Please use this identifier to cite or link to this item:
Title: Predicting safety hazards and safety behavior of underground coal mines
Authors: Hongxi, Di
Asma, Sbeih
Shibly, FHA
Keywords: Safety hazards and safety behavior
Underground coal mines
Prediction framework
Issue Date: 14-Aug-2021
Publisher: Springer Berlin Heidelberg
Citation: Soft Computing; 2021
Abstract: Most coal mine accidents are attributed to the miner’s unsafe behavior. Regulating the safety attitude and thus enhancing miners’ safety behavior are significant for accident prevention. Capturing the interrelations between risks is important to understand and promote coal mining safety thoroughly. Therefore, this paper proposes the intelligent accident predictive framework (IAPF) for monitoring and analyzing the safety hazards and safety behavior of underground coal mines. The most significant variables involved in occupational accidents and their association rules have been identified. These rules are composed of numerous predictor variables that cause accidents, describing their characteristics and environment. The accident model path analysis demonstrates that adverse effects, risk-taking behaviors predict and job dissatisfaction an increased number of injuries in mines. The IAPF model gives an outcome as an indicative risk score linked with the identified accident-prone situation, based upon which an appropriate mitigation plan can be established. The results show the most typical instant causes and the percentage of accidents with a basis in every connotation rule. The experimental results of IAPF show the highest prediction ratio of 97.5%, safety rate of 96.3%, security rate of 95.4%, and lowest accident rate of 22.6%, energy consumption ratio of 28.6%, carbon management ratio of 25.3% and hazard risk ratio of 20.2% compared to other methods.
ISSN: 1433-7479
Appears in Collections:Research Articles

Files in This Item:
File Description SizeFormat 
Abstract_Accident_Paper.pdf82.71 kBAdobe PDFThumbnail

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.