Conference Agenda

In this paper, the distribution of electromagnetic and temperature from the two-dimensional (2-D) finite element analysis of the simplified transformer is estimated by the convolutional neural network (CNN) model U-net. By changing the geometric dimension, material property and current excitation, the image dataset with different parameters is obtained and additionally expanded by the operation of scaling, flipping and rotating. Based on it, the weight parameters of CNN model are trained and the optimization for the model is carried out by hyperparameters searching, which improves the estimation precision of the electromagnetic and temperature field distribution of transformer. The estimation results prove the effectiveness of the method.

This paper presents a nonlinear analytical model (NAM) with the direct coupling of Maxwell’s equations and magnetic equivalent circuits (MECs) for the prediction of the electromagnetic field distribution for the dual-rotor segmented-stator axial flux permanent magnet (DRSSAFPM) machine. The local saturation of the stator core can be modeled with a mesh-based MEC. The key to coupling is the effective connection between the air-gap magnetic density calculated by the Maxwell’s equations and the corresponding branches of the MECs. It is shown that the NAM allows to combine the advantages of analytical and MECs modelling and agrees with the results of finite element method (FEM).

This paper proposes an analytical model for the prediction of flux fringing effect for surface mounted permanent magnet (SMPM) machine. In this paper, the main assumption, that is the end effect is independent of the main flux, is made firstly. Consequently, the calculation model is built and the sub-domain technique is adopted to solve this model. The end effect for surface permanent magnet machines are obtained accordingly based on Schwarz-Christoffel mapping and Maxwell’s equations. The results show that the proposed approach agrees with the 3D finite element (FE) model, which also verify the main assumption. The model is validated by experiments, showing good agreements. Main contribution of the work is to model the end-effect, which can improve the analysis accuracy of both radial flux permanent magnet (RFPM) machine and axial flux permanent magnet (AFPM) machine in fast analytical calculation.

In this paper, a novel magnetic field calculation method for considering the static eccentricity of axial flux permanent magnet motor (AFPMM) rotor is proposed. The method in this paper is based on the principle of permanent magnet equivalent, which is simple and universal applicability. Analytical expressions for the no-load magnetic field distribution of AFPMM with static rotor eccentricities are presented in full paper. The analytical method was validated through finite element method and experimental results to show the effectiveness of the analytical technique proposed in the paper. The main contribution of this paper is presenting a simple and accurate approach for permanent magnet machine, not limit to AFPMM.

A hybrid model containing offline and online calculation procedures, for synchronous reluctance motor is presented in this paper. Via this modeling procedure, the initial design parameters of a synchronous reluctance motor can be obtained iteratively, with only a fraction of time involved as compared to the commercial finite-element-method based design software. This modeling procedure involves firstly computing the machine’s circuit parameters by running a MATLAB-script and storing them in a 3D-lookup table. The performance parameters are then obtained by solving the electromagnetic coupling model, in the online environment of MATLAB-SIMULINK.

This paper presents a novel air-cored linear-rotary induction machine (LRIM) in which its columnar movable armature surfaced with the conductor. The analytical magnetic distribution in the conductor and air gap region of the LRIM is predicted by employing a group of H-formulations in the conductor region and Laplacian equations with magnetic scalar potential in others in the Cartesian coordinate. The coils are equivalent to several permanent magnets. The analytical magnetic field distributions are derived and verified by the numerical computational results using 3D finite element method.

In the paper a Finite Element (FE) analysis for investigating the electric properties of a Wireless Power Transfer System (WPTS) devoted to charge the batteries of electric vehicles is performed. In particular, the dynamic-WPTS, which is particularly challenging because of the position-varying properties of the system, is considered. The field analysis is computationally heavy because of thin conductive layers modelling the car chassis: an effective analytical approximation for the field calculation in thin layers is applied to both car frame bottom and the shielding aluminum layer. This approach allows for an accurate solution and, meanwhile, for a reduction of the computational costs, making the repeated simulations feasible.

This thesis is about the design of the electronic brake used in the wafer transfer device during the semiconductor process. In the case of a robot for transferring wafers, which is a basic material for semiconductor production, 5 or more independent axes are required, and an electromagnetic brake is required for operation and stopping. The electromagnetic brake generates a stopping force through the attraction force by the magnetic flux generated from the permanent magnet located in the stator, and is driven in such a way that the magnetic flux of the permanent magnet is offset through the application of current during operation. In this paper, for the design of the electronic brake, the friction coefficient of the boundary where the stopping force occurs is calculated and the target suction force is derived through the size of the inner and outer diameters of the brake. In addition, in consideration of the current density, the winding design is carried out to offset the magnetic flux of the permanent magnet to drive the device. In addition, by arranging slits in the stator and rotor, magnetic flux changes according to rotation were induced, thereby generating iron loss. The stopping power of the brake was improved by using the iron loss as the stopping power when the brake is operating. At the same time, the optimal model was derived by optimizing the number and shape of the slits through the optimization technique. Finally, the feasibility of the design method in this paper is verified through fabrication and testing.

In this paper, the harmonic characteristics according to the combination of the number of poles and the number of slots of an external type synchronous generator for drones were compared. Four models of abductor generators with fractional slots were selected to analyze and analyze the harmonic content of the no-load back EMF. Through comparison using FEM, one pole number slot combination with the best THD characteristics was selected among the four models, and the response surface method was applied to the selected model to design an optimal design for further harmonic reduction. The reliability of the optimal design result was verified by comparing the optimal design result with the FEM result of the basic model.

Solid-state transformers (SST) are a promising technology that allows decreasing the volume and weight of the magnetic device due to its high operating frequency. SST are used together with power electronic converters based on switching devices for bidirectionally converting between the involved low- and high- frequency signals. Unlike SST, the conventional transformers are large due to their low operating frequency 50/60 Hz. Nevertheless, the high-frequency transformer (HFT) needs to be accurately designed so that it has a reliable power transfer. An accurate knowledge of the leakage inductance is one of the most important transformer parameters because it defines the power transfer capability from each side of the HFT. This digest presents an optimal design of the HF transformer, by successfully integrating finite element analysis (FEA), design of experiments, response surface modeling and multiobjective optimization.

The insulation of power transformers needs to be appropriately designed to withstand the voltages under normal o transient conditions.

Paper and pressboard are the most used materials in oil-filled transformers. Both insulating materials are made of cellulose. Another

insulation component used is oil, which works as an insulator and cooling fluid. Due to the complex geometry of the solid insulating

materials, it is advisable to use numerical methods to compute the electric scalar potential and, hence, determine the stress gradient and

assess the transformer's scalar electric potential. However, insulation design rules are found only with the manufacturers, and it is not

easy to access them. This digest proposes developing a knowledge base to facilitate the design and assessment of the transformer

insulation before it is built. The knowledge-based system (KBS) will gather numerical information of the electrostatic fields obtained by

using FEA. The proposed KBS is developed by applying it to a 34.5 kV power transformer.

Electric vehicle traction drive motors are a major application area for electromagnetic simulations. More attention is being paid to complete drivetrain analysis using multi-physics simulations. Multi-body dynamic mechanical simulations rely on export of accurate electromagnetic force and torque values. The authors demonstrate methods that improve accuracy. The benefit of determining the effect of geometric deformations by repositioning mesh nodes rather than meshing a new geometry is shown. Errors caused by different discretization in each tooth are calculated using averaging algorithms for concentric configurations and used to reduce discretization errors in deformed geometry results. Improvement is also obtained by minimizing cancellation errors.

A homogeneous and low residual environmental magnetic field is required for the proper operation of most optically pumped magnetometers (OPMs). This is achieved using a combination of passive and active magnetic shielding. Passive magnetic shielding often uses multiple layers of highly permeable materials. The effect of a one-layer highly permeable wall on the distribution of the magnetic field inside a shielded room can be conveniently studied using the method of images. We use the method of images for the case of a realistic two-layer shielded room, approximated in two dimensions, and compare the results with a finite element method solution. We found that mirroring level 10 ensures a normalized root mean square deviation of the magnetic flux density <0.05 % in a circular area placed centrally in the shielded room with a radius of 29 cm which is desired from the point of Magnetoencephalography using OPMs. The achieved accuracy of the method of images makes it suitable for the optimization of active shielding coils.

Anatomical realistic voxel models of human being are commonly used in numerical dosimetry to evaluate human exposure to low-frequency (LF) electromagnetic fields. The downside of these models is that they are unable to correctly reproduce the boundaries of curved surfaces. These stair-casing approximation errors introduce computational artifacts in the evaluation of the induced electric field and the use of post-processing methods is essential to try to eliminate these errors. This paper, using a suitable exposure scenario, shows that tetrahedral meshes makes it possible remove stair-casing errors. However, in real exposure scenarios, other sources of artifacts are still present and must be filtered out with the known techniques.

Low-frequency (LF) electromagnetic fields generate induced currents in the tissues of the human body, but their direct evaluation, let alone measurements, is difficult or more often impossible. For this reason numerical dosimetry is fundamental for the evaluation of human exposure to LF electromagnetic fields. In the classical approach all dosimetric computations are performed with a postured human phantom, but there are situations (e.g. on the workplace) in which the position of the human body is known only approximately and the source of the electromagnetic field must be characterized in realistic conditions. In this paper, a new approach based on the evaluation of human exposure to electromagnetic fields by using a non-postured human model through a source term transformation is validated on a realistic scenario: a whole-body female human phantom with postured arms exposed to LF magnetic field.

Evoked compound action potential (ECAP) represents the synchronous response of a population of nerve fibers excited by the electrical stimulation. ECAP acts as a clinical indicator for an objective assessment of a compound neural response. This paper uses a cochlear implant (CI) model to simulate ECAP based on the volume conduction method. ECAPs obtained using alternating polarity recorded by different sensing electrodes were compared. Besides, we also examine the relationship between ECAP amplitude and given stimulation current. Preliminary results show that the ECAP amplitude becomes larger when the stimulation current increases, which is consistent with typical clinical outcomes. The simulated ECAP of the CI model could be used as a theoretical reference for patient-specific cochlear implant studies for objective programming for infant CI patients.

Wireless power transfer systems can be used to charge batteries of electrically powered cars. These systems emanate a lowfrequency magnetic field exposing humans in the vicinity of the car and inducing electric fields inside the body. The body-internal electric fields should not exceed limits given by the International Commission on Non-Ionizing Radiation Protection. In this paper these electric fields are to be calculated from in-situ measurements of the magnetic flux density in near real-time with the so-called Co-Measurement Scalar Potential Finite Difference scheme. A method is presented to determine the magnetic flux density in the area of an exposed human from few measurement points with an optimized assembly of the discrete Poisson system of linear equations of the Scalar Potential Finite Difference scheme to calculate the body-internal electric fields in near real-time.

The problem of the electromagnetic forces modeling is one of the most important and complicated topics in the electromagnetic field theory. Especially interesting is the interaction between magnetized objects and external magnetic field. Our experience in teaching students of technical specializations shows that these topics are especially difficult for understanding. We developed an experimental setup used now in laboratory work for students. The work is included in the general course of electromagnetic field theory given for students specialized in power electro engineering in Peter the Great St. Petersburg Polytechnic University. Simple research work together with consequent analysis of the experimental results help to understand better the nature and main properties of the mechanical interaction between magnetized objects.

Electromagnetic devices have induced voltages due to the movement of magnetic fields, and the electrical currents generate forces that cause vibrations and deformations. The above are the principal source of equipment malfunction or the need for corrective maintenance in the moving parts of electromagnetic machines. The coupling between an electromagnetic device's magnetic and structural fields is studied in two dimensions in this digest. The partial differential equations are solved with the finite element method. The modeling is made by using FEniCS in the programming language Python, and then the problem is validated with a commercial software .