Research On Multi-source Disturbance Suppression Method For Permanent Magnet Synchronous Motor Drive System

Thanks to the characteristics of high efficiency,high power density,large torque inertia ratio,low noise,and maintenance-free,permanent magnet synchronous motors(PMSM)drive system has been applied in many industrial applications.Affected by external disturbances and internal uncertainties,PMSM high-precision control and wide-range smooth speed regulation face severe challenges.It is necessary to seek breakthroughs in control strategies or control methods.In this thesis,the PMSM drive system is taken as subject investigated,the existing disturbances in the system are summarized.Further research is carried out on three disturbances: electrical parameters uncertainties,current-loop coupling,and cogging torque.Aiming at the internal model control(IMC)-based PI current controller’s parameters sensitivity,a current controller considering parameter uncertanties is proposed in this thesis.Uncertainty and disturbance estimator(UDE)is used to estimate and compensate for the influence of parameter uncertanties on the system.The compensated system can be regarded as a system without parameters variation,and then an IMC-based PI current controller is designed for this system.The method has high parameters adjustability and low sensitivity to the motor electrical parameters.Theoretical analysis,simulation,and experimental all prove the proposed current controller’s robustness to parameters variation.Aiming at the parameter dependence of the current feedback decoupling and deviation decoupling control strategies,a UDE-based decoupling control strategy is proposed in this thesis.The theoretical derivation proves that the proposed method has a decoupling network similar to the current feedback decoupling,while the proposed method is not affected by parameters variation.Both simulation verification and experimental verification prove the proposed method’s decoupling performance and its robustness to parameters variation.The limitations of IMC-based observer,extended state observer(ESO),and modelcompensation-based ESO when observing cogging torque are analyzed.A IMC-based seriestype ESO is proposed in this thesis.The proposed observer solves the shortcomings of the IMC observer and reduce the requirement for ESO’s bandwidth.Simulations and experiments prove that the proposed observer can observe the cogging torque and load torque under the condition of limited bandwidth and obtain good speed response performance.