Estimation of smoke arrival time in tunnel fires

https://doi.org/10.1016/j.tust.2013.08.002Get rights and content

Abstract

The estimation of smoke arrival time in tunnel fires is helpful to comprehensive fire risk assessment and effective fire evacuation, while few studies focused on this topic. A model to estimate the arrival time of fire smoke in tunnels is derived based on the smoke temperature distribution along the tunnel ceiling. The predictions from the model are compared to experimental data from one past study, which shows good agreements. The influencing factors of the smoke arrival time are studied based on the model. Results show that the Stanton number is the main influencing factor. The smoke arrival time increases with the increase of the Stanton number.

Introduction

Tunnel or long corridor fires have gained much concern recently in the field of fire safety science since several fatal tunnel fires occurred (Brousse et al., 2001, Hong, 2004, Kirkland, 2002, Rail Accident Investigation Branch, 2010). Smoke control is one of the most concerned topics as fire smoke spread along the tunnels will threaten the lives inside and the tunnel structure, which may cause disasters such as injuries, loss of lives, severe property damage, service disruption, or even loss of public confidence in tunnels as a safe means of transportation (Kang, 2010).

In a tunnel or long corridor fire, the spread process of the smoke flow under the tunnel ceiling could be described in three steps (Delichatsios, 1981, Kunsch, 2002, Li et al., 2012c) as shown in Fig. 1:

  • 1.

    The formation of the axisymmetric radial fire plume, i.e., from (a) to (b).

  • 2.

    The occurrence of a density jump. In this step, the fire plume first contacts the tunnel ceiling and turns to ceiling jet (this transition part is called turning region, i.e., from (b) to (c)) and then it intercepts the tunnel sidewalls (from (c) to (d)), and a transition from radial to one-dimensional flow occurs. The density jump occurs during this process, either before or after the fire plume intercepts the sidewalls depending on the tunnel height-to-width ratio.

  • 3.

    The formation of the one-dimensional flow. After the density jump, a one dimensional flow along the tunnel is generated, which moves toward the tunnel exits under the ceiling.

The main hazards come from the spread of the one-dimensional smoke flow due to its toxicity and high temperature. Thus, some studies focused on the smoke control, especially the critical velocity of the longitudinal ventilation (Chow et al., 2010, Hu et al., 2008, Ko et al., 2010, Li et al., 2012a, Li et al., 2010), while some others focused on the most representative fire parameter – temperature, i.e., the maximum temperature (Hu et al., 2006, Kurioka et al., 2003, Li et al., 2012b, Li et al., 2011, Li and Ingason, 2012) and the temperature decay (Chen, 2008, Hu et al., 2007b, Li et al., 2012c, Li et al., 2012d). These are helpful studies to tunnel fire safety design. However, few studies have been reported to investigate the arrival time of the fire smoke. Knowledge of smoke arrival time in tunnel fires could help to estimate the fire risk and to assist appropriate fire evacuation.

Based on the authors’ previous work on smoke temperature distribution under natural ventilation, a model for smoke arrival time will be derived in this paper. The predictive ability of the model will be justified using a set of experimental data. The main influencing factors will also be discussed.

Section snippets

Model derivation

The authors’ previous study presented a model for the temperature distribution of the one-dimensional fire smoke flow along tunnel ceiling under natural ventilation (Li et al., 2012c). This model comes to the following form if the start point of the one-dimensional smoke flow is taken as the reference position:ΔTΔT0=exp-St(2ϕ+20.3ϕ-1)(x-W/2)H

When the density difference between the smoke flow current and the ambient airflow is small, Bailey et al. (2002) proposedu=0.7ghΔTTa

This correlation

Experimental data

The experimental data used here are from Hu et al. (2007a). The experiments were carried out in a tunnel with the scale of 88 m × 8 m × 2.65 m, as shown in Fig. 2. Two diesel pool fires were used near the north end, which had the heat release rate of 0.8 MW and 1.5 MW, respectively. No mechanical ventilation systems were used. The south end was open providing natural ventilation while the north end was closed. 10 pairs of infrared beams and 49 points of thermocouples were set along the tunnel, as shown

Results and discussion

Fig. 3 compares the model in Eq. (6) with the experimental data by Hu et al. (2007a). The authors’ previous work has discussed the appropriate range of the Stanton number in tunnel fires, which showed a result of 0.0027–0.0073 (Li et al., 2012d). The minimum value St = 0.0027 is adopted in Fig. 3. The ±25% relative error bars of the experimental data are shown. The figure shows that the model in this paper can fit the experimental data well, especially for the dimensionless distance values that

Conclusions

The toxicity of smoke in a fire causes the most deaths, thus to estimate the smoke arrival time in a tunnel fire could help to organize effective evacuation and reduce the injuries and deaths. This paper presents a model for estimating the smoke arrival time in tunnel fires, the main conclusions are as follows:

  • 1.

    The estimations from the model agree well with the existed experimental data, and the predictive ability is justified with engineering accuracy.

  • 2.

    The Stanton number and fire heat release

Acknowledgement

The study was conducted under the project 2012CB719705 supported by the National Key Basic Research and Development Plan.

Cited by (0)

View full text