Research On Generalized Predict Control Of PMSM With Constraints

High quality requirement of servo system is raised with the development of industry and extensive application of servo system. The mature control strategies of Permanent - Magnetic AC Servo - Mechanism, such as Vector Control and Torque Direct Control, are effective and successfully applied in process industry. Alone with the development of computer technology and the further study on control theory, researchers pay more attention to Intelligent Control Strategy. Predict control algorithm which is suitable for computer realization, shows a great vitality and has been applied in process industry because of its robustness and low request to precise model.Permanent Magnet Synchronous Motor(PMSM) is a nonlinear system with significant coupling. Traditional PID control may probably cause decline of dynamic performance when the load or parameters change, which is unavoidable in practical, so a robust controller is needed. Meanwhile, the performance index, which can be achieved best by adjusting properties to certain limitation, must be taken into account,thus it is necessary to make further research on constrained Generalized Predict Control(GPC). This paper takes the PMSM as a research object. First, its methematics models in several coordinate are set up based on the analysis of electromagnetism restricted relation. Then, vector control method of PMSM is presented, in the same time, PWM technology based on current comparing to selecting space voltage vector and predictive control technology used in speed-loop is presented in detail. As for Single Input Single Output (SISO) System Controlled auto-regressive intergrated moving average (CARIMA) model is setup in the rotor's reference frame. Combining feedback mechanism of PID and predict mechanism of GPC, an improved model GPC algorithm is proposed, and as a result, the performance is improved by inhibiting the overshoot. Constrained optimization is added to performance index on basis of matrix decomposition. Finally, the experiment result shows the effectiveness of the proposed control algorithm.