Research on fire risk quantification for extralong highway tunnels based on Wuli–Shili–Renli theory, dempster–shafer theory, and bayesian network

The fire risk of extralong highway tunnels (ELHTs) is a significant concern because of their severe consequences. To achieve a quantitative assessment of ELHT fire risk, this paper proposes a fire risk assessment model that integrates Wuli–Shili–Renli (WSR) theory, Dempster–Shafer (DS) evidence theory, and Bayesian network (BN). First, on the basis of WSR theory, fire risk factors are identified from three dimensions—human, physical, and managerial—and an evaluation index system is established. The factors are then mapped into a BN structure, with DS evidence theory and expert knowledge used to calculate conditional probabilities. The prior probabilities are determined using a risk quantification table, and reasoning analysis and sensitivity analysis are conducted to identify the risk level and key indicators. The model is validated through axioms and a case study. Finally, the method is applied to an ELHT, the results reveal that the fire risk of the tunnel is 13.5 %, which is classified as low risk, with "tunnel length" and "traffic volume" identified as key indicators. The study demonstrated that the WSR–DS–BN model has good applicability and interpretability, providing theoretical support and technical assistance for fire risk assessment of ELHTs.