Optimization of Thermal Radiation Source for High Temperature Infrared Thermometer Calibration

Transcription

1 Optimization of Thermal Radiation Source for High Temperature Infrared Thermometer Calibration Speaker/Author: Frank Liebmann, Fluke Calibration, 799 Utah Valley Dr., American Fork, Utah, 84003, USA , FAX Author: Tom Kolat, Fluke Calibration Abstract Industrial level infrared thermometers are being increasingly designed to measure temperatures above 500 C. A thermal radiation source is needed to calibrate these instruments. The infrared thermometers designed to measure these temperatures generally measure with a smaller field-ofview. This means there is a possibility of using a blackbody cavity as the thermal radiation source. A previous attempt was made to use a cavity mounted inside of a furnace meant for thermocouple calibration for a thermal radiation source. However, when the cavity s emissivity was measured, it was found the emissivity was not constant for different wavelengths and was very dependent on cavity position. This paper discuses a new attempt to mount a cavity inside a thermocouple furnace with much better temperature uniformity. It discusses the measurements made to verify the emissivity of the cavity using industrial radiation thermometers. It also talks about measurements made with industrial handheld infrared thermometers and compares these measurements to measurements made using other thermal radiation sources and the associated measurement uncertainty. Learning Objectives Learn about IR thermometer calibration Be informed about available calibration sources for radiation thermometer Learn about attributes of a blackbody cavity Learn how to qualify a calibration source 1 Introduction In 2008, Fluke Calibration released a series of flat plate infrared calibrators for the calibration of infrared thermometers. The temperature range of the flat plates is -15 C to 500 C. Since the release of these products, there has been increased demand for a calibrator to cover a higher temperature range. Previously, a blackbody cavity was used inside a thermocouple calibration furnace for evaluation purposes. This solution was found to be marginal. This was due to the poor temperature uniformity along the cavity walls. Since then, Fluke Calibration has come out with a new thermocouple calibration furnace with much better uniformity on its walls.

2 2 Background To understand the developments for this new design, one must understand the work that lead up to the research presented in this paper. The aforementioned need for a source was principle in doing this research. There are several commercial solutions on the market that did not meet the need due to issues with traceability and emissivity in the 8 14 µm band. The availability of a more uniform furnace would solve many of these problems. Such a furnace became available, and testing was done to verify that it would be a good solution for the customer. 2.1 Calibration Sources for Radiation Thermometry There are a number of calibration sources for infrared (IR) thermometers. These sources take on one of two forms, a flat-plate thermal radiation source or a cavity thermal radiation source [1]. These sources have their own advantages and disadvantages. The cavity source theoretically provides an emissivity of close to unity throughout the entire electromagnetic spectrum [2]. However, there are two factors that can diminish the effective emissivity. First, this emissivity can be diminished when there is poor temperature uniformity, especially poor temperature uniformity on the cavity walls. Second, this effective emissivity is also diminished when the temperature of the cavity bottom is poorly estimated [3]. This is especially the case when using Scheme I traceability which means traceability comes through contact thermometry [4]. The solution for many laboratories is to calibrate the cavity with a radiation thermometer, otherwise known as Scheme II traceability [4]. This is only valid if the radiation thermometer has the same spectral bandwidth as the IR thermometers to be calibrated using the cavity. In other words, the units under test must have the same spectral bandwidth as the transfer standard. The other limitation with using a cavity is difficulty due to the wide field-of-view of less expensive IR thermometers. The depth of the cavity is often so deep that the IR thermometer field-of-view is larger than the cavity bottom. For these types of measurements a flat plate calibrator is often used [1]. Flat plate calibrators often have a range of up to 500 C. Flat plates with higher temperature ranges are costly. 2.2 Previous attempts Previously a cavity was made using a Fluke Calibration Model 9112B thermocouple calibration furnace [5]. This attempt of using a cavity inside a thermocouple furnace gave marginal results. The poor uniformity of the furnace caused two effects that caused large uncertainty. First, the heat flow between the cavity bottom and the reference probe was large enough that there was a considerable temperature drop between these two locations. Second, the temperature on the cavity walls was poor enough that the effective emissivity of the cavity was not well estimated. This was evident when measuring the cavity at two different wavelengths. This is a problem common to many industrial level cavities. After doing this research, it was decided that this was not a good solution for IR thermometer calibrations above 500 C. 2.3 Availably of a more uniform furnace In 2014, Fluke Calibration released the Model 9118A Thermocouple Calibrator furnace. The furnace has improved axial uniformity over the Model 9112B. It was hoped that this improvement in uniformity would result in a better solution for a blackbody cavity. The one drawback with the 9118A was the diameter of the tube, 40 mm. This was thought not to be large enough for work

3 with IR thermometers. As a result, the 9118A used for this investigation was modified to accommodate a cavity of 50 mm in diameter. 2.4 Description of cavity The cavity is shown in Figure 1 as it is installed in the furnace. The cavity used is a cylindroconical design with an apex angle of 120 [2]. The cavity walls were designed to extend well beyond the cavity bottom. In Figure 1, the cavity is shown with the reference probe in place. The intent was to place the sensor of the reference probe as close as possible to the conical apex of the cavity bottom. The reason for this is to minimize temperature measurement error due to heat transfer between the probe s sensor and the cavity bottom surface [3]. The cavity may be positioned at various axial depths. These axial depths are measured from the reflector plate by the cavity opening to the opening of the cavity walls. These depths are referred to as cavity positions for the rest of this paper. 3 Contact measurements 3.1 Uniformity Figure 1: 9118A blackbody cavity diagram. In order to calculate data needed for the blackbody emissivity model and to estimate uncertainty due to cavity bottom heat exchange, temperature gradient was measured along the cavity axis at cavity positions of 5, 10, and 15 cm and at three cavity temperatures. These results are shown in Figure 2.

4 Stability / K Heat Exchange / K / cm C 660 C 960 C Cavity Position / cm 3.2 Short Term Stability Figure 2: 9118A axial uniformity in blackbody cavity configuration. The short term stability of the 9118A blackbody cavity measured by a radiation thermometer was observed and recorded. The stability measurements took place for a 5 minute period. The stability reported is 2 standard deviations of the samples over this 5 minute period. In Figure 3, the results of the 9118A stability as measured by a radiation thermometer are compared to the Fluke Calibration IR Laboratory (AFL) in American Fork s HT cavity s stability as measured by a radiation thermometer [6][7] Temperature / C µm µm HT 8-14 µm HT 3.9 µm Figure 3: Cavity stability (2σ) as measured by a radiation thermometer.

5 3.3 Traceability of contact measurements The traceability of the contact measurements comes through the American Fork Primary Temperature Laboratory, through NIST, to the Système International d'unités (International System of Units or SI). The traceability of each of the measurements is covered more in the next section. 4 Non-contact measurements In order to test the cavity design, measurements were made on the cavity. These measurements involved comparing the traceable readout temperature of a contact probe located close to the conical apex at the cavity bottom to that of a radiation thermometer measuring the cavity bottom. The radiation thermometers used for the comparison ranged from hand held instruments with thermopile detectors to precision radiation thermometers with pyroelectric detectors [8]. In all measurements two chains of traceability to the SI were maintained. These are the traceability by contact thermometry and traceability by radiation thermometry. In all cases the two measurements were compared using normal equivalence (En) [9][10]. The traceability for each of the measurements and the comparison made is covered in Figure 4. Scheme A Scheme B Scheme C Scheme D Figure 4: Traceability schemes. The measurements were divided into six sets for organizational purposes. Only a portion of the results are reported in this paper due to the sheer number of tests. The data was taken at cavity positions of 5, 10, and 15 cm. Data for each of these positions was consistent. In all but Set 6, the 15 cm cavity position data is reported. In the case of Set 6, the 5 cm cavity position data is reported, since the 10 and 15 cm positions were found to be unusable with one of the infrared thermometers. A summary of the sets of measurements performed with their traceability is shown in Table 1. Details on these measurements are discussed in the following sections.

6 εeff Table 1: Sets of measurements with their traceability scheme. Data Set RT λ / µm Scheme Ref Source 9118A Ref Cav Pos. 1 KT B HT Cav F cm 2 TRTII 3.9 B HT Cav F cm 3 TRTII 8 14, 3.9 C PTB F cm 4 KT B HT Cav F cm 5 F568, 561, D F4181 F cm 6 F62M D F4181 F cm 4.1 Initial measurements Initial measurements were made using an IR thermometer model Fluke 568 compared to a type K thermocouple using traceability Scheme A [11]. Two cavity designs were measured. These are termed as a thin cavity and a thick cavity for the sake of discussion. The thick cavity is the one shown in Figure 1. The measured emissivity of the thin cavity showed a large dependence to its axial position. This was believed to be a result of heat flow in and around the mass of the thin cavity. This design was abandoned and not considered any further. The thick cavity was measured at various axial positions. The results of one of these tests is shown in Figure 5. This testing showed that the design was promising, since the emissivity was maintained close to unity for a wide range of cavity axial positions. This meant that the furnace s temperature uniformity did not adversely influence the emissivity of the cavity as was observed in the previous study Thin PRT No I Thin PRT I Thin TC No I Thin TC I Thick TC No I d / cm Figure 5: Results of initial measurements.

7 4.2 Intercomparison with cavities The next set of tests compared the cavity s temperature as measured by a radiation thermometer with a pyroelectric detector to that of a thermocouple Fluke Model 5650 shown as Sets 1 and 2 in Table 1. Additional measurements were made using a Fluke Model 5628 PRT as a reference. These measurements are shown as Set 4. It should be noted that the measurements using a PRT showed larger uncertainty. This was due mainly the cavity bottom heat exchange uncertainty [3]. This was due to the length of the PRT sensor and the heat flow mentioned previously. An additional set of data, shown as Set 3, was taken using a radiation thermometer calibrated at PTB. For these sets of data, measurements were made both in the 8 14 µm and 3.9 µm bands. The radiation thermometers used were Heitronics models KT19.82II and TRTII. The traceability for these measurement is shown as Schemes B and C in Figure Intercomparison with flat-plates Tests involving handheld IR thermometers were performed. These are shown in data Sets 5 and 6. The IR thermometers were measured using calibrated 4181 flat plates and simultaneously measured using the 9118A cavity with a contact thermometer reference probe. These measurements were done with several different models of IR thermometers. The traceability for these tests is shown in Figure 4 under Scheme D. Measurements were made above 500 C on the cavity only. These measurements were done to compere these consistency of these results with the results at lower temperatures. The experimental uncertainties for the infrared thermometer measurements are shown in Table 2. Table 2: IR thermometer experimental measurement uncertainty. Thermal Radiation Source: 9118A Cavity U / K (k = 2) IRT 300 C 390 C 480 C 540 C 600 C 680 C Fluke Fluke NA NA NA Fluke 62 Max NA Fluke NA NA NA Thermal Radiation Source: Fluke Calibration 4181 U / K (k = 2) IRT 300 C 390 C 480 C 540 C 600 C 680 C Fluke NA NA NA Fluke NA NA NA Fluke 62 Max NA NA NA Fluke NA NA NA 4.4 Modeling of the cavity The cavity was modeled for emissivity using STEEP3 [12]. The results of this modeling are shown in Table 3 for the cavity positions 5 and 15 cm. It should be noted that the paint used on the cavity walls and cavity bottom have a significantly lower emissivity in the 3.9 µm band than in the 8 14 µm band [7].

8 Table 3. Results of STEEP3 modeling. Cavity Position Temperature Isothermal Non-Isothermal 3.9 µm 8 14 µm cm cm Results Table 4 gives the results of comparison of the AFL cavity with the 9118A cavity. In addition the comparisons with the PTB calibration are shown. Results are given for both the 8 14 µm band and the 3.9 µm band. Table 5 gives the results of the comparisons between the 9118A cavity and the 4181s using IR thermometers. All values in these tables were normalized to the nominal temperature. Table 4: Results of radiation thermometer comparisons. λ / µm LAB CAV / C 9118A / C Diff / C U CAV / K U 9118A / K En Set Set Set Set Set Set Set Set Set Set Set Set Set Set Set Set

9 Table 5: Results of infrared thermometer comparisons. IRT 4181 / C 9118A / C Diff / C U CAV / K U 9118A / K En Set 5 F Set 5 F Set 5 F Set 5 F Set 5 F Set 5 F Set 5 F Set 5 F Set 5 F Set 5 F Set 5 F Set 5 F Set 6 F62M Set 6 F62M Set 6 F62M Set 6 F62M Set 6 F62M Set 6 F62M Tables 4 and 5 show that all of the measurements passed the normal equivalence test of being below unity [9] [10]. Another look at the tests is shown in Figure 6 showing the amount of normal equivalencies above and below µm RT 3.9 µm RT IRT < > 1.00 < > 1.00 < > 1.00 Figure 6: Normal equivalence distributions. Additional measurements were made above 500 C with the IR thermometer models Fluke 568 and 62 Max+. These measurements were not compared to another traceable measurement. However, they were plotted to determine their consistency. The results are shown in Figure 7 with a 2 nd order polynomial fit of their data. The uncertainty of the measurements are shown with a coverage factor of 2 (k = 2) by the error bars.

10 measurement error / K measurement error / K measurement error / K measurement error / K F568 #1 F568 # Temperature / C Temperature / C F62M+ #1 F62M+ # Temperature / C Temperature / C Figure 7: Self consistency of measurements above 500 C. 5 Conclusion The testing showed that the Fluke 9118A modified for use with a cavity is a viable option for IR thermometer calibrations over 300 C. The extensive testing completed and presented in this paper proved this out. The test uncertainties were favorable and would meet the needs of IR thermometer calibrations. In addition, dependence on spectral wavelength was not observed. It is possible for a calibration laboratory to reach similar uncertainties as presented in this paper with use of the proper contact thermometer and procedure. References 1. ASTM Standard E , Standard Guide for Selection and Use of Wideband, Low Temperature Infrared Thermometers, ASTM International, West Conshohocken, Pennsylvania, 2010, DOI: /E DeWitt, D., Nutter, G., Editors, Theory and Practice of Radiation Thermometry, 1988, pp , , Fischer, J., Saunders, P., Sadli, M., Battuello, M., Park, C., Yuan, Z., Yoon, H., Li, W., van der Ham, E., Sakuma, F., Yamada, Y., Ballico, M., Machin, G., Fox, N., Hollandt, J., Ugur, S., Matveyev, M., Bloembergen, P., Uncertainty budgets for calibration of radiation thermometers below the silver point, CCT-WG5 working document CCT-WG508-03, BIPM, Sèvres, France, May 2008.

11 4. ASTM Standard E e1, Standard Test Method for Calibration and Accuracy Verification of Wideband Infrared Thermometers, ASTM International, West Conshohocken, Pennsylvania, 2010, DOI: /E Liebmann, F., Kolat, T., Use of a Furnace for a Thermal Radiation Source, Proceedings of NCSLI Liebmann, F., Kolat, T., Traceability and Quality Control in a Radiation Thermometry Laboratory, NCSLI Measure, vol. 7, no. 1, pp 72 77, Liebmann, F., Infrared calibration development at Fluke Corporation Hart Scientific Division, Proceedings of SPIE Thermosense XXX, 6939, 5, VDI/VDE Guideline 3511 Blatt 4: Temperature Measurement in Industry Radiation Thermometry, ISO/IEC 17043:2010 Conformity assessment General requirements for proficiency testing, ILAC-G22:2004 Use of proficiency testing as a tool for accreditation in testing, ASTM Standard E230/E230M 12, Standard Specification and Temperature-Electromotive Force (emf) Tables for Standardized Thermocouples, ASTM International, West Conshohocken, Pennsylvania, 2012, DOI: /E0230_E0230M Prokhorov, A., Monte Carlo Method in Optical Radiometry, Metrologia, Vol. 35, pp , 1998.