Application Of Improved Kalman Filtering Theory On Surface Deformation Prediction

With the rapid development of economy, resource dried up increasingly, the problemthat mining under three more and more. Compared to coal mining under the roadwaysand the buildings, coal mining under the railway does not allow excessive deformation,and has a higher control standards ground movement and deformation. When miningunder the railway, if the influence range of underground mining on railway line, the thedeformation values and sinking speed of any point on the railway line and so on, can bepredicted in real time, we can plan to and targeted to repair the railroad, make railroadmaintain in good condition.Therefore, the processing method of ground movement anddeformation data is correct or not, has a directly effect on the accuracy and reliability ofrail deformation’s real-time prediction. The domestic and foreign have existing methodand and algorithm of deformation analysis model, there are some limitations on theeffect and the scope of application.Aiming at solving the problem of improving the reliability of movement anddeformation data of surface mining under the railway, and the accuracy of the real-timeforecasts, using the adcantage of wavelet analysis and Kalman filtering in dataprocessing areas, using the methlod of combined with wavelet transform and Kalmanfiltering, the paper improved the traditional Kalman filtering model. First, By usingmethods of wavelet denoising for preprocessing observational data. Then, using Kalmanfilter process the results once more. Through the comparison and the analysis of theexperimental, it is obvious that processing the original observed deformation data byusing improved Kalman filter theory, which can overcome the insufficient that processdeformation observation datas using only single method, can improve the accuracy ofdeformation analysis. According to the high precision of surface movement anddeformation data originate from mining under the railway, using the improved Kalmanmodel for prediction. Through comparied to the measured values, validate it is feasibleto use the improved Kalman filter theory in order to accomplish the prediction of miningunder the railway, this method has higher prediction accuracy and can realize thereal-time prediction.