Crosstalk analysis of printed circuits with many uncertain parameters using sparse polynomial chaos metamodels

摘要

This paper presents a metamodel based on the sparse polynomial chaos approach, well adapted to high dimensional uncertainty quantification problems, applied for the analysis of crosstalk in printed circuit board microstrip traces. It enables to estimate, with a low computational cost compared to Monte Carlo (MC) simulation, statistical quantities and provides a sensitivity analysis of the crosstalk effects considering numerous uncertain variables. The approach is validated against MC simulation and shows a good efficiency and accuracy.