Temperature measurement. Influence of the conditions of use on uncertainties

When a measurement is made, it is necessarily accompanied by an uncertainty. It's a good idea to answer the question: what are the main components of measurement uncertainty in a given context? Once these causes are understood, any necessary improvements can be made.

Unlike sensitivity, the notion of accuracy in thermometry depends not only on the measuring instrument, and in particular the sensor, but also and sometimes essentially on the nature of the sensor-medium coupling under study. In other words, it's not enough to install the right instrument to obtain the right measurement. Thermometer set-ups are highly varied in nature, and it is not possible to identify all the problems likely to be encountered. In the following paragraphs, however, we shall try to alert the experimenter to certain precautions to be taken when setting up a thermometer, which also depend on the purpose for which it is intended.

The term "safety in use" is deliberately vague, and seeks to address the following concern: won't the thermometer equipment pose too many maintenance and replacement problems, when all we want to do is avoid highly erroneous measurements? With regard to the problem of replacing an element in the chain, we can ask ourselves about the uncertainty that each element brings separately.

The rather subjective assessments given to the various thermometric devices in the [R 2518] brochure are clear in themselves, but they cover quite different, more or less overlapping and sometimes contradictory realities. A sensor is liable to go out of adjustment for a variety of accidental or permanent causes: the environment under study can create mechanical, pneumatic or chemical stresses in inhomogeneous materials, thermal stresses due to leakage with the outside environment, magnetic, electromagnetic or even more simply electrical stresses.

Digital thermometers, like any other measurement chain, have their own sources of uncertainty at each link in the chain. In case of doubt, there is always recourse to a global calibration of the instrument.

We all have a tendency to check equipment less frequently the more complex it is. This attitude is particularly dangerous in thermometric studies, because sensors evolve considerably over time. Device ageing phenomena, linked to external pollution, internal transformations or even accidents, are particularly serious for temperature sensors linked to controllers. These phenomena...