Research on key points deficiency of three subsidence time series models combined with DInSAR technology

In view of the problem that it is difficult to comprehensively study the dynamic process of deformation in subsidence area and some key points may be missed by single monitoring method in the stage of surface subsidence,exponent Knothe model, Logistic model and Weibull model were established respectively bycombining InSAR and leveling data ofsurface single point to analyze the dynamic subsidence law in mining area. The detailed steps can be described as follows. Firstly, the deformation characteristics of surface observation points in subsidence area and the factors affecting the fitting of the above three time function curves were analyzed by simulation data. Then, taking a mining area of Yuaner coal mine in Huaibei as an experimental zone, the deformation valuestime series of InSAR at each benchmark were obtained by differential synthetic aperture radar interferometry (DInSAR) technology. Finally, the InSAR and leveling monitoring data were combined to compare and analyze the difference of the fitting curves of the three models based on the multi-source or single moni〖LM〗toring method. The test results show that:① The parameters to be estimated for the surface subsidence of each time series model have a great influence on the curve fitting shape.Only by ensuring a certain monitoring period and inverting the appropriate subsidence parameters can the subsidence be fitted more accurately. For the subsidence time series of observation points in the area, the root mean square error can be increased by up to 95.56%. ② The changing process of the power exponent Knothe time model, Logistic model and Weibull model changing with time series is consistent with the time series process of the physical subsidence of the observation points in the surface subsidence area. Although the three models have little difference in the fitting effect of subsidence curve, there are obvious differences in the fitting of subsidence velocity curve and the fitting accuracy has been improved. It can be seen that the settlement time series combined with InSAR and leveling has the advantages in the fitting method of the dynamic settlement of the mining area surface. ③ The fitting three models based on multi-source measurement are more reliable than those based on single leveling data, and have a better convergence effect on the subsidence curve about each model.