序贯高斯模拟在矿石品位估计中的应用研究

Application of Sequential Gaussian Simulation in Ore Grade Estimation

随机模拟是地质统计方法的重要内容。在矿石品位估计方法中克里格方法作为一种无偏估计方法,常被用于矿石品位的估计。但克里格法估值存在平滑效应。作者在分析序贯高斯模拟和普通克里格法基本原理的基础上,运用序贯高斯模拟方法和普通克里格方法对某铁矿体内全铁(TFe)品位进行估计,给出了品位估计结果模型。研究从勘探线方向、垂直勘探线方向和竖直方向分别计算变差函数,对球状模型、指数模型、高斯模型的变差函数拟合效果进行了优选,结果表明球型模型拟合效果最好。针对序贯模拟和克里格品位估值效果进行了分析,结果显示:序贯高斯模拟结果在品位分布形态上更接近样品品位分布形态,其平滑效应更小;克里格方法估计与序贯高斯模拟方法相比仅在品位均值方面更接近样本品位均值。因此,认为序贯高斯模拟方法可以更好地刻画矿体内品位分布状态。

Stochastic simulation is an important method of geological statistics.The Kriging method is a unbiased estimation method,ften used for ore grade estimation with smoothing effect.Based on analysis of principle of Sequential Gaussian Simulations and Ordinary Kriging method they are used to estimate the TFe grade of a iron ore body and the grade model is given.The variation function is calculated from the direction of the exploration line,the vertical exploration line direction and the vertical direction.The variation function is fitted with the spherical model,the exponential model and the Gaussian model.The fitting effect of he spherical model is the best.The result show that distribution of the Sequential Gaussian Simulation is closer to the distribution of sample grade,with smaller smoothing effect.Compared with Sequential Gaussian method only mean grade value of the Kriging method is closer to to that of the sample.Therefore,the authors believe that the Sequential Gaussian Simulation can better characterize the distribution of ore in the ore body.