Prediction of rock burst risk in deep coal mining based on time-frequency domain analysis

ACKNOWLEDGEMENTS: This work was supported by ministry of education industry-university cooperative education project (Grant No.: 231106441092432) and special research project on teaching reform (Grant No.: 30120300100-23-yb-jgkt03)

Abstract

Coal is an important energy and industrial raw material in China, and its exploitation plays a vital role in the stability and development of the national economy. This paper will use the data given in the title to explore more efficient data analysis methods and more advanced monitoring equipment to improve the accuracy and reliability of the early warning system. For the identification of interference signals, a detailed data preprocessing and analysis process is adopted. Through data cleaning, duplicate and invalid data are removed and data quality is ensured. Wavelet analysis is used to extract the features of interference signals in electromagnetic radiation and acoustic emission signals by using time-domain, frequency-domain and time-frequency domain methods, effectively identify and further eliminate interference data, and improve the accuracy of signal processing. In addition, support vector machine (SVM), neural network and random forest comparison model are constructed to effectively classify the interference signals. Precursor characteristic signal analysis, analyze precursor characteristic signal and identify the time interval of precursor characteristic signal. First of all, the same data processing method as work 1 is used to clean the data involved in work 2. Combining wavelet transform and statistical analysis method, the key statistical features of the signal are extracted, and a prediction model based on the sliding window of time series analysis is established to calculate the average value, standard deviation and energy of each window. In order to identify precursory feature signals as early as possible, a mathematical model based on historical and real-time data is developed. Smooth moving technology is used to capture the basic characteristics of signals in each window, and combined with the probability of having a small standard deviation in the prediction results, the method that the standard deviation is less than the average predicted value is calculated, so as to establish a prediction probability model and obtain the probability of precursor characteristics at the time of data collection. This model can realize timely early warning of disasters such as rock burst. It provides important guarantee for mine safety.