Fuzzy Fault Tolerant Controller Applied in an Iron Ore Concentrate Dewatering Plant

摘要

Over the last few years, the mineral processing industry has started to experiment the benefits that advanced control techniques can bring. These techniques were tested mainly in flotation and grinding processes. However, in other operations like crushing, classification, dewatering and others, SISO regulatory control loops are still predominant. The use of regulatory control cannot guarantee optimum global performance to the operation by itself. Furthermore, when a fault occurs in the system, regulatory control may hide it from being observed until the plant eventually reaches a state in which loop failure is inevitable. One possible solution is to use fault tolerant control, a closed-loop control system which can tolerate component malfunctions while maintaining desirable performance and stability properties. In an iron ore concentrate dewatering plant of VALE S.A in Serra Norte, Carajas - Brazil, many tons of concentrate are lost every year due to vacuum disk-filters fails. This paper presents the dynamic simulation and fault tolerant control of a three tank system in this filtering plant. First a dynamic model of the process is created, next, for fault accommodation, a fuzzy layer is employed. The proposed set of fuzzy rules supports the top layer to change regulatory controllers' setpoints, responding to filter faults and other disturbances, avoiding losses of concentrate. The proposed controller smoothly compensated faults on filters, faults on thickener and disturbances in the process. The level variability in tanks was reduced by using the fuzzy controller, helping to increase lifespan of valves and pumps due to a smoother operation; the buffer tanks were operated inside its operational limits and a larger time of operation was obtained.