Estimation of Permanent Magnet Synchronous Machine Performance for Pre-Design Using a Reluctance Network

Abstract

Flexible and fast development processes are required to adapt to changing user and market requirements. This is also true for the design of electric machines. Typically, many different parameter setups like rotor and stator diameter, teeth width, and number of windings, have to evaluated until a suitable motor design is found. As performance estimation is often conducted using finite element analysis (FEA), which is computationally and time intensive, the process takes a long time. Reluctance networks (RNs), on the other hand, simplify the model description but are less accurate than FEA. The aim of this paper is to show that an RN might show larger deviations to the actual motor performance, but can be used in a pre-design step to limit the parameter domain for FEA simulations if the influence of parameter changes is predicted well. Analytical dependencies between motor geometry and model parameters are derived based on a simplified RN and validated using FEA.