Numerical Simulation on Fire Phenomena During Simultaneous Fires in Two Rooms Connected by Side Opening

Yun Young Kim; Chi Young Lee

doi:10.7731/KIFSE.14741e25

Abstract

Numerical simulations were conducted to analyze fire phenomena in two rooms connected by a side opening, each containing a single fire source. The two rooms (Room-A and Room-B) were connected via Opening-A, with Room-B also connected to the outside via Opening-B. Fire sources with heat release rates of 2.605 kW (Case1) and 5.21 kW (Case2) were placed at the centers of both rooms. The results were compared with those of previous numerical simulations of a single 5.21 kW fire source placed in either Room-A (Case3) or Room-B (Case4). Flow velocities through Opening-A were lower in Case1 and Case2 than in Case3, with Case2 exhibiting higher velocity than Case1. The flow velocity through Opening-B was highest in Case2. When a fire source was in Room-A, mass flow rates through Opening-A were lower in Case1 and Case2 than in Case3, while Case2 showed the highest mass flow rate through Opening-B. When a fire source was in Room-B, Case2 exhibited the highest mass flow rates through both openings, and Case4 the lowest. The highest hot gas layer (HGL) temperature rise and thickness were observed in Case2, while the lowest HGL temperature rise was observed in Case1. The thinnest HGL was observed in Case3 and Case4 when the fire source was in Room-A and Room-B, respectively.

Keywords: Side opening-connected rooms; Simultaneous fires; Fire phenomena; Numerical simulations

1. Introduction

Fires stemming from multiple fire sources may exhibit more complex fire phenomena than those derived from a single source [1]. Interactions between multiple fire sources can give rise to phenomena such as stronger combustion, flame merging, and flame whirls, resulting in increased fire intensity, flame height, and flame temperatures [1-4]. Given the potentially significant risk to life and property, research on compartment fires involving multiple fire sources is essential.

Several studies have explored the fire phenomena in a single room with multiple sources. Ji et al. [1] conducted experiments to investigate the influence of fire source size and separation distance on mass loss rates, flame geometry, and heat flux, and developed a correlation to predict ceiling flame spread. Using numerical simulations based on ISO 9705 fire tests conducted at the National Institute of Standards and Technology (NIST) [5], Chen et al. [6] analyzed how the number, location, and separation of fire sources influenced opening flow velocity, temperature, and the concentrations of carbon monoxide and carbon dioxide within the compartment. Vigne et al. [7] carried out experiments and numerical simulations to compare compartment fires with four fire sources to single-source scenarios. They found that the complex flow patterns generated by multiple fire sources induced turbulence, which significantly affected compartment temperatures. Using experiments and numerical simulations, Tsai et al. [8] investigated the influence of fire source separation distance and heat release rates (HRRs) on two-source corridor fires, determining critical ventilation velocities for each configuration and comparing them to single-source scenarios. Fukuda et al. [9] experimentally measured the heat flux, flame height, and centerline temperature under varying fire conditions, reporting that the number of fire sources significantly influenced flame-merging dynamics.

While many previous studies have focused on fire phenomena within a single compartment with multiple fire sources, studies of interconnected compartments have received less attention. Amokrane and Sapa [10] conducted numerical simulations of the cases where one or two of three-interconnected compartments contain a single fire source, comparing their results with experimental data from the PRISME3 project [11] and evaluating the Peatross and Beyler correlation [12] for mass loss rate per unit area. Similarly, Lee et al. [13] examined the predictive capabilities of the fire dynamics simulator (FDS) for scenarios where two of three interconnected compartments contained a fire source, also using PRISME3 data [11]. They analyzed temperature, heat flux, and oxygen concentration to derive bias factors and relative standard deviations.

In the previous study [14], we investigated fire scenarios involving a single fire source located within one of two interconnected rooms (R-1 or R-2). We varied key parameters such as the HRR, side-opening width, and wall material, analyzing the resulting mass flow rate and flow velocity through the connecting opening, as well as the hot gas layer (HGL) thickness and temperature rise within both the room containing the fire source and the adjacent room. Based on the numerical simulation results, we proposed a correlation for predicting the HGL temperature rise in both rooms. The previous study [14] demonstrated the significant influence of the fire source location (R-1 or R-2) on the fire phenomena of both rooms. However, it did not consider simultaneous fires in both rooms, which could significantly alter fire phenomena owing to fire interaction.

Extending the previous study [14] on compartment fire phenomena, this study investigates two connected rooms (Room-A and Room-B), each containing a central fire source. We numerically simulated two scenarios with different HRRs, analyzing the resulting flow velocity and mass flow rate through the opening, as well as the HGL thickness and temperature rise within both rooms. We then compared our results to the previous study [14], where only one compartment contained a fire source.

2. Method and Conditions

2.1 Method

In this study, we used the same numerical simulation method as in the previous work [14]. The numerical simulation schematic and measurement positions are shown in Figures 1(a) and 1(b). The dimensions are 1065 mm (width) × 1000 mm (depth) × 1100 mm (height) for Room-A and 975 mm (width) × 1000 mm (depth) × 1100 mm (height) for Room-B. Room-A and Room-B are connected via Opening-A, and Room-B is connected to the outside by another opening, Opening-B. Both openings have identical dimensions of 330 mm (width) × 825 mm (height).

Fire sources were placed at the centers of Room-A and Room-B. The HRRs of the fire sources were set to be identical, and fire growth was modeled using the t-squared fire model (Q̇=αt²) [15], where Q̇, t, and α represent the HRR, time, and fire growth factor, respectively. Methanol was used as the fuel, and the fire growth factor was set to ultrafast [15].

The setup for measuring compartment temperatures, opening flow velocities, and mass flow rates was identical to that of our previous study [14]. Four thermocouple trees (TC1-TC4) were used to the measure temperature distribution in Room-A (TC1 and TC3) and Room-B (TC2 and TC4). Sixteen temperature measurements were taken at each TC: one at a height of 200 mm from the floor, and 15 more at 50 mm intervals from 300 mm to 1000 mm above the floor, totaling 64 data points. The flow velocities through the side openings (Opening-A and Opening-B) were measured along the vertical centerline of each opening, and the V-velocity (y-axis velocity) was acquired at 21 points spaced at 40 mm intervals, from 12.5 mm to 812.5 mm above the floor. Additionally, the mass flow rates through Opening-A and Opening-B were measured.

The HGL thickness and temperature were derived using the same method as in the previous study [14], which is summarized below. The HGL height within the compartments was determined using Eqs. (1)-(3) [16,17].

(1)

∫0HT(z)dz=(H−zi)Tu+ziTl=I1

(2)

∫0H1T(z)dz=(H−zi)Tu+ziTl=I2

(3)

zi=Tl(I1I2−H2)I1+I2Tl2−2TlH

where H represents the compartment height, and T_u and T_l are the hot upper layer and the cool lower layer temperatures, respectively. The HGL thickness was calculated by subtracting the HGL height (z_i) from the compartment height (H), and the HGL temperature rise was determined by averaging the temperature measurements within the HGL. In this study, the HGL thickness and temperature rise are the average values from TC1 and TC3 for Room-A, and TC2 and TC4 for Room-B.

In this study, we used FDS (version 6.6.0) for the numerical simulations. The grid size was set to 0.01 m based on the grid-sensitivity analysis conducted in the previous study [14]. Similarly to the previous work [14], the numerical simulations were run for 200 s, and average values and standard deviations were calculated using data from the steady-state period (150-200 s).

2.2 Conditions

Figure 2 illustrates the numerical simulation conditions investigated in this study. In Case1 (Figure 2(a)) and Case2 (Figure 2(b)), both Room-A and Room-B contained fire sources, with HRRs of 2.605 kW and 5.21 kW, respectively. To analyze the results of this study, we compared them with those of our previous study [14] (Case3 and Case4, shown in Figure 2(c) and 2(d)), where a single 5.21 kW fire source was located in Room-A (Case3) or Room-B (Case4). The total HRR (the sum of the HRRs of the fire sources in Room-A and Room-B) was 5.21 kW for Case1, Case3, and Case4, and 10.42 kW for Case2.

3. Results and Discussion

3.1 Velocity profiles on side openings

Figures 3(a) and 3(b) illustrate the flow velocity distributions through Opening-A and Opening-B with measurement height. For Case1, Case2, and Case3, where a fire source was in Room-A, a flow from Room-A to Room-B was observed in the upper region of Opening-A, with a counter-flow in the lower region. Compared with Case3, where there was no fire source in Room-B, flow velocities in Case1 and Case2 were lower, with Case2 (higher HRR) exhibiting higher velocities than Case1. By contrast, Case4, where there was no fire source in Room-A, showed a different velocity profile. As reported in the previous study [14], a flow from Room-B to Room-A was observed in the upper region and a flow from Room-A to Room-B was observed in the middle region. In addition, a slow and weak flow was observed in the lower region [14].

On the other hand, all cases exhibited the same flow profile through Opening-B: an outward flow from Room-B at the upper region and a counter-flow in the lower region. Among all cases investigated, Case2 exhibited the highest flow velocity.

3.2 Mass flow rates on side openings

Figure 4 compares the mass flow rates through Opening-A and Opening-B. Figure 4(a) shows the results for when a fire source was placed in Room-A (Case1, Case2, and Case3). "Opening-A" refers to the mass flow rate from Room-A to Room-B through Opening-A, and "Opening-B" from Room-B to the outside through Opening-B. As reported previously [14], in Case3, where Room-B had no fire source, the mass flow through Opening-B was slightly higher than through Opening-A. However, in Case1 and Case2, where a fire source was in Room-B, the mass flow through Opening-B was significantly higher than through Opening-A. The mass flow rate through Opening-A was lower in Case1 and Case2 (fire source in Room-B) than in Case3 (no fire source in Room-B). This, combined with the velocity distributions shown in Figure 3(a), indicates that the presence of a fire source in Room-B suppresses the discharge flow from Room-A to Room-B. By contrast, Case1 showed a higher mass flow rate through Opening-B than Case3. The mass flow rate discharged through Opening-B is the same as the mass flow rate of fresh air entering from the outside. With the same total HRR, two fire sources (as in Case1 and Case2) provide a larger surface area for entrainment, leading to increased entrainment of fresh air into the fire plumes [7] and resulting in a greater discharge mass flow rate through Opening-B. Meanwhile, Case2, with two fire sources and the highest total HRR, exhibited the greatest mass flow rate.

Figure 4(b) shows the results for Case1, Case2, and Case4, in which a fire source was placed in Room-B. Here, "Opening-A" refers to the mass flow rate from Room-B to Room-A through Opening-A. As reported in the previous study [14], in Case4 (no fire source in Room-A), the mass flow rate through Opening-A was lower than through Opening-B. This trend was also observed in Case1 and Case2 (fire source in Room-A). The mass flow rate through Opening-A was lower in Case4 than in Case1 and Case2. This is because, as shown in Figure 3(a), discharge flow from Room-B to Room-A through Opening-A is limited to the upper region of the opening. Meanwhile, the mass flow rate through Opening-B was lowest in Case4, followed by Case1 and then Case2, owing to the higher HRR and number of fire sources.

3.3 HGL temperature rise and thickness

Figure 5 presents the HGL temperature rise within the compartments. When a fire source was in Room-A (Figure 5(a)), Case2 exhibited the highest temperature rise, despite having the same HRR (5.21 kW) in Room-A as Case3. This indicates that the presence of a fire source in Room-B influenced the HGL temperature rise in Room-A. Meanwhile, Case1, with a lower HRR in Room-A, showed the lowest HGL temperature rise. A similar trend was observed when a fire source was in Room-B (Figure 5(b)): Case2 had the highest HGL temperature rise, followed by Case4 and then Case1. This is attributed to the combined effects of the HRR in Room-B and the presence of the fire source in Room-A.

Figure 6 shows the HGL thickness in the compartments. Among the three cases (Case1, Case2, and Case3) where Room-A contained a fire source (Figure 6(a)), Case2 exhibited a thicker HGL than Case3. Despite its lower HRR in Room-A, Case1 also showed a greater HGL thickness than Case3. A similar trend was observed when the fire source was in Room-B (Figure 6(b)): the HGL was thickest in Case2, followed by Case1 and then Case4. In other words, the conditions where both Room-A and Room-B have the fire source (Case1 and Case2) showed a thicker HGL than the conditions where only Room-A or Room-B has the fire source (Case3 and Case4, respectively). Meanwhile, Figures 5 and 6 show that both the HGL temperature rise and thickness were lower in Room-B than in Room-A. This difference is because Room-B is connected to the outside through Opening-B.

While this study provides valuable insights that complement the previous study [14] investigating the effects of a central fire source in two rooms, it is important to acknowledge that fire phenomena can be influenced by several factors, including fire location. In real-world scenarios, fires can ignite at various locations within a compartment, potentially influencing flow behavior and HGL characteristics. Therefore, future research should experimentally and numerically investigate the effects of varying the fire source location, along with factors such as the number, size, and HRR of the fire source.

4. Conclusion

In this study, we conducted numerical simulations to analyze fire phenomena in two rooms connected by a side opening, each containing a single fire source. The two rooms (Room-A and Room-B) were connected via Opening-A; Room-B was also connected to the outside via Opening-B. Fire sources with HRRs of 2.605 kW (Case1) and 5.21 kW (Case2) were placed at the center of both rooms. The numerical simulation results, including flow velocity, mass flow rate, and HGL temperature rise and thickness, were compared with those of previous numerical simulations [14], where a single 5.21 kW fire source was placed in either Room-A (Case3) or Room-B (Case4). The key findings of this study are summarized below.

(1) In Case1 and Case2, flow through Opening-A was bidirectional: from Room-A to Room-B in the upper region and from Room-B to Room-A in the lower region. Compared with Case3, Case1 and Case2 had lower flow velocities through Opening-A, with Case2 showing higher velocities than Case1. Similarly, flow through Opening-B was bidirectional: from Room-B to the outside in the upper region and from the outside to Room-B in the lower region, with Case2 exhibiting the highest velocities.

(2) When Room-A contained a fire source, the mass flow rate through Opening-A was lower in Case1 and Case2 than in Case3. However, Case2 had the highest mass flow rate through Opening-B. In Room-B with a fire source, Case2 had the highest mass flow rates through both Opening-A and Opening-B, while Case4 had the lowest.

(3) The highest HGL temperature rise and thickness were observed in Case2, while Case1 had the lowest HGL temperature rise. The thinnest HGL was observed in Case3 and Case4 when the fire source was in Room-A and Room-B, respectively.

Notes

Author Contributions

Conceptualization, Y.Y. and C.Y.; methodology, Y.Y.; software, Y.Y.; formal analysis, Y.Y.; investigation, Y.Y.; writing—original draft preparation, Y.Y. and C.Y.; writing—review and editing, C.Y.; visualization, Y.Y.; supervision, C.Y. All authors have read and agreed to the published version of the manuscript.

Conflicts of Interest

The authors declare no conflict of interest.

Acknowledgments

This paper was supported by the National Fire Agency R&D program (grant number 20016433). This study was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. 2019R1F1A1062867).

Figure 1.

Two rooms connected by side opening: (a) schematic and (b) top-view for measurement positions[14].

Figure 2.

Numerical simulation conditions for (a) Case1, (b) Case2, (c) Case3[14], and (d) Case4[14].

Figure 3.

V-velocity profiles with measurement height on (a) Opening-A and (b) Opening-B.

Figure 4.

Mass flow rates on side openings in (a) Case1, Case2, and Case3; and (b) Case1, Case2, and Case4.

Figure 5.

HGL temperature rise in (a) Room-A and (b) Room-B.

Figure 6.

HGL thickness in (a) Room-A and (b) Room-B.

References

1. J. Ji, H. Wan, Z. Gao, Y. Fu, J. Sun, Y. Zhang, K. Li and S. Hostikka, “Experimental Study on Flame Merging Behaviors from Two Pool Fires along the Longitudinal Centerline of Model Tunnel with Natural Ventilation”, Combustion and Flame, Vol. 173, pp. 307-318 (2016), https://doi.org/10.1016/j.combustflame.2016.08.020.

2. W. G. Weng, D. Kamikawa, Y. Fukuda, Y. Hasemi and K. Kagiya, “Study on Flame Height of Merged Flame from Multiple Fire Sources”, Combust Science and Technology, Vol. 176, No. 12, pp. 2105-2123 (2004), https://doi.org/10.1080/00102200490514949.

3. N. Liu, Q. Liu, J. S. Lozano, L. Zhang, Z. Deng, B. Yao, J. Zhu and K. Satoh, “Multiple Fire Interactions: A Further Investigation by Burning Rate Data of Square Fire Arrays”, Proceedings of the Combustion Institute, Vol. 34, No. 2, pp. 2555-2564 (2013), https://doi.org/10.1016/j.proci.2012.06.098.

4. Z. Liu, N. Liu, J. Lei, X. Miao, L. Zhang and D. X. Viegas, “Evolution from Conical to Cylindrical Fire Whirl: An Experimental Study”, Proceedings of the Combustion Institute, Vol. 38, No. 3, pp. 4579-4586 (2021), https://doi.org/10.1016/j.proci.2020.05.001.

5. A. Lock, M. Bundy, E. L. Johnsson, A. Hamins, G. H. Ko, C. H. Hwang, P. Fuss and R. Harris, "Experimental Study of the Effects of Fuel Type, Fuel Distribution, and Vent Size on Full-scale Underventilated Compartment Fires in an ISO 9705 Room", NIST Technical Note 1603 (2008).

6. Q. Chen, A. C. Y. Yuen, T. B. Y. Chen, R. F. Cao, H. Liu and G. H. Yeoh, “A Large-eddy Simulation Study on the Effect of Fuel Configuration and Pan Distance towards Chemical Species for Under-ventilated Compartment Fire Scenario”, International Journal of Heat and Mass Transfer, Vol. 184, (2022), https://doi.org/10.1016/j.ijheatmasstransfer.2021.122306.

7. G. Vigne, W. Węgrzyński, A. Cantizano, P. Ayala, G. Rei and C. G. Montes, “Experimental and Computational Study of Smoke Dynamics from Multiple Fire Sources Inside a Large-volume Building”, Building Simulation, Vol. 14, pp. 1147-1161 (2021), https://doi.org/10.1007/s12273-020-0715-1.

8. K. C. Tsai, H. H. Chen and S. K. Lee, “Critical Ventilation Velocity for Multi-source Tunnel Fires”, Journal of Wind Engineering and Industrial Aerodynamics, Vol. 98, No. 10-11, pp. 650-660 (2010), https://doi.org/10.1016/j.jweia.2010.06.006.

9. Y. Fukuda, D. Kamikawa, Y. Hasemi and K. Kagiya, “Flame Characteristics of Group Fires”, Fire Science and Technology, Vol. 23, No. 2, pp. 164-169 (2004), https://doi.org/10.3210/fst.23.164.

10. A. Amokrane and B. Sapa, “Large Scale Multi-compartment and Multi-source Fire Modelling Using the Peatross & Beyler Correlation”, Fire Safety Journal, Vol. 138, (2023), https://doi.org/10.1016/j.firesaf.2023.103810.

11. S. Suard, P. Zavaleta and H. Pretrel, "Overview of the OECD PRISME 3 Project", Fire Safety and Emergency Preparedness for the Nuclear Industry, FSEP, OTTAWA, Canada (2019).

12. M. J. Peatross and C. Beyler, “Ventilation Effects on Compartment Fire Characterization”, Fire Safety Science, Vol. 5, pp. 403-414 (1997), https://doi.org/10.3801/IAFSS.FSS.5-403.

13. J. H. Lee, B. J. Kim, S. K. Lee and W. G. Shin, “Validation of the Fire Dynamics Simulator (FDS) Model for Fire Scenarios with Two Liquid Pool Fires in Multiple Compartments”, Fire Safety Journal, Vol. 141, (2023), https://doi.org/10.1016/j.firesaf.2023.103892.

14. Y. Y. Kim and C. Y. Lee, “Numerical Investigation on Upper Layer Thickness and Temperature Rise in a Fire in Two Rooms Connected through a Side Vent”, Case Studies in Thermal Engineering, Vol. 58, (2024), https://doi.org/10.1016/j.csite.2024.104370.

15. P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, L. P. Richard, J. R. Hall and J.M. Watts, "SFPE Handbook of Fire Protection Engineering", 3rd ed., National Fire Protection Association, Society of Fire Protection Engineers (2002).

16. K. McGrattan, S. Hostikka, R. McDermott, J. Floyd, M. Vanella, C. Weinschenk and K. Overholt, "Fire Dynamics Simulator Technical Reference Guide Vol. 3: Validation", 6th ed., NIST Special Publication 1018-3 (2017). http://dx.doi.org/10.6028/NIST.SP.1018.

17. Y. He, A. Fernando and M. Luo, “Determination of Interface Height from Measured Parameter Profile in Enclosure Fire Experiment”, Fire Safety Journal, Vol. 31, pp. 19-38 (1998), https://doi.org/10.1016/S0379-7112(97)00064-7.