Optimal Modulation for Automotive Electrical Drives

Abstract

This thesis focuses on optimal modulation techniques for power converters in automotive applications and their role in improving energy conversion efficiency and electrical drive performance. As the automotive industry strives to enhance the efficiency of electric vehicles, high-speed electrical machines with non-linear magnetic anisotropy are increasingly being utilized. The anisotropydependent harmonic distribution of the stator currents, however, poses challenges, especially when operation at low switching-to-fundamental frequency ratios is required, rendering conventional modulation techniques inadequate. Optimized pulse patterns (OPPs), emerge as a promising solution for automotive drives, due to their exceptional performance even at low pulse numbers and their versatility that allows for customization based on specific load characteristics.

In this work, the OPP optimization problem is developed to enable the computation of OPPs that can achieve lower current total demand distortion (TDD) and improved performance in practical applications. To achieve this, the electrical synchronous machine is first modeled in detail by accounting for magnetic anisotropy, high-frequency winding resistance, and the impact of winding distribution on back-electromotive force (EMF) harmonics. Second, to enhance torque characteristics, OPPs are designed by constraining specific harmonics of the stator current. Finally, for a meaningful impact on the efficiency of the drive train, the optimization process takes a holistic approach, minimizing the stator current TDD while simultaneously constraining converter and winding losses.

Overall, this thesis presents advancements in optimal modulation aiming to enhance the performance of automotive drives, with OPPs specifically tailored for anisotropic synchronous machines. The key contributions include the applicability to synchronous machines with a non-sinusoidal back-EMF, reduced electromagnetic torque ripple, and improved overall drive efficiency. The proposed methods generate pulse patterns that achieve optimal trade-offs, increasing efficiency by up to 3.4% and reducing stator current harmonic content by up to 50% compared to conventional modulation techniques. As a result, they significantly enhance the overall performance of the automotive drive system. Numerical and experimental results on industrial drives validate the effectiveness of the proposed methods.