Evaluation of regional key comparison data: two approaches for data processing

Abstract

This paper deals with the problem of establishing degrees of equivalence for national measurement standards participating in regional key comparisons (RMO KC), with respect to a key comparison reference value (KCRV), using measurement results obtained by linking national metrology institutes (NMIs). Two types of key comparisons are considered, which differ in applying either a material measure or a measuring instrument as a travelling measurement standard.The results obtained by linking NMIs are used to determine transformation procedures for measurement results of the RMO KC so that the transformed results can be compared directly with CIPM KC results.Formulae are given for evaluating the degrees of equivalence of the national measurement standards of laboratories participating in the RMO KC, with respect to the KCRV, taking into account possible correlations between measurement results.

Linking of a RMO key comparison to a related CIPM key comparison using the degrees of equivalence of the linking laboratories

“…We show that for single stable travelling standards in the two key comparisons the proposed proceedure is in line with the previously proposed linking procedures [2,5,8]. Just as in [2,5,8], linking is in this case established by employing an additive shift to the measured quantity values of the RMO key comparison before relating them to the CIPM KCRV. The shift is determined as a weighted mean of differences of the measured quantity values of the linking laboratories in the two key comparisons.…”

“…, L, is valid. Under assumption (A) the x α − y α are meaningful estimates of X − Y , and they can be employed for linking as proposed in [2,5,8]. Assumption (A) is also the assumption underlying the method proposed in this paper.…”

“…To establish the according degrees of equivalence (DOEs) 3 for the RMO laboratories several linking procedures have been proposed in recent years, cf e.g. [2][3][4][5][6][7][8]. Yet, this task is still a topic of current debate.…”

“…For instance, in [6,7] constrained least-squares estimation on the basis of models that constitute an extension of (1) and ( 2) is suggested for linking, using the assumption that the linking laboratories' biases for the RMO key comparison equal their estimates obtained from an analysis of the CIPM key comparison data. In the context of the methods [2,5,8] the difference X − Y is estimated by a weighted mean of the differences x α − y α = (X − Y ) + (δx α − δy α ), α = 1, . .…”

Linking the results of CIPM and RMO key comparisons with linear trends

“…As pointed out in [ 3 ], however, this approach will generate a completely new analysis, which obviously will influence the existing results. Kharitonov and Chunovkina [ 6 ] and Decker et al [ 7 ] have also discussed linking of CIPM and RMO key comparisons.…”