A Study On Identification And Control Methods Of Permanent Magnet Synchronous Motor Multirate Sampled System

Permanent magnet synchronous motors(PMSM) due to their high efficiency, high power/torque density, high starting torque, etc., are widely used in high-performance precision servo control and other fields. Development of power electronics and microprocessor technology provides a solid foundation for the application of advanced control methods to PMSM. Field oriented control and direct torque control are two main high performance control strategies for PMSM. However. PMSM is a nonlinear multi-variable system with time-varying parameters, and the actual system performance is often affected by the changes of the electrical/mechanical parameters and the external load disturbance. Real-time identification of important parameters and controller design are two focuses of current research for PMSM system. Classical system identification methods and control algorithms generally are designed with the assumption that the system input and output are sampled (updated) at the same frequency. However, for complex industrial plant, using the same frequency for sampling and controlling is impractical,or even unattainable. Aiming to solve the problems of the parameters identification and control with PMSM, multirate system theory,sliding mode control theory,model predictive control theroy and state estimation methods are merged together and the following aspects are discussed and studied:Firstly, based on polynomial transformation technology, dual-rate recursive least square algorithm is proposed for PMSM system without noise. And then,dual-rate recursive extended least square algorithm is proposed for polynomial transformated PMSM system with colored noises. To the single-rate output error model of dual-rate sampled system, dual-rate auxiliary model recursive least squares algorithm is proposed for PMSM system. Simulation results show the effectiveness of those algorithms.Secondly, to stochastic gradient algorithm’s slow convergence, based on the multi-innovation identification theory, the multi-innovation stochastic gradient algorithm is proposed for PMSM. Because of introducing an innovation length and expanding a single innovation vector into an multi-innovation matrix, multi-innovation stochastic gradient algorithm has the improvement on the convergence and accuracy. With the length of the innovation of PMSM increased, the convergence performance of the proposed algorithm identification process is further enhanced. Least square algorithm needs to calculate the covariance matrix,and then the computation of least square algorithm is too large, especially in multirate systems, the number of estimated parameters increases, the computation will increase dramatically. Based on the polynomial transformation technique, stochastic gradient algorithm with convergence factor is proposed for PMSM.With the convergence factor, the convergence speed and accuracy are significantly improved in the identification process.Thirdly, extended Kalman filter is the modern filtering method developed from an optimal prediction with the minimum variance,and has been widely used in motor speed estimation. Based on the electrical parameters and the speed of PMSM, third-order model of the motor is established.With the "lift" technology, PMSM multirate-input state-space description is built. Multirate-input extended Kalman filter algorithm is used to estimate the motor speed under the data loss situation. Simulation and experimental results show that the convergence accuracy and stability of speed estimation are enhanced with the multirate-input extended Kalman filter algorithm.Fourthly, based on multirate digital control theory and discrete-time quasi sliding mode control theory, multirate input quasi-sliding mode control algorithm is proposed for the systems with limited bandwidth interference, and the algorithm can effectively suppress the influence of interference on the system, and the system steady chattering can be effectively weakened. Considerating the typical multirate feature to PMSM’s mechanical and electrical subsystems, multirate-input quasi-sliding mode controller is proposed with terminal attractor, and works as a motor speed regulator. With the use of extended input vector in one speed sampling unit, the bandwidth of PMSM system steady chattering is effectively reduced, and the system dynamics is increased. Simulation and experimental results show the effectiveness of the algorithm.Finally, based on multirate digital control theory and model predictive control theory, in consideration of the dynamic response speed mismatch between the current loop and the speed loop of permanent magnet synchronous motor field oriented control system, multirate-input model predictive controller of PMSM in the single-loop is proposed. Simulation and experimental results show the effectiveness of the algorithm,and furthermore the algorithm can ensure the whole control system performance under the data loss situation.