Full-Scale Burning Tests For Retail Shop Fires
Full-Scale Burning Tests For Retail Shop Fires
Full-Scale Burning Tests For Retail Shop Fires
94-121, 2004
ABSTRACT
High combustible contents are stored in small retail shops including those located in crowded shopping malls.
As the number of non-accidental fires appeared to be increasing, the general public is now quite concerned
about the hidden fire hazard. For assessing the consequences of a fire, the possible heat release rates should be
studied experimentally. Not only that, conducting full-scale burning tests for local retail shops under flashover
conditions is necessary and had been carried out recently.
In this paper, the resultant heat release rates by burning those combustibles in a shop fire before and after
flashover will be studied. Experimental data on heat release rate, oxygen concentration, temperature and
radiative heat flux are presented. Whether the heat release rate can be reduced by discharging water mists will
also be investigated.
94
International Journal on Engineering Performance-Based Fire Codes
this will not give the contribution of materials, nor Flame spreading of building materials used would
their assemblies, to a fire under flashover condition. be controlled. Fibreglass composites and
The HRR measured would not be too high as only a flammable aerosols are usually not allowed.
small amount of the combustibles were ignited.
And for most cases, fire suppression systems might The likelihood of flashover in the shop had been
be operated to reduce the resultant HRR. The tests pointed out with points of concern identified:
then become a demonstration on how the system
would act at the tested fire, but not for y Consequences of flashover [e.g. 16] in a shop
understanding the actual HRR and the possibility of should be watched. The HRR of
igniting the combustibles under flashover condition. combustibles stored in the shop are key
The situation should be reviewed as more fires factors [17]. Storing products of HRR higher
other than due to accidents were reported. than the value giving flashover under a certain
ventilation condition should be watched
Effect of high thermal radiation heat flux (such as carefully.
20 kWm-2 for flashover) on combustibles in a shop
y Effectiveness of operating sprinkler system in
should be included in studying the possible HRR.
controlling a fire should be watched and water
Missing this point might be quite serious for retail
mist system [e.g. 18] might be used if
shops. Combustible items stored there such as
necessary.
plastic dolls, especially those without quality
control through standard fire tests, might be ignited y Fire safety management schemes [e.g. 19]
easily by the incident heat flux if there is an including training of the storekeepers and the
accidental fire. Note that flashover would occur building management staff are to be worked
easily in such small retail shops as raised before [2]. out carefully. For example, keeping a fire to
Upon ignition of those combustibles, much larger be smaller than the design fires for sizing the
quantity of heat and smoke would be liberated and sprinkler and smoke extraction systems
spread to the hall space outside the shop. What will should be ensured.
happen if the big hall is overcrowded?
All points should be analyzed carefully to provide
Better understanding on the probable HRR under total fire safety. Performing full-scale burning tests
flashover is strongly recommended [14]. on some selected shops is necessary to observe key
Preliminary burning tests of several fire scenarios fire issues. Performance of fire safety provisions
in small retail shops under flashover condition were such as water mist fire suppression system on those
performed in a new full-scale burning facility, the shops in big halls [14,20] can also be demonstrated.
Chinese Assembly Calorimeter [15]. That facility Consequences of flashover fires for some
was just developed as a collaboration project dangerous arrangements of concern must be studied.
between the Harbin Engineering University (HEU). It is difficult to work out fire safety regulations and
Results on HRR in flashover shop fires will be recommend good fire safety management without
reported in this paper. in-depth studies on fire dynamics.
As reviewed before for local small retail areas [1,2], A set of preliminary full-scale burning tests were
combustibles include plastic toys in blister packs carried out to study the HRR resulted from burning
stored in vertical stacks; paper boxes and hanging combustible items in small retail shops. These were
arms; soft toys, pillows on wire display baskets; measured in the Chinese Assembly Calorimeter
video cassettes and CD displays; sweets and [14,15] located in Lanxi, a remote area in Harbin,
candies with plastic bags; and sports wares, jackets Heilongjiang, China as shown in the first paper of
on plastic hangers with plastic bags for keeping out this issue. An exhaust hood with a fan-duct system
the dust. was developed for measuring HRR using the
oxygen consumption method [e.g. 5]. The tests
But for those small retail shops located in terminal were carried out in a room of length 3.6 m, width
halls or public transport interchanges, only selected 2.4 m and height 2.4 m with a door of height 2 m
combustibles are allowed to store. Examples are and width 0.8 m as shown in Fig. 1.
newspapers and magazines; cigarettes and tobacco;
alcohols up to 75 l; shoes, not selling, but repairing Six fire scenarios labelled as B1 to B6 were tested
and polishing; clothes; furniture including for assessing boutiques. Goods with an iron clothes
polyurethane foam sofas or cushions; coffee tables hanger with short sleeves T-shirts made of 100%
with wood or other timber products and chairs. cotton were taken as combustibles. Three fire
95
International Journal on Engineering Performance-Based Fire Codes
scenarios C1 to C3 were tested for retailing shops vertically on each iron rack and therefore 96 T-
selling CDs. Those scenarios displaying CDs are shirts would be hung on the iron hanger. The T-
arranged in the test room and under pre-flashover shirts were of the same size and each T-shirt
and post-flashover condition. Flashover was set off was covered by a transparent plastic bag.
by burning sufficient amount of gasoline. That is
similar to testing materials by a cone calorimeter The layout of clothes is shown in Fig. 3a. The
[e.g. 5] where the samples are exposed under a clothes on rack 1, which were the farthest away
radiative heat flux of values varying about 20 from the door, would be ignited initially. There
kWm-2, while in pre-flashover condition, fire would were a total of 22 thermocouples utilized in this
only be started manually by an igniter. experiment. Thermocouples T1 to T6, M1 to
M6, C1 to C6 and D1 to D4 were placed at
ceiling, near the left wall, near the door and in
4. TESTS FOR BOUTIQUE SELLING the door side as in Fig. 3b.
CLOTHES
The air temperature curves are shown in Fig. 4,
• Test B1: The clothes were ignited by an igniter while the graphs of heat flux, HRR and oxygen
consumption are shown in Fig. 5. It is observed
The diagram of an iron clothes hanger of length that the peak HRR can go up to 1.2 MW. The
2.1 m, width 0.7 m and height 1.2 m is shown values are higher than 1 MW within a burning
in Fig. 2. The hanger was put at the centre of period of 40 s. Air temperatures are up to
the room. There are four iron racks on the 830°C as shown in Fig. 4a.
clothes hanger. Twenty-four T-shirts were hung
Flow, smoke density, and gas
concentrations measurement
2.4 m
Exhaust
m 1m
2.4
Gas collection
2.4 m hood
2m
m
3.6 m 0.8
2.1 m
Rack 1
0.6 m 0.9 m 0.6 m
0.7 m
Rack 3
Rack 2
Rack 4
1.2 m
96
International Journal on Engineering Performance-Based Fire Codes
0.9 m
C1
T5
C2
m
T2 T4 T6
6
0.
C3
0.8 m T3 C4
0.25 m
C5
m m
2.1 m
0.1 m
m
27 3
T1
0. 0.0
M1
4
0.
1.72 m
0.3 m
m
C1
1.57 m
25
M2 D1 C6
m
0.
1.42 m
3
0.
1.27 m
0.4 m C2 D2
m
0.97 m
3
M3
0.
C3 D3
0.4 m
C4 D4
M4
C5
0.4 m
M5
C6
0.4 m
0.3 m
M6 HeatHeat
Flux flux meter
Heat Flux
m
3
0.
Thermocouple
Thermocouples
0.3 m
• Test B2: The clothes were ignited by a gasoline The air temperature curves are shown in Fig. 6,
pool fire while the heat flux, HRR and oxygen
consumption curves are shown in Fig. 7. Note
The set up of the experiment was the same as in that the maximum HRR of both the clothes and
Test B1, except the clothes were ignited by a gasoline, and gasoline only is up to 2.4 MW
pool fire. Gasoline of 4000 ml was placed in a and 0.47 MW, which is much higher than Test
pool of diameter 0.5 m and located at the centre B1. The values of HRR are above 1.5 MW
of room to get flashover. Locations of within a long period of over 170 s of burning
thermocouples were the same as in Test B1. time, while the maximum air temperatures
The T-shirts were ignited spontaneously by the would be up to 880 °C as illustrated in Fig. 6a.
heat flux due to the flashover fire.
97
International Journal on Engineering Performance-Based Fire Codes
98
International Journal on Engineering Performance-Based Fire Codes
1.8
1.6
1.4
0.8
0.6
0.4
0.2
0
0 500 1000 1500 2000
Time / s
(a) Heat flux
1400
1200
Heat release rate / kW
1000
800
600
400
200
0
0 100 200 300 400 500 600 700
Time / s
(b) HRR
21.5
21
Oxygen concentration / %
20.5
20
19.5
19
0 100 200 300 400 500 600 700
Time / s
(c) Oxygen level
99
International Journal on Engineering Performance-Based Fire Codes
100
International Journal on Engineering Performance-Based Fire Codes
0
0 500 1000 1500 2000
Time / s
(a) Heat flux
3000
C: Clothes
2500 C G: Gasoline only
Heat release rate / kW
2000
1500
1000
G
500
0
0 200 400 600 800 1000 1200
Time / s
(b) HRR
22
21
Oxygen concentration / %
20
19
18
17
16
15
0 200 400 600 800 1000 1200
Time / s
(c) Oxygen level
101
International Journal on Engineering Performance-Based Fire Codes
• Test B3: The room was installed with Water be installed in the fire test room. Sprinkler head
Mist Fire Suppression System (WMFSS) without the sensing element would be
connected to the pipe sections, water pump and
The set up of the experiment was the same as in water tank. Pressure and flow rate of the system
Test B2. A pool fire of 0.5 m diameter with were set at 11 bars and 2.9 ls-1 respectively. The
4000 ml gasoline was used to set off flashover. sprinkler was installed near the centre of ceiling.
A WMFSS was installed in the fire test room. Locations of thermocouples were the same as
The system comprised a water mist nozzle, in Test B1. Although the activation temperature
water tank, pipes and other components. of the sprinkler head was set at 68oC, it was not
Pressure and flow rate of the system were set at the real actuation temperature. Therefore, a
15 bars and 0.24 ls-1 respectively. The nozzle hypothesis would be taken such that the
was installed near the centre of ceiling. sprinkler system was activated manually with a
Locations of thermocouples were the same as delay of 30 s after the thermocouple T4 reached
in Test B1. The T-shirts were ignited 68 oC. This activation temperature of T4 would
spontaneously by the heat flux due to the then be compared with the actuation
flashover fire. WMFSS would be activated temperature acquired in Test B3.
manually once the clothes were ignited.
The results of air temperatures and heat flux,
During the experiment, a sprinkler head HRR, oxygen consumption are shown in Figs.
without connecting to pipes and water tank 10 and 11 respectively. As the clothes were
would also be installed in ceiling, where the only ignited by an igniter and the sprinkler
position was totally above the pool fire, so as to system would be manually actuated at 125 s, i.e.
find out the operating temperature of the 30 s later after the thermocouple T4 reached 68
o
sprinkler. A thermocouple was attached to the C, the maximum HRR and maximum air
sensing element of the sprinkler head for temperatures are only up to 0.8 MW and 360
detecting the accurate breaking temperature. °C respectively.
This breaking temperature was exactly the
actual actuation temperature of the sprinkler. The predicted delay time of a sprinkler head in
As the sprinkler system could not be the hypothesis is too long. The sprinkler system
automatically operated in Test B4, it would be is manually actuated when T4 reached 164 oC,
manually controlled. The activation which is much higher than the actual activation
temperature of the sprinkler obtained in this temperature of 90oC as indicated in Test B3.
experiment would then be compared with the Actually, the sprinkler head is only required 8 s
actuation temperature hypothesized in Test B4, to go up to 90oC after T4 reached 68 oC.
so as to verify the accuracy of the assumption. Therefore, the delay time should be reduced to
8 s to match the actuation temperature, so as to
Transient results on air temperatures and heat decrease the values of HRR and air
flux, HRR, oxygen consumption are shown in temperatures as measured in Test B4.
Figs. 8 and 9. As WMFSS was activated
manually once the clothes were ignited, i.e. at • Test B5: The clothes were ignited under the
about 100s, the maximum HRR of the clothes hood
is only up to 1.6 MW. The values of HRR are
above 1 MW over 100 s of burning time, while In order to realize the HRR under free burning
the maximum air temperature is only up to 720 condition, the clothes were ignited under the
o
C as illustrated by the thermocouple T2. hood by an igniter. As it is needed to prevent
the fire become too vigorous, only eight T-
According to Test B3, the breaking temperature shirts would be ignited. The results of the HRR
of the sensing element in a sprinkler head was would be times a factor of twelve, so as to
90 oC. It will be used as a reference in Test B4 predict the value of HRR when 96 T-shirts
for finding the actual delay time of the were ignited.
activation of a sprinkler head once the
temperature of thermocouple T4 reached 68oC. The hanger was put at the centre of the hood.
The photo of hood was shown in Fig. 12. Eight
• Test B4: The room was installed with sprinkler T-shirts were hung vertically together on rack 1.
system Each T-shirt was covered by a transparent
plastic bag. The clothes, which were hung in
The set up of the experiment was the same as in the middle of rack 1, would be ignited initially.
Test B1. The clothes in rack 1, which were the As only HRR was concentrated in this
farthest away from the door, would be ignited experiment, temperatures and heat flux would
initially by an igniter. Sprinkler system would not be measured such that thermocouples and
heat flux meter were not set.
102
International Journal on Engineering Performance-Based Fire Codes
103
International Journal on Engineering Performance-Based Fire Codes
1.2
0.6
0.4
0.2
0
0 500 1000 1500 2000
Time / s
(a) Heat flux
1800
C C: Clothes
1600
G: Gasoline only
1400
Heat release rate / kW
1200
1000
800
600
G
400
200
0
0 100 200 300 400 500 600 700
Time / s
(b) HRR
21
Oxygen concentration / %
20.5
20
19.5
19
18.5
18
0 100 200 300 400 500 600 700
Time / s
(c) Oxygen level
104
International Journal on Engineering Performance-Based Fire Codes
105
International Journal on Engineering Performance-Based Fire Codes
0.16
0.14
0.12
Heat flux / kWm-2
0.1
0.08
0.06
0.04
0.02
0
0 200 400 600 800 1000
Time / s
(a) Heat flux
90
80
70
Heat release rate / kW
60
50
40
30
20
10
0
0 100 200 300 400 500 600
Time / s
(b) HRR
21
20.98
Oxygen concentration / %
20.96
20.94
20.92
20.9
20.88
20.86
20.84
20.82
20.8
0 100 200 300 400 500 600
Time / s
(c) Oxygen level
106
International Journal on Engineering Performance-Based Fire Codes
The HRR and oxygen consumption curves are predicted maximum HRR would be up to 4.6
shown in Fig. 13. Although the maximum HRR MW, while the values are higher than 2 MW
is only up to 0.4 MW in free burning condition, within a burning period of 70 s.
it is only for eight clothes. For 96 clothes, the
5000
C2
C1: HRR of 8 clothes
4500
C2: The predicted HRR
4000 of 96 clothes
Heat release rate / kW
3500
3000
2500
2000
1500
1000
500 C1
0
0 100 200 300 400 500
Time / s
(a) HRR
21
20.95
Oxygen concentration / %
20.9
20.85
20.8
20.75
20.7
20.65
20.6
20.55
20.5
0 100 200 300 400 500
Time / s
(b) Oxygen level
107
International Journal on Engineering Performance-Based Fire Codes
• Test B6: The clothes were ignited without 5. TESTS FOR CD SHOPS
covered by transparent plastic bags
• Test C1: The room was installed with WMFSS
The set up of the experiment was similar to
Test B2, except the clothes were ignited CD displays of width 1.5 m and height 1.6 m
without covered by the transparent plastic bags. were put at the left and back walls of the room.
Gasoline of 4000 ml was placed in a pool of The diagram of display and layout of CDs are
diameter 0.5 m and located in the centre of shown in Fig. 17a and 17b respectively. There
room to get flashover. The T-shirts were ignited were ten rows in each display. Each row was
spontaneously by the heat flux due to the put in an array of 8 x 3 CDs and 240 CDs were
flashover fire. put in each display. Therefore, 480 CDs were
utilized in the experiment. Gasoline of 4000 ml
The layout of thermocouples are shown in Fig. was placed in a pool of diameter 1 m to get
14. The positions of thermocouples M1 to M6 flashover. Twenty-two thermocouples were
were changed to the centre of room. For other utilized in this experiment. Thermocouples T1
thermocouples, the locations were the same as to T6, M1 to M6, C1 to C6 and D1 to D4 were
in Test B1. placed at ceiling, centre of room, near the door
and in the door side as in Fig. 14, which were
The air temperature curves and heat flux, HRR the same positions as those in Test B6. The
and oxygen consumption curves are shown in CDs were ignited by the thermal radiative heat
Figs. 15 and 16 respectively. Note that the flux.
maximum HRR is up to 1.9 MW, which is
much lower than the value in Test B2, i.e. 2.4 WMFSS was installed in the fire test room and
MW. The values of HRR above 1.5 MW would it was operated manually. Pressure and flow
be within a burning period of 120 s. However, rate of the system were set at 15 bars and 0.24
the maximum air temperatures of the ls-1 respectively. A water mist nozzle was
experiment would be up to 915 oC as measured installed near the centre of ceiling. In order to
by thermocouple M3, which is higher than the prevent the damage of the exhaust duct, water
maximum temperature in Test B2. mist would be discharged when the
temperatures of exhaust duct and ceiling were
larger than 200oC and 1000oC. It would be
activated and turned off periodically during the
experiment, so as to evaluate the performance
of WMFSS in controlling the fire.
0.9 m
C1
T5
C2
m
T2 T4 T6
6
0.
C3
0.1 m
T3 M1 C4
0.25 m 0.3 m
C5
m m
M2
2.1 m
27 3
T1
0. 0.0
1.72 m
0.4 m
m
C1
1.57 m
25
M3 D1 C6
m
0.
1.42 m
3
0.
1.27 m
0.4 m C2 D2
m
0.97 m
3
0.
M4 C3 D3
0.4 m D4
C4
M5
C5
0.4 m
M6
C6
0.3 m
Heat
Heat Fluxflux meter
Heat Flux
m
Thermocouples
3
0.
Thermocouple
0.3 m
Thermocouple trees
108
International Journal on Engineering Performance-Based Fire Codes
109
International Journal on Engineering Performance-Based Fire Codes
3.5
1.5
0.5
0
0 200 400 600 800 1000 1200 1400
Time / s
(a) Heat flux
2000
C
1800 C: Clothes
G: Gasoline only
Heat release rate / kW
1600
1400
1200
1000
800
600 G
400
200
0
0 100 200 300 400 500 600 700
Time / s
(b) HRR
21.5
21
Oxygen concentration / %
20.5
20
19.5
19
18.5
18
17.5
0 100 200 300 400 500 600 700
Time / s
(c) Oxygen level
110
International Journal on Engineering Performance-Based Fire Codes
1.5 m
0.05 m
1.6 m
The air temperature curves are shown in Fig. 18, ceiling was larger than 1000 oC. After a short
while the heat flux, HRR and oxygen period of about 70 s, the air temperatures
consumption curves are shown in Fig. 19. It is decreased tremendously. For example, the
observed that the peak HRR for both the CDs temperatures of thermocouples M1 to M6,
and gasoline, and gasoline only can go up to 4 which were located in the centre of room,
MW at 290 s and 1.1 MW at 180 s respectively. decreased from about 1000 oC to around 60 oC.
The values of HRR are higher than 2 MW over The HRR also reduced to lower values.
100 s of burning time. The maximum air
temperatures are also increased to 1130°C as After turning off the water supply, the
measured by thermocouple M2, which is much temperatures and HRR increased again.
higher than the highest temperatures measured However, such increments are not so large as
in the experiments for clothes. shown by the air temperatures and HRR curves
in Figs. 18 and 19b, i.e. only about 130 oC and
Water mist was discharged manually at 170 s 0.18 MW.
after ignition when the temperature of the
111
International Journal on Engineering Performance-Based Fire Codes
112
International Journal on Engineering Performance-Based Fire Codes
4.5
3.5
2.5
1.5
0.5
0
0 500 1000 1500 2000
Time / s
(a) Heat flux
4500
C: CDs
4000 CD G: Gasoline only
Heat release rate / kW
3500
3000
2500
2000
1500
G
1000
500
0
0 200 400 600 800 1000 1200 1400
Time / s
(b) HRR
22
21
Oxygen concentration / %
20
19
18
17
16
15
0 200 400 600 800 1000 1200 1400
Time / s
(c) Oxygen level
113
International Journal on Engineering Performance-Based Fire Codes
114
International Journal on Engineering Performance-Based Fire Codes
115
International Journal on Engineering Performance-Based Fire Codes
1.4
1.2
0.8
0.6
0.4
0.2
0
0 200 400 600 800 1000
Time / s
(a) Heat flux
1200
G C: CDs
G: Gasoline only
1000
Heat release rate / kW
800
CD
600
400
200
0
0 100 200 300 400 500
Time / s
(b) HRR
21
20.9
Oxygen concentration / %
20.8
20.7
20.6
20.5
20.4
20.3
20.2
0 50 100 150 200 250 300 350 400
Time / s
(c) Oxygen level
116
International Journal on Engineering Performance-Based Fire Codes
117
International Journal on Engineering Performance-Based Fire Codes
1.8
1.6
1.4
0.8
0.6
0.4
0.2
0
0 100 200 300 400 500 600 700 800
Time / s
(a) Heat flux
1200
G C: CDs
1000 G: Gasoline only
Heat release rate / kW
800 CD
600
400
200
0
0 100 200 300 400 500 600
Time / s
(b) HRR
21
20.9
Oxygen concentration / %
20.8
20.7
20.6
20.5
20.4
20.3
20.2
20.1
20
0 100 200 300 400 500 600
Time / s
(c) Oxygen level
118
International Journal on Engineering Performance-Based Fire Codes
y For the clothes which were ignited by an covered by the CD boxes. As the CDs were
igniter in Test B1, only the T-shirts on racks 1 made of polycarbonate, they were extremely
and 2 were ignited. As the clothes on those tough and had high temperature resistance [24].
racks were not burnt thoroughly and the Therefore, only a few CDs were broken and
distance between racks 2 and 3 is quite long, melted. On the other hand, the CD boxes were
the clothes on racks 3 and 4 could not be made of polystyrene (PS) [25]. They were not
ignited. Therefore, only the clothes on two able to withstand the temperature as high as
racks would be ignited in Test B1. By CDs. Most of the boxes were broken, melted
observation, flashover was said to have or even stuck with other CD boxes. Those
occurred at around 195 s, as thermocouple T2 photos of CDs and CD boxes are shown in Fig.
reached 550 oC and flame was coming out 24.
from the door at that time.
119
International Journal on Engineering Performance-Based Fire Codes
y Due to the induced air-flow by fire, many CDs shops of the mall. Further experiments must be
were turned over and dropped to the floor or carried out for digging out more in future.
even the pool. Those CDs would be ignited Support from fire industry and government is
much easily. The fire became very vigorous in requested.
a short period. As the CDs and CD boxes were
made of plastics, i.e. carbon products, large y It appears that the sprinkler system is good in
amount of black soot was given out and spread suppressing the fire in those tests. If the
through the whole environment system is activated at the right time, the fire
instantaneously. The flame was also covered may probably be controlled. However, there
by the ‘black’ smoke and could not be seen. might be water damage in discharging
excessive water. On the other hand, the
y The fire was controlled very quickly once the arrangement of sprinkler heads should also be
sprinkler system or WMFSS was activated. As carefully designed, so as to enable the
the system was activated once a row of CDs discharged water reaching the burning objects
was ignited, the HRR given out would still be effectively and not blocked by goods.
low such that it is easy to be cooled down
when water or mists was discharged. However, y Water mist system appears to be useful in
the HRR of Test C3 is much lower than Test suppressing the fire. As demonstrated, the fire
C2. This is because the water discharged by tested as in above was extinguished within a
the sprinkler system, i.e. 2.9 ls-1, is much reasonable time after discharge. However, the
larger than the WMFSS, i.e. 0.24 ls-1, the heat HRR would be increased again once the water
liberated would be much easier to be absorbed supply was cut off. Water should be kept on
and cooled down by water in Test C3. discharging for some time. Again, more
studies on different arrangements should be
carried out as performance of the system
7. CONCLUSION depends on the fire scenarios.
The possible heat release rates of burning Fire safety provisions are normally designed for
combustibles under flashover condition in a small protection against accidental fires. However, the
retail shop were measured. Flashover was set off by number of arson fires over the world appears to be
a liquid pool fire with the amount of gasoline increasing [10,11]. The general public is now very
adjusted carefully. Once flashover occurred, say concerned about the hidden fire hazard, even when
satisfying the criteria of 550°C near the ceiling or they are travelling on an underground railway [11].
20 kWm-2 of radiative heat flux recorded at floor Perhaps, it is the right time to review the fire safety
level, the pool fire would only be sustained further codes and the government is doing that at the
for a short time afterward. Observations are: moment. Obviously, a heat release rate database for
local combustibles should be developed first [e.g.
y The peak heat release rate was up to 4 MW as 5,6] and carrying out full-scale burning tests is
in Test C1 for CDs, with flame filling up the necessary. The results are useful for designing
entire room. If the burning retail shop is in a smoke management system [26] in big halls such as
large hall, this 4 MW burning object with the shopping malls or public transport interchanges.
same size as the shop will give a big thermal Fire safety management [19] can be recommended
plume [2]. based on such studies. For example, the clothes are
suggested not to be covered with transparent plastic
y For the preliminary full-scale burning tests bags.
carried out as reported in above, only very
small amount of combustibles were burnt. It is also proposed that the performance of fire
Only one vertical clothes hanger and two suppression systems such as sprinkler system and
vertical CD displays were tested for boutiques WMFSS [27] to be demonstrated by similar burning
and CD shops respectively. In real situations, tests. Anyway, further studies on burning those
many different types with much larger combustibles under flashover retail shop fires are
quantities of goods would be kept in a shop. necessary to give a better understanding of the
Clothing stocks of a boutique would not be possible heat release rates.
only T-shirts. Other goods such as dresses,
trousers, jeans and scarves are also displayed.
Burning such huge amount of combustibles ACKNOWLEDGEMENT
under flashover condition would give very
high HRR, might be several times higher than This project was partly funded by the PolyU
the values measured in the above experiments. President with account number A-078; and by a
Fire would spread out quickly to adjacent PolyU research studentship with number G-W097.
120
International Journal on Engineering Performance-Based Fire Codes
3. Code of Practice for Minimum Fire Services 17. B. Hume, “Sprinkler for life safety in shops: Survey
Installation and Equipment and Inspection on of shops”, Research Report Number 70, Fire
Testing of Installation and Equipment, Fire Services Protection Association, Home Office Fire Research
Department, Hong Kong Special Administrative and Development Group, UK (1997).
Region (1998). 18. G. Grant, J. Brenton and D. Drysdale, “Fire
4. W.K. Chow and Philip C.H. Yu, “Simulation on suppression by water sprays”, Energy and
energy use for mechanical ventilation and air- Combustion Science, Vol. 26, pp. 79-130 (2000).
conditioning (MVAC) systems in train 19. D.E. Della-Giustina, The fire safety management
compartments”, Energy – The International Journal, handbook, American Society of Safety Engineers,
Vol. 25, No. 1, pp. 1-13 (2000). 2nd edition, Des Plaines, Illinois, USA (1999).
5. V. Babrauskas and S.J. Grayson, Heat release in 20. W.K. Chow and B. Yao, “The potential application
fires, Elservier Applied Science, London (1992). of water mist systems for fire protection in atria”,
6. R.D. Peacock, R.W. Bukowski, W.W. Jones, P.A. ASHRAE Transactions: Research, Vol. 107, Part 1,
Reneke, V. Babrauskas and J.E. Brown, “Fire safety pp. 171-177 (2001).
of passenger trains: A review of current approaches 21. W.D. Walton and P.H. Thomas, “Estimating
and of new concepts”, NIST Technical note 1406, Temperatures in Compartment Fires”, The SFPE
National Institute of Standards and Technology, Handbook of Fire Protection Engineering, 2nd
Maryland, USA (1994). Edition, USA, pp. 3-134 – 3-140 (1995).
7. G. Garrad and D.A. Smith, “The characterisation of 22. D. Drysdale, An introduction to fire dynamics, 2nd
fires for design”, Interflam’99, Proceedings of 8th edition, Edinburgh, UK, pp. 233-254 (1999).
International Fire Science & Engineering
Conference, 29 June - 1 July, Edinburgh 23. D.T. Gottuk and R.J. Roby, “Effect of combustion
Conference Centre, Scotland, Vol. 1, pp. 555-566 conditions on species production”, The SFPE
(1999). Handbook of Fire Protection Engineering, 2nd
edition, USA, pp. 2-64 – 2-81 (1995).
8. G.C.H. Lui and W.K. Chow, “Survey on the fire
safety requirements in karaoke establishments”, 24. E.A. Lemaster, Compact disc manufacturing –
International Journal on Engineering Performance- Procedures and processes, University of
Based Fire Codes, Vol. 2, No. 1, pp. 1-13 (2000). Washington, USA, pp. 1-10 (1994).
9. South China Morning Post, “21 killed as they flee a 25. S.M. Zhang, “Architectural decorating materials”,
Chicago nightclub in panic”, 18 February (2003). China Machine Press, pp. 120-124 (2003).
10. W.K. Chow, “Instant responses – On the attack fire 26. NFPA 92B, Guide for smoke management systems
at World Trade Centre”, International Journal on in malls, atria, and large areas, 1995 edition,
Engineering Performance-Based Fire Codes, Vol. 3, National Fire Protection Association, Quincy, MA,
No. 3, pp. 128-129 (2001). USA (1995).
11. South China Morning Post, “120 die in Korea 27. W.K. Chow, Y. Gao, H. Dong, G.W. Zou and L.
subway arson attack”, 19 February (2003). Meng, “Will water mist extinguish a liquid fire
rapidly?”, Architectural Science Review, Vol. 46,
12. South China Morning Post, “14 injured in peak- No. 2, pp. 139-144 (2003).
hour MTR arson attack”, 6 January (2004).
13. M. Månsson, M. Dahlberg, P. Blomqvist and A.
Ryderman, “Combustion of chemical substances:
Fire characteristics and smoke gas components in
large-scale experiments”, Swedish National Testing
and Research Institute, Fire Technology and
121