Conference Agenda

In this research, we applied different analytical and numerical dynamic magnetization models on Non-oriented (NO) steel to reproduce the magnetization field H(B) and energy loss in both Rolling and Transverse Directions. The conventional analytical dynamic magnetization model applied is the well-known Bertotti’s three-term model, in which a simplified equation for the classical eddy current loss is used. And the numerical model applied is the one-dimensional (1D) finite element model, which is a nonlinear eddy current analysis with modified conductivity, hence nonlinearity, skin effect, and excess loss are all included. Both models can provide a reliable basis for energy loss through choosing parameters fitting with the measured loss. But use of the Bertotti’s model provides too much straight dynamic loop shapes compared with the measured ones, vice versa for the use of the 1D model. Therefore, the combination method of the two models is proposed. It shows the best performance among the three to represent the dynamic loop shape, and hence energy loss for both the NO steel in rolling and transverse directions.

This paper presents calculations of self and mutual inductances using an integral approach between long carbon fibers in unidirectional composites materials. The random distribution of fibers is considered via an algorithm designed specifically for this purpose. The algorithm takes into account repulsion constraints and ensures no overlap between the fibers. Simulation results, presented as images, depict the random distributions of fibers with variable volume fractions (from 40% to 70%). The final calculations reveal mutual inductances in the actual state of the fibers and provide insights into values at increasing distances. The magnitude of the inductances, along with resistance calculations included in the extended format of the paper, enables a better understanding of the effects of frequency on these types of materials.

This paper deals with numerical modeling of current density and magnetic field in a superconducting wire carrying AC transport current. For that, we use the finite element method. We observe the current density distribution and including magnetic field distribution in order to study the electromagnetic behavior in superconducting materials. Then we compare the results obtained by using different numbers of filaments in the superconducting wire.

This paper examines the computation accuracy of the play models with B-input and H-input, whose shape functions are identified from the Everett function, by comparing their computational hysteresis loops with the measured ones. As a result, it was revealed that there is a significant difference in the computation accuracy between the play models with B-input and H-input for the asymmetric hysteresis loop, though their computation accuracy is equivalent for the symmetric one.

This paper proposes a novel hysteresis modeling approach using a recurrent neural network (RNN) to accurately describe the hysteresis behavior of magnetic materials, incorporating eddy current losses through a Cauer circuit. The RNN-based hysteresis model is trained with measured hysteresis loops and effectively predicts untrained loops with high accuracy. The proposed model is integrated into Cauer circuit to extend its applicability to high-frequency regions, where eddy current effects become significant. Experimental results validate the model's predictions for magnetic losses and hysteresis loops, demonstrating its reliability and potential for advanced magnetic component design.

Internal microcracks, which are microscopic fissures within materials, can significantly impact their mechanical, thermal, and electrical properties. This paper addresses the challenges in inspecting these defects by employing eddy current nondestructive testing.

It focuses on numerically modeling a steel sheet under varying stress states and characterizing eddy current indices influenced by factors such as delamination and defect geometry. The study explores the relationships between microcrack characteristics like

location, orientation, and size within single and multilayer conductive plates. The researchers compiled a comprehensive database of model signals from defects of varying geometric parameters, as well as experimental signals from artificially created defects with

known characteristics in multilayer structures.

A biaxially anisotropic vector hysteresis model is developed by the superposition of scalar play models. Its identification method was derived analytically. BH loops of non-oriented silicon steel sheets are accurately reconstructed along the rolling, transverse, and 45-degree directions. The proposed model is applied to the representation of anisotropic pinning field distribution for physical magnetization analysis.

The combination of the play model and the Cauer circuit is an efficient model to represent the dynamic hysteretic properties of magnetic materials. It requires the numerical solution of a nonlinear ordinary differential equation. To avoid the nonlinear iterations, we introduce a new predictor-corrector (PC) scheme called Newton PC scheme to the hysteretic Cauer circuit. We compared the Newton PC scheme with another PC scheme called Heun PC scheme. We found that the Newton PC scheme improves the accuracy successfully even when the Heun PC scheme is not very effective.

In inductors for power electronics, the driving source frequency becomes higher to the MHz range by applying new semiconductors (GaN/SiC). To represent the high frequency (HF) response of the inductor with the iron core of laminated electrical steel sheets by the electromagnetic field analysis (FEA), not only the stray capacitance of winding but also the frequency dependance of magnetic permeability of the steel plate is considered in this paper. The original complex permeability without the influence of the skin effect of the steel plate is determined so that the HF response of inductor obtained from FEA is matched with the measurement. It is shown that the HF response can be represented by taking account of the frequency dependence of magnetic permeability. Moreover, it is also suggested that the real part of complex permeability decreases with the frequency and its imaginary part has the peak at 10 kHz.

This paper proposes a vector extension of the Loss Surface model used for modelling static and dynamic hysteresis phenomena that appear in electromagnetic devices containing soft ferromagnetic materials. The aim being to include the behavior of the material during a solving process based on the finite element method (FEM). The vectorization strategy is based on the Mayergoyz approach. The developed model has been implemented in the 2D FEM formulation. The nonlinear problem is solved using the Newton-Raphson algorithm with an exact approximation of the magnetic reluctivity tensor. Several test cases of the TEAM 32 problem have been studied for validation purposes. The results are in agreement with the measurements combining good accuracy and saving of computation time.

This paper explores the implementation of a phenomenological anisotropy model, which relies solely on measurements along the rolling and transverse directions, within a finite element method (FEM) simulation environment. Currently, grain-oriented laminations are widely used to enhance the efficiency of electrical machines, but their complex behavior demands accurate modeling. Nonlinear simulation for this model are performed on a simplify case of a transformer, and compared with a model based solely on measurements.