Multi-material topology optimization of a flux switching machine

This paper investigates the topology optimization of the rotor of a 3-phase flux-switching machine with 12 permanent magnets located within the stator. The objective is to find the steel distribution within the rotor that maximizes the average torque for a given stator, permanent magnets, and electrical currents. The optimization algorithm relies on a density method based on gradient descent. The adjoint variable method is used to compute the sensitivities efficiently. Since the rotor topology depends on the current feedings, this approach is tested on several electrical periods and returns alternative topologies. Then, the method is extended to the multi-material case and applied to optimize the non-magnet part of the stator. When dealing with 3 phases, the algorithm returns the reference topology as well as a theoretical machine with no return conductor according to the set current angle. To illustrate the creativity of the method, the optimization is finally performed with a single-phase and returns a new topology.