This paper presents the application of a homogenisationtechnique, based on the multi-scale expansion theory, to theanalysis of heterogeneous materials constituted of magneticinclusions dispersed in a dielectric lattice. The role of the shapeand dimensions of the inclusions is analysed with reference to theeffective electromagnetic properties and energy losses. Theinvestigation is extended to the influence of flux waveforms withharmonic distortion, focusing the attention on the energy lossdependency on the harmonic content.

New microwave electronic technology challenges require the integration of many passive components onto chips. Among them, isolators and circulators are non-reciprocal passive devices that contain magnetic materials. It is therefore important to know the magnetic properties of suitable materials. The permeability tensor is the most important parameter to define, because it governs the interaction between the electromagnetic wave and the material, and is thus the origin of all magnetic phenomenons. Our work consists of performing the characterization of thin films of BaM and YIG over a frequency range of 0.4 GHz to 65 GHz. The work shows the presence of a gyroresonance phenomenon and locates the resonance frequencies. The technique is based on S parameter measurements on the magnetic materials performed with a network analyser and a probe tester. Polder's model is used to calculate the different elements of the permeability tensor, which consider the ferrite in its saturated state.

For designers, calculation of local fluxes can be very useful. Inthe vector potential formulation, the local fluxes can be easilydeduced. In the scalar potential formulation, the determination ofthese fluxes presents some difficulties. In this paper, we presentthree methods to compute a flux through any surface in the scalarpotential formulation. These are compared with the one used in thevector potential formulation for two application examples.

Solute dispersion is studied experimentally in periodic or disordered arrays of beads in a capillary tube. Dispersion is measured from light absorption variations near the outlet following a steplike injection of dye at the inlet. Visualizations using dye and pure glycerol are also performed in similar geometries. Taylor dispersion is dominant both in an empty tube and for a periodic array of beads: the dispersivity l d increases with the Péclet number Pe respectively as Pe and Pe 0.82 and is larger by a factor of 8 in the second case. In a disordered packing of smaller beads (1/3 of the tube diameter) geometrical dispersion associated to the disorder of the flow field is dominant with a constant value of l d reached at high Péclet numbers. The minimum dispersivity is slightly higher than in homogeneous nonconsolidated packings of small grains, likely due to wall effects. In a weakly disordered packing with the same beads as in the periodic configuration, l d is up to 20 times lower than in the latter and varies as Pe γ with γ = 0.5 or = 0.69 (depending on the fluid viscosity). A simple model accounting for this latter result is suggested.

In this paper, the modeling and study of a new non-linear hybrid current controller is presented. It ensures high dynamic response of the current loop with a fixed switching frequency operation mode, a zero static error and high robustness properties as regard to system parameters variations. To model the proposed non-linear current controller, different tools are developed. In a first step a high frequency average model is proposed. It allows studying the average dynamic properties such as bandwidth, time response and overflow. To investigate the behavior of the current ripple due to the switching effect, a second model, based on the construction of a three-dimensional bifurcation diagram and the definition of a form function is established. This model allows studying the nature of the cycle described by the state trajectory and proving that the system operates with a fixed switching frequency. Design rules of the control parameters of this controller are explained and its robustness properties are tested by numerical simulations and experimental tests.

In this paper, we present a modeling of a captured ferromagnetic particles quantifying method based on the detection and evaluation of the magnetic field variation caused by the particle static buildup. The detecting element is a Hall-effect sensor (probe) associated to the particle capture element which is an electromagnet with iron core. From a known value of the magnetic field measured at the particle static buildup locality, we predict the particle global concentration in the treated material sample (powder or liquid). This prediction is achieved by solving the inverse problem for the captured particles magnetic field variation. Such a resolution is based on an iterative resolution of the forward problem until obtaining the global particle concentration related to the measured value of the field variation. Because an exact value can not be obtained we retain the value that minimizes an error function. To minimize such a function we have used the Tabu searching method.

We introduce pairing matrices on simplicial cell complexes in discrete electromagnetism as a means to avoid the explicit construction of a topologically dual complex. Interestingly, the Finite Element Method with first-order Whitney elements — when it is looked upon from a cell-method perspective — features pairing matrices and thus an implicitly defined dual mesh. We show that the pairing matrix can be used to construct discrete energy products. In this exercise we find that different formalisms lead to equivalent matrix representations. Discrete de Rham currents are an elegant way to subsume these geometrically equivalent but formally distinct ways of defining energy-products.

A study of compressibility effects on structure and evolution of a sheared homogeneous turbulent flow is carried out using rapid distortion theory (RDT). The focus of this paper consists in validating RDT code by considering the direct numerical simulation (DNS) and RDT results of Simone [Simone, Coleman and Cambon, Fluid Mech. 330, 307 (1997)]. We will use this theory to clarify the physics of the compressible turbulent flows. An analysis of the behavior of different terms appearing in the turbulent kinetic energy and the Reynolds stress equations permit to well identify compressibility effects which allow us to analyze performance of the compressible model of Fujiwara and Arakawa concerning the pressure-dilatation correlation. The evaluation of this model stays in the field of RDT validity.

This paper compares the hybrid FEM-BEM and FEM-DBCImethods for the solution of open-boundary electrostatic fieldproblems. Both methods couple a differential equation for theinterior problem with an integral equation for the exterior one. Thecomparison shows that FEM-BEM is more accurate than FEM-DBCI butrequires more computing time.

The aim of this research is the elaboration of a method allowing simple simulation of electromagnetic interferences (EMI) of complex electric systems. The method presented in this paper allows the replacement of the electromagnetic active objects by simple closed surfaces that surround each source. The flux density components are calculated on these surfaces using standard 3D finite element software separately for each object. The magnetic flux density in every point of the free space around the system is calculated with an analytical formula. The main topic of the paper is the definition of the limits of the discussed method.

Steady-state operation modes of three-phase induction machines canbe efficiently simulated by 2D nonlinear time-harmonic finite-element models, although only induced currents with respect to the fundamental air-gap field are correctly taken into account.This technique does not generalise to single-phase inductionmachines. An approach based on multiple rotor models and a spectraldecomposition of the air-gap field enables to consider higherharmonic air-gap field contributions. In a capacitor-motor model,the first, third and fifth forward and backward rotating componentsgive raise to different frequencies in the rotor which result indifferent eddy-current effects. The torque dip due to the thirdharmonic is accurately simulated.

In this paper, differential conductivity and reluctivity matricesare constructed for formulations discretised by the finiteintegration technique and linearised by the Newton method. It turnsout that these matrices are nondiagonal and unsymmetric, even in thecase that orthogonal grids are used. A diagonal approximation of thedifferential matrices leads to an approximate Newton method wherecross-magnetisation effects are no longer considered, except in theupdate between the successive nonlinear iteration steps. The approximate method does not guarantee second order convergence but outperforms the true Newton method when only a moderate accuracy is required.

The spectroscopic electrical characterization of biological tissues between 10 Hz and 10 MHz is generally based on four-electrode impedance measurement. An array of four aligned needle-electrodes is commonly used for the probe. The electrode/sample interface impedances and the parasitic capacitances of the system induce measurement inaccuracies respectively in the low and high parts of the studied frequency range. These inaccuracies cannot be removed by any usual calibration method but can be minimized by modifying the geometry of the probe. Using a 3D finite element model of the probe coupled with a deterministic optimization algorithm, it is shown that the configuration of the electrodes should be different at 10 Hz and 10 MHz to improve the performances of the measurement system.

The computation of induced currents in a conductor whose shape isa dome is achieved iteratively by dividing this conductor intoseveral elementary layers to take into account the tridimensionalbehaviour of the induced currents.

This study deals with the wall characterization of a glass furnace based on an electrical resistance measurement between two buried electrodes. The relationship between the wall geometry and the electrical resistance is established. This paper presents the modelling aspects of this inspection technique.

A dynamical model of a deformable circuit submitted to an induction source is described in infinite dimension and in its discrete version. One explains in particular how to practically carryout the calculation of the inductance of the circuit, which is one necessary condition to obtain an exploitable model.

In this paper, a coupled magneto-mechanical finite-element model is briefly presented and then used to determine the different electromagnetical strains of a structure mechanically closed to a single-phase stator of a switched reluctance motor. Magnitudes of the magnetostrictive forces and of the magnetic forces, such as their relative importance in the structure strain, are given for different current excitations.

The aim of this study is to determine the complex permittivity of dielectric materials using a coaxial discontinuity and the combination of neural networks (NN) with the finite element method. Two types of measurement cells are used. One is for solid samples and the other one for liquids. Data sets used to train neural networks are created using the finite element method. The number of hidden neurons of the NN is determined by the split-sample method. The designed NN are used for the estimation of the permittivity of several materials and their results compared with the ones obtained with a gradient inversion method.