COMPUTER AIDED EVALUATION OF MEASUREMENT UNCERTAINTY BY CONVOLUTION OF PROBABILITY DISTRIBUTIONS

Giovanni B. Rossi, Francesco Crenna, Michele Codda

Abstract

The evaluation of measurement uncertainty, from the final user’s standpoint, involves several issues which are only partially addressed by the GUM. An evaluation procedure assisted by a computer program can be of great support and provide the user with detailed results useful for further analysis. In this paper a code is presented for the assisted evaluation of measurement uncertainty, based upon the convolution of probability distributions. So the method allows expression of the final result of measurement as a probability distribution, on which it is possible to evaluate useful parameters such as expanded uncertainty, at whatever confidence level. It also permits to evaluate, in probabilistic terms, the risk associated to matching tolerances. Moreover it is possible to manage the uncertainty on dispersion parameters (sometimes called ‘uncertainty of uncertainty’) considering a modified probability distribution. In the paper the code is presented together with some case studies showing the support available to the operator, the GUM compatibility and the application importance of the final probability distribution.

Keywords

Measurement uncertainty, Probability distributions, Computer-assisted uncertainty evaluation

Download

PWC-2003-TC8-012.pdf

DOI

-

IMEKO TC

TC8 - Traceability in Metrology

Back to the proceedings