An analytical methodology of rock burst with fully mechanized top-coal caving mining in steeply inclined thick coal seam

Shan, Pengfei and Yan, Zhongming and Lai, Xingping and Xu, Huicong and Hu, Qinxin and Guo, Zhongan (2024) An analytical methodology of rock burst with fully mechanized top-coal caving mining in steeply inclined thick coal seam. Scientific Reports, 14 (1). 651. ISSN 2045-2322 (

[thumbnail of Shan-etal-SR-2024-An-analytical-methodology-of-rock-burst-with-fully-mechanized-top-coal-caving-mining]
Text. Filename: Shan-etal-SR-2024-An-analytical-methodology-of-rock-burst-with-fully-mechanized-top-coal-caving-mining.pdf
Final Published Version
License: Creative Commons Attribution 4.0 logo

Download (3MB)| Preview


Rock burst disaster is still one of the most serious dynamic disasters in coal mining, seriously restricting the safety of coal mining. The b value is the main parameter for monitoring rock burst, and by analyzing its changing characteristics, it can effectively predict the dangerous period of rock burst. This article proposes a method based on deep learning that can predict rock burst using data generated from microseismic monitoring in underground mining. The method first calculates the b value from microseismic monitoring data and constructs a time series dataset, and uses the dynamic time warping algorithm (DTW) to reconstruct the established b value time series. A bidirectional short-term and short-term memory network (BiLSTM) loaded with differential evolution algorithm and attention mechanism was used for training, and a prediction model for the dangerous period of rock burst based on differential algorithm optimization was constructed. The study used microseismic monitoring data from the B1+2 fully mechanized mining face and B3+6 working face in the southern mining area of Wudong Coal Mine for engineering case analysis. The commonly used residual sum of squares, mean square error, root mean square error, and correlation coefficient R2 for time series prediction were introduced, which have significant advantages compared to basic LSTM algorithms. This verifies that the prediction method proposed in this article has good prediction results and certain feasibility, and can provide technical support for the prediction and prevention of rock burst in steeply inclined thick coal seams in strong earthquake areas.