Licentiate thesis defense

Qixuan Wang, Electric Power Engineering

Modelling and Mitigation of Torque Ripple in Permanent Magnet Synchronous Machines

Overview

Electrical machines as the key components of EVs, provide the torque for the powertrain and determine the dynamic performance.
Nevertheless, this is accompanied by torque ripple and the electromagnetic excitation of noise, vibration, and harshness (NVH).
To minimize the torque ripple without hardware modification, the control-oriented approach is the focus in this study.

This study starts from a design and analysis process, which establish the base of the analysis and mitigation of torque ripple.
In this process, a reference permanent magnet synchronous machine (PMSM) is designed and its characteristics including efficiency map, torque, flux linkage, back-EMF, radial flux and force density, core loss, as well as reduced order model (ROM) are compared.

Moving towards torque ripple mitigation, to develop the control-oriented approach, several designed-oriented approaches including skewing, notching and optimization are investigated first and set as the reference.
To reduce manufacturing cost and increase flexibility, harmonic current injection method is investigated.
To extract and inject harmonic components, the stochastic gradient descent algorithm and phase-locked loop are implemented to identify the amplitude and phase of the torque ripple, respectively.
The results show that the control-oriented method can effectively reduce the torque ripple, and reveal the potential of this method, which is competitive with the design-oriented method but without hardware modification.

To further analyze the torque ripple when the PMSM coupled with the inverter, a circuit-based PMSM model needs to be implemented.
In this study, a time-efficient spatial harmonics modelling method of a PMSM in polar coordinates is proposed to reduce the flux mapping time.
The results show a significant mapping time reduction utilizing the proposed method, and it captures the saturation
and spatial harmonics while maintains high fidelity.
Qixuan Wang
  • Doctoral Student, Electric Power Engineering, Electrical Engineering
Qixuan Wang, Electric Power Engineering | Chalmers