A fast and parallel stroud-based stochastic collocation method for statistical EMI/EMC analysis

• Published in: 2008 IEEE International Symposium on Electromagnetic Compatibility pp. 1 - 5
• Main Authors: Bagci, H., Yavuz, C., Yucel, A.C., Hesthaven, J.S., Michielssen, E.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.08.2008
• Subjects: Artificial neural networks; Couplings; Electromagnetic compatibility; electromagnetic coupling; Electromagnetic interference; electromagnetic interference and compatibility; Equations; fast solvers; hybrid solvers; Mathematical model; parallelization; stochastic collocation; Stroud integration rules; time domain integral equations; Transmission line matrix methods
• ISBN: 9781424416998; 142441699X
• ISSN: 2158-110X
• DOI: 10.1109/ISEMC.2008.4652151

A fast and parallel Stroud-based stochastic collocation method for statistically characterizing electromagnetic interference and compatibility (EMI/EMC) phenomena on loaded multiscale platforms with uncertain system configurations and subject to variable electromagnetic excitations is described. The proposed method uses a previously developed hybrid time domain integral equation based field-cable-circuit to carry out deterministic EMI/EMC simulations permitting the statistical characterization of pertinent observables. The number of simulations required by the proposed method is far fewer than those needed by Monte-Carlo methods. The proposed method is used to characterize cable-induced coupling onto PC cards located in shielding enclosures. Both the hybrid simulator and the stochastic collocation code execute with near-full efficiency on distributed memory clusters.