Optimal Design of Multi-Airgap Electrical Machines: An Unknown Size Mixed-Constrained Global Optimization Formulation

Abstract

This paper solves, in a rational way, more sophisticated electrical machine design problems by extending the work done by Fitan et al. It addresses the special case of multi-airgap cylindrical concentric machines. The solution requires a reformulation of the inverse problem associated with the design of electrical machines. Thus, a new mixed-constrained global optimization problem has to be solved. From a mathematical point of view, a major difficulty is that the number of variables and constraints varies during the resolution, depending on the number of considered airgaps. Moreover, this involves extensions of analytical models used for permanent-magnet machines. The paper uses numerical tests for concentric rotor machines with one, two, and three (in one case) mechanical airgaps to validate this methodology of design