HYBRID NON-LINEAR MODEL PREDICTIVE CONTROL OF A RUN-OF-MINE ORE

A run-of-mine (ROM) ore milling circuit plays a crucial role in processing ores containing precious metals to achieve a particle size suitable for the liberation of valuable minerals. The effective management of the ROM milling circuit, which includes components such as a mill, sump, and hydrocyclone, poses significant challenges due to unmeasured process outputs, inherent non-linearities, time delays, and disturbances. This study aims to enhance the control of ROM ore grinding circuits by integrating hybrid non-linear model predictive control (MPC) strategies. The methodology involves the formulation of a control model that accommodates multiple manipulated variables to address various process disturbances while adhering to quality specifications and maximizing throughput. Our findings reveal that hybrid MPC significantly improves performance metrics compared to traditional control methods, especially in complex scenarios characterized by variability and uncertainty. This research contributes to the ongoing evolution of control strategies in the mining sector, addressing the gaps found in existing literature, and proposes that hybrid approaches, which adapt to the dynamic characteristics of milling circuits, can lead to more efficient and stable operations.