Abstract
Metallurgical testwork is a key component of processing facility design for the mining industry, and is used at all stages of project development, from Preliminary Economic Assessment to Detailed Engineering. Sample selection for metallurgical testwork programs is typically done manually, selecting the best combination of available samples to represent the potential future plant feed or variation thereof. Due to the difficulty of finding representative groups of samples, which increases with the number of characteristics (grades, lithology, extraction, etc.) to consider, manual selection tends to be subjective, slow, difficult to peer review, and may lead to suboptimal results. Few software packages exist for metallurgical sample selection and present their own challenges. A new metallurgical sample selection mathematical model was developed to address the shortfalls of both the manual selection method and existing software. A case study compares the new model’s selected sampled to those chosen by hand (traditional) as well as using available software is presented.