Experimental Thermal and Fluid Science 62 (2015) 5257

Contents lists available at ScienceDirect

Experimental Thermal and Fluid Science

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

A new hybrid heating method used in re test

Jingsi Huo a, Yan Xiao a,b,, Xiaohu Ren a, Xiang Zeng a,c
a
China Ministry of Education Key Laboratory of Building Safety and Energy Efciency, College of Civil Engineering, Hunan University,
Yuelu Mountain, Changsha 410082, China
b
Department of Civil Engineering, University of Southern California, 3260 S. Vermont Ave., Los Angeles, CA 90089-2531, USA
c
College of Civil Engineering and Architecture, Hainan University, No.58, Renmin Ave., Haikou 570228, China

a r t i c l e i n f o a b s t r a c t

Article history: Based on a common electric furnace, a newly-developed built-up re furnace equipped with liqueed gas
Received 28 June 2014 burners is used to heat a structural column by electricity and blowing liqueed gas ame into the furnace
Received in revised form 25 November 2014 simultaneously. Fire test results indicate that the hybrid heating method can be used to heat the test
Accepted 25 November 2014
specimens closely following the ISO-834 standard re curve, and can be used in re test with excellent
Available online 3 December 2014
convenience and cost-effectiveness because it combines both merits of electric heating and liqueed gas
re heating. The proposed method provides a convenient way to simulate re effects.
Keywords:
2014 Elsevier Inc. All rights reserved.
Fire resistance test
Electric furnace
Liqueed gas burner
Hybrid heating

1. Introduction
Structural damage due to re is a major concern in high-rise
building design. Understanding the re performance and post-re
behaviour of structures and structural components is an important
task for structural designers. Fire furnaces have typically been used
to carry out re tests all over the world since the beginning of the
last century. The International Standard Organization (or ISO) has
established a re test standard and a standard curve to simulate
real re attacks [1]. In the case of China, more than ten re furnaces
are in use to conduct tests on the re resistances of various types of
structural components, such as beams and columns. Due to the
limitation of standard re testing furnaces, custom-built furnaces
and test set-ups have recently been used to conduct re tests on
restrained structural members at elevated temperatures or at
non-uniform temperature exposure. Some researchers have used
specially-built electric furnaces to heat the specimens owing to
the advantages of safety, easy manipulation and reduced environ-
mental pollution.
Spyrou and Davison [2] developed a purpose-built electric fan-
assisted furnace which was designed with image acquisition and
processing system in order to capture the deformations of the test-
ing specimen. The fan can not only avoid the presence of ames
Corresponding author at: China Ministry of Education Key Laboratory of
Building Safety and Energy Efciency, College of Civil Engineering, Hunan Univer-
sity, Yuelu Mountain, Changsha 410082, China. Tel./fax: +86 731 88821395.
E-mail address: yanxiao@hnu.edu.cn (Y. Xiao).
within the eld of view of the video cameras but also ensure a uni-
form distribution of surrounding temperature up to 1100 C. Wang
and Davies [3] developed a gas furnace which consists of six panels
bolted at their junctions for easy assembly. A non-sway loaded
steel column, rotationally restrained by two loaded steel beams,
was exposed to re to evaluate how bending moments in
restrained columns would change and how these changes might
affect the column failure temperatures. Al-Jabri et al. [4] present
test results of twenty semi-rigid steel beam-to-column connec-
tions at ambient and elevated temperatures, with a series of ele-
vated temperature tests conducted on momentrotation
temperature curves of beam-to-column connections. Tests were
performed in a gas-red portable furnace lined with ceramic bre
specially designed for testing connections. An electric rebrick kiln
was used by Bailey and Toh [5] to experimentally study the failure
modes of horizontally unrestrained two-way spanning reinforced
concrete slabs at elevated temperatures. Tan et al. [6] developed
L-shaped electric furnaces in order to provide various heating
schemes, such as three-face or four-face heating for different spec-
imen lengths. Each furnace had four units of heating modules
made of metallic spiral heating elements embedded in foamed
ceramic bre insulation. Gao et al. [7] and Yuan et al. [8] developed
an oil-red furnace to carry out experimental studies on re resis-
tance of unbonded pre-stressed concrete continuous slabs in re.
Wang et al. [9] carried out an experimental study on the re-resis-
tant capacity of extended end-plate joints. Tests were performed in
an oil-red furnace designed for testing joints. The furnace was
specially built according to the shape of tested joints as well as

http://dx.doi.org/10.1016/j.expthermusci.2014.11.015
0894-1777/ 2014 Elsevier Inc. All rights reserved.
 J. Huo et al. / Experimental Thermal and Fluid Science 62 (2015) 5257
the requirements of gas burners and exhaust opening
arrangements.
The literature above shows that researchers tend to assemble
the built-up heating units into a re chamber in order to exibly
enclose different types of testing specimens. It is also noted that
most researchers prefer electric heating to gas or oil re because
of the merits of electric heating, such as safety, reliability, easy
operation, high-accuracy control and reduced pollution. However,
the heating rate of electricity is relatively low compared with that
of gas or oil re exposure. Gas or oil-red furnaces can produce a
speedy increase of temperature required to simulate real re
attacks, however they need complex temperature, air pressure
and exhaust management systems which are generally difcult
to control.
Based on the relative merits of electric furnace and gas or oil-
red furnaces, herein a hybrid heating method is newly developed
to heat test specimens by a stackable electric furnace, in which a
liqueed petroleum gas ame is introduced into the furnace simul-
taneously in order to increase the heating rate to simulate the pre-
scribed re-temperature curve. The furnace heating can be
controlled easily and safely in such a hybrid system using gas
and electricity, such that the average temperature in the furnace
closely follows the ISO-834 standard re curve.
2. Hybrid furnace structure, heating control method and
heating efciency
2.1. Electric furnace
A stackable electric furnace was built rst in order to conduct
re tests on structural columns. The shape and structure of the fur-
nace was determined according to the geometrical dimensions of
test specimens in order to be installed conveniently and exibly.
Fig. 1 shows a schematic view of the stackable furnace, which is
composed of six semi-cylindrical shells with an inner diameter of
590 mm and a height of 630 mm. The furnace was built with small
segments in order to exibly provide different heights of testing in
the chamber. Thus, the furnace can be used to heat columns with
heights of 630 mm, 1260 mm and 1890 mm. The furnace wall is
made of an exterior steel shell, a thick layer of insulation bre
and interior honeycomb ceramic wall. Each semi-cylindrical shell
has a unit of heating module made of metallic spiral heating ele-
ments embedded in foamed ceramic bre insulation. The furnace
rating is 60 kW/380 V/3 phase, with a maximum operating tem-
perature of 1200 C and a maximum heating rate of 100 C/min.
Three Type K thermocouples (chromelalumel) are xed to the
upper, middle and lower interior furnace wall at the middle section
Fig. 1. Schematics of electric furnace (left plan view; right elevation view).
53
of each segment (as shown in Fig. 1), and used to monitor the fur-
nace temperature so that the stackable electric furnace can be used
to provide a uniform heating condition. Three units of heating
module of the upper, middle and lower segments, i.e., two semi-
cylindrical heating elements, can be automatically controlled inde-
pendently or share one heating control signal. The heating control
system is mainly composed of a XMT2100 intelligent temperature
controller, voltage control triggering switch, silicon-controlled
switch. The XMT2100 intelligent temperature controller is made
by Changzhou Nuohai Electronics Co., Ltd., China. The furnace heat-
ing can be close-loop controlled by comparing and processing the
temperature difference between the re furnace temperature and
the pre-determined value in real time, as well as by transmitting
the control signals to continuously adjust the output power rates.
Fig. 2(a) shows the heating control system.
In order to know the maximum heating capacity and the heat-
ing control accuracy, trial re tests without a test specimen were
carried out. Fig. 3 shows a typical temperature versus time curve
labelled as Electric heating, which was obtained with the heating
rate at the highest setting. Obviously, the electric furnaces heating
rate is too lower as it took about 20 min to reach the ISO-834 re
standard curve. There is a remarkable difference between the mea-
sured temperature curve and the ISO-834 standard re curve
before the two joins. However, the electric heating can easily and
closely follow the ISO-834 re standard curve during the temper-
ature increase and during the cooling phases, as shown in Figs. 4
and 6.
2.2. Gas re burner
Since the heating capacity of the electric furnace is not suf-
cient to follow the ISO-834 re standard during the initial 20 min
of the ISO-834 re exposure, a liqueed gas red facility was devel-
oped with the intent to accelerate the initial heating process. This
facility was achieved by installing gas red burners in two re
holes at the bottom of the furnace, thereby allowing for additional
energy to be introduced while the furnace is being heated by elec-
tricity. The liqueed gas facility, as shown in Fig. 2(b), includes the
temperature control unit, proportional valve, gas re burner and
ame igniter, liqueed gas conducts and air-blower. The liqueed
gas facility is equipped with four gas re burners. The furnace
heating can also be controlled in a close-loop fashion by regulating
the gas pressure valve in real time to change the ratio of air to liq-
ueed gas according to the predetermined temperature versus
time history.
In order to know the maximum heating capacity and the heat-
ing control accuracy, trial tests without a test specimen were also
carried out using the gas red facility. Fig. 3 shows the comparison
of different re heating modes with the ISO standard curve. Obvi-
ously, the heating capacity of the gas red facility is a little higher
than that of electric heating during the initial 5 min, but sharply
lower thereafter. Apparently, it would take much more time to
intersect with the ISO-834 re standard curve. The comparison
between temperature versus time histories for electric heating
and the gas red heating demonstrates that the gas re heating,
compared with the electric heating, can result in rapid temperature
rise in the initial several minutes but lower heating capacity at
later stages.
3. Hybrid heating control method and validity of application in
re tests
Although the heating rates of the electric furnace and the gas
red burners have their limitations as shown in Fig. 3, gas red
heating can provide the required rapid temperature rise during
54 J. Huo et al. / Experimental Thermal and Fluid Science 62 (2015) 5257

Fig. 2. Electric heating and gas burner controllers of the re furnace.

t (min)
Fig. 3. Comparisons of measured temperature vs. time curves using different
heating methods with the ISO-834 standard re curves.
(a) temperature rise duration time of 30 min

the initial several minutes but the electric power can be used to
maintain stable heating of the furnace. Therefore, tests were car-
ried out by operating the electric heaters and gas burners simulta-
neously in order to double the heating rate during the initial
phases of re exposure. The electric and gas red heating are both
automatically controlled in real time, and the resulting tempera-
ture is compared with the ISO-834 re standard curve (or other
pre-determined heating curve). When the temperature versus time
curve intersects with the ISO-834 re standard curve, and the elec-
tric or the gas red heating rate at one moment is larger than the
required heating rate, either of the two power sources can be
adjusted gradually to zero, and then the furnace is heated by elec-
tricity or gas re only.
Fig. 3 shows that the hybrid heating method can remarkably t (min)
quicken the temperature rise in the furnace so that the tempera-
(b) temperature rise duration time of 60 min
ture versus time curves can reach the ISO-834 re standard curve
at about 3 min. The curve named as Hybrid heating_E denotes Fig. 4. Comparisons of measured heating and cooling temperature vs. time curves
the measured temperature versus time curve in a test where the with the ISO-834 standard re curves.
 J. Huo et al. / Experimental Thermal and Fluid Science 62 (2015) 5257 55

1200 denotes a case where the furnace was heated by the hybrid method
CF029-1 before the temperature line intersected with the ISO-834 re stan-
dard curve and heated by the gas burners after the intersection
1000 CF029-2
point. Fig. 4 shows the test results for the entire re exposure
CF29 including temperature rise and cooling phases. It is indicated that
800 the hybrid heating method can be used to heat the furnace closely
CF06
following the ISO-834 standard re curve, with the exception of
CF26 the portion where the temperature decreases below 200 C. The
600
CF22 re tests were carried out by using the hybrid heating method with
no specimen in the furnace.
400 CF42
Huo et al. [10] and Zhang [11] carried out an experimental
Scheduled study on the cyclic behaviours of concrete-lled steel tubular col-
200 umns and concrete columns after exposure to re under sustained
axial load. The columns were heated by the hybrid method
0 described in the paper. Figs. 5 and 6 show the curves for the aver-
age measured temperature in the furnace versus time which are in
good agreement with the assigned heating curves. As shown in
Fig. 6, there is a relatively larger difference between the testing
Fig. 5. Heating and cooling temperature vs. time curves measured by Huo et al. curve of Specimen CF13 and the prescribed heating curve due to
[10]. a malfunction in the heating system just before the cooling phase.
Generally, it can be concluded that the newly developed hybrid
heating method can be used to heat the testing specimens effec-
1200
tively, safely and economically.
CF20 According to ISO-834 (1975) [1] and GB/T9978 (1999) [12], the
1000 area difference as well as the temperature difference between the
CF22
measured temperature versus time curve and the prescribed curve
800 CF13 were used to evaluate the heating performance and validate the
newly developed hybrid method. The temperature difference dur-
CF15 ing the initial 10 min should not be larger than 100 C, and the rel-
600
Scheduled ative area difference during the temperature increase phase should
not exceed the corresponding values shown at the bottom of
400 ISO-834 Table 1. The relative area difference d can be determined according
to Eq. (1).
200
A  As
d  100% 1
As
0
where A is the area under the measured temperature versus time
t (min) curve and As is the area under the ISO-834 standard re curve.
Table 1 shows the relative differences of the integrated values
Fig. 6. Heating and cooling temperature vs. time curves measured by Zhang [11]. for the temperature relationships measured by the authors in the
paper, Huo et al. [10] and Zhang [11], as well as the ISO-834 curve.
furnace was heated by the hybrid method before the temperature It can be shown from Table 1 that, generally, the temperature error
line intersected with the ISO-834 re standard curve and heated by (d) decreased gradually with increasing re duration time, and
electricity after the intersection point. The Hybrid heating G line most of the d values are within the allowable values specied by

Table 1
Relative area difference value, d of tested temperature versus time curves (%).

No. Fire tests Fire duration time, t (min) References

10 30 60 90 120 180 240
1 Hybrid heating_E1 2.5 2.4 1.4 1.8 7.4 This paper
2 Hybrid heating_G1 9.2 5.4 2.1 1.0 4.9
3 Hybrid heating_E2 1.4 1.7 1.2 0.6 0.5 4
4 Hybrid heating_G2 9.0 5.6 1.3 0.2 0.14 4.4
5 CF029-1 11 5.8 0.4 0.8 6.2 1.2 1.5 Huo et al. [10]
6 CF029-2 12.5 6.1 3.7 1.8 1.8 2.8 5.4
7 CF29 11 1.8 0.4 0.8 0.7 1.1 1.5
8 CF06 20 8.9 5.3 7.1 2.7 1.8
9 CF26 17 4.7 1.1 2.8 3 1.9
10 CF22 5.2 2.2 4.9 2.3 1.6
11 CF42 10.5 2.0 2.4 8.5 10.2
12 CF20 17.8 8.6 3.5 3.0 1.1 1.2 0.1 Zhang [11]
13 CF22 9.1 9.6 5.4 3.3 2.3 1.6 1.1
14 CF13 1.9 7.4 5.9 2.5 3.7 18.6
15 CF15 9.2 9.0 7.5 3.2 0.1 14.4

Allowable value 15 10 5 5 5 5 5

Note: The underlined area difference value exceed the allowable value.
56 J. Huo et al. / Experimental Thermal and Fluid Science 62 (2015) 5257

1200 1200
Welch et
Foster et al.(2007)
al.(2007)
1000 Test
1000 Test

800 800

600 600

400 400

200 200

0 0

t (min) t (min)

(a) Foster et al. [13] (a) Welch et al. [15]

1000
1000
Wald et al.(2004) Abecassis-Empis et al.(2008)
800 Test
800 Test

600
600

400
400

200
200

0
0
t (min)
t (min)
[14] (b) Abecassis-Empis et al. [16]
(b) Wald et al.
Fig. 8. Heating and cooling temperature vs. time curves measured by (a) Welch
Fig. 7. Heating and cooling temperature vs. time curves measured by (a) Foster et al. [15] and (b) Abecassis-Empis et al. [16].
et al. [13] and (b) Wald et al. [14].

ISO-834 (1975) [1]. Only three temperature errors during the tem-
perature rise phase for Specimens CF06, CF26 and CF20, and two
temperature errors during the cooling phase for Specimens CF13
and CF15 in the 15 re tests did exceed the allowable values. The
difference between the measured temperature and the pre-
assigned value, except for Specimens CF13 and CF15, did not
exceed 100 C. Generally, the re tests using the hybrid heating
method met the requirements of ISO-834 (1975) [1]. Obviously,
the hybrid heating method can be used to heat the re furnace fol-
lowing the prescribed ISO standard temperature time history with
acceptable accuracy.
After the re tests on the BRE large compartment were con-
ducted, a new re safety concept based on the observed behaviour
of natural (or real) res is starting to be gradually accepted by
researchers and engineers [1316]. In this new perspective,
researchers are becoming more and more interested in the re
resistance and behaviour of structures under natural res. There-
fore, the newly-developed hybrid heating method was used to sim-
ulate some typical temperature versus time histories [1316].
Fig. 7 shows the time variation of the simulated average tem-
perature measured in the furnace, which is in good agreement with
the heating and cooling curves established by Foster et al. [13] and
Wald et al. [14]. The Hybrid heating_E method was used to sim-
ulate the temperature versus time relations of two typical natural
res. In addition to the initial 10-min heating stage, the measured
temperature versus time curves are in good agreement with the
pre-assigned heating curves. Owing to the higher gas red heating
rate during the initial heating phase, the temperature rise is clearly
faster than that of the pre-assigned heating curve. Therefore, the
electric heating method was used to heat the furnace during the
rst 10 min, while the hybrid method was used to simulate two
other typical natural res which are described in Welch et al.
[15] and Abecassis-Empis et al. [16]. Fig. 8 shows the simulated
average measured temperature in the furnace versus time curves
which are in good agreement with the heating and cooling curves
measured by Welch et al. [15] and Abecassis-Empis et al. [16]. It
can be concluded that the newly developed hybrid heating method
can be used to simulated natural res accurately and effectively.
Therefore, this method can combine the advantage of electric heat-
ing and that of gas-red heating so as to remarkably increase the
heating rate. The simple heating facility can easily and safely heat
the furnace accurately following the ISO-834 standard re curve
and simulate some real re attacks.
4. Conclusions
Based on a common electric furnace, a newly-developed built-
up re furnace equipped with liqueed gas burners was used to
heat a structural column by electric power and by gas ames
simultaneously. Fire test results indicated that the hybrid heating
 J. Huo et al. / Experimental Thermal and Fluid Science 62 (2015) 5257
method can be used to heat the test specimens closely following
the ISO-834 standard re curve, and can be used in re test with
excellent convenience and cost-effectiveness. The method com-
bines the merits of electric heating, i.e., safety and reliability, high
measurement precision, control convenience with little pollution,
with the merits of fast response and more realistic re conditions
associated with gas re heating. The hybrid heating method can
also be used to simulate other re situations with different tem-
perature time histories.
Acknowledgements
The research reported in the paper is part of Project 50778069
and 50678065 supported by the National Natural Science Founda-
tion of China, the Program for Changjiang Scholars and Innovative
Research Team Project (Project No. IRT0619) and the Program for
New Century Excellent Talents in University China (NCET-11-
0123). All these sources of nancial support are highly appreciated.
References
[1] ISO-834. Fire resistance tests-elements of building construction, ISO, Geneva,
1975.
[2] S. Spyrou, J.B. Davison, Displacement measurement in studies of steel T-stub
connections, J. Constr. Steel Res. 57 (6) (2001) 649661.
[3] Y.C. Wang, J.M. Davies, Fire tests of non-sway loaded and rotationally
restrained steel column assemblies, J. Constr. Steel Res. 59 (3) (2003) 359383.
[4] K.S. Al-Jabri, I.W. Burgess, T. Lennon, R.J. Plank, Momentrotationtemperature
curves for semi-rigid joints, J. Constr. Steel Res. 61 (2005) 281303.
57
[5] C.G. Bailey, W.S. Toh, Small-scale concrete slab tests at ambient and elevated
temperatures, Eng. Struct. 29 (10) (2007) 27752791.
[6] K.H. Tan, W.S. Toh, Z.F. Huang, G.H. Phng, Structural responses of restrained
steel columns at elevated temperatures. Part I: Experiments, Eng. Struct. 29 (8)
(2007) 16411652.
[7] L.T. Gao, Y.L. Dong, A.M. Yuan, Experimental investigation of the behaviour of
continuous slabs of unbonded prestressed concrete with the end span under
re, J. Build. Struct. 25 (2) (2004) 118123.
[8] A.M. Yuan, Y.L. Dong, H. Dai, Y.H. Li, L.T. Gao, Experimental investigation of
unbonded pre-stressed concrete continuous slab subjected to re, J. Build.
Struct. 27 (6) (2006) 6066.
[9] W.Y. Wang, G.Q. Li, Y.L. Dong, Experimental study and spring-component
modeling of extended end-plate joints in re, J. Constr. Steel Res. 63 (8) (2007)
11271137.
[10] J.S. Huo, X. Zeng, Y. Xiao, Cyclic behaviours of concrete-lled steel tubular
columns with pre-load after exposure to re, J. Constr. Steel Res. 67 (4) (2011)
727739.
[11] J.G. Zhang, Experimental research on mechanical behaviours of reinforced
concrete columns with sustained axial load after exposure to re, Master
thesis, Hunan University, Changsha, China, 2009.
[12] Chinese national standard GB/T9978-1999. Fire-resistance testselements of
building construction, The State Bureau of Quality and Technical Supervision,
Beijing, 1999.
[13] S. Foster, M. Chladna, C. Hsieh, I. Burgess, R. Plank, Thermal and structural
behaviour of a full-scale composite building subject to a severe compartment
re, Fire Saf. J. 42 (3) (2007) 183199.
[14] F. Wald, L.S.d. Silva, D. Moore, A. Santiago, Experimental behaviour of steel
joints under natural re, Connections Steel Struct. V. (2004) 393402.
[15] S. Welch, A. Jowsey, S. Deeny, R. Morgan, J.L. Torero, BRE large compartment
re testscharacterising post-ashover res for model validation, Fire Saf. J.
42 (8) (2007) 548567.
[16] C. Abecassis-Empis, P. Reszka, T. Steinhaus, A. Cowlard, H. Biteau, S. Welch, G.
Rein, J.L. Torero, Characterisation of Dalmarnock re test one, Exp. Thermal
Fluid Sci. 32 (7) (2008) 13341343.