Abstract
With the increased use of digital technologies in the mining industry, the amount of centrally stored production data is continuously growing. However, datasets collected in mines and processing plants are not fully utilized to build links between extracted materials and the performance of processing plants. Moreover, innovative solutions must be found to integrate data-driven insights into decision-making, such as mine production scheduling, to achieve quantifiable improvements.
A new geometallurgical framework is presented which builds empirical prediction models of key performance indicators of the plant, affecting processing capacities, processing costs, and revenues. Applied case studies in an Australian gold mining complex are shown, including prediction models of steel ball consumptions, ball mill throughput, and reagent consumptions in a gold leaching operation. The underlying datasets are gathered from production processes such as measurement while drilling data, records of fleet management systems, sensors, and measured metallurgical responses of the processing plant. The prediction modes can thus be easily constructed and updated from centrally stored databases and give a potential financial advantage over standard geometallurgical laboratory tests such as grindability, hardness, leachability, recovery, etc.
Finally, the integration of these geometallurgical prediction models into a stochastic short-term mine production scheduling optimization tool is demonstrated to monetize the gained insights. Compared with conventional planning, the resulting mine plans increase profit and are more likely to align with short-term and long-term production targets.