The Primary-Response Framework for Geometallurgical Variables

Geometallurgy is an emerging, cross-disciplinary field that integrates spatial models of rock properties with time-based outcomes of mining and treatment processes. Geometallurgical variables include any rock property that has a positive or negative effect on the business. Some of the more critical geometallurgical variables include recovery, grindability, throughput, power consumption, mineralogy and content of deleterious materials. These variables drive project costs and revenues in a fundamental way and thus geometallurgy has potential to positively impact on the value of strategic and tactical decisions across many mining organisations. Geometallurgical models, like traditional grade-only models are based on samples collected from the mineralised and non-mineralised regions of the deposit. The nature of some non-grade variables, particularly those that are dependent on the ore treatment processes, requires special consideration during sampling and subsequent spatial modelling. Failing to objectively consider the characteristics of variables may seriously compromise the validity of the sample selection, sample treatment, data models and the subsequent decisions. In addition to addressing the vexed question of collecting an æunbiased and representativeÆ sample, the sampling and modelling approach must consider issues of scale, both geostatistical support and also the relationship between bench-scale data and operational-scale mineral processing performance. This paper proposes a framework (the Primary-Response framework) for the classification of geometallurgical variables. This framework is designed to assist with developing sampling approaches and identifying the most appropriate spatial modelling approach. The proposed framework can also help identify the risks associated with designing, sampling and modelling of both types of geometallurgical variables. The Primary-Response framework divides variables into two categories; primary variables and response variables. How a variable is classified in this framework depends on the degree to which the variable reflects either an intrinsic attribute of the rock (æprimaryÆ) or its response to measurement processes (æresponseÆ). Depending on the classification in this framework, appropriate sampling and modelling decisions can be made to minimise the risk associated with the incorrect treatment of the variable. In particular, the authors argue that for response variables, the approach taken to building spatial models may need to be different to conventional linear averaging approaches (for example kriging).