Reflection errors in radiation thermometry
A radiation thermometer infers the temperature of an object by measuring the amount of radiation that is emitted by the object within a specified spectral band. This radiation is directly related to Planck’s law and the emissivity of the surface of the object. However, as well as emitting radiation, objects also reflect a certain fraction of the radiation that originates from other objects in its surroundings. This reflected radiation is detected along with the emitted radiation and leads to an error in the thermometer’s reading. The higher the difference between the temperature of the surroundings and the target, the higher this ‘reflection error’.
Reflection errors are a problem in almost all applications of radiation thermometry. In high-temperature industries, such as metal processing and petrochemicals, the objects whose temperature must be measured to control or optimise the process are often surrounded by heaters, flames, or hot furnace walls. The very high temperature of these surrounding objects can lead to errors of, typically, several tens of degrees Celsius, and sometimes up to 200 °C. In low-temperature applications, such as food processing and storage, the surrounding walls are often at a similar temperature to the target, and represent a significant source of reflections.
At MSL, our research into reflection errors has focused on developing methods for minimising or eliminating reflection errors in difficult industrial environments, most notably in the petrochemical industry. We have identified three measurement regimes and developed strategies for managing the errors in each regime:
Strategy 1: When the surroundings are much colder than the target, reflection errors are negligible and the instrumental emissivity on the radiation thermometer should be set to the value of the target emissivity. This is recognised as ‘normal’ usage, as recommended by most radiation thermometer manufacturers. However, this regime is only rarely encountered in practice.
Strategy 2: When the surroundings are at a similar temperature to the target, the target behaves as if it were a blackbody (having an emissivity of 1.0). This is because, no matter what the actual emissivity of the target, the reflected radiation exactly compensates for the reduced emission (compared to a true blackbody) from the target. Thus, setting the instrumental emissivity on the radiation thermometer to 1.0 provides a good measurement of the target’s temperature. A small error arises when the surroundings and target are not exactly at the same temperature, but this error is easily quantified. In many situations, this strategy can be taken advantage of by making measurements inside gaps or cracks between objects at similar temperatures instead of on flat surfaces.
Strategy 3: When the surroundings are at a much higher temperature than the target, significant reflection errors can occur. Setting the instrumental emissivity to 1.0 minimises the error, but in most cases additional corrections for the error need to be made. We have developed a methodology for modelling these reflection errors based on the geometry of the space surrounding the target, in some cases including the bi-directional reflectance distribution function (BRDF) of the target’s surface. The BRDF describes the angular dependence of the target’s reflectance, and becomes increasingly significant as surfaces become smoother and smoother.
Our research into reflection errors has led to the development of a consultancy service for the petrochemical industry, whereby we provide computer modelling and temperature surveys of large processing furnaces, such as reformers, platformers, etc. A crucial part of the modelling is a complete uncertainty analysis for the measurement process. This allows plant operators to assess the risks associated with operational decisions based on the temperature measurements, enabling optimum plant operation.
An interesting consequence of the uncertainty analysis is that we have discovered that when reflection errors are large, there exists an optimum operating wavelength for the radiation thermometer that minimises the uncertainty. This is contrary to the traditional shortest-wavelength-is-best advice for most laboratory measurements. The reason for the optimum is that at shorter wavelengths, the reflected radiation is a larger fraction of the total detected radiation, so corrections for reflection errors are larger. The uncertainty associated with large corrections is correspondingly large, and in fact increases exponentially with decreasing wavelength. At long wavelengths, the usual proportional increase of uncertainty with wavelength occurs. Therefore, there exists a minimum uncertainty between the exponential increase at short wavelengths and the linear increase at long wavelengths. This minimum is relative broad, so there is some flexibility in the choice of suitable wavelength.
While we have mainly focused our research efforts on high-temperature applications, and published a book aimed at the petrochemical industry. We have also begun analysis of reflection errors for low-temperature radiation thermometers. These thermometers utilise an algorithm that automatically compensates for reflection errors under specific conditions, which are not always achievable in practice (see Low-Temperature Radiation Thermometers). Manufacturers do not provide documentation on this algorithm, and without a full understanding of its consequences, many users are left wondering about the reliability of such instruments. We have initiated a programme to quantify the errors and uncertainties for these instruments with the aim of improving measurement confidence, mainly in the food industry.
For further information contact Peter Saunders
P Saunders, D R White, “A theory of reflections for traceable radiation thermometry”, Metrologia, 32, 1–10, 1995.
P Saunders, D R White, “A model for reflection errors in radiation thermometry: Application to tube misalignment in reformer furnaces”, in Proceedings of TEMPMEKO ’96, Sixth International Symposium on Temperature and Thermal Measurements in Industry and Science, edited by P Marcarino, Levrotto & Bella, Torino, 395–400, 1997.
G Beynon, P Saunders, “Comparative infra-red measurements of tube temperatures in a reformer furnace”, in Proceedings of TEMPBEIJING ’97, Proceedings of the International Conference of the Temperature and Thermal Measurements, edited by Z Baoyu, H Lide, Z Xiaona, Standards Press of China, Beijing, 84–89, 1997.
D R White, P Saunders, S J Bonsey, J van de Ven, H Edgar, “Reflectometer for measuring the bidirectional reflectance of rough surfaces”, Applied Optics, 37, 3450–3454, 1998.
P Saunders, S J Bonsey, D R White, “Determination of reformer-tube temperature by means of a CCD camera”, High Temperatures – High Pressures, 31, 83–90, 1999.
P Saunders, “Reflection errors in industrial radiation thermometry”, in Proceedings of TEMPMEKO ’99, Seventh International Symposium on Temperature and Thermal Measurements in Industry and Science, edited by J F Dubbeldam, M J de Groot, IMEKO/NMi Van Swinden Laboratorium, Delft, 631–636, 1999.
P Saunders, “Reflection errors and uncertainties for dual and multiwavelength pyrometers”, High Temperatures – High Pressures, 32, 239–249, 2000.
P Saunders, A B Trotter, H Edgar, D M J Cochrane, “In situ measurement of catalyst tube emissivity by means of a portable solid integrating sphere reflectometer”, Measurement Science and Technology, 12, 622–626, 2001.
P Saunders, “Reflection errors for low-temperature radiation thermometers”, in Proceedings of TEMPMEKO 2001, 8th International Symposium on Temperature and Thermal Measurements in Industry and Science, edited by B Fellmuth, J Seidel, G Scholz, VDE Verlag GmbH, Berlin, 149–154, 2002.
P Saunders, “Absorption and emission effects on radiation thermometry measurements in reformer furnaces”, in Temperature: Its Measurement and Control in Science and Industry, Vol. 7, edited by D. C. Ripple et al., AIP Conference Proceedings, Melville, New York, 825–830, 2003.
D R White, P Saunders, H Edgar, “On the utility of laser pyrometers for measuring reformer tube-skin temperatures”, in Proceedings of TEMPMEKO 2004, 9th International Symposium on Temperature and Thermal Measurements in Industry and Science, edited by D Zvizdic, LPM/FSB, Zagreb, 1249–1254, 2005.
P Saunders, “Accurate temperature measurement in reformers”, in Proceedings of AIChE 50th Annual Safety in Ammonia Plants and Related Facilities Symposium, 343–352, 2005.
P Saunders, Radiation Thermometry: Fundamentals and Applications in the Petrochemical Industry, SPIE Press, Bellingham, 2007.
D R White, P Saunders, “A graphical method for calculating reflection errors in radiation thermometry”, International Journal of Thermophysics, 29, 395–402, 2008.