A geometallurgical model for a pre-feasibility study

The study needed recovery, concentrate quality and grinding behavior in the block model for process design.

The study needed recovery, concentrate quality and grinding behavior inside the block model to support process design. Mineralogy and metallurgical tests were sparse and unevenly sampled, and geometallurgical domains had not yet been defined.

We reviewed data coverage and imputed missing variables stochastically, then defined domains with machine learning on the multivariate data. Metallurgical variables were estimated and process equations built for throughput, energy and recovery.

The result is a geometallurgical block model ready for the study, with a reliability label on every parameter and a handover workshop for the team.