Effects of the uncertainty contributions on the methods used for measurement uncertainty evaluation

Abstract

Adequate measurement uncertainty evaluation is crucial for supporting basic measurement purposes. However, the most prevalent approach, the uncertainty propagation, may not be validly applicable under certain conditions which require the use of less restricted alternative method (MCM-based propagation of distributions). We demonstrate the effects of major conditions, e.g. non-linear measurement model, non-Gaussian input quantities, on the reliability of uncertainty evaluation methods, highlighting the importance of the less frequently examined impact of uncertainty component contributions to the standard uncertainty of measurand.