Study on large-gradient deformation of mining areas based on InSAR-PEK technology

Abstract

To solve large-gradient deformation in mining areas unavailable by SAR data, a method combining PIM Exponent Knothe (PEK) model and InSAR technology (InSAR-PEK) was proposed to predict the mining-induced subsidence and obtain the large-gradient deformation dynamically. Firstly, the maximum subsidence value predicted by the probability integration method was combined with SAR data, and the subsidence values in the initial and residual periods were obtained. Secondly, three groups of power exponent Knothe function parameters were obtained, including csar and ksar based on SAR data, clevel_wz, and klevel_wz based on leveling data over a complete observation period, and clevel_bf and klevel_bf based on the elimination of the leveling data in the main period. Finally, the predicted values of the three groups of parameters were compared with the measured data, respectively, and the root mean square errors (RMSE) were obtained. The engineering example verified that RMSEs were 28.1mm~91.7mm in the main period and 30.9mm~58.7mm in the whole period estimated by the InSAR-PEK method. The results showed that the subsidence values in the main period were relatively stable by the InSAR-PEK method, and some points' prediction accuracy was better than that of leveling data. The predicted values obtained by the InSAR-PEK method and those extracted by SAR were compared with the measured values. In the main period, the values extracted by SAR differed greatly from the measured values, which were false values. However, the predicted values by the InSAR method were close to the measured values, which can be used to independently get subsidence values in the main period from SAR data.