煤矿三维地质模型是指导煤矿安全高效生产的重要前提，对实现煤炭精准开采和地质保障具有重要意义。特别是在复杂地质条件下，地质模型若不能准确描述构造和地层之间的关系，将会影响到煤矿生产过程中的采区规划、采掘工艺及生产效率等方面。

以内蒙古桌子山煤田棋盘井煤矿为例，结合现场的地质资料、断层成果和验证结果，开展高精度的煤矿三维地质建模技术研究。首先，通过空间配准、交叉验证和联合反演等手段对地质数据进行融合，以获取更为精确的建模数据；其次，通过沉积构造规律研究，掌握沉积演化特征和构造发育规律；然后，通过不规则三角网−广义三棱柱(TIN-GTP)算法构建三维地质模型。

结果表明：基于该算法的建模技术能够有效、准确且快速地模拟任意复杂地质体，为处理地层不整合、地层尖灭、煤层分叉和断层切割等问题提供了便利。同时，构建了棋盘井煤矿矿井、采区、工作面及其顶底板等不同尺度地质模型，为实现矿井地质透明化智能开采提供了重要数据支撑。经验证，模型误差值均处于0.2 m以下，且0~0.1 m范围内占整体95%，表明该建模技术具有先进性与合理性。研究成果对于提升煤矿生产的安全性、效率性和可持续性具有重要意义。

A 3D geological model of a coal mine serves as an essential prerequisite to guiding the safe and efficient production of coal mines, holding great significance for precise coal mining and geological guarantee. Especially in complex geologic conditions, the failure of geological models to accurately describe the relationship between structures and strata will affect the mining area planning, mining processes, and production efficiency during the production of coal mines.

With the Qipanjing Coal Mine in the Zhuozishan Coalfield, Inner Mongolia, as an example, this study investigated the technology for high-accuracy 3D geological modeling of coal mines based on field geological data, fault results, and verification outcomes. First, geological data were integrated using methods like spatial registration, cross-validation, and joint inversion to obtain more accurate modeling data. Second, the sedimentary evolutionary characteristics and structural developmental patterns were determined by investigating the sedimentary tectonic patterns. Third, 3D geological models were constructed using the triangulated irregular network-generalized tri-prism (TIN-GTP) algorithm.

The results show that the TIN-GTP algorithm-based technique for modeling can simulate any complex geological bodies in an effective, accurate, and rapid manner, creating favorable conditions for addressing issues such as stratigraphic unconformities, stratigraphic pinch-out, coal seam bifurcation, and fault cutting. Furthermore, geological models on varying scales like the mine, mining area, and mining face combined with its roof and floor were built for the Qipanjing Coal Mine, providing significant data for the geological transparency and intelligent mining of coal mines. It has been verified that the errors of all the models are below 0.2 m, with 95% falling within the range of 0 to 0.1 m, demonstrating the advanced and rational nature of the modeling technique. The research findings are of significant importance for enhancing the safety, efficiency, and sustainability of coal mining operations.