Hybrid-driven risk assessment and emergency effect analysis of coal mine gas explosion: Integration of complex network, Bow-tie and dynamic Bayesian modeling

Aiming at the problems such as the uncertainty of the gas explosion risk and the fuzziness of the emergency response effects, taking 94 coal mine gas explosion accidents as examples, a complex topological network was established to explore the multi-level causes. For the four dimensions, the evaluation index system of gas explosions was determined by using clustering and inductive reasoning methods. By applying the Bow-tie model, Dynamic Bayesian Network (DBN) and fuzzy set theory, a risk assessment and emergency effect analysis model was established. The temporal evolution law of the risk of gas explosion was revealed, and the quantitative control effect of the risk when multiple safety barriers were jointly implemented was obtained. Research shows that based on the current safety control conditions and emergency capabilities of this mine, the probability of a gas explosion occurring at present and in the future (30 weeks later) is 4.8% and 9.3% respectively, and the accident probability has increased by 95.2% compared to the baseline state. Eight key disaster-causing paths were determined. Among them, combined paths 1-6 were the most severe, and the accident probability of this paths increased by 388.2%. To verify the correctness, a verification analysis was conducted taking the investigation of the gas explosion accident at Xintai Coal Mine as an example. The results show that, based on the actual situation of Xintai Coal Mine, the accident probability, consequences and key cause chain are basically consistent with the accident investigation report. The model proposed can help managers effectively prevent and control gas explosion accidents.