COMPARISON BETWEEN GUM, MONTE CARLO AND BAYESIAN UNCERTAINTY EVALUATION APPROACHES FOR AN AREAL SURFACE TEXTURE MEASURING INSTRUMENT

摘要

The National Physical Laboratory (NPL) has recently developed a traceable areal surface texture measuring instrument. This paper compares three methods of estimating the uncertainties associated with NPL's Areal Instrument. The first is the conventional GUM approach, the second is the Monte Carlo approach and the third is the Bayesian. It is shown that the three methods differ only when the uncertainties arising from alignment influence factors are significant.