geometallurgical& geochemical correlation modeling
GRE applies a unique approach to mapping the relationship between process plant performance and ore body characteristics. Our methodology allows for a better overall representation of the local variability of metallurgical response and how it relates to geology, geochemistry, and ultimately financial metrics. GRE employs geometallurgical and geochemical modeling using cluster analysis, recursive partitioning, and multiple adaptive regression to help develop metallurgical/orebody relationships. This type of nonparametric statistical modeling allows the flexibility to analyze multiple variables/parameters within the deposit to produce a best fit process model that is utilized to produce an optimum mine plan, production schedule and economic model. This goes beyond the typical univariate analysis that uses one parameter to estimate a predicted value. This type of statistical analysis and modeling involves database construction, data analysis, non-parametric statistical modeling, geostatistics, and is followed by block modeling. In a recent project, GRE used elevation (approximation of the redox, supergene boundary), correlation between metal grades (lead, zinc, and copper), along with the percentage of 4 separate minerals types to create a robust predictive recovery model of a complicated poly-metallic deposit. GRE substantially improved the understanding of the flotation response and removed much of the metallurgical uncertainty of the project.