Data Assimilation And Prediction Model Of Mining Subsidence Monitoring Data Based On Moving Least Squares

Land subsidence affects people's production and life.It will cause serious harm to people's property and regional economic development.The land subsidence process caused by mining is complicated.It is necessary to collect and collate the ground subsidence monitoring data,study and establish the ground settlement calculation and prediction model,and process and analyze the monitoring data.The moving least squares method is widely used in data smoothing,numerical analysis.Based on the moving least squares method,the data assimilation and prediction methods of mine land subsidence monitoring data are studied,and the ground subsidence monitoring data is calculated,analyzed and predicted.It is of great significance to monitor and forecast ground disasters in mining areas.In this paper,the moving least squares primary function and weight function are analyzed.The polynomial primary function,radial basis function,Gaussian weight function,spline weight function and the effect of radius on the moving least squares are discussed.And summarized the relevant characteristics.The moving least square method is used to calculate and analyze the ground subsidence data in the space and time domains.In the space domain,interpolate the ground subsidence monitoring data based on the moving least squares method.The distance attenuation parameters of the Gaussian weight function are determined according to the spatial distribution of the monitoring points.Gaussian weight functions and polynomial basis functions and radial basis functions are used to construct shape functions in the form of points for spatial interpolation calculations.With the increase of the number of primary function,the interpolation precision is improved,but the shape function is also complicated,the calculation amount is increased,and the ill-conditioned matrix is easy to be generated.Using radial basis functions can avoid a large number of matrix inversions and matrix calculations.The interpolation accuracy is enough high and the calculation efficiency is improved.In the time domain,the ground subsidence monitoring data is fitted based on the moving least square method.Divide the monitoring area.Determine the impact area of each node and the monitoring points involved in the calculation.Use the moving least squares method to estimate the amount of settlement at each node.Trend analysis of land subsidence monitoring data was performed using cumulative settlement and settlement rate.The comparison and analysis between the predicted value and the monitored value verified the feasibility of the method.