Measurement uncertainty propagation through the Feature Selective Validation method

The Feature Selective Validation (FSV) is the standard method used for validation assessment in Computational Electromagnetics, and it uses both quantitative and qualitative indicators to measure de similarity between a pair of data sets. However, standardized FSV rely on a heuristic procedure for graphical comparison that does not include considerations about the uncertainty of the data sets involved. The reliability of the validation results, and therefore of the model under validation, depends on the uncertainty of the data sets used as input for the FSV, even more considering that some measurements associated to electromagnetic compatibility tests are characterized by a large uncertainty. Nonetheless, the FSV algorithm makes the propagation of such uncertainties a difficult and cumbersome task through the conventional approaches. This paper presents the application of the Monte Carlo Method as an approach to propagate the uncertainty of the input data sets in order to estimate a confidence interval for each FSV indicator. Finally, a numerical example is presented and discussed.