92 / 2023-06-14 19:46:49

Research on dynamic trend warning methods for rockburst in coal mines

rockburst,early warning,microseismic monitoring,temporal trend test,data fusion

As a frequent coal-rock dynamic hazard, the research on prevention and monitoring of rockburst is of great engineering significance. In order to further improve the monitoring and early warning capability of rockburst, a dynamic warning method of rockburst with multiple indicators using Mann-Kendall trend test is proposed. Firstly, cases of rockburst at different mining levels in a large mining depth span mine is analyzed, and the inducing factors of rockburst are obtained, including the large horizontal tectonic stresses generated by the compressive and prying action of the hard roof and rock pillar and the dynamic load disturbance generated by the destabilization deformation, and the rockbursts gradually changed from being triggered by the rock pillar’s "prying action" at shallow depths to being triggered by the rock pillar and roof "compressive and prying action" at deeper depths. Microseismic monitoring data gradually transformed from "low energy and high frequency" to "high energy and low frequency", and the high energy density zone tended to be locally concentrated, with the danger degree of rockburst showing a nonlinear and rapid increase. Subsequently, the evolution patterns and response sensitivity of precursor warning indicators for rockbursts at different mining depths were analyzed, and it was found that there would be continuous abnormal trends of increasing or decreasing before each rockburst event, which could serve as the basis for rockburst precursor warning, and the high warning efficiency indicators for each mining level differed greatly. Based on this, a multi-parameter dynamic trend warning method for rockbursts was proposed, which used the Mann-Kendall trend test method to determine the temporal changes of each warning indicator and whether it conforms to the characteristic of rockburst precursor as the trigger for alarm, combined with the confusion matrix to judge the warning efficiency of each indicator, and selected the indicators based on the principle of maximizing efficiency, and finally, the unified and quantitative rockburst risk index P_RB was obtained by data fusion based on the warning efficiency as the weight. The results of on-site application showed that the early-warning efficiency of P_RB can reach 0.431, higher than any single indicator. At the same time, this method can realize self-feedback updates driven by the real-time monitoring data on the site, which has good adaptability and scalability for the complex stress conditions of deep mining and can provide efficient and accurate decision-making basis for the prevention and control of underground rockburst.