FUZZY SELF-TUNING PI CONTROL OF PM SYNCHRONOUS MOTOR-DRIVES

In this paper, a new gain self-tuning method for PI controllers based on a fuzzy inference mechanism has been proposed and implemented. The essential idea is: (1) define a dynamically changed reference trajecto ry in the error and error derivative plane, called a 'sliding trajecto ry'; (2) compute the error area between the sliding trajectory and the effective one with a special algorithm, and use this quantity as a pe rformance index to evaluate the system response; (3) draw a fuzzy rela tionship between a PI gain correction parameter and the performance in dex; (4) based on the fuzzy inference mechanism, PI gains are calculat ed and tuned. The proposed method has been verified through simulation s, using the speed control of a PM synchronous motor drive system as a testbed. The results obtained demonstrate the effectiveness of this n ovel method.