A Level-Set-Based Topology Optimization for Maximizing the Torque of Synchronous Reluctance Machines

The article presents a level-set (LS)-based topology optimization method for improving the magnetic torque of electrical motors such as synchronous reluctance machines (SynRMs) using a 2-D magnetostatic finite-element context modeling. The novelty is the use of a fully analytical expression for the torque sensitivity analysis (SA) based on the virtual work principle which is more accurate and general than other methods and its application to an LS implementation. The numerical study of a synchronous reluctance machine is presented. Then, the LS method is detailed in a general context. The corresponding analytical continuous SA using the adjoint method to maximize the torque is expressed and validated. Finally, the application to the optimization of the rotor of a synchronous reluctance machine is shown to illustrate the efficiency of the method.

I. Introduction

Synchronous reluctance machines (SynRMs) have many benefits. They offer interesting performances and do not need permanent magnets that are rare and expensive. However, their torque is mainly determined by the topology of their rotor. This topology can be hard to guess and requires an expensive design time for engineers [1], [2]. Topology optimization is a method that allows the finding of an optimal material distribution in a target region. Thus, it seems to be a really interesting approach to discover new designs in a finite-element modeling context since it completely avoids the need of any prior intuition on the optimal design of the machine [1], [2].