Generalized Reluctance Network Framework for Fast Electromagnetic Analysis of Radial-Flux Machines

Abstract

This article proposes a generalized reluctance network framework for analysis and non-parametric design optimization of rotating, radial-flux, electric machines, combining published state-of-the-art contributions and some original improvements. A novel element geometry and a novel airgap permeance function are proposed to support the automatic and systematic rotating radial-flux machine meshing without requiring the knowledge of geometrical parameters. An extensive and detailed validation of the proposed method was carried out on synchronous and asynchronous three-phase motor technologies, for a wide range of operating points, which constitutes an important contribution of this work. The numerical results obtained with the proposed method and with finite element analysis are compared for relevant design optimization quantities, such as input power, winding losses, core losses, power factor, permanent magnet Joule losses, and harmonic analysis. A satisfactory compromise between accuracy and computational speed was achieved, making the proposed method relevant not only for fast design/pre-design optimization purposes, enabling the evaluation of a large number of solutions in a relatively short time, but also for real-time, low-computational-intensity, reasonable-precision simulations.