China Safety Science Journal ›› 2022, Vol. 32 ›› Issue (11): 134-139.doi: 10.16265/j.cnki.issn1003-3033.2022.11.2387

• Safety engineering technology • Previous Articles     Next Articles

Human factor association rules mining and analysis for hazardous chemical accidents

LI Xin(), YANG Fuqiang**()   

  1. College of Environment and Safety Engineering, Fuzhou University, Fuzhou Fujian 350116, China
  • Received:2022-06-14 Revised:2022-09-11 Online:2022-11-28 Published:2023-05-28
  • Contact: YANG Fuqiang

Abstract:

Apriori algorithm was used to mine the association rules of human factors in hazardous chemical accidents. 217 accident reports were employed as data samples. Based on the accident statistics, 895 association rules were obtained by Apriori algorithm. The rules cover the correlation between human factors, accident time and month and four types of accidents:leakage, explosion, fire and poisoning and asphyxiation. Based on the analysis results, a graph of non-human factors and illegal fire operation lift direction was drawn in this study. Furthermore, a graph of non-human factors and human factors lift direction was drawn. The results show that explosion and fire accidents, leakage and poisoning and asphyxiation accidents have strong correlation rules with each other. Each accident has a strong correlation with the accident time and month. Illegal fire work and three types of accidents (explosion, fire, leakage) are forming strong correlation rules. Nine types of non-human factors, such as inadequate management of work procedures, mixed explosive gas/physical environment, and removed/uninstalled blind, have strong association rules with illegal fire work.

Key words: hazardous chemicals accidents, human factor, Apriori algorithm, association rules, illegal fire work