Experimental investigation on the characteristics and propagation of fire inside subway train

Highlights

• The performance and propagation of fires in subway train were studied.

• The flame propagation length was analyzed and compared to previous results.

• Correlation for predicting two-dimensional ceiling temperature rise was proposed.

• The proposed correlation was compared with literature and showed a good agreement.

Abstract

With the rapid development of subway systems, the related fundamental knowledge is critically needed to benefit their designs and fire hazard assessment. The experimental study was performed in a 1/5th scale subway train model to characterize the fire inside subway train, and the experiment in open space was also conducted as a control group. Results show that in the case where the ceiling flame has not yet propagated to the sidewalls of the subway train, the normalized ceiling flame length varies as 1/4 power of the non-dimensional ceiling heat release rate. In addition, the temperature distributions at different transverse distances from the fire source follow the exponential decay law effectively, the decay rate is indeed a function of normalized transverse distance from the longitudinal centerline. An empirical model involving dimensionless transverse and longitudinal coordinates is proposed to predict the two-dimensional temperature underneath the ceiling, which is reasonably well fitting with experimental results from this study and literatures.

Introduction

The research of fire in complex confined structures has gradually attracted the considerable attention of researchers for decades. For example, Du et al. (Du et al., 2015) used CFD to perform the simulation of the whole urban traffic link tunnel, and studied the design method of smoke control. Yang et al. (Yang et al., 2018) conducted brine water experiments in the inclined tunnel and analyzed the back-layering flow under this simulated fire. Ng Y W et al. (Ng et al., 2019) performed experiments in train car model to explore the flame color under poorly ventilated condition. As one of the complex confined structures, subway system exists widely around the world due to its great convenience for the people in daily life. If a fire breakout in the subway train with lateral openings, it is hazardous for the life of the commuter and the infrastructure. In fact, many major subway fire accidents happened in the history (Li and Ingason, 2018, Peng et al., 2020). Hence, it is important to study the characteristics of the fire inside the subway train, especially the flame length and temperature distribution under ceiling, which are considered to be the two most crucial and representative characteristics for fire detection and protection system.

Numerous research work have been carried out on the flame extension in the ceiling jet under various confinement conditions. For unconfined horizontal ceiling, You and Faeth (You and Faeth, 1979) were the first to experimentally measure the impinging flame lengths, and a formula for predicting the flame length under the ceiling was developed. Then, Zhang et al. (Zhang et al., 2014, Zhang et al., 2017) measured and correlated the impinging flame lengths beneath the horizontal and inclined ceiling for a free fire jet impingement. Further, the study (Zhang et al., 2019a) on the fire impingement scenario of the wall-attached fire were carried out, and the flame lengths in different directions were correlated based on the physical nature of such impinging flow structure. Moreover, Poreh and Garrad (Poreh and Garrad, 2000) studied the measurements of flame heights under different confined conditions. The effect of wall on the flame height was discussed and the proposed model overestimated the experimental value of the flame length induced by fires near the wall. Lattimer et al. (Lattimer et al., 2013) reviewed the flame lengths in various confined configurations and compared the dependence of the flame length on the HRR in different configurations. Recently, Gao et al. (Gao et al., 2019) conducted multiple experiments to characterize the ceiling jet under different flame-wall interactions. They found that the ceiling impingement flames under different confinement strengths followed different patterns, and obtained the correlations to predict the ceiling flame for different impinging flows. The above researches clearly show that different scenarios, especially different confinement conditions, can have an effect on the flame length under the ceiling. However, the flame length in a long - narrow confined structure with lateral openings has been rarely studied systematically.

In terms of smoke temperature distribution under the ceiling, many correlations under various conditions were formulated. Alpert (Alpert, 1975, Alpert, 2004) and Heskestad and Delichatsios (Gunnar et al., 1979) proposed the classical models of the temperature distribution underneath the large flat ceiling, and the dimensionless temperature $\mathrm{\Delta }T/\mathrm{\Delta }{T}_{\mathit{max}}$ could be represented as a function of the non-dimensional location $x/H$. Oka et al. (Oka et al., 2010) carried out experiments in an inclined unconfined ceiling apparatus to study the effect of the inclination angel of the ceiling on the temperature distributions in different directions along the ceiling. Zhang et al. (Zhang et al., 2019b) performed experiments of wall-attached fire in a configuration consisting of wall and inclined ceiling, and the correlation for the temperature profile underneath the ceiling was developed for various ceiling inclination angles. In addition, there are also many studies on fire scenarios in long-narrow space. Zhao et al. (Zhao et al., 2019) performed a mass of experiments in tunnels with various scales to research the ceiling temperature distribution. To study the effect of the tunnel width, Ji et al. (Ji et al., 2016) carried out experiments in a model tunnel with variable aspect ratio. The models for predicting the temperature distributions involving the tunnel width and HRR were proposed. Zhao et al. (Zhao et al., 2017) established a reduced-scale corridor to analyze the effect of openings on longitudinal and vertical smoke temperature, and concluded that the longitudinal temperature under the ceiling in the far-field for different openings also submitted to the exponential decay. These studies perform an important role in the exploration of temperature distribution in confined space fires. However, studies focusing on fires in long -narrow structures with lateral openings are rarely reported, although the existence of lateral openings may have an effect on the propagation of smoke in the structure. In addition, the current temperature prediction models are generally one-dimensional relationships such as transverse or longitudinal, and less attention is paid to the prediction of two-dimensional temperature distribution.

Moreover, some researches on fires in confined structures with opening have also been carried out. Ji et al. (Ji et al., 2015) performed experiments in a compartment with two opposite openings to investigate the effect of wind velocity on compartment fires. For compartment fire under cross wind, Gao et al. (Gao et al., 2016a) performed experiments to focus on the spill plume and the new length scale was proposed to correlate the temperature of facade plume. Zhang et al. (Zhang et al., 2020) conducted fire experiments in compartment with circular openings, and discussed the internal temperature and the extension mechanism of the façade flame. The smoke movement in the above scenario is different from that in the subway train. Because these experimental structures are cubic compartments rather than long-narrow structures, and these studies are more focused on the characteristics of the spill plume rather than the fire characteristics inside the structure. Recently, Ingason (Ingason, 2007) conducted the fire experiments in a model railcar to study the fire development under different parameters. Li et al. (Li et al., 2014) performed the fire experiments in the train carriage models with various scales to analyze the fire development and concluded that the fire development in different scales was similar. Lönnermark et al. (Lönnermark et al., 2017) studied the impact of different ignition source-related variables on the fire development by conducting experiments in the train model. Although train fires have been qualitatively studied in many aspects, the fundamental knowledge of the quantitative research on train fire needs to be supplemented.

The above literature review shows that the state of the art on this subject provides inadequate knowledge about fire inside subway train, especially the important features that characterize the fire hazard, the flame length and temperature distribution under ceiling. Besides, most of the work was focused on the one-dimensional temperature distribution along the longitudinal or transverse centerlines, but very few studies have been conducted for overall ceiling temperature contour, even though they are the determining factors for the fire hazard. Therefore, in order to investigate the flame length and 2D ceiling temperature profile induced by fire inside the subway train, three series of experiments were carried out (fire source located in the open space or inside the train). Correlations were developed to quantitatively describe the flame length and a predictive model for the temperature rise under the ceiling files was also proposed, which could provide a fundamental base for the fire hazard assessment inside the subway train.