Risk analysis of stampede in sporting venues based on catastrophe theory and Bayesian network

With the continuous increase of large-scale activities, it is easy to occur serious stampedes. This study focuses on sporting venues to study stampede accidents in large-scale activities for quantitative and dynamic risk analysis. A comprehensive risk analysis model for stampede accidents is established containing 96 nodes that take into account human error, site failure, environmental impact, and management deficiency, and it is applicable for most sports venues as the relevant inputs can be consistent with reality. First, the bow-tie method is applied for risk identification, and the catastrophe theory is introduced to better handle the expert system and statistics to establish the Bayesian network. Then, scenario analysis is implemented to dynamically predict the consequences, and sensitivity analysis is conducted to quantitatively identify the critical factors of the stampedes. It is identified that the failure of emergency lighting and indicating signs contribute more to the occurrence of a stampede, and the effective emergency measures are critical to reducing the casualties and economic losses. This paper could quantitatively assess the stampede risks and find out the critical hazards of stampede accidents in detail to provide technical and managerial support for most sports venues.