Description
The potential hazards from fires in tunnels that directly impact on individual life safety, emergency response, and tunnel integrity must be meticulously considered. However, due to numerous influencing parameters of road tunnel fire design, universally acceptable solutions and road standards remain unachievable at this time. This paper presents an in-depth numerical investigation using the Fire Dynamic Simulator (FDS) model, focusing on critical aspects in naturally ventilated road tunnels, including tunnel length, slope, fire location, fire size, and cross-section dimensions. A comprehensive literature review was undertaken to evaluate the gap of knowledge on how each component independently affects a fire scenario. The combined influence of relevant components was then explored numerically. To obtain generalized findings, the numerical simulation results are fed into a machine learning algorithm known as Multilayer Perceptron (MLP). Subsequently, based on the FDS and MLP results, a probabilistic analysis is performed to assess the fire risk in naturally ventilated road tunnels in terms of human safety. Results demonstrate that the probability of failure (𝑃𝑃𝑓𝑓) for small fires (5 MW) is low; however, on steeper slopes it can increase to over 30% due to heat accumulation caused by the stack effect. Large fires (25 MW), have higher failure probabilities on slopes over 3%, approaching 50% on those where tunnel length has less of an effect due to pressure differentials and heat dissipation; for those up to 75 MW with extreme heat and smoke production, shorter tunnels experience up to 87% failure. The results of this study may pave the way for a better understanding of the combined effect of various essential aspects on the risk of fire in such tunnels. Furthermore, these findings may provide specific recommendations for each factor, contributing to a well-defined set of safety measures for road tunnels, thereby enhancing the current standards for tunnels.
