Modification And Prediction Of Surface Deformation Monitoring Data In Mining Area Based On Improved Grey Neural Network Model

Due to the effects of surface deformation,climate,and human factors,monitoring points may be damaged or observations may not be performed,which results in the lack of monitoring data of surface deformation in mining areas.When the discontinuous data is longer,it will have a greater impact on the prediction of surface deformation trends.At present,there are many methods for processing deformation monitoring data,such as gray system theory,regression analysis,artificial neural networks,etc.The grey model is suitable for small sample data,while regression analysis and neural network are better for large sample data processing,but each treatment method has its limitations.According to the finiteness and development trend of surface deformation data in mining area,this paper studies the grey neural network combination model.The paper first analyzes the deficiencies of unequally spaced grey models and BP neural network models.According to the characteristics of large-scale dispersion of surface deformation data and small sample data,a grey neural network combination model was proposed.The ground deformation data of the 22001 working face in the Chaohua Mining Area was used to simulate the discontinuity of mine monitoring data.The grey model,neural network and combined model were used to interpolate the discontinuous data.The interpolation model of the combined model has higher precision,basically reflecting the trend of surface deformation in the mining area,and has obtained satisfactory results.Then,the improved combination model is applied to the processing of discontinuous data of railway deformation monitoring in the mining area,the interpolation data is filled into the original data,and the surface deformation map of the mining area is plotted to analyze the surface deformation characteristics of the mining area.Check the applicability of the improved combination model to the processing of deformation data in mining areas.The research shows that the gray neural network combination model is feasible for the intermittent data interpolation of surface deformation monitoring in mining areas.The short-term prediction of surface deformation data conforms to the surface deformation law of the mining area.