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

net/publication/302404945

Finite Element Modelling Of Compartment Fire Using ABAQUS

Article in Jurnal Teknologi · April 2016

CITATIONS READS

0 636

1 author:

Mariyana Aida Ab Kadir

Institution Noise and Vibration, Faculty of Civil Engineering, Universiti Teknologi Malaysia
96 PUBLICATIONS 380 CITATIONS

SEE PROFILE

All content following this page was uploaded by Mariyana Aida Ab Kadir on 09 May 2016.

The user has requested enhancement of the downloaded file.

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/279819489

Finite Element Modelling Of Compartment Fire

Using ABAQUS

Article · April 2015

DOI: 10.11113/jt.v74.4606

READS

21

2 authors, including:

Mahmood Md Tahir
Universiti Teknologi Malaysia
64 PUBLICATIONS 100 CITATIONS

SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate, Available from: Mahmood Md Tahir
letting you access and read them immediately. Retrieved on: 09 May 2016
 Jurnal Full paper
Teknologi

Finite Element Modelling Of Compartment Fire Using ABAQUS

Mariyana A. A. K.a*, A. S. M. Abdul Awala, Mahmood Md. Tahirb
aDepartment of Structure and Materials, Faculty of Civil Engineering, Universiti Teknologi Malaysia, 81310, UTM Johor Bharu, Johor, Malaysia
b
UTM-Construction Research Centre, Faculty of Civil Engineering, Universiti Teknologi Malaysia, 81310, UTM Johor Bharu, Johor, Malaysia

*Corresponding author: mariyanaida@utm.my

Article history Abstract

Received: 10 November 2014 This paper presents finite element modelling (FEM) of a reinforced concrete (RC) frame subjected to
Received in revised form: elevated temperature. The work presented is part of the UK-India Education and Research Initiative
23 January 2015 (UKIERI) project. In this project, an experimental test of sub-assemblage frame with elevated
Accepted: 12 April 2015 temperature has been performed at Indian Institute of Technology (IIT) Roorkee, India. The finite
element model using ABAQUS software has been used to validate the increased in temperature
Graphical abstract distribution on reinforced concrete frame exposed to fire. The idea of this study is to design a
compartment fire, and determination of emissivity value at different height. And composition of hot gases
was calculated. Gas temperatures used was based on the average temperature-curve obtained in the fire
test. The validity of the finite element model was established by comparing the predicted values from the
FEM with test data direct from fire test results. The results obtained indicate that suggested FEM analysis
procedure is capable of modelling temperature in compartment fires.

Keywords: Finite element model; heat transfer analysis; reinforced concrete frame; compartment fire;
ABAQUS

© 2015 Penerbit UTM Press. All rights reserved.

1.0 INTRODUCTION of Cardingtion frame reveals using heat heat transfer mode is
sufficient, efficient and gain results in the short time[1].
The need to consider fire loading into structural design has Compartment fires are defined as fire in enclosed spaces.
been of great concern since the World Trade Centre disaster Theoretically the fires are often discussed in terms of five
in 11 September 2001. An investigation carried out by the growth stages-ignition, growth, flashover, fully developed fire
Building Performance Assessment Team has indicated that and decay as shows in Figure 1 [2]. This idealization may
the fire issues were vital in the collapse of the twin towers. provide useful information to understand further the
Reinforced concrete (RC) structures commonly exposed to compartment effect due to fire. Flashover is not a stage of
fire have not been taken into consideration for thermal development, but simply a rapid transition between the
analysis during fire conditions. Rather, code provisions growth and fully developed stages. Sometimes the fire may
generally are considered relevant for detailing and cover fail just before the flashover without experiencing the
requirements which provide an acceptable fire-rating in terms development into all stages. Therefore, it is important to
of the length of time that the structure can sustain its determine the stages of fire growth.
mechanical loads in the presence of fire without collapsing. There are many factors that may affect the fire growth.
Therefore, there has been a growing interest in research on the For example, type of fuel, thermal properties, size of
heat transfer analysis and design of RC structures subjected to compartment and ventilation that influence fire to develop in
fire. a compartment fire. The type of fuel, indicated as a primary
A full scale fire test is not a practical solution as it is factor, can be defined as fire growth during ignition stage.
very expensive and need a state-of-the-art furnace with an Nevertheless, as the fire moves into the growth stage, it may
appropriate capability. Therefore, a computer simulation is a be controlled by ventilation.
better option for researchers to demonstrate the heat transfer Compartment fires can be modelled to predict the
in RC structures by modeling the specimen in a proper temperature generation based on the type of fire and smoke
manner. The use of finite element modelling with high-speed movement. Klote and Milke have studied about smoke
electronic computers in heat transfer analysis began in the movement in a compartment [3]. Fire may occur at any parts
mid 1970’s and has gained wide acceptance throughout most below a ceiling in the compartment, furthermore this will
of major research works in fire engineering. Lamont et al. in releases energy and product of combustion. The hot products
their investigation of temperature distribution within steel slab of combustion form a plume which rises towards the ceiling

74:4 (2015) 33–38 | www.jurnalteknologi.utm.my | eISSN 2180–3722 |

34 Mariyana A. A. K, A. S. M Abdul Awal & Mahmood Md. Tahir / Jurnal Teknologi (Sciences & Engineering) 74:4 (2015), 33–38

due to its buoyancy. As the plume rises, it draws cool air from characteristics of concrete [5,6]. In practice, fire resistance of
within the compartment, decreasing the plume's temperature structural members used to be evaluated mainly through fire
and increasing its volume flow rate. The interchange between tests.
the hot upper layer and the air in the lower part of the This paper attempts to implement the capability of 3D
compartment assumed within the plume. As the hot layer heat transfer analysis onto a finite element modeling program
moves and reaches ventilation in the compartment walls, hot for single-storey of RC frame. The temperature-dependent
gas flows out the ventilation and outside air flow into the material properties followed EC2 are considered in this heat
ventilation. Figure 2 illustrate the compartment fire behaviour transfer analysis [7].Taking advantage of zone model, mainly
also known as a two layers or zone model which has been based on gas composition of each layers in compartment,
developed by Klote and Milke in 2002 [3]. In this model the transferred between a fires, the fire gases, the fuel bed and the
compositions of the layers are assumed uniform and the surfaces in an enclosure can be estimated [8]. While,
temperature of the upper layer remain greater than the lower temperature in the finite element modelling is used based on
layer. the data from the UKIERI fire test [9, 10].
The following discussion will provide descriptions on
calculation of emissivity values, finite element modelling
procedure, and temperature distribution within elements of the
UKIERI frame.

2.0 EMISSIVITY VALUE AND SOOT HEIGTH

Heat transfer analysis by means of radiation and convection

mode is considered in this study. The convective mechanism
in this study used a convection heat transfer coefficient of 25
kW/m2. This is a typical value used in structural fire analysis
[6]. Furthermore, Yung Lee et al. [11] have shown the
convective components as coefficient in heat transfer
modelling can be regarded as the crucial material property of
concrete with respect to the prediction of behaviour of
Figure 1 Compartment fires growth stages [2] thermal cracking [12]. The model
e layer radiation mode is affected by the
soot height in the compartment fire. In real gas properties,
combustion gases are complex in term of radiative properties
due to the presence of gas, H20 and C02, which radiate
discretely over wavelength bands that can overlap while soot
radiates continuously over a wide range of wavelengths.
When heat transfer in radiation mode is analysed using
ABAQUS the emissivity, must be defined which simulates
the behavior of soot in the compartment. In reality
combustion gases are complex in terms of their radiative
properties, however, due to the presence of gases such as H2O
and CO2, which radiate discretely over wavelength bands can
overlap and soot radiates continuously over a wide range of
wavelengths. In a room smoke layer, the products of H20 and
C02 could contribute about g ~ 0.3 and s range from 0 to
0.7 taking up to 1.0. The total depends on the path
length, if long enough it can take up to 1.0 [13]. Combining
Figure 2 Schematic of the zone model [3]
the soot, s and the gas, g together has contributed to the
flame and smoke radiation, as shows in Equation 1.
Heat transfer analysis is a process of energy transmitted
due to a temperature difference. The transmission of the [1]
energy happened in two mechanisms known as conduction H 2O CO2 correction
and radiation. The conduction is a molecular energy transport
which can be dominating mechanism for the heating up of In this analysis the emissivity is derived in accordance to
small devices such as detectors. The radiation mode is very Babraukas 2 as shown in Equation 2.
often dominated for transferring heat in enclosure fire in the
case of higher temperatures of flames and fire gases. D
(1 e ) [2]
Babraukas showed in his study that pool fire with diameter is
1.0 m or bigger the radiation mode effect is more important The equation consists of an extinction coefficient, κ, a
than convection [4]. This mode of heat transfer is also ‘mean beam length corrector', β and D is the diameter of the
important for a target which is located laterally of the pool fire. Particularly κβ is an empical constant (m-1) in an
exposure fire source. area of fuel where effectively heats of combustible are known.
Venkatesh Kodur in his studied, explained the concrete Basically the emissivity value is defined depending on the
is best fire resistance properties compared to any building fuel type used in a pool of fire and for this research the oil
material due to unique characteristic of concrete which, when type is kerosene.
chemically combined from hydration proses it form a material Based on Equation 2 where kerosene parameter was
that has low thermal conductivity, high heat capacity, and given by Blinov [14] and Kutcha et al. [15], which represent
slower strength degradation subjected to elevate temperature the emissivity in the whole compartment (highest point of
15. However, the fire resistance rates of concrete vary as a smoke level) was taken as 0.9829 (AHT3). Figure 3 illustrates
function of temperature and depend on the composition and the behaviour of smoke of well-mixed case in an enclosure
35 Mariyana A. A. K, A. S. M Abdul Awal & Mahmood Md. Tahir / Jurnal Teknologi (Sciences & Engineering) 74:4 (2015), 33–38

with uniform temperature, Tg, which is higher than the

outside temperature, Ta. In this figure, ρa = ambient air
density (kg/m3) and ρg = gas density in enclosure (kg/m3), H
is the height of the window opening with 1meter height, Ho is
the heat of combustion of the fuel (MJ/kg). From the figure,
the height of smoke and air flow were then estimated using
Equation 3 to determine the stages of fire started with the first
stage; H1= HN with the reference point represent the bottom of
the opening.

Figure 4 3D continuum brick element to model the frame using

ABAQUS 6.8

Figure 3 The well-mixed case: An enclosure with uniform The heat was introduced into the element uniformly
temperature, Tg, which is higher than the outside temperature, Ta since the flume is burned in the middle of the
compartment.The compartment of the frame in this study is
H0 divided into three areas. The areas at different heights in the
h1 compartment are referred to as AHT1, AHT2 and AHT3, as
1/ 3
[3] shown in Figure 5, where AHT1 is the area located near to the
a
1 opening, AHT3 is the area located at the top of the
g compartment and AHT2 is the area considered between
AHT1 and AHT3. From Figure 5, AHT3 is assumed to be
The smoke layer in the compartment was calculated filled with smoke with estimation emissivity value to be
based on Equation 4 by Klote and Milke [3]. 0.9829 at the highest smoke level for kerosene oil. Hence by
considering AHT3 in this study, the area of AHT2 is then
being estimated by assuming that the smoke only in its upper
353
T [4] region (until the opening) and area AHT1 is assumed to
contain little smoke. The emissivity value varies with
Based on the frame configuration used in this study, different smoke layers as discussed earlier. Table 1 shows the
values of Tg was 1400oC and Ta was 20oC. Therefore, emissivity values applied in the heat transfer analysis.
according to Equation 4, the HN = h1 = 0.357m was obtained.

3.0 FINITE ELEMENT MODELLING OF HEAT

TRANSFER

The heat transfer analysis was performed using ABAQUS

version 6.8 to get the temperature distribution within the
elements of the RC frame. 3D continuum elements (DC3D8)
defined as 8-noded linear (hexahedral) elements were used to
model the UKIERI frame [16]. Temperature as a single
degree of freedom at each node was defined in the modelling
o recorded the temperature behaviour in the elements. The
total nodes in the model is 4452 with 84 nodes at concrete
columns, 672 nodes at the reinforced columns, slab contains
of 3298 nodes, reinforced beam with 270 nodes and beams
with 128 nodes respectively. Figure 4 shows the configuration Figure 5 Compartment fire modelled in ABAQUS 6.8
of the 3D frame modelled for heat transfer analysis in
ABAQUS. Table 1 Parameter of emissivity for different areas used in heat
Material properties of concrete at elevated temperature transfer analysis using ABAQUS
used was in accordance to Eurocode 2 [7]. In typical concrete
it is known as non-homogeneous, anisotropic medium
composed of aggregate, cement paste and water. For Area in the compartment at AHT1 AHT2 AHT3
simplicity, concrete can be treated as a homogeneous different height
isotropic material in heat transfer analysis. The temperature
curves based on the fire test were applied at different height in Emissivity applied at each area , 0.3 ( 0 ) 0.6415 0.9829
elongation and area in plan. s
g s
36 Mariyana A. A. K, A. S. M Abdul Awal & Mahmood Md. Tahir / Jurnal Teknologi (Sciences & Engineering) 74:4 (2015), 33–38

4.0 RESULTS AND DISCUSSION The temperature distribution is recorded by thermocouples

embedded at 5 points within the elements. The 5 points are
The idea of this study is to design a compartment fire, and measured as 5 mm, 25 mm, 115 mm, 205 mm and 225 mm
determination of emissivity value at different height. The from the base of the beam whiles the dotted lines represent
temperature distribution within the frame from numerical test data and solid line for numerical analysis as shown in
analysis and test data on real temperature using following graph. At lower floor, the temperature distributions
thermocouples and data logger is then compared. The though beam B1 provides the most detailed results obtained
temperature distribution of the RC frame after 1 hour of fire is from the test; therefore, a comparison is made between the
illustrated in Figure 6. Figure 7 shows the area where each results of the test and the numerical analysis in this beam
time-temperature curve relates to, with the notation of the (Figure 8). From the graphs, the temperature distribution
beam and column given for further reference. Due to failure through beam B1 at mid span exhibit good agreement in both
of thermocouples at elevated temperature, most of the the numerical analyses and the test up to 50 minutes.
temperature distributions data within elements are not
presented. Therefore, only temperature distribution on B1, 1000
B5, C4 and slab is compared with the test data. 900
800

Temperature, T (ͦC)
700
600
500
400
300
200
100
0
0 10 20 30 40 50 60
Time (min)
5mm a 25mm a 115mm a 205mm a
225mm a 5mm b 25mm b 115mm b
205mm b 225mm b

Figure 8 Temperature distributions in the beam, B1 at the middle

side along the height (a) Test data (b) Heat Transfer using ABAQUS
Figure 6 Temperature distribution of the frame from the heat transfer
analysis
At the top floor, B5 is compared with the test data as
shown in Figure 9. From the observation, the temperature
distribution in the test slowly increased while in the heat
transfer analysis increased about 50% faster than that obtained
in the test. The temperature distribution as it can be seen there
again very good aggrement between predicted and test data
for the first 20 minutes.

1000
900
800
Temperature,T (°C)

700
600
500
400
300
200
100
0
0 10 20 30 40 50 60
Time (min)
5mm a 25mm a 115mm a 205mm a
225mm a 5mm b 25mm b 115mm b
205mm b 225mm b

(a)

Figure 9 Temperature distributions in the top beam, B5 at the middle

side along the height (a) Test data (b) Heat Transfer using ABAQUS

The temperature gradient in the columns obtained from

the test shows a gradual increase as the time of exposure
increased. Then, in the last 10 minutes, most of the points
showed a decrease in temperature gradient (Figure 10). This
Figure 7 Notation of beams, columns and slab in a compartment (a)
top floor (b) lower floor (Plan view)
may be due to the fuel running out at the end of the test. It is
37 Mariyana A. A. K, A. S. M Abdul Awal & Mahmood Md. Tahir / Jurnal Teknologi (Sciences & Engineering) 74:4 (2015), 33–38

interesting to note that the temperature gradient in the column fire test are higher as compared to the emissivity value
from the heat transfer analysis rises quickly from the calculated in the numerical analysis. The difference in results
beginning of the analysis until a time of 30 minutes. Then the between finite element modelling and test may be due to the
temperatures remain steady until the end of 1 hour. presence of thick smoke layer in the test as the emissivity
value used in the heat transfer analysis is based on the effect of
1000 the smoke layer in the compartment. In other hand, moisture
. 900 content of concrete in numerical modelling not properly
800 defined caused the high temperature distribution in the
Temperature,T (°C)

700 numerical modelling of the frame elements.

600
500
Acknowledgement
400
300 The authors wish to acknowledge the financial support
200 provided by Ministry of Higher Education Malaysia and
100 Universiti Teknologi Malaysia. The authors also express their
0 thanks to all researchers from the Department of Civil
Engineering, IIT Roorkee, India for providing the tests data for
0 10 20 30 40 50 60
full-scale UKEIRI frame.
Time (min)
5mm a 40mm a 150mm a 260mm a
Nomenclature
295mm a 5mm b 40mm b 150mm b
260mm b 295mm b
Where:
(a) density of gas in enclosure
g
Figure 10 Temperature distributions in the column, C4 at the middle
along the height (a) Test data (b) Heat Transfer using ABAQUS density of gas at ambient
a
Tg temperature in enclosure
Figure 11 reveals the temperature distribution through
the slab SLA obtained from both the test and the heat transfer Ta ambient temperature
analysis. Increment pattern of the temperature distribution .
within the slab agree well between the test and the numerical ma mass of the gas at ambient temperature
analysis at the end of test. However, as it can be seen the
.
temperatures obtained from the test are increased rapidly than
those taken from the heat transfer analysis, after 20 minutes of mg mass of the gas in enclosure
heating. H0 opening of enclosure

1000 HN reference height of the gas

900 hu height of outgoing gas
800
Temperature,T (°C)

700 h1 height of incoming gas

600 T temperature in Kelvin [K]
500 density in [kg/m3]
400 correction correction factor due to overlap wave length
300
200
100
References
0
0 10 20 30 40 50 60 [1] Lamont, S., Usmani, A. S., Drysdale, D. D. 2001. Heat Transfer
Time (min) Analysis of the Composite Slab in the Cardington Frame Fire Tests.
5mm a 30mm a 60mm a 90mm a
Fire Safety Journal. 36(8): 815–839.
115mm a 5mm b 30mm b 60mm b
90mm b 115mm b [2] Drysdale, D. 1998. An Introduction to Fire Dynamic. Second
Edition. John Wiley & Sons, Chichester, UK.
(a) [3] Klote, J. H., Milke, J. A. 2002. Principles of Smoke Movement.
American Society of Heating, Refrigerating and Air-Conditioning
Figure 11 Temperature distributions in the slab (SLA) (a) Test data Engineers, Inc., Atlanta, GA.
(b) Heat Transfer using ABAQUS [4] Babrauskas, V. 1983. Estimating Large Pool Fire Burning Rates.
Fire Technology.
[5] Kodur, V. R. 2014. Properties of Concrete at Elevated Temperature.
ISRN Civil Engineering. Article ID 468510.
5.0 CONCLUSIONS [6] Kodur, V. R. and Harmathy, T. Z. 2008. Properties of Building
Materials, SFPE Handbook of Fire Protection Engineering, P. J.
In this work, a compartment fire is developed and DiNenno, Ed. National Fire Protection Association, Quincy, Mass,
implemented into a 3D finite element analysis procedure for USA.
reinforced concrete frame. The gas composition within the [7] EN1992-1-2: Eurocode 2. 2004. Design of Concrete Structures-Part
1-2: General Rules-Structural Fire Design.
compartment is calculated based on the emissivity values [8] Robert, P., Magdalena, P. 2012. Pool Boiling on Surfaces with
provided from soot height due to heat transfer analysis using Mini-Fins and Micro-Cavities. Proceedings of 6th European Thermal
general purposed analysis programs, ABAQUS. From the Sciences Conference (Eurotherm).
results, the temperatures from the heat transfer analyses are [9] Sharma, U. K., Bhargava, P., Singh, B., Singh, Y., Kumar, V.,
generally higher than that of corresponding temperatures from Kamath, P., Usmani, A., Torero, J., Gillie, M., Pankaj, P., May, I. &
Zhang, J. 2012. Full Scale Testing of a Damaged RC Frame in Fire.
the test data. This is due to the fact that hot gases properties in Structures and Buildings.Proceedings of the ICE.
 38 Mariyana A. A. K, A. S. M Abdul Awal & Mahmood Md. Tahir / Jurnal Teknologi (Sciences & Engineering) 74:4 (2015), 33–38

[10] Praveen, K., Sharma, U. K., Bhargava, Kumar, V., P., Singh, B., [13] Hottle, H. C. 1954. Heat Transmission. MacGraw-Hill, New York.
Singh, Y., Kamath, P., Usmani, A., Torero, J., Gillie, M., Pankaj, P., [14] Blinov, V. I., and Khl, G. N. 1961. Diffusion Burning of Liquids.
2015. Full Scale Fire Test of an Earthquake-Damaged Reinforced U.S. Army Translation, NTIS No. AD296762.
Concrete Frame. Fire safety Journal. 73: 1–19. [15] Kuchta, J. M., Furno, A. L., Martindill, G. H. Imhof, A. C. 1969.
[11] Lee, Y., Choi, M. S., Yi, S.T., Kim, J. K. 2009. Experimental Study Crash Fire Hazard Rating System for Controlled Flammability
on the Convective Heat Transfer Coefficient of Early-age Concrete. Fuels. Report NA-69-17, Bureau of Mines, Pittsburgh, NTIS No.
Cement and Concrete Composites. 31(1): 60–71. AD684089.
[12] Tien, C. L., Lee, K. Y., and Stretton, A. J. 1995. Radiation Heat [16] ABAQUS. 2008. Abaqus Analysis User's Manual. Providence,
Transfer, SFPE Handbook of Fire Protection Engineering. National Dassault Systemes Simulia Corp.
Fire Protection Association, Quincy, MA.

View publication stats