Towards topology optimization of a hybrid-excited machine using recursive material interpolation

TL;DR

This paper introduces a recursive material interpolation method for topology optimization of hybrid-excited electrical machines, improving design performance by effectively integrating diverse materials without compromising physical realism.

Contribution

It extends density-based topology optimization to hybrid-excited machines using recursive material interpolation, enabling better material integration and performance.

Findings

Optimized hybrid-excited rotors outperform existing designs.

Recursive interpolation improves physical realism in material modeling.

Method demonstrates clear superiority over traditional techniques.

Abstract

Hybrid-excited electrical machines aim to combine the advantages of permanent magnet machines (high efficiency and torque density) with those of separately excited machines (ease of flux-weakening at high speed). These machines are of interest to electric vehicles, and only parametric approaches are available in the literature for their optimization. This work proposes a more general topology optimization methodology by extending the formalism of density methods. The difficulty lies in integrating the numerous natures of materials (conductors, permanent magnets, ferromagnetic material, air...) without strongly deconvexifying the optimization problem, which leads to non-physical results with unsatisfactory performance. To address this issue, a recursive material interpolation is introduced. The hybrid-excited rotors optimized by this approach are compared with those of existing techniques, demonstrating a clear superiority of the recursive interpolation.