Data science for the SI

We are improving the traceability of data for realising and defining the SI units, and the international comparison of data.

Measurement traceability, which underpins international trade and our quality of life, requires an uncertainty to be associated with all measurements. The traditional approach to measurement uncertainty evaluation is stated in the Guide to the Expression of Uncertainty in Measurement (GUM) published by the BIPM Joint Committee for Guides in Metrology (JCGM). This approach assumes a linear model, or a model that can be linearised. Two accompanying JCGM guides extend the approach to multivariate measurement and to the use of a Monte Carlo method, which applied to non-linear measurement models in general.

Our research provides innovative solutions to many challenges in the area. In demanding areas such as energy, environment and healthcare we are using domain-specific knowledge to improve the models. Development of Bayesian methods delivers reliable results for these challenging problems, and we are continuing to extend this research as new measurement modalities are developed. We support the JCGM and we are playing a key role in the expansion of the GUM suite of documents.

Our work includes the following themes:

• Pre-normative research by contributing worked uncertainty evaluation examples, using GUM, Monte Carlo and Bayesian methods, to many ISO and CEN committees, Eurachem and the JCGM
• Development of new documentary standards such as for polynomial calibration, conformity assessment, and experimental design for the evaluation of uncertainty
• Development of new JCGM guides such as on the development and use of measurement models, and for evaluating inter laboratory comparison data
• Applications of our research to many areas in physics, chemistry, biology, engineering and technology

Contact us