Assessing Clay-Related Geometallurgical Dilution in Open-Pit Copper Deposits Using Geostatistical Modeling

Abstract

The thesis explores the geometallurgical performance of porphyry copper deposit with regard to spatial distribution of clay and its implications to open-pit mining. Clay-bearing material has been considered as one of the key factors that affect the efficiency of mineral processing, especially in flotation, because it affects slurry rheology and flotation recovery. Conventional ore appraisal models generally focus on metal grade and tonnage, and overlook the heterogeneity of important geometallurgical variables like clay that can cause large discrepancies between geological estimates and the actual performance in minerals processing. This study fills the gap between sample-based data and mining block-based geometallurgical interpretation. In order to do this, a geostatistical workflow was created based on the drillhole based clay data. The analysis entailed exploratory data analysis, capping, declustering, and variogram modeling to determine spatial continuity. The kriging of clay values was done on a fine grid of 2 x 2 x 5 m, which was then upscaled to a coarse block model of 20 x 20 x 10 m. The findings were tabulated in two important products: average clay content and a probability map of clay ranges in each block. The mean map provides a smooth average of the value, and the probability map indicates the percentage of the block that meets a chosen clay cutoff, giving an idea of the internal heterogeneity of the block and the possibility of clay-related dilution. The main results are the validation of the spatial variability of clay, which has been proved by the anisotropic spatial continuity and the existence of various continuity ranges in both horizontal and vertical directions. The effect of extreme values was successfully minimized by capping, and the declustering was not essential since the effect of sample spacing on global statistics was very small. The cross-validation showed that the interpolation model was very accurate, and there was a high correlation between estimated and actual values of clay (0.87) and low bias. The thesis provides a useful geostatistical model that can be used to obtain block support representations of clay estimate that can be used in mine planning and processing. The results can be used to make more informed geometallurgical choices in open-pit mining by taking into account the average clay content and the fraction of the block affected by clay. The study emphasizes the need to consider the heterogeneity of clay during mining and offers a methodology, which can be used to consider other similar geometallurgical variables in other deposits. Future studies may combine metallurgical response information with spatial clay models to make better predictions of the processing and consider other modeling methods that can better represent local variability.