In seeking to improve efficiency in the mine to mill process, mining companies are incorporating ore processing variables during ore planning and scheduling stages to improve decision making. Therefore, the accurate modelling of these ore processing variables has become more critical.

This case study presents a D-vine copula modelling process applied to predict recovery (Rec) from three geometallurgical variables, that is, Bond Work index test (BWi), resistance to abrasion and breakage index (A*b), and semi-autogenous grinding (SAG) power index test (Spi), using a dataset consisting of 775 diamond drill core samples of the four geometallurgical variables from the Tarkwa Mine in Ghana. The distributions of the four variables were used to create a geometallurgical model to predict values of Rec based on SPI, BWi, and A*b using a 10-D version of copula-based quantile functions, which can be integrated in the mine planning and scheduling process. The module produced an average predicted recovery of 89.26% compared to an observed average recovery of 88.64%. A mean absolute error of 4.41 was obtained indicating an acceptable module performance considering the observed recovery values, which suggests that the model's predictions closely align with the observed values.

This modelling framework enables the creation of a geometallurgical block model that not only improves the prediction of metallurgical variables but also significantly improves the integration of geological and metallurgical information, thereby optimizing processing outputs, project revenue, and cash flows for mining companies.