Research Of Fuzzy Self-tuning 2DOF-PID Controller For PMSM

The PMSM servo-system is widely applied to high performace driven field,such as flexible manufacture,CNC machinary tools,elevator adjust-speed etc. the PMSM has many individual features because of its structure and operation. A this paper the mathematical model and controller of PMSM are studied.The traditional PID control is abroad used to the motion control field for its simple mechanism and its flexible structure.But between the system tracking and the anti-disturbance performance,it can not find the better balance, an its parameters adjustment depend on the accurate model of the controlled process: the more clear machanism analysis of the controlled process is recognized, the more accurate mathematical model is developed, the more appropriate parameters are tuned, and the better control results are reached.However, the PMSM variables have been strongly coupled,at this time the system is non-linear, time varying process, it is difficult to describe the system with mathematical model. For these features of PMSM, in this thesis, the two-degree-of-freedom(2-DOF)PID control is introduced, one of the intelligent control-Fuzzy technique is added. Several control schemes are introduced and presented as follows:(1) To improve the traditional PID control which can not keep the system tracking and the anti-disturbance performance in good condition simultaneously, the necessity to use the new type of 2-DOF PID control in drive systems is illustrated. furthermore, the advanced-two-degree-of-freedom PID and the super-two-degree-of-freedom PID are provided. The simulation results demonstrate that these two methods solved the mentioned problem of the conventional PID control.(2) Integrate the Super-two-degree-of-freedom PID control with fuzzy techniques, a self-tuning PID controller bassed on fuzzy inference is designed. On-line automatically tuned parameters are realized at different states. Many simulation experiments have done are proven the controller can overcome model mismatch, and it is strongly adaptability and robustness.