Optimization of multiphase single-layer winding end-connections by differential evolution

In this paper, the authors propose a flexible automatic procedure for optimizing multiphase single-layer winding end-connections by the differential evolution algorithm. By using the winding distribution in slots as a predefined input, connections between phase conductors located in different slots are determined to minimize the winding average coil pitch and to simultaneously maximize the number of coils with identical coil pitch. For an adopted geometrical shape and number of turns of individual coils, minimization of winding average coil pitch results in minimized end-connection length, winding resistances, end-leakage inductances, winding mass and copper losses, while machine efficiency and power factor are maximized. However, maximizing the number of identical coils simplifies and speeds up the process of coil fabrication during machine assembly, but potentially leads to higher average coil pitch. The proposed method is applied to single-layer windings with different slot, pole and phase combinations, where it is proven that the method automatically provides end-connection arrangements which are better or equal in terms of copper usage and simplicity of fabrication than the ones found in the literature. Verification is conducted by comparing overall performance of several Rotor Permanent Magnet Flux Switching motors with different stack lengths, outer diameters and optimized end-windings by means of 2D finite element analysis and 3D analytical approach. The proposed method is conceived as an automatic tool for optimization of single-layer winding end-connections during comprehensive electrical machine design optimization.