A new prediction model for mining subsidence deformation: the arc tangent function model

Abstract

Ground subsidence in underground coal mining areas causes environmental damage and creates hazards on the ground surface, which is long-term, widely distributed, and can lead to large-scale geological disasters. Achieving a high-precision method to predict mining subsidence deformation is very important for assessing environmental damage and countermeasures. In this paper, based on the “S”-type settlement curves of the monitoring points in the collapsed pit and the failure mechanism of rock strata on the goaf, the arc tangent function model was proposed and applied to the Taihe coal mine in Fushun, Liaoning Province, China. Using the Levenberg–Marquardt algorithm for nonlinear curve fitting of the data, the parameters of the model are obtained, and extending it in time, the prediction function will be obtained. Using different monitoring data to validate the model shows that the accuracy of the medium- and short-term forecasting is very good. With continuous updating of the monitoring data, the forecasting achieves higher accuracy and the function of dynamic track forecasting is achieved. A very high correlation coefficient was obtained (0.996) using all the available data from the monitoring point for the best-fit curve. This prediction model provides a reference for the evaluation and treatment of ground subsidence in the Taihe coal mining area.