Conference Agenda

Circuit type simulation softwares (like SPICE) are routinely used in power electronics systems to analyze a large variety of topologies and parameter combinations (duty cycle, frequency, input/output voltages, \dots). The magnetic components are usually considered linear in these simulations. This article discuss the feasability of implementing the Energy Based Hysteresys Model (EBHM) into a harmonic balance simulation in order to obtain an efficient tehnique to obtain the steady response of a power electronic topology includin the saturation and hysteresis effect of the magnetic component.

Cracks in concrete affect its structural integrity and related performance. Accordingly, concrete diagnosis results in a task of primary importance in making informed decisions about reconstructing or repairing buildings. Among the different approaches for evaluating concrete characteristics, methods based on electromagnetic waves have been proposed. In this context, the characterization of the concrete complex dielectric permittivity $\epsilon_{eff}(\omega)$ has received significant attention, and within this framework the Debye, the Jonscher model and its variants have been intensively used. Although the above are consolidated models, recently, data-driven techniques have begun to emerge as alternatives for material property modeling because of their regression and generalization potential. In this study, we investigate the ability of Gaussian Process Regression to effectively model $\epsilon_{eff}(\omega)$ by comparing its performance with that of the standard Debye and Jonscher models.

In this paper, the stepwise monotonic hysteresis model is used to evaluate the hysteresis losses in synchronous machines under the influence of mechanical stress, thereby needing only a single magnetostatic simulation. The mechanical stress is considered by using a magneto-elastic vector hysteresis model within the finite element formulation. The loss evaluation is applied on a permanent magnet synchronous machine affected by a mechanical stress state induced by manufacturing. The results are compared to a full transient model and show a good agreement while only needing a fraction of the computational time related to a transient simulation.

In this work, the Simplified Multiscale Model (SMSM) is integrated into the energy-based vector hysteresis framework and represents anhysteretic material behavior, enabling the inclusion of effects such as mechanical stress, magnetostriction, material anisotropy, and crystallographic texture into hysteretic simulations. This combination allows for detailed material information to be utilized while accounting for dissipative effects. This methodology is applied to a permanent magnet synchronous machine (PMSM), including rotor rotation and the influence of mechanical stresses, demonstrating its effectiveness in modeling complex magnetic phenomena in electric machines.

Circuit type simulation softwares are routinely used in power electronics systems to analyze a large variety of topologies and parameter combinations (duty cycle, frequency, input/output voltages, ...). The inductors present in such systems very often contain magnetic cores exhibiting a significant hysteresis behaviour. Accurate hysteresis models are therefore needed to predict magnetic quantities and losses in those cores. This paper shows how to implement the Energy-Based hysteresis model into the LTspice software, in order to account for hysteresis phenomena in the magnetic cores of inductor components in transient simulations. Some results are compared with measurements on a DC/DC buck converter.

This paper introduces a new anhysteresis curve for electromagnetic simulations such as transformers, rotating machines, or magnetic devices. Unlike existing anhysteresis curves, this curve has several advantages, including continuous differentiability, smoothness for numerical solvers, and characterization using four parameters, each responsible for a specific magnetization region. The new anhysteresis curve is validated by fitting various magnetic materials, and its accuracy is compared to well-known curves such as the (double) Langevin function, the inverse tangent function, and the hyperbolic tangent function. The incorporation into an energy-based hysteresis model is also discussed briefly.

Having accurate material models is crucial for performing high fidelity simulations of electrical machines.

Obtaining the parameters of such material models usually requires a high number of simulations to be performed.

This is often not feasible due to the complex nature of the underlying models. To this end, machine learning

methods, specifically Gaussian Process Regression (GPR) is used to approximate certain relationships obtained

from simulation data.

In this paper, we investigate the application of data driven methods for the parameter identification of

magnetic properties in steel sheets. First, we show how we apply the methodology to the nonlinear case without the

hysteresis and then give an outlook for the final results including the hysteresis models which will be included

in the full paper.

The development of models for the modeling and numerical simulation of the electromagnetic behavior of electronic-based systems

(EBS) - in order to predict their electromagnetic compatibility (EMC) with other electrical devices - requires a good knowledge of

the specific materials to be used in such EBS. Ferrite materials are often used for inductors, such as those found in electromagnetic

interference (EMI) filters. Common mode chokes (CMC) are a prominent example of such inductors. Unfortunately, the frequencydependent

dispersive permeability of such ferrite materials is often unknown, so the complex permeability must be determined by

measurement such that certain model parameters can be fitted to the observed behavior. These material models can then be used

for developing a digital twin of the EBS to pre-estimate the EMC behavior. In the present work a time-domain measurement based

approach is proposed which enables the determination of the frequency-dependent permeability of a CMC by using the fast Fourier

transform (FFT) of the time domain measurement signal. A comparison of the proposed method to a frequency-domain based method

is presented.

The presented paper focuses on the influence of manufacturing, i.e., laser cutting, on the magnetic behavior of soft magnetic

steel sheets. A material degradation model (MDM) on the basis of an analytical eddy current (EC) model is introduced, and the

corresponding identification routine is presented. For the model parameter’s determination ring shaped specimen (RS) in two different

material grades, cut with two different laser cutting techniques, in five different geometries respectively are characterized for their

magnetization behavior. Furthermore, to verify the results wire eroded samples, unaltered in their magnetic properties, are utilized

as reference. Moreover, scanning electron microscopy (SEM) is employed to investigate the changes in the material’s micro-structure,

as well as a verification for the modeling results.

The aim of this paper is to present a homogenization method that is suitable for obtaining the transversal conductivity of a unidirectional CFRP material made of fibers with a non-metallic coating. This conductivity is dominated by the connection of the fibers; with a fictitious compression of the medium, it can be obtained through the conductance per unit length. The coating, which has a particularly important role in the connections, has a poor conductivity; its modeling is also included in the study.

In effective medium theory, a material parameter of a heterogeneous structure is replaced by an effective material, which is derived from solving a cell problem and used in a simulation with a homogenised core. With results of this simulation, the iron losses and the reactive power of machines or transformers with laminated cores are determined. For this work, a nonlinear magnetic material considering magnetic vector hysteresis is employed to simulate a single-phase transformer, providing insights into its performance.

The energy-based vector hysteresis model of Francois-Lavet et al. establishes an implicit relation between magnetic fields and fluxes via internal polarizations each of which is determined by a convex but non-smooth minimization problem. As shown in a recent contribution by the authors, the inverse relation can be defined in a similar manner. The computation of the internal variables however now leads to a coupled convex and non-smooth minimization problem whose dimension depends on the number of the internal variables. We show that this problem can be solved efficiently. After regularization of the non-smooth terms a Newton method with line search can be applied. By introducing the total polarization as an additional internal unknown, the Newton step can be reduced to solving independent small linear systems for each internal variable. The computational complexity for evaluating the inverse hysteresis operator thus essentially reduces to the same as that for computing the forward hysteresis operator. The theoretical results are illustrated by comparing the computation times for some standard benchmark problems.

This paper proposes a simple interpolation method employing parametric polynomial approximation to identify the shape function of a play model. Among various hysteresis models, the play model provides relatively high accuracy; however, its precision is heavily influenced by the number of play hysterons, which depends on the number of measured hysteresis loops. Achieving this level of accuracy in applications such as accelerators requires both appropriate hysteresis loop divisions and approximately 1,000 hysteresis loops, which makes such extensive measurements highly demanding. This effort is generally unnecessary for standard applications like motors but is critical for high-precision applications. To address this issue, this study proposes a method for interpolating and generating a large number of symmetric hysteresis loops from a small number of measurement data. The proposed method successfully reduces the burden of data collection while maintaining high accuracy through the interpolation of hysteresis loops with equidistant maximum values. Normalization effects brought this success. The proposed approach is expected to have future applications in the characterization and design of accelerator magnets.