Logo PTB

Measurement Uncertainty

Philip M. Fleischmann

„Sicher is‘, dass nix sicher is‘, drum bin ich vorsichtshalber  misstrauisch.“  “It is certain that nothing is certain, so I‘d rather be sceptical.“  (Karl Valentin)

It is a horrible idea to assume that measurements might be uncertain. Even more so in the case of calibration laboratories! But unfortunately, even the most accurate method has the unpleasant characteristic that its measurement uncertainty cannot be zero. And that is exactly what the Technical Committee “Measurement Uncertainty“ deals with.

In addition to the exchange of information in this area, the balancing act between a practice-oriented handling and a mathematically correct formulation has to be managed: “root sum of squares“, “correlation coefficients“ and “Monte Carlo simulation“ are the names of the tools. But since the Technical Committee consists of participants from all disciplines, it is often very difficult to agree on general methods in order to find the best solution for a problem. But perhaps it is in the nature of things that only “approximately accurate“ information can be provided? Currently, the Technical Committee deals with creating a checklist for the validation of measurement uncertainty budgets: a sort of cookery book  which classifies, summarizes and thoroughly explains the “ingredients“ for the most common types of measurement uncertainty calculations from measurement tasks through to process and model measurement, taking into account normative references and basic principles.

Thus, the users in the calibration laboratories will be provided with a viable tool which enables them to create process descriptions including the respective measurement uncertainties or to validate them, fulfilling all the essential requirements of the DIN EN ISO/ IEC 17025 and the GUM. Moreover, recommendations for the realization of interlaboratory comparisons shall soon be developed and handed over to the sector committee of the DAkkS. Besides, another guide for the determination of calibration intervals with regard to the calculation of measurement uncertainty is planned and shall be discussed during the next meetings. 

The Technical Committee has recently published the following guide:  • DKD-L 5: Guide to the practical determination of measurement uncertainty – basics and procedures for simple models without correlation