An MPC approach to deep level mine rock winder hoist control

This paper presents a model predictive control approach to the control and scheduling of rock winders found in South African deep level mines in the context of demand-side management and time-of-use tariffs. Half hourly predictive control is applied on a discrete dynamic and constrained linear programming model of the winder system taking into account simulated system delays in determining an optimal hoist schedule whereby a minimum hoist target is achieved at a minimum energy cost. The hoist schedule, which changes continuously based on the feedback and controller output, states the number of hoists to complete during each half hour over a D-day period, ensuring minimal hoisting during expensive high demand periods.