Temperature and Humidity Standards Research
Temperature-related research within MSL covers a very wide set of topics ranging through difficult industrial measurements, the characterisation of sensors, new calibration techniques and instruments, physical modelling and uncertainty analysis, and the fundamental definitions of the standards.
Traditionally, uncertainty associated with the emissivity of target objects has been considered to be the greatest source of error in industrial radiation thermometry measurements. However, in our experience, the errors caused by reflections from other hot objects surrounding the target can be several times the magnitude of any emissivity errors. Reflection problems occur in all high-temperature processing industries, such as the petrochemical, power generation and metal processing industries, with errors as large as 200 ºC. MSL has been working on modelling and solving this problem, and has developed large computer models of industrial furnaces that allow furnace operators and users of the radiation thermometers to apply corrections for the reflections. Read more >>
When we calibrate a thermometer for a client, we make an estimate of the uncertainty in temperatures measured using that thermometer. The uncertainty is based in part on measurements made during the calibration and in part on our understanding of the causes of errors. As industry requires ever increasing accuracies, we must improve our understanding of the various effects that lead to errors or uncertainty in the temperature measurements. Recently, we have decided to explore in detail three effects that occur in industrial thermometers:
- heat leaks up the thermometer stem (immersion effects);
- hysteresis or memory effects in platinum resistance thermometers; and
- ageing effects in thermocouples. All of these effects are well known, but not so
It has been said that the accuracy of radiation thermometers far exceeds our ability to use them intelligently. This is certainly true of the huge number of low-temperature radiation thermometers finding application in the food processing, storage and transport industries. Not only do they suffer severely from reflection and emissivity problems, they have additional errors due to the radiation emitted by the thermometer itself. Worse, thermometer manufacturers have designed the instruments to deliberately obscure some of the errors from the user. MSL has been developing simple measurement strategies for overcoming these errors during both use and calibration. Read more>>
Relative humidity depends on the concentration of water vapour in the air (usually measured as the dew-point temperature) and the air temperature. The humidity sensor can give a different reading on rising and falling humidity (hysteresis). Therefore, to calibrate a sensor at 4 temperatures and 5 different relative humidities for example, 40 separate calibration measurements may be required. The two-dimensional aspect of humidity combined with the long response times and complex behaviour of most sensors means relative humidity calibrations can take many days and the large data-set is not easily interpreted by the user.
MSL is investigating ways of
- identifying and quantifying factors that influence the sensor response,
- reducing the complexity of the data set by fitting calibration equations, and
- reducing the calibration times through more efficient use of transient calibration data
All temperature and humidity measurements rely on sensors; i.e. we depend on some device such as an electrical resistance or capacitance to respond to the temperature. We then have to relate measurements of resistance or capacitance etc. back to temperature. That is, all thermometers require a calibration equation. This has led us to investigate not only the suitability of different calibration equations for different sensors, but also how uncertainties associated with the calibration process propagate and affect the uncertainties in temperatures measured at some time later. Such research has been carried out for standard platinum resistance thermometers, radiation thermometers, and humidity sensors. Read more >>
A good standard platinum resistance thermometer (SPRT) covers the range from about –190 °C to 960 °C with an accuracy of about 1 mK. Such temperature measurement routinely requires the measurement of the resistance of the platinum sensor to an accuracy of about 1 part per million (ppm). Less routine measurements that support research may require resistance measurements with an accuracy of better than 0.01 ppm. Since the 1960s, when the first automatic resistance bridges were developed, we have faced the problem of proving that the bridges are as accurate as we have required and as the manufacturers have claimed. For a long time the best we could do was to confirm the accuracy at a few well chosen points, and to do a few basic 'health checks' on the bridges. MSL finally solved the bridge calibration problem with the development of a specially designed resistance network. Read more >>
All radiation thermometers and thermal imaging systems rely on lenses or mirrors to focus an image of the target onto a radiation detector. Imperfections in the optical system inevitably lead to some of the infrared radiation being scattered within the lenses. This means the boundaries of the target become blurred, and the indicated temperature becomes sensitive to the size of the target. MSL has been refining methods for measuring and modelling the size-of-source effect for the high-accuracy thermometers used to maintain the temperature scale. Read more>>
The SI unit of temperature, the kelvin, is currently defined as 1/273.16 of the temperature of the triple point of water and, in principle, this defines the kelvin exactly. However, there are a number of small effects that limit the practical accuracy of a triple-point-of-water cell. MSL has researched the impacts of the isotopic composition of the water, the slow dissolution of the glass cell by the water, as well as other smaller effects. Read more>>
Ideally, the temperature scale should be purely thermodynamic, i.e., based on equations of state, such as PV = NkT, which relates the pressure and volume of N moles of an ideal gas to temperature. Similar equations of state exist that enable us to measure temperature in terms of the speed of sound, the radiance of a blackbody, or the dielectric constant of a gas. However, despite decades of research into such thermodynamic thermometers, none have been found to offer sufficient accuracy, convenience, and low cost to satisfy our needs of a temperature standard. The lack of a good thermodynamic thermometer is the reason we resort to a recipe for a temperature scale (the International Temperature Scale of 1990), using interpolation and defined temperatures for the freezing and triple points of various pure substances. MSL has been exploring the possibility of measuring Johnson noise, caused by the random movement of electrons in a conductor, to measure thermodynamic temperature. Read more>>
In principle, humidity is simply a measure of the concentration of water vapour in air. In practice, it is measured by cooling the air until condensation occurs (the dew-point temperature), or as a fraction of the vapour pressure of water at the same temperature (relative humidity). The relationship between these two quantities and the concentration of water are conventionally given by empirical formulae. The use of these expressions raises several problems:
- what is the uncertainty in the measurements arising from inaccuracies or uncertainties in the expressions?
- what happens when the expressions are used more than once in a calculation (commonly the case) – do the errors cancel? , and
- what happens if we want to measure humidity outside the normal range e.g., relative humidity of air at 120 ºC.