VALUABLE ADVICES ON COMPARISON RULES FOR UNAMBIGUOUS EVALUATION OF MEASUREMENT RESULTS

F. Härtig, K. Kniel

Abstract

International comparison measurements are the backbone for validating the competence of metrology institutes among one another. An informative interpretation of the results requires, however, intensive studies of the individual reports and great expertise of the readers. This is due to the lack of a clear procedure for the assessment of such comparison measurements which leads to different values which are difficult to compare. The selection of the reference value formation is, for example, open. Likewise, an equivalence value, the normalized error value (<i>E</i>_n value) – which is usual for comparisons – can be calculated on the basis of the standard measurement uncertainty or on the basis of the expanded measurement uncertainty. The third degree of freedom relates to the elimination of outliers which is handled in different ways. A decision of how a participant contributes to the reference value, or whether his measurement results are regarded as comparable, may – depending on the calculation carried out – one time turn out in his favour and another time to his disadvantage. Studies of international comparison measurements from the field of length show – as examples – that the different procedures applied in comparison measurements are commonplace. In addition to the usual assessment possibilities, examples are shown which clearly show the problems. With the example of gear measurement, a proposal is presented of how comparison measurements should in future be carried out clearly and transparently.

Keywords

intercomparison, evaluation rules, precision metrology, comparability, En-value, normalized error

Download

IMEKO-WC-2012-TC21-O11.pdf

DOI

-

IMEKO TC

TC21 - Mathematical Tools for Measurements

Back to the proceedings