Research On Model Predictive Current Control Of Open-end Winding Permanent Magnet Synchronous Motor With Common Bus

Dual-inverter fed open-end winding permanent magnet synchronous motor can greatly expand the speed range when the bus voltage is limited,because of a higher DC bus voltage utilization rate,so it is suitable for the occasions where a wide speed range is required,like traffic motors.Dual-inverter fed open-end winding permanent magnet synchronous motor with common DC bus link faces the challenge of zero-sequence current suppression,however,traditional PI/PR regulators based current control method is difficult to design because of too many parameters,and it suffers from high switching frequency.On the basis of analyzing and improving the traditional control method,this dissertation studies the model predictive control based current control method to relieve the current-loop design efforts,improve the current control performance,and lower the switching frequency.First,the traditional current control method is improved based on disturbance observation and compensation.Due to the limited modulation range,the zero-sequence current cannot be effectively suppressed at high modulation ratios,and it will interact with the third harmonic back-EMF to generate the torque ripple,which affects torque control performance.On the basis of analyzing the zero-sequence voltage modulation range of ZVR-PWM method,a torque ripple suppression method based on disturbance observation and compensation is proposed.A general-integral extended state observer is designed to observe the torque ripple caused by the zero-sequence current,and compensate it through the active disturbance rejection control law,so that the torque ripple caused by the zero-sequence current can be offset by the active torque generated by the quadrature-axis current.This method can effectively suppress the torque ripple caused by the residual zero-sequence current and improve the torque control performance.Second,a current control method with low switching frequency is designed based on the finite control set model predictive control.The finite control set method can realize the integrated design of control and modulation algorithms,and has the characteristics of low switching frequency,but the problems of zero-sequence current suppression and increased computational burden caused by more voltage vectors need to be considered.Using the three-phase voltages decoupling feature of the open-end winding permanent magnet synchronous motor,the distribution laws of the voltage vectors in ABC threephase natural coordinate system are studied,three-phase reference voltages containing the zero-sequence voltage information are predicted,and certain geometrical principle is used to simplify the voltage vector selection process.At the same time,switch states redundancy feature of dual-inverter system is used to optimize the switch states,and further lower switching frequency or reduce the change rate of common-mode voltage.Third,a current control method with zero-sequence model robustness is designed based on the dual-vector model predictive control.Zero-sequence voltages multi-polarity of the plane voltage vectors is studied,and a dual-vector model predictive current control method is proposed.First of all,plane geometric relationship is used to select the two vectors and calculate the dwell time;then,according to the polarity of the zero-sequence current,proper switch states are selected to form the hysteresis suppression for the zerosequence current.When the switch states adjustment cannot meet the hysteresis control requirement,the zero-vector insertion measure is supplemented to ensure the suppression of zero-sequence current.Finally,an optimal current control method is designed based on the continuous control set model predictive control.First of all,a phase voltage direct modulation method is proposed to modulate the reference voltage.It can modulate the zero-sequence voltage and make each bridge arm alternately in clamping and modulation states to reduce the switching frequency.A multi-step predictive current control method including zerosequence current is designed,and a quadratic cost function is defined.Benefiting from the regular voltage boundary of above modulation method,a spatial geometric solution is proposed to deal with the voltage constraint,so that the global voltage optimal solution with the voltage constraint is solved.At the same time,in order to enhance the parameters robustness of the algorithm,an adaptive Kalman filter based on the augmented state model is designed to observe the disturbance introduced by the parameters mismatch.This method can improve the current steady-state performance,as well as keep the low switching frequency,and has strong robustness against parameters mismatch.