Early warning of deep coal miners' unsafe behavior based on the HFACS-CM-BP neural network.

Abstract

Preventing miners' unsafe behavior and reducing accidents in deep coal mines are crucial. This study comprehensively used methods such as the human factor analysis and classification system for China mines (HFACS-CM) model, grounded theory and the back propagation (BP) neural network model to construct an early warning index system for miners' unsafe behavior. A three-layer feed-forward BP neural network warning model with a structure of 13-14-4 layers was developed to predict miners' unsafe behavior. The results showed that the model can accurately predict miners' unsafe behavior and reflect the complex non-linear relationship between the driving factors and unsafe behavior. Unsafe supervision was the most critical driving factor affecting miners' unsafe behavior, followed by organizational influence, miners' unsafe state and environmental factors. This study can help mining enterprises formulate more effective management measures for miners' unsafe behavior so as to improve the efficiency of coal mine safety management.