Application of fuzzy control to a non-linear thermal process

An overview of a fuzzy controller is presented. A fuzzy control algorithm has been developed for a single-input single dynamic process. This algorithm chooses the set of applicable rules (four rules) from the rules base (49 rules) to derive a fuzzy control action. Fuzzy control rules are derived by referring to an operator's experience and a control engineer's knowledge. A real-time application to a nonlinear thermal process has been implemented. The results obtained are satisfactory for achieving tracking and regulation objectives. Different defuzzification methods, such as mean of max, center of gravity methods, are applied to obtain crisp control action. Also, a novel fuzzy reasoning method is proposed to directly synthesize the crisp control action. A comparison between fuzzy control and a classical control approach (multimodel control) is made to show the potential of the applied algorithm to control the process.<
>