Coal and Gas Outburst Risk Prediction and Management Based on WOA-ELM

Abstract

A gas outburst risk level prediction method, based on the Whale Optimization Algorithm (WOA) Improved Extreme Learning Machine (ELM), is proposed to predict the coal and gas outburst hazard level more accurately. Based on this method, recommendations are given according to the gas outburst risk level with the help of the Case-Based Reasoning (CBR) method. Firstly, we analyze the accident reports of gas outburst accidents, select the gas outburst risk prediction index, and construct the gas outburst risk prediction index system by combining the gas outburst prevention and control process. The WOA-ELM model was used to predict the gas outburst risk level by selecting data from 150 accident reports from 2008 to 2021. Again, based on the coal and gas outburst risk level, CBR is used to match the cases and give corresponding suggestions for different levels of gas outburst risk conditions to help reduce the gas outburst risk. The results show that the WOA-ELM algorithm has better performance and faster convergence than the ELM algorithm, when compared in terms of accuracy and the error of gas outburst hazard prediction. The use of CBR to manage prediction results can be helpful for decision-makers.