at 90–105 ◦ C in air were often concurrent (23/27 cases), at developments of the present research; the increase in PP percentage
any volume percent of steel and/or PP fibres (Fig. 11). Air drying at to 0.44–0.66% - with adequate aspect ratio of fibres, i. e. 200–300 - is
ambient conditions seem to be more beneficial in reducing the likeliness foreseen.
of spalling (17/22 cases). • Available literature points at hybrid reinforcement with values of S
higher than 1%, like the upper values considered in the present
4. Conclusions research. This is as well an indication for future investigations.
• Beyond 200–400 ◦ C, as literature reports, the residual compressive
The paper has presented an experimental research aimed at assessing strength of UHPC decreases; the trend depends on the PP fibre dos­
the high temperature behaviour of three UHPC mixes – one (A) of 150 ages. This is especially clear by observing θ0.5fc (the temperature at
and two (B) of 180 N/mm2 basic strength – with polypropylene fibres which the strength is reduced to 50%). A very high PP fibre content
(PP, 0.27% in volume) and steel fibre (S, various percentages in volume (2%) shows significant benefit, but such a high dosage should ac­
from 0% to 1.92%). The compressive strength was measured under and count for possible issues of fresh material’s workability. This also
after high temperature exposure. The aspect ratio of the PP fibres (12 indicates an important field for future investigations.
mm long) is 240, while steel fibres (of two shapes, length 20 and 30 mm) • About the heating of UHCP samples with hybrid fibre contents,
have aspect ratios of 50 and 43. Results during (hot) and after (residual) literature reports the occurrence of spalling at very different tem­
exposure to a cycle to 200 ◦ C are given. peratures, from 200 to 300 ◦ C (like it happened in the present
research) on. For higher temperatures, up to 600–700 ◦ C, cases of
• At 200 ◦ C, in hot conditions, CLS-A shows limited strength varia­ disruptive spalling involved any range of steel and PP fibre per­
tions, in line with literature information. At 90 days, the material centage; high fibre percentages seem worth investigation.
strength slightly decreases in absence of steel fibres, while it remains
at the original level for both steel fibre dosages. At the same condi­
Declaration of Competing Interest
tion, the concrete age is not relevant for CLS-A, while the increasing
steel fibre dosage brings on a slight performance improvement.
The authors declare that they have no known competing financial
• At the same condition, CLS-B undergoes small strength reductions at
interests or personal relationships that could have appeared to influence
28 days, both for PP-only and hybrid reinforcements; the strength
the work reported in this paper.
remains at the original levels for the mix which attains the desired
strength at 90 days of age.
• After 200 ◦ C, the residual strength of both CLS-A and CLS-B is higher Acknowledgements
than the original values through all the steel fibre dosages and con­
crete ages. The contribution of steel fibre is relevant, in fact S = The authors warmly thank the Italian Fire Brigade, Buzzi Uni­
1.25% and S = 1.92% increases the strength of CLS-A and CLS-B up calcestruzzi, Federbeton (Italy) for financial support for “Seismic
to about 200 and 225 N/mm2 respectively. Behaviour of Integral Bridges” and DPC-ReLUIS consortium for the
• After 200 ◦ C, the effect of the steel fibre content on CLS-A is more financial support within the framework of the 2014-2018 and 2019-
intense than in hot conditions at the same temperature. 2021 Research Projects. The research was supported by the National
• The benefit of hybrid (PP + steel) fibre content is significant. In the Natural Science Foundation of China (Grant No. 51778148).
present research, a steel fibre content of 1.25% in volume for CLS-A
has increased both hot (+11%) and residual strength (+36%) at 28 Appendix A. Supplementary data
days of age. S = 1.92% for CLS-B-Lt (90 days) has increased the re­
sidual strength (+27%).

15
F. Sciarretta et al. Construction and Building Materials 304 (2021) 124608

Table A1
Tests on UHPCs with PP fibres: UHPC mix design.
Reference Tests Country ID p c sf sp qp fl s va st bl g mk ss ca fa w/c fc (N/
mm2)

present research M Italy CLS-A x x x x x x x 0.195* 150

CLS-B x x x x x x x 0.165* 200
Chen, Yu & Tang 2020 M Taiwan UHPC-1 x x x x x x 0.195 156
[14]
Liu & Zhang 2020 [32] S China UHPCPP-1 x x x x x 0.29 100
UHPCPP-2 x x x x x 0.29 117
Yang et al. 2019 [16] M, S China HF-CA x x x x x x 0.18 157
HF-noCA x x x x x 0.18 145
Li, Tan & Yang 2019 S Singapore – x x x x x 0.2* 150
[33]
Ozawa et al. 2019 [31] S Japan S x x x 0.14* 213
IC x x x 0.14* 213
Liang et al. 2018 [15] M Australia - quartz sand x x x x x 0.21 112
China steel slag x x x x x 0.21 90
Zhang, Dasari & Tan S Singapore UHPC/O x x x x x 0.2* 160
2018 [34]
Sarwar 2017 [28] S USA UHPC F x x x x x x x 0.14 165
Xiong & Liew 2015 M, S Singapore Ducorit®- x 0.076* 167
[21] D4
Choe at al. 2015 (PP + S USA 200 MPa x x x x x x x x 0.227 205
nylon fibres) [35] 150 MPa x x x x x x x x 0.23 149
Pimienta et al. 2012 M France BSI®-fire x 0.19* 148–165
(review paper) [17] France Ductal®-AF x 0.14* 160
Italy Italcementi x x 0.31* 121
France CERIB_I N/A 170
France CERIB_II N/A 200
France BCV® x 0.104* 155
Ye at al. 2012 [29] S China – x x x x 0.22 N/A
Hosser, Kampmeier & S Germany B5Q x x x x x x 0.2 N/A
Hollmann 2012 [36]
Burke 2011 [18] M USA Ductal®-AF x x x 0.186 142
Diederichs & Mertsch M Germany, FIR/LA/PF/ x x x x x 0.2 158
2008 [41] Finland H
Mindeguia et al. 2007 S France BSI® x x 0.2* N/A
[37]
Heinz, Dehn & S Germany – x x x x 0.18 180
Urbonas 2004 [30]
Felicetti et al. 2000 M Italy - UK RPC x x x 0.18 160
[42] (mortar)

M: mechanical tests, S: spalling tests; pr: premix, c: cement, sf: silica fume, sp: silica powder, qp: quartz powder, fl: fly ash, s: superplasticizer, va: viscous agent, st: steel
slag, bl: blast furnace slag, g: gypsum, mk: metakaolin, ss: silica sand, ca: coarse aggregates, fa: fine aggregates; w/c: water-concrete ratio, *) water/premix ratio.

16
F. Sciarretta et al. Construction and Building Materials 304 (2021) 124608

Table A2
Fibres’ geometry for the considered references.
study steel fibres PP fibres type of tests
% vol. ∅ (mm) L (mm) AR % vol. ∅ (mm) L (mm) AR

Present research 0.0 0.64 1.28 1.92 0.4 0.7 2030 5043 0.27 0.05 12 240 M
Chen. Yu and Tang (2020) [14] 0.5 0.2 13 65 0.03 0.05 12 240 M
Liu & Zhang (2020) [32] – – – – 0.22 0.06 10 167 S
Yang et al. (2019) [16] 1.0 0.2 16 80 0.15 0.02 19 950 M, S
Li. Tan & Yang (2019) [33] 0 1.0 2.0 3;0 0.22 13 59 0 0.22 0.44 0.66 0.03 12 400 S
Ozawa et al. (2019) [31] 2.0 0.2 15 75 0 0.3 0.5 0.042 12 286 S
Zhang, Dasari & Tan (2018) [34] – – – – 0 0.33 0.033 12 364 S
Liang et al. (2018) [15] 0.0 1.0 2.0 0.12 10 83 0.0 2.0 0.031 10 323 M. S
Sarwar (2017) [28] 1.5 N/A N/A N/A 0.22 N/A N/A N/A S
Xiong & Liew (2015) [21] – – – – 0 0.10 0.25 0.50 0.03 13 433 M. S
Choe et al. (2015) [35] – – – – 0.075 0.15 0.25 N/A N/A N/A S
Burke (2011) [18] 1.94 0.2 14 70 0.55 N/A N/A N/A M
Pimienta et al. (2012) - BSI®-fire [17] 2.5 0.3 20 67 0.33 0.018 12 667 M
Pimienta et al. (2012) - Ductal®-AF [17] 1.87 0.2 13 65 0.46 N/A N/A N/A M
Pimienta et al. (2012) – Politecn.-Italcementi [17] – – – – 0.57 0.02 20 1000 M
Pimienta et al. (2012) – CERIB UHPC_I [17] 1.0 0.16 13 81 0.33 0.018 12 667 M
Pimienta et al. (2012) – CERIB UHPC_II [17] 1.7 0.16 13 81 0.33 0.018 12 667 M
Pimienta et al. (2012) - BCV® [17] 2.0 0.175 12.7 73 – – – – M
Ye et al. (2012) [29] – – – – 0.1 0.2 N/A 19 N/A S
Hosser, Kampmeier & Hollmann (2012) [36] 2.6 0.19 9 47 0 0.082 0.0165 0.25 N/A N/A N/A S
Diederichs & Mertsch (2008) [41] – – – – 0.06 N/A N/A N/A M
Mindeguia et al. (2007) [37] 2.5 0.3 20 67 0.033 N/A N/A N/A S
Heinz. Dehn & Urbonas (2004) [30] 2.5 3.5 0.15 6 40 0 0.33 0.66 0.016 4.9 306 S
Felicetti et al. (2000) [42] 2.0 0.16 16 100 2.0 N/A N/A N/A M

∅ : diameter, L: length, AR: aspect ratio (L/Ø).

Table A3
Mean values and intervals of confidence for the collected data (graphs fc,θ/fc - θ).
PP fibre quantity(% vol.) ◦
C n. data mean fcθ (N/mm2) standard deviation confidence 95% confidence 65%
max min max min

0 – NULL 200 5 1,366 0,454 1,764 0,968 1,556 1,176

400 4 0,198 0,395 0,585 − 0,190 0,382 0,013
0.03–0.17 - LOW 200 4 1,123 0,106 1,226 1,019 1,172 1,073
400 5 1,000 0,268 1,235 0,765 1,112 0,888
600 5 0,476 0,326 0,762 0,190 0,612 0,340
800 4 0,273 0,094 0,365 0,180 0,316 0,229
0.20–0.33 - MEDIUM 150 3 0,957 0,129 1,103 0,811 1,026 0,887
450 3 1,070 0,471 1,603 0,537 1,324 0,816
600 5 0,656 0,321 0,937 0,375 0,790 0,522
0.44–0.66 - HIGH 200 5 1,066 0,069 1,126 1,006 1,095 1,037
300 4 1,065 0,192 1,253 0,877 1,155 0,975
400 6 0,932 0,216 1,104 0,759 1,014 0,849
500 5 0,878 0,079 0,947 0,809 0,911 0,845
600 5 0,599 0,199 0,773 0,424 0,682 0,516
2.00% - MAXIMUM 200 3 1,090 0,030 1,124 1,056 1,106 1,074
400 3 1,350 0,062 1,421 1,279 1,384 1,316
600 4 0,953 0,245 1,193 0,712 1,067 0,838
800 3 0,487 0,367 0,902 0,071 0,685 0,288
1000 3 0,413 0,241 0,686 0,141 0,543 0,283

17
F. Sciarretta et al. Construction and Building Materials 304 (2021) 124608

Table A4
Mean values and intervals of confidence for the collected data (graphs θ0.5fc - vPP).
steel % ID PP% steel % T50%fc mean standard deviation interval of confidence 95% interval of confidence 65%
max min max min

0% QS 0 0 330 530.4 125.3 601.3 459.5 564.2 496.6

SS 0 0 350
Ducorit®D4, 5 ◦ C/min 0 0 445
FIR/LA/PF/H 0.06 0 470
Ducorit®D4, 5 ◦ C/min 0.10 0 600
Ducorit®D4, 30 ◦ C/min 0.10 0 560
Ducorit®D4, 5 ◦ C/min 0.25 0 570
Ducorit®D4, 30 ◦ C/min 0.25 0 570
Ducorit®D4, 5 ◦ C/min 0.5 0 550
Ducorit®D4, 30 ◦ C/min 0.5 0 540
Italcementi 0.57 0 570
QS 2.00 0 810
0.5–1.5% UHPC-1 0.03 0.50 530 738 137.8 858.8 617.2 795.6 680.4
HFCA 0.15 1.00 725
HFnoCA 0.15 1.00 905
CERIB-I 0.33 1.00 725
QS 2.00 1.00 805
1.5–2.5% QS 0 2.00 310 750.5 193.8 870.6 630.3 807.8 693.2
BCV® 0 2.00 655
BSI®-fire 0.33 2.50 695
CERIB-II 0.33 1.70 875
Ductal®-AF 0.46 1.90 900
Ductal®-AF 0.55 1.90 920
Ductal®-AF 0.55 1.90 750
Ductal®-AF 0.55 1.90 995
Ductal®-AF 0.55 1.90 725
RPC 2.00 2.00 680

18
F. Sciarretta et al. Construction and Building Materials 304 (2021) 124608

Fig. A1. 2D and 3D graphs of the residual strength ratio vs temperature of exposure.

References with steel or polymer fibers, Appl. Sci. 8 (9) (2018) 1671, https://doi.org/
10.3390/app8091671.
[8] M. Empelmann, M. Teutsch, G. Steven, Expanding the application range of RC-
[1] B. Graybeal, Ultra High Performance concrete. Federal Highway Administration
columns by the use of UHPC, in: J.C. Walraven, D. Stoelhorst (Eds.), Tailor Made
Research and Technology, publication n◦ FHWA-HRT-11-038, 2011.
Concrete Structures, Taylor and Francis Group, London, 2008, pp. 461–468.
[2] Recommendations (2013).
[9] T. Stengel, P. Schiessl, Sustainable construction with UHPC – from life cycle
[3] V. Perry, What really is Ultra-High Performance Concrete? Towards a global
inventory data collection to environmental impact assessment, in: E. Fehling, M.
definition, in: C. Shi, B. Chen (Eds.), Proceedings of 2nd International Conference
Schmidt, S. Stürwald (Eds.), Proceedings of the 2nd International Symposium on
on UHPC Materials and Structures, Fuzhou, 2018.
Ultra High Performance Concrete, Kassel, 2008, pp. 461-468.
[4] T. Stengel, P. Schiessl, 22 – Life cycle assessment (LCA) of Ultra-High Performance
[10] M. Rebentrost, G. Wight, Experience and applications of ultra-high performance
Concrete (UHPC) structures, in: F. Pacheco-Torgal, L.F. Cabeza, J. Labrincha, A. de
concrete in Asia, in: E. Fehling, M. Schmidt, S. Stürwald (Eds.), Proceedings of the
Magalhães (Eds.), Eco-efficient Construction and Building Materials, Woodhead
2nd International Symposium on Ultra High Performance Concrete, Kassel, 2008,
Publishing Ltd., 2014, pp. 528–564, https://doi.org/10.1533/
pp. 461-468.
9780857097729.3.528.
[11] J. Xue, B. Briseghella, F. Huang, C. Nuti, H. Tabatabai, B. Chen, Review of ultra-
[5] H.G. Russell, B. Graybeal, Ultra-high performance concrete: a state-of-the-art-
high performance concrete and its application in bridge engineering, Construction
report for the bridge community, Federal Highway Administration Research and
and Building Materials 260 (2020) 119844, https://doi.org/10.1016/j.
Technology, publication n◦ FHWA-HRT-13-060 (2013).
conbuildmat.2020.119844.
[6] E. Denarié, E. Brühwiler, Structural rehabilitations with ultra-high performance
[12] F. Sciarretta, Modeling of mechanical damage in traditional brickwork walls after
fibre reinforced concretes (UHPFRC), Restor. Build. Monum. 12 (5/6) (2006)
fire exposure, Advanced Materials Research 919-921 (2014) 495-499. https://doi.
453–468.
org/10.4028/www.scientific.net/AMR.919-921.495.
[7] J. Xue, D. Lavorato, A.V. Bergami, C. Nuti, B. Briseghella, G.C. Marano, T. Ji,
[13] Fire resistance – Tests for thermophysical and mechanical properties of structural
I. Vanzi, A.M. Tarantino, S. Santini, Severely damaged reinforced concrete circular
meterials at elevated temperatures for fire engineering design. ISO/TR 15655,
columns repaired by turned steel rebar and high-performance concrete jacketing
International Standard Organisation, 2020.

19
F. Sciarretta et al. Construction and Building Materials 304 (2021) 124608

[14] D. Cree, P. Pliya, M.F. Green, A. Noumowé, Thermal behaviour of unstressed and [37] J. C. Mindeguia, P. Pimienta, A. Simon, N. Atif, Experimental and numerical study
stressed high strength concrete containing polypropylene fibers at elevated of an UHPFRC at very high temperature, in: F. Toutlemonde, K. Sakai, O. E. Gjorv,
temperature, JSFE 8 (4) (2017) 402–417. N. Banthia (Eds.), Proceedings of CONSEC’07, Concrete Under Severe Conditions:
[15] N. Yermak, P. Pliya, A.-L. Beaucour, A. Simon, A. Noumowé, Influence of steel and/ Environment and Loading, Tours, 2007, ID 175.
or polypropylene fibres on the behaviour of concrete at high temperature: Spalling, [38] L. Missemer, E. Ouedraogo, Y. Malecot, C. Clergue, D. Rogat, Fire spalling of ultra-
transfer and mechanical properties, Construction and Building Materials 132 high performance concrete: From a global analysis to microstructure
(2017) 240–250. investigations, Cement and Concrete Research 115 (2019) 207–219.
[16] M.-X. Xiong, J.Y. Richard Liew, Spalling behavior and residual resistance of fibre [39] Ch. Kahanji, F. Ali, A. Nadjai, in: Experimental study of ultra-high performance
reinforced Ultra-High performance concrete after exposure to high fibre reinforced concrete under ISO 834 fire, Princeton University, 2016,
temperaturesSpalling behavior and residual resistance of fibre reinforced Ultra- pp. 165–173.
High performance concrete after exposure to high temperatures, Mater. construcc. [40] C. Nuti, J. Xue, S. Fava, P. Castelli, L. Euzor, M. Francini and L. Ponticelli,
65 (320) (2015) e071, https://doi.org/10.3989/mc.2015.v65.i32010.3989/ Experimental behavior of UHPC under fire action, in: C. Shi, B. Chen (Eds.),
mc.2015.00715. Proceedings of 2nd International Conference on UHPC Materials and Structures,
[17] Y. Li, K.H. Tan, in: Effects of polypropylene and steel fibers on permeability of Fuzhou, 2018.
ultra-high performance concrete at hot state, Princeton University, 2016, [41] U. Diederichs, O. Mertsch, Behaviour of Ultra High Strength Concrete at high
pp. 145–152. temperatures, in: E. Fehling, M. Schmidt, S. Stürwald (Eds.), Proceedings of the 2nd
[18] I.L. Larsen, R.T. Thorstensen, The influence of steel fibres on compressive and International Symposium on Ultra High Performance Concrete, 2008, pp. 347–354.
tensile strength of ultra high performance concrete: A review, Construction and [42] R. Felicetti, P.G. Gambarova, M.P. Natali Sora, G.A. Khoury, Mechanical behaviour
Building Materials 256 (2020) 119459, https://doi.org/10.1016/j. of HPC and UHPC in direct tension at high temperature and after cooling, in: in:
conbuildmat.2020.119459. Proceedings of BEFIB – 5th Symposium on Fibre-Reinforced Concrete, 2000,
[19] H.-J. Chen, Y.-L. Yu, Ch.-W. Tang, Mechanical properties of Ultra-High pp. 749–758.
Performance Concrete before and after exposure to high temperatures. Mater. 13 [43] R.T. Way, K. Wille, Effect of heat-induced chemical degradation on the residual
(2020), 770. https://doi.org/10.3390/ma13030770. mechanical properties of ultrahigh performance fibre-reinforced concrete,
[20] X. Liang, C. Wu, Y.u. Su, Z. Chen, Z. Li, Development of ultra-high performance J. Mater. Civ. Eng. 28 (4) (2016) 04015164, https://doi.org/10.1061/(ASCE)
concrete with high fire resistance, Construction and Building Materials 179 (2018) MT.1943-5533.0001402.
400–412. [44] Vahid Afroughsabet, Luigi Biolzi, Paulo J.M. Monteiro, The effect of steel and
[21] J. Yang, G.-F. Peng, J. Zhao, G.-S. Shui, On the explosive spalling behavior of ultra- polypropylene fibers on the chloride diffusivity and drying shrinkage of high-
high performance concrete with and without coarse aggregate exposed to high strength concrete, Composites Part B: Engineering 139 (2018) 84–96.
temperature, Construction and Building Materials 226 (2019) 932–944. [45] J.-C. Mindeguia, P. Pimienta, A. Noumowé, M. Kanema, Temperature, pore
[22] Felicetti, P. Bamonte, P. G. Gambarova, Literature review on the behaviour of pressure and mass variation of concrete subjected to high temperature –
UHPFRC at high temperature, in: M. Schmidt, E. Fehling, C. Glotzbach, S. Frölich, experimental and numerical discussion on the spalling risk, Cem. Concr. Res. 40
S. Piotrowski (Eds.), Proceedings of Hipermat – 3rd International Symposium on (2010) 477–487, https://doi.org/10.1016/j.cemconres.2009.10.011.
UHPC and Nanotechnology for High Performance Construction Materials, Kassel, [46] Dorjan Dauti, Alessandro Tengattini, Stefano Dal Pont, Nikolajs Toropovs,
2012, pp. 549-556. Matthieu Briffaut, Benedikt Weber, Analysis of moisture migration in concrete at
[23] B. T. Burke, Residual strength of Ultra-High Performance Concrete after exposure high temperature through in-situ neutron tomography, Cement and Concrete
to elevated temperatures, Master’s thesis, University of Connecticut, 2011. URL: Research 111 (2018) 41–55.
https://opencommons.uconn.edu/gs_theses/44. [47] Sun Bei, Lin Zhixiang, Investigation on spalling resistance of ultra-high-strength
[24] Eurocode 2: Design of concrete structures – Part 1-2: General rules – Structural fire concrete under rapid heating and rapid cooling, Case Studies in Construction
design, Comité Européen de Normation, EN 1992-1-2, 2004. Materials 4 (2016) 146–153.
[25] F. Lu, M. Fontana, in: Effects of polypropylene fibres on preventing concrete [48] X. Liu, G. Ye, G. De Schutter, Y. Yuan, L. Taerwe, On the mechanism of
spalling in fire, Princeton University, 2016, pp. 241–248. polypropylene fibres in preventing fire spalling in self-compacting and high-
[26] Ultra-high performance Concrete: an Emerging Technology Report. American performance cement paste, Cem. Concr. Res. 38 (2008) 487-499. https://doi.org/
Concrete Institute, C-239 Committee, PRC-239-18, 2018. 10.1016/j.cemconres.2007.11.010.
[27] Concrete materials and methods of concrete construction/Test methods and [49] Xiangwei Liang, Chengqing Wu, Yekai Yang, Zhongxian Li, Experimental study on
standard practices for concrete. CSA Group, A23.1/A23.2, 2019. ultra-high performance concrete with high fire resistance under simultaneous
[28] M. A. Sarwar, Characterizing temperature-induced strength degradation and effect of elevated temperature and impact loading, Cement and Concrete
explosive spalling in ultra-high performance concrete. Civil Engineering MS thesis, Composites 98 (2019) 29–38.
Michigan State University, 2017. URL: https://d.lib.msu.edu/etd/6756. [50] Fadzli Mohamed Nazri, Ramadhansyah Putra Jaya, Badorul Hisham Abu Bakar,
[29] Hao-wen Ye, Nai-qian Feng, Yan Ling-hu, Zhi-wei Ran, Li-xun Lin, Shi-kun Qi, Raudhah Ahmadi, A. Hasan, A.A. Khan, Md. A. Mannan, C.N. Hipolito,
Yi Dong, Research on Fire Resistance of Ultra-High-Performance Concrete, N. Mohamed Sutan, Al-K. Hj. Othman, M.R. Kabit, N. Abdul Wahab, Fire resistance
Advances in Materials Science and Engineering 2012 (2012) 1–7. of ultra-high performance fibre reinforced concrete due to heating and cooling,
[30] D. Heinz, F. Dehn, L. Urbonas, Fire resistance of Ultra High Performance Concrete MATEC Web Conf. 87 (2017) 01021, https://doi.org/10.1051/matecconf/
(UHPC) – Testing of laboratory samples and columns under load, in: M. Schmidt, 20178701021.
E. Fehling, C. Geisenhanslüke (Eds.), Proceedings of the International Symposium [51] H. Huang, R. Wang, X. Gao, Improvement effect of fiber alignment on resistance to
on Ultra High Performance Concrete, 2004, pp. 703–715. elevated temperatures of ultra-high performance concrete. Composites, part B 177
[31] Mitsuo Ozawa, Sirjana Subedi Parajuli, Yuichi Uchida, Bo Zhou, Preventive effects (2019),107454. https://doi.org/10.1016/j.compositesb.2019.107454.
of polypropylene and jute fibers on spalling of UHPC at high temperatures in [52] G. F. Peng, J. Yang, Q. Q. Long, X. J. Nin and Q. P. Zeng, Comparison between
combination with waste porous ceramic fine aggregate as an internal curing Ultra-High-Performance Concretes with recycled steel fiber and normal industrial
material, Construction and Building Materials 206 (2019) 219–225. steel fiber, in: E. Ganjian, N. Ghafoori and P. Claisse (Eds.), Proceedings of 4th
[32] Jin-Cheng Liu, Zhigang Zhang, Neural network models to predict explosive spalling International Conference on Sustainable Construction Materials and Technologies,
of PP fiber reinforced concrete under heating, Journal of Building Engineering 32 Las Vegas, 2016, paper S312.
(2020) 101472, https://doi.org/10.1016/j.jobe.2020.101472. [53] J.-H. Lee, Y.-S. Sohn, S.-H. Lee, Fire resistance of hybrid fibre-reinforced, ultra-
[33] Ye Li, Kang Hai Tan, En-Hua Yang, Synergistic effects of hybrid polypropylene and high-strength concrete columns with compressive strength from 120 to 200 MPa,
steel fibers on explosive spalling prevention of ultra-high performance concrete at Magazine of Concrete Research 64 (6) (2012) 539–550.
elevated temperature, Cement and Concrete Composites 96 (2019) 174–181. [54] Jian Yang, Baochun Chen, Camillo Nuti, Influence of steel fiber on compressive
[34] Dong Zhang, Aravind Dasari, Kang Hai Tan, On the mechanism of prevention of properties of ultra-high performance fiber-reinforced concrete, Construction and
explosive spalling in ultra-high performance concrete with polymer fibers, Cement Building Materials 302 (2021) 124104, https://doi.org/10.1016/j.
and Concrete Research 113 (2018) 169–177. conbuildmat.2021.124104.
[35] Gyeongcheol Choe, Gyuyong Kim, Nenad Gucunski, Seonghun Lee, Evaluation of [55] Istruzioni per la progettazione, l’esecuzione e il controllo di strutture in
the mechanical properties of 200 MPa ultra-high-strength concrete at elevated calcestruzzo fibrorinforzato. CNR DT 204/2006, Italian Council for Research, 2008
temperatures and residual strength of column, Construction and Building Materials [in Italian]. https://www.cnr.it/it/node/2624.
86 (2015) 159–168. [56] Testing hardened concrete - Part 3: Compressive strength of test specimens. Comité
[36] D. Hosser, B. Kampmeier, D. Hollmann, Behaviour of Ultra High Performamce Européen de Normation, EN 12390-3, 2009.
Concrete (UHPC) in case of fire, in: M. Schmidt, E. Fehling, C. Glotzbach, S. Frölich, [57] C. Gallé, Effect of drying on cement-based materials pore structure as identified by
S. Piotrowski (Eds.), Proceedings of Hipermat – 3rd International Symposium on mercury intrusion porosimetry, Cement and Concrete Research 31 (10) (2001)
UHPC and Nanotechnology for High Performance Construction Materials, Kassel, 1467–1477.
2012, pp. 573-582.

20