MMF Harmonics Effect on the Embedded FE Analytical Computation of PM Motors

Abstract

This paper aims to define a minimum set of finite-element (FE) solutions to be used in the design and analysis of saturated permanent-magnet motors. The choice of the FE solutions belonging to this set is strictly associated with the classical d-q axis theory, and it is described in terms of key points on the flux-magnetomotive-force diagram. When synchronous machines are considered, such a diagram has a regular shape, so that a huge reduction in FE field solutions is possible with no loss of accuracy. It is also shown that the torque computed by using the d-q axis theory is almost independent of the variation of the flux linkage with the rotor position. At last, this paper describes a technique in which few FE solutions allow the identification not only of the average torque but also of the main torque harmonics. As a result, the torque behavior versus rotor position can be rapidly predicted

Performance comparison of the permanent magnet assisted synchronous reluctance motor and the double magnet synchronous motor

“…For these positions, the sum of the principal flux linkage harmonics is zero [4]. Therefore, the average torque is calculated by (4).…”

A review of efficient FE modeling techniques with applications to PM AC machines

“…Additionally, the MEC-FEA model requires special pre-processing, which unlike automatic FE mesh generators, maybe dependent on the motor topology and design details. [16,17,18,19,20] The flux-mmf diagram was proposed by Staton et al as a method for computing the average torque of electric motors [16]. The typical algorithm consists of magnetostatic FEA for calculating the flux linkage in the abc reference frame and plotting the results against the current on a per phase basis.…”

Analysis of asynchronous machines for direct drive wind power generation

“…Although both computations present the same average value, the first method considers the complete torque developed by the AG, while the second method inherently does not consider all torque oscillations [12]. Therefore, for an accurate analysis of the AG, the computation by means of Maxwell's stress tensor has to be carried out.…”