SRM torque computation from 3D finite element field solutions

Force and torque computation from finite element field solutions is a key step in the design of some electromechanical devices, and a successful simulation of such devices demands an accurate and reliable method of force or torque calculation. There are four established ways of predicting EM forces...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Benhama, A, Williamson, A.C, Reece, A.B.J
Format: Tagungsbericht
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Force and torque computation from finite element field solutions is a key step in the design of some electromechanical devices, and a successful simulation of such devices demands an accurate and reliable method of force or torque calculation. There are four established ways of predicting EM forces or torques: (1) the Lorentz or the JxB method; (2) the Maxwell stress tensor (MST) method; (3) the classical virtual work method; and (4) the Coulomb virtual work (CVW) method. The first three methods have limitations when used in the finite element context. The CVW method, however, provides a more general algorithm which is easier to implement especially in 3D finite element electromagnetic field analysis. The theory of the CVW method is now well known but no results have been published which illustrate its application to practical problems which require 3D finite element analysis. This paper corrects this deficiency and shows that the CVW method can be superior in accuracy and implementation to the MST method. To highlight this superiority, a switched reluctance motor is considered and investigated using both 2D and 3D finite element analyses. The motor, which has six poles on the stator and four poles on the rotor, has been extensively and carefully tested. The torque-angle characteristics are obtained by the application of the CVW method in both 2D and 3D for various excitations at different positions of the rotor. Comparison between the 2D, 3D and the experimental data shows that the torque computed from the 3D analysis is in much better agreement with the experimental data than that obtained from 2D.
ISSN:0537-9989
DOI:10.1049/cp:19971038