1、OR-05-4-1 Laboratory Infrared Thermography Technique for Window Surface Temperature Measurements A.H. Elmahdy, PhD, PEng Member ASHRAE ABSTRACT Infrared thermography is a process that uses an infrared scanner toproduce thermal images ofsurfaces by detecting the radiosity of these surfaces. The use o
2、f infrared thermography (IR) as a diagnostic tool in building science has been known particularly for qualitative assessment of building envelopes. Its use as a quantitative means to measure surface tempera- tures was demonstrated in a major CanadaIUSA joint research project that used IR thermograph
3、ic techniques to measure the surface temperature of insulated glass window units. The results of thatproject werepublished in 1996. The earlier work has now been extended in a recent study of complete window assemblies (including frames). Thispaper describes the use ofIR thermography to deter- mine
4、surface temperatures of the glass andframe members of two window units and a calibration transfer standard unit. The IR results were compared with thermocouple temperature measurements atspecijk locations on the surface of the assem- blies. Other participants in this project have presented their fin
5、dings in ASHRAE publications, and a summary paper will be prepared to compare all the experimental and simulation results. This paper focuses on the results obtained from the work performed at the Institute for Research in Construction of the National Research Council Canada (IRUNRCC). INTRODUCTION
6、Infrared thermography is a process that uses an infrared scanner to produce thermal images of surfaces by detecting the radiosity of the surfaces (e.g., emitted and reflected radiation). The use of infrared thermography (IR) as a diagnostic tool in building science has been known particularly for qu
7、alitative Frank Devine assessment of building envelopes. Its use as a quantitative means to measure surface temperature was documented at a 1996 ASHRAE symposium that presented results of a major Canada/SA joint research project on the use of IR thermo- graphic techniques to determine glass surface
8、temperature. Several papers reported on the testing (Elmahdy 1996; Griffith et a1.1996) and simulation (de Abreu et al. 1996; Zhao et al. 1996) of seven insulating glass units. In addition, Sullivan et al. (1996) presented a summary report on the results of all test- ing and simulation. This paper d
9、iscusses an extension ofthe previous research project to develop a test method to measure glass and frame surface temperatures without disturbing the heat transfer boundaries adjacent to those surfaces. Testing (at IRCNRCC and Lawrence Berkeley National Laboratory, LBNL) and simulation (at LBNL and
10、the University of Waterloo) have been performed. The three units included in this project (which are described later in this paper) were tested at LBNL and the results were presented earlier by Griffith et al. (2002). A computer simulation of the three units was also performed and the results were p
11、resented by Kohler et al. (2003). Wright and McGowan (2003) presented a comparison of the measured and calculated temperature profiles of the three units and indicated that the IR thermographic procedures need to be refined. It should be noted that the same (three) units were tested and were shipped
12、 from LBNL labs to NRC labs to ensure the proper comparison at the end of this project. This paper focuses on reporting the results of the tests conducted at IRC/ NRCC. A.H. Elmahdy is a principal research officer and Frank Devine is a technical officer at the Institute for Research in Construction,
13、 National Research Council Canada, Ottawa, Ontario. 02005 ASHRAE. 561 Specimen ID Type Frame Material 1 Calibration Transfer Standard (CTS) NIA 2 Fixed Casement Wood 3 Fixed Casement Wood Glass Pane 7 Overall Size, Glazing Configuration WxH, mm (ft) Foam Core 610 x 914 12.5 mm EPS Dual, air-filled,
14、610 x 914 clear-clear, (2 O“ x 3 o“) (2 O“ x 3 W) 16.5 mm gap Dual, air-filled, 610 x 914 clear low-e, (2 O“ x 3 V) 16.5 mm gap i Pol y styrene P3WV COLD SIDE Thermocouples and mounting rh im c Figure 1 Cross section of the CTS. WARM SIDE - 914 min - I Figure 2 Mounting of the CTS. - surroi panel mm
15、 md TEST SPECIMENS measuring 609.6 mm (24 in.) wide by 914.4 mm (36 in.) long. - Each glass pane was fitted with 12 type-T thermocouples brazed to 5o mm square copper shims and then mounted on the shown in Figure 3. Additional type-T thermocouples (20 on each side) were added on both outer surfaces
16、of the unit to facilitate glass surface temperature measurement and were arranged as shown in Figure 3. Three units were supplied for this set of tests: a calibration (the units have the same perimeter dimensions). Other details as shown in Table 1. Specimen #1 (CTS) was constructed of two sheets of
17、 glass separated by 12.5 mm polystyrene foam (see Figure 1) and mounted in the surround panel of a guarded hot box (see transfer standard (CTS) and two wood-frame window units inner face ofthe glass between the glass and the EPS core, as Figure 2). More details about the design and the characteiza-
18、tion of the CTS are given in Goss et al. (1991). The expanded polystyrene foam core was positioned (sandwiched) between two sheets of 4.76 mm thick glass Samples #2 and #3 are wood casement windows. The insulating glass units (dual seal) were made of two sheets of glass separated by an aluminum spac
19、er bar (incorporating silica-gel desiccant) with a polyisobutylene sealant (primary 562 ASHRAE Transactions: Symposia Figure 3 Calibration transfer standard (CTS) specimen no. 1-thermocouple locations and scan lines. seal) and polysulphide sealant (secondary seal). Sample #2 is constructed of clear
20、glass, and sample # 3 is constructed of low-emissivity coated glass on surface 3 (= O. 10). The wood- frame windows were mounted (one at a time) in the surround panel of the guarded hot box, similar to the configuration illus- trated in Figure 2. The two wood-framed window units were dimensionally i
21、dentical. Thermocouples were attached to the frame and glass surfaces on each side of the units (20 on each side) to measure frame and glass temperatures. These were positioned as shown in Figure 4. INFRARED THERMOGRAPHY SYSTEM The infrared scanning system used for these tests was composed of a long
22、wave infrared imaging scanner having a sensitivity wave band range of 8 to 12 pm and a dedicated (control) computer equipped with specialized IR post- processing and analysis software. This proprietary software program enabled the operator to analyze the thermal images generated by the scanner and t
23、o produce temperature profiles, line temperatures, point temperatures, and histograms, as well as reproducible thermograms. ASHRAE Transactions: Symposia 563 1s Figure 4 Wood-framed units specimen no. 2 and 3-location of thermocouples und scan lines. Prior to testing, the complete Infrared system (s
24、canners and computer system) was upgraded and recalibrated by the manufacturer. The Swedish Testing and Research Institute (Swedish National Laboratory O 1, “Temperature”) performed the calibration. This laboratory is accredited by NIST (corre- spondences were obtained form the lab to certify these
25、facts). The IR analysis software (Thermovision 1992) was also upgraded. The scanner (camera) used for these tests was equipped with a 20” horizontal by 10” vertical field-of-view lens, which limited the field of view (area) that can be captured within the room-side confines of the guarded hot box. T
26、he image processing and analysis depended on the knowledge of accurate emissivity of the surfaces under inves- tigation as well as other variables that could affect the final thermal image (e.g., ambient temperature, relative humidity of air in the optical path, and the optical path length). The IR
27、scanner recorded all the radiation (both direct and reflected) received either from the intended target or from any other radi- ative surfaces in its field of view (FOV). The IR system was used in the hot box in a way similar to the way the manufacturer calibrated it, that is, placing the IR camera
28、at 90” in front of the target. No attempt was made to tilt the IR camera in order to look at the corners of the test spec- imen. Therefore, all the images presented in this report were taken with the camera at 90” in front of the surface of the CTS or window. 564 ASHRAE Transactions: Symposia Figure
29、 5 Infrared camera setup. To scan the entire surface of the CTS and windows, the specimen area to be viewed was divided into three vertical sections measuring 610 mm horizontal and 305 mm vertical (one-third the area of the specimens) in order to obtain the best IR resolution possible within the ava
30、ilable room-side area of the hot box (see Figure 5). This was also necessary to reduce the effects of any undesired radiation from the surrounding surfaces. The following emittance values were used for the samples tested: glass, 0.86; wood, 0.90; and tape (vinyl), 0.9. It is worth noting that the ot
31、her researchers also used those values for test- ing and simulation. The surface emittance of all materials was measured at LBNL (Griffith et al. 2002) and all participants in this work used the same values. Other variables, such as the air temperatures and relative humidity in the optical pass, wer
32、e locally measured and used during the IR imaging process. EXPERIMENTAL SETUP Prior to mounting any unit in the surround panel, the guarded hot box thermocouple monitoring system together with the surround panel, specimen, and air temperature ther- mocouples were calibrated. The testing was carried
33、out using a guarded hot box in which the environmental conditions on both sides of the spec- imens were controlled and monitored by a separate computer control and data acquisition system. The system was set up to monitor and record the temperature measurements made by up to 200 thermocouples in gro
34、ups of 20. In this case, groups of 20 type-T (copper-constantan) thermocouples were mounted on each side of the specimen, the surround panel, the wind machine cold air tubes, and the baffle assembly (both sides). The 24 CTS thermocouples were also included in the system when required as two groups o
35、f 12 thermocouples. The room-side and weather-side temperature control system used signals from resistance temperature detectors mounted in appropriate locations on each side of the surround panel. For these tests, the system was set to control tempera- ture conditions of+21*lC ontheroomsideand-18 p
36、ower measurement is accurate within *0.01%; CTS core material is determined within 42%; film heat transfer coefficient is determined within *2.3%. More details about the calibration of the NRC guarded hot box are given in Elmahdy (1 992). The baMe assembly is equipped with three removable sections,
37、each measuring 610 mm wide by 305 mm high (representing one-third of the test specimen area). Each of these sections, when removed, allowed the scanner to view a one-third (top, middle, or lower) section of the test specimen. This was necessitated by the field-of-view limitations of the scanner and
38、available space in the chamber (see Figure 5). Post-Processing The images generated by the infrared scanner were processed utilizing a proprietary software program (Thermo- vision 1992), and the analysis software program was installed on the scanners computer. This software enabled the system to pro
39、duce a temperature profile for any specified line imposed on the produced thermogram (thermal image). This was utilized to obtain horizontal and vertical temperature line profiles. The profile data were then downloaded to a spread- sheet for further analysis. The following is a summary of the techni
40、cal data of the IR system and the associated accuracy of measurements. More details about the IRC infrared system, accuracy, calibration, and resolution are given in Elmahdy (1 996) and Thermovision (1 992). Longwave, Sterling-cooled detector type: mercuryicadmiuml Spectral response: 8 to 12 pm Temp
41、erature range: user-selected from -30C to 1500C Sensitivity: 0.08“C at 30C telluride (MCT) Absolute accuracy: 61C (over the range -30 “C to 80C) Relative accuracy: weather side temperature = -18“. tude of such errors is a considerable task and it is certainly beyond the scope of this paper. As a rem
42、inder, the intent of this paper is to report on the use of IR thermography as a tool for surface temperature measurement and compare the results with other test labs and other theoretical calculations (performed by other authors). In-depth error analysis and critiques of the software of the IR syste
43、m were not the subject of this study. All thermal images were recorded with the IR camera at 90“ in front of the target. It is possible that background radi- ation may have affected the recorded images. Because all measurements were made within the field of view that does not extend beyond the targe
44、t area, it was possible that the error of background radiation was minimal. Furthermore, before processing any image, air temperature and relative humidity of the environment in the optical path were entered into the IR software program to account for such effects. Nevertheless, corrections were mad
45、e automatically in the system software to compensate for any of these effects. Further details about these corrections are given in Thermovision (1992). It was obvious from the thermal images of the glass areas close to the spacer bar that they exhibited cooler temperatures, while the surface temper
46、ature increased toward the center-of- glass area. Figures 6 and 7 also demonstrated the sensitivity of the IR camera to any temperature changes in the field of view. For example, the thermocouple wires were shielded with plas- tic cover, which was slightly shiny. This resulted in a display of bright
47、er color lines in the thermal images. In addition, the tape used to affix the thermocouples to the different surfaces had an emissivity of E = 0.9 compared to that of the glass, which was E = 0.84. The result was a slight shade appearing ASHRAE Transactions: Symposia Figure 7 Infrared thermal image-
48、Clear glass window in woodframe room. Temperature = 21 “C; weather side temperature = -18“. where the tape was placed. The three brighter circles shown in the IR thermal images are the result of the reflection of the cameras lens off the surface of the glass (narcissus effect)- see Figures 6 and 7.
49、This would have produced localized error in the surface temperature measurements over an area equiv- alent to the area of the lens. The shaded lines in Figures 6 and 7 are the boundaries of the three images that compose each figure. As indicated earlier, each image was split into three portions (top, middle, and bottom) due to the limited space in the warm-side chamber of the hot box. Several IR themograms were taken during the test and the line temperature profiles were produced as explained earlier. Figures 8 and 9 are samples of the line temperature profiles at specifi
