Research On Surface Subsidence Monitoring And Trend Prediction In Mining Area Based On InSAR

Coal resources are an important support for the development of industrial modernization.With the increasing demand for energy and the continuous exploitation of coal resources,the induced disaster problems have become increasingly prominent.Mining surface subsidence is one of the typical disasters.The problem of mining subsidence makes the geological problems such as soil erosion,farmland damage and vegetation damage in the mining area increasingly serious,resulting in the collapse of buildings,railways and highways,which restricts the sustainable development of resource-based cities.Therefore,it is of great significance to realize the accurate analysis and prediction of the development law of surface deformation,and it is also an urgent problem to be solved in the current industry.In order to solve the shortcomings of traditional surface subsidence monitoring methods,this paper applies D-In SAR technology to the subsidence of 6306 working face in Dongtan Coal Mine,and combines leveling technology to achieve complementary advantages between data.The subsidence law and subsidence prediction of the study area are analyzed and studied.The main research results are as follows :(1)By analyzing the mechanism of surface subsidence caused by coal seam mining and summarizing its basic characteristics,it is found that the surface subsidence caused by coal seam mining is the final manifestation of the process of surrounding rock stress transfer and rock fragmentation on the surface.According to the different causes and subsidence areas,it is manifested as subsidence basin,collapse pit,ground fissure and step deformation.The numerical simulation method is used to obtain the surrounding rock stress,plastic zone and vertical displacement.It is found that the vertical stress is positively correlated with the coal seam advancing distance.The roof of the coal seam initially experienced tensile failure,and then gradually turned into shear failure.Finally,the roof overburden experienced bending and collapse,and the plastic zone of the roof and floor of the coal seam continued to expand.The simulation results of vertical displacement are compared with the measured data of leveling,and the two are in good agreement with the trend of vertical displacement.(2)This paper introduces the application background and advantages of In SAR technology,expounds the basic principles of In SAR technology and D-In SAR technology,and analyzes the processing methods,platforms and processes of D-In SAR technology.The results show that D-In SAR technology can monitor the surface deformation with high precision by analyzing the phase difference between two or more synthetic aperture radar images,and the surface subsidence monitoring based on DIn SAR technology has high feasibility.In the actual operation of this paper,ENVI and SARscape software are used to process Sentienl-1A data and ASTER GDEM data.The process includes data clipping,baseline estimation,interferogram generation and deleveling effect,adaptive filtering and coherence calculation,orbit refinement and phase unwrapping,phase-to-deformation and geocoding.(3)The interference processing of 18 scenes of Sentinel-1A image data in the survey area was completed by using the ’ two-track method ’,and 17 interference results were obtained for time series analysis to obtain the surface subsidence law of the mining area.The results show that the surface subsidence value is small in the early stage of working face mining,the development of subsidence basin is relatively gentle,and the maximum settlement is only 10 mm;with the continuous advancement of underground working face mining,the maximum subsidence magnitude of surface subsidence has been increasing to 1900 mm,and the influence range of subsidence has been expanding.The subsidence funnel is formed in the area with serious subsidence,and the funnel moves from southwest to northeast,and the extension direction is consistent with the mining direction of the working face.By comparing and analyzing the measured data of the qualified level and the corresponding D-In SAR technology monitoring results,it is found that the changing trend of the two is basically consistent,showing a strong positive correlation.There is only millimeter error in accuracy.(4)The deformation monitoring and prediction of mining area based on D-In SAR are carried out by grey GM(1,1)algorithm,SVR algorithm and LSTM algorithm respectively,and the prediction results are evaluated by root mean square error RMSE and mean absolute error MAE.The results show that the fitting trend of each model is basically consistent with the actual trend,and the prediction results are acceptable.The maximum error of the grey GM(1,1)model in the prediction process is 33.3 mm,the maximum error of the SVR algorithm in the prediction process is 17.2 mm,and the maximum error of the LSTM algorithm in the prediction process is 3.6 mm,showing a good prediction effect.The fitting degree of the three models is LSTM model > SVR >grey GM(1,1),and the RMSE and MAE of the LSTM algorithm are 2.27 and 2.08,respectively,which are the minimum values of the three models.It is proved that the prediction results of LSTM are obviously better than SVR model and grey GM(1,1)model.