QUADRATIC WEIGHING, APPLIED IN HARMONIZATION, A TOOL TO COMPARE THE PERFORMANCE OF TEST-FACILITIES

R. van den Brink, M. P. van der Beek, I. J. Landheer
Abstract:
Measurement-data of laboratories participating in comparisons are often treated equivalently with respect to operational conditions and statistical impact. When analysing the comparison, this may distort some conclusions because no attention is paid to deviations caused by differences in operational conditions (e.g. pressure). A better way to perform a comparison is to transpose the sets of flow-rate data to Reynolds numbers, to make it easier to see whether or not the points are in the same Reynolds-region and thus allow for a valid comparison.
Next, not all participating laboratories contribute equally to establish a Mean (Average) Value, because the laboratories with a low uncertainty must have a greater impact on the final result than laboratories of a lesser quality. So the data -sets must be weighed and to do so, weighing-factors are needed. Finally, a common way to analyse the comparison-results is by using Youden-plots.
This paper shows the "flow-rate to Reynolds-transformation" to compare test-data, and demonstrates which weighing-factors must be used to get the "Best Known Mean Value" with the smallest uncertainty. In the end, this method is also applied in Youden-plots.
Download:
IMEKO-TC9-2003-017.pdf
DOI:
-
Event details
IMEKO TC:
TC9
Event name:
FLOMEKO 2003
Title:
11th Conference on Flow Measurement
Place:
Groningen, NETHERLANDS
Time:
12 May 2003 - 14 May 2003