Fuzzy adaptation of the PI controller parameters applied for induction motor

Several works have been studied the use of the fuzzy logic confirming that the classical controllers have advantages compared to the fuzzy controllers: very short time of execution and mathematical precision. But, the classical controllers have a certain limitation, if rather significant variations of disturbance factors act on the system to be controlled. The classical controller does not always react in an optimal way. Thus the abilities of adaptation and robustness of this controller are limited. In the present paper, an adaptation mechanism (supervisor) for the PI gains is proposed allowing the improvement of the classical controller, introducing of a certain intelligence degree in the control strategy of the controller. Indeed, this approach associating the PI controller and a supervisor composed of fuzzy rules, offers the possibility to use the mathematical accuracy of the PI algorithm with the adaptability, the flexibility and the simplicity of the fuzzy linguistic formalism.