Descending of smoke layering interface in long channel hot smoke tests
L.H. Hu, Y.Z. Li, R. Huo
State Key Laboratory of Fire Science, University of Science and Technology of China, Hefei, Anhui, China;
Abstract Descent of smoke layer in a long channel when carrying out hot smoke tests with small fires was presented. An experimental method using infrared beams is used to track the smoke layer interface front. The arrival time of the smoke layer interface front to a designated position would be measured by the abrupt attenuation of the infrared beam intensity. Field measurements were carried out in an underground long channel of length 96 m. Diesel pool fires up to 1.5 MW were studied. The results on tracking the smoke layer interface front were compared with the measurements by using thermocouples, thermal resistors and visual observations. This method is demonstrated to be useful for tracking the smoke layer interface front under small fires. Therefore, the method is suitable for studying the smoke layer descending in hot smoke tests for evaluating smoke management systems in long channels. The Descending of smoke layer was tracked at 7 m, 39 m and 79 m away from the fire. Results showed that the smoke layer at cross section of 79 m away from the fire descended much faster than that of 7 m and 39 m. The smoke layer seemed to descend faster at positions further away from the fire due to longitudinal air entrainment at the smoke layer interface. Keywords: smoke layer interface front tracking, infrared beam, thermocouple
Introduction Hot smoke tests  are commonly requested by the Authority in the Far East to evaluate the smoke management system in long channels  and atriums. This has been applied mostly to performance-based design  where both zone and field fire models  are used for predicting the critical time for smoke spreading to adjacent occupied zones. The main reason is that the modeling results are not so convincing . Many hot smoke tests had been carried out in big channels, at least in the Far East . But there were problems in carrying out the tests in long channels with a small fire. Earlier studies showed that the propagation of smoke in long channels or corridors appears to have two phases: the ceiling jet forming phase and the smoke descending phase . The descending of smoke layer is a key point when evaluating the smoke management system [7-9]. It is important to judge whether the smoke management system can provide enough Available Safely Egress Time (ASET) for human evacuation in case of a fire. Smoke layer interface fronts and ceiling jet fronts were usually tracked by visual observations by human eyes; and verified to some extent by thermocouples [10-12]. Advanced techniques such as optical fibres  and laser sheets [12,14] were introduced. All these methods have problems in larger-scale tests in long channels with testing fires of low heat release rate. In common size room fires, the smoke layer interface will have uniform heights from the floor, thus the two-layer zone model can give good prediction of the descending of the smoke layer interface [15,16]. But in long channels, longitudinal entrainment will happen at the smoke layer interface along with the traveling of the buoyancy-driven smoke flow. Although the longitudinal entrainment coefficient was reported to be very small , the accumulated entrained mass should be considerable after long distance traveling in a long channel. The smoke layer interface height will be different at different longitudinal distance from the fire. This point should be considered to improve the current zone models to be applicable in long channels. In this paper, a new method based on infrared beams was introduced to track the smoke layer interface front. Experiments were carried out in an underground channel of length 96 m. An infrared beam system was set up at some designated positions. The arrival of the smoke layer interface front was indicated by the abrupt attenuation of the infrared beam intensity received by the receiver. The results were compared with traditional measurements by thermocouples and visual observations by human eyes. The descending of smoke layer interface front at three different locations, being 7 m , 39 m and 79 m away from the fire, was recorded by the infrared beam system. They are compared to see the different descending speed of smoke layer at different distances from the fire in such long channels.
Field measurements were carried out in an underground channel of length 96 m and height 2.65 m in South China. The plane view of the experimental layout and the channel is shown in Figure 1. The width of straight part was 8m. The north end was closed and there was an opening of size 4 m (wide) x 2.65 m (high) at the south end. A diesel pool fire was placed at floor level about 9 m from the north end at the centre of the two sidewalls. Two sets of tests were conducted with four and eight pans of diesel respectively. The peak heat release rates for the pool fires were about 0.8 MW and 1.5 MW. A thermocouple tree and an infrared beam section were positioned at 39 m from the fire source to track the smoke layer interface front at the middle part of the channel. And two other infrared beam section alone were also installed at 7 m and 79 m from the fire, for tracking the smoke layer descending at the near and far fire positions. The positions of the three infrared beam section positions are also shown in Figure 1. The vertical thermocouple tree composed of eight thermocouples with height of 2.65 m, 2.4 m, 2.15 m, 1.9 m, 1.65 m, 1.4 m, 1.15 m and 0.9 m from the floor level. The infrared beam section was composed of five infrared beam emitters with height of 2.45 m, 1.95 m, 1.45 m, 0.95 m and 0.45 m and two infrared beam receivers with height of 2.45 m and 0.45 m. The receiver #1 at 2.45 m high receives the infrared light from the emitter of 2.45 m, and the receiver #2 at 0.45 m high can receive the other four infrared lights from the emitters of 1.95 m, 1.45 m, 0.95 m and 0.45 m respectively. The thermocouple tree and the infrared beam section were shown in Figure 2. 3. The infrared system
The intensity of an infrared beam passing through smoke would be attenuated by refraction, scattering and absorption due to the smoke particles . The relationship between the infrared intensity before (Io) and after (I) passing through smoke of optical length L is described in terms of the light extinction coefficient K as follows: I = I 0 exp(? KL) (1)
Note that K is given in terms of the specific extinction coefficient Km and smoke mass concentration Ms as: K = KmM s (2)
North 96 m 4m
IR Section 3 IR Section 2 IR Section 1
16 m 9m 88 m
Figure 1: Schematic Plane view of the design of the channel
Smoke layer Thermocouple tree
Emitter1 Emitter2 Emitter3 Emitter4 Emitter5 Infrared beam section Smoke layer interface
Figure 2: Design of the infrared beam section and the thermocouple tree
Earlier studies showed that the value of Km depends on the physical properties of the smoke particles and the wavelength λ of the infrared beam. Taking D as the diameter of the smoke particle, ρ s as its density and Qext as the extinction coefficient of a single smoke particle, K m is:
Km = 3 2ρ s
1 dM s D ? ? Qext ( , n r )dD λ D dD
Qext is a function  of the ratio D/λ and a compound refraction index nr. The infrared system was set up in the channel at designated positions as shown in Figure 1b. Infrared beams launched from the emitter on one side of the wall would incident onto the receiver on the opposite side as in Figure 3. The emitter was composed of an array of infrared emitting diodes. The corresponding receiver was an Infrared Charge Coupled Device (CCD) camera. Signals received by this camera would be collected by a parallel video processor, then sent out to a central processor unit for signal sampling and processing.
Infrared CCD Camera Signal sampling and processing
Parallel video processor
Infrared diode array
Figure 3: Schematic of the infrared beam system When the smoke layer interface front reached the infrared beam, the light intensity at the receiver would suddenly be attenuated. The arrival of the smoke layer interface front was then identified. Typical signals achieved by three pairs of infrared beams at different positions are shown in Figure 4. It is in the middle part of about 39 m from the fire with different heights from the floor level.
Signal sampled (Beam intensity)
200 180 160 140 120 100 80 60 40 20 0 0 50 100 150 200 250 300 350 400 450 500 550 600
Figure 4: Arrival of smoke layer interface front indicated by infrared beams 4. Results and Discussion
Typical smoke temperature curves measured by the thermocouple tree are shown in Figure 5. The arrival of the smoke layer interface front, at each thermocouple position was identified by a sudden temperature rise. The traveling time for the smoke layer interface fronts to the five certain heights could thus be found.
70 60 50 40 30 20 0 100 200 300 400 500
2.65 m 2.4 m 2.15 m 1.9 m
Figure 5: Temperature measured by the thermocouple tree The arrival time of the smoke layer interface front to some certain heights deduced from temperature curve and the infrared beam system are compared in Figure 6. It can be seen that at higher positions, such as at height of 2.45 m or 1.95 m, the arrival times by the thermocouple tree and that by the infrared beams are very near. But when smoke layer descended to the lower part of the channel, the arrival times achieved by the infrared beams are shorter than that by the thermocouple tree. The smoke temperature measured by the
thermocouple at the lower part of the channel was also found to be very low, being a little higher than that of the ambient, just as that of the ceiling jet front being far away from the fire source. And according to one visual observation for the smoke layer interface descending in
the test of 0.8 MW, it took about 300 s for the smoke layer interface front to descend to 0.5 m, this was nearer to that achieved by the infrared beams. The smoke particles at the lowest part of the smoke layer would also have not enough temperature to active a thermocouple, but would be enough to attenuate the light intensity received by the receiver.
350 300 250
0.8M W , 1.5M W , 0.8M W , 1.5M W ,
tem perature tem prature infrared beam infrared beam
200 150 100 50 0.0 0.5 1.0 1.5 2.0 2.5 3.0
H eight from the floor level (m )
Figure 6: Arrival times of smoke layer interface at different heights (39 m from the fire) Therefore, using infrared beams is better than using thermocouples for field measurement in long channels with a small fire. The differences between the traveling times measured by the three methods can be very high for much smaller fires in long channels in hot smoke tests. The arrival times of smoke layer interface to the heights of the emitters recorded by the infrared beam system are summarized in Table 1. As the times taken for the smoke flow to reach the cross sections at 7 m, 39 m and 79 m from the fire are different, the descending time of smoke layer in these three locations are normalized by t descending =t arrival ?t travel to get a comparison of descending speed of the smoke layer there, where tdescending, tarrival and ttravel are the descending time, the arrival time recorded by the infrared beam system and the time taken for the longitudinal smoke flow to travel. The value of ttravel is identified by the arrival time of 2.45m high recorded by the infrared beam system in Table 1. The descending times the smoke layer interface fronts at the three locations for the two fire sizes are compared in Figure 7. It can be seen that the descending of smoke layer at 39 m is a bit faster than that at 7 m, while that at 79 m was much faster than at the other two positions. The descending of smoke layer at further positions was faster. The descending speeds of smoke layer interfaces were nearly the same at near fire positions, this was just the
same as what was seen in normal sized room fires. Table 1: Smoke layer descending time tracked by infrared beam section at different distance Height from floor level (m) 2.45 1.95 1.45 0.95 0.45 7m 0.8 MW 1.5 MW 20.2 17.7 20.3 17.8 24.4 17.9 214 148 263.7 202.5 39 m 0.8 MW 1.5 MW 80 63 84 64 85 64 265 172 320 238 79 m 0.8 MW 1.5 MW 180 132.6 180 132.9 180 137.8 240 182.7 274 210.9
Descending time (s)
7 m away from fire 39 m away from fire 79 m away from fire
Smoke layer interface height (m)
Descending time (s)
7 m away from fire 39 m away from fire 79 m away from fire
Smoke layer interface height (m)
(b) 1.5 MW Figure 7: Descending times of smoke layer interface fronts at different distance
There are two factors determining the longitudinal development of the smoke layer interface as shown in Figure 8: the diffusion factor and the buoyancy factor. The diffusion due to the different velocities of the up smoke flow and the lower air flow drags the smoke particle to the lower position of the current interface and entrains the fresh air into the smoke layer. The buoyancy due to the higher temperature of the smoke flow than the ambient drives the smoke particle back to the upper position of the current interface. The diffusive entrainment coefficient would be determined by the velocity difference of the upper smoke flow and the lower air flow. The changes of this velocity differences seemed to be small along the channel. However, the buoyancy seemed to vary largely due to the fast exponential smoke temperature decay along the channel. At a certain distance from the fire, the smoke temperature is still high enough, although decaying along the channel, to give comparable buoyancy force to the diffusive force. After that, when the smoke temperature is too low, the diffusive force will be only dominate factor and the smoke layer will descend faster. Smoke layer Smoke layer interface Diffusive Buoyancy entrainment
Figure 8: Effecting factors at the smoke layer interface 5. Conclusions In evaluating the performance of smoke management systems in long channels by hot smoke tests using small testing fires of low heat release rates, there are difficulties in tracking the smoke layer interface front. A new method of using infrared beams was proposed. The descending of smoke layer interface at different distances from the fire in long channels was recorded and studied by this method. Field tests were carried out in an underground channel of length 96 m. The descending time of the smoke layer interface at 7 m, 39 m and 79 m away from the fire was compared. Results on arrival times of the smoke layer interface front tracked by infrared beams were compared with those by thermocouples and by visual observations. The infrared beam method was demonstrated to be good in tracking the smoke layer
interface front. This technique is better than using the temperature method. Such infrared system is more suitable for hot smoke tests in long channels but with testing fires of low heat release rates. The comparison of descending speed of smoke layer interface at 7 m, 39 m and 79 m away from the fire indicated that the smoke layer descended faster at further positions away from the fire in such long channels. The descending of smoke layer will be much faster after traveling a certain distance away from the fire, due to the buoyancy contribution decayed to be relative small and not comparable to the diffusive entrainment.
Acknowledgement This work was supported by Specialized Research Fund for the Doctoral Program of Higher Education (SRFDP) under Grant No. 20030358051. References  Chow WK. On carrying out atrium hot smoke tests. Architectural Science Review 2005;48(1):105-107.  Chow WK, Li Jojo SM. Safety requirement and regulations reviews on ventilation and fire for tunnels in the Hong Kong Special Administrative Region. Tunnelling and Underground Space Technology 1999;14(1):13-21.  Chow WK. Review on fire models and their application to design based on 'fire engineering approach'. Keynote Lecture – International Conference on Building Fire Safety, 20-21 November 2003, Gardens Point Campus, Queensland University of Technology, Brisbane, Queensland, Australia; appeared in the proceedings, p. 12-24.  Jones WW, Matsushita T, Baum HR. Smoke movement in corridors - adding to the horizontal momentum equation to a zone model. Proceedings of the 12th Joint Meeting of the UJNR Panel on Fire Research and Safety, 1992, pp. 42-54.  Hu LH, Huo R, Li YZ, Wang HB, Chow WK. Full-scale burning tests on studying smoke temperature and velocity along a corridor. Tunnelling and Underground Space Technology 2005;20(2):223-229.  Bailey JL, Forney GP, Tatem PA, Jones WW. Development and validation of corridor flow submodel for CFAST. Journal of Fire Protection Engineering 2002;22:139-161.  He, Y.P, Wang, J., Wu, Z.K., Hu, L., Xiong, Y., Fang, W.C., 2002. Smoke venting and fire safety in an industrial warehouse, Fire Safety Journal 33, 191-215.  Chow, W. K. 1997. Fire hazard assessment in a big hall with the multi-cell zone modeling concept. Journal of Fire Science, 15, 14-28.
 Hu L.H., Li Y.Z., Huo R. and Wang H.B., Smoke Filling Simulation in a BoardingArrival Passage of an Airport Terminal Using Multi-cell Concept, Journal of Fire Sciences, 2005,Vol. 23, No. 1, 31-53.  Stroup DW, Madrzykowski D. Modeling smoke flow in corridors. International Conference on Fire Research and Engineering, September 10-15, 1995. Orlando, FL. Proceedings.  Kim MB, Han YS, Yoon MO. Laser-assisted visualization and measurement of corridor smoke spread. Fire Safety Journal 1998;31:239-51.  Myung Bae Kim, Yong Shik Han. Tracking the smoke front under a ceiling by a laser sheet and thermocouples. Fire Safety Journal 2000;34:287-295.  Chow WK, Wan Eric TK, Cheung KP. Possibility of using laser-fibre optics as a fire detection system. Optics and Laser in Engineering 1997;27(2):201-210.  Chow WK, Wong William C.W. Experimental studies on the sensitivity of fire detectors. Fire and Materials 1994;18(4):221-230.  Fu, Z.M., Hadjisophocleous, G., 2000. A two-zone fire growth and smoke movement model for multi-compartment buildings. Fire Safety Journal 34, 257-285.  Jones, W.W, Quintiere, J.G., 1984. Prediction of corridor smoke filling by zone models. Combustion Science and Technology 35, 239-253.  Kunsch J.P. Critical velocity and range of a fire-gas plume in a ventilated tunnel. Atmospheric Environment, 1999 33, 13-24.  Wu L.B. and Yuan H.Y. Fire detection and control engineering. Press of University of Science and Technology of China, Hefei, Anhui, China (in Chinese). 1999: 245-246.