Research and Development and On-site Application of Intelligent Early Warning System for Coal and Gas Outburst

Abstract

To address the challenge of early warning for coal and gas outburst disasters, taking Shanxi Xinjing Mine as the research context and based on the statistical analysis of 209 gas dynamic phenomena in the mine, an early warning indicator system covering three dimensions—production system, region and working face—was constructed. A multi-source information fusion early warning model based on association rules and evidence theory was developed, integrating four core modules: monitoring and data collection, professional analysis, early warning dissemination, and database management. This culminated in the formation of an intelligent early warning system for coal and gas outbursts. Field application results demonstrate that the early warning model exhibits self-learning and self-evolving characteristics. During the evolutionary improvement phase, the overall accuracy of early warnings for heading working faces increased from 81.78% to over 91%, while the overall accuracy for mining working faces remained stable at over 91%. In the early warning application phase, the system achieved an overall early warning accuracy of 91.52%, with a 0% rate of missed reports for non-outburst hazards and an 8.54% rate of false reports for non-outburst hazards, enabling precise early warning of outburst disasters and providing technical support for safe coal mine production.