关键词: 克里金插值, 粒子群算法, 三维地质建模, 地质统计学, 空间插值

Abstract: The construction of three-dimensional geological models is crucial for understanding and predicting underground structures. Geological drilling data can reflect the rock mass Based on the spatial distribution and geological structure characteristics, this study uses 23 geological drilling data from the 11th panel of Xiaobaodang No.1 coal mine as the basis, and adopts The method of adding virtual strata solves the problems of missing and duplicated strata, and constructs a three-dimensional geological model of a total of 27 strata, as well as two Dimensional profile model. In addition, in response to the difficulty of parameter selection in processing complex geological data using traditional kriging methods, this paper adopts a granular approach The subgroup algorithm is applied to the block values (C0) in traditional kriging interpolation methods Optimize the three key parameters of off base station value (C) and variable range (a), In order to overcome the subjectivity and uncertainty of parameter selection in ordinary kriging interpolation, the actual verification method was used to select four boreholes in the study area To compare the interpolation results, the results show that the Kriging algorithm optimized by PSO performs well in X3-1 The four boreholes X3-2, K3-4, and K3-5 The RMSE values decreased to 1 184, 1 267, 1 606, 1 560, an average reduction of 31% in RMSE compared to Kriging, and PSO Kriging algorithm applies to 2 at four drilling locations -The interpolation results of coal seam 2 have an error of 1 compared to the actual values 00 meters 0.01 meters 0.11m and 0 03 m, which is closer to the actual value than Kriging interpolation, indicates the effectiveness and superiority of the proposed method.