中文摘要：

为了克服概率积分法原始下沉曲线公式与实测数据拟合度低、求取下沉预计参数偏差大、曲线边缘部分收敛快等缺点,通过理论对比分析,推导出概率积分法下沉曲线修正公式,并进一步提出新的下沉曲线公式（G＆L下沉曲线函数）。以哈拉沟煤矿22407工作面实测数据为例,利用Matlab软件中参数拟合工具箱求得上述3种下沉曲线的预计参数,并进行比较分析。结果表明：G＆L下沉曲线函数提高了数据拟合度及下沉预计参数精确度,解决了理论曲线边缘部分收敛过快的问题,预计结果更接近于现场实际,具有较高的理论和实用价值。

英文摘要：

In order to overcome the disadvantages such as low fitting degree with actual data, big deviationwithin prediction subsidence parameters and rapid convergence at the margin of subsidence prediction curvewhen adopting traditional subsidence prediction based on probability integral method, by theoretical analysis,modified formula of subsidence curve based on probability integral method was concluded, and it is put forwarda new subsidence prediction formula （G＆L subsidence prediction function） further. Based on the actual datafrom the 22407 panel of Halagou coal mine,and using the parameters fitting toolbox within Matlab software,theprediction parameters in terms of these three kinds of prediction formula were obtained, and a further contrastsanalysis was carried out. The results indicate that G＆L subsidence prediction function improves the fitting degreeand parameters prediction accuracy as well as solves the problem of rapid convergence at the margin ofsubsidence prediction curve,which enables the prediction result conform to surveying data, thus G＆L subsidenceprediction function has relatively high theoretical and practical value.