Based On Wavelet Packet Construction And Application Of Mine Pressure Prediction Model

Based on the large amounts of data obtained from practice, Wavelet denoising method is applied in this article, the scientific research methods to the production of the mine pressure monitoring system is studied. In order to reduce the inevitable artificial factors in the process of the monitoring which has the negative impact on the system of data analysis in the mine pressure on the basis of using wavelet denoising to extract the useful information to reflect the actual circumstances of the mine pressure data; And nonlinear wavelet neural network combination forecast method is introduced into the mine pressure prediction to maximize the accuracy of pressure prediction. The major work of paper is as follow:1, Collecting a large amount of information on wavelet denoising, carries on the induction analysis, research for this article USES the method of wavelet.2, In the practice of a large number of mine pressure monitoring data, remove the outliers. Summarize the useful information together, the use of wavelet packet denoising processing, Through selecting suitable wavelet coefficients, the decomposition level and threshold value of the data signal processing. The processing of the wavelet signal is analyzed, according to the variation of amplitude to determine the location of the discontinuities, and through the theoretical analysis of discontinuous points, to determine the location of the mine pressure mutations.3,Roadway displacement of daily observation data processing, through data interpolation to be perfect. The roadway displacement prediction model based on LIBSVM is established. And through the project example to validate the reliability of the model.