Fire Dynamics Simulator Analysis of Smoke Confinement and Radiation Shielding using Water Mist Curtains

Article information

Abstract

1. Introduction

Fire can be defined as a large-scale combustion reaction that is not appropriately controlled, the results of the reaction capable of causing damage to the surroundings in the form of combustion heat, combustion products, and flames. The radiation emitted from the flames heats surrounding combustibles, increases the surface temperature, and can ignite and expand the fire. At times, it can cause direct harm to people. The smoke that is propagated to surroundings along with a buoyant plume flow due to the fire may include toxic combustion gases and soot. Heavy smoke that includes considerable volumes of soot reduces visibility and can send evacuees into a state of panic, magnifying fire damage by inhibiting efficient evacuation, firefighting, and rescue operations [1].

Water-based fire protection systems suppress fires through cooling of the flame by water and suffocation of it by evaporation and expansion of the water. Additionally, the fine droplets generated by sprays absorb or scatter the radiation emitted from flames, blocking the heat transfer, and limiting the impact of fire by dissolving and adsorbing the soot and harmful components. Moreover, water mist systems use less water compared to existing sprinklers, forming much finer droplets so that the effect of radiation heat shielding and smoke confinement can be maximized. Consequently, water mist systems can be applied in a greater variety of fire scenarios than sprinkler systems—for example, they have been developed for firefighting applications in cultural-property facilities that are required to minimize water damage or facilities with high explosion risks [2-5].

Accordingly, various research has been conducted on thermal-radiation shielding and the effect of controlling smoke using water-mist or water-curtain systems. Dembele et al. (2001) conducted experiments to identify the impact of the flow rate, atomization pressure, and water-mist droplet size on thermal-radiation shielding [6], the relationship between thermal-radiation shielding, atomization pressure, and optical thickness based on the analysis of the thermalradiation spectrum transmission rate being studied. Dombrovsky and Dembele (2022) considered a water curtain arranged in a double layer to improve the thermal-radiation shielding effect and derived a 2D-analytical model that reflected the optical thickness of the water curtain, the absorption rate, and the scattering rate based on the velocity and size of the droplets to predict the attenuation of thermal radiation [7]. Their model required less time and computational overhead compared to 3D computational fluid dynamics (CFD) simulations, making it applicable to real-time analysis of full-scale fires. Zhu et al. (2015) analyzed the thermal radiation attenuation effect of water-curtain nozzles under full-scale fire conditions using diesel oil [8]. For this study, a water-curtain nozzle complex comprised nine water-mist injectors, with various data—including the droplet characteristics, thermal radiation flow rate, attenuation rate, and temperature distribution based on various pressurization conditions—being provided. You and Ryou (2016) conducted experiments to identify the impact of the droplet size characteristics that constitute the water mist and the optical thickness on the reduction of radiant intensity. A correlation formula of the optical thickness based on the water-mist load was proposed, which could be used in the calculation of radiant intensity [9]. Jo and Lee (2021) used a twin-fluid nozzle supplied with water and air to perform experimental research on the attenuation of radiation heat, the optimal conditions for the attenuation effect being studied using the results based on the flow ratio between air and water and the average droplet size [10].

Fire analysis using CFD can provide precise data on the behavior of flame and smoke, fire extinguishment, and heat and combustion-gas generation and propagation, so it is widely used in fire engineering. In particular, the fire dynamics simulator (FDS) developed by the National Institute of Standard and Technology (NIST) in the United States [11] is a representative analytical tool which has been broadly applied in the analysis of radiation-heat shielding using water-mist droplets and the consequent smoke-confinement phenomenon. Mehddi et al. (2020) conducted experiments and performed numerical analysis to identify the radiation and smoke inhibition effect in tunnel fires, from which they found that water-mist curtains effectively shielded the expansion of radiation heat and smoke [12]. However, they pointed out that as the smoke that traveled along the tunnel ceiling mixed with inflowing air due to the spraying of the water mist before descending, the smoke concentration at the bottom of the tunnel increased.

Ko (2020) and Jo and Lee (2021) conducted FDS analysis for quantitatively analyzing the effect of attenuation as the radiation heat emitted from high-temperature panels passed through the water mist, presenting attenuation-rate data based on the droplet size, flow rate, and spraying angle, amongst others [13,14]. However, when considering the case of high-temperature panels as the source of radiation heat, the plume occurring from circulating fire and smoke did not exist, so the complex circulation characteristics that occur in actual fires could not be accounted for.

In this research, numerical analysis using FDS was conducted to analyze the smoke confinement characteristics and radiation shielding due to a water-mist curtain in full-scale fires. The spraying location of the water-mist curtain, radiation-heat attenuation rate based on the droplet size, and optical-density distribution were reviewed, the fire and smoke propagation characteristics being studied based on these findings.

2. Methods: Simulation Method

2.2 Simulation condition

The analysis domain selected in this research is as shown in Figure 1, where the corridor space has a length of 10 m, width of 3 m, and height of 9 m.

Figure 1

Schematic of the simulation domain.

The upper plane of the vertical space is open, the remaining planes being treated as walls. The 1 x 1 m heptane fire source is located at the 2.5-m point within the corridor space, the size of the fire being set to 500 kW. The water mist curtain sprays through two nozzle complex installed at 1-m intervals on the ceiling in the width direction. The location of the length direction (x) of the water-mist curtain varies from 6, 8, to 10 m, the impact of each being studied. The nozzle complex comprises a set of nozzles where nine direct-spray injectors are arranged at even angles on a semicircular circumference of 0.05 m, the water curtain being formed using a 32 L⋅min-1 flow rate through the two nozzle complex. Each nozzle has 30° spraying angle, the average size of the droplets sprayed being 200, 500, and 1,000 μm for the purposes of analysis. The grid size (δx) used in the analysis is set to 0.05 m to determine the initial plume flow and the combustion in the section that includes the fire source. In the downstream smoke flow area, the grid size is set to 0.1 m.

The ratio of the dimensionless length to the grid size(D*/δx), which was deemed as the FDS grid setting standard [15], was 14.4 and suitable. The total number of grids used in the grid system applied in this research was 486,000.

3. Results and Discussion

In this research, numerical analysis was conducted to analyze the smoke confinement and radiant-heat shielding characteristics due to a water mist curtain during the plume flow propagation process which occurred at the fire source and propagated through the ceiling. The flow field based on the location of the spray, the size of the water-mist droplet, the temperature distribution, and the soot distribution was reviewed, and the optical density was extracted at a height of 1.5 m after which the thermal-radiation attenuation rate was analyzed.

Figure 2 shows the average value during the 40-50 s of the velocity and temperature distribution calculation results at the intermediate cross-section (y = 1.5 m) of the analysis area.

Figure 2

Predictions of the velocity vector field and temperature distribution averaged during t = 40 to 50 s in the middle section (y = 1.5 m) for (a) no water-mist case and (b) water mist of D_m = 200 μm case.

Figure 2(a) shows the characteristics of the general compartment fire where water-mist operation is not involved. The plume that ascends in the fire source collides with the ceiling and propagates downstream. The ceiling jet flow forms a stable layer, moves along the corridor space, collides with the vertical wall, and ascends again, the temperature distribution being similar to that of the flow field. The high-temperature ceiling jet flow that moves along the corridor ceiling and the cold air inflow at the bottom toward the fire source are clearly evident.

Figure 2(b) shows the temperature and velocity distributions for the case of water-mist operation at x_inj = 8 m, the average diameter of the sprayed droplets being 200 μm. The jet flow that streams along the ceiling meets the water curtain, where its movement is inhibited, exhibiting a downward curve. Owing to the interaction of the jet flow and the spray flow, the flow layer is disturbed, and a change to a more complex distribution is evident. Additionally, rapid cooling can be identified as the high-temperature smoke (250 °C or above) mixes with the low-temperature spray flow.

Figure 3 shows the soot distribution analysis results for the same conditions shown in Figure 2.

Figure 3

Predictions of the soot mass-fraction distribution averaged during t = 40 to 50 s in the middle section (y = 1.5 m) for (a) no water-mist case and (b) water mist of D_m = 200 μm case.

Figure 3(a) shows the state of no operation of the water mist, exhibiting a similar distribution to the temperature distribution of Figure 2(a). Thus, in the corridor space, the highest soot concentration is at the ceiling, gradually decreasing downstream. In the vertical space, a concentration distribution—where the upper and lower layers gradually decrease—is evident. In conclusion, it is clear that the temperature distribution and soot concentration are determined by the movement of the plume flow generated from the fire source without the operation of the water mist.

Figure 3(b) shows the soot distribution when the water curtain is operational. The soot generated from the fire source moves as it is carried into the plume flow and meets the water-mist curtain. It is evident that its horizontal distribution is inhibited. However, the soot distribution as shown in Figure 3(b) displays a different tendency when compared to the temperature distribution of Figure 2(b)—that is, it is evident that the soot moves downwards owing to the spray flow, moving toward the fire source with the cold air flow, and being distributed evenly across entire the corridor space. From these results, it can be inferred that the soot distribution is more sensitive to the air-flow field than is the temperature distribution.

Figures 4(a) and 4(b) shows the results for the velocity and temperature distributions when the spray location is identical at x_inj = 6 m, the average droplet size being at D_m = 200 μm and D_m = 1,000 μm, respectively. Figure 4(a) shows that the ceiling jet flow meets the D_m = 200 μm water curtain, the velocity rapidly decreases and changes direction (downwards), exhibiting a vortex flow form that returns toward the fire source. As the high-temperature plume flow mixes with the cold-water curtain droplets, it cools rapidly, the high-temperature part being restricted to near the water curtain. Conversely, Figure 4(b) (with D_m = 1,000 μm) shows that despite the operation of the water-mist curtain, it does not greatly impact the progress of the jet flow, and a high-velocity jet flow is maintained at the lower layers of the ceiling. The hot air currents traveling toward the vertical space along the ceiling collide with the wall and move upwards, escaping through an opening in the ceiling. Additionally, it is evident that the high temperature state at the ceiling formed along the temperature distribution flow continues to the downstream areas. Given these results, it can be considered that the smoke-flow inhibition effect will be greater with smaller droplet sizes.

Figure 4

Predictions of the velocity vector field and temperature distribution in the middle section for the cases of water mist location of x_inj = 6 m; (a) D_m = 200 μm and (b) D_m = 1000 μm.

Figure 5 shows the soot distribution analysis results for the same conditions as Figure 4. When compared to the case of D_m = 1,000 μm, as shown in Figure 5(b), D_m = 200 μm in Figure 5(a) shows that the soot distribution is greatly limited. However, when compared to the temperature distribution shown in Figure 4(a), a major part of the soot propagates downward from the location of the water curtain. As mentioned above, the cooling effect owing to the water curtain can be considered to be greater than the smoke confinement effect. Through Figures 4 and 5, it is evident that the size of the droplets comprising the water curtain greatly impact the plume flow and smoke confinement. With relatively samll droplet spraying, the initial momentum can be easily lost, so the tendency to remain near the ceiling is high, the surface area is greater, facilitating heat absorption and increasing cooling. Conversely, high momentum can be gained if the droplet size is large, but the mist passes through the upper smoke layer at high velocity, so continuous impact is not considered possible.

Figure 5

Predictions of the soot mass fraction distribution in the middle section for the cases of water mist location of x_inj = 6 m; (a) D_m = 200 μm and (b) D_m = 1000 μm.

Figures 6(a) and 6(b) show the analytical results for the velocity field and soot distribution when the average diameter of the spray droplet is constant at D_m = 200 μm, the location of the water-mist curtain being x_inj = 8 and x_inj = 10 m, respectively. In the case of the velocity field, when analyzed with the velocity results as shown in Figure 4(a), it is evident that downstream movement of the location where the velocity change occurs can be observed based on the location of the water curtain. However, the tendency of the spray flow and smoke flow that occur near the water curtain to interact is similar regardless of the spraying location. Consequently, owing to interaction with the water curtain, the flow moving downstream meets the floor plane and forms a vortex as it moves toward the fire source, the size of the vortex increasing based on the location of the water curtain.

Figure 6

Predictions of the velocity vector field and soot mass fraction distribution in the middle section for the cases of droplet mean diameter of D_m = 200 μm; (a) x_inj = 8 m and (b) x_inj = 10 m.

The soot distribution shows that the soot layer thickens owing to the interaction of the smoke flow and spray flow near the nozzle, smoke progress downstream being inhibited. In the case of x_inj = 10 m, as shown in Figure 6(b), the distance between the water-curtain nozzle and the vertical wall is close, so the smoke confinement effect is not substantial. Here, one consideration point that needs to be reviewed is the impact of the water-curtain device on the ease of evacuation for residents. The analysis results show that reduced damage owing to smoke in the downstream region and the water-curtain inhibiting smoke propagation are evident. However, when compared with Figure 3(a), the smoke concentration in the upstream area from the water curtain to the fire source is higher than the case when the water curtain is not operated. Consequently, in the upstream region of the water curtain, the stable smoke layer distribution is disrupted, which can cause the smoke to spread across the entire space, greatly diminishing the evacuation efficiency. Therefore, when the water-mist curtain location is determined, a thorough review of its disturbance to the evacuation plan is essential.

Figure 7 shows the prediction results of the optical density that changes from the floor, at z = 1.5 m height, and the downstream area during t = 20-50 s, as an average value.

Figure 7

Spatial variation of the predicted optical density, at z = 1.5 m; (a) with different droplet mean diameters for the water-mist injection at x_inj = 10 m and (b) with different water-mist locations for the droplet mean diameter of D_m = 200 μm.

Figure 7(a) compares the average spray droplet size when the location of the water mist spray is constant at x_inj = 6 m and shows the optical density change based on the location of the water curtain when the average size of the spray droplet is constant at D_m = 200 μm. Here, the optical density is determined and impacted upon by photosorption characteristics such as the soot concentration and droplets with similar characteristic values to the extinction coefficient used during the process of determining visibility. In general, the visibility is inversely proportional to the optical density. Figure 7(a) shows that when D_m = 1,000 μm, similar values are evident with the case of a non-existing water curtain. When the average diameter is D_m = 200 μm, the optical density at downstream areas lower than x = 9 m is lower than the case with no water curtain. However, in upstream areas higher than x = 9 m, the optical density is higher. In particular, the optical density increases considerably near the water curtain. In this region, soot, water-mist droplets, and water vapor exist in large quantities, so the optical density increases considerably, implying that the visibility will be low.

Figure 7(b) shows that this increase in optical density tends to move downstream along the spraying location. In particular, in the case of x_inj = 10 m—with a short distance between the water curtain nozzle and the vertical wall—higher optical density values are evident than the case with no water curtain, even at the maximum downstream points.

Figure 8 shows the average value of the predicted radiation-attenuation rate changes from the floor, at z = 1.5 m height, and downstream, during t = 20-50 s. The attenuation rate of thermal radiation can be defined as follows using the heat flux ratio (I₀) in the case of no existing water-mist curtain and the heat-flux ratio (I) at the state of operation:

Figure 8

Spatial variation of the radiation attenuation rate, at z = 1.5 m; (a) with different droplet mean diameters for the water-mist injection at x_inj = 10 m and (b) with different water-mist locations for the droplet mean diameter of D_m = 200 μm.

Figure 8(a) shows the impact of the average droplet size when the location of the water-mist curtain is constant at x_inj = 6 m. Overall, when passing through the x = 6-8 m area—where most spray droplets are distributed—the thermal radiation attenuation rate greatly increases. In the case of the average spray droplet size being D_m = 200 μm, the attenuation rate at x = 6-8 m section increases from 14.5% to 65.1%, the attenuation rate increasing up to 79.0% at x = 13 m. As the average spray droplet size increases, the effect of thermal radiation attenuation rapidly decreases. When D_m = 500 and 1,000 μm, the attenuation rate at x = 13 m decreases at 29.7% and 17.0%, respectively.

Figure 8(b) shows the prediction results of the thermal radiation attenuation rate based on the water-curtain location when the average spray droplet size is constant at D_m = 200 μm. In all three cases, rapid attenuation rate increase is evident near the spraying location, and when the spraying locations are x_inj = 6, 8, and 10 m, the attenuation rates at x = 13 m are predicted to be 79.0%, 80.3%, and 60.7%, respectively. Along with the spraying location, the location at which the thermal radiation attenuation rate increases moved downstream, although the tendency for attenuation rate change at the front and rear of the water curtain can be considered to be similar. Consequently, the impact on the thermal radiation attenuation rate can be considered to be greater from the spray droplet size than the location of the water-mist curtain.

4. Conclusions

In this study, numerical analysis was conducted for precise analysis of the radiant-heat shielding and smoke confinement characteristics owing to the water-mist curtain during the plume flow along the ceiling after its occurrence in the fire source. The flow field due to the water curtain, the temperature distribution, and the soot distribution were reviewed to precisely analyze the interaction between the smoke flow and spray flow. The optical density extracted from a 1.5 m height from the ground and the predicted value for the thermal radiation attenuation rate were compared to assess the impact of the average water mist droplet size and the location of the water curtain.

In the case of not operating a water curtain, the soot concentration stably progressed downstream from the ceiling of the corridor in layers. When the water curtain was operated, the spray flow impeded the ceiling jet flow and cause significant mixture. As the average size of the sprayed droplet decreased, the extent of mixture increased, and smoke confinement was efficient. While the cooling effect due to the water mist curtain was immediate, a slight spatial delay occurred in the decrease of soot concentration. In addition, the smoke spread entirely across the upstream area of the water curtain due to the effect of the vortex flow that developed from the spray flow, and the soot concentration in this area was higher than in the case of not operating a water mist.

At z = 1.5 m height from the ground, the optical density prediction review showed that reduction could be identified due to the water mist curtain, but the reduction effect was not significant overall compared to the case of not operating a water curtain. In particular, high optical density was shown in the upstream region and near the water curtain even when the average spray droplet size was small at D_m = 200 μm. This was considered to be because significant fine droplets along with the soot contributed to the optical quenching phenomenon. In contrast, the attenuation extent of thermal radiation was significantly impacted by the average droplet size. When the water mist curtain was located at x_inj = 6 m, the thermal radiation attenuation rate tended to decrease to 79.0%, 29.0%, and 17.0% as the average spray droplet size changed from D_m = 200, 500, and 1,000 μm, respectively, at x = 13 m.

Article information Continued