地质与勘探2026，Vol.62Issue(4)：797-808,12.DOI:10.12134/j.dzykt.2026.04.010
多矿化元素地质冶金学模型动态模拟
Dynamic Simulation of Geometallurgical Model for Multi-Mineralization Elements:A Case Study of the Bayan Obo Deposit
摘要
Abstract
Geometallurgical models underpin the optimization of mineral resource exploitation and production.However,for complex paragenetic deposits with multi-element assemblages and strong spatial heterogeneity,conventional models poorly characterize harmful element constraints,leading to inaccurate ore domain delineation and insufficient quantitative support for ore blending,which fails to meet the integrated precise production demands of exploration,mining and beneficiation.Taking the Bayan Obo deposit as a case study,this work develops a multi-element dynamic geometallurgical simulation method incorporating harmful element evaluation.Within the unified Surpac platform framework,geological models,reserve estimation models and real-time stope production data are integrated.Ordinary Kriging is used for the spatial interpolation of TFe,mFe,F and S,and a Bayesian inference framework based on Gaussian Mixture Model and Markov Random Field is established to achieve iterative optimization of ore domain delineation.A total of 61,748 three-dimensional grid blocks(2 m×2 m×14 m,75%unknown)are adopted,and the orebody is classified into seven ore domains using TFe cut-off gradients(15%,20%,30%,45%)and a 14%sulfur threshold.Iterative optimization is implemented via an 18-neighborhood system with an optimal smoothing parameter β of 0.3.The initial ore domain variation rate of 0.8 converges to below 0.38 after six iterations.The average TFe and S contents of optimized domains range from 12.24%～45.48%and 0.882%～2.082%,respectively,with high-sulfur and high-iron ore domain 2 possessing the highest sulfur gradient in the mining area.Cross-validation on 566 typical samples confirms that iterative optimization significantly improves ore delineation accuracy.Successfully applied to intelligent production design,precise ore blending and unmanned equipment collaborative scheduling,the proposed model realizes the transition from static delineation to dynamic refined modeling.It provides a technical reference for the refined exploitation and intelligent construction of mines for complex paragenetic deposits subject to harmful element constraints.关键词
地质冶金学/多矿化元素/有害元素评价/矿域划分/动态模拟/白云鄂博矿床Key words
geometallurgy/multiple mineralization elements/deleterious element evaluation/ore domain delineation/dynamic simulation/Bayan Obo deposit分类
天文与地球科学引用本文复制引用
杨楠,张丽,田琦,刘峰,王国栋,尹世滔,刘小玲,陈进,李以科..多矿化元素地质冶金学模型动态模拟[J].地质与勘探,2026,62(4):797-808,12.基金项目
中国地质科学院矿产资源研究所科技成果转化项目"白云鄂博铁矿精细三维地质动态模型构建及其智能化采矿综合研究"(编号:HE2547)资助. （编号:HE2547）
