R. Van Keer

An Iron Loss Model For Electrical Machines Using The Preisach Theory

In this paper we deal with a mathematical model for the evaluation of the electromagnetic iron losses in rotating electrical machines under no load conditions. The presented problems of electromagnetic field computations are coupled with refined material models based on the Preisach theory. The model is validated by the comparison of numerical results and experimental values from measurements.

Finite element approximation for 2nd order elliptic eigenvalue problems with nonlocal boundary or transition conditions

In this paper, we use two symplectic schemes to simulate the Ablowitz-Ladik model associated to the cubic nonlinear Schrodinger equation and we compare them with nonsymplectic methods.

On a magnetodynamic model for the iron losses in non-oriented steel laminations

In this paper we deal with a numerical model for the evaluation of the magnetic iron losses in one steel lamination. The magnetodynamic model for the magnetic field strength H is coupled with the Preisach model. The resulting boundary value problem for H is solved numerically by a modified finite element - finite difference method. The model is validated by the comparison of numerical experiments and measurements. The dynamic behaviour of two materials is investigated.

Computational methods for the evaluation of the electromagnetic losses in electrical machinery

SummaryThis paper deals with the numerical modelling of electromagnetic losses in electrical machines, using electromagnetic field computations, combined with advanced material characterisations. Due to the complexity of this objective, simplified settings, deliberately choosen by physical arguments, must be considered first. The idea is to proceed gradually to the actual problem of an electrical machine through intermediate models showing physical relevance on their own. The numerical methods used to solve these various problems mainly involve modified finite element-finite difference discretisations, which properly take into account the nonlinear and memory properties of the magnetic material. To this end, one must start from suitable variational formulations of the underlying magnetic field problems. The combined magnetodynamic-hysteresis models are validated in several ways, particularly by comparison of numerically obtained electromagnetic losses with experimental results.

On a numerical method for the evaluation of electromagnetic losses in electric machinery.

SUMMARY The paper deals with a numerical method for the evaluation of the magnetic iron losses in steel laminations used in rotating electric machinery. The magnetic hysteresis and the eddy current effects are directly and simultaneously taken into account. Hereby commonly used analytic expressions for the distribution function in the widely adapted Preisach hysteresis model are found to be not quite accurate. The magnetic circuit is decomposed into magnetic and air gap network elements, connected by fundamental loops. The magnetic network elements show a finite element structure. The kinematics of the electric machine is deliberately taken into account by an interpolation technique. Although the model retains the essential features of a cumbersome 3-D problem, a relatively simple algorithm may be developed. For the resulting algebraic system, we propose a suitable decoupling technique, which is efficient from the computational point of view. Numerical experiments show that the results obtained by our numerical approach are in good agreement with the known behaviour of the magnetic material.

Parameter identification by a single injection?extraction well

In this paper, we present numerical modelling techniques supporting the determination of parameters for the contaminant transport by underground water flow. The parameter identification is based on measurements obtained by a single injection–extraction well. The underground water flow is modelled using a Dupuit–Forchheimer approximation for the unsaturated–saturated aquifer.

Global and local Trefftz boundary integral formulations for sound vibration

The sound-pressure field harmonically varying in time is governed by the Helmholtz equation. The Trefftz boundary integral equation method is presented to solve two-dimensional boundary value problems. Both direct and indirect BIE formulations are given. Nonsingular Trefftz formulations lead to regular integrals counterpart to the conventional BIE with the singular fundamental solution. The paper presents also the local boundary integral equations with Trefftz functions as a test function. Physical fields are approximated by the moving least-square in the meshless implementation. Numerical results are given for a square patch test and a circular disc.

On a numerical method for 2D magnetic field computations in a lamination with enforced total flux

The paper deals with a numerical method for the evaluation of the electromagnetic loss in a lamination of an electric machine, based upon the Maxwell equations. The underlying problem consists in the computation of the unidirectional magnetic field in a cross-section of the laminate orthogonal to the enforced flux. The method starts from a suitable variational formulation of the governing parabolic problem with a nonlocal Neumann boundary condition accompanied by a Dirichlet side condition, involving nonlinear and hysteresis effects through the differential permeability coefficient. The variational problem is solved numerically by a finite element method, combined with a finite difference technique, which is deviced so as to take into account the nonlinear and hysteresis behaviour of the material. The numerical method is found to be effective and reliable.

Modelling of microstructural effects on magnetic hysteresis properties

In this paper, the relationship between microstructural properties of steels and the material parameters in the Preisach model and in the Jiles-Atherton (JA) model is discussed, in the instance where both models describe quasi-static hysteretic magnetic behaviour. It is shown how the material parameters in both hysteresis models should be modified to reflect their dependence on dislocation density and grain size. The dependence of the Preisach material parameters on these microstructural features is identified starting from hysteresis loops calculated by the microstructurally dependent modified JA model. For the Preisach model, a Lorentzian distribution function is used for the distribution function. This makes it possible to compare predictions here to results of an earlier paper in which the Lorentzian distribution was used for Preisach fits to experimental data for steels of different grain sizes. Also, in a different earlier paper, it was shown how the Lorentzian distribution can be formulated so that it connects with salient features of the JA model. The procedure in this paper enables one to examine and predict microstructural variations of Preisach parameters in steels not only for the case of grain size variation but also for the case of variation in dislocation density.

Reduced Conductivity Dependence Method for Increase of Dipole Localization Accuracy in the EEG Inverse Problem

The EEG is a neurological diagnostic tool with high temporal resolution. However, when solving the EEG inverse problem, its localization accuracy is limited because of noise in measurements and available uncertainties of the conductivity value in the forward model evaluations. This paper proposes the reduced conductivity dependence (RCD) method for decreasing the localization error in EEG source analysis by limiting the propagation of the uncertain conductivity values to the solutions of the inverse problem. We redefine the traditional EEG cost function, and in contrast to previous approaches, we introduce a selection procedure of the EEG potentials. The selected potentials are, as low as possible, affected by the uncertainties of the conductivity when solving the inverse problem. We validate the methodology on the widely used three-shell spherical head model with a single electrical dipole and multiple dipoles as source model. The proposed RCD method enhances the source localization accuracy with a factor ranging between 2 and 4, dependent on the dipole location and the noise in measurements.