Comparison of magnetic circuit design methods of transverse flux reluctance machines

Abstract

The magnetic circuit of transverse flux reluctance machines causes magnetic flux in three spatial directions. 3-D FEM can be used designing the magnetic circuit, but it requires a relatively long computation time. Therefore 3-D FEM is not useful for an optimization tool. This paper compares four much faster design methods: A simple analytical approach, a magnetic equivalent network, a modified 2-D FEM model, and a simplified and very coarse 3-D FEM. Thereby 36 different motor designs (with different pole pitch, diameter and tooth geometry) and 2 different current-values (one causes strong saturation) are used to evaluate the accuracy and limits of those methods. The deviations of magnetic flux and torque from 3-D FEM results (ANSYS® Maxwell®) are discussed based on statistical probability distributions. The results will show that the equivalent network and the 2-D FEM are not useful for all of the investigated motor designs.