Integrated Artificial Neural Network and Discrete Event Simulation Framework for Regional Development of Refractory Gold Systems

Abstract

Mining trends in the gold sector indicate a growing imbalance in global supply and demand chains, especially in light of accelerated efforts towards industrial electrification and automation. As such, it is important that research and development continue to focus on processing options for more complex and refractory ores. Unlike conventional (i.e., free-milling) ore feeds, refractory gold is not amenable to standard cyanidation, and requires additional pretreatment prior to leaching and recovery. With recent technological advancements, such as sensor-based ore sorting, there is opportunity to advance the development of smaller untapped refractory resources with marginal economics, particularly those in proximity to processing infrastructure within major gold districts. However, it will be critical that the necessary tools are developed to capture the potential system-wide effects caused by varied ore feeds and improve related decision-making processes earlier in the value chain. Discrete event simulation (DES) is a powerful computational technique that can be used to monitor the interactions between important processes and parameters in response to random natural variations; the approach is thus suitable for the modelling of complex mining systems that deal with significant geological uncertainty. This work implements an integrated artificial neural network (ANN) and DES framework for the regional coordination of conventional and preconcentrated refractory gold ores to be processed at a centralized plant. Sample calculations are presented that are based on a generated dataset reflective of sediment-hosted refractory gold systems.

Keywords

• artificial neural network
• ore classification
• discrete event simulation
• mass balance
• modes of operation
• ore sorting