Article
Scale modeling and numerical simulation of smoke control for rescue stations in long railway tunnels
Journal of Fire Protection Engineering 22(2) 101–131 ! The Author(s) 2012 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1042391512445409 jfe.sagepub.com
Ying Zhen Li Fire Technology, SP Technical Research Institute of Sweden, Bora˚s, Sweden
Bo Lei School of Mechanical Engineering, Southwest Jiaotong University, China
Haukur Ingason Fire Technology, SP Technical Research Institute of Sweden, Bora˚s, Sweden
Abstract Rescue stations are usually provided in very long railway tunnels. Those stations already constructed or under construction worldwide are reviewed and the basic pattern of smoke control during a rescue station fire is identified. A total of 54 model scale tests were carried out to investigate smoke control issues in rescue station fires. The effects of heat release rate, train obstruction, fire source location and ventilation condition on smoke control in the cross-passages of a rescue station were tested and analyzed. A critical smoke layer temperature near the fireproof door protecting the rescue station was investigated theoretically and experimentally and a simple equation for this temperature is obtained. A height of 2.2 m is proposed for the fireproof doors in crosspassages of rescue stations. Keywords Rescue station, tunnel fires, model scale, smoke control, critical velocity, gas temperature
Corresponding author: Ying Zhen Li, Fire Technology, SP Technical Research Institute of Sweden, Bora˚s, Sweden Email:
[email protected]
102
Journal of Fire Protection Engineering 22(2)
Introduction The number of very long railway tunnels, several tens of kilometers, that are designed and constructed is steadily increasing. Fire safety issues are one of the major topics in the design of these very long tunnels. The safety level in a very long tunnel can be improved significantly by use of a rescue station since passengers can evacuate efficiently through multiple crosspassages with a short spacing. Gerber [1] studied the possibility of a train involved in an accident not able to arrive at any of the emergency exits, i.e. tunnel exits or emergency rescue stations, assuming that the train can still travel over 20 km after an accident. The results show that the possibility is around 30% if the interval between the emergency exits is 30 km and about 0.01% if the interval is 20 km [1]. Limited research has been published on smoke control issues in rescue station fires, although several rescue stations have been built and used for many years. Kim and Park [2] performed experiments using cold gas buoyant smoke in the 1:48 scale models of a rescue station and a train. However, rather limited information was obtained, and the main objective of that study was to verify the ventilation scheme of the Sol-An Rail Tunnel in Korea. The method of smoke control in a rescue station fire is to supply fresh air toward the tunnel where the accident occurs to keep the cross-passages free of smoke. A very similar case is smoke control in the cross-passages in the vicinity of a fire. The critical velocity in a cross-passage, defined as the minimum ventilation velocity through the fireproof door that can prevent smoke ingress, was investigated by several authors [3–5]. Tarada [4] proposed a performance-based method to calculate a specific critical velocity for a cross-passage located downstream of the fire, and simply regarded the critical Froude number as 4.5. Li et al. [5] carried out a series of model scale tests, and proposed specific equations based on the critical Froude number method proposed by Tarada [4], however, their results show that the critical Froude number is not a constant. This means that the critical Froude number method may not be reasonable to estimate the critical velocity for smoke control in a cross-passage. Further, Li et al. [5] proposed a robust equation to predict the critical velocity for smoke control in a cross-passage, based on a dimensional analysis and a parametric study of the influence factors on the critical velocity. The objective of this article is to investigate smoke control issues in a rescue station fire in a long railway tunnel. Most of the rescue stations already constructed or under construction all over the world are reviewed with the aim of finding some basic pattern of smoke control in these rescue stations. A simplified method to carry out smoke control tests is proposed based on numerical simulations. Based on these findings, a large number of model scale tests are carried out to investigate smoke control for a rescue station, including a determination of critical velocities in different cross-passages and the critical smoke layer temperature near doors in the cross-passages. The work carried out is considered the best way of realizing this objective.
34
16.2
2012/2013
2007
2000c
2016 2007 2007
Seikan [6]
Gotthard Base [7] Lo¨tschberg [7]
Young Dong [3]
Koralm [8] Guadarrama [9] Taihangshan
32.8 28.4 39.4d
Twin Twin Twin
Single
Twin
Twin
Single
Tunnel typea
1 1 2
1
1
2
2
Number
Internal Internal + Service Internal
Internal
Internal
External + Service
External + Service
Typeb
400 500 540
–
450
440/450
480
Length (m)
50 50 60
–
90
90
40
Interval of crosspassages (m)
Rescue stations
8 11 9
8
6
6
13
Number of crosspassages
Service tunnel supply Shafts suppy/exhaust Shafts suppy/exhaust Shafts suppy/exhaust Shafts suppy Service supply Shafts suppy/exhaust
Ventilation
b
‘Single’ corresponds to a single-bore double-track tunnel, and ‘twin’ corresponds to a twin-bore single-track tunnel. ‘External’ and ‘internal’ correspond to external rescue station and internal rescue station respectively, and ‘service’ means there is at least one additional service tunnel besides the main tunnels. c The construction began in 2000. d One 27.8 km long tunnel and one 11.6 km long tunnel are connected through a viaduct; thus, these two tunnels are considered together as one tunnel from the point of view of safety.
a
57
1988
Tunnels
53.9
Construction year
Tunnel length (km)
Table 1. The rescue stations constructed or under construction.
Li et al. 103
104
Journal of Fire Protection Engineering 22(2)
exit
escape route fresh air
supply fans room inclined shaft
fireproof door 1 exhaust fans room exhaust duct jet fans
fireproof door 2
cross-passage
jet fans left line
fireproof valves
vertical shaft
right line
tunnel
tunnel supply duct
Taiyuan
Shijiazhuang
Figure 1. A schematic view of the ventilation in rescue station of Taihangshan tunnel.
exhaust fire
accident tunnel
accident train
platform
cross-passges
cross-passages
safe tunnel
Safe tunnel
platform supply
inflow evacuation
Figure 2. The internal rescue station in a long tunnel.
Rescue stations in long tunnels Table 1 gives a summary of the rescue stations constructed or under construction worldwide. All these tunnels are very long, with a length from 16.2 km to 53.9 km. It is shown clearly in Table 1 that most of these tunnels are twin-bore single-track tunnels, except the Seikan tunnel. The length of the rescue stations ranges from 400 m to 540 m. At least six cross-passages are available or planned in each rescue station in these very long tunnels (Table 1). According to the placement of platforms and safe regions, the rescue stations in the long tunnels in Table 1 can be categorized into two types: internal rescue station and external rescue station, as shown in Figures 2 and 3. Note that only six crosspassages are plotted here, in practice there could be more. In an internal rescue station, the two neighboring tunnels are connected together using cross-passages, and the neighboring tunnel is regarded as safe region. In an external rescue station, the two running tunnels are independent and extra regions connected to the
Li et al.
105
supply Safe region
(Service tunnel or pilot tunnel)
cross-passages
cross-passages accident tunnel fire
accident train
platform
accident tunnel
exhaust
safe tunnel
platform
safe tunnel
cross-passages
cross-passages Safe region (Service tunnel or pilot tunnel) inflow evacuation
Figure 3. The external rescue station in a long tunnel.
incident tunnel are regarded as safe regions. The extra region could be a service tunnel or a pilot tunnel or a large space particularly built for the rescue station. Note that for both types of rescue stations, the safe region, i.e. safe tunnel or safe region, is pressurized to push fresh air flow into the incident tunnel to prevent smoke spread from the incident tunnel to the escape path or the safe region. Meanwhile, the smoke in the incident tunnel could be displaced by the exhaust shaft available in the incident tunnel or be moved away along a preferred direction by the forced longitudinal flow in the incident tunnel. This basic pattern of smoke control in a rescue station in a very long tunnel suggests that it is possible and reasonable to carry out model scale rescue station fire tests using a test rig consisting only of an incident tunnel and several cross-passages. Nowadays, rescue stations are specified only in long railway tunnels. The main reason is the difficulty in evacuating large numbers of passengers from a train during a short period. In a rescue station in a long tunnel, emergency communication, emergency lights, fire hose and other safety facilities are also required. In case of a fire, the rescue station provides routes for evacuees at the evacuation stage and also provides the fire brigade a shortcut to attack the fire during the firefighting stage. It can also be used as a construction site during the construction stage and as a maintenance station under normal conditions. In summary, a rescue station can be used for emergency evacuation, fire-fighting, maintenance and construction.
Numerical study of test method It is difficult and also not necessary to use a scale model of the entire rescue station with the associated tunnels. Some simplifications are needed to carry out scale model tests efficiently. The primary idea is to investigate smoke control in cross-passages of the rescue stations using a scale model rescue station with a limited length of tunnel and
106
Journal of Fire Protection Engineering 22(2)
several cross-passages. If a model rescue station with a limited length is used in the tests, the thermal boundaries on both sides of the incident tunnel should also be modeled by some means, since in reality these two boundaries are not open to the ambient. In order to model the thermal boundaries, two blocks (solid blockage elements) are used in the tests. This method is validated using numerical simulations discussed in the following section. The simulations are conducted using Fire Dynamics Simulator (FDS) developed by NIST [10].
Mesh size analysis The fire properties are directly related to the fire characteristic diameter, which can be expressed as follows [10]: D ¼
Q pffiffiffi r0 c p T o g
2=5 ð1Þ
It is shown in equation (1) that the characteristic diameter D* is directly related to the heat release rate (HRR). Hu et al. [11] carried out simulations of a full-scale tunnel fire and good agreement was obtained between the simulation results and the tests data. The mesh size used in their study was in a range of 0.1D* to 0.12D*. Hwang and Edwards [12] also carried out simulation of tunnel fires and the mesh size used was in a range of 0.06D* to 0.12D*. In the authors’ previous study [13], the effect of mesh sizes on the simulation results of longitudinally ventilated tunnel fires was also investigated. The simulated tunnel was 100 m long, 4.5 m wide and 5 m high. The parameters investigated include mesh sizes, HRRs and ventilation velocities. The results show that there is almost no difference in the results if the mesh sizes are smaller than 0.075D*. The above analysis suggests that a mesh size of 0.075D* is a reasonable value in the simulation. Note that a smaller fire corresponds to a smaller mesh size based on equation (1) and the above analysis. In a model scale fire, the mesh size is much smaller than that in full scale. For a tunnel with a height of H, equation (1) can be transformed into: 2=5 D Q ¼ ¼ Q2=5 r0 cp To g1=2 H5=2 H
ð2Þ
This means that at the same dimensionless HRR, the fire characteristic diameter is directly related to the tunnel height, i.e. the reasonable mesh size is proportional to the tunnel height. This means that the ratio of reasonable mesh sizes between model scale and full scale equals the scale ratio. In other words, the mesh numbers required for model scale and full scale are about the same. Note that this conclusion is deduced based on a similar flow mode and the same dimensionless HRR.
Li et al.
(a)
107
(b)
0.9 m upstream of the fire
0.25
0.20 vertical height (m)
vertical height (m)
0.20
0.5 m upstream of the fire
0.25
0.15 0.10 Test data FDS
0.05
0.15 0.10 Test data FDS
0.05
0.00
0.00 0
30
60
90
120
150
0
40
80
T(°C)
(d)
0.25 m upstream of the fire 0.25
0.25
0.20
0.20 vertical height (m)
vertical height (m)
(c)
0.15 0.10 Test data FDS
0.05
160
200
fire center
0.15 0.10
Test data FDS
0.05 0.00
0.00 0
50
100
150
200
250
300
100
200
300
T(°C)
(e)
400
500
600
T(°C)
(f)
0.85 m downstream of the fire 0.25
0.25
0.20
0.20 vertical height (m)
vertical height (m)
120
T(°C)
0.15 0.10 Test data FDS
0.05
1.8 m downstream of the fire
0.15 0.10 Test dta FDS
0.05 0.00
0.00 0
50
100
150
T(°C)
200
250
300
0
50
100
150
200
250
T(°C)
Figure 4. A comparison of gas temperatures between the simulations and the tests.
Verification of modelling Data from one of the authors’ previous model scale tests [14,15] was used to verify the modeling. The model scale tests were carried out in a model tunnel with a dimension of 12 m (L) 250 mm (W) 250 mm (H). The propane fire burner was set at the floor level and 6 m away from the downstream exit. In the simulation, the tunnel is closely the same as that used in the model-scale tests except a length of 8 m rather than 12 m. The ambient pressure is 95,590 Pa and the ambient temperature is
108
Journal of Fire Protection Engineering 22(2)
Symmetrical plane
L=480m L=240m
L=120m 60m x=0m
60m
60m
Computation domain L=600m
Computation domain L=720m
L=360m
60m
120m
120m
fire
120m
120m x=720m
30m
fire proof door air flow Cross-passages
x
Figure 5. A schematic diagram of the effective length analysis.
23.8 C. The HRR is 9.45 kW and the longitudinal ventilation velocity is 0.51 m/s. A fire source was placed in the center of the tunnel. Only half of tunnel was simulated due to full symmetry. The mesh size used in the simulation is 0.075D*, which has proved to be a reasonable mesh size according to the above analysis. The simulated vertical temperature distribution at different places was compared to the tests data, as shown in Figure 4. It is shown that the simulated results correlate well with the test data measured upstream and downstream of the fire. The relative error is within 20% except the data at 10 mm right above the fire source.
Numerical study of simplification of rescue station fires Effect of lengths. Although it is not possible to model the whole rescue station with the tunnels, it is known that the simulation results get closer to reality by increasing the tunnel length. Note that the smoke spread along the tunnel takes some time, thus the effective tunnel length, i.e. the tunnel length occupied by the smoke, is not so long during the evacuation stage of a tunnel fire. It can be expected that quite reasonable results can be obtained by increasing the tunnel length to a certain value. Numerical simulation of fires in the rescue station and the connected tunnels is carried out to investigate the effect of tunnel length. The simulated rescue station corresponds to a realistic rescue station in the Taihangshan tunnel. The geometry
Li et al.
x=0m
8.4
(b)
7.0
H (m)
5.6 4.2
5.6
2.8
1.4
1.4
0.0 100
200
300
400
500
0.0
600
L=120 L=240 L=360 L=480 L=600 L=720
4.2
2.8
0
x=30m
8.4 7.0
L=120 L=240 L=360 L=480 L=600 L=720
H (m)
(a)
109
0
50
100
T (°C)
(c)
x=60m
8.4
(d) 8.4
200
250
300
x=90m
7.0
7.0
5.6 H (m)
5.6 H (m)
150
T(°C)
dx=120 dx=240 dx=360 dx=480 dx=600 dx=720
4.2 2.8 1.4
L=120 L=240 L=360 L=480 L=600 L=720
4.2 2.8 1.4 0.0
0.0 0
50
100 T(°C)
150
200
0
50
100
150
200
T (°C)
Figure 6. Temperature distribution at different locations (Q ¼ 10.5 MW).
of the 1:20 scale rescue station can be found in Experimental Setup. The rescue station is 600 m long and consists of 9 cross-passages with an interval of 60 m. The fire source in the simulation is placed at the center of the rescue station and there is no mechanical ventilation in the incident tunnel except the flows through the cross-passages. This corresponds to the worst case, which will be discussed later. Half of the rescue station and the connected tunnels are simulated to significantly reduce the computation time. The fire source is a rectangular propane fire with dimensions of 3 m 3 m. The computation domain is 120, 240, 360, 480, 600 and 720 m, respectively, as shown in Figure 5. The HRRs for the whole rescue station are 10.5 MW and 20 MW respectively. These values can be considered as HRRs for a metro car or railcar fire [16]. To prevent smoke from spreading into the cross-passages, a ventilation flow with a velocity of 1.5 m/s is supplied through each fireproof door. Figures 6 and 7 show the temperature distribution at different locations at a HRR of 10.5 MW and 20 MW, respectively. The gas properties in the vicinity of the fire, i.e. within 60 m, are focused on in this study. Note that the 60 m corresponds to 5.5 m in the following model scale tests. It is shown in Figures 6 and 7 that there is almost no difference in the temperature distribution if the length is over 360 m. This may suggest that if the total length of 1:20 model-scale rescue station and the
110
Journal of Fire Protection Engineering 22(2)
(a)
(b)
x=0m
8.4
7.2
7.2
3.6
4.8
2.4
1.2
1.2 0
0.0
100 200 300 400 500 600 700 800 900 ΔT (°C)
x=60m
(d)
7.2
6.0
6.0
L=120m L=240m L=360m L=480m L=600m L=720m
3.6 2.4 1.2 0.0
0
100
200 ΔT (°C)
0
100
200 300 ΔT (°C)
500
4.8 L=120m L=240m L=360m L=480m L=600m L=720m
3.6 2.4 1.2
300
400
m x=90
8.4
7.2
4.8
L=120m L=240m L=360m L=480m L=600m L=720m
3.6
2.4
(c) 8.4
H (m)
6.0 H (m)
L=120m L=240m L=360m L=480m L=600m L=720m
4.8
H (m)
H (m)
6.0
0.0
x=30m
8.4
0.0
0
100
200
300
ΔT (°C)
Figure 7. Temperature distribution at different locations (Q ¼ 20 MW).
connected tunnel is over 36 m, the temperatures are close to the realistic values in the vicinity of the fire, i.e. within the tested three cross-passages. Effect of block height. Since the length of the test bed still needs to be shorter for practical reasons, blocks are used at the two exits of the tunnel to model the thermal boundaries while doing tests in a shorter rescue station. The 1:20 scale model rescue station is 12 m long including 3 cross-passages, corresponding to 240 m in full scale. Note that the symmetry plane is also used here to reduce computation time and thus only half of the domain is simulated. The blocks are installed at the both exits to block the upper layer, as shown in Figure 8. The parameter investigated here is the open height, hop, defined as the vertical distance between the tunnel floor and the lower edge of the block, see Figure 8. The results corresponding to a length of 720 m, as shown in Figures 6 and 7, are used as the reference for comparison. Figures 9 and 10 give the results with different open heights at HRRs of 10.5 MW and 20 MW. The results show that a good correlation can be found between the reference and the results with an open height of 0.65 times tunnel height (0.65 H). It is also shown that there is no difference in the temperature distribution after the open height gets lower than 0.2 H.
Li et al.
111
79
block
hop
7.6 m 9m
Figure 8. A schematic diagram of the block on the sides of the rescue station.
x=0m
8.4
(b) reference 0.1H 0.2H 0.4H 0.6H 0.65H 0.7H 0.8H 1.0H
vertical height (m)
7.0 5.6 4.2 2.8
x=30m
8.4 7.0
vertical height (m)
(a)
reference 0.1H 0.2H 0.4H 0.6H 0.65H 0.7H 0.8H 1.0H
5.6 4.2 2.8 1.4
1.4
0.0
0.0 0
100
200
300
400
0
500
50
100
T(°C)
x=60m
(c) 8.4
(d)
200
250
300
x=90m
8.4 7.0
vertical height (m)
vertical height (m)
7.0 reference 0.1H 0.2H 0.4H 0.6H 0.65H 0.7H 0.8H 1.0H
5.6 4.2 2.8 1.4 0.0
150 T(°C)
0
50
100
150 T (°C)
200
250
reference 0.1H 0.2H 0.4H 0.6H 0.65H 0.7H 0.8H 1.0H
5.6 4.2 2.8 1.4
300
0.0
0
50
100 T (°C)
Figure 9. Temperature distribution at different locations (Q ¼ 10 MW).
150
200
112
Journal of Fire Protection Engineering 22(2)
x=0m
8.4
(b)
vertical height (m)
x=30m
8.4 7.2
7.2
reference 0.1H 0.2H 0.3H 0.4H 0.6H 0.65H 0.7H 0.8H 1.0H
6.0 4.8 3.6 2.4 1.2
vertical height (m)
(a)
reference 0.1H 0.2H 0.3H 0.4H 0.6H 0.65H 0.7H 0.8H 1.0H
6.0 4.8 3.6 2.4 1.2 0.0
0.0 0
100
0
100 200 300 400 500 600 700 800 900
200
(c)
x=60m
8.4
(d)
400
500
x=90m 8.4
7.2
7.2 reference 0.1H 0.2H 0.3H 0.4H 0.6H 0.65H 0.7H 0.8H 1.0H
6.0 4.8 3.6 2.4 1.2
vertical height (m)
vertical height (m)
300 T(°C)
T(°C)
reference 0.1H 0.2H 0.3H 0.4H 0.6H 0.65H 0.7H 0.8H 1.0H
6.0 4.8 3.6 2.4 1.2 0.0
0.0 0
100
200
300
T(°C)
400
0
50
100
150
200
250
300
T(°C)
Figure 10. Temperature distribution at different locations (Q ¼ 20 MW).
In summary, the numerical results show that a 12 -m long model rescue station with three cross-passages and blocks with open heights of 0.65 H can give results similar to that for the real rescue station under the tested conditions.
Experimental procedure A total of 54 tests, including 33 tests without a model train placed inside and 21 tests with the train, are carried out in a 1:20 scale model rescue station.
Scaling The model used in the present study is built to a scale of 1:20, which means that the size of the tunnel is scaled geometrically according to this ratio. Scaling for the tests follows the widely used Froude modelling method. Clearly, it is impossible and not necessary to preserve simultaneously all the terms obtained through this modeling theory in model scale tests compared to full scale. However, the terms that are most important and most related to the study can be preserved. The thermal inertia of the involved material, turbulence intensity and radiation are not explicitly scaled,
Li et al.
113
Table 2. A list of scaling correlations for the model rescue station. Type of unit
Scaling
Equation number
Heat release rate (HRR) (kW) Velocity (m/s) Time (s) Energy (kJ) Mass (kg) Temperature (K)
QM =QF ¼ ðLM =LF Þ5=2 VM =VF ¼ ðLM =LF Þ1=2 tM =tF ¼ ðLM =LF Þ1=2 EM =EF ¼ ðLM =LF Þ3 Hc,F =Hc,M mM =mF ¼ ðLF =LM Þ3 TF =TM ¼ 1
(3) (4) (5) (6) (7) (8)
L is the length scale, index M is related to the model scale and index F to large scale (LM/LF ¼ 1/20 in the current study).
12000 3000
1500
1500
3000
3000
Fire source main tunnel
model train
F2
main tunnel
model train
F1 150
block
A
A Cross-section A-A
B
B Cross-section B-B
850
fire proof door
block
1250
air flow
Cross-passage 1
V
V
Cross-passage 3
4000
Cross-passage 2
V
fan
Figure 11. A schematic drawing of the model scale rescue station (dimensions: mm).
(a)
main tunnel
(b)
cross-section A-A
(c) cross-section B-B
175
218
393
block
hop 380
door 170µ112.5 200
300
450
Figure 12. Cross-sections of the model main tunnels and the cross-passages (dimensions: mm).
114
Journal of Fire Protection Engineering 22(2)
B
A 50mm 40mm
main tunnel
20mm
Pile A T1
Pile B T3 T4 T5 T6
T2
air flow cross-passage
door
T2 Pile A
T5
70 mm
10mm
T3 T4
B
A 25mm 45 mm
70 mm
T1
Pile B
10mm
Pile A
Gas analysis
thermocouple
T6 Pile B
Figure 13. Instrumentation beside the cross-passages (dimensions: mm).
but the HRR, the flow time, flow rates, the energy content and mass are scaled properly, as shown in Table 2. The scenario of a tunnel fire is quite different from an open fire or enclosure fire. The ratio of tunnel length to tunnel height should be great enough to scale a realistic tunnel fire. Therefore, it is very expensive to build a large-scale tunnel. In the authors’ opinion, the scaling ratio should not be smaller than about 1:20. Experience with model tunnel fire tests at this scale shows there is a good agreement between model scale and large scale on many focused issues [14,15,17–19]. This type of scaling is widely used in model tunnel fire tests all over the world. More information about scaling theories can be obtained from the examples in references [20,21].
Experimental setup The 1:20 model rescue station consists of a 12 -m long main tunnel and three crosspassages, as shown in Figure 11. The cross-sections of the model main tunnel and the cross-passages are shown in Figure 12. The model tunnel was made from stainless steel with a thickness of 1 mm covered by 23 mm of concrete. The crosspassage was made from stainless steel with a thickness of 1 mm. A stainless-steel door was set up at the junction between the model tunnel and the cross-passage, as shown in Figures 11 and 12. The doors are 0.17 m wide and 0.1125 m high, corresponding to 3.4 m wide and 2.25 m high at full scale. Two series of rescue station fire tests were carried out, including 33 tests without a model train and 21 tests with a train in the tunnel. The model train was 8 m
Li et al.
115
Table 3. Tests results in rescue station fire tests without model train inside. Model-scale critical flow rate (m3/h) Test no.
Ventilationa
Location of fire source
Open height, hop (mm)b
Full-scale Q (MW)
Cross 1
Cross 2
Cross 3
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 33
3 3 3 3 3 3 3 3 2+1 3 3 2+1 3 3 3 2+1 2+1 3 3 3 3 2+1 2+1 2+1 2+1 2+1 2+1 3 3 3 3 3 3
F1 F1 F1 F1 F1 F1 F1 F1 F1 F1 F1 F1 F1 F2 F2 F2 F2 F2 F2 F2 F2 F2 F2 F2 F1 F1 F1 F1 F1 F1 F2 F2 F2
393 300 220 140 85 40 closed + 140 closed + 85 closed + 85 closed + 140 closed + 85 closed + 85 85 85 closed + 85 closed + 85 closed + 85 closed + 85 closed + 85 closed + 85 closed + 85 closed + 85 closed + 85 closed + 85 closed + 85 closed + 85 closed + 85 85 85 85 85 85 85
10.5 10.5 10.5 10.5 10.5 10.5 10.5 10.5 10.5 20 20 20 20 10.5 10.5 10.5 20 20 5 15 25 5 15 25 5 15 25 5 15 25 5 15 25
– – 9.75 12.75 15.25 16.25 16.25 16.75 15.75 19.75 20.75 19.75 19.75 15.25 16.25 15.75 19.25 20.25 8.75 17.75 22.25 7.75 16.25 21.25 8.75 16.25 21.75 7.75 16.75 19.75 7.75 16.25 19.75
7.25 8.25 12.75 13.75 16.75 17.25 16.25 16.75 15.75 20.75 21.25 20.25 21.75 12.75 12.25 11.75 16.25 16.25 7.75 14.25 18.75 6.25 13.75 17.75 9.75 16.75 22.75 9.25 17.75 22.25 7.25 14.75 *
– – 9.75 12.25 14.25 14.75 13.25 13.25 – 16.25 16.25 – 17.75 12.75 12.0 – – 15.25 6.75 13.75 17.75 – – – – – – 7.75 16.25 19.25 6.75 13.25 *
a ‘3’ indicates 3 cross-passages with the same ventilation conditions, ‘2 + 1’ indicates cross-passage 1 and 2 with the same ventilation condition and cross-passage 3 with a constant volumetric flow rate of 30 m3/h. b ‘85’ corresponds an open height of 85 mm on the both sides of the rescue station, ‘closed + 85’ corresponds a open height of 85 mm at one exit and closed at the other exit. ‘–’ indicates that no smoke was observed in the cross-passage and ‘*’ indicates no measurement in the cross-passage.
116
Journal of Fire Protection Engineering 22(2)
Table 4. Tests results in rescue station fire tests with a model train inside. Model-scale critical flow rate (m3/h) Test no.
Ventilation
Location of fire source
Open height, hop (mm)
Full-scale Q (MW)
Cross 1
Cross 2
Cross 3
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
3 3 3 3 3 3 3 3 3 3 3 3 3 2+1 3 3 3 2+1 3 2+1 2+1
F1 F1 F1 F1 F1 F1 F1 F1 F1 F1 F2 F2 F2 F2 F2 F2 F2 F2 F2 F1 F1
85 closed + 85 85 closed + 85 140 closed + 140 30 393 140 closed + 140 140 closed + 140 closed + 85 closed + 85 85 85 closed + 85 closed + 85 closed + 140 closed + 85 closed + 85
10.5 10.5 20 20 10.5 10.5 10.5 10.5 20 20 10.5 10.5 10 10.5 10.5 20 20 20 20 10.5 20
12.75 13.75 17.25 17.75 12.75 13.75 13.25 – 16.25 17.75 12.75 12.75 12.75 13.25 12.75 17.25 19.75 19.25 19.25 13.25 18.75
13.75 13.75 17.75 16.75 13.75 13.75 14.25 6 16.75 17.75 11.75 10.25 10.75 9.25 11.25 15.25 15.75 13.25 15.75 13.25 18.75
11.75 11.25 16.75 13.75 11.75 11.75 12.75 – 14.75 15.25 10.75 9.75 9.75 – 10.75 15.25 13.25 – 13.75 – –
long, 0.2 m high and 0.15 m wide and placed 30 mm above the floor, as shown in Figure 11. The parameters tested were HRRs, fire source locations, open heights and ventilation conditions. According to the previous analysis, the blocks with open heights of 0.65 H give the best fit to the reference data. However, since computational error had to be taken into account and also as a safety precaution, the open heights used were mainly in a range of 0.2 H to 0.35 H during the model-scale tests. The fire source was either placed at F1 or F2, as shown in Figure 11, to analyze the effect of fire source location on the smoke control in a rescue station fire. The interval between the fire source locations was 1.5 m. The fire source was a 150 mm diameter porous bed burner with its top surface set at floor level. Propane was used as fuel, and its gas flow rate was metered by a rotameter with 1% accuracy.
Li et al.
117
0.32 Cross-passage 1(no train) Cross-passage 2(no train) Cross-passage 3(no train) Cross-passage 1(train) Cross-passage 2(train) Cross-passage 3(train)
Vcc (m/s)
0.24
0.16
0.08
0.00 0.0
0.2
0.4
0.6 h*
0.8
1.0
1.2
Figure 14. Effect of open height below the block (Q ¼ 5.9 kW).
Gas temperatures were measured by using K-type stainless steel-sheathed thermocouples with a diameter of 1.0 mm. The layout of thermocouples beside the cross-passages is given in Figure 13. Ceiling gas temperatures along the cross-passages were measured 10 mm below the centerline of the ceiling in the cross-passages. The interval between the ceiling thermocouples was 50 mm. Adjacent to each door of the cross-passages, two piles of thermocouples (Pile A and Pile B, see Figure 13) were placed to measure the gas temperature. Note that the lowest thermocouples at both piles, i.e. T2 and T6, were adjacent to the upper edge of the doors. During the tests, the doors were open. The ventilation flow rates through the cross-passage were measured by vortex flow meters with measurement range of 8–60 m3/h and 1% accuracy. The ventilation velocity across the door in a cross-passage was calculated by dividing the volumetric flow rate by the door’s cross-sectional area. During each test, the volumetric flow rate was gradually (by 0.5 m3/h) adjusted from high to low every measurement interval, until smoke ingress occurred. Gas was sampled inside the fireproof door (beside T3) to determine whether smoke ingress occurred in the cross-passage, see Figure 13. Since the design volume flow rate through each cross-passage should approximately be the same, the flow rate was so adjusted in most of the tests, except in some special tests, where the flow rate in cross-passage 1 was set as 30 m3/h. In practice, since the volume flow rate on one side of the rescue station could approach zero, this side of the tunnel was completely closed using a block to model this situation in some tests. To reduce experimental error, the average value of the volume flow rate when the smoke just spread into the cross-passage and the volume flow rate of the previous test (+0.5 m3/h) was considered as the critical flow rate. The estimated experimental errors for the measurement of critical velocities in cross-passages range from approximately 2% to 5% in the tests.
118
Journal of Fire Protection Engineering 22(2)
Vcc (closed+85, m/s)
0.4
0.3
no train, fire source F1 train, fire source F1 train, fire source F2 equal line y=x
0.2
0.1
0.0 0.0
0.1
0.2 0.3 Vcc (closed+140, m/s)
0.4
Figure 15. Effect of open height below the block with one closed end (Q ¼ 5.9 kW).
Test results Results are given in Tables 3 and 4 without and with a model train in the tunnel, respectively. Generally, the ventilation conditions in three cross-passages are the same, called uniform ventilation or ‘3’ in the tables. However, in some tests, the volumetric flow rate in cross-passage 3 was kept as a constant volume flow rate of 30 m3/h and meanwhile the tunnel exit close to the cross-passage 3 was closed, called non-uniform ventilation or ‘2+1’ in the tables. The two open heights below the blocks were the same in most tests except in some tests the tunnel exit close to cross-passage 3 was closed, called ‘closed +140/85’ in the tables. The HRR tested ranged from 2.8 kW to 14 kW, corresponding to 5 MW to 25 MW at full scale. Note that the HRRs of 5.9 kW and 11.2 kW at model scale corresponds to 10.5 MW and 20 MW at full scale, respectively.
Discussion of results In the following, the effect of different parameters on the critical velocities for smoke control in the cross-passages and the gas temperatures beside the doors were investigated.
Critical velocity for smoke control in the cross-passages Open height. The open height of block was analyzed here using the tests data. Figure 14 shows the critical velocity in the cross-passages as a function of the open height at a HRR of 5.9 kW at model scale, corresponding to 10.5 MW at full scale. It is shown that, when the model train was placed inside, the critical velocities in the cross-passages increases linearly with decreasing open height at a dimensionless open height above 0.2 and are almost the same below 0.2. When a model train
Li et al.
119
1.5
Cross-passage 1 Cross-passage 3 Cross-passage 1(train) Cross-passage 3(train)
reduction ratio η
1.2 0.9 0.6 0.3 0.0
0
1
2
3
4
5 6 test no.
7
8
9
10
Figure 16. The reduction ratio Z for fire sources located in front of the cross-passage (F1).
1.6
Cross-passage 1(closed end) Cross-passage 3(closed end) Cross-passage 1(train,closed end) Cross-passage 3(train,closed end)
1.4 reduction ratio η
1.2 1.0 0.8 0.6 0.4 0.2 0.0
0
1
2
3
4
5 6 test no.
7
8
9
10
Figure 17. The reduction ratio Z for fire sources located in front of the cross-passage (F1) with one closed end.
was placed inside, the same trend can be found, however, the transition point changes to about 0.35. In other words, under such conditions, the critical velocities in the cross-passages increase linearly with decreasing open height at a dimensionless open height above 0.35 and approaches a constant below 0.35. Note that the dimensionless open heights, hop ¼ hop =H, of 0.2 and 0.35 approximately correspond to block heights of 85 mm and 140 mm, respectively, in the model scale tests. Figure 14 does not include the results from tests with one exit closed. In these tests with one closed exit, an open height of 85 mm or 140 mm was used in the tests.
120
Journal of Fire Protection Engineering 22(2)
1.5
Cross-passage 2 Cross-passage 3 Cross-passage 2(train) Cross-passage 3(train)
reduction ratio, ξ
1.2 0.9 0.6 0.3 0.0
0
1
2
3 test no.
4
5
6
Figure 18. The reduction ratio x for fire sources placed between the cross-passages (F2).
Figure 15 shows the effect of open heights on the critical velocities in the crosspassages under these conditions. Clearly, the critical velocities in these cross-passages for these two open heights are nearly the same. Based on the previous analysis, it is clear that the critical velocities in the crosspassages are almost the same if the open height is below 85 mm, independent of the fire source location and whether one exit is closed. This correlates well with the conclusion from the numerical simulation results. Location of fire source relative to the cross-passages. It is shown in Tables 3 and 4 that the critical velocity in cross-passage 2 is the highest for fire sources placed in front of the cross-passage and the critical velocity in cross-passage 1 is the highest for fire sources placed between the cross-passages. To analyze the relation of critical velocities in these cross-passages, two location reduction ratios are defined: Z¼
Vcc, 1 or 3 Vcc, 2 or 3 ,x¼ Vcc, 2 Vcc, 1
ð9Þ
where Vcc,1 or 3 is the critical velocity in cross-passage 1 or 3, Vcc, 2 is the critical velocity in cross-passage 2, Z is the ratio of critical velocity in cross-passage 1 or 3 to that in cross-passage 2 for fire sources placed in front of the cross-passage, x is the ratio of critical velocity in cross-passage 2 or 3 to that in cross-passage 1 for fire sources placed between the cross-passages. Figure 16 gives the reduction ratio Z for fire sources located in front of the crosspassage 2. The tested HRRs are in the range of 2.8 to 14 kW at model scale, corresponding to 5 to 25 MW at full scale. It is shown in Figure 16 that for the fire sources placed in front of the cross-passage, the reduction ratio Z is about 0.9, regardless of whether the model train is placed inside or not. This suggests that the
Li et al.
121
1.4
Cross-passage 2(closed end) Cross-passage 3(closed end) Cross-passage 2(train,closed end) Cross-passage 3(train,closed end)
1.2
reduction ratio, ξ
1.0 0.8 0.6 0.4 0.2 0.0 0
1
2
3
4
5
6
7
8
9
10
11
test no.
Figure 19. The reduction ratio x for fire sources placed between the cross-passages (F2) with one closed end.
critical velocities in cross-passage 1 and 3 are the same and further, the critical velocity in the neighboring cross-passages is 90% of that in front of the crosspassage. Figure 17 gives the reduction ratio Z in tests with one closed end. It is shown that the reduction ratio Z is about 1.0 for cross-passage 1 and 0.8 for crosspassage 3 in the model rescue station with one closed end. This suggests that the critical velocities in cross-passage 1 and 2 are the same and the critical velocity in the upstream cross-passage, i.e. cross-passage 3, is smaller. Figure 18 shows the reduction ratio x for fire sources placed between the crosspassages under different test conditions. The HRRs tested are also in a range of 2.8 to 14 kW in model scale. It is shown that for fire source placed between the crosspassages, the average reduction ratio of cross-passage 2 is approximately 0.9 and that of cross-passage 3 is a little lower, i.e. about 0.85. Figure 19 shows the reduction ratio x in tests with one closed end. It is shown that the average reduction ratios of both upstream cross-passages are about 0.8 in the rescue station tests with one closed end. Based on the above study, it is shown that the reduction ratios of the neighboring two cross-passages were mainly in a range of 80% of 90%, regardless of the fire source location and existence of model train. Further, it can be expected that the critical velocity in a cross-passage decreases with increasing distance from the fire source. Ventilation condition. Comparing the critical volumetric flow rates under uniform ventilation and non-uniform ventilation conditions shows that the critical velocities in the cross-passages are lower under non-uniform ventilation conditions. This is mainly due to the higher ventilation which results in lower gas temperature. This means that the critical velocity decreases with the ventilation velocity across the fire. This is also verified by the previous study [5].
122
Journal of Fire Protection Engineering 22(2)
0.5 0.4
Cross-passage 1 Cross-passage 2 (worst cases) Cross-passage 3 1/3 power line, Eq. (12) 1/3 power line
* Vcc
0.3 0.2 0.1 0.0 0.00
0.05
0.10
0.15
0.20
Q*
Figure 20. Dimensionless critical velocities in the worst cases as a function of dimensionless heat release rate (no train).
The worst case. For smoke control in a rescue station, the key point is to determine the maximum critical velocity, i.e. the critical velocity in the worst cases. First, the worst cases needs to be identified. Based on the above analysis, the effect of open height below the blocks has no significant influence on the critical velocities in the cross-passages after the open height is below 85 mm. The critical velocity in the vicinity of the fire source is the highest. Comparing the critical velocities for fire sources located in front of crosspassage 2 and between the cross-passages shows that the former are higher. Further, the critical velocities decrease with the ventilation velocity. In other words, the worst case is when the fire source is placed in front of crosspassage 2 and the open height is about 85 mm. Heat release rate. In this section, the critical velocities in the worst cases are correlated with the HRRs. Based on the previous study [5], the dimensionless critical velocity is a function of dimensionless HRR, for a given fireproof door and tunnel. The dimensionless critical velocity for smoke control in a cross-passage is defined as: Vcc Vcc ¼ pffiffiffiffiffiffiffiffiffi gHd
ð10Þ
and the dimensionless HRR: Q ¼
Q ro cp To g1=2 H5=2
ð11Þ
Li et al.
123
critical velocity ratio ψ
2.0 Cross-passage 1 Cross-passage 2 Cross-passage 3 1/3 power line
1.6 1.2 0.8 0.4 0.0
0
4
8
12
16
20
Q (kW)
Figure 21. Critical velocities in all the tests without train as a function of the heat release rate.
Figure 20 shows the dimensionless critical velocities in the worst cases as a function of the HRR. The critical velocities in the worst cases can be expressed in the following form: Vcc ¼ 0:58Q1=3
ð12Þ
To fully utilize all the tests data, the data with a HRR of 5.9 kW are used as reference values and a critical velocity ratio is defined as: c¼
Vcc ðQÞ Vcc ðQ ¼ 5:9 kWÞ
ð13Þ
that is, the critical velocity ratio is the ratio of the critical velocity at a given HRR to that at Q ¼ 5.9 kW at model scale. Figure 21 plots all the critical velocity ratios in the worst cases as a function of the HRR. Since data from Q ¼ 5.9 kW are used as references, the curve passes through the point (5.87, 1). It is shown that all the data correlate relatively well with a 1/3 power law, following the same trend as shown in Figure 20. Train. The fire in a rescue station is generally related to an incident train. The existence of the train reduces the cross-sectional area of the tunnel and changes the flow pattern of hot gases and air flow. As a consequence, the train itself has a significant influence on the critical velocities for smoke control in the cross-passages.
124
Journal of Fire Protection Engineering 22(2)
1.2 1.0
φ
0.8 0.6 Cross-passage 1 Cross-passage 2 Cross-passage 3 φ =0.86
0.4 0.2 0.0 0
3
6 9 Different test no.
12
15
Figure 22. The critical velocity ratio due to train obstruction in different cross-passages.
To analyze the effect of a train on the critical velocities in different cross-passages, a critical velocity ratio due to obstruction is defined as the ratio of the critical velocity with a model train placed inside to the critical velocity without the model train, which can be expressed as follows: f¼
Vcc, tr Vcc, ntr
ð14Þ
where Vcc, tr and Vcc, ntr are the critical velocities with a model train placed inside and without, respectively. Figure 22 shows the critical velocity ratio due to train obstruction in tests with different cross-passages, fire source locations and ventilation conditions. Clearly, it is shown that the critical velocity ratio due to a train obstruction is in a range of 0.8 to 1.0. This means that the critical velocity decreases due to the obstruction by the vehicle. Further, it is shown that the average critical velocity ratio due to obstruction is about 0.86. This suggests the critical velocity due to the obstruction is reduced by about 14%. Note that the blockage ratio of the model vehicle, a, is about 20%. This suggests that the reduction ratio of critical velocity due to obstruction is a little lower than the blockage ratio of the vehicle, i.e. (1 a) < f < 1. There are two reasons for the lower critical velocity with a train inside. Firstly, the obstruction of the train reduces the cross-sectional area of the tunnel and thus the hot gases mainly are present in the upper layer above the train. Secondly, the existence of train carriage suppresses the entrainment of the fire plume and the temperature of hot gases in the upper layer increases. The gas
Li et al.
125
0.12 no train train Eq. (20)
2 Vcc
0.09
0.06
0.03
0.00 0.0
0.1
0.2
0.3 0.4 2gHdΔ Tcc/Tcc
0.5
0.6
Figure 23. Determination of critical smoke layer depth coefficient k.
temperature decreases more rapidly along the tunnel due to low mass flow rate of the hot gases.
Critical smoke layer temperature near the fireproof door The gas temperature and smoke depth in the vicinity of the fireproof door under critical conditions are important in estimating whether evacuees can escape through the doors. Here, a critical gas temperature is defined as the temperature of the smoke layer outside the fireproof door under the critical condition, i.e. at the critical velocity. The corresponding measuring points were placed immediately beside the fireproof door, e.g. thermocouple T2 in Figure 13. Under the critical condition, the thermal pressure of the smoke flow at the door height is just balanced by the dynamic pressure of fresh air through the door, which suggests 1 rcc gh ¼ ro V2cc 2
ð15Þ
Note that the smoke depth beside the door, h, is the vertical distance between the bottom of the smoke layer beside the door and the upper edge of the door. According to the ideal gas equation p ¼ rRT, equation (15) can be transformed into: Tcc V2cc ¼ Tcc 2gh
ð16Þ
126
Journal of Fire Protection Engineering 22(2)
120
Δ Tcc (°C)
90
no train trian Eq. (21)
60
30
0 0.0
0.1
0.2 * Vcc
0.3
0.4
Figure 24. Critical smoke temperatures as a function of critical velocities in the cross-passages.
Assume that the smoke depth beside the door, h, is proportional to the door height, i.e. h ¼ kHd , equation (16) can be transformed into: V2cc ¼ 2kgHd
Tcc Tcc
ð17Þ
or Tcc ¼
2kgHTo 2kTo ¼ 2kgH V2cc 2k V2 cc
ð18Þ
Note that Tcc ¼ Tcc To , so equation (18) is equivalent to: Tcc ¼
To V2 cc 2k V2 cc
ð19Þ
The critical smoke depth coefficient k can be determined based on the test data, as shown in Figure 23. The results show that all the data from tests with and without train correlate well with the proposed straight-line fit, which can be expressed as: V2cc ¼ 0:42gHd
Tcc Tcc
ð20Þ
A correction coefficient of 0.971 was found for equation (20). Comparing equations (17) and equation (20) shows that the critical smoke depth
Li et al.
127
100 Cross-passage 1 Cross-passage 2 Cross-passage 3
T (°C)
80 60 40 20 0 12
13
14 15 16 Volumetric flow rate q(m3/h)
17
18
Figure 25. The critical gas temperature and the volume flow rates in the cross-passages.
coefficient k equals 0.21, i.e. h ¼ 0:21Hd . Thus, equation (19) can be transformed into: Tcc ¼
To V2 cc 0:42 V2 cc
ð21Þ
Equation (21) suggests that the critical gas temperature can be calculated if the critical velocity in a cross-passage is known or can be estimated. Figure 24 shows the critical smoke temperatures as a function of critical velocities in the cross-passages. It is shown that all the data correlate well with equation (21). In addition, the critical gas temperature of the cross-passage increases rapidly with the dimensionless critical velocity. In general, the ventilation velocity through the fireproof doors in the crosspassages is a little greater than the critical velocity for smoke control in the cross-passages. Figure 25 gives the gas temperatures as a function of the ventilation flow rate in the three cross-passages. The flow rates in the cross-passages 1 to 3 are 15.25 m3/h, 16.75 m3/h and 15.25 m3/h, respectively. It is shown clearly that the gas temperature decreases very slowly with the increasing ventilation flow rate. In other words, the gas temperature is almost independent of the flow rate through the fireproof door. The reason is that the gas temperature outside the fireproof door mainly depends on the gas temperature inside the incident tunnel and the much smaller air flow in the cross-passages has only a little influence on it. This suggests that the critical gas temperature can be regarded as the characteristic gas temperature outside the fireproof door even if the actual ventilation velocity through the door is little greater or less than the critical value. In other words, the gas temperature outside the fireproof door can be calculated if the critical velocity for smoke control in the cross-passage is known.
128
Journal of Fire Protection Engineering 22(2)
Table 5. Fireproof door height and smoke layer height beside the door. Door height
Smoke layer height beside the door
m 1.8 2.0 2.2 2.4 2.6 2.8 3
m 1.42 1.58 1.74 1.90 2.05 2.21 2.37
Critical door height Here, the heights of the fireproof doors in the cross-passages are investigated based on the critical smoke layer depth coefficient. According to the preceding analysis, the smoke layer depth beside the door, i.e. the vertical distance between the bottom of the smoke layer beside the door and upper edge of the door, can be expressed as: h ¼ 0:21Hd
ð22Þ
This means that the smoke layer height beside the door and above the floor, Hd-h, is equivalent to 0.79 times door height. It is shown in Figure 24 that the gas temperature at the door height could be as high as 60 C. During the evacuation stage, the escape path should be free of smoke, especially beside the door. A height of 1.75 m could be a reasonable value for the heads of evacuees. This suggests that the height of the fireproof door should not be too low. Table 5 gives the relation between the fireproof door height and the smoke layer height beside the door. It is shown in Table 5 that the fireproof door height should not be lower than 2.2 m to ensure a smoke layer height above 1.74 m. From this view point, the fireproof door height should be as high as possible. However, in order to reduce the critical velocity for smoke control in the cross-passages, the fireproof door height should be as low as possible. As a consequence, a fireproof door height of 2.2 m is proposed as a reasonable compromise value. Based on the previous study [5], the width of the fireproof door has only insignificant influence on the critical velocity for smoke control in a cross-passage. It is also shown in equation (21) that the width of the fireproof door has no influence on the critical gas temperature of the fireproof door, since critical gas temperatures are correlated well with critical velocities. This suggests that increasing the width of the fireproof door is a good measure to increase the evacuation capacity.
Li et al.
129
Conclusions Use of solid blockage elements, or blocks, installed at both exits to model the thermal boundaries is proposed to carry out the model scale rescue station fire tests based on two series of simulations. The method is also verified using the model test results. Based on these findings, a total of 54 scale model tests are carried out to investigate smoke control issues in the rescue station fires. The effect of HRR, train obstruction, fire source location, ventilation condition and open height below blocks on smoke control of the cross-passages in a rescue station is investigated. The test results show that the critical velocity in the cross-passage beside a fire source is the highest, and the critical velocity in cross-passages decreases with distance away from the fire source. The critical velocities in the two neighboring crosspassages are about 80% to 90% of the maximum critical velocity. The critical velocities in the cross-passages vary as the 1/3 power law of the HRR. The critical velocity decreases due to the obstruction effect by a train. The average critical velocity ratio due to obstruction is about 0.86. This suggests the critical velocity due to the obstruction is reduced by about 14% on average, slightly lower than the blockage ratio of the vehicle. The critical gas temperature, i.e. smoke layer temperature beside the fireproof door under critical conditions, was investigated theoretically and experimentally. The critical gas temperature is correlated well with the critical velocities in the cross-passages. A simple equation, which correlates very well with the experimental data, is obtained to estimate the critical gas temperature. The results show that the gas temperature beside the door is not sensitive to the air velocity through the door and the critical gas temperature can be regarded as a characteristic gas temperature beside the door in a cross-passage. The smoke layer height beside the door under critical conditions was also investigated and it has been shown that the smoke layer height beside the door is directly related to the critical gas temperature. Based on these findings, a door height of 2.2 m is proposed as a reasonable value for crosspassages in rescue stations. Acknowledgements Acknowledgements to SP Tunnel and Underground Safety Centre for their support. We would also like to thank associate Professor Zhihao Xu and Zhihui Deng at Southwest Jiaotong University for their help in the tests and valuable comments to the work.
Funding This study was sponsored by the National Natural Science Foundation of China under Grant 51078315.
Nomenclature
cp ¼ heat capacity of air (kJ/kgK) D ¼ fire characteristic diameter (m)
130
Journal of Fire Protection Engineering 22(2)
E ¼ energy content (kJ) g ¼ Gravity acceleration (m/s2) H ¼ tunnel height (m) Hc ¼ heat of combustion (kJ/kg) Hd ¼ door height (m) h ¼ Vertical distance between smoke layer bottom and upper edge of the door (m) hop ¼ open height below the block (m) k ¼ critical smoke depth coefficient L ¼ length scale (m) m ¼ fuel mass (kg) P ¼ pressure (Pa) q ¼ Volumetric flow rate (m3/h) Q ¼ Heat release rate (kW) Q ¼ dimensionless heat release rate R ¼ ideal gas constant (J/molK) T ¼ gas temperature (K) To ¼ ambient temperature (K) Tcc ¼ critical gas temperature beside door (K) Tcc ¼ critical excess gas temperature beside the door (K) t ¼ time (s) V ¼ velocity (m/s) Vcc ¼ critical velocity in cross-passage (m/s) Vcc ¼ dimensionless critical velocity in cross-passage
Greek symbols o ¼ ambient density (kg/m3) cc ¼ Critical density difference of smoke beside the door (kg/m3) a ¼ blockage ratio Z ¼ location reduction ratio in equation (9) x ¼ location reduction ratio in equation (9) c ¼ critical velocity ratio in equation (13) f ¼ critical velocity ratio due to obstruction in equation (14)
Subscripts F ¼ full scale M ¼ model scale tr ¼ with train ntr ¼ without train
References 1. Gerber P. Quantitative risk assessment and risk-based design of the Gotthard base tunnel. In: Proceedings of the 4th International Conference Safety in Road and Rail Tunnels, April 2–6, Madrid, Spain, 2006, pp.395–404.
Li et al.
131
2. Kim DH and Park WH. Experiment by using reduced scale models for the fire safety of a rescue station in very long rail tunnel in Korea. Tunnel Underground Space Technol 2006; 21: 303. 3. Tarada F, Bopp R and Nyfeler S. Ventilation and risk control of the Young Dong Rail Tunnel in Korea. In: 1st international conference on major tunnel and infrastructure projects, May 22–24, Taibei, Taiwan, 2000. 4. Tarada F. Critical velocities for smoke control in tunnel cross-passages. In: 1st international conference on major tunnel and infrastructure projects, May 22–24, Taibei, Taiwan, 2000. 5. Li YZ, Lei B and Ingason H. Theoretical and experimental study of critical velocity for smoke control in a tunnel cross-passage. Fire Technol 2010, doi: 10.1007/s10694-0100170-0. 6. Ozawa S. Ventilation and fire countermeasure in Seikan tunnel. In: 6th aerodynamics and ventilation of vehicle tunnels, Sep 27–29, Durham, UK, 1988, pp.481–493. 7. Rudin C. Fires in long railway tunnels – the ventilation concepts adopted in the AlpTransit projects. In: 10th international symposium on aerodynamics and ventilation of vehicle tunnels, Nov 1–3, Boston, MA, 2000, pp.915–925. 8. Bassler A, Bopp R, Scherer O, et al. Ventilation of emergency station in the Koralm tunnel. In: 12th international symposium on aerodynamics and ventilation of vehicle tunnels, Jul 11–13, Portoroz, Slovenia, 2006. 9. Hilar M and Srb M. Long railway tunnels – comparison of major projects. In: WTC 2009, May 23–28, Budapest, Hungary, 2009. 10. McGrattan K, McDermott R, Hostikka S, et al. Fire dynamics simulator (Version 5), user’s guide. NIST Special Publication 1019-5. Gaithersburg, MD: National Institute of Standards and Technology, 2010. 11. Hu LH, Huo R, Peng W, et al. On the maximum smoke temperature under the ceiling in tunnel fires. Tunnel Underground Space Technol 2006; 21(6): 650–655. 12. Hwang CC and Edwards JC. The critical ventilation velocity in tunnel fires—a computer simulation. Fire Safety J 2005; 40: 213–244. 13. Li YZ and Lei B. Model of critical velocity in a tunnel cross-passage. J China Railway Soc 2008; 30(3): 87–90. 14. Li YZ, Lei B and Ingason H. The maximum temperature of buoyancy-driven smoke flow beneath the ceiling in tunnel fires. Fire Safety J 2011; 46: 204–210. 15. Li YZ, Lei B and Ingason H. Study of critical velocity and backlayering length in longitudinally ventilated tunnel fires. Fire Safety J 2010; 45: 361–370. 16. Ingason H. Design fires in tunnels. In: Proceedings of the 2nd International Symposium Safe & Reliable Tunnels – Innovative European Achievements, May 30–31, Lausanne, Swiss, 2006. 17. Li YZ and Ingason H. The fire growth rate in ventilated tunnel fires. In: Fire safety science – Proceedings of the 10th International Symposium, Jun 19–24, Maryland, USA, 2011, pp.347–358. 18. Ingason H and Li YZ. Model scale tunnel fire tests with longitudinal ventilation. Fire Safety J 2010; 45: 371–384. 19. Ingason H and Li YZ. Model scale tunnel fire tests with point extraction ventilation. J Fire Protect Eng 2011; 21(1): 5–36. 20. Quintiere JG. Scaling applications in fire research. Fire Safety J 1989; 15: 3–29. 21. Heskestad G. Physical modeling of fire. J Fire Flammability 1975; 6: 253–273.
