Design of an optimal fuzzy logic controller using response surface methodology

When the fuzzy logic controller (FLC)-designed based on the plant model-is applied to the real control system; satisfactory control performance may no t be attained due to modeling errors from the plant model. Also, a controll er that is designed under specific circumstances may not show satisfactory control performance when applied to other circumstances. In such cases, the control parameters of the controller must be adjusted to enhance control p erformance. Until now the trial and error method has been used, consuming m uch time and effort. Also, the set of adjusted values is not guaranteed to be optimal. To resolve such problems, response surface methodology (RSM), a new method of adjusting the control parameters of the controller, is sugge sted. This method is more systematic than the previous trial and error meth od, thus, optimal solutions can be provided with less tuning. First, the in itial values of the control parameters were determined through the plant mo del and the optimization algorithm. Then designed experiments were performe d in the region around the initial values, determining the optimal values o f the control parameters that satisfy both the rise time arid overshoot sim ultaneously.