Current reviewed methods that are used to build relationships between measured orebody characteristics and process responses (e.g., blast fragmentation, comminution rates, and recovery) rely primarily on laboratory tests carried out on samples deemed to be representative of the characteristics of domains within an orebody. Using these data to predict the long run performance of a mine requires many samples and many tests and may produce predictions that are vulnerable to biases that can be inadvertently introduced during sample selection, sample testing, estimation and process modelling.
Increasingly combinations of high resolution (core scale) and deposit scale data (sensed and measured) can be used to improve the methods used to characterise the ore body and use multivariate estimates to predict process performance. Improving the sequence of measurement and testing to characterise the ore body, will reduce the time and cost to arrive at more robust predictions of process performance, and facilitate the evaluation of operational strategies that have improved probabilities of profitable outcomes.
Additional data sources such as the historical performance of existing mines can be associated with the characteristics of the ore that those mines are processing can also be used to improve the models and predictions of future performance.
High resolution scan data can be variously used to associate a few tests with measured laboratory scale responses. Once these relationships have been validated it may well be possible to generate scan data from cores and estimate the sensed data into volumes of larger scale and use these estimates to directly predict process performance and mining project outcomes.