AUTHOR=Wang Chong , Wei Like , Hu Haiyong , Wang Jiren , Jiang Mengfeng 

TITLE=Early Warning Method for Coal and Gas Outburst Prediction Based on Indexes of Deep Learning Model and Statistical Model

JOURNAL=Frontiers in Earth Science

VOLUME=Volume 10 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/earth-science/articles/10.3389/feart.2022.811978

DOI=10.3389/feart.2022.811978

ISSN=2296-6463

ABSTRACT=In recent years, as coal and gas outburst occurrence continues, more and more scholars have begun to study early warning models for it. However, when implementing these methods, their theoretical models are often so complex that they cannot be used to guide field applications due to the unavailability of the real time parameters. Based on the evolving mechanism of gas outburst, this paper selects the gas emission that can not only be obtained in real time through the online safety monitoring system, but also can characterize the whole process of outburst as the main controlling factor for analysis. In order to characterize the variation of gas emission in the outburst process and distinguish it from the normal mining process, we employed four statistical methods (moving average, deviation, dispersion, and volatility), as well as Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE) which are used to gauge the accuracy of the deep learning model when predicting the variation of the gas emission, to form the six characteristic indicators to quantify the variation of gas emission. The multi-factor fuzzy comprehensive evaluation model established from these six indicators then forms the outburst early warning system by calculating the combined index of the difference of each indicator. Finally, we take the "3.25" gas outburst in Shigang Coal Mine as an example to verify the accuracy of the early warning system. By analyzing the result, we can prove the advantages of the comprehensive evaluation model established from the six characteristic indicators in predicting outburst.