- Email: [email protected]
Burning behavior of two adjacent pool fires behind a building in a cross-wind
ARTICLE IN PRESS Fire Safety Journal 44 (2009) 989–996
Contents lists available at ScienceDirect
Fire Safety Journal journal homepage: www.elsevier.com/locate/firesaf
Burning behavior of two adjacent pool fires behind a building in a cross-wind Zhibin Chen a,b,, Kohyu Satoh b, Jennifer Wen a, Ran Huo b, Longhua Hu b a b
Centre for Fire and Explosion Studies, Faculty of Engineering, Kingston University, Friars Avenue, Roehampton Vale, London SW15 3DW, United Kingdom State Key Laboratory of Fire Science, University of Science and Technology of China, Hefei, Anhui 230026, PR China
a r t i c l e in fo
abstract
Article history: Received 10 November 2008 Received in revised form 9 June 2009 Accepted 15 June 2009 Available online 4 July 2009
With the global move towards performance based fire design, fire safety assessment in and around buildings becomes increasingly important. However, key knowledge gaps still exist concerning the behavior of fire swirling, which may be generated if one or more accidental fires are in the passage of the vortices behind an adjacent tall building. The present study is focused on the experimental investigations of the burning behavior of two pool fires behind 1/50 scaled tall buildings with heights varying from 0.565 to 1.165 m in a cross-wind. The objective is to gain insight of the effect of the distance between the two fires (D2), the distance between the fires and the building (D1), wind speed (V), and the height of the scaled building (H) on the burning behavior. Important conclusions have been drawn about the influence of D1 and D2 on the fuel mass loss rate, the influence of D1 on fire swirling, the influence of D2 on the possible merging of the two fires and the effect of wind speed on the mass loss rate. The results suggested the existence of a critical velocity for the cross-wind on the initiation of fire swirling and an approximate value was identified for the conditions in the tests. The investigations also covered the effect of height of the scaled building on the fuel mass loss rate and the occurrence of fire swirling. This relationship was found to be also dependent on the wind speed. Analysis of the results has led to some important recommendations to enhance the fire protection of tall buildings. & 2009 Elsevier Ltd. All rights reserved.
Keywords: Fire swirling Fire merging Mass loss rate Critical velocity
1. Introduction Increasing number of tall buildings are being constructed across the world. In the event of fire, safe evacuation of occupants in and away from the immediate vicinity of such buildings is a major concern. Fire safety assessment in and around these tall buildings is of significant importance. Previous studies [1,2] in this context have addressed the issues of fire spread, smoke movement and control, evacuation of occupants, and building structural analysis against fires. One example was the recent investigations on the collapse of the World Trade Center carried out by NIST [3]. Some other studies have identified the possible formation of fire swirling, which may be generated in city fires due to the complexity of air flow patterns in the vicinity of tall buildings [4–6]. When fluid flow is disturbed by an object, a so-called Karman Vortex Street may form behind this object and propagate downstream. If a fire exists in the passage of these vortices, it may develop into fire swirling due to accelerated fire growth from the rapid entrainment of air into the combustion zone. In a dense
Corresponding author. Centre for Fire and Explosion Studies, Faculty of Engineering, Kingston University, Friars Avenue, Roehampton Vale, London SW15 3DW, United Kingdom. E-mail address: [email protected] (Z. Chen).
0379-7112/$ - see front matter & 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.firesaf.2009.06.006
urban area, buildings of varying heights are often constructed with small gaps between them. If an accidental fire in one building happens to be in the passage of the vortices behind an adjacent tall building, fire swirling could develop in this process. Such fire swirling will not only intensify the burning of the original fire but may also cause secondary fire in the tall building as the elongated flame would tilt against the wind direction and towards the tall building as shown in Fig. 1. This phenomenon can lead to rapid fire spread from one building to another and escalate fire accidents in dense urban area. It is therefore of great importance to gain insight into such phenomena and devise measures to prevent and mitigate related accidents. However, few studies [7,8] have been carried out in this context. In-depth understanding about the generation and behavior of such fire swirling is still lacking. There are also knowledge gaps on the mutual influence of adjacent fires, wind condition and relative distances between the fires and buildings. In the present study, experimental investigations have been carried out to investigate the fire behavior. The generic scenario of two pool fires behind a tall building is considered while the distances between the two pool fires and their relative distance to the building were systematically varied. In the following, the general behavior of two pool fires will be presented. Detailed analysis about the effect of key parameters, such as the distance between the two pools, wind speed, and height of the building, will also be analyzed.
ARTICLE IN PRESS 990
Z. Chen et al. / Fire Safety Journal 44 (2009) 989–996
Nomenclature
Abbreviations
D1 D2 V H m
HSV FPS
the distance between the fires and the building, m the distance between the two fires, m wind speed, m/s the height of the scaled building, m total fuel mass, kg
2. Experiments Experiments were conducted in the combustion wind tunnel (6 m long, 1.8 m wide, and 1.8 m high), of State Key Laboratory of Fire Science (SKLFS), University of Science and Technology of China (USTC). The ambient temperature was about 16 1C. Fresh air was blown into the wind tunnel from right to left (Fig. 2) in order to produce uniform velocity on the tunnel cross-section. The key parameters in the experiments are listed in Table 1. In each test, a 1/50 scale building (0.48 m long and 0.26 m wide) with varying heights from 0.565 to 1.165 m at 0.2 m interval was set in the upstream direction. The wind speed (V) controlled from 0 to 2.0 m/s was chosen based on Froude-number [2] to ensure that the corresponding velocity range (0–14 m/s) is of practical interest in the real scale. The gasoline fuel was contained in two square trays of size 0.1 m 0.1 m placed at equal distance behind the scaled building and equal distance to its centerline as shown in Fig. 2. The distance between the trays’ centre and the nearest rim of the scaled building, D1, was varied from 0.12 m or 0.24 m, while the distance between the two trays (D2, rim to rim) was varied from 0.15 m or 0.30 m. In Cases 1–12, the total fuel mass (m) in each tray was about 0.1 kg for the analysis of the effect of D2 and D1 on the burning behavior only, while in Cases 1332 this was about 0.2 kg to make the duration of quasi-steady combustion be sufficient for the comparative study with concern of probably positive effect of wind speed on combustion. A digital scale (0.01 g accuracy) was employed to measure the transient fuel masses of the two trays to estimate the fuel mass loss rates during the quasi-steady period. A High Speed Video (HSV) camera manufactured by the Photron Limited Company, offering high-speed recording from 60 to 2000 Frames Per Second (FPS), was placed outside the wind tunnel to record variations of flame structure and the fire swirling phenomena. For the protection of the HSV lens, photos were taken in the direction perpendicular to wind speed as shown in Fig. 2. In the experiments, the recording rate was set as 1000FPS. A Charge
Wind
flow
Fig. 1. Schematic of fire behavior behind a tall building in the cross-wind.
High Speed Video Frames Per Second
Coupled Device (CCD) camera located inside the wind tunnel was used to record fire merging and swirling phenomena from the front view.
3. Results and discussions Fig. 3 describes the schematic of the flow around the scaled building in the wind tunnel. Two rows of air vortices with different swirling directions (clockwise and anti-clockwise) can be generated as shown in this figure. When the two pans located on two sides behind the building are ignited, the vaporized fuel gas above the pool surface mixes with the swirling air following its spinning direction, resulting in swirling flame in the same circulation direction as the local vortices. The typical characteristic of swirling flame is its elongated shape, which is used to define whether swirling phenomena happens or not. The measured instantaneous flame shape (Fig. 4) got from Case 18, in which D2 and D1 were 0.24 and 0.15 m, respectively, the wind speed 0.5 m/s and the scaled building 0.765 m, are taken as example to display this phenomenon. It is also seen in Fig. 4 that the two flames are showing a tendency to merge at one point. This could possibly be due to the pressure drop in the space between the two flames [9]. Note that a pool fire is sustained by entraining fresh air into the combustion zone. In the region between the two fires, the flames compete to entrain limited amount of air to sustain combustion. The air entrainment rates on the outer edges of the two pool fires are therefore much bigger than the inner sides, and this has a positive effect to push the two flames together.
3.1. General behavior of the flames Case 18 was also chosen as an example and the HSV recordings are presented in Fig. 5 with a time interval of 0.012 s between the frames. The images in Figs. 5(a–w) are taken by the HSV camera located as shown in Fig. 2. The left pool tray corresponds to Tray2 in Fig. 4. It is seen that during the initial phase, the flame heights above the two trays are almost the same (Fig. 5a) and then the flame height above Tray2 decreases while that above Tray1 increases with time. Then, the flame height above Tray2 starts to increase after it reaches the minimum height (Fig. 5c). Further on, the flame above Tray1 starts to decrease in height (Fig. 5d) while that above Tray2 continues to increase. After a while, the flame height above Tray1 starts to increase (Fig. 5f) while that above Tray2 continues to increase. After the flame above Tray2 reaches its maximum height (Fig. 5i), its height starts to decrease while the flame above Tray1 continues to climb higher. After the flame above Tray1 reaches its maximum height (Fig. 5m), it begins to decrease in height along with the flame above Tray2. Then, after the flame above Tray2 reaches its minimum height (Fig. 5r), it begins to increase in height while the flame height above Tray1 continues to decrease. Finally, the flame heights above both trays become almost the same (Fig. 5w) again.
ARTICLE IN PRESS Z. Chen et al. / Fire Safety Journal 44 (2009) 989–996
991
V
H
t igh He
D2
Width
D1
CCD ra came
V HS era m a c
th Leng
O Fig. 2. Schematic of the experimental setup.
Table 1 Key parameters in the experiments. Cases
m (kg)
V (m/s)
D2, D1 (m)
H (m)
Cases
m (kg)
V (m/s)
D2, D1 (m)
H (m)
Case Case Case Case Case Case Case Case Case Case Case Case Case Case Case Case
0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.2 0.2 0.2 0.2
0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.0 0.0 0.0 0.0
0.24, 0.30 0.24, 0.30 0.24, 0.30 0.12, 0.30 0.12, 0.30 0.12, 0.30 0.12, 0.15 0.12, 0.15 0.12, 0.15 0.24, 0.15 0.24, 0.15 0.24, 0.15 0.24, 0.15 0.24, 0.15 0.24, 0.15 0.24, 0.15
0.565 0.765 0.965 0.565 0.765 0.965 0.565 0.765 0.965 0.565 0.765 0.965 0.565 0.765 0.965 1.165
Case Case Case Case Case Case Case Case Case Case Case Case Case Case Case Case
0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2
0.5 0.5 0.5 0.5 1.0 1.0 1.0 1.0 1.5 1.5 1.5 1.5 2.0 2.0 2.0 2.0
0.24, 0.24, 0.24, 0.24, 0.24, 0.24, 0.24, 0.24, 0.24, 0.24, 0.24, 0.24, 0.24, 0.24, 0.24, 0.24,
0.565 0.765 0.965 1.165 0.565 0.765 0.965 1.165 0.565 0.765 0.965 1.165 0.565 0.765 0.965 1.165
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15
scaled building
wind
pool trays
Fig. 3. Schematic of the flow dynamics.
The above characteristics of the flames can probably be explained by the following reasons. When the vortices move into the flame region and interact with the flame, the fire will entrain and mix with the swirling fresh air. On one hand, the buoyancy forces would lead to an increase in flame height. On the other hand, the flame height may decrease to a lower level than that in no-wind conditions due to the horizontal movement of air and the competition for fresh air between the two fires. Because of asymmetry distribution of vortices on the two sides where the trays are located, the vortices may not always reach the trays at the same time. The variations in the heights of the two flames are, hence, controlled by whether there are vortices intertwining with one, both or none of them.
Tray1
Tray2
Fig. 4. Flame structure from CCD camera in Case 18.
Previous investigators have studied the puffing phenomenon of pool fires and developed empirical formulas to predict the puffing frequency [10–12]. However, those studies were restricted to a
ARTICLE IN PRESS 992
Z. Chen et al. / Fire Safety Journal 44 (2009) 989–996
0.68 0.64 0.60
a
b
c
d
e
f
Mass loss rate (g/s)
0.56 0.52 0.48 0.44 0.40 0.36
D2=0.24m, D1 =0.30m
0.32
D2=0.12m, D1 =0.30m D2=0.12m, D1 =0.15m
0.28
D2=0.24m, D1 =0.15m
0.24
g
h
i
j
k
l
0.20 0.5
0.6
0.7
0.8 H (m)
0.9
1.0
1.1
Fig. 6. The measured fuel mass loss rate.
m
s
n
t
o
u
p
v
q
r
w
Fig. 5. Measured flame images in Case 18.
single pool fire in open environment and without swirling. Hence the findings are not directly applicable to the present scenarios. The coupling between flame oscillation and vortices shedding renders the present cases more complicated. Furthermore, unlike the case for flow passing a cylinder, for flow passing a rectangular object like the scaled building, no quantitative relationship has been developed between wind speed and vortex shedding frequency. Therefore, further study should be covered in the future.
3.2. Effect of relative positions of the two trays (D2, D1) The effect of D2 and D1 on the burning behavior was investigated in Cases 1–12 in which the wind speed was maintained at 0.8 m/s with three different heights (0.565, 0.765, and 0.965 m) of the scaled buildings. Four different positions of the trays were considered: (D2 ¼ 0.24 m, D1 ¼ 0.30 m), (D2 ¼ 0.12 m, D1 ¼ 0.30 m), (D2 ¼ 0.12 m, D1 ¼ 0.15 m), and (D2 ¼ 0.24 m, D1 ¼ 0.15 m). Fig. 6 shows the mass loss rates, which were obtained by differentiating curves of fuel mass during the quasi-steady period. It is seen that for the same height of the scaled building, the mass loss rates are greatly influenced by positions of the two trays (D2, D1). For example with the scaled building height of 0.965 m, the mass loss rates for the four different positions of the trays were 0.475, 0.572, 0.631, and 0.498 g/s, respectively. When the trays were located at position (D2 ¼ 0.12 m, D1 ¼ 0.15 m), the fuel mass loss rate was the largest while it was the smallest when the trays were at position (D2 ¼ 0.24 m, D1 ¼ 0.30 m). Hence it can be concluded that the increase of D2 and D1 both result in the
decrease of fuel mass loss rate. It also appears that D2 has greater effect on the mass loss rate than D1. This means that if D2 decreases, mass loss rate will increase regardless of the variation in D1. The above finding may lead to the wrong conclusion that the fact that D2 has greater effect on the mass loss rate than D1 is responsible for the different swirling characteristics of the two flames. In order to understand this better, typical pictures of the flames from the 0.965 m high-scaled building tests with different locations of the trays are shown in Fig. 7. It is seen that for position (D2 ¼ 0.12 m, D1 ¼ 0.15 m) there is no sign of fire swirling at all, while at position (D2 ¼ 0.24 m, D1 ¼ 0.15 m) the swirling phenomenon is most obvious and the flame heights are the highest as well. It is also seen in Fig. 7 that at positions (D2 ¼ 0.12 m, D1 ¼ 0.15 m) and (D2 ¼ 0.12 m, D1 ¼ 0.30 m), the two flames have merged. Flame merging can enhance fuel vaporization through radiation feedback to the pool surface and convective heat transfer, and hence increases the mass loss rate. It is concluded that the chance of fire merging increases with the decrease of D2 when the two flames get closer, while that of fire swirling increase with the decrease of D2. Fire merging, not fire swirling, is the main determining factor for the fuel mass loss rate. In addition, the heat feedback from the scaled building may also influence the combustion process and the burning rate. The interaction of the two fires in close proximity further increases the radiation feedback and mixing, leading to greater overall mass burning rate. The effect of fire swirling was further investigated through the influence of variation in D1 on the mass loss rate. It can be seen from Fig. 6 that the increase of D1 led to decrease in mass loss rate. The movement of vortices would also influence the combustion process. It is thought that there might be few vortices near the centerline of the scaled building. The vortices formed on the two sides behind the building and propagated in the downstream direction. The chance of swirling vortices close to the two trays is expected to decrease with the increase of D1. This explains why the fire swirling phenomenon is most obvious at position (D2 ¼ 0.24 m, D1 ¼ 0.15 m) in Fig. 7. Further study is still necessary to better understand the nature of vortex shedding behind a rectangular object such as a building.
3.3. Effect of wind speed In order to understand the effect of wind speed and height of the scaled building on the swirling and burning behavior, the
ARTICLE IN PRESS Z. Chen et al. / Fire Safety Journal 44 (2009) 989–996
D2 =0.24m,D1=0.30m
D2 =0.12m,D1=0.30m
D2 =0.12m,D1=0.15m
D2 =0.24m, D1 =0.15m
993
Fig. 7. Measured flame structures for the cases with 0.965 m high-scaled building.
1.5 1.4 1.3
Mass loss rate (g/s)
1.2 1.1 1.0 0.9 0.8
H = 0.565m H = 0.765m H = 0.965m H = 1.165m
0.7 0.6 0.5 0.4 0.3 -0.25 0.00
0.25
0.50
0.75
1.00
1.25
1.50
1.75
2.00
2.25
V (m/s) Fig. 8. The measured mass loss rate.
mass loss rates measured for Cases 13–32 under different wind conditions and for different heights of the scaled building, are plotted in Fig. 8. In these tests, the two trays were located at position (D2 ¼ 0.24 m, D1 ¼ 0.15 m). This position was chosen because it was found to have the most obvious swirling in the configurations tested. The wind speed was varied from 0 to 2.0 m/s. In Fig. 8, it can be seen that if the wind speed was within the range of 0–1.5 m/s, the mass loss rate would increase with the increase of wind speed for any heights of the scaled building tested. However, if the wind speed was greater than 1.5 m/s, for the 0.565 m or 0.765 m high-scaled building cases, the mass loss rates of the two trays would also increase with the wind speed but the rate of increase was slower than the cases for wind speed less than 1.5 m/s. In contrast, for the 0.965 and 1.165 m high-scaled
building cases, the mass loss rate decreased sharply with the increase of wind speed. Ideally it would be clearer if we had more data for the wind speeds between 1.5 and 2 m/s. While more tests have already been planned to cover this range, the present results suggest that the critical velocity for the effect of wind speed on the variation of mass loss rate in this condition is approximately 1.5 m/s. For wind speed lower than this, the mass loss rate will increase with the increase of wind speed, and it will decrease with further increase of wind speed beyond this value. Fig. 9 presents the images of the two pool fires in the case of 0.965 m high-scaled building. It can be seen that the size of the flame zones and flame luminosity both increased with the increase of wind speed from 0 to 1.5 m/s. On the contrary, when the wind speed was 2.0 m/s, no swirling was observed and the heights of the flames were much lower in comparison with the other cases with wind speed at and below 1.5 m/s. From the above analysis, it can be concluded that fire swirling has important effect on the mass loss rates of the two trays, and the exact influence depends on the magnitude of the wind speed. The critical velocity for initiating fire swirling behind the 0.965 m or 1.165 m high-scaled building considered here was found to be approximately 1.5 m/s. When the wind speed is below 1.5 m/s, the shear stress will increase with the increase of wind speed, resulting in stronger swirling at two sides of the scaled building which will lead to flame acceleration. However, if the wind speed increases further beyond the critical value, the horizontal movement of the swirling air is so fast that there is very limited time for the vaporized fuel to mix with the air. As a result, both flames tilt to the downstream direction leading to reduced feedback of radiation and convection to the fuel surface and hence lower mass loss rate. Also as shown in Fig. 8, for the 0.565 and 0.765 m high-scaled buildings, the mass loss rate increases with the increase of the wind speed for the whole tested range and no critical wind speed value was found within the tested wind speed range. Further experiments with a wider range of wind speed should be conducted to determine if there is critical lateral velocity in that conditions.
ARTICLE IN PRESS 994
Z. Chen et al. / Fire Safety Journal 44 (2009) 989–996
V=0m/s
V=0.5m/s
V=1.0m/s
V=1.5m/s
V=2.0m/s Fig. 9. Flame structure for the cases with 0.965 m high-scaled building.
3.4. Effect of the height of the scaled building In Fig. 6, it can also be seen that when the wind speed was 0.80 m/s, the fuel mass loss rate remained almost constant with variations in the height of the scaled building for all the four positions tested, implying that in that condition the building height has little effect on the mass loss rate for this wind condition. Further evidence in Fig. 8 then suggests that when the wind speed was within the range of 0–1.0 m/s, there were only marginal differences between the measured mass loss rates for different heights of the scaled building. Additionally, in Fig. 8 the fuel mass loss rates for all scaled building heights at the wind speed of 0.8 m/s are expected to range from 0.80 to 0.85 g/s, while Fig. 6 nearly gives the constant value of 0.50 g/s for the same parameter (D2 ¼ 0.24 m, D1 ¼ 0.15 m). It can be explained that the same tray with more fuel gives larger fuel evaporation rate according to fuel boiling phenomena observed during the quasisteady combustion of experiments. However, when the wind speed was varied from 1.0 to 2.0 m/s, the effect of scaled building height became obvious. The increase in the height of the scaled building will result in the decrease of the mass loss rate. This finding is in line with the changing patterns of the flame structures under 2.0 m/s wind speed as shown in Fig. 10. This figure also shows that swirling becomes less obvious with the increase of height and the flame height
decreases with the increase of scaled building height. If the wind flow passes the scaled building over its roof, swirling vortices will also be produced there. Consequently, there is ample swirling fresh air near the roof region of the scaled building. When the scaled building height is lower, this region comes closer to the pool fires. As a result, the flames can also entrain the swirling air from this region, which may then accelerate combustion and increase the mass loss rate. When the scaled building is taller, this region is further away above the fires and hence has less effect on the combustion process.
4. Summary Experimental investigations have been carried out on the burning behavior of two pool fires behind the scaled building in the cross-wind. The study has uncovered general behavior of the two flames and the effect of D2, D1, V as well as H on the fuel mass loss rate, fire swirling and fire merging. In particular, the following conclusions can be drawn concerning the influence of D1 and D2. The increase of both D2 and D1 will result in the decrease of fuel mass loss rate. However, D2 has greater influence on the fuel mass loss rate than D1 while D2 has more influence on fire merging which can accelerate combustion. D1 was also found to be the determining factor for the occurrence
ARTICLE IN PRESS Z. Chen et al. / Fire Safety Journal 44 (2009) 989–996
H=0.565m
995
H=0.765m
H=0.965m
H=1.165m
Fig. 10. Flame structure for the cases in 2 m/s cross-wind.
of fire swirling. With the decrease of D1, more swirling fresh air is entrained into the flame, the flame height will increase. In terms of the effect of cross-wind, it is found that the increase of wind speed will not always lead to the increase of fuel mass loss rate, and a critical wind speed exists. If the wind speed exceeds this critical value, the mass loss rate will decrease and there would be no fire swirling. For the cases of 0.965 and 1.165 m highscaled building, 1.5 m/s was found to be likely the critical velocity for the initiation of fire swirling. It was also found that when the wind speed varied from 0 to 1.0 m/s, the change in the height of the scaled building had little effect on the fuel mass loss rate while this effect became obvious for the cases with a cross-wind in the range between 1.0 and 2.0 m/s. Based on the above findings, the following recommendations can be drawn in the context of fire protection of tall buildings and fire fighting
The main directions of cross-wind in the region should be
identified, and the potential of fire swirling should be taken into account when designing the maximum height of tall buildings in a dense urban area. Specific measures should be taken to avoid the placement of combustible materials in the vicinity of tall buildings, and especially they should not be placed at two sides next to them on the leeward side of cross-wind in order to avoid fire swirling and elongated flames. In terms of fighting multiple fires next to a tall building, the priority should be given to those situated at two sides of it to prevent rapid increase of flame heights, which may lead to fire spreading from lower to higher floors or from one building to another, and furthermore when resource is limited, it may be beneficial to focus effort on attacking fires on one side first in order to avoid mutual influence and fire merging.
With help of dimensionless analysis, further study is recommended to quantify fire swirling and merging. Such study should
include a wider range of wind conditions to help establish the value of the critical wind speed in different conditions. Further investigations on the differences of the fuel mass loss rate between one-tray and two-tray fires under different combinations of D1 and D2 should lead to insight on fire merging and help to establish the critical merging distance for different size fires. Finally, it should be pointed out that the present experimental investigations were conducted in reduced scale geometrical configurations. Further study should also include scaling analysis and consider the possible block effect of scaled buildings in order to put the quantitative findings into to practice.
Acknowledgements This work was sponsored by National Natural Science Foundation of China under Grant no. 50676090. The authors gratefully acknowledge the Nation Funding Scholarship sponsored by China Scholarship Council (CSC) and Oversea Research Students Awards Scheme (ORSAS) funded by Higher Education Funding Council for England (HEFCE), and the contributions of Naian Liu, Linhe Zhang, Guangyue Dai, Yu Zhong, Bole Xu, Shaohua Mao, Zhaoyu Xu, Shi Zhu, et al., who give great supports of experiments. References [1] D. Drysdale, An Introduction to Fire Dynamics, second ed, Wiley, Boca Raton, FL, 1998. [2] B. Karlsson, J.G. Quintiere, Enclosure Fire Dynamics, CRC Press, 2000. [3] J.G. Quintiere, M. diMarzo, R. Becker, A suggested cause of the fire-induced collapse of the world trade towers, Fire Safety Journal 37 (2002) 707–716. [4] Q.S. Li, Y.Q. Xiao, J.Y. Fu, Full-scale measurements of wind effects on the Jin Mao building, Journal of Wind Engineering and Industrial Aerodynamics 95 (2007) 445–466. [5] J. Burnett, M. Bojic, F. Yik, Wind-induced pressure at external surfaces of a high-rise residential building in Hong Kong, Building and Environment 40 (2007) 765–777.
ARTICLE IN PRESS 996
Z. Chen et al. / Fire Safety Journal 44 (2009) 989–996
[6] W.K. Chow, Wind-induced indoor-air flow in a high-rise building adjacent to a vertical wall, Applied Energy 77 (2004) 225–234. [7] A.Yu. Snegirev, J.A. Marsden, J. Francis, G.M. Makhviladze, Numerical studies and experimental observations of whirling flames, International Journal of Heat and Mass Transfer 47 (2004) 2523–2539. [8] K. Kuwana, K. Sekimoto, K. Saito, F.A. Williams, Scaling fire whirls, Fire Safety Journal 43 (2008) 252–257. [9] P.H. Thomas, R. Baldwin, A.J.M. Heselden, Buoyant Diffusion Flames: Some Measurements of Air Entrainment, Heat Transfer and Flame
Merging, in: Tenth Symposium (International) on Combustion, 1965, pp. 983–996. [10] P.J. Pagni, Pool Vortex Shedding Frequencies, Applied Mechanics Review 43 (1990) 153–170. [11] B.J. McCaffrey, Purely Buoyant Diffusion Flames: Some Experimental Results, Gaithersburg, MD, 1979. [12] B.M. Cetegen, T.A. Ahmed, Experiments on the periodic instability of buoyant plumes and pool fires, Combustion and Flame 93 (1993) 157–184.
Contents lists available at ScienceDirect
Fire Safety Journal journal homepage: www.elsevier.com/locate/firesaf
Burning behavior of two adjacent pool fires behind a building in a cross-wind Zhibin Chen a,b,, Kohyu Satoh b, Jennifer Wen a, Ran Huo b, Longhua Hu b a b
Centre for Fire and Explosion Studies, Faculty of Engineering, Kingston University, Friars Avenue, Roehampton Vale, London SW15 3DW, United Kingdom State Key Laboratory of Fire Science, University of Science and Technology of China, Hefei, Anhui 230026, PR China
a r t i c l e in fo
abstract
Article history: Received 10 November 2008 Received in revised form 9 June 2009 Accepted 15 June 2009 Available online 4 July 2009
With the global move towards performance based fire design, fire safety assessment in and around buildings becomes increasingly important. However, key knowledge gaps still exist concerning the behavior of fire swirling, which may be generated if one or more accidental fires are in the passage of the vortices behind an adjacent tall building. The present study is focused on the experimental investigations of the burning behavior of two pool fires behind 1/50 scaled tall buildings with heights varying from 0.565 to 1.165 m in a cross-wind. The objective is to gain insight of the effect of the distance between the two fires (D2), the distance between the fires and the building (D1), wind speed (V), and the height of the scaled building (H) on the burning behavior. Important conclusions have been drawn about the influence of D1 and D2 on the fuel mass loss rate, the influence of D1 on fire swirling, the influence of D2 on the possible merging of the two fires and the effect of wind speed on the mass loss rate. The results suggested the existence of a critical velocity for the cross-wind on the initiation of fire swirling and an approximate value was identified for the conditions in the tests. The investigations also covered the effect of height of the scaled building on the fuel mass loss rate and the occurrence of fire swirling. This relationship was found to be also dependent on the wind speed. Analysis of the results has led to some important recommendations to enhance the fire protection of tall buildings. & 2009 Elsevier Ltd. All rights reserved.
Keywords: Fire swirling Fire merging Mass loss rate Critical velocity
1. Introduction Increasing number of tall buildings are being constructed across the world. In the event of fire, safe evacuation of occupants in and away from the immediate vicinity of such buildings is a major concern. Fire safety assessment in and around these tall buildings is of significant importance. Previous studies [1,2] in this context have addressed the issues of fire spread, smoke movement and control, evacuation of occupants, and building structural analysis against fires. One example was the recent investigations on the collapse of the World Trade Center carried out by NIST [3]. Some other studies have identified the possible formation of fire swirling, which may be generated in city fires due to the complexity of air flow patterns in the vicinity of tall buildings [4–6]. When fluid flow is disturbed by an object, a so-called Karman Vortex Street may form behind this object and propagate downstream. If a fire exists in the passage of these vortices, it may develop into fire swirling due to accelerated fire growth from the rapid entrainment of air into the combustion zone. In a dense
Corresponding author. Centre for Fire and Explosion Studies, Faculty of Engineering, Kingston University, Friars Avenue, Roehampton Vale, London SW15 3DW, United Kingdom. E-mail address: [email protected] (Z. Chen).
0379-7112/$ - see front matter & 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.firesaf.2009.06.006
urban area, buildings of varying heights are often constructed with small gaps between them. If an accidental fire in one building happens to be in the passage of the vortices behind an adjacent tall building, fire swirling could develop in this process. Such fire swirling will not only intensify the burning of the original fire but may also cause secondary fire in the tall building as the elongated flame would tilt against the wind direction and towards the tall building as shown in Fig. 1. This phenomenon can lead to rapid fire spread from one building to another and escalate fire accidents in dense urban area. It is therefore of great importance to gain insight into such phenomena and devise measures to prevent and mitigate related accidents. However, few studies [7,8] have been carried out in this context. In-depth understanding about the generation and behavior of such fire swirling is still lacking. There are also knowledge gaps on the mutual influence of adjacent fires, wind condition and relative distances between the fires and buildings. In the present study, experimental investigations have been carried out to investigate the fire behavior. The generic scenario of two pool fires behind a tall building is considered while the distances between the two pool fires and their relative distance to the building were systematically varied. In the following, the general behavior of two pool fires will be presented. Detailed analysis about the effect of key parameters, such as the distance between the two pools, wind speed, and height of the building, will also be analyzed.
ARTICLE IN PRESS 990
Z. Chen et al. / Fire Safety Journal 44 (2009) 989–996
Nomenclature
Abbreviations
D1 D2 V H m
HSV FPS
the distance between the fires and the building, m the distance between the two fires, m wind speed, m/s the height of the scaled building, m total fuel mass, kg
2. Experiments Experiments were conducted in the combustion wind tunnel (6 m long, 1.8 m wide, and 1.8 m high), of State Key Laboratory of Fire Science (SKLFS), University of Science and Technology of China (USTC). The ambient temperature was about 16 1C. Fresh air was blown into the wind tunnel from right to left (Fig. 2) in order to produce uniform velocity on the tunnel cross-section. The key parameters in the experiments are listed in Table 1. In each test, a 1/50 scale building (0.48 m long and 0.26 m wide) with varying heights from 0.565 to 1.165 m at 0.2 m interval was set in the upstream direction. The wind speed (V) controlled from 0 to 2.0 m/s was chosen based on Froude-number [2] to ensure that the corresponding velocity range (0–14 m/s) is of practical interest in the real scale. The gasoline fuel was contained in two square trays of size 0.1 m 0.1 m placed at equal distance behind the scaled building and equal distance to its centerline as shown in Fig. 2. The distance between the trays’ centre and the nearest rim of the scaled building, D1, was varied from 0.12 m or 0.24 m, while the distance between the two trays (D2, rim to rim) was varied from 0.15 m or 0.30 m. In Cases 1–12, the total fuel mass (m) in each tray was about 0.1 kg for the analysis of the effect of D2 and D1 on the burning behavior only, while in Cases 1332 this was about 0.2 kg to make the duration of quasi-steady combustion be sufficient for the comparative study with concern of probably positive effect of wind speed on combustion. A digital scale (0.01 g accuracy) was employed to measure the transient fuel masses of the two trays to estimate the fuel mass loss rates during the quasi-steady period. A High Speed Video (HSV) camera manufactured by the Photron Limited Company, offering high-speed recording from 60 to 2000 Frames Per Second (FPS), was placed outside the wind tunnel to record variations of flame structure and the fire swirling phenomena. For the protection of the HSV lens, photos were taken in the direction perpendicular to wind speed as shown in Fig. 2. In the experiments, the recording rate was set as 1000FPS. A Charge
Wind
flow
Fig. 1. Schematic of fire behavior behind a tall building in the cross-wind.
High Speed Video Frames Per Second
Coupled Device (CCD) camera located inside the wind tunnel was used to record fire merging and swirling phenomena from the front view.
3. Results and discussions Fig. 3 describes the schematic of the flow around the scaled building in the wind tunnel. Two rows of air vortices with different swirling directions (clockwise and anti-clockwise) can be generated as shown in this figure. When the two pans located on two sides behind the building are ignited, the vaporized fuel gas above the pool surface mixes with the swirling air following its spinning direction, resulting in swirling flame in the same circulation direction as the local vortices. The typical characteristic of swirling flame is its elongated shape, which is used to define whether swirling phenomena happens or not. The measured instantaneous flame shape (Fig. 4) got from Case 18, in which D2 and D1 were 0.24 and 0.15 m, respectively, the wind speed 0.5 m/s and the scaled building 0.765 m, are taken as example to display this phenomenon. It is also seen in Fig. 4 that the two flames are showing a tendency to merge at one point. This could possibly be due to the pressure drop in the space between the two flames [9]. Note that a pool fire is sustained by entraining fresh air into the combustion zone. In the region between the two fires, the flames compete to entrain limited amount of air to sustain combustion. The air entrainment rates on the outer edges of the two pool fires are therefore much bigger than the inner sides, and this has a positive effect to push the two flames together.
3.1. General behavior of the flames Case 18 was also chosen as an example and the HSV recordings are presented in Fig. 5 with a time interval of 0.012 s between the frames. The images in Figs. 5(a–w) are taken by the HSV camera located as shown in Fig. 2. The left pool tray corresponds to Tray2 in Fig. 4. It is seen that during the initial phase, the flame heights above the two trays are almost the same (Fig. 5a) and then the flame height above Tray2 decreases while that above Tray1 increases with time. Then, the flame height above Tray2 starts to increase after it reaches the minimum height (Fig. 5c). Further on, the flame above Tray1 starts to decrease in height (Fig. 5d) while that above Tray2 continues to increase. After a while, the flame height above Tray1 starts to increase (Fig. 5f) while that above Tray2 continues to increase. After the flame above Tray2 reaches its maximum height (Fig. 5i), its height starts to decrease while the flame above Tray1 continues to climb higher. After the flame above Tray1 reaches its maximum height (Fig. 5m), it begins to decrease in height along with the flame above Tray2. Then, after the flame above Tray2 reaches its minimum height (Fig. 5r), it begins to increase in height while the flame height above Tray1 continues to decrease. Finally, the flame heights above both trays become almost the same (Fig. 5w) again.
ARTICLE IN PRESS Z. Chen et al. / Fire Safety Journal 44 (2009) 989–996
991
V
H
t igh He
D2
Width
D1
CCD ra came
V HS era m a c
th Leng
O Fig. 2. Schematic of the experimental setup.
Table 1 Key parameters in the experiments. Cases
m (kg)
V (m/s)
D2, D1 (m)
H (m)
Cases
m (kg)
V (m/s)
D2, D1 (m)
H (m)
Case Case Case Case Case Case Case Case Case Case Case Case Case Case Case Case
0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.2 0.2 0.2 0.2
0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.0 0.0 0.0 0.0
0.24, 0.30 0.24, 0.30 0.24, 0.30 0.12, 0.30 0.12, 0.30 0.12, 0.30 0.12, 0.15 0.12, 0.15 0.12, 0.15 0.24, 0.15 0.24, 0.15 0.24, 0.15 0.24, 0.15 0.24, 0.15 0.24, 0.15 0.24, 0.15
0.565 0.765 0.965 0.565 0.765 0.965 0.565 0.765 0.965 0.565 0.765 0.965 0.565 0.765 0.965 1.165
Case Case Case Case Case Case Case Case Case Case Case Case Case Case Case Case
0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2
0.5 0.5 0.5 0.5 1.0 1.0 1.0 1.0 1.5 1.5 1.5 1.5 2.0 2.0 2.0 2.0
0.24, 0.24, 0.24, 0.24, 0.24, 0.24, 0.24, 0.24, 0.24, 0.24, 0.24, 0.24, 0.24, 0.24, 0.24, 0.24,
0.565 0.765 0.965 1.165 0.565 0.765 0.965 1.165 0.565 0.765 0.965 1.165 0.565 0.765 0.965 1.165
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15
scaled building
wind
pool trays
Fig. 3. Schematic of the flow dynamics.
The above characteristics of the flames can probably be explained by the following reasons. When the vortices move into the flame region and interact with the flame, the fire will entrain and mix with the swirling fresh air. On one hand, the buoyancy forces would lead to an increase in flame height. On the other hand, the flame height may decrease to a lower level than that in no-wind conditions due to the horizontal movement of air and the competition for fresh air between the two fires. Because of asymmetry distribution of vortices on the two sides where the trays are located, the vortices may not always reach the trays at the same time. The variations in the heights of the two flames are, hence, controlled by whether there are vortices intertwining with one, both or none of them.
Tray1
Tray2
Fig. 4. Flame structure from CCD camera in Case 18.
Previous investigators have studied the puffing phenomenon of pool fires and developed empirical formulas to predict the puffing frequency [10–12]. However, those studies were restricted to a
ARTICLE IN PRESS 992
Z. Chen et al. / Fire Safety Journal 44 (2009) 989–996
0.68 0.64 0.60
a
b
c
d
e
f
Mass loss rate (g/s)
0.56 0.52 0.48 0.44 0.40 0.36
D2=0.24m, D1 =0.30m
0.32
D2=0.12m, D1 =0.30m D2=0.12m, D1 =0.15m
0.28
D2=0.24m, D1 =0.15m
0.24
g
h
i
j
k
l
0.20 0.5
0.6
0.7
0.8 H (m)
0.9
1.0
1.1
Fig. 6. The measured fuel mass loss rate.
m
s
n
t
o
u
p
v
q
r
w
Fig. 5. Measured flame images in Case 18.
single pool fire in open environment and without swirling. Hence the findings are not directly applicable to the present scenarios. The coupling between flame oscillation and vortices shedding renders the present cases more complicated. Furthermore, unlike the case for flow passing a cylinder, for flow passing a rectangular object like the scaled building, no quantitative relationship has been developed between wind speed and vortex shedding frequency. Therefore, further study should be covered in the future.
3.2. Effect of relative positions of the two trays (D2, D1) The effect of D2 and D1 on the burning behavior was investigated in Cases 1–12 in which the wind speed was maintained at 0.8 m/s with three different heights (0.565, 0.765, and 0.965 m) of the scaled buildings. Four different positions of the trays were considered: (D2 ¼ 0.24 m, D1 ¼ 0.30 m), (D2 ¼ 0.12 m, D1 ¼ 0.30 m), (D2 ¼ 0.12 m, D1 ¼ 0.15 m), and (D2 ¼ 0.24 m, D1 ¼ 0.15 m). Fig. 6 shows the mass loss rates, which were obtained by differentiating curves of fuel mass during the quasi-steady period. It is seen that for the same height of the scaled building, the mass loss rates are greatly influenced by positions of the two trays (D2, D1). For example with the scaled building height of 0.965 m, the mass loss rates for the four different positions of the trays were 0.475, 0.572, 0.631, and 0.498 g/s, respectively. When the trays were located at position (D2 ¼ 0.12 m, D1 ¼ 0.15 m), the fuel mass loss rate was the largest while it was the smallest when the trays were at position (D2 ¼ 0.24 m, D1 ¼ 0.30 m). Hence it can be concluded that the increase of D2 and D1 both result in the
decrease of fuel mass loss rate. It also appears that D2 has greater effect on the mass loss rate than D1. This means that if D2 decreases, mass loss rate will increase regardless of the variation in D1. The above finding may lead to the wrong conclusion that the fact that D2 has greater effect on the mass loss rate than D1 is responsible for the different swirling characteristics of the two flames. In order to understand this better, typical pictures of the flames from the 0.965 m high-scaled building tests with different locations of the trays are shown in Fig. 7. It is seen that for position (D2 ¼ 0.12 m, D1 ¼ 0.15 m) there is no sign of fire swirling at all, while at position (D2 ¼ 0.24 m, D1 ¼ 0.15 m) the swirling phenomenon is most obvious and the flame heights are the highest as well. It is also seen in Fig. 7 that at positions (D2 ¼ 0.12 m, D1 ¼ 0.15 m) and (D2 ¼ 0.12 m, D1 ¼ 0.30 m), the two flames have merged. Flame merging can enhance fuel vaporization through radiation feedback to the pool surface and convective heat transfer, and hence increases the mass loss rate. It is concluded that the chance of fire merging increases with the decrease of D2 when the two flames get closer, while that of fire swirling increase with the decrease of D2. Fire merging, not fire swirling, is the main determining factor for the fuel mass loss rate. In addition, the heat feedback from the scaled building may also influence the combustion process and the burning rate. The interaction of the two fires in close proximity further increases the radiation feedback and mixing, leading to greater overall mass burning rate. The effect of fire swirling was further investigated through the influence of variation in D1 on the mass loss rate. It can be seen from Fig. 6 that the increase of D1 led to decrease in mass loss rate. The movement of vortices would also influence the combustion process. It is thought that there might be few vortices near the centerline of the scaled building. The vortices formed on the two sides behind the building and propagated in the downstream direction. The chance of swirling vortices close to the two trays is expected to decrease with the increase of D1. This explains why the fire swirling phenomenon is most obvious at position (D2 ¼ 0.24 m, D1 ¼ 0.15 m) in Fig. 7. Further study is still necessary to better understand the nature of vortex shedding behind a rectangular object such as a building.
3.3. Effect of wind speed In order to understand the effect of wind speed and height of the scaled building on the swirling and burning behavior, the
ARTICLE IN PRESS Z. Chen et al. / Fire Safety Journal 44 (2009) 989–996
D2 =0.24m,D1=0.30m
D2 =0.12m,D1=0.30m
D2 =0.12m,D1=0.15m
D2 =0.24m, D1 =0.15m
993
Fig. 7. Measured flame structures for the cases with 0.965 m high-scaled building.
1.5 1.4 1.3
Mass loss rate (g/s)
1.2 1.1 1.0 0.9 0.8
H = 0.565m H = 0.765m H = 0.965m H = 1.165m
0.7 0.6 0.5 0.4 0.3 -0.25 0.00
0.25
0.50
0.75
1.00
1.25
1.50
1.75
2.00
2.25
V (m/s) Fig. 8. The measured mass loss rate.
mass loss rates measured for Cases 13–32 under different wind conditions and for different heights of the scaled building, are plotted in Fig. 8. In these tests, the two trays were located at position (D2 ¼ 0.24 m, D1 ¼ 0.15 m). This position was chosen because it was found to have the most obvious swirling in the configurations tested. The wind speed was varied from 0 to 2.0 m/s. In Fig. 8, it can be seen that if the wind speed was within the range of 0–1.5 m/s, the mass loss rate would increase with the increase of wind speed for any heights of the scaled building tested. However, if the wind speed was greater than 1.5 m/s, for the 0.565 m or 0.765 m high-scaled building cases, the mass loss rates of the two trays would also increase with the wind speed but the rate of increase was slower than the cases for wind speed less than 1.5 m/s. In contrast, for the 0.965 and 1.165 m high-scaled
building cases, the mass loss rate decreased sharply with the increase of wind speed. Ideally it would be clearer if we had more data for the wind speeds between 1.5 and 2 m/s. While more tests have already been planned to cover this range, the present results suggest that the critical velocity for the effect of wind speed on the variation of mass loss rate in this condition is approximately 1.5 m/s. For wind speed lower than this, the mass loss rate will increase with the increase of wind speed, and it will decrease with further increase of wind speed beyond this value. Fig. 9 presents the images of the two pool fires in the case of 0.965 m high-scaled building. It can be seen that the size of the flame zones and flame luminosity both increased with the increase of wind speed from 0 to 1.5 m/s. On the contrary, when the wind speed was 2.0 m/s, no swirling was observed and the heights of the flames were much lower in comparison with the other cases with wind speed at and below 1.5 m/s. From the above analysis, it can be concluded that fire swirling has important effect on the mass loss rates of the two trays, and the exact influence depends on the magnitude of the wind speed. The critical velocity for initiating fire swirling behind the 0.965 m or 1.165 m high-scaled building considered here was found to be approximately 1.5 m/s. When the wind speed is below 1.5 m/s, the shear stress will increase with the increase of wind speed, resulting in stronger swirling at two sides of the scaled building which will lead to flame acceleration. However, if the wind speed increases further beyond the critical value, the horizontal movement of the swirling air is so fast that there is very limited time for the vaporized fuel to mix with the air. As a result, both flames tilt to the downstream direction leading to reduced feedback of radiation and convection to the fuel surface and hence lower mass loss rate. Also as shown in Fig. 8, for the 0.565 and 0.765 m high-scaled buildings, the mass loss rate increases with the increase of the wind speed for the whole tested range and no critical wind speed value was found within the tested wind speed range. Further experiments with a wider range of wind speed should be conducted to determine if there is critical lateral velocity in that conditions.
ARTICLE IN PRESS 994
Z. Chen et al. / Fire Safety Journal 44 (2009) 989–996
V=0m/s
V=0.5m/s
V=1.0m/s
V=1.5m/s
V=2.0m/s Fig. 9. Flame structure for the cases with 0.965 m high-scaled building.
3.4. Effect of the height of the scaled building In Fig. 6, it can also be seen that when the wind speed was 0.80 m/s, the fuel mass loss rate remained almost constant with variations in the height of the scaled building for all the four positions tested, implying that in that condition the building height has little effect on the mass loss rate for this wind condition. Further evidence in Fig. 8 then suggests that when the wind speed was within the range of 0–1.0 m/s, there were only marginal differences between the measured mass loss rates for different heights of the scaled building. Additionally, in Fig. 8 the fuel mass loss rates for all scaled building heights at the wind speed of 0.8 m/s are expected to range from 0.80 to 0.85 g/s, while Fig. 6 nearly gives the constant value of 0.50 g/s for the same parameter (D2 ¼ 0.24 m, D1 ¼ 0.15 m). It can be explained that the same tray with more fuel gives larger fuel evaporation rate according to fuel boiling phenomena observed during the quasisteady combustion of experiments. However, when the wind speed was varied from 1.0 to 2.0 m/s, the effect of scaled building height became obvious. The increase in the height of the scaled building will result in the decrease of the mass loss rate. This finding is in line with the changing patterns of the flame structures under 2.0 m/s wind speed as shown in Fig. 10. This figure also shows that swirling becomes less obvious with the increase of height and the flame height
decreases with the increase of scaled building height. If the wind flow passes the scaled building over its roof, swirling vortices will also be produced there. Consequently, there is ample swirling fresh air near the roof region of the scaled building. When the scaled building height is lower, this region comes closer to the pool fires. As a result, the flames can also entrain the swirling air from this region, which may then accelerate combustion and increase the mass loss rate. When the scaled building is taller, this region is further away above the fires and hence has less effect on the combustion process.
4. Summary Experimental investigations have been carried out on the burning behavior of two pool fires behind the scaled building in the cross-wind. The study has uncovered general behavior of the two flames and the effect of D2, D1, V as well as H on the fuel mass loss rate, fire swirling and fire merging. In particular, the following conclusions can be drawn concerning the influence of D1 and D2. The increase of both D2 and D1 will result in the decrease of fuel mass loss rate. However, D2 has greater influence on the fuel mass loss rate than D1 while D2 has more influence on fire merging which can accelerate combustion. D1 was also found to be the determining factor for the occurrence
ARTICLE IN PRESS Z. Chen et al. / Fire Safety Journal 44 (2009) 989–996
H=0.565m
995
H=0.765m
H=0.965m
H=1.165m
Fig. 10. Flame structure for the cases in 2 m/s cross-wind.
of fire swirling. With the decrease of D1, more swirling fresh air is entrained into the flame, the flame height will increase. In terms of the effect of cross-wind, it is found that the increase of wind speed will not always lead to the increase of fuel mass loss rate, and a critical wind speed exists. If the wind speed exceeds this critical value, the mass loss rate will decrease and there would be no fire swirling. For the cases of 0.965 and 1.165 m highscaled building, 1.5 m/s was found to be likely the critical velocity for the initiation of fire swirling. It was also found that when the wind speed varied from 0 to 1.0 m/s, the change in the height of the scaled building had little effect on the fuel mass loss rate while this effect became obvious for the cases with a cross-wind in the range between 1.0 and 2.0 m/s. Based on the above findings, the following recommendations can be drawn in the context of fire protection of tall buildings and fire fighting
The main directions of cross-wind in the region should be
identified, and the potential of fire swirling should be taken into account when designing the maximum height of tall buildings in a dense urban area. Specific measures should be taken to avoid the placement of combustible materials in the vicinity of tall buildings, and especially they should not be placed at two sides next to them on the leeward side of cross-wind in order to avoid fire swirling and elongated flames. In terms of fighting multiple fires next to a tall building, the priority should be given to those situated at two sides of it to prevent rapid increase of flame heights, which may lead to fire spreading from lower to higher floors or from one building to another, and furthermore when resource is limited, it may be beneficial to focus effort on attacking fires on one side first in order to avoid mutual influence and fire merging.
With help of dimensionless analysis, further study is recommended to quantify fire swirling and merging. Such study should
include a wider range of wind conditions to help establish the value of the critical wind speed in different conditions. Further investigations on the differences of the fuel mass loss rate between one-tray and two-tray fires under different combinations of D1 and D2 should lead to insight on fire merging and help to establish the critical merging distance for different size fires. Finally, it should be pointed out that the present experimental investigations were conducted in reduced scale geometrical configurations. Further study should also include scaling analysis and consider the possible block effect of scaled buildings in order to put the quantitative findings into to practice.
Acknowledgements This work was sponsored by National Natural Science Foundation of China under Grant no. 50676090. The authors gratefully acknowledge the Nation Funding Scholarship sponsored by China Scholarship Council (CSC) and Oversea Research Students Awards Scheme (ORSAS) funded by Higher Education Funding Council for England (HEFCE), and the contributions of Naian Liu, Linhe Zhang, Guangyue Dai, Yu Zhong, Bole Xu, Shaohua Mao, Zhaoyu Xu, Shi Zhu, et al., who give great supports of experiments. References [1] D. Drysdale, An Introduction to Fire Dynamics, second ed, Wiley, Boca Raton, FL, 1998. [2] B. Karlsson, J.G. Quintiere, Enclosure Fire Dynamics, CRC Press, 2000. [3] J.G. Quintiere, M. diMarzo, R. Becker, A suggested cause of the fire-induced collapse of the world trade towers, Fire Safety Journal 37 (2002) 707–716. [4] Q.S. Li, Y.Q. Xiao, J.Y. Fu, Full-scale measurements of wind effects on the Jin Mao building, Journal of Wind Engineering and Industrial Aerodynamics 95 (2007) 445–466. [5] J. Burnett, M. Bojic, F. Yik, Wind-induced pressure at external surfaces of a high-rise residential building in Hong Kong, Building and Environment 40 (2007) 765–777.
ARTICLE IN PRESS 996
Z. Chen et al. / Fire Safety Journal 44 (2009) 989–996
[6] W.K. Chow, Wind-induced indoor-air flow in a high-rise building adjacent to a vertical wall, Applied Energy 77 (2004) 225–234. [7] A.Yu. Snegirev, J.A. Marsden, J. Francis, G.M. Makhviladze, Numerical studies and experimental observations of whirling flames, International Journal of Heat and Mass Transfer 47 (2004) 2523–2539. [8] K. Kuwana, K. Sekimoto, K. Saito, F.A. Williams, Scaling fire whirls, Fire Safety Journal 43 (2008) 252–257. [9] P.H. Thomas, R. Baldwin, A.J.M. Heselden, Buoyant Diffusion Flames: Some Measurements of Air Entrainment, Heat Transfer and Flame
Merging, in: Tenth Symposium (International) on Combustion, 1965, pp. 983–996. [10] P.J. Pagni, Pool Vortex Shedding Frequencies, Applied Mechanics Review 43 (1990) 153–170. [11] B.J. McCaffrey, Purely Buoyant Diffusion Flames: Some Experimental Results, Gaithersburg, MD, 1979. [12] B.M. Cetegen, T.A. Ahmed, Experiments on the periodic instability of buoyant plumes and pool fires, Combustion and Flame 93 (1993) 157–184.