A human and organizational analysis method for Chinese high-speed railway accidents/incidents based on Bayesian Network

摘要

Recent study has observed that the number of accidents due to the human errors is increasing in the high-speed railway industry. Therefore, the Human Factor Analysis and Classification System for Railway Accidents (HFACS-RAs) is modified to analyzing the casual factors in accident reports. The correlation analysis of human and organizational factors identified from 106 accident investigation reports is conducted, and the BN structure is established. DS/AHP evidence fusion method depending on the expert knowledge is adopted to infer the conditional probability tables CPTs in the BN. The proposed model is employed to the real accident cases, and numerous relationships between factors in different categories are explored. The BN model indicates that a violation behavior raises the probability of breakdown 83%. Factors related to the task and individual conditions influence unsafe acts the most. The model can be used to address problem areas and enhance safety management in the railway industry.