Search

A method is presented for the optimal choice of the fixed-point reluctivity in nonlinear eddy current problems solved for steady-state in the frequency and the time domain. The optimal selection is shown to ensure a minimal contraction number for the fixed-point method. The proposed technique is compared to earlier approaches using a 2-D example. Index Terms--Convergence of numerical methods, eddy currents, Fourier transforms, numerical analysis.

Abstract: Two different techniques for the analysis of nonlinear, periodic eddy‐current problems are compared using a three‐dimensional benchmark problem. The methods are the parallel time periodic finite element method and the time periodic fixed‐point technique. [ABSTRACT FROM AUTHOR]

In ultra-high-frequency radio-frequency-identification (UHF-RFID) the conjugate complex matching of the packaged UHF-RFID transponder IC and the tag antenna structure is important to increase the reading range of the transponder tag. Quite often half-wavelength dipole-like antenna structures are used. Due to the fact that the impedances of the packaged IC and the dipole antenna structure are in different domains, a matching structure is needed to accomplish the conjugate complex matching. A common technique to carry out the matching is the so called T-match. Since the nominal input impedance of the IC depends on the applied antenna voltage (and in a strongly non-linear way), the backscatter abilities of the tag are also functions of the applied voltage. In the present work those dependencies are investigated in terms of full-wave finite element simulations and measurements. [ABSTRACT FROM PUBLISHER]

A method is presented for obtaining the steady-state solution of three-dimensional eddy current problems with nonlinear material properties and periodic excitations using the T,\Phi-\Phi formulation. The problem is formulated in the time-domain and the application of a fixed-point approach combined with a discrete Fourier-transformation technique ensures that, during one nonlinear iteration step, there is no coupling between the time-steps within one period. The steady-state result is hence obtained by solving half as many time steps as their number in one period. Since the time steps are decoupled, their computation can be readily parallelized resulting in a fast method. Comparison with results obtained by a transient time-stepping technique is presented. [ABSTRACT FROM PUBLISHER]