Experimental tuning of fuzzy PID controllers: Electrical applications

Electrical and electromechanical systems have to satisfy to more and more constrained specifications. Therefore, non-linear control structures must be spread out. Among them, fuzzy logic control can be one interessant and performant alternative. The main handicap of this kind of stucture resides in the fact that the tuning parameters are very numerous. In this paper, we first propose an on-site tuning strategy of this set of parameters in the case of a fuzzy proportionnal-integrative controller based on the experimental designs methodology and on a limited number of pre-defined closed-loop experiments. Then a complete set of predetermined parameters for a fuzzy proportionnal-integrative-derivative controller will be given. These parameters have been optimized on a specified benchmark according to an IAE criterion. They are calculated like the Ziegler-Nichols or Broida methodology on conventional controllers; that is, using a single open-loop step response to obtain a model of a first-order plus delay transfert function. Validity limits for this method are provided.