Deformation Monitoring And Prediction In InSAR Mining Area Considering Single Point Evolution Law

In the past 40 years of reform and opening up,the coal industry has undergone tremendous changes,and coal mining has gradually turned to large-scale development.However,during the process of coal extraction,the structure of the overburdened rock formation can easily be damaged,which in turn causes the surface to move and deform,threatening the ecological environment on the ground and the safety of the buildings above.Therefore,accurate monitoring or even predicting the surface deformation caused by coal mining in advance has positive significance for the prevention and control of damage to ground structures and ecological environmental protection.At present,to overcome the problems of low spatial density,time-consuming and laborious monitoring by traditional measurement methods,and high cost,Interferometric Synthetic Aperture Radar(InSAR)technology has been widely used in the monitoring of surface deformation in mining areas.Its advantages,such as all-day and all-weather,provide a new perspective for monitoring the timing and surface area of the surface of the mining area.However,a large number of mining subsidence studies have found that surface deformation caused by underground mining has a certain time and spatial distribution law.As a small unit for the study of the law of time and spatial distribution,the surface single point is the basis for the understanding and reliable prediction of the law of subsidence in mining areas through reasonable modeling of the surface single point.However,at present,there are many problems in the research on single-point mobile deformation of the surface.The modeling adaptability is not strong,and it is difficult to combine with related monitoring/forecasting methods,which restricts its further application in the modeling,monitoring,and forecasting of surface deformation in mining areas.Because of this,this paper has carried out a series of mining subsidence monitoring and prediction studies based on the law of single-point evolution and combined with InSAR technology.Its purpose is mainly to optimize the selection of time functions that can describe the deformation process of single-point movement of the surface,improve the surface deformation monitoring method of the InSAR mining area,and construct a suitable dynamic prediction model of InSAR mining subsidence.The main work and innovation points of this article are:(1)To solve the problem of "over-fitting" due to the lack of consideration of function model complexity in existing single point modeling research of coal mining area,an optimization method of single point dynamic settlement time function in coal mining area based on AIC criterion was proposed.The optimal selection of time functions was analyzed in this paper using two indicators of fitting residual and model complexity.More specifically,time-series subsidence observations at 103 field points in seven coal mining areas with different geological mining conditions were selected to be observation samples for ensuring the applicability of the optimal time function.Then,12 common “S-shaped”growth models were chosen for candidates,and the theoretical analysis and Akaike Information Criterion(AIC)were further used to analyze the optimal selection of time function from the chosen 12 “S-shaped” models.The research results deepen the understanding of the evolution law of ground single point and time function while providing suggestions for the application of time function in mining areas with different geological conditions.(2)Since the current monitoring of surface deformation in the timing InSAR mining area rarely considers the nonlinear factors inherent in the deformation of the mining area,a timing InSAR mining area deformation monitoring method based on the optimization time function is proposed.The method first optimizes and selects the time function based on the geological mining conditions of the mining area,establishes a function model between the single-point time function and the InSAR interference pair,uses the relevant nonlinear algorithm to estimate the model parameters,and finally calculates the time series deformation of the mining area.This part takes a mining area in Datong City,Shanxi Province as the research object.The verification of GPS single-point monitoring results shows that the algorithm’s single-point maximum deformation value monitoring accuracy is 15 mm higher than that of the traditional timing InSAR algorithm.In addition,based on the proposed method,the spatiotemporal distribution law and physical properties of the model parameters are fully explored.By drawing the distribution of the model in time and space,the interrelationship between the parameters of the model and the propulsion of the working surface is visually demonstrated,and the mathematical connection between it and the eigenvalues of the surface movement deformation is determined by combining regression analysis.(3)Given the current use of empirical time function and probability integral model to construct a dynamic prediction model of mining subsidence in mining areas,there are problems such as limited forecasting efficiency and accuracy due to the weak adaptability of the time function and the complexity of the model.Constructed a dynamic prediction model of subsidence in mining areas which based on the optimization time function and probability integral model,and the sensitivity of each parameter of the model is determined based on the global sensitivity analysis method,and the dynamic prediction model is optimized firstly.Then based on the InSAR monitoring information,the model parameters that cannot be directly obtained are inverted,and the three-dimensional deformation characteristics of the mining area caused by subsequent mining are predicted.On the one hand,the final research results show that global sensitivity analysis can simplify the constructed model and facilitate subsequent parameter inversion work.On the other hand,the next stage of deformation predicted by the model is consistent with InSAR observations,the root mean square error is 0.034 m after extracting the profile,which proves the feasibility of this method.