Search

The paper deals with the design by optimization in pre-sizing step of a drive system that combines a switched capacitor system and a permanent magnet synchronous machine (PMSM). The sizing of the converter and the machine are considered strongly dependent due to the series quasi-resonant topology. The system modelling is based on a semi-analytical approach validated by finite element (FE) simulations. It involves an analytical modelling of the electrical machine and a numerical time solving of the electrical circuit sub-model because of the natural switching events of diodes. The optimization is carried out by SQP (Sequential Quadratic Programming). The model gradients are computed combining mathematical laws, symbolic derivation and automatic differentiation. Finally, trade-offs and gains on the machine power density, the magnet volume and the efficiency are investigated through a Pareto approach.

International audience; This paper deals with system level EMC models for power electronics converter, able to investigate the conducted emissions in embedded grid, for instance. The model topology presented in this work is not symmetric and its identification does not rely on the usual Differential / Common Mode separation. The identification method is presented, and the model is validated in different asymmetric environments: a specific configuration with a modified LISN, and another one with minus connected to ground. Both approaches are carried out in simulation and experiment, for the simple example of a boost converter.

In a few years, Tesla Motors Inc. has succeeded in establishing itself as one of the market leaders in the electric car market. The Tesla Model S, its first flagship vehicle, still arouses the interest and curiosity of engineers around the world, thanks to the manufacturer’s technological secrets and bold choices, in particular about the unusual use of an induction machine in a mass-produced car. Thus, this article aims to present an accurate 2D finite element model of the Tesla Model S 60 induction motor. For this purpose, all the needed data are gathered and completed. Then, a model of the machine is made with Altair Flux2D™ software and several operating points are simulated in order to validate the created model, according to data available in literature. Finally, an analysis of the machine design or operating conditions is made in order to propose possible justifications to some of the technical choices made by Tesla engineers.

We study the relation between the induced anisotropy obtained on nanocrystalline Fe 73.0 Cu 1.1 Nb 3.1 Si 15.7 B 7.1 toroidal cores crystallized under transverse field at 570 °C, and the polarization J c of nanograins. To do this, an original disposition is used, mixing treated cores with cores made of electrotechnical SiFe alloy, featuring same geometry, using the interfacial magnetic charges to increase the field in treated cores. The field supplied by the coil being μ 0 H a = 33 m T , playing on the proportion between treated and SiFe cores, effective field ranging between 33 m T and 230 m T were obtained inside treated cores. As a result, the induced anisotropy at the scale of the alloys at the end of crystallization range from K u = 12 J / m 3 to K u = 20 J / m 3 . To simulate the evolution of J c during the crystallization process, the configuration of treated and SiFe cores in the sample holder was implemented in Altair FluxT FEM simulation environment with the dedicated polarization laws. Dealing with treated cores, the law J cor ( H ) at the scale of the core reflects the law J c H featured by nanograins during the crystallization process. J c H was modelized as a function of the crystallized fraction f c . The relation between the simulated J c and the experimental final values of K u / f c (i.e. the anisotropy in nanograins) was then studied. This was done considering the polarization 〈 J c 〉 averaged on the entire crystallization process, wondering about a possible memory effect.

The increase in active component switching frequency, via GaN technology, allows downscaling the inductor size as soon as the magnetic core-losses are controlled. However frequency induced dissipations in the magnetic material will create thermal heating, consequently complex core shapes allowing a better thermal exchange are needed. Here we propose a specific dimensioning methodology to determine the best form factor to optimise the exchange surface and verify constraints on the inductor. An inductor made of NiZn spinel ferrite powder has been manufactured by tape casting using results of calculations based on magnetic core-loss measurements.

International audience; This paper describes the evolution of a powerelectronics class at Master level (second year), to comply withthe new trends of this science. Indeed, the switching speed ofpower semiconductor devices –especially wide bandgap devices–is continually increasing and, the knowledge and comprehensionof switching mechanisms is absolutely necessary to design powerconverters. Apart from theoretical lectures, several lab workshave been specifically developed, to use these devices at theirrated operating point. Five different themes are studied: devicecharacterization (bare dies), switching waveforms, lossescharacterization, thermal design, ElectroMagneticCompatibility and magnetics design.

Switched reluctance machine has many advantages in both its simple structure and the wide speed range at which it can operate. Nerveless, one of its main flaws is its poor power factor especially in high speed, which leads to over-sizing of the power supply. The purpose of this article is to show the impact of an electronic impedance based on a "Floating Switched Capacitor (FSC)" on the performance of a switched reluctance machine (SRM) controlled via a voltage inverter. Time simulations carried out with Psim™ software show that by associating this device with the machine, the dynamics of growth and decay of the magnetizing current is improved, which in other words improves the power factor. Thus, the machine is able to provide more torque on all of its operating points.

This paper investigates the effect of the cobalt addition on the magneto-crystalline anisotropy parameter of some polycrystalline ferrites. Magnetization curves have been experimentally determined at different temperatures on Ni1−xZnxFe2O4 sintered samples with or without cobalt addition. In each case, a law of approach to saturation has been fitted to the measured data in order to extract the anisotropy parameter K1 representative for the polycrystalline material. Besides, measurement of the complex susceptibility and power core-loss up to 5 MHz has been performed. This approach allows a discussion on the role of cobalt addition on the power core-loss mitigation. It has been found that the substitution with 0.02 mol. of Co leads to a small decrease in the K1 values. This evolution is consistent with the rise in the rotational permeability. However, the change in K1 cannot account for the core-loss reduction with Co. Rather, it is confirmed that Co hinders the domain wall displacement and subsequent dissipation up to an induction level that depends on the grain size.

WOS:000458776900018

In this paper, we propose to control the output power of a surface mounted permanent magnet generator connected to a load with a three phase Magnetic Energy Recovery Switch system (MERS). Each switch consists of four MOSFET or IGBT elements and one capacitor working at machine frequency. Both effects of switch parameter, capacitor value and control variable are described and analyzed with a system model and verified by Finite Element Method (FEM) simulations. It is shown that output power can be not only well modulated but also increased significantly, compared with standard series inverter solution. Also, an optimum operating point can be found regardless of load. A discussion about possible drawbacks of such system is finally done.

This article focuses on the modeling of the hysteresis loop featured by Fe-Cu-Nb-Si-B nanocrystalline alloys with transverse induced anisotropy. The magnetization reversal process of a magnetic correlated volume (CV), characterized by the induced anisotropy Ku, and a deviation of the local easy magnetization direction featuring the effect of a local incoherent anisotropy Ki, is analyzed, taking account of magnetostatic interactions. Solving the equations shows that considering a unique typical kind of CV does not enable accounting for both the domain pattern and the coercivity. Actually, the classical majority CVs obeying the random anisotropy model explains well the domain pattern but considering another kind of CVs, minority, mingled with classical ones, featuring a magnitude of Ki comparable to Ku, is necessary to account for coercivity. The model has been successfully compared with experimental data.

This paper focuses on the modeling of the hysteresis of Fe–Cu–Nb–Si–B nanocrystalline alloys with transverse induced anisotropy $K_{u}$ based on the behavior of a correlated volume (CV). Only coherent rotation mechanism was considered, given the chosen value of $K_{u}$ . The formulations account for $K_{u}$ , the incoherent anisotropy $K_{i}$ , and the magnetostatic energy. The solving of the equations shows the existence of two kinds of CVs: the first one is responsible for the domain pattern, while the other causes the coercivity. The model has been compared with experimental data, and improved by considering the existence of a minority of CVs featuring strong incoherent anisotropy, causing the coercivity, mingled with classical CVs obeying the random anisotropy model.

International audience; At high temperature, FeSi nanograins obtained after partial crystallization of amorphous Fe-Si-Cu-Nb-B precursor are magnetically decoupled and superparamagnetic, due to the disappearance of the magnetism of the residual amorphous matrix. At a transition temperature Ttr typically ranging between 600 K and 850 K, still above the amorphous Curie temperature, a spontaneous polarisation JSsf appears, despite a much lower blocking temperature. Magnetostatic interactions, rendered by the Lorentz field (L.f.), or a residual ferromagnetic coupling between grains through the amorphous medium, have been invoked to explain this behavior. We investigate here the magnetostatic hypothesis. An improvement of the L.f. model is carried out, based on the idea that the field acting on a nanograin is screened by a surrounding soft shell. The permeability of this shell in the vicinity of Ttr is determined, leading to an attenuation factor S = 2.8 and a decrease of Ttr compared to the L.f. approximation. Moreover, a description of the spontaneous magnetization curve JSsf(T) / JS(T) is obtained, with a slope near Ttr much sharper than predicted by the (Ttr - T)1/2 law associated to an invariant molecular field factor. Comparison with experiments shows good agreement.

In the design of power electronic converters, the choice of the suitable magnetic alloys and the measurement of the variations of their magnetic properties versus the temperature are fundamental issues. A theoretical model based on Neel theory and that takes into account wall displacement and coherent rotation was elaborated to derive the expression of the complex permeability versus the frequency. The static and the dynamic behavior are considered. The dependence of temperature is implicit in this model, i.e. the parameters involved in the expression of the complex permeability may depend on the temperature. An impedance analyzer has been used to measure the open circuit and short circuit complex impedances of a toroidal core versus the frequency at several temperatures ranging from 20°C to 180°C. The complex permeability of the material is then deduced. Finally, the theoretical model is used to identify the domain susceptibility and the anisotropy energy of the materials. The effect of temperature on these parameters is analyzed. The result obtained are valuable for the design of a power electronic converters from the magnetic and thermal viewpoints. The results prove the need for the combination of heat transfer and magnetic mechanisms at the design.

Ni-Zn-Cu ferrites are exhibiting excellent magnetic properties at large frequencies and are consequently produced as low-loss magnetic materials used for magnetic components in power electronics. Here we present an experimental study of (Ni0.31Zn0.49Cu0.20)0.979Co0.021Fe1.9-dO4 ferrites synthesized by the Pechini type sol-gel method. Their structural and magnetic properties are characterized and compared to the ones of reference ferrites synthesized by classical solid-state reactions. Optimized parameters for Pechini method allow obtaining comparable low magnetic core-losses for both synthesis methods. The cores obtained by this sol-gel method present higher static permeability and lower coercive field than the ones obtained by classical solid-state reactions. Results suggest that the microstructure of ferrites synthesized by Pechini method can be considered as a promising alternative route for the design of low-loss magnetic components.

Behaviour of the Nanocrystalline multilayer cores are investigated by use of a 1D electromagnetic propagation model. The magnetic properties of conductive ferromagnetic materials are considered using a domain-wall magnetization model. The physical parameters of the model are domain's width, d, and the initial susceptibility χ 0 . Numerical solutions of the problem taken up with impedance measurements are used to correlate d and χ 0 to macroscopic behaviour of the material.

International audience; the paper describes the application of a brand newapproach to evaluate magneto harmonic effects in roundgeometries. Largely employed regular modeling approaches suchas Dowell or homogenization methods show important lacks ofaccuracy for 2D modeling. They are not able to reflect manyphysically existing phenomena like side effects and empty layersof conductors e.g, what require Finite Elements simulation. Newproposed analytical model brings all FEM simulation capabilitiesand often outrivals them for high frequencies and computingperformance. Any geometry of stacked conductors can besimulated and all electromagnetic values can be retrieved in 2Dspace (potential vector, current density, electric and magneticfield). The mathematical basics are reminded, then the extensionfor several conductors in 2D space without any meshing ispresented as well as comparisons with Finite Element analysisresults are shown.

International audience; the paper describes an application of themagnetic potential based so-called “A-model” to evaluatemagneto harmonic effects in round hollow and clad metalsconductors. These technologies of conductors showinteresting properties to maximize the energy efficiency ofwindings. However, beside Finite Element Methodsimulation no analytical modeling makes it possible tosimulate with accuracy those windings. The A-methodbased on analytical solution without any object meshing, isextended for these two new geometries and reflectsaccurately the electromagnetic behavior of the analyzedconductors. The mathematical basics are reminded, thenthe extension for hollow and clad metal conductors in 2Dspace is presented and chosen applications are also shown.The results have been validated by comparison with FEMsimulation.

International audience; The magnetic behavior under applied mechanical stress at audible and inaudible frequencies is investigated. A focus is given on how remanent induction, coercive field strength and permeability values that describe hysteresis-loop shape are involved. To this aim, an experimental setup is introduced to both measure mechanical and magnetic behaviors. The mechanical resonant modes of a magnetic ring core sample are tracked to allow large longitudinal bendings. It results that the coercive field strength and remanent induction are greatly reduced under these mechanical solicitations. Those results are then set forth to show that elastic mechanical applied stress is relevant to describe these effects for both steady and transient operating conditions. The experimental results are then focused on how specific magnetic losses and complex permeability spectra depend on the applied stress amplitude. These results are finally discussed to highlight future applicative developments.

International audience; The decrease of coercive magnetic field strength according to applied mechanical stress at audible and inaudible frequencies is investigated. To this aim, an experimental setup is introduced to both measure mechanical and magnetic behaviors. The mechanical resonant modes of a magnetic ring core sample are tracked to allow large longitudinal bendings and to reduce significantly the hysteresis loop width. Results are then set forth to show that elastic mechanical applied stress is relevant to describe the reduction of coercive field strength measured for both steady and transient states.

Thermal models are more and more used to simulate power electronic devices. For example, in aeronautic domains, the numerical validation of prototype in high temperature environment is very important. Thermal modeling is generally performed using finite element or finite volume methods, which are difficult to integrate in the design procedure. These simulations must then combine different software and lead to important computation times. To develop reduced models, many methodologies are proposed as: the thermal quadrupoles method [1][2][3], the nodal method [4], Pole Analysis via Congruence Transformation (PACT) method [5] and thermal matrix description. All these methods allow developing conductive thermal models with a limited number of unknowns and are called compact thermal models. This paper focuses on how the thermal matrix description methodology can be used with finite element method to generate reduced models that are boundary-condition-independent. This method is successfully used in the present study to generate reduced models for 1D and 2D.

International audience

International audience; Ultrasoft magnetic materials, such as amorphous, nanocrystalline, and polycrystalline alloys, have been successfully used for power electronic applications during recent years. However, enhancements are needed for the integration of power electronic features, which involves high power densities and operating frequencies up to a few megahertz. Complex permeability spectra, which are used to describe material behavior in these applications, depend mainly on domain-wall motions and coherent magnetization rotation mechanisms. In this paper, we present a model describing these mechanisms, in terms of magnetic anisotropies and domain structure of the material. To validate the model, we measured permeability spectra of polycrystalline Ni-Fe alloys under mechanical stress using a specific setup. These measurements are suitable for comparing the physical model according to different magnetoelastic anisotropies. Results are useful for correlating high-frequency magnetic behavior and the annealing process of materials.

International audience; An equivalent circuit has been developed to model the linear electrical behaviour of any two windings common mode chokes. Its topology is independent of size and technology and a general method of characterization, exclusively based on external impedance measurements has been presented.

This study deals with the characterization and modeling of a soft ferromagnetic sample of Cobalt-Iron under vibratory mechanical stimulation. The characterization test bench allows to control the magnitude and the frequency of the mechanical solicitation through piezoelectric actuators. The magnetic hysteresis loop and the magnitude of the mechanical deformations are measured simultaneously when the ring sample is placed under longitudinal bending. The experimental results show a significant reduction of the static magnetic hysteresis. Decreases of coercivity are linearly correlated to the amplitude of elastic strain applied on the torus regardless of the mechanical vibration frequency. These reductions, observed under a cyclic external stress, are associated with the depinning of Bloch domain walls when they encounter residual stress defects. To clarify these results, a modeling of the interaction between domain wall and defect is proposed using an energy approach. Energy profiles interaction is used to express the coercivity attributed to residual stress defects in the presence of an external stress of both variable amplitude and/or variable direction(s). The model highlights the need to explore a large number of mechanical states for each magnetic field value. Finally, applications raised by this work are proposed from an energy balance analysis performed on the magneto-mechanical excitation system. This study opens new opportunities for vibratory mechanical energy harvesting devices and systems for demagnetization by mechanical activation., La thèse porte sur la caractérisation et la modélisation d'un échantillon ferromagnétique doux de Fer-Cobalt sous stimulation mécanique vibratoire. Le banc de caractérisation permet de maîtriser l'amplitude et la fréquence de la sollicitation mécanique au travers d’actionneurs piézoélectriques. Le cycle d'hystérésis magnétique ainsi que l'amplitude des déformations mécaniques sont mesurés simultanément lorsque l'échantillon torique est sollicité en flexion longitudinale. Les résultats expérimentaux montrent une réduction importante de l'hystérésis magnétique statique. Les réductions de coercivité sont corrélées linéairement à l'amplitude des contraintes élastiques exercées sur le tore quelle que soit la fréquence mécanique des vibrations. Ces réductions, observées en présence d'une contrainte externe cyclique, sont associées à des décrochements de parois de Bloch lorsque celles-ci sont piégées par des défauts de contraintes internes. Pour interpréter ces résultats, une modélisation de l'interaction paroi - défaut est proposée à l'aide d'une approche énergétique. Les profils d'interaction énergétique entre paroi et défaut permettent alors d'exprimer la coercivité attribuée aux défauts de contrainte interne en présence d'une contrainte externe d'amplitude et/ou de direction variable(s). Le modèle met en évidence la nécessité d'explorer un grand nombre d'états mécaniques pour chaque valeur de champ magnétique. Les portées applicatives de ces travaux sont proposées à l'issue d'un bilan énergétique effectué sur le système d'excitation magnéto-mécanique. Cette étude ouvre de nouvelles voies pour des dispositifs de récupération d'énergie mécanique vibratoire ainsi que pour des systèmes de désaimantation par activation mécanique.

The aim of this thesis is to implement the method μPEEC to assess, without using simulations or measurements, the reluctance of magnetic circuits and simple 2D to predict their behavior from electromagnetic design phases . It's just asking the field created by a blade straight conductor placed in a round window or rectangular magnetic circuit with or without an air gap. This study is part of a larger research aimed at developing the equivalent circuit of transformers, before prototyping to optimize these components in their applications. Before reaching this goal, several steps were taken gradually, with the help of analytical solutions and finite element simulations to validate our approaches., Le but de cette thèse est donc de mettre en œuvre la méthode µPEEC pour évaluer, sans avoir recours à des simulations ni à des mesures, la réluctance de circuits magnétiques simples en 2D et ce, afin de prédire leur comportement électromagnétique dès les phases de conception. Cela revient à chercher le champ créé par un brin conducteur rectiligne placé dans une fenêtre ronde ou rectangulaire de circuit magnétique comportant ou non un entrefer. Cette étude s'inscrit dans le cadre d'une recherche plus vaste, visant à élaborer le circuit équivalent de transformateurs, avant la réalisation de prototypes, afin d'optimiser ces composants au sein de leurs applications. Avant d'atteindre ce but, différentes étapes ont été franchies progressivement, en s'aidant de solutions analytiques et de simulations par éléments finis pour valider nos approches.

The aim of this thesis is to implement the method μPEEC to assess, without using simulations or measurements, the reluctance of magnetic circuits and simple 2D to predict their behavior from electromagnetic design phases . It's just asking the field created by a blade straight conductor placed in a round window or rectangular magnetic circuit with or without an air gap. This study is part of a larger research aimed at developing the equivalent circuit of transformers, before prototyping to optimize these components in their applications. Before reaching this goal, several steps were taken gradually, with the help of analytical solutions and finite element simulations to validate our approaches.; Le but de cette thèse est donc de mettre en œuvre la méthode µPEEC pour évaluer, sans avoir recours à des simulations ni à des mesures, la réluctance de circuits magnétiques simples en 2D et ce, afin de prédire leur comportement électromagnétique dès les phases de conception. Cela revient à chercher le champ créé par un brin conducteur rectiligne placé dans une fenêtre ronde ou rectangulaire de circuit magnétique comportant ou non un entrefer. Cette étude s'inscrit dans le cadre d'une recherche plus vaste, visant à élaborer le circuit équivalent de transformateurs, avant la réalisation de prototypes, afin d'optimiser ces composants au sein de leurs applications. Avant d'atteindre ce but, différentes étapes ont été franchies progressivement, en s'aidant de solutions analytiques et de simulations par éléments finis pour valider nos approches.

In the context of the more electric aircraft, the industry replaces pneumatic or/and hydraulic actuators for electromechanicalactuators. The ultimate goal is to reduce the weight and improve the overall energy efficiency of an aircraft to reduce fuelconsumption. This thesis aims the mass reduction converters while providing criteria EMC interference (electromagneticcompatibility). In fact, EMC standards are becoming more stringent, particularly in sensitive areas such as aeronautics. EMCfilters are more numerous. In these filters, the magnetic components have more than 50% of the space and weigth. Theobjective of this thesis is to design innovative EMC filters solutions, using a single magnetic component, to reduce conducted disturbances (common and differential modes). Following a review of the literature, we conducted a comparative study of single magnetic component topologies. We have developed a modeling tool frequency magnetic single component. From a modeling <> we have to take into account the environment of the magnetic component (Impedance Stabilization Network Line (LISN), Source disturbances and capacities) and a finer modeling of magnetic properties of component. With a gyrator-capacitor model, we could predict the performance of the complete filter in its environment and optimize it. This study has allowed us to achieve a unique magnetic topology component more compact and efficient than the existing topology. In addition, we have demonstrated the feasibility of integrating PCB (Printed Circuit Board) of nanocrystalline alloy by performing another single component topology based on the same factors as above but integrated into a PCB. The technological obstacles which we faced and solutions we have proposed to solve them are detailed in the report., L'industrie aéronautique, dans le cadre de l'avion plus électrique, met en oeuvre le remplacement d'actionneurs pneumatiques ou hydrauliques par des actionneurs électromécaniques. Cela permet de réduire le nombre de réseaux de distribution présents dans l'avion, de faciliter la maintenance et d'améliorer le rendement. Le but final est de réduire la masse et d'améliorer l'efficacité énergétique globale d'un avion pour réduire sa consommation en carburant. Les travaux de ce mémoire de thèse s'inscrivent donc la logique de réduction de la masse des convertisseurs tout en assurant des critères de perturbations CEM (Compatibilité ElectroMagnétique). En effet, les normes CEM sont de plus en plus contraignantes, en particulier les domaines sensibles tel que l'aéronautique. Les filtres CEM sont donc de plus en plus nombreux. Dans ces filtres, les composants magnétiques occupent plus de 50 % de la place et surtout de la masse. L'objectif de cette thèse est de faire émerger des solutions innovantes, en dotant les filtres CEM passifs d'un composant magnétique unique pour les perturbations de mode conduites afin de réduire sa masse. Afin de comprendre leur fonctionnement, suite à une étude bibliographique, nous avons réalisé une étude comparative de topologies à composant magnétique unique. Nous avons développé un outil de modélisation fréquentielle du composant magnétique unique. A partir d'une modélisation <> nous avons pu prendre en compte l'environnement du composant magnétique (Réseau Stabilisateur d'Impédance de Ligne (RSIL), Source de perturbations, Capacités) et une modélisation plus fine des propriétés magnétiques du composant. Grâce à un modèle gyrateur-capacité, nous avons pu prédire les performances du filtre complet dans son environnement et l'optimiser. Cette étude nous a permis de réaliser une topologie de composant magnétique unique plus compacte et plus performante que la topologie existante. Par ailleurs, nous avons démontré la faisabilité de l'intégration sur PCB (Printed Board Circuit) d'alliage nanocristallin en réalisant une autre topologie de composant unique basée sur les mêmes facteurs que la précédente mais intégrée dans un PCB. Les verrous technologiques auxquels nous avons été confrontés et les solutions que nous avons proposées pour les résoudre sont détaillés dans le mémoire.