An empirical study of early warning model on the number of coal mine accidents in China. (March 2020)

• Record Type: Journal Article
• Authors: Liu, Qing; Liu, Jian; Gao, Jinxin; Wang, Jingjing; Han, Jing
• Is Part Of: Safety science. Volume 123(2020)
• Publication Date: 2020-03
• Subjects: Coal mine accidents -- BP model -- VAR model -- Early warning; Industrial accidents -- Periodicals; Accident Prevention -- Periodicals; Safety -- Periodicals; Travail -- Accidents -- Périodiques; 363.11
• DOI: 10.1016/j.ssci.2019.104559
• Languages: English
• ISSNs: 0925-7535

Abstract

Abstract: Coal mining is an important energy industry as well as high risk industry with high accident rate. The objective of this study is to establish an early warning model for coal mine accidents. Based on the accident data in the recent 11 years, it is concluded that the coal mine accidents data has the characteristics of large fluctuations, small sample size, small value and large numerical nonlinear characteristics. After literature consulting, data analysis and variable comparison, Policy intervention degree (PID), Manufacturing Purchasing Managers' Index (PMI), Producer Price Index for Industrial Products (PPI), Main raw material purchase price index (RMPPI), Employment index (EI) are selected as the auxiliary variables to construct the early warning model. VAR model is applied to determine the model structure. Results show that BP neural network model is suitable for the prediction of the number of coal mine accident.