Research On Monitoring And Forecasting Method Of Dynamic Mining Subsidence In Mining Areas Based On SAR Image

Underground coal mining leads to the movement and deformation of overlying rock layers and surface,causing geological disasters and environmental pollution,such as damage to surface buildings,land subsidence,soil erosion,etc.,which seriously affects the production and life of mining area residents.Therefore,monitoring the deformation caused by coal mining is crucial.Differential interferometric SAR(D-InSAR),as a ground observation technology,has the characteristics of all-weather,unaffected by clouds and rain,high accuracy,and low monitoring cost.It is an effective method for obtaining large-scale deformation.However,due to the influence of image incoherence,when the subsidence of the mining area is too large,the monitoring results may be too small or deformation information cannot be obtained.At the same time,traditional time-series InSAR mainly uses linear deformation phase models to calculate surface deformation,but the surface subsidence caused by mining in mining areas is mainly nonlinear,so the final deformation obtained by existing methods will underestimate some of the nonlinear deformation.In response to the above issues,this thesis combines D-InSAR and OffsetTracking(OT)methods to study dynamic mining subsidence monitoring and parameter inversion methods,improving the accuracy of traditional surface deformation monitoring.The main research work and achievements are as follows:(1)Summarized the research status of SAR(Synthetic Aperture Radar),InSAR(Interferometric Synthetic Aperture Radar),and mining subsidence prediction models,elaborated on the basic principles of obtaining deformation using D-InSAR,time-series InSAR,and Offset-Tracking methods,analyzed several time function models and their numerical characteristics at feature points,and pointed out the problems of existing surface deformation research methods.(2)The method of obtaining time series settlement of large deformation area in mining area is studied.According to the coherence and prior weights,the deformation levels of large deformation areas were divided,and the surface deformation data obtained by D-InSAR and OT were fused,so as to obtain the temporal deformation monitoring results of mining subsidence basins.Taking Daliuta Coal Mine with image time span from November 25,2011 to April 24,2012 as the research object,the surface temporal deformation of the mine was obtained and analyzed,and the effectiveness of the fusion method was verified.(3)A dynamic prediction parameter inversion and prediction method for mining subsidence based on SAR temporal deformation is proposed.This method uses the Weibull time function as the surface single point time series deformation function,and the surface settlement results obtained by SAR technology as constraints.The PSO(Particle Swarm Optimization)algorithm is used to solve the prediction model parameters and predict the cumulative settlement of different time periods.Taking Daliuta Coal Mine as the research object,the root mean square error between the predicted trend and dip deformation using this method and the GPS monitoring results is 0.064 m and 0.015 m,with a maximum difference of 0.472 m and 0.321 m.This indicates that the proposed method has good prediction accuracy and provides a new method for dynamic prediction of mining subsidence in mining areas.(4)A nonlinear deformation model for time-series InSAR based on dynamic mining subsidence function is constructed to address the issue of underestimating some nonlinear deformations due to the current use of linear deformation phase models in time-series InSAR.This model is based on the Weibull dynamic mining subsidence function and constructs a time-series InSAR nonlinear deformation model based on the relationship between differential phase and deformation.At the same time,the differential phase obtained from the distributed scatter InSAR(DS-InSAR)is used as the true value of the model,and the model parameters are inverted using the PSO+LM(Levenberg Marquardt)algorithm.The constructed model is used to obtain the surface deformation caused by mining in the mining area.Taking the Nantun coal mine and Shilawusu coal mine as research objects,the method presented in this thesis is used to calculate the surface point deformation of the two coal mines.After comparing the results of the traditional linear model DS-InSAR and the measured data,it was found that compared to the DS-InSAR method for calculating deformation,the relative error between the deformation values of the mining area obtained by the proposed model and the measured values is relatively small.The correlation between the surface deformation obtained from the Nantun coal mine and the measured data reached 0.96.The root mean square errors between the deformation obtained by this method and DSInSAR and the measured data were 13.4mm and 39 mm,respectively.The effectiveness of the proposed method has been verified.This thesis has 36 figures,7 tables,and 114 references.