Industrial applications of fuzzy identification and control
BP Graham, RB Newell, GP Lepage, LC Stonehouse · 1986
In the last ten years fuzzy control has shown much promise as a method for tackling the automatic control of complex industrial processes that cannot be satisfactorily controlled by more traditional techniques. One of the major problems hampering the further development of fuzzy control is the difficulty in developing an adequate set of control rules from a source such as a human operator. In this paper we demonstrate the use of a fuzzy algorithm that makes decisions based on predictions obtained from a fuzzy process model, rather than from an explicit control rule set. The use of a fuzzy process model allows the use of fuzzy identification techniques in model development. This lessens the reliance on human process knowledge. The application of this fuzzy algorithm to the control of silicain-product at ALCOA's alumina refinery at Kwinana is presented.