Parameter Identification And Torque Ripple Reduction Of Permanent Magnet Synchronous Motor System

Permanent magnet synchronous motor(PMSM)has been widely applied in high-end equipment manufacturing industries like aviation,marine propulsion,rail traction,high precision instrument and so on,due to its advantages of high power density,high efficiency,and fast dynamic response.However,in the motor system,model parameter errors and periodic disturbances including rotor flux harmonics,cogging torque,inverter dead-time effect and current measurement error,deteriorate the dynamic and steady state performance of the system,and consequently limit the application of PMSM in high performance fields.Aiming at improving the performance of PMSM system,this paper carries out a series of researches focusing on parameter identification and torque ripple reduction against the impact of model parameter errors and periodic disturbances.The inverter nonlinearities including dead-time effect exerts significant influences on the accuracy of online motor parameter identification.In this paper,an accurate model of the disturbance voltage caused by dead-time effect is established,and an online parameter identification strategy of PMSM considering the disturbance voltage is proposed.Ignoring the current harmonics causes estimation error of the disturbance voltage,and based on this,two methods are proposed to reduce the estimation error of the disturbance voltage.One employs an iterative learning controller in the current control loop to suppress the current harmonics,while the second estimates the disturbance voltage using the average value of the variables in the voltage equation,so that the estimation result is immune to current harmonics.The proposed method,independent of motor and VSI parameters,effectively improves the estimation accuracy of the disturbance voltage,and motor parameter identification accuracy is consequently improved.The estimated motor parameters are utilized to adjust the controller parameters of predictive current control,so that the steady state error is eliminated,and steady state performance of the system is improved.Focusing on the problem that the inverter nonlinearities cause estimation errors of the motor parameters at standstill,an accurate model of the inverter disturbance voltage at standstill is established,and a parameter identification strategy of PMSM at standstill considering the disturbance voltage is proposed.Appropriate voltage excitations are selected to implement the proposed identification method,which requires neither mechanical equipment for rotor locking,nor model parameters of motor or inverter device.Meanwhile,the influence of the disturbance voltage is eliminated,and the motor parameter identification accuracy is improved.The motor parameters estimated at standstill are utilized for controller parameter tuning of the traditional vector control,and the start-up performance of the system is improved.Due to the torque ripple induced by the periodic disturbances,a parameter optimization based iterative learning control(ILC)strategy is put forward for torque ripple reduction of PMSM.Stability analysis is carried out according to the Nyquist stability criterion,and based on this,the feasible domain of ILC parameters are obtained.Moreover,an adaptation mechanism of ILC parameter is established,and the adaptation law is derived to adjust the controller parameter online.Research results show that the parameter optimization based iterative learning control proposed in this paper achieves lower torque ripple during steady-state operation and shorter regulating time of dynamic response,thus satisfying the demands for both steady state and dynamic performance of the PMSM system.