An online early warning method for coal mine rockburst disaster based on feature drift

Abstract

The invention discloses an online early warning method for coal mine rockburst disaster based on feature drift, which belongs to the field of information processing technology. The invention applies the nonlinear signal analysis theory to process the microseismic data stream in the time window, and extracts each microseismic event. Eight time-frequency domain data are obtained to form the feature vector, which can more accurately describe the drift law of the microseismic data before and after the rock burst disaster; the least squares support vector machine (LS‑SVM) is introduced to learn and train the sample data, and the LS ‑SVM classifier, using LS‑SVM classifier to classify microseismic monitoring data streams, and then discover the precursors of rock burst disasters and give online warnings, which are more robust than conventional threshold warning methods.

Description

technical field [0001] The invention belongs to the technical field of information processing, and in particular relates to an online early warning method for coal mine rock burst disasters based on feature drift. Background technique [0002] Rock burst is a typical coal-rock dynamic disaster, which has the characteristics of long disaster-pregnant process, sudden occurrence, and strong destructive power, and poses a huge threat to coal mine safety production. At present, microseismic monitoring has become one of the important means of monitoring and early warning of rock burst in coal mines. Research in this area has been carried out for many years at home and abroad, and fruitful results have been achieved. However, due to the complexity of the mechanism of rock burst, the existing Most of the microseismic monitoring and early warning methods are based on the static statistics of the number and energy of microseismic events, and the monitoring and early warning is carried...

AI Technical Summary

Problems solved by technology

At present, microseismic monitoring has become one of the important means of monitoring and early warning of rock burst in coal mines. Research in this area has been carried out for many years at home and abroad, and fruitful results have been achieved. However, due to the complexity of the mechanism of rock burst, the existing Most of the microseismic monitoring and early warning methods are based on the static statistics of the number and energy of microseismic events, and the monitoring and early warning is carried out by setting the threshold, and the threshold setting needs to comprehensively consider various factors, and sometimes even need to continuously adjust relevant parameters, which greatly limits the microseismic monitoring and early warning. The popularization and application of the system also greatly reduces the accuracy of the early warning

In other words, these early warning methods do not fully consider the process characteristics of rock burst disaster formation, especially the drift characteristics of microseismic monitoring data streams, which makes their popularization and application difficult and the early warning accuracy is low

[0003] In the process of microseismic monitoring, a large amount of microseismic monitoring data is generated. These microseismic monitoring data exist in the form of data streams and are constantly expanding. However, the use of these data is currently only at the level of microseismic event counting and energy calculation. There is no in-depth analysis and research on the characteristics and laws contained in the data flow, and there is an embarrassing situation where the data is rich but the available information and knowledge are poor.

Images