Conference Agenda

The plasmonic modulator antennas with high field enhancement values are essential to detect weak fields at high speeds for future communication systems. Consequently, such antennas with high field enhancement values at frequencies reaching up to 500 GHz are rising interest in the research. These antenna systems enable direct mapping of electromagnetic waves onto the optical signals; thereby, eliminate the radio frequency (RF) losses that are introduced by the conventional RF to optical conversion systems. The high field enhancement values over a broad frequency bandwidth at millimeter-wave frequencies are aimed for the antennas to obtain efficient, robust, and accurate operation. The geometry of the antenna, as well as the dimensions of the geometric parameters, determine the field enhancement response with respect to frequency. In order to reach the desired characteristics, the optimization of the geometric design of the plasmonic modulator antennas via genetic algorithm is carried out, and promising results are presented in this digest.

Microspeakers are used to generate sound in portable devices, such as smartphones. The space for microspeakers is decreasing, owing to the recent trends in smartphone designs, which leads to a reduction of back volume from 0.6 cm3 to 0.25 cm3. This reduction results in an increase in the stiffness of the microspeaker, which affects the sound pressure level (SPL) at low frequencies. The current paper proposes a dual coil microspeaker design for SPL enhancement at low frequencies while having a similar exterior size to a prototype microspeaker. The analysis method, which solves the multiphysics system, was verified through experimentation in advance. The proposed dual coil microspeaker was designed and analyzed using three-dimensional finite element analysis (FEA). Samples were manufactured and tested; Compared to prototype, novel dual coil microspeaker in serial connection presented an SPL improvement of 2.27 dB at low frequencies.

The ring magnet used in the safety window motor is to detect the position of the window with a hall sensor. The dead zone with weak magnetic flux appears between the poles of this ring magnet, and the smaller the angle of the dead zone, the more precise position control is possible. In this paper, to reduce the dead zone of a ring bonded magnet, the shape of polar anisotropic magnetizing yoke with a magnetic flux concentration structure is proposed. And the thickness of the inner and outer teeth and the magnetization angle of the yoke were used as the design parameter. The results were analyzed through finite element analysis and the validity of this paper was verified.

A non-linear reluctance network, coupled with an electrical circuit, are used to study the impact of geomagnetically induced currents on three-phase power transformers. It’s about a meshed approach that allows to consider different topologies of transformers. This model is self-adapting and respects the distribution of magnetomotive forces and electromotive forces, even in the faulty state. The model is first defined in MatLab, then automatically implemented in EMTP. The behavior of the obtained model was compared to calculations carried out by finite elements. Finally, a study of the behavior of power transformers under the effect of geomagnetically induced currents was carried out.

Lighting phenomena are rather common for Photo-Voltaic Arrays, as these plants are usually deployed in open areas. A numerical assessment of the effects due to either direct hits or indirect strikes is very useful to preventively investigate possible damages to the structures or to electronic circuits connecting the array to the power grid. A full simulation using numerical methods typically adopted for electromagnetic fields computation would require huge amounts of computational power, due to different scales present in the model, ranging from the millimeter size of some circuital component to the kilometer scale of the lighting channel. Therefore, lumped models have been proposed to compute the overvoltage and overcurrent developing during lightning strikes. Such models comprise both actual electronic components (e.g., the solid-state devices for power conditioning) and equivalent elements, describing lighting to photovoltaic panels coupling, but also interactions between neighboring panels in the array. In this work, the computation of such inter-panel coupling

parameters is described, making use of a new approach combining semi-analytical expressions for stray mutual inductances and finite element models for stray capacitances.

In the context of electromagnetic compatibility ventilation panels with small quasi-periodic openings exist. The standard finite element method leads to a large number of unknowns for ventilation panels. An efficient approach to consider the multiple scales is introduced. Therefore, the multiscale finite element method is applied with special functions, in this work finite element solutions of a cell problem with support localised at the apertures computed by higher order. The proposed technique should be able to cope with varying thickness of the metallic sheet of the ventilation panels and arbitrary shapes of apertures as well as lossy materials and frequencies in a wide range. The introduced method improves the accuracy of coarse finite element simulations for ventilation panels in 3D.

This paper presents the neural network (NN), trained via electromagnetic field simulations, to identify existence of foreign metal objects in the wireless power transfer system which has magnetic cores and possible misalignment between the primary and secondary coils. The training data for NN consist of the frequency-dependent differential voltages in the detection coils together with the input voltage of the primary coil and existence of the foreign metal object. Although the coil misalignment and magnetic core induce confusing voltages, the trained NN is shown to have accuracy over 90% for the validation dataset.

The eddy current sensor using direct current (dc) biased magnetization has been developed, and sensor signals are investigated numerically considering hysteresis effects. This testing method can detect the opposite side defect (metal loss) through permeability measurement around the defect, which overcomes the limit of classical eddy current sensor due to the skin effect on the specimen. In this paper, we introduce a calculation method of sensor signals considering a high frequency ac magnetic field and a gradual change in a dc magnetic field. The entire calculations are divided into two sequential analysis in shortening the computation time. First, a time-dependent simulation calculates the dc magnetic field including magnetic hysteresis according to each sensor position. Then, the results of magnetic field are stored and used to calculate the incremental permeability based on the history-dependent hysteresis model. Then, the pre-calculated permeability is used for time-harmonic field analysis, and the sensor signal can be finally obtained. We implemented the calculation method to a prototype sensor and the effect of dc magnetic field and remanent magnetization on sensor signals can be effectively discovered.

A new design of a flow meter based on the principle of inductive measurement of the flow of carbonated beverages is presented. During operation, the flow, liquid homogeneity and chemical composition of the fluid often change and it is very difficult to design a sufficiently accurate flow meter. An advanced model of a nonlinear electromagnetic field and a flow field using online diagnostics to monitor the change in velocity and chemical composition is presented. This model working on the principle of a digital twin is then experimentally verified on a measuring stand.

The core of the presented work is the application of the quasi-stationary Darwin model, which is highly beneficial, compared to simulating the full set of Maxwell’s equations. This model includes resistive, inductive and capacitive effects but neglects wave propagation and enables to take the electric material parameter (permittivity) into account, which classical eddy current models cannot achieve. The finite element simulation of the Darwin model is validated by simulating the impedance of a standard coaxial cable, with well known behavior. Furthermore, the finite element results are compared to a circuit simulation with parameters from the datasheet, showing good accordance. This enables the application of the Darwin model to a wider variety of problems regarding electromagnetic compatibility, up to a frequency range where wave propagation does not need to be considered.

Computation of electric field induced by low frequency magnetic field in the human body usually requires to provide as source term a magnetic vector potential, which may be computed numerically.

In this case, the problem arise of projecting this potential on an anatomical phantom, which in our case is discretized by using a mesh composed of several millions of elements.

By using standard approaches, projecting directly the vector potential onto the mesh of the phantom requires solving a very large linear system.

We propose a solution which allows to mitigate the computational cost of the projection, with no penalization in term of accuracy.

This paper shows the useful combination of stochastic tools with 3D finite element analysis in order to build accurate predictor at a low computation cost in the case of human exposure for a high power wireless power transfer system. This surrogate model can be used to compute accurate sensitivity analysis regarding the various positions of the human body for the electromagnetic problem. Such an analysis enabled us to find the worst case scenario for the posture of the human body around the WPT3/Z3 system from SAE J2954.