Hybrid Simulation of Electromagnetic Field Interaction with Metallic Structures Showing Massive Nonlinear Loading
Final Report Abstract
In the course of the project a method was investigated and a complete tool chain was implemented for the hybrid simulation of nonlinearly loaded linear structures. The main advantages of the hybrid method is that the computationally expensive full-wave coupling dictated by the structure of the ports is modeled only once upfront. Once present, the same model can be used to compute the response of the 3D structure for variable excitation waveforms and terminations. This method outperforms traditional full-wave solvers based in the time domain when investigating many different excitations and terminations with a fixed structure. This advantage could be used to its full extent for finding optimization guidelines for nonlinear shielding against high intensity radiation fields (HIRFs). The method was validated in different applications besides nonlinear shielding such as distribued ESD protection and reconfigurable meta-surfaces. Challenges of the hybrid method were identified as well, most notably a strong dependence of the computational effort with respect to the number of ports as well as the number of resonances of the structure. Future work will try to address these challenges but it is clear that full-wave solvers will likely remain the tool of choice for structures with a numer of ports starting at several thousands and up. In the course of this project it became clear (after many investigations) that the problem required a very accurate model of the passive structure. Small errors introduced by the adapted Model Order Reduction technique when fitting the scattering parameter frequency data for specific reference resistance are likely to be amplified when renormalizing the model to different reference resistance, which is a required step in the proposed iteration-dependent decoupling framework. Specifically, this means that the reduced order model is only guaranteed to be accurate for terminations which are similar to the reference resistance. If the terminations differ from the reference by two or more orders of magnitude the response of the model can be significantly different from the response of a model built on frequency data referenced to resistances similar to the actual terminations. This does not pose a problem if the scattering data can be renormalized to a reference which is predictably in the range of the terminations which are to be connected to the model in future investigations. For the applications investigated in the course of this project, this was not the case as the terminations exhibited strongly nonlinear behavior with the effective differential resistances of e.g. anti-parallel diodes ranging from 1 Ohm to 1 MOhm. This is a solver independent problem, as the solver will produce accurate solutions for the model only and does not have access to the original frequency data. These problems became particularly clear when using solver schemes which use models with multiple references like the iteration dependent decoupling.
Publications
-
“Numerical Complexity Study of Solving Hybrid Multiport Field-Circuit Problems for Diode Grids,” 2019 International Conference on Electromagnetics in Advanced Applications (ICEAA), Granada, Spain, 2019
T. Wendt, C. Yang, C. Schuster and S. Grivet-Talocia
-
“A Macromodeling-Based Hybrid Method for the Computation of Transient Electromagnetic Fields Scattered by Nonlinearly Loaded Metal Structures,” IEEE Transactions on Electromagnetic Compatibility, vol. 62, no. 4, pp. 1098-1110, Aug. 2020
T. Wendt, C. Yang, H. D. Brüns, S. Grivet-Talocia and C. Schuster
-
“A Multi-Stage Adaptive Sampling Scheme for Passivity Characterization of Large-Scale Macromodels,” in IEEE Transactions on Components, Packaging and Manufacturing Technology
M. De Stefano, S. Grivet-Talocia, T. Wendt, C. Yang and C. Schuster
