Abstract

The control of rougher flotation circuits represents a challenging control problem due to the non linearities, multiple inputs-multiple outputs and the wide variety of disturbances acting on the system. Many concentrators rely on regulatory control loops to maintain a stable operation and on the plant operators to find the best operational results. As a mean of using the operator’s knowledge in a consistent manner Expert Systems have been proposed to emulate the current practices in flotation plants. In this work, an algorithmic supervisory system, as opposed to rule-based systems, is proposed and analyzed using simulations of an industrial rougher flotation circuit.