Hybrid model predictive control for mineral grinding

The mining industry is in search of control strategies that allow considering a global optimization of their processes, ensuring their stability. This has led to the application of recently developed control techniques on large-scale systems in the mining processes. One of the most important process of the mining industry is the mineral grinding, as the product particle size impacts, in a significant manner, the recovery rate of the valuable mineral in the separation stages. As a solution, a centralized hybrid model predictive control (HMPC) is presented; this control approach maintains the product particle size on a defined range and minimizes energy consumption, maintaining the grinding plant in a stable operation. As a comparison, a conventional model predictive control with an expert system to handle discrete events was developed, showing HMPC is a suitable solution that can used in the grinding process, with the benefit that it handles in the same controller both discrete and continuous dynamics and events.