A finite element method framework for modeling rotating machines with superconducting windings

TL;DR

This paper introduces a finite element method framework for modeling rotating electrical machines with superconducting windings, addressing the challenges of implementing power-law resistivity laws in FEM simulations.

Contribution

It proposes a novel two-part modeling approach using H- and A-formulations with boundary continuity conditions and Lagrange multipliers for superconducting and conventional regions.

Findings

Effective boundary condition implementation between regions

Stable solutions despite power-law resistivity complexities

Applicable to simple and complex machine configurations

Abstract

Electrical machines employing superconductors are attractive solutions in a variety of application domains. Numerical models are powerful and necessary tools to optimize their design and predict their performance. The electromagnetic modeling of superconductors by finite-element method (FEM) is usually based on a power-law resistivity for their electrical behavior. The implementation of such constitutive law in conventional models of electrical machines is quite problematic: the magnetic vector potential directly gives the electric field and requires using a power-law depending on it. This power-law is a non-bounded function that can generate enormous uneven values in low electric field regions that can destroy the reliability of solutions. The method proposed here consists in separating the model of an electrical machine in two parts, where the magnetic field is calculated with the most appropriate formulation: the H-formulation in the part containing the superconductors and the A-formulation in the part containing conventional conductors (and possibly permanent magnets). The main goal of this work is to determine and to correctly apply the continuity conditions on the boundary separating the two regions. Depending on the location of such boundary -- in the fixed or rotating part of the machine -- the conditions that one needs to apply are different. In addition, the application of those conditions requires the use of Lagrange multipliers satisfying the field transforms of the electromagnetic quantities in the two reference systems, the fixed and the rotating one. In this article, several exemplary cases for the possible configurations are presented. In order to emphasize and capture the essential point of this modeling strategy, the discussed examples are rather simple. Nevertheless, they constitute a solid starting point for modeling more complex and realistic devices.