Validation Study of Linear and Nonlinear Kriging Estimator on Ore-waste Grade Control Process

Facing to unsatisfied results in grade-tonnage estimation especially in dynamic programming is always being a great problem in mining revenue operation. If the problem is to estimate a grade point only, linear kriging estimators can show accurate results. But if target is to achieve to probability distribution estimation of a spatial zone for considering ore-waste block mixing control, using linear kriging methods with minimum estimation variance cant be applied for an appropriate results. Most of probability function is nonlinear, therefore estimation of these function by nonlinear estimator showed an accurate results. Main target of this paper is to achieve to the most exact ore-waste boundaries in 2462.5 benchmark of Sarcheshmeh copper mine using indicator kriging (IK) as nonlinear estimator and comparing with ordinary kriging (OK) as linear estimator to evaluate validity of linear estimator. Because of OK dependency to normal distribution data for a given minimum estimation variance, utility data have been separated to ore and waste group using geological map and mine-sight. After this separation ore groups was approached to normal distribution and OK estimator can be applied for estimation. 25629 blocks were estimated by these two kinds of estimators. IK estimator classified 2905 blocks of total blocks as waste blocks, but OK estimator showed 2475 blocks as waste block. Finally IK estimator recommended as best estimator for ore and waste block separation and after this process using ordinary kriging estimator almost gave more confident estimation in ore blocks grade control process.