Conference Agenda

The multiscale problem of computing ohmic losses in steel reinforced concrete slabs protecting HVAC cables is solved with an integral equation approach. The approximations at the basis of the proposed method guarantee accurate results when compared with FEM reference solutions for a small test case. Losses are then computed for a realistic structure.

This work presents a procedural method for analytically computing losses in solid conductor air-core coils. The approach was developed for the coaxial subset of air-core coil systems in which hoop-turn, axisymmetric assumptions can be made. It encapsulates loss generally associated with magnetic diffusion, including that induced by rotating magnetic fields from multi-phase current sources, without computing the complex eddy current distribution in a conductor cross-section. Each turn in the system is replaced with an infinitesimally thick circular current loop and is independently excited according to a peak current and phase. Field contributions are summed up at each loop center, with the inclusion of phase to calculate the major and minor components of the resultant magnetic field at each turn. Well established analytic formulae are then used to compute loss accordingly. The approach is significantly faster than 2D axisymmetric FEA and greatly improves approximations for losses in isolated coils.

The Cauer ladder network (CLN) method enables to construct a reduced model of an eddy current based on an equivalent electrical circuit. A method is proposed in order to derive a parametric CLN from only one construction of a CLN for a given set of parameters. The proposed approach is tested on a 3D benchmark with the conductivity and permeability as parameters.

A coupled FEM-Fourier formulation is proposed to compute the ac winding losses in MFTs with foil windings. The proposed system of equations utilizes FEM for the conductive cross section area of winding window, to acquire the flexibility and precision of FEM on modeling the induced current density distribution. The Fourier method is assigned to estimate the magnetic field distribution on the non-conductive domains, due the simplicity. This work investigates the effect of such coupling on computational efficiency of ac loss modelings in MFTs.

The quench simulation of superconducting accelerator magnets imposes a nonlinear multi-physical multi-scale problem, for which a fully coupled 3D simulation is by far out of reach. This work proposes a numerical method, which combines 2D finite-elements and 1D spectral-elements into a hybrid quasi-3D method, which is able to resolve the 3D physics in the magnet, while achieving a superior computational efficiency compared to conventional 3D simulation. For the first time, the quench development in a superconducting coil model is computed in a magneto-thermal nonlinear transient simulation using this approach.

The paper deals with the optimization of the air gap field distributions for inverter-fed high speed synchronous machines. The task of the determination of the magnetic pole contour of the machine or its magnetization distribution has been reduced to the solution of an over-determined, non-linear, and ill-conditioned set of algebraic equations. This methodology has been implemented into the finite element method. The resulting algorithm was modified to be able to take into account the displacement of the mesh nodes without destroying the mesh structure. The application of the sparse matrix topology allowed for quick iterative determination of required solutions without changing the structure of finite element matrices.

Applying Green's theorem to Maxwell’s equation in the frequency-domain, we get boundary integral equations (BIEs) for the eddy current analysis. The original BIEs are composed of surface electric and magnetic currents denoted as Js, Ks. Introducing loop electric current Jl for Js on the surface element ΔS, and constant Kc for Ks on ΔS, we get new BIEs composed of Jl and Kc. These variables possess beneficial characteristics for the analysis, i.e., Jl guarantees that the eddy current doesn’t leak to the air, and guarantees that the total magnetic charges are zero. In the nonlinear eddy current analysis, as the BEM is based on superposition principle, the nonlinear analysis is mathematically rejected. Fourier series expansion can solve the difficulty as follows. The magnetic flux density (B) on the air-conductor interface can be given according to the B-H curve, and a nonlinear permeability (μn) is derived from the fundamental B. Improving μn, we carry out the nonlinear eddy current analysis. Using the improved μn, fundamental and harmonics of eddy current are analyzed.

Complementarity for static electromagnetic problems has the peculiarity to provide upper and lower energetic bounds depending

on the adopted formulation. For instance, the presence of these bounds is not questionable if electrostatic problems are considered,

for which theoretical foundation justify this property. Here we investigate whether these bounds are also present in electroquasistatic

hypothesis.

This paper presents a fast 2.5D loss calculation method for round Litz wires. Losses due to excitation currents and external magnetic fields are calculated independently. Litz wire is sliced into several sections per pitch. Each section’s calculation is based on the solutions of the magnetic vector potential, which is obtained from solving Maxwell equations under two-dimensional (2D) quasi magneto-statics. The proposed method is compared with numerical simulations. The results show the proposed method have good accuracy and fast computational speed.

Characterization of red blood cells is important for biological, biomedical and bioengineering applications. Impedance spectroscopy in a microchannel can be used to examine single-cell electrical properties. In this work, we employ 3D numerical calculation to investigate the conductance measured between a pair of microelectrodes fabricated on the bottom of a microchannel. Electrical properties of a red blood cell contribute to the difference in conductance when the cell is present in the electrode gap. The effects of the cell orientation and microchannel geometries on the measurement are analyzed where the red blood cell is modeled as a biconcave disc to reflect its actual profile. The numerical results give the conductance as a function of electrical frequency. The orientation of the cell in the sensing gap affects the conductance-frequency characteristic. The microchannel geometries also have a considerable effect on the measurement result. We investigate the variation of the conductance values with the position of cell in the sensing gap.

In this study, we established the surface potential decay experiment to analyze the trap characteristics of a representative polymeric insulator, cross-linked polyethylene. Surface potential decay measurement is the representative process among the various methods to deduce the trap characteristics. By employing this method, trap characteristics for electrons and holes can be evaluated separately. In this paper, we introduced the detailed experimental setup to measure the surface potential. The experimental results were analyzed based on the iso-thermal relaxation theory. The bipolar charge transport model was used to describe the surface potential decay mechanism with induced deep trap characteristics from experimental results. The established numerical analysis results were compared with experimental results to verify. The numerical analysis model well explained the tendency of surface charge decay phenomena for the initial voltage and thickness of the sample. Furthermore, we analyzed the effects of various parameters related to the characteristics.

In some electrical machines, normal flux and pulse width modulation (PWM) can have significant impact on the loss of laminated steels, e.g. stator end-core loss of converter-fed synchronous generators. This paper presents a method for computing the loss of laminated steels with impacts of normal flux and PWM switching considered. A dedicated lab test setup is developed to emulate normal field on laminations of converter-fed electrical machines and validate the loss model. Compared to measured loss with sinusoidal voltage supply, the largest error of the loss model is -8.8%. Computation shows that the core loss with PWM voltage supply is 1.35-6.42 times the loss with sinusoidal voltage supply. Measurement under PWM voltage supply is ongoing, and more results are to be provided in the final paper.