Assessing mineralogical indicators for ore hardness

Liberation is a key requirement for efficient separation in mineral processing. As a consequence of this an understanding of the inherent textural features of an ore is no longer possible since during the process of liberation, texture is destroyed. Knowledge of key textural features, such as mineral grain size and distribution, prior to breakage processes provides the opportunity to understand the textural drivers for ore hardness. Modern characterisation methods, including automated SEM-based techniques, provide the possibility of measuring mineral characteristics at a range of scales. This paper discusses how textural features of two different ore types in a porphyry copper deposit can be measured and interpreted at a number of scales. Standard breakage tests conducted on the ores indicate that one of the ores can be characterised as hard and the other as soft. The impact of mineral grain size and the presence and scale of vein type structures in the two ores have been used to determine the extent to which these features influence ore hardness.INTRODUCTIONIn order to effectively convert ore into metal, knowledge of both the processing stages and the characteristics of the ore are essential. Prior to the advent of technologies that enabled the fundamental understanding of the sub-process at each stage of processing and the measurement of appropriate ore properties the most effective way to understand how an ore would respond to a process was to put it through that process, for example the way to determine how an ore would break was to first break it. Existing comminution models seek to predict the size distribution of rock fragments as the result of applied energy based on knowledge of the rock strength. It is common for these models to consider the rock as homogeneous and despite providing important data for the design and optimisation of industrial comminution circuits these models are only applicable within a narrow range of operation. In order for reliable predictive models to be developed that account for the inherent ore characteristics, a step change is required in the type of measurements that are carried out to characterise an ore and the type of data that is extracted and used as model inputs