Study on the Law of Subsidence of Overburden Strata above the Longwall Gob

Abstract

In recent years, with the increase of depth and intensity of coal mining, geological disasters in deep engineering occur frequently. It is essential to study the law of subsidence of overburden after mining for analyzing the mechanism and control of geological disasters in deep engineering. However, the value of the subsurface subsidence factor and other overburden subsidence parameters could not be obtained easily. In this paper, the strata bulking of the caving zone, the bed separation of the fractured zone, and the dilatation of bending zone were considered a kind of expansion in the vertical direction uniformly. An assumption on vertical expansion value distribution of overburden strata above the longwall gob was proposed. Based on the assumption, a formula of the subsurface subsidence factor was deduced according to such parameters as the bulking factor, mining depth, mining height, and surface subsidence factor. Moreover, many field data were used to verify the correctness of the assumption by the data fitting method. The results show that the exponential function could well describe the expansion distribution characteristic of overburden strata and subsurface subsidence in the vertical direction. The parameter $a$ in the deduced formula influences the shape of the subsurface subsidence curve; it is called the subsurface subsidence influence factor (SSIF). The SSIF, maximal vertical expansion value, and the mining depth-to-mining height ( $\text{MD}/\text{MH}$ ) ratio influence the subsurface subsidence factor. The expansion distribution (E-D) factor is defined to describe the subsidence characteristic of overburden strata above the longwall gob by the surface subsidence factor and bulking factor of the caving zone. The E-D factor presents the logarithm relation with the $\text{MD}/\text{MH}$ ratio, and the range of its maximum is 2.84~3.40. The case study demonstrates that the subsurface subsidence factor calculated by the proposed method has higher precision, and the mean errors between the calculated value and the measured value are less than 2%.