Fire Safety Journal 45 (2010) 283–297

Contents lists available at ScienceDirect

Fire Safety Journal

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

Experimental study on concrete spalling in prestressed slabs subjected to fire

W.Z. Zheng, X.M. Hou n, D.S. Shi, M.X. Xu
School of Civil Engineering, Harbin Institute of Technology, Harbin, China

a r t i c l e in fo abstract

Article history: Experiments on the fire resistance of 15 prestressed concrete (PC) simply-supported slabs and 9 two-span
Received 24 October 2008 unbonded PC continuous slabs were conducted. Preliminary conclusions drawn are that concrete spalls
Received in revised form more easily or more seriously when the compressive stress is higher or the tensile stress is lower on the
20 May 2010
surface exposed to fire, and when the strength and water content of concrete at ambient temperature are
Accepted 3 June 2010
Available online 20 June 2010
higher. Based on the test results, an expression for the top envelope surface of the concrete spalling is
presented by taking the average standard concrete cube strength at ambient temperature as the x-axis,
Keywords: the concrete water content at ambient temperature as the y-axis, and the normal stress level of concrete
Fire extreme fiber in the pre-compression zone of a slab at elevated temperature as the z-axis. To facilitate
Spalling
design and application, an expression for the top envelope surface of the concrete spalling is presented by
Prestressed slabs
keeping the x-axis and the y-axis the same, and the normal stress level of concrete extreme fiber in the
Stress
Concrete pre-compression zone of a slab at ambient temperature as the z-axis.
Water content & 2010 Elsevier Ltd. All rights reserved.

1. Introduction Although important understanding of explosive spalling of

reinforced concrete members subjected to fire has been gained,
Explosive spalling is a disastrous phenomenon that can occur and experiment and analyses on the mechanical behaviour of
when concrete is subjected to fire. It can expose reinforcement prestressed concrete (PC) and reinforced concrete (RC) members
directly to fire and quickly reduce mechanical performance. This subjected to fire have been completed [14–17], there has been no
means a significant reduction of the fire resistance of the concrete systematic report on the spalling of PC members. PC members are
structure occurs due to the direct exposure of the reinforcement to more susceptible to spalling than ordinary concrete members
high temperature. Data from the studies show that concrete subjected to fire because the compressive stress at the pre-
spalling, in general, is affected by a range of factors, including compressive zone is higher or the tensile stress level is lower. Data
concrete strength, concrete density, load intensity and type, from various studies show that predicting fire-induced spalling of PC
permeability, water content, fire intensity, aggregate type, and members is very complex since it is affected by a number of factors.
lateral reinforcement [1–4]. Some researchers have suggested that It is the intention of the authors to use the experiments conducted
spalling is caused by the build-up of pore pressure during heating here to study the influence on spalling of key parameters including
[5–8]. The production of steam in concrete at elevated temperatures concrete strength, water content, and stress levels at the pre-
could increase the possibility of the occurrence of spalling [9]. The compressive zone. Experiments on the fire resistance of 15 PC
higher the strength and density of concrete, the lower the simply-supported slabs and 9 two-span unbonded PC continuous
permeability which increases fire-induced spalling. Permeability slabs were conducted. The range of average values of standard
of concrete varies significantly with temperature and pore pressure. concrete cube strength at ambient temperature was 22.8–56.9 MPa
The higher the permeability at elevated temperatures, the lower the and the range of water content was 1.83–3.98% (by weight). The
developed vapor pressure which reduces fire-induced spalling [10]. nominal tensile stress of concrete extreme fiber in pre-compression
A review of the literature indicates that there is a lack of test data on zone of slab at ambient and elevate temperatures was calculated
the variation of permeability as a function of temperature [11,12]. separately to propose a reference criterion for preliminary judging
Dwaikat and Kodur [13] used a numerical model to predict fire- the risk of spalling of concrete subjected to fire.
induced spalling in concrete structures: the studies show that
permeability, tensile strength of concrete, and heating rate have a
2. Design and fabrication of the specimens
significant influence, but relative humidity also has a marginal
influence on fire-induced spalling.
2.1. Selection of materials

n
Corresponding author.
The concrete mixture ratio by weight is as follows: cement:
E-mail addresses: [email protected] (W.Z. Zheng), sand:coarse aggregate:water ¼1:1.41:2.74:0.42. Ordinary Port-
[email protected] (X.M. Hou). land cement (OPC) of Strength Class 42.5 MPa is used throughout

0379-7112/$ - see front matter & 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.firesaf.2010.06.001
284 W.Z. Zheng et al. / Fire Safety Journal 45 (2010) 283–297

Nomenclature L length of the slab, mm

L0 span of the slab, mm
Ap area of prestressing steel wire, mm2 T heating time, min
As area of non-prestressed steel bar, mm2 PPR prestressing degree
b slab width, mm o water content
c concrete cover thickness, mm b elasticity modulus ratio of the strip concerned and the
mfcu measured standard concrete cube strength at ambient extreme strip on the surface exposed to fire
temperature, MPa H load level
mft average value of axial tensile strength limit of sct nominal tensile stress of concrete extreme fiber in
concrete at ambient temperature, MPa pre-compression zone of slab at ambient tempera-
mft(T) average value of axial tensile strength limit of ture, MPa
concrete at elevated temperature, MPa sct(T) nominal tensile stress of concrete extreme fiber in
fptk characteristic tensile strength limit of prestressing pre-compression zone of slab at elevated tempera-
steel wire, MPa ture, MPa
fy yield strength of non-prestressed steel bar, MPa spe effective prestress, MPa
h slab depth, mm sp calculated ultimate stress in unbonded prestressing
hs distance between point for resultant of forces for steel wire at ultimate state in flexure, MPa
longitudinal tensile non-prestressing steel bars to Dsp stress increase in the unbonded prestressing steel
compressive extreme fiber of section, mm wire beyond the effective prestress, up to ultimate
hp distance between point for resultant of forces for state at ambient temperature, MPa
longitudinal tensile prestressing steel wires to com-
pressive extreme fiber of section, mm

the investigation. The coarse aggregate is calcium carbonate and The dimensions of the furnace chamber are 3  4 m2. Standard
the fine aggregate is silica. The main chemical composition of weights were used to apply the loads on the test slabs. First, the
coarse aggregate is calcium carbonate. Coarse aggregate of design loads were imposed on the slabs to be tested. Heat was
maximal size 20 mm is used. The density of the concrete is generated in the test furnace in accordance with the ISO834
2400 kg/m3. The prestressing steel tendons are low relaxation jP5 standard heating curve [18] by burning oil. The furnace tempe-
steel wires with a fptk value of 1670 MPa and low relaxation ratures were measured by 8 thermocouples. The heating
jS15.24 steel strands with fptk value of 1860 MPa. The prestres- curve monitored by the thermocouples in the furnace is shown
sing steel strands jS15.24 are used in slab UPSA-S, and the in Fig. 4.
prestressing steel wires jP5 are used in the other slabs; the The temperature field was measured by thermocouples
details are listed in Table 1.The anchorage for the prestressing embedded on the surfaces of the bars and the concrete.
steel wires is a homemade button-head anchorage and that for The temperature on the unexposed surface was monitored
the prestressing strands is a single hole wedge-type anchorage. by the postpositional thermocouples. The deflections were
The stresses in prestressing steel wires by load cells installed measured at mid-span of each slab by a differential displacement
at the anchorage ends are shown in Fig. 1. transducer (LVDT) (range: 7150 mm) and the stresses in the
prestressing steel wires were measured by load cells installed
2.2. PC simply-supported slabs at the anchorage ends. The support reactions of the continuous
slabs when subjected to fire were measured by load cells at the
two side supports. These spans were not restrained against
The details of 15 PC simply-supported slabs for the experiments
thermal expansion. The whole intermediate support of the two-
are listed in Table 1. The specimens were cured outdoors, the
span PC continuous slabs was located within the furnace chamber
ambient temperature ranged from 2 to 18 1C, and the ages of the
and is also exposed to the fire which is similar to actual fire case.
specimens were 40–71 days. Variability of concrete strength and
The test furnace for testing continuous slabs after being
water content was found by testing the specimens before the
reconstructed with the intermediate supporting pier is shown in
spalling experiments. The layout of the non-prestressed steel bars
Fig. 5.
and prestressing steel wires in UPSS-1 is shown in Fig. 2. Slab PSS-1
Water and vapor escaping from the surfaces of the slabs,
is a pretensioned PC slab and the other slabs are unbonded PC slabs.
as well as through cracks and the nearby intermediate
support could be observed on the platforms at two sides of the
2.3. Two-span PC continuous slabs furnace. Concrete spalling at the bottom of the slabs was
monitored through observation holes at two sides of the furnace
The details of 9 two-span unbonded PC continuous slabs for [19–21].
the experiments are listed in Table 2. The profiles of prestressing A general picture of the simply-supported slabs installation
steel wires in slab UPCS-1 are shown in Fig. 3. is shown in Fig. 6. A steel beam can be seen to be firmly installed
over each of the test slabs. Some hangers are placed between
each of the slabs and the corresponding steel beam in order
3. Test scheme and spalling phenomena to prevent the slab dropping into the furnace on failure.
In addition, the steel beam was used to fix displacement
3.1. Test scheme transducers (LVDT). A general picture of the continuous slabs
installation is shown in Fig. 7. The experimental program of
The loaded slab tests under fire conditions were carried out in explosive spalling behaviour of PC members subjected to fire is
the test furnace at the Tianjin Fire Research Institute (Tianjin, China). shown in Fig. 8.
 W.Z. Zheng et al. / Fire Safety Journal 45 (2010) 283–297 285

Table 1
Key parameters of simply-supported slabs.

Specimens c (mm) PPR spe (MPa) Z Ap As b  h (mm) L/L0 (mm) mfcu (MPa) o (%) Ages of the
specimens (d)

UPSS-1 15 0.41 655 0.42 3jp5 2j6.5 + 3j8 600  80 3500/3300 56.9 3.98 55
UPSS-2 15 0.54 780 0.45 3jp5 3j8 600  80 3500/3300 48.2 3.50 51
UPSS-3 15 0.70 1013 0.52 5jp5 3j8 600  80 3500/3300 56.9 3.98 55
UPSS-4 25 0.42 806 0.54 3jp5 5j8 600  90 3500/3300 52.1 2.36 49
UPSS-5 25 0.56 988 0.34 3jp5 2j6.5 + 2j8 600  90 3500/3300 52.0 3.29 48
UPSS-6 25 0.68 651 0.44 5jp5 2j6.5 + 1j8 600  90 3500/3300 52.1 2.36 49
UPSS-7 30 0.42 1022 0.43 2jp5 3j6.5 + 2j8 600  95 3500/3300 52.0 3.29 48
UPSS-8 30 0.56 701 0.52 4jp5 2j6.5 + 2j8 600  95 3500/3300 47.3 3.49 46
UPSS-9 30 0.72 842 0.36 5jp5 2j6.5 + 1j8 600  95 3500/3300 22.8 1.83 44
PSS-1 25 0.55 900 0.35 3jp5 2j6.5 + 3j8 600  90 3500/3300 41.3 2.67 40
UPSA-1 15 0.40 678 Only the weight 3jp5 2j6.5 + 3j8 600  95 3500/3300 40.9 2.77 54
of the slab
UPSA-2 15 0.71 1036 5jp5 3j8 600  90 3500/3300 40.9 2.77 54
UPSA-3 30 0.40 989 2jp5 3j6.5 + 2j8 600  95 3500/3300 47.3 3.49 46
UPSA-4 30 0.68 674 6jp5 1j6.5 + 2j8 600  90 3500/3300 22.8 1.83 44
UPSA-S 15 0.70 900 js15 4j8 600  95 3500/3300 48.2 3.50 52

Note:
(1) Non-prestressed steel bars and prestressing steel wire of the critical section at mid-span of the slabs have the same concrete cover thickness. The concrete cover at the
critical section is represented by c.
(2) PPR ¼ Ap sp =ðAp sp þ As fy Þ (prestressing degree); sp is the predicted ultimate stress in unbonded prestressing steel wire at the limit state at ambient temperature;
sp ¼ spe + Dsp, where spe is the effective prestress at ambient temperature and Dsp is the stress increase in the unbonded prestressing steel wire beyond the effective
prestress, up to ultimate state at ambient temperature; fy is the average measured yield strength of the non-prestressed bar.
(3) Z (load level) is the ratio of the imposed load (including the weight of the slab) in the fire condition to the maximum load that the slab can carry at ambient
temperature.
(4) mfcu is the measured standard cube (150  150  150 mm3) strength of concrete. The concrete strengths were measured just prior to the spalling tests.
(5) The average measured tensile strength limit, proportional limit, and yield strength of jP5 prestressing steel wire are 1722, 1343, and 1498 MPa, and 1918, 1496, and
1668 MPa for jS15.24. The average measured yield strength and tensile strength limit of HPB235-j6.5 steel bars are 328 and 446 MPa, and 320 and 478 MPa for
HPB235-j8 bars.
(6) The profile of the wires in the unbonded PC slabs is a parabola. The wires stretch out straight from the centre line of the slab depth at the ends, and the distance
between the contra-flexure point and the centre line of the support is 15% of the span.
(7) L and L0 are the length and span of the slab, respectively. The exposed span of the simply-supported slab is 3050 mm.
(8) o is water content at ambient temperature.

ignition and within the concrete temperature range (tested by

embedded thermocouples in the slabs) 200–500 1C.
External loads on the slabs were kept constant during each fire
test, and deflections of the specimens were found to increase with
heating. Variation of the stress of the unbonded prestressing steel
wires in the slabs subjected to fire can be divided into 3 stages. In
the first stage, corresponding to the initial stage of heating, the
stresses in the unbonded prestressing steel wires were increasing
because the increase in stress values in the unbonded prestressing
steel wires caused by the deflection of the slabs were greater than
the reducing value caused by elastic modulus deterioration,
swelling, and stress relaxation of the tendons in the initial
stage when subjected to fire. In the second stage, the reverse of
stage 1 occurred because the members were heated for a certain
time and the stresses reduced until the heating stopped. In the
third stage, the stresses were retrieved slightly because the
modulus of the steel recovered and its swelling was reduced.
Support reactions measured by load cells at the side supports
indicated that the side support reactions decreased rapidly in
the initial stages and increased after that. The reason for this
Fig. 1. Load cell installed at the anchorage end of unbonded prestressing steel wire.
was that the side supports of the continuous slabs had tended
to rise because the slabs expanded in the initial stages of
heating and after that they would increase because cracks on
3.2. Test phenomena the unexposed surface at the intermediate support and plastic
internal force redistribution were occurring. The experimental
In 15 simply-supported slabs and 9 continuous slabs, spalling stress–time curves of the unbonded prestressing steel wires of the
of the concrete occurred on 8 simply-supported slabs and 3 simply-supported slabs are shown in Fig. 9. The experimental
continuous slabs to different degrees. The number of spalling stress–time curves of the unbonded prestressing steel wires of the
slabs, first and last times of spalling, first time of water overflow, continuous slabs are shown in Fig. 10. The experimental side
the total heating time, and the depth and area of spalling are support reaction force–time curves of the continuous slabs are
listed in Table 3. Concrete spalling occurred during 4–20 min after shown in Fig. 11.
286 W.Z. Zheng et al. / Fire Safety Journal 45 (2010) 283–297

Fig. 2. Layout of non-prestressed steel bars and prestressing steel wires in UPSS-1: (a) layout of non-prestressed steel bars and (b) profile of prestressing steel wires.

Table 2
Key parameters of continuous slabs.

Specimens c (mm) PPR spe (MPa) Z Ap As b  h (mm) L/L0 (mm) mfcu (MPa) o (%) Ages of the
specimens (d)

UPCS-1 15 0.37 657 0.36 3jp5 2j8 + 2j10 600  70 5300/2550  2 56.9 3.98 57
UPCS-2 15 0.53 823 0.50 3jp5 2j6.5 + 2j8 600  70 5300/2550  2 56.9 3.98 57
UPCS-3 15 0.69 1019 0.59 3jp5 3j6.5 600  70 5300/2550  2 48.2 3.50 53
UPCS-4 25 0.39 918 0.58 2jp5 4j8 600  80 5300/2550  2 40.1 2.77 71
UPCS-5 25 0.56 972 0.33 3jp5 4j8 600  80 5300/2550  2 52.1 2.36 68
UPCS-6 25 0.74 760 0.36 4jp5 3j6.5 600  80 5300/2550  2 48.2 3.50 69
UPCS-7 30 0.41 1000 0.42 2jp5 5j8 600  85 5300/2550  2 52.1 2.36 68
UPCS-8 30 0.59 753 0.48 3jp5 3j6.5 + 1j8 600  85 5300/2550  2 52.0 3.29 65
UPCS-9 30 0.73 830 0.32 4jp5 2j6.5 + 1j8 600  85 5300/2550  2 52.0 3.29 66

Note: Non-prestressed steel bars and prestressing steel wire of the critical section at mid-span of slabs have the same concrete cover thickness, and the concrete cover at
intermediate support section is 5 mm. The exposed span of slabs UPCS-1,UPCS-7, and UPCS-8 is 2350 mm, and the exposed span of slabs UPCS-2 UPCS-6, and UPCS-9 is 2120 mm.

Fig. 3. Profiles of prestressing steel wire in slab UPCS-1.

It should be mentioned here that the crack distribution on the in the continuous slabs UPCS-3 and UPCS-6. Two jp5 prestressing
unexposed surface at the intermediate support of the continuous steel wires in the continuous slab UPCS-3 were broken, as
slab was similar to that at ambient temperature. Water began shown in Fig. 13. In the continuous slab UPCS-6, not only were
spilling over the surface of the slabs which was not exposed to two jp5 prestressing steel wires broken, but spalling also
fire, yielding a little water vapor after 11–18 min of heating. occurred at the non-exposed surface after 43 min heating, and
Water appeared near the embedded parts first, and then in more so a hole of 50 mm diameter and 860 mm from the tensioning
and more places, but reduced after the initial stages, and end appeared. After 45 min heating, the slab was broken in a
disappeared gradually after 50 min of heating. section of 920 mm from the tensioning end, and at 55 min
Simply-supported slab UPSA-2 suddenly disrupted after another break of the slab occurred 900 mm from the anchorage
26 min of heating because of concrete spalling at the exposed end; spalling occurred at the exposed surface and progressed
surface and exposure of the reinforcement to the fire, as through the depth of the entire slab, as shown in Fig. 14. Concrete
shown in Fig. 12. Four jp5 prestressing steel wires and one j8 spalling and the crack distribution of the other slabs are shown in
reinforcement were broken in the fire. Concrete spalling occurred Figs. 15–22.
 W.Z. Zheng et al. / Fire Safety Journal 45 (2010) 283–297 287

1000

800
Temperature (°C)

600
ISO834 Standrad heating curve
Test average temperature in furnce
400

200

0
0 10 20 30 40 50 60
Time (min)

Fig. 4. ISO834 heating curve and tested heating curve in furnace.

Fig. 7. General fire test picture of PC continuous slabs.

4. Top envelope curve and surface of concrete spalling based

on test results

4.1. Top envelope curve and surface of concrete spalling based on

normal stress of concrete extreme fiber in pre-compressive zone at
elevated temperature

There are thermal gradients in the slabs from the exposed to

the unexposed surfaces subjected to fire, and the elastic modulus
of concrete is different at different temperatures. To calculate
the nominal normal stress of concrete extreme fiber in the
pre-compressive zone at elevated temperature, the section of
the slab is divided into strips along its depth according to the
distribution of the temperature field. On the basis of invariable
width of the slab on the surface exposed to fire, the transformed
section of the slab subjected to fire can be obtained by
multiplying the width of each strip by the elasticity modulus
ratio b of the strip concerned and in the extreme strip in the
Fig. 5. Test furnace for testing continuous slabs after being reconstructed: A fire-
surface exposed to fire, the reinforcement transformation is the
resistant steel intermediate supporting, Bfirebrick intermediate supporting pier,
C side support, D thermocouples monitoring temperature of furnace, E fuel same as for concrete. The transformed section is shown in Fig. 23.
injection, and F aluminosilicate refractory fibers. The measured effective stress in the prestressing steel wire
was used to calculate the normal stress of concrete extreme fiber
in the pre-compressive zone at elevated temperature. Inner force
in the continuous slabs was calculated using the measured
support reaction and actual load. The influence of the load
induced thermal strain (LITS) occurring during the fire was
included in the stress and support reaction, and the LITS was
indirectly included in the calculations of the normal nominal
stress. The heating time for a non-spalling slab or a non-spalling
region in the slab is considered to be 20 min. Every specimen has
a multi-region as shown in Fig. 24.
Concrete spalls more easily or more seriously when the
compressive stress is higher on the surface exposed to fire
and when the strength and the water content of the concrete
at ambient temperature are higher. Taking average standard
concrete cube strength at ambient temperature as the x-axis
and the difference (which is called the normal stress level)
s(T)¼ sct(T)  mft(T) between nominal tensile stress sct(T) (tensile
stress is positive) at an extreme fiber in the transformed section
subjected to fire and the tensile strength of the concrete exposed
surface at elevated temperature as the y-axis, points with
different water content are drawn using these coordinates, as
shown in Fig. 25. The spalling envelopes were derived by curve
Fig. 6. General fire test picture of PC simply-supported slabs. fitting. It should be mentioned here that the nominal tensile
288 W.Z. Zheng et al. / Fire Safety Journal 45 (2010) 283–297

Explosive spalling behaviour of

PC members subjected to fire

PC simply-supported slabs PC continuous slabs

Temperature Stress of Spalling, Support

 fcu, 
of concrete prestressing steel vapor reaction
measurements
test wire, deflection observation test
of mid-span test

Temperature Top envelope curve and Spalling of concrete

range of spalling surface of spalling and crack
occurred distribution of slabs

Effect of nominal stress ,concrete

strength and water weight content
on spalling analysis

Fig. 8. Experimental program of explosive spalling behaviour of PC members subjected to fire.

Table 3
Schedule of spalling and overflowing in surface subjected to fire.

Specimens First time of Last time of Beginning time of water Total heating time Maximum depth Area of spalling
spalling (min) spalling (min) overflowing (min) (min) of spalling (mm) (%)

UPSS-2 15 20 15 68 25 13.1
UPSS-3 17 20 18 68 10 2.92
UPSS-7 11 18 13 112 6 0.65
UPSS-8 11 19 14 112 10 1.65
UPSS-9 12 19 14 112 20 11.6
UPSA-2 10 20 13 28 47 63.1
UPSA-S 11 22 13 28 23 48.4
PSS-1 12 18 15 112 16 2.19
UPCS-3 10 22 16 60 49 21.3
UPCS-6 7 36 13 55 Perforation 27.0
UPCS-9 4 11 11 90 15 2.19

stress sct(T) is caused by the equivalent load of prestressing steel (

sðTÞ ¼ 0:7126o2:654 ð1:6 r o r3:5Þ
wires and by loads in fire conditions. ð2Þ
sðTÞ ¼ 2:083o7:452 ð3:5 o o r4:2Þ
Figs. 25 and 26 show that all spalling points are subjacent to
the top envelope curves of concrete spalling. The top envelope
curves of concrete spalling are expressed by The expression for the top envelope surface of concrete
( spalling is presented by taking the average standard concrete
sðTÞ ¼ 0:04692mfcu 2:419 ð20 r mfcu r47:5Þ
ð1Þ cube strength at ambient temperature as the x-axis, the water
sðTÞ ¼ 0:1144mfcu 5:671 ð47:5o mfcu r 65Þ
content in concrete at ambient temperature as the y-axis, and the
 W.Z. Zheng et al. / Fire Safety Journal 45 (2010) 283–297 289

1200
1200 1200
Heating time(68min)
1000 Heating time(68min)
Slab being suspended by screw(57min) 1000 Slab being suspended by screw(53min) 1000
Stress (MPa)

800

Stress (MPa)
800

Stress (MPa)
800 firing stopped
firing stopped
Heating time(68min)
600 600
firing stopped 600 Slab being suspended by screw(58min)

400 400 400

200 200 200

0 0 0
0 20 40 60 80 100 120 0 20 40 60 80 100 120 0 20 40 60 80 100 120
Time (min) Time (min) Time (min)
1200 1200 1200
Heating time(68min)
Heating time(68min)
1000 1000 1000
Stress (MPa)

Stress (MPa)

Stress (MPa)
800 firing stopped 800 800
firing stopped firing stopped
600 600 Heating time(112min)
600
Slab being suspended by screw(94min)
400 400 400

200 200 200

0 0 0
0 20 40 60 80 100 120 0 20 40 60 80 100 120 140 160 0 20 40 60 80 100 120
Time (min) Time (min) Time (min)
1200 1200
Heating time (112min) 1200
1000 1000 Slab being suspended by screw (99min)
1000 Heating time (112min)
firing stopped
Stress (MPa)

Stress (MPa)

firing stopped

Stress (MPa)
800 800 800
firing stopped
Heating time (112min)
600 600 600
400 400 400
200 200 200
0 0 0
0 20 40 60 80 100 120 140 160 0 20 40 60 80 100 120 140 160 0 20 40 60 80 100 120 140 160
Time (min) Time (min) Time (min)

1200 1200 1200

The prestressing steel wire Heating time (28min)
Heating time (28min) was broken at 21min
1000 1000 1000
firing stopped
Stress (MPa)

Stress (MPa)
Stress (MPa)

800 800 800

600 600 600

400 400 400

200 200 200

0 0 0
0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40 0 10 20 30 40 50 60 70 80 90
Time (min) Time (min) Time (min)

1200
Heating time (28min)
1000
Stress (MPa)

800

600

400

200

0
0 10 20 30 40 50 60 70 80 90
Time (min)

Fig. 9. The stress–time curves of the unbonded prestressing steel wires in simply-supported slabs subjected to fire. (a) UPSS-1, (b) UPSS-2, (c) UPSS-3, (d) UPSS-4,
(e) UPSS-5, (f) UPSS-6, (g) UPSS-7, (h) UPSS-8, (i) UPSS-9, (j) UPSA-1, (k) UPSA-2, (l) UPSA-3, and (m) UPSA-4.
290 W.Z. Zheng et al. / Fire Safety Journal 45 (2010) 283–297

1200 1200
1000 1000 Heating time (60min)
Heating time (60min)

Stress (MPa)
Stress (MPa)
800 800
600 600
400 400
firing stopped firing stopped
200 200
0 0
0 20 40 60 80 100 120 140 160 0 20 40 60 80 100 120 140 160
Time (min) Time (min)
1200 1200
1000 1000 Heating time(60min)
Stress (MPa)

Stress (MPa)
800 The prestressing steel wire
800
was broken at 29min
600 600
400 400
firing stopped
200 200
0 0
0 20 40 60 80 100 120 140 160 0 20 40 60 80 100 120 140 160
Time (min) Time (min)
1200
1200
1000 Heating time(90min)
1000
Stress (MPa)

800
Stress (MPa)

800 The prestressing steel wire

was broken at 35min
600 600
400 400
firing stopped
200 200
0 0
0 20 40 60 80 100 120 140 160 0 20 40 60 80 100 120 140 160
Time (min) Time (min)
1200 1200
Heating time(90min)
1000 1000 Heating time(55min)
Stress (MPa)
Stress (MPa)

800 800

600 600

400 400
firing stopped
firing stopped
200 200
0 0
0 20 40 60 80 100 120 140 160 0 20 40 60 80 100 120 140 160
Time (min) Time (min)
1200

1000 Heating time(90min)

Stress (MPa)

800

600

400
firing stopped
200

0
0 20 40 60 80 100 120 140 160
Time (min)

Fig. 10. The stress–time curves of the unbonded prestressing steel wires in continuous slabs subjected to fire. (a) UPCS-1, (b) UPCS-2, (c) UPCS-3, (d) UPCS-4, (e) UPCS-5,
(f) UPCS-6, (g) UPCS-7, (h) UPCS-8, and (i) UPCS-9.
 W.Z. Zheng et al. / Fire Safety Journal 45 (2010) 283–297 291

5.0 5.5

Side support reactrion (kN)

Side support reactrion (kN)
4.5
5.0
4.0 The anchorage end The anchorage end
The tensioning end 4.5 The tensioning end

3.5
4.0 Heating time (60min)
3.0 Heating time (60min)

2.5 3.5
0 40 80 120 160 200 0 40 80 120 160 200
Time (min) Time (min)

4.0
Side support reactrion (kN)

Side support reactrion (kN)

5.5
3.9 The anchorage end
The tensioning end
5.0
3.8
The anchorage end
3.7 4.5 The tensioning end

3.6 Heating time (55min)

The slab was disrupted at 30min 4.0
3.5
0 10 20 30 40 0 20 40 60 80 100 120 140
Time (min) Time (min)
3.5 Side support reactrion (kN) 4.5
Side support reactrion (kN)

The anchorage end

The tensioning end The anchorage end
3.0 The tensioning end
4.0

2.5
3.5
2.0 Heating time (90min) The slab was disrupted at 35min

1.5 3.0
0 40 80 120 160 200 0 10 20 30 40
Time (min) Time (min)
5.5 5.5
Side support reactrion (kN)

Side support reactrion (kN)

The anchorage end

5.0 The tensioning end 5.0

4.5
4.5 The anchorage end
The tensioning end
4.0
Heating time (90min)
4.0 Heating time(55min)
3.5
3.5
0 40 80 120 160 200 0 20 40 60 80 100 120 140
Time (min) Time (min)

5.0
Side support reactrion (kN)

The anchorage end

4.5 The tensioning end

4.0

3.5

3.0 Heating time (90min)

2.5
0 40 80 120 160 200
Time (min)

Fig. 11. Support reaction–time curves of the continuous slabs subjected to fire. Note: The anchorage end means the side support reaction at the anchorage end, while the
tensioning end means the side support reaction at the tensioning end. (a) UPCS-1, (b) UPCS-2, (c) UPCS-3, (d) UPCS-4, (e) UPCS-5, (f) UPCS-6, (g) UPCS-7, (h) UPCS-8, and
(i) UPCS-9.
292 W.Z. Zheng et al. / Fire Safety Journal 45 (2010) 283–297

Fig. 12. Concrete spalling and reinforcement breaking in UPSA-2 simply-supported slab: (a) spalling of exposed surface, (b) reinforcement breaking, and (c) area and depth
of concrete spalling (hatched means spalling, number in the hatched area means the maximum depth of spalling).

Fig. 13. Concrete spalling, reinforcement breaking, and crack distribution in UPCS-3 continuous slab: (a) spalling of exposed surface, (b) reinforcement breaking, and
(c) area and depth of concrete spalling and crack distribution.
 W.Z. Zheng et al. / Fire Safety Journal 45 (2010) 283–297 293

Fig. 14. Spalling of concrete, reinforcement breaking, and crack distribution in UPCS-6 continuous slab: (a) spalling of exposed surface, (b) slab breaking subjected to fire,
(c) spalling hole of unexposed surface, (d) reinforcement breaking, and (e) area and maximum depth of spalling of concrete and crack distribution.

Fig. 15. Spalling of concrete and crack distribution in UPSS-2 simply-supported slab.

normal stress level s(T)¼ sct(T)  mft(T) of concrete extreme fiber Fig. 27 shows that all spalling points are subjacent to the top
in the pre-compression zone of the slab at elevated temperature envelope surface of concrete spalling, and the top envelope
as the z-axis. Points with different concrete strengths and water surface of concrete spalling is expressed as
contents are shown in Fig. 27, and the top envelope surface of
concrete spalling is also drawn in Fig. 27. sðTÞ ¼ ð0:0068mfcu þ1:7763Þ ð0:5053o1:6100Þ ð3Þ
294 W.Z. Zheng et al. / Fire Safety Journal 45 (2010) 283–297

Fig. 16. Spalling of concrete and crack distribution in UPSS-3 simply-supported slab.

Fig. 17. Spalling of concrete and crack distribution in UPSS-7 simply-supported slab.

Fig. 18. Spalling of concrete and crack distribution in UPSS-8 simply-supported slab.

Fig. 19. Spalling of concrete and crack distribution in UPSS-9 simply-supported slab.

Fig. 20. Spalling of concrete in UPSA-S simply-supported slab.

4.2. Top envelope curve and surface of concrete spalling based on temperature as the y-axis, and by taking the water content in
normal stress of concrete extreme fiber in pre-compressive zone at concrete at ambient temperature as the x-axis and the normal
ambient temperature stress level of concrete extreme fiber in the pre-compression zone
of the slab at ambient temperature as the y-axis.
To facilitate design and application, two expressions for the Following the method of drawing the top envelope curve of
top envelope curve of concrete spalling are presented by taking concrete spalling at elevated temperature, the top envelope
the average standard concrete cube strength at ambient tem- curves of concrete spalling based on the normal stress level of
perature as the x-axis and the normal stress level of concrete concrete extreme fiber pre-compressive zone at ambient tem-
extreme fiber in the pre-compression zone of the slab at ambient perature are obtained, as shown in Figs. 28 and 29. Corresponding
 W.Z. Zheng et al. / Fire Safety Journal 45 (2010) 283–297 295

Fig. 21. Spalling of concrete and crack distribution in PSS-1 simply-supported slab.

Fig. 22. Spalling of concrete and crack distribution in UPCS-9 continuous slab.

EXPOSED SURFACE EXPOSED SURFACE

SECTION AT NORMAL TRANSFORMED SECTION

TEMPERATURE AT ELEVATED TEMPERATURE

Fig. 23. Original and transformed section of the slab.

4
Top envelope curve of spalling
3.50
2 2.67
3.98
0 1.83
3.29
 (T) (MPa)

-2 3.49
2.77
-4

-6 Spalling points
No-spalling points
-8

Fig. 24. Section for checking computations of spalling.

-10
20 25 30 35 40 45 50 55 60 65
 fcu (MPa)

expressions are given as Eqs. (4) and (5). It should be mentioned Fig. 25. Spalling points, non-spalling points, and top envelope curve of spalling in
here that the values used for s ¼ sct  mft in the y-axis of Figs. 28 o  s(T) coordinates. (Numbers near spalling points are water weight contents at
ambient temperature of the points. The results are for ISO834 heating and
and 29 are the values at ambient temperature.
calcareous concrete.)
(
s ¼ 0:001943mfcu 2:319 ð20 r mfcu r47:5Þ
ð4Þ
s ¼ 0:04784mfcu 4:500 ð47:5 o mfcu r65Þ The expression of the top envelope surface of concrete spalling
is presented by taking the average standard concrete cube strength
( at ambient temperature as the x-axis, the water content in concrete
s ¼ 0:02869o2:328 ð1:6 r o r 3:5Þ
ð5Þ at ambient temperature as the y-axis, and the normal stress level
s ¼ 0:9383o5:511 ð3:5 o o r 4:5Þ s ¼ sct  mft of concrete extreme fiber in the pre-compression zone
296 W.Z. Zheng et al. / Fire Safety Journal 45 (2010) 283–297

4 of the slab at ambient temperature as the z-axis; points with

Top envelope curve of spalling
different concrete strengths and water contents are shown in
2 41.3 Fig. 30, and the top envelope surface of concrete spalling is also
41.3
48.2 drawn in Fig. 30.
0 Fig. 30 shows that all spalling points are subjacent to the top
envelope surface of concrete spalling, and the top envelope
52.1 47.3
 (T) (MPa)

-2 surface of concrete spalling is expressed as

22.8
41.0 s ¼ ð0:0068mfcu þ1:7763Þð0:1053o1:2411Þ ð6Þ
-4
The parameter effective ranges in Eqs. (1)–(6) are mfcu ¼
20–65 MPa and o ¼1.6–4.2%.
-6 Spalling points
Figs. 27 and 30 (Eqs. (3) and (6)) show that the nominal stress
No- spalling points
of concrete and concrete strength have significant influence on
-8
the extend of spalling in concrete. Higher nominal tensile stress
on the exposed surface and lower concrete strength lead to less
-10 spalling in concrete. The water content has a minor effect on
1.6 2.0 2.4 2.8 3.2 3.6 4.0
 (%) spalling.
It should be mentioned here that the number of non-spalling
Fig. 26. Spalling points, non-spalling points, and top envelope curve of spalling in points inside the spalling envelope (as shown in Figs. 25–30)
mfcu  s(T) coordinates. (Numbers near spalling points are average standard are the points E and F (as shown in Fig. 24) in continuous slabs
concrete cubic strengths at ambient temperature of the points. The results are
UPCS-1 and UPCS-9. The points E and F are near the intermediate
for ISO834 heating and calcareous concrete.)

1 Top envelope curve of spalling

0
5 41.3
-1
48.2
-2
0 52.1
(T) (MPa)

22.9 47.3
-3 41.3
 (MPa)

-4
-5 41.0
-5

-6
-10 Spalling points
-7
60 5 No-spalling points
4 -8
μf 40 3
c (M -9
Pa 2 (%)
) 20 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2
1
 (%)
Fig. 27. Distribution of spalling points, non-spalling points, and top envelope
surface of spalling in mfcu wsðTÞ coordinates (‘‘  ’’ is non-spalling points and Fig. 29. Spalling points, non-spalling points, and top envelope curve of spalling in
‘‘,’’spalling points. The results are for ISO834 heating and calcareous concrete). o  s coordinates. (Numbers near spalling points are average standard concrete
cubic strengths at ambient temperature of the points. The results are for ISO834
heating and calcareous concrete.)

1
0
Top envelope curve of spalling
-1 3.50
1.83 5
-2 3.98
2.67 3.29
-3
 (MPa)

3.49 0
-4
(MPa)

2.77
-5
-5
-6
Spalling points
-7
No-spalling points -10
-8
60 5
-9 4
20 25 30 35 40 45 50 55 60 65 μfc 40
(M 3
fcu (MPa) Pa 2
) 20 1 (%)
Fig. 28. Spalling points, non-spalling points, and top envelope curve of spalling in
Fig. 30. Distribution of spalling points, non-spalling points, and top envelope
mfcu  s coordinates. (Numbers near spalling points are water weight contents at
surface of spalling in mfcu ws coordinates (‘‘  ’’ is no-spalling points, and ‘‘,’’
ambient temperature of the points. The results are for ISO834 heating and
spalling points. The results are for ISO834 heating and calcareous concrete).
calcareous concrete.)
 W.Z. Zheng et al. / Fire Safety Journal 45 (2010) 283–297 297

support; the nominal compressive stress of concrete extreme fiber (No. 2001-8), and the Postdoctoral Science Foundation of Harbin
in exposed surface of slabs UPCS-1 and UPCS-9 at elevated Institute of Technology.
temperature is bigger, but the nominal tensile stress of concrete
extreme fiber on non-exposed surface of slabs UPCS-1 and UPCS-9 References
near the intermediate support is bigger. So the cracks near the
intermediate support of the non-exposed surface increased the [1] K.D. Hertz, L.S. Sorensen, Test method for spalling of fire exposed concrete,
permeability of the concrete, and water overflows on the non- Fire Safety Journal 40 (5) (2005) 466–476.
exposed surface of slabs UPCS-1 and UPCS-9 more easily than the [2] K.D. Hertz, Limits of spalling of fire exposed concrete, Fire Safety Journal 38
(2) (2003) 103–116.
other slabs, which leads to the development of lower pore
[3] V.K.R. Kodur, T.C. Wang, F.P. Cheng, Predicting the fire resistance behaviour of
pressure and in turn less spalling. Overall, the expression of the high strength concrete columns, Cement & Concrete Composites 26 (2)
top envelope surface of concrete spalling leads to conservative (2004) 141–153.
[4] F. Kelly, J. Purkiss, Reinforced concrete structures in fire: a review of current
estimates for spalling predictions.
rules, The Structural Engineer 86 (19) (2008) 33–39.
[5] Z.P. Bazant, W. Thonguthai, Pore pressure in heated concrete walls:
theoretical prediction, Magazine of Concrete Research 101 (31) (1979) 67–76.
5. Conclusions [6] V.K.R. Kodur, L. Phan, Critical factors governing the fire performance of high
strength concrete systems, Fire Safety Journal 42 (6-7) (2007) 482–488.
[7] V.K.R. Kodur, Guidelines for fire resistance design of high strength concrete
1. Experiments on the fire resistance of 15 PC simply-supported columns, Journal of Fire Protection Engineering 15 (2) (2005) 93–106.
slabs and 9 two-span PC continuous slabs have been [8] P. Kalifa, F.D. Menneteau, D Quenard., Spalling and pore pressure in HPC at
completed. Spalling of concrete, breaking of the tension high temperature, Cement and Concrete Research 30 (12) (2000) 1915–1927.
[9] Y.N.S. Chan, G.F. Peng, M. Anson, Fire behavior of high performance concrete
reinforcements, and the distribution and development of made with silica fume at different moisture contents, ACI Materials Journal
cracks for 8 simply-supported slabs and 3 continuous slabs 96 (3) (1999) 405–409.
are described. Spalling occurred within the concrete tempera- [10] A.N. Noumowe, R. Siddique, G. Debicki, Permeability of high-performance
concrete subjected to elevated temperature (600 1C), Construction and
ture range 200–500 1C.
Building Materials 23 (5) (2009) 1855–1861.
2. The nominal stress of concrete and concrete strength have [11] D. Gawin, C.E. Majorana, B.A. Schrefler, Numerical analysis of hygro-thermic
significant influence on the extend of spalling in concrete, but behaviour and damage of concrete at high temperature, Mechanics of
the water content has a minor effect on spalling. However, Cohesive-Frictional Materials 4 (1999) 37–74.
[12] P. Kalifa, G. Chene, C. Galle, High-temperature behavior of HPC with
concrete strength is an indirect contributing factor; it is related to polypropylene fibers: from spalling to microstructure, Cement & Concrete
permeability. Preliminary conclusions drawn are that concrete Composites 31 (10) (2001) 1487–1499.
spalls more easily or more seriously when the compressive stress [13] M.B. Dwaikat, V.K.R. Kodur, Hydrothermal model for predicting fire-induced
spalling in concrete structural systems, Fire Safety Journal 44 (3) (2009)
is higher or the tensile stress is lower on the surface exposed to 425–434.
fire and when the strength of the concrete and the water content [14] E. Ellobody, C.G. Bailey, Modelling of unbonded post-tensioned concrete slabs
in the concrete at ambient temperature are higher. under fire conditions, Fire Safety Journal 44 (2) (2009) 159–167.
[15] W.Z. Zheng, X.M. Hou, Experiment and analysis on the mechanical behaviour
3. Based on the test results, expressions for the top envelope of PC simply-supported slabs subjected to fire, Advances in Structural
curve and the surface of concrete spalling are presented by Engineering 11 (1) (2008) 71–89.
curve fitting for ISO834 heating and calcareous concrete only. [16] C.G. Bailey, E. Ellobody, Fire tests on unbonded post-tensioned one-way
concrete slabs, Magazine of Concrete Research 61 (1) (2009) 67–76.
The criteria can be used for preliminary judging the possibility [17] G.M.E Cooke, Behaviour of precast concrete floor slabs exposed to
of spalling in PC members subjected to fire. standardized fires, Fire Safety Journal 36 (4) (2001) 459–475.
[18] International Standard ISO 834-1-1999(E), Fire-resistance tests: elements of
building construction—Part 1: general requirements, International Organiza-
Acknowledgements tion for Standardization, Switzerland, 1999.
[19] Eurocode 2: Design of concrete structures, Part 1–2: general Rules–structural
fire design, prEN 1992-1-2, British Standards Institution, London, British,
The work presented in this paper is supported by the National 2005.
Natural Science Foundation of China (No. 50678050), the [20] GB/T9978, Fire-resistance tests—elements of building construction, 1999.
[21] BS. Fire tests on building materials and structures—Part 20: method from
Changjiang Scholars Program of China (No. 2009-37), the Out- determination of the fire resistance of elements of construction (general
standing Youth Science Foundation of Heilongjiang Province principles), BS 476-3:1987, BSi, UK, 1987.