Development of a new inversion method for detecting spatiotemporal characteristics of coal mines based on earth observation technology

Abstract

We propose a novel methodology for detecting the spatiotemporal characteristics of coal mines based on earth observation technology. Based on the dynamic probability integral model, the spatiotemporal relationship model between underground mining and ground surface response is constructed by fitting dynamic probability integral parameters. Considering the highly non-linearity of the constructed model, an improved fireworks algorithm (IFWA) is introduced to solve the target parameters. Taking the surface above a specific working panel of Guqiao Coal Mine in Huainan, China as the experimental area, based on the short-term surface monitoring data of the observation station, the proposed method was applied to detect the spatiotemporal characteristics of underground mining in six periods. The results show that the detected spatiotemporal characteristics were consistent with the measured ones. However, the large fluctuation of the dynamic probability integral parameters, and subcritical extraction may affect the method. In the simulation experiment, the model error of the construction method for multiple periods obtained was in the range of 0–21.0%, with an average of 5.0%. Furthermore, the sensitivity of the inverted spatiotemporal characteristic parameters to the input probability integral parameters was studied, and the results show that when error percentages of less than 20%, the inversion of target parameters are closer to the simulation (the mean relative errors is approximately 7.7%). Due to the advantages of short-term earth observation, this study provides a practical, convenient method for detecting spatiotemporal characteristics of underground mining, and avoids time-consuming and laborious long-term monitoring.