Credibility evaluation of uncertainty analysis results of EMC simulation

As an important part of credibility evaluation, uncertainty analysis has been taken into consideration in Electromagnetic Compatibility (EMC) simulation. Frequently, the quantitative validation of the uncertainty analysis results is performed by comparing a group of results of Mont Carlo Method (MCM) and reliable references. In this case, the multiple data sets comparison method needs to be developed. In this paper, a method named FSV-kmedoids is proposed to solve this problem, which is an effective combination of Feature Selective Validation (FSV) and a kind of clustering method, k-Medoids. Generally, FSV method gives the quantitative description of discrepancy between data sets, and k-Medoids clusters elements in metric spaces. So the k-Medoids is used to reduce the size of MCM results and the FSV is applied to assess the credibility of the reduced results. Consequently, the proposed method provides quantitative assessment for the credibility of uncertainty analysis of EMC simulation.