Analysis of human and organizational factors that influence mining accidents based on Bayesian network.

Purpose. The present study aimed to analyze human and organizational factors involved in mining accidents and determine the relationships among these factors. Materials and methods. In this study, the human factors analysis and classification system (HFACS) was combined with Bayesian network (BN) in order to analyze contributing factors in mining accidents. The BN was constructed based on the hierarchical structure of HFACS. The required data were collected from a total of 295 cases of Iranian mining accidents and analyzed using HFACS. Afterward, prior probability of contributing factors was computed using the expectation-maximization algorithm. Sensitivity analysis was applied to determine which contributing factor had a higher influence on unsafe acts to select the best intervention strategy. Results. The analyses showed that skill-based errors, routine violations, environmental factors and planned inappropriate operation had higher relative importance in the accidents. Moreover, sensitivity analysis revealed that environmental factors, failed to correct known problem and personnel factors had a higher influence on unsafe acts. Conclusion. The results of the present study could provide guidance to help safety and health management by adopting proper intervention strategies to reduce mining accidents.