Association Rule Analysis of Petrochemical Fire Accidents Based on the Apriori Algorithm

To analyze the characteristics of petrochemical fire accidents and emergency disposal rules, firstly we did a survey of 947 petrochemical accidents which happened in China from 2013 to 2018. Then, for analyzing the association rules systematically and comprehensively between fire accidents and the causes, a data mining method, Apriori algorithm, was applied to mine the potentially rules. The results indicated that: (1) There were 22 causes summarized from all investigation reports, but by quantitatively calculation, only 7 of them were key disaster-causing factors. (2) Five disaster-causing factors can be classified as corporate governance problems, so “Business management has loopholes” was one main risk factors of fire accidents. (3) Some key disaster-causing factors were related with each other. “Failure to implement the main responsibility of production safety”, “Security management has bugs” and “Inadequacy of safe education and training” were related to “Inadequacy of government supervision”.