Selection of Geometric Design Variables for Fine Numerical Optimizations of Electrical Machines

Abstract

This paper proposes an extended set of variables for numerical optimizations of tooth-coil permanent magnet synchronous machines (TC-PMSMs). The design variables for the extended set are chosen based on the magnetic asymmetries introduced by the saturated core under the loading conditions. The extended set allows to reach considerably higher values of torque density and significant reduction of torque ripple in comparison with a conventional set of optimization variables. The conventional and extended sets are compared on an example of an finite element method simulation of a double-layer winding surface permanent magnet-rotor TC-PMSM. A considerably better optimal machine is obtained when the extended set of variables is utilized. The paper showcases the design and the optimization of electrical machines using only free and open-source software.