Analysis of Key Comparison Data: Unstable Travelling Standards
- 43 Downloads
A procedure for the analysis of key comparison data is proposed in the case where the travelling standard shows a deterministic or random drift. It is assumed that some parametric model such as a linear trend is appropriate for modelling the deterministic part of the drift. The random part of the drift is assumed to be adequately described by a random walk model. The chosen drift model is determined by least squares and a model check is carried out. If the model check is satisfied, the derived drift model is applied to determine degrees of equivalence. A star-type sequence is required for the comparison in order to check the adequacy of the random drift model. The procedure is illustrated by its application to simulated data.
Key wordskey comparison stability random walk least squares
Unable to display preview. Download preview PDF.
- 1.“Mutual recognition of national measurement standards and of calibration and measurement certificates issued by national metrology institutes,” in: Technical Report, International Committee for Weights and Measures (1999).Google Scholar
- 2.Guide to the Expression of Uncertainty in Measurement, 2nd ed., Geneva, International Organization for Standardization (1995).Google Scholar
- 3.M. G. Cox, “The evaluation of key comparison data,” Metrolog., 39, 589–595 (2002).Google Scholar
- 4.M. Desenfant, “Drift problems with the circulated standard during comparisons,” in: Proceedings of the International Workshop on Data Analysis of Key Comparisons, D. Richter, W. Woger, and W. Haesselbarth (eds.), PTB Bericht IT-10, 55–62 (2003).Google Scholar
- 5.L. Nielsen, “Evaluation of the calibration history of a measurement standard,” in: Proceedings of the International Workshop on Data Analysis of Key Comparisons, D. Richter, W. Woger, and W. Haesselbarth (eds.), PTB Bericht IT-10, 135–140 (2003).Google Scholar
- 6.R. B. Frenkel, “Comparison of artefact standards by more than two laboratories,” Metrolog., 39, 21–29 (2002).Google Scholar