采煤沉陷是指由煤炭井工开采引起的煤矿区地面沉陷、塌陷和山地开裂等地质环境破坏及其诱发的地质灾害。采煤沉陷已经成为导致煤矿区生态环境恶化和制约煤矿区经济、社会可持续发展的主要因素。在采前对采煤沉陷进行准确预计，是通过限制开采控制采煤沉陷的前提。为此，开发出基于地质观点的采煤沉陷预计系统软件。本文主要介绍该系统针对任意形状工作面及急倾斜煤层采煤沉陷预计的设计思路和实际应用。 采煤沉陷预计系统在考虑采矿因素对采煤沉陷影响的同时，将对采煤沉陷有控制作用的主要地质因素引入采煤沉陷预计模型，可以更准确地仿真模拟随着工作面推进，地表移动变形的发生、发展过程。 本系统可以在研究区地质数据库的基础上，自动统计计算影响采煤沉陷的地质因素数据，并引入概率积分模型，实现对任意形状工作面地表采煤沉陷的预计，自动绘制地表任意方向剖面上的下沉、倾斜、曲率、水平移动、水平变形预测曲线图，并可绘制采煤沉陷盆地地表移动变形等值线。将经验公式皮尔森Ⅲ型函数用于急倾斜煤层的采煤沉陷预计，可绘制急倾斜煤层开采时主断面上的各种移动变形曲线。 本文还探讨了遗传规划方法在采煤沉陷最大值预计中的应用。认为在地质和岩移资料足以构建出高质量训练样本的煤矿区，可以将遗传规划作为预计采煤沉陷的一条新途径。 通过利用典型算例进行验算以及矿区实例应用对比，证明本系统预计精度在允许范围之内。本项研究成果，可为采煤沉陷预计和控制提供技术支持，具有一定的推广应用前景。

Large scale mining of underground coal resources causes or brings out hazards of geological environment including sinking, fault and fracture of the ground, which are called coal-mining-induced subsidence. It has become one of main factors which worsen ecological environment and restrict the development of economy and social sustainable of the coal mining areas. The precise prediction of coal-mining-induced subsidence before mining is a premise to control coal-mining-induced subsidence by limiting mining. Therefore, a software to predict coal-mining-induced subsidence based on geological viewpoint is worked out. The design idea and the practical application of the software are mainly introduced in this thesis. The pivot in the thesis is about the prediction of coal-mining-induced subsidence under the condition of random-shape faces and steep coal seams. Being introduced into the prediction model not only mining factors but also main geological factors influencing coal-mining-induced subsidence, the software can simulate accurately the development process of surface movement and distortion with a working face advancing. The software possesses the function of automatical counting the data of geological factors which influence coal-mining-induced subsidence base on the geological database of research area. The software is based on the probability integral principle. With the help of this software any user can obtain the prediction results about ground subsidence induced by coal mining of random-shape faces, and the prediction graphs such as surface subsidence, slant, curvature, level movement and level distortion in random orientation section. The coal-mining-induced subsidence of an arbitrary spot within the subsidence basin can be calculated automatically. The empirical formula Piersen III function is used to predict the surface subsidence induced by steep coal seam mining, and draw the prediction graphs of surface movement and distortion on main sections. It is also discussed in this thesis to use Genetic Programming method in predicting the maximum value of coal-mining-induced subsidence. Genetic Programming is a new feasible way to predict coal-mining-induced subsidence under the condition of geological information and observation data of surface movement are enough to construct a training sample of high quality. Checking with praxises in text books and the typical examples from coal mining areas, the prediction precision of the software can satisfy demands. The software can supply technical support for the prediction and control coal-mining-induced subsidence in mining areas, with definite application value.