P. Thomas

Comparison between two approaches to model induction machines with skewed slots

To study electric machines with skewed slots, it is necessary to use 3D models. However, to limit computation time, some approaches based on 2D or 2D-3D models are proposed. In this paper, we compare both approaches. The first one is based on a 2D sliced model and the second one, on a 3D using the 2D-invariant hypothesis. In both models, we take into account the coupling with the electric circuit, the magnetic nonlinearity and the mechanical motion. The results obtained are compared to experimental ones.

A time-stepped 2D-3D finite element method for induction motors with skewed slots modeling

To model an induction motor with skewed rotor bars, we propose a hybrid 3D approach using the 2D invariant hypothesis. Our method is coupled with the electric circuit equations for both the stator and the rotor. We take into account the magnetic materials nonlinearity. The mechanical rotation is modeled, using the slipping surface technique. The obtained results are compared with measurements and results of the 2D standard model, which was applied to a similar motor but with an unskewed rotor.

A Three-Dimensional Electromagnetic Shell Finite Element for Coupled Vector-Scalar Potential Formulations

In many eddy current problems solved by FEM, meshing of the skin depth leads to CAD difficulties and large computing. In order to avoid such problems, one currently use surface impedance conditions: boundary conditions in the case of massive conductors or transmission conditions or shell elements in the case of thin conducting layers. In this paper, we propose a new shell finite element with arbitrary thickness that is compatible with the (A-φ) and (T-Ω) formulations. Both formulations exist in some general purpose finite element electromagnetic codes like Code_Carmel3D. To validate the approach, two elementary numerical tests are treated.

An Arbitrary Thick Shell Finite Element for Eddy-Current Dual Vector-Scalar Potential Formulations

In many eddy current problems, surface impedance conditions are currently used to avoid the modelling of the skin depth in volumes. In the finite element method context, such conditions lead to surfaces elements at boundaries of massive conductors, or to shell elements representing thin layers, which unfortunately require a complex specific implementation in the codes. We propose here a new shell finite element easy to implement in the standard (A-φ) and (T-Ω) formulations. The advantage of this element is to be arbitrarily thick, so it takes into account some end-effects at its perimeter such as the connection between shell an 3D regions. The basic principles of the thick shell as well as a three-dimensional validation test dealing with a magnetic core-fault are presented.