Keywords

33.1 Introduction

The tunnel structure has long-narrow and closed characteristics, fires can easily cause congestion and vehicle collisions, resulting in very serious consequences. Compared with the traditional water-based fire extinguishing system, the water mist system is atomized into small-sized droplets by spraying water through a special structure nozzle under a certain pressure, which can fill the space to achieve the effects of cooling, suffocation, and blocking radiant heat. At the same time, it can reduce impact on circuits, flammable and explosive materials, etc. Moreover, the water mist system has the advantages of fast heat absorption, no pollution and low water consumption. Therefore, the application of water mist system has attracted the attention of many scholars [1, 2].

The direct action on fire source with water system has been studied extensively by scholars. Chen et al. [3] conducted a 1:10 reduced-scale study referring to the experiment of Siemens and proposed that the establishment of a coupling system of water mist and ventilation could significantly improve the efficiency of fire extinguishing. Under longitudinal ventilation, the interaction between water mist and smoke has been carried out in a numerical simulation (Blanchard et al. [4]), and verified the effectiveness of water mist to absorb heat and radiation. The experimental results of Li et al. [5] in a small-scale tunnel showed that with longitudinal ventilation, water mist can effectively suppress the n-heptane pool fire, and the cooling effect was enhanced with the ventilation velocity. The experiments by Wang et al. [6] showed that the maximum smoke temperature decreased with the increase of the water mist flow, and the exponential equation of the longitudinal smoke temperature distribution was modified. Yang et al. [7] studied the variation of particle size on smoke migration with longitudinal velocity in road tunnel and indicated that there was an optimal water mist fire extinguishing particle size. Li et al. [8] found that the fire extinguishing efficiency of fixed fire extinguishing system can be improved to a certain extent by longitudinal ventilation, and the ventilation system should operate earlier than FFFS. Liu et al. [9] investigated the combined effect of a water mist system and longitudinal ventilation, the results showed that the internal tunnel temperature decreased significantly under longitudinal airflow, especially in the upstream area and near fire source. Magdolenova [10] carried out a numerical simulation of a 1/3 scale model and pointed out that the system with double-row nozzle arrangement was more stable.

There are relatively few studies on the situation that no direct contact between the fire source and the water mist. Sun et al. [11] indicated that when the wind speed reached the critical value, the water system will have less role in preventing smoke diffusion would be reduced, although the temperature decrease phenomenon in the downstream area still existed. A new system with simultaneous water mist screen and transverse ventilation system (WMSTV system) is proposed by Liang et al. [12], they found that the visibility in the closed area between the two water mist screens was low, but the temperature at 2 m above the ground was relatively low and would be suitable for personnel evacuation. Wang et al. [13] conducted a numerical simulation study on the coupling of longitudinal ventilation and water mist. Results showed that the downstream temperature decreased with the increase in the length of water mist zone, while the temperature did not change significantly with the water pressure. Li et al. [14] revealed that the water mist segment system should not be used in combination with longitudinal ventilation. In the study of Mehaddi et al. [15], they concluded that the smoke flow could not be prevented by the water curtain, but it could achieve greater radiation attenuation. Fan et al. [16] conducted a full-scale experiment in the railway tunnel rescue station, and the results showed that longitudinal ventilation could help enhance the visibility and reduce temperature. In the area covered by water mist, the dimensionless maximum temperature rise under longitudinal ventilation was lower than that without it.

Based on the above analysis, the research of water mist screen in tunnels is mainly based on long and straight tunnels. There is no specific research result on the application of water mist screen to bifurcated tunnels. Compared with the long straight tunnel, the bifurcated tunnel is more complex and traffic accidents are relatively frequent at the fork. This paper mainly studies the smoke temperature and the CO concentration distributions in bifurcated tunnel fire by changing heat release rate of fire source, longitudinal ventilation velocity, number of row of water mist nozzle and particle size of water mist.

33.2 Numerical Simulation

33.2.1 Numerical Model

Based on the similarity criterion [17, 18], a 1:10 bifurcation tunnel numerical model is established in this study. In Fig. 33.1, the dimension of the main tunnel is 10 m (length) × 1.5 m (width) × 0.6 m (height), and the branch is 5 m (length) × 1 m (width) × 0.6 m (height). The angle between the main tunnel and the branch tunnel is 45°. A series of thermocouples and carbon monoxide measuring points are arranged longitudinally along the center of the tunnel ceiling. A fire source with a size of 0.2 m × 0.2 m × 0.1 m is set at the center of the main tunnel, and propane is used as fuel. The longitudinal ventilation is set on the left side of the main tunnel. The main tunnel takes the fire source center as the origin OA and the branch tunnel takes the fork center as the origin OB. The tunnel is provided with three sections of water mist screen, each section has three rows of water mist nozzles with an interval of 0.5 m. The upstream water mist screen section of the main tunnel is set at −2 to −3 m, the downstream water mist screen section of the main tunnel is set at 2–3 m, and the water mist screen of the branch is set at 1.5–2.5 m. The row of water mist closest to the fire source is defined as the first row, and so on. During the process of simulation, longitudinal ventilation is activated firstly, the same number of row of water mist nozzle are activated at the same time in the main road and branch at 150 s, and the water mist system stops working at 300 s.

Fig. 33.1
A diagram of the top view of the main tunnel, provided with a branch tunnel, at an angle of 45 degrees with the dimensions marked. The longitudinal ventilation, thermocouple, and carbon monoxide measuring points, as well as the location of the water mist nozzles, are indicated.

Top view diagram of numerical model

Full size image

33.2.2 Simulation Conditions

Table 33.1 shows the details of the simulation conditions, considering the heat release rates of fire source (HRR), the longitudinal ventilation velocity (v), the number of row of water mist nozzle and the particle size of water mist (D) as variables. A total of 60 working conditions are simulated.

Table 33.1 Summary of the simulation conditions
Full size table

33.2.3 Grid Independence Analysis

The setting basis of grid size used in numerical simulation of this study is as follow [19]:

$$D^{*} = \left( {\frac{HRR}{{\rho_{\infty } c_{p} T_{\infty } \sqrt g }}} \right)^{\frac{2}{5}}$$
(33.1)
$$4 < \frac{{D^{*} }}{\delta x} < 16$$
(33.2)

where \(D^{*}\) is the characteristic diameter of fire source, \(c_{p}\) is the specific heat capacity of ambient air (kJ/(kg K)), \(T_{\infty }\) is the ambient air temperature (K), \(\delta_{x}\) is the grid size, \(\rho_{\infty }\) is the ambient air density (kg/m3).

When the heat release rate of fire source is 31.62 kW, the calculated grid size is between 0.01 and 0.06 m. When the heat release rate of fire source is 63.25 kW, the calculated grid size is between 0.02 and 0.08 m. As shown in Fig. 33.2, the grid sizes of 0.04, 0.05 and 0.06 m are selected for temperature distribution analysis beneath the tunnel ceiling. Taking the accuracy and calculation time into consideration, a multi-grid method is used for numerical simulation, i.e., the grid size within 1 m2 area of the fire source is 0.025 m × 0.025 m × 0.025 m, and the grid size in other areas is 0.05 m × 0.05 m × 0.05 m.

Fig. 33.2
A line graph plots temperature versus distance from the fire source for grid sizes, 0.04 by 0.04 by 0.04, 0.05 by 0.05 by 0.05, and 0.06 by 0.06 by 0.06 meters. All lines have an increasing trend on the upstream side from negative 5 to the fire source at 0 and start decreasing on the downstream side.

Grid adaptability curve

Full size image

33.3 Results and Discussions

33.3.1 The Heat Insulation and Smoke Control Effect of the Main Tunnel

33.3.1.1 Temperature Distribution

Figure 33.3 shows the ceiling temperature variation curves on both sides of the water mist screen section in the main tunnel (i.e., at the positions of ±1 and ±4 m away from the fire source). It shows that the water mist nozzles are activated at 150 s and reach stability at about 200 s. After the water mist is sprayed into the fire scene, the droplets evaporate rapidly, which has a strong effect of gasification and cooling. The effect of different number of water mist nozzle rows on the upstream area near the fire source is not obvious in Fig. 33.3a. At 4 m upstream from the fire source (i.e., the outside area of the water mist screen section), the initial ceiling temperature is low due to the influence of longitudinal ventilation. After activating the water mist system, the ceiling temperature drops sharply to ambient temperature, and the cooling effect reaches the best under all the different number of water mist nozzle rows. Due to the existence of water mist screen, the heat transfer at 1 m upstream from the fire source (i.e., the area near the fire source) is inhibited. For the downstream area, the cooling effect of water mist screen section is significant at 4 m downstream from the fire source. When the 3 rows of water mist nozzles come into play, the ceiling temperature at 4 m downstream rapidly decreases to close to the ambient temperature. Increasing the number of water mist nozzle rows is equivalent to increasing the number of droplets, which has a better cooling effect, seen in Fig. 33.3b, d. The ceiling temperature at 1 m downstream from the fire also increases slightly, which is similar with the result of 1 m upstream from the fire.

Fig. 33.3
4 line graphs of temperature versus time for heat release rates of 31.62 and 63.25 kilowatts for upstream and downstream with 8 lines each. Ceiling temperatures have an initial steep increase that stabilizes with fluctuations. Water mist activation is at 150 with steady-state duration after.

Ceiling temperature variations on both sides of the water mist screen section in the main tunnel (±1 and ±4 m away from the fire source)

Full size image

Figure 33.4 shows the ceiling temperature variation along the main tunnel with different number of nozzle rows activated. It can be observed that, the coupling effect of longitudinal ventilation and water mist has a significant influence on the ceiling temperature in the upstream area. The increase of wind speed inhibits the smoke diffusion in the upstream area. For the same wind speed, no matter how many rows of water mist nozzles act, the ceiling temperature drops to ambient temperature at 2 m upstream away from the fire source. As the wind speed and the number of water mist nozzle rows increase, the sharp drop point of ceiling temperature is closer to the center of the fire source. The ceiling temperature variation of the downstream area is basically consistent with that of the upstream area. The wind speed reduces the maximum ceiling temperature. For the same wind speed, the ceiling temperature decreases obviously after the water mist nozzles are activated. The more the number of water mist nozzle rows is, the greater the reduction of ceiling temperature. The coupling effect of longitudinal ventilation and water mist on the ceiling temperature in the downstream area is not obvious.

Fig. 33.4
3 line graphs of temperature versus distance from fire source at 0.3, 0.4, and 0.5 meters per second wind speed with lines for 0, 1, 2, and 3 rows. All lines have an increasing trend on the upstream side, which starts decreasing on the downstream side. A vertical line at 0 indicates the fire source.

Ceiling temperature along the main tunnel with different water mist nozzle rows

Full size image

In order to study the influence of longitudinal ventilation velocity and particle size of water mist on the ceiling temperature, the stable ceiling temperature variation along the downstream of main tunnel with the same number of water mist nozzle rows activated is shown in Fig. 33.5. It can be seen that the variation of wind speed mainly affects the ceiling temperature in the area near the fire source. The higher the wind speed, the lower the ceiling temperature near the fire source. Under the same wind speed, the smaller the particle size of water mist causes the more droplets and the larger specific surface areas, which can obtain the better cooling effect on ceiling temperature in the outside area of the water mist screen section.

Fig. 33.5
3 line graphs of temperature versus distance from fire source for 1, 2, and 3 rows of water mist nozzles for ventilation velocities of 0.3, 0.4, and 0.5 meters per second with particle sizes of 50, 100, and 200 microns. All the lines have an initial spike followed by a downward trend.

Influence of longitudinal ventilation velocity and particle size of water mist on the ceiling temperature along the downstream of main tunnel

Full size image

Figure 33.6 visually shows the typical temperature flow field distribution in the downstream of main tunnel at 200 s when different number of water mist nozzle rows activated. The influence of the water mist section is strengthened with the increasing rows of water mist nozzles, more smoke are captured and cooled down. Therefore, the ceiling temperature in the area near the fire source increases, while the ceiling temperature in the outside area of the water mist screen section decreases, which is consistent with the previous study by Wang et al. [13].

Fig. 33.6
Three spectral graphs a to c of the temperature distribution for one, two, and three rows of water mist nozzles. The maximum intensity of 170 at the left reduces to medium range of 93 as it approaches the mist, which further reduce to 46.8 in 1 or 2 rows and to 16.8 in 3 rows of nozzles downstream.

Typical temperature flow field distribution in the downstream of main tunnel with different number of water mist nozzle rows activated (t = 200 s)

Full size image

33.3.1.2 CO Concentration

Figure 33.7 shows the ceiling CO concentration distribution along the main tunnel during the stable period. It shows that the water mist screen section has the obvious blocking effect on upstream ceiling CO concentration. When the wind speed is 0.3 m/s, activating 1 row of water mist nozzles can reduce upstream ceiling CO concentration to a certain extent, and the upstream ceiling CO concentration in the outside area of the water mist screen section is basically blocked when 2 rows or 3 rows of water mist nozzles activated. As the wind speed increase to 0.4 and 0.5 m/s, the upstream ceiling CO concentration is completely blocked by water mist screen section in the area near the fire source. It can also be seen that the downstream ceiling CO concentration decreases under the action of water mist screen, and increasing the number of water mist nozzle rows has a better effect on reducing downstream ceiling CO concentration.

Fig. 33.7
Three line graphs from a to c plot C O concentration versus distance from fire source for 0.3, 0.4, and 0.5 meters per second wind speeds for 0, 1, 2, and 3 rows with a particle size of 50 microns. All lines increase on the upstream, which start decreasing downstream. Fire source is at 0.

Ceiling CO concentration distribution along the main tunnel with different number of water mist nozzle rows activated

Full size image

The ceiling CO concentration variation with time at 4 m downstream from the fire source of main tunnel is plotted in Fig. 33.8. It can be seen that, the downstream ceiling CO concentration decreases rapidly and remains in a stable fluctuation state after activating the water mist system. When 1 row of water mist nozzles are activated, the smaller the particle size of water mist, the lower the ceiling CO concentration. For all the wind speeds, the particle size of water mist of 200 μm has the worst reducing effect on ceiling CO concentration. The main mechanism of the downward movement of smoke is caused by the drag force from the water droplets [20]. When the 3 rows of water mist nozzles are activated, due to the large amount of droplets, the particle sizes of water mist of 50 and 100 μm have the similar reducing effect on ceiling CO concentration.

Fig. 33.8
2 line graphs of C O concentration versus time for 1 and 3 rows of water mist nozzles for wind velocities of 0.3, 0.4, and 0.5 meters per second with particle sizes of 50, 100, and 200 microns. All lines rise steeply, move horizontally, drop after water mist activation at 150 seconds, and stabilize.

Ceiling CO concentration variation with time at 4 m downstream in main tunnel

Full size image

33.3.2 The Heat Insulation and Smoke Control Effect of the Branch Tunnel

33.3.2.1 Temperature Distribution

According to the above analysis, the coupling effect of longitudinal ventilation and water mist has a significant influence on the upstream ceiling temperature of the main tunnel, but the variation of wind speed has no obvious effect on the downstream ceiling temperature of the main tunnel. It should be noted that due to the bifurcation structure, the wind speed entering the branch tunnel is smaller than the downstream of the main tunnel, so the working condition under the wind speed of 0.5 m/s is selected to analyze the ceiling temperature in the branch.

As shown in Fig. 33.9, under different heat release rates of fire source, the ceiling temperature along the branch tunnel shows a similar variation trend. When there is no water mist, the ceiling temperature along the branch tunnel is exponentially attenuated. After activating the water mist screen, the ceiling temperature drops sharply in the outside region of the water mist section. When 2 rows of water mist nozzles are activated, the ceiling temperature can be reduced to close to the ambient temperature. The cooling effect on ceiling temperature when 3 rows of water mist nozzles activated is almost the same as that of 2 rows.

Fig. 33.9
Two line graphs a and b plot temperature versus distance from the center of the branch tunnel entrance for heat release rates of 31.62 and 63.25 kilowatts, respectively. The lines are plotted for 0, 1, 2, and 3 rows. Both graphs have a downward concave-up trend.

Influence of the number of water mist nozzle rows on ceiling temperature distribution along the branch

Full size image

To verify the influence of particle size of water mist on ceiling temperature, the ceiling temperature variations with time at 3.5 m downstream from the entrance of branch tunnel (i.e., the outside area of the water mist screen section) are compared, as shown in Fig. 33.10. Under the condition that only 1 row of water mist nozzles are activated, the smaller the particle size of water mist results in the lower the ceiling temperature, which further confirms the results obtained in downstream area of the main tunnel. When 2 rows of water mist nozzles are activated, all the three particle sizes of water mist can rapidly reduce the ceiling temperature to the ambient temperature, the cooling effect is already acceptable. At this time, further increasing the number of water mist nozzle rows has no obvious change in the cooling effect on ceiling temperature.

Fig. 33.10
A line graph plots temperature versus time. The lines are plotted for rows 1 and 2 with particle sizes of 50, 100, and 200 microns. All lines rise steeply, move horizontally between 30 and 150 seconds, drop after water mist activation at 150 seconds, and stabilize.

Ceiling temperature variation with time at 3.5 m downstream from the entrance of branch tunnel

Full size image

33.3.2.2 CO Concentration

Similar to the ceiling temperature analysis of the branch tunnel, the working condition under the wind speed of 0.5 m/s is selected to analyze the ceiling CO concentration in the branch, seen in Fig. 33.11. After activating the water mist nozzle, the CO concentration near the fire source area increased compared with the original. At the same time, when only one row of nozzles is activated, the farther away from the branch tunnel entrance (i.e., the outside area of the water mist screen section), the more obvious CO concentration decreases along the way. Under this condition, there is little difference in the effect of activating two or three rows of water mist. Activating more nozzles prevents more smoke diffusion than activating a single row of nozzles. As the increase in the number of water mist nozzle rows, the amount of droplets also increases, so more smoke will move downward by the drag force from the droplets, resulting in the better effect on reducing ceiling CO concentration.

Fig. 33.11
A line graph plots C O concentration versus distance from the center of the branch tunnel entrance. The lines are plotted for 0, 1, 2, and 3 rows. Th lines have a downward trend. The location of the first nozzle is indicated at approximately 1.5 meters.

Ceiling CO concentration distribution along the branch tunnel with different number of water mist nozzle rows activated

Full size image

To study the influence of particle size of water mist on the ceiling CO concentration, the ceiling CO concentration variation at 1 and 3.5 m downstream from the entrance of branch tunnel with the same number of water mist nozzle rows activated is shown in Fig. 33.12. When only 1 row of water mist nozzles are activated, the three different particle sizes of water mist have the similar blocking effect on ceiling CO concentration. When the particle size was 50 μm, the inhibition effect was better, seen in Fig. 33.12a. As the increase in the number of water mist nozzle rows, reducing effect on ceiling CO concentration of the three particle sizes of water mist begins to differ, and the recovery speed is accelerated. Therefore, compared with the main tunnel, the water mist screen section of branch tunnel has the limited ability to block the smoke diffusion.

Fig. 33.12
Two line graphs a and b plot C O concentration versus time for 1 and 2 rows of water mist nozzles. The lines are plotted for distances of 1 and 3.5 meters with particle sizes of 50, 100, and 200 microns. Both graphs have an initial spike, which stabilizes with minor fluctuations.

Influence of the particle size on the ceiling CO concentration at 1 and 3.5 m downstream from the entrance of branch tunnel

Full size image

33.4 Conclusions

In this paper, simulation are carried out to study the heat insulation and smoke control effect of water mist screen under the action of longitudinal ventilation in bifurcated tunnel fire. The major conclusions are:

  1. (1)

    Under the action of water mist system, the longitudinal ventilation has a wonderful cooling effect on the ceiling temperature in the upstream area of main tunnel, but the variation of wind speed has no obvious effect on the downstream ceiling temperature of main tunnel and the ceiling temperature of branch tunnel.

  2. (2)

    For the ceiling temperature of the main tunnel and branch, the more the number of water mist nozzle rows, the lower the ceiling temperature in the outside area of the water mist screen section, and the better the cooling and heat insulation effect. In general, the smaller particle size also has a better cooling and heat insulation effect.

  3. (3)

    In the main tunnel, the smaller the particle size of water mist, the lower the ceiling CO concentration. The coupling effect of longitudinal ventilation and water mist on the upstream ceiling CO concentration is obvious, increasing the number of water mist nozzle rows or the wind speed can effectively block the upstream smoke diffusion and reduce the upstream CO concentration. The downstream ceiling CO concentration is less affected by the variation of wind speed, and increasing the number of water mist nozzle rows has a better effect on reducing downstream ceiling CO concentration.

  4. (4)

    In the branch, due to the influence of the bifurcation structure, the wind speed entering the branch tunnel is smaller and approximates natural ventilation. After activating the water mist system, the ceiling CO concentration decreases firstly and then begins to increase. The more number of water mist nozzle rows results in the better effect on reducing ceiling concentration. Compared with the main tunnel, the water mist screen section of branch tunnel has the limited ability to block the smoke diffusion.