College of Energy Engineering, Xi’an University of Science and Technology
Shaanxi Xiaobaodang Mining Co. , Ltd.
Shaanxi Coal Industry and Chemical Technology Research Institute Co. , Ltd.

三维地质模型的构建对于理解和预测地下结构至关重要。地质钻孔数据能够反映岩体空间分布和地质构造特征,本研究以小保当一号煤矿11盘区内的23个地质钻孔数据为基础,采用添加虚拟地层的方法解决了地层缺失与地层重复现象,构建共计27层地层的三维地质模型以及二维剖面模型。此外,针对传统的克里金方法在处理复杂地质数据参数选择困难的问题,采用粒子群算法对传统克里金插值方法中的块金值(C0)、偏基台值(C)和变程(a)三个关键参数进行寻优,从而克服普通克里金插值参数选择的主观性和不确定性,采用实际验证法选取了研究区内四个钻孔来对比插值结果,结果表明经过PSO优化的Kriging算法在X3-1、X3-2、K3-4、K3-5四个钻孔的RMSE值分别降低至1.184、1.267、1.606、1.560,相比于Kriging的RMSE平均降低了31%,且PSO-Kriging算法在四个钻孔处对2-2煤层的插值结果与实际值相比较误差分别为1.00m、0.01m、0.11m和0.03m,比Kriging插值结果更接近实际值,表明了所提方法的有效性和优越性。

The construction of 3D geological models is crucial for understanding and predicting underground structures. Geological borehole data can reflect the spatial distribution of rock bodies and geological structure characteristics. Based on geological boreholes data in the panel areas of Xiaobaodang No. 1 Coal Mine, the method of adding virtual strata is applied to solve the phenomenon of missing and repeated strata, and a total of layers of 3D geological model as well as 2D section model are constructed. In addition, aiming at the difficulty in parameters selection of the traditional Kriging method in complex geological data processing, particle swarm algorithm is used to optimize the three key parameters of traditional Kriging interpolation method, namely, the block gold value (C ), the biased abutment value (C), and the variable range (a), so as to overcome the subjectivity and uncertainty of the parameter selection of the ordinary Kriging interpolation; and actual validation method was utilized using data of four boreholes from the study area to compare the interpolation data. The results show that the RMSE values of the PSO-optimised Kriging algorithm at the four boreholes X3-1, X3-2, K3-4, and K3-5 are reduced to 1. 184, 1. 267, 1. 606, and 1. 560, respectively, which is an average reduction of 31% compared with that calculated by Kriging RMSE. The interpolation results of PSO-Kriging algorithm for No. 2-2 coal seam at four drill holes have an error of 1. 00 m, 0. 01 m, 0. 11 m and 0. 03 m compared with the actual values, which are closer to the actual value than Kriging interpolation results, indicating the effectiveness and superiority of the proposed method.

国家自然科学基金面上项目(52074209,51874232);陕西省自然科学基础研究计划联合基金项目(2021JLM-06)

丁自伟,刘江,王小勇,等.基于PSO-Kriging算法的三维地质建模技术研究[J].煤炭工程,2024,56(10):82-89.