Study On The Related Theory And Modeling For Coal Mining Deformation Hazard Monitoring

Deformation has been one of the most serious potential hazards in the mine. There are many methodologies to monitor this hazard. We present a method that mainly focus on large area monitoring, and built the datum of 3D datum monitoring algorithm, and proposed a GPS datum construction method. We concluded a model to monito the deformation based on GPS observation, and had a good result when put it for practice in the mine production field.Based on the dynamical deformation monitoring data, we present a modified model of multi-path EMD-WAVELET, and studied the variants and fractual characters for more information. Then we built a gray forecast model and a GA-BP model. Finally we used it in the shaft deformation monitoring successfully.The main research results we made include:(1) We studied the key theory and technology of the 3D deformation datum construction. Reliability theory was introduced to evaluate the reliability of the datum. We studied the unrestricted robust model and robust coordinates translation model of the GPS baseline network, and assembled all program works to a software, and introduced the data of GPS control network in Yanzhou, Inner Mongolia and Shanxi mine to test this software.(2) We studied different algorithms to establish the datum model and present a geoid purify model based on support vector machine and GA-BP algorithm. We established the height datum of Inner Mongolia and Shanxi with the proper parameters.The support vector machine algorithm based on the restriction of externally coincident precision was present in these experiments.(3) The quick update technology based on the collocated deformation parameters from the height datum and DEM algorithm was present in this paper. Firstly we proved the practicability of GPS height datum deformation monitoring in theory, We studied the GPS mine deformation parameter collection technology to monitor the deformation for Inner Mongolia mine.(4) GPS muti-path real-time dynamic extract model was proposed, and found that the model of EMD-Wavelet is more efficient. We successfully used this technology to modify the real-time multi-path affection in a shaft tower.(5) We present a variance function, and analyzed the characters dimension based on the deformation data. Then we put it used in the forecast of deformation. GA-BP model was adapted to forecast the deformation of the shaft tower in the mine, and the result showed that the precision had been significantly improved.