- Email: [email protected]
Methodology for spectral emissivity measurement by means of single color pyrometer
Measurement 82 (2016) 403–409
Contents lists available at ScienceDirect
Measurement journal homepage: www.elsevier.com/locate/measurement
Methodology for spectral emissivity measurement by means of single color pyrometer Carlo Purpura a, Eduardo Trifoni a, Marilena Musto b,⇑, Giuseppe Rotondo b, Roberto della Ragione b a b
CIRA Italian Aerospace Research Centre, Via Maiorise, 81043 Capua, Italy Dipartimento di Ingegneria Industriale, Università degli Studi di Napoli Federico II, P.le Tecchio, 80, 80125 Napoli, Italy
a r t i c l e
i n f o
Article history: Received 1 September 2015 Received in revised form 8 January 2016 Accepted 11 January 2016 Available online 14 January 2016 Keywords: Emissivity CFD Temperature measurement Single color pyrometer Semi-empirical formula
a b s t r a c t The application of non-intrusive optical devices, such as infrared pyrometers able to measure the temperature of surfaces, makes possible the evaluation of emissivity curve of the tested materials at different temperature values. In this paper the authors propose a methodology for the spectral emissivity measurement by means of a single color pyrometer providing a semi-empirical formula, obtained experimentally at CIRA’s laboratory. The semi-empirical formula allows to know the actual emissivity value of the sample’s surface for whatever emissivity value set up on the pyrometer. The agreement between the experimental emissivity and the emissivity predicted by semi-empirical formula was verified. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction The measurement of the temperature of the material’s surface solicited by thermal loads, has always been the target of experiments developed in laboratory. The development of optical devices operating in the infrared region of the spectrum has given the opportunity to measure temperatures up to several thousands of kelvin. In fact the use of non-contact devices such as pyrometers or thermocameras has given the possibility to measure temperatures of the material’s surface without the insertion of temperature sensors in the samples. The only problem occurring with this type of application is the knowledge of the material’s surface characteristics, in particular the emissivity. Purpura et al. [1] have performed a technique to determine the experimental emissivity of a material during a test in the PWT-SCIROCCO by comparing the temperature values obtained by a thermocamera operating with the ⇑ Corresponding author. Tel.: +39 081 7682290; fax: +39 0812390364. E-mail address: [email protected] (M. Musto). http://dx.doi.org/10.1016/j.measurement.2016.01.018 0263-2241/Ó 2016 Elsevier Ltd. All rights reserved.
known emissivity and by a dual color pyrometer. They have found an increasing of about the 10% of the material’s emissivity during the development of the test campaign. The use of a dual color pyrometer to determine the temperature of the sample was performed also by Teodorescu et al. [2], to measure the spectral emissivity of nickel by a Fourier transform infrared spectrometer. But in that analysis the sample was located in a vacuum chamber, so it was not considered the oxidation influence on emissivity. The introduction of dual color optical devices has allowed the temperature measurement by-passing the knowledge of the material’s emissivity and considering the material as a gray-body in the two operating wavelengths k. The use of thermocouples to calibrate a pyrometer to measure the temperature of a sample was performed by Hagqvist et al. [3], which determined that the temperature uncertainty was less than 2.5%. In that analysis the sample was located in air at ambient pressure so the effect of oxidation phenomena was considered. The use of thermocouples, and the effect of surface oxidation, were described also by Shi et al. [4] to affect the spectral emissivity.
404
C. Purpura et al. / Measurement 82 (2016) 403–409
To evaluate the spectral emissivity values, the authors were performed experiments in air at ambient pressure on metallic sample using optical single color pyrometers, at different wavelengths and temperatures. In particular, a sample of OFHC Copper was installed on a quartz beam and then inserted inside the graphite spherical cavity of a black body in the temperature range from 400 °C to 900 °C. A K-type thermocouple was installed in the sample to measure its temperature. This temperature was assumed as the temperature of the entire sample in the hypothesis of uniform temperature distribution. The hypothesis of uniform temperature distribution was assessed by a preliminary CFD analysis in steady state condition. By combining the temperatures measured by the pyrometers and by the thermocouple, the spectral emissivity of the sample’s surface as function of temperature at three different wavelengths was determined, maintaining fixed the emission angle. The authors have been provided a semi-empirical formula, obtained experimentally at CIRA’s laboratory for three different of pyrometer’s models, that allows to know the value of the emissivity ek;th of the sample’s surface for whatever emissivity value ek;SC set on the pyrometer. Furthermore, the ek;th was compared with the experimental emissivity value, ek;exp , that was evaluated directly by changing the emissivity value on the pyrometer until the temperature provided by pyrometer was equal to the ones provided by the thermocouple. The agreement between the experimental emissivity determined and the emissivity predicted by semi-empirical formula was verified.
2. Equipment used for the investigation To carry out the investigation, a blackbody with spherical cavity was used. As shown in Fig. 1, the sample was installed on a quartz holder with an inner pipe for the passage of a K-type thermocouple to measure the temperature of the sample. The blackbody is a MIKRON M305 with a temperature range of 100–1000 °C and error equal to ±0.2% RDG + 1 °C. The pyrometer system is consisting of one IMPAC device and two MIKRON ones. The IMPAC pyrometer is
the IGAR-12LO with focusable optics MB10, operating at k1 = 1.52 lm and k2 = 1.64 lm, at a temperature range 300–1000 °C, error equal to ±0.5% RDG + 1 °C. The pyrometer is controlled by a remote PC and operates both in single and dual color modes. The MIKRON pyrometers are both the M67S model operating in single color mode only, at a temperature range 400–800 °C with k = 1–1.16 lm, while the other is operating at temperature range 600–1000 °C with k = 0.78–1.06 lm; the error is equal to ±0.5% FS, or ±1 °C whichever is greater. The sample is of OFHC copper, with nominal surface emissivity equal to 0.80. In Fig. 2 (on the right side) it is possible to observe the presence of a little hole, £ 3 mm and 3 mm in depth located on the lower flat face, in order to allow the insertion of the K-type thermocouple for the measurement of its temperature. The K-type thermocouple measures in the range of temperature 200 to 1100 °C with an error of ±2.5 °C or 0.0075 ⁄ T (°C) and its wire is protected by insulation material. The end of the thermocouple is the sensor and it is at direct contact with the inner material of the sample. 3. Mathematical approach and experimental setup 3.1. Mathematical approach An infrared thermometer (IRT) consists of two parts, the optical system and the detector. The output of the detector may be different depending on the wavelength, and it is proportional to the amount radiated by the target at the specific wavelength. The calibration function of the thermometer, that is the thermometer output U, varies with the temperature and wavelength. U(T) is the integral of the Plank’s law that in the Wien’s hypothesis, can be written as:
Z UðTÞ ¼ kC 1
k2
k1
C2
k5 e kT dk
ð1Þ
with C1 the first radiation constant 3:7415 1016 W=m2 , C2 the second radiation constant 1:43879 102 m K, and k is a constant depending on the construction of the thermometer. At a single temperature or over a narrow range of temperatures, the calibration function may be expressed as: C2
UðTÞ ¼ ek kT LT
ð2Þ
where T is the temperature of the target with emissivity ek, and L = L(T) is a characteristic of the thermometer defined during the calibration. Now, let assume as ‘‘a” the ratio:
a¼
Fig. 1. Sample installed on the support.
C2 UðTÞ ¼ ek T LT k
ð3Þ
which is constant when the target is at a fixed temperature. Hence, by setting a blackbody at a fixed temperature TBB, and pointing the optical thermometer (pyrometer) at the center of the opening of the black body cavity, it is possible to measure the corresponding temperature at the given emissivity e = 0.995. So, the a(T) and L(T) functions
405
C. Purpura et al. / Measurement 82 (2016) 403–409
Fig. 2. The sample of standard materials used for the investigation and the hole for the installation of the K-thermocouple.
were obtained and a correlation function (Eq. (4)) was found:
ek ¼
a
ð4Þ
C
LT 2 T pyropyro
The Eq. (4) allows to obtain the emissivity ek as function of the temperature measured by the pyrometer (Tpyro). In order to evaluate the ek value in the temperature range considered, it was assumed L ¼ L that is the average value between two successive calibration points. So, being ‘‘a” constant at fixed temperature, the following equation can be written:
C2
a ¼ ek;th T ref ref LT
a ¼ ek;SC T SC
ð5Þ
C2
ð6Þ
LT SC
from which, being ‘‘a” constant, for each pyrometer considered can be written:
C2
C2
ek;th T ref ref ¼ ek;SC T SCLT SC LT
ð7Þ
where Tref is the temperature of reference measured by means a thermocouple installed inside the sample and ek,th is its actual emissivity. The semi-empirical formula (7), obtained experimentally at CIRA’s laboratory for three different pyrometer models, allows to know the actual emissivity value of the sample surface for whatever emissivity value ek;SC was set on the pyrometer. The semiempirical formula was obtained for the single-color pyrometer MIKRON M67/S 400–800 °C. Such method may be applied for every single color pyrometer. 3.2. Experimental setup To perform the investigation, the pyrometers system was installed on a tripod with micro heads for a fine pointing. The pyrometers are simultaneously pointing the sample at the same point as shown in Fig. 3. The sample is located inside the spherical cavity of the blackbody with the K-thermocouple installed. In this way, the convective and radiative fluxes coming from the black body, will heat the specimen. Then the Kthermocouple shall indicate the temperature (Tref/K) of the sample and each single color pyrometer shall indicate the temperature of sample’s surface (TSC/K). When the tem-
Fig. 3. Experimental setup of pyrometers, sample support system and blackbody.
perature becomes stable and the steady state condition is achieved, it is possible to match the temperature provided by the pyrometers with that of the thermocouple by tuning the emissivity value of each single color pyrometer. It is also possible to evaluate the proper emissivity by means pyrometer by applying the following semiempirical formula (8):
ek;th ¼ ek;SC
C2 LT SC
T SC
ð8Þ
C2 LT ref
T ref
where C2 is equal to 1.439 104 lm K, and L/lm is a constant different for every pyrometer and provided in Table 1. Table 1 Constant L (lm) of the semi-empirical formula. Pyrometer model
L (lm)
MIKRON M67S SC 400–800 °C MIKRON M67S SC 600–1000 °C IMPAC IGAR12-LO 300–1000 °C
8.75 5.66 9.15
C. Purpura et al. / Measurement 82 (2016) 403–409 Table 2 Comparison between the ek,exp and the ek,th emissivity values concerning the IGAR12-LO at k = 1.52 lm, [300, 1000 °C].
4. Test results: emissivity variation with the temperature To validate the proposed formula, some tests were carried out on a sample of OFHC copper material. The investigation was carried out for different temperatures for three different range of wavelengths. The sample was mounted on a homemade support and introduced inside the cavity of a blackbody. A K-thermocouple installed inside the sample measured the inner temperature of the sample at same time a pyrometers system measured the surface temperature of the sample. After the achievement the thermal equilibrium condition, the temperature of the sample was measured in order to characterize the material surface emissivity. The experimental emissivity value ek,exp was evaluated directly changing the emissivity value on pyrometer until the temperature provided by pyrometer matching the ones provided by thermocouple, instead the experimental/theoretical emissivity value ek,th provided by means formula (8), in which the temperature TSC was obtained setting pyrometers with emissivity value of 0.80. Fig. 4 shows the temperature values obtained: (i) at fixed value of emissivity chosen in literature (e = 0.80), (ii) experimentally; (iii) by means the semi-empirical formula. Furthermore, in Fig. 4, the experimental emissivity values, obtained by means the pyrometer ‘‘IMPAC IGAR12-LO 300–1000 °C” operating at the wavelength of k = 1.52 lm, are reported. In Table 2 the ek,exp and the emissivity ek,th values concerning the IGAR12-LO, are summarized at different temperature values. It is possible to note a good agreement between the semi-empirical and experimental emissivity values. In particular the largest percentage emissivity error is lower than 3% in correspondence of the temperature of the surface of 657 °C. In Fig. 5, the results of measurement performed by using the MIKRON pyrometer M67S in the range 400– 800 °C operating at wavelength between 1 and 1.6 lm
Single-Colour Temperature/ °C
1000 950
ελ,th=0.81 ελ,exp=0.81
900 850 800
ελ,th=0.79 ελ,exp=0.79
ελ,th=0.75 ελ,exp=0.74
750 700 650 600 550 500
ελ,th=0.62 ελ,exp=0.61
ε=0.80
ελ,th=0.66 ελ,exp=0.65
=
450
500
550
600
650
IGAR12-LO, SC (°C), fixed e = 0.80
ek,exp
ek,th
462 559 657 754 854
447 545 652 753 855
0.61 0.65 0.74 0.79 0.80
0.62 0.66 0.76 0.79 0.81
900 850
ελ,th=0.85 ελ,exp=0.85
800 750
ελ,th=0.77 ελ,exp=0.77
700 650 600 550
ελ,th=0.58 ελ,exp=0.60
500
ε=0.80
ελ,th=0.67 ελ,exp=0.67
450
ελ,exp ελ,th
400 400
450
500
550
600
650
700
750
800
K-Thermocouple/ °C Fig. 5. MIKRON M67S pyrometer at k = 1–1.6 lm [400, 800 °C]: temperature measurement comparison between the K-type thermocouple and the SC pyrometer by setting the proper material surface emissivity of the sample.
are shown. Fig. 5 shows the temperature values obtained: (i) at fixed value of emissivity chosen in literature (e = 0.80), (ii) experimentally; (iii) by means the semiempirical formula. In Table 3 the ek,exp and the emissivity ek,th values concerning the pyrometer M67S in the range 400–800 °C, are summarized at different temperature values. The two methods, semi-empirical and experimental, for the emissivity evaluation, show deviations at low temperature, in particular the largest percentage emissivity error is in correspondence of the temperature of the surface of 462 °C, where it is lower than 3%. It is possible to observe that the temperature profile, obtained with the fixed emissivity value at 0.80, exhibits a large deviation at low temperature of the surface and the deviation decreases as temperature increases. The application of the single color pyrometer MIKRON M67S 600–1000 °C shows a good agreement in the measurement of the temperature with the theoretical and experimental emissivity, as shown in Fig. 6. In Table 4
ελ,exp ελ,th
450 400 400
K-TC (°C)
Single-Colour Temperature/ °C
406
700
750
800
850
900
K-Thermocouple/ °C Fig. 4. IGAR 12-LO pyrometer at k = 1.52 lm, [300, 1000 °C]: temperature measurement comparison between the K-type thermocouple and the SC pyrometer by setting the proper material surface emissivity of the sample.
Table 3 Comparison between the ek,exp and the ek,th emissivity values concerning the M67S SC pyrometer at k = 1–1.6 lm [400–800 °C]. K-TC (°C)
M67S 400–800 (°C), fixed e = 0.80
ek,exp
ek,th
462 559 657 754
446 547 656 761
0.60 0.67 0.77 0.85
0.58 0.67 0.77 0.85
407
C. Purpura et al. / Measurement 82 (2016) 403–409
Single-Colour Temperature/ °C
900
of the sample in the black body cavity to obtain a uniform temperature distribution in the copper sample [6]. The grid used for the sample is composed by 53,709 tetrahedral elements with a spacing of 1 mm, while the surface has a spacing of 0.5 mm. The grid of the air inside the blackbody is composed by 89,767 tetrahedral elements with spacing, on the wall, of 5 mm. A parallelepiped air box, with dimension of 8 7 16.5 cm3, simulates the environmental. It was been assumed, for the quartz holder:
850
ελ,th=0.84 ελ,exp=0.84
ελ,th=0.85
800
ελ,exp=0.85
750 700
ελ,th=0.80 ελ,exp=0.80
ε=0.80 ελ,exp
650 600 600
ελ,th 650
700
750
800
850
900
– – – –
constant emissivity equal to 0.93; density q = 2200 kg/m3; specific heat at constant pressure cp = 670 J/(kg K); conductivity, k = 1.4 W/m K.
K-Thermocouple/ °C Fig. 6. MIKRON M67S pyrometer at k = 0.78–1.06 lm [600, 1000 °C]: temperature measurement comparison between the K-type thermocouple and the SC pyrometer by setting the proper material surface emissivity of the sample.
Table 4 Comparison between the ek,exp and the ek,th emissivity values concerning the MIKRON M67S pyrometer at k = 0.78–1.06 lm [600–1000 °C]. K-TC (°C)
M67S 600–1000 (°C) fixed e = 0.80
ek,exp
ek,th
657 754 854
659 758 858
0.80 0.85 0.84
0.80 0.85 0.84
Table 5
ek,exp variation of the material sample with k at fixed temperature. K-TC (°C)
k (lm)
ek,exp
462.3
1.30 1.52
0.60 0.61
559.2
1.30 1.52
0.67 0.65
657
0.92 1.30 1.52
0.80 0.77 0.74
754
0.92 1.30 1.52
0.85 0.85 0.79
854
0.92 1.52
0.84 0.80
the ek,exp and ek,th values concerning the pyrometer M67S in the range 600–1000 °C, are summarized at different temperature values. It is possible to note that the percentage emissivity error tends to zero in correspondence of temperature values larger than 650 °C. Hence, by means of the measurement performed on the copper sample, it is possible to realize the emissivity characterization as a function of the pyrometer wavelength. In the following Table 5 the experimental emissivity values obtained by the analysis are reported. 5. Numerical analysis A preliminary CFD analysis (by means commercial code FluentÒ [5]) was employed to evaluate the optimal position
Turbulent and radiative models – The k–x Standard model was used to predict the turbulence, and the motion of the air near the sample [5]. – For the radiative heat exchange has been used the Discrete–Ordinate model [5]. Boundary conditions – The cavity walls were modeled as wall at constant temperature, with an emissivity of 0.995. – Temperature of the black body cavity (TBB) is equal to 500 °C. – For the air box, the upper surface was modeled as ‘pressure-outlet’ while the others were modeled as ‘pressure-inlet’. – To the surface of the sample, the emissivity has been set as the result of experimental investigation of Table 5. – The air was assumed as ideal-gas. Mesh analysis – Fluid dynamic mesh analysis: Table 6 shows the results obtained for different mesh spacing. It was chosen the type A mesh characterized by 5.0 mm air grid spacing and 1.0 mm sample grid spacing. – Radiative mesh analysis: discretization refinement was performed to verify the discretization independent results. In Table 7 the discretization used for the
Table 6 Temperature of the sample for different mesh spacing. Mesh
Air grid spacing (mm)
Sample grid spacing (mm)
T (°C)
A B C
10.0 5.0 4.0
1.0 1.0 0.5
475 467 467
Table 7 Temperature of the sample for different discretization. Theta Division
Phi Division
Theta Pixels
Phi Pixels
T (°C)
2 6 8
2 6 8
1 3 5
1 3 5
478 467 467
408
C. Purpura et al. / Measurement 82 (2016) 403–409
simulation was reported. Mesh with Theta Division = Phi Division = 6 and Theta Pixels = Phi Pixels = 3 was chosen. In order to verify that the quartz holder doesn’t influence significantly the temperature’s field in the area where it’s placed the sample, a numerical analysis was carried out. The numerical analysis shows that the presence of the holder influences only the lower part of the cavity, while in the zone around the simple there is a maximum temperature difference of 7.0 °C (see Fig. 7). In particular in Fig. 7, the comparison between the examined cases with and without the quartz holder in terms of temperature fields in the cavity, was reported. The computational domain and its grid model are shown in Fig. 8.
Fig. 8. Grid model.
6. Numerical results Figs. 9 and 10 show, respectively, the temperature contour on the symmetry section when the temperature of the cavity (TBB) is equal to 500 °C and the temperature contours on the front and the symmetry side of the sample. It is possible to note that there are not significant temperature gradients in the measurement spot of pyrometers. In Fig. 11, temperature profiles on internal line of the sample, in function of the x-coordinate of the grid in the thickness of the sample, are shown. It is possible to note that the temperature value, in the zone closed to the sample, varies of 0.20 °C at most. For the other temperature, since that the experimental spectral emissivity values are more different, the model used for the radiative heat transfer include three gray bands. The entire thermal spectrum of wavelength (0.3–1000 lm) was divided in three or two gray bands. For each band the emissivity value was imposed equal to the experimental one.
Fig. 9. Temperature contours of the symmetry section for TBB = 500 °C.
In Table 8, a comparison between the numerical and experimental temperature values (Tsimul and K-TC, respectively) inside the sample are reported.
Fig. 7. Cavity temperature fields and temperature profiles along position from the center to the exit of the black body with and without the quartz holder.
C. Purpura et al. / Measurement 82 (2016) 403–409
409
The comparison shows that Tsimul values are always greater than the K-TC values; this is due to the assumption of the gray bands, in fact the copper exhibits an emissivity that reduces as the wavelength increases, while in the simulation the emissivity was imposed constant for the entire wavelength in the gray-bands. For all the TBB values considered, the temperature variations inside the sample are less than 0.5 °C, so the hypothesis of thermal equilibrium is numerically verified. 7. Conclusions
Fig. 10. Temperature on the front and symmetry face of the sample.
An experimental investigation has been performed to establish a methodology for the determination of the emissivity curve of a surface of material. Different optical pyrometers in the low infrared region have been used. A preliminary CFD analysis was carried out to obtain the optimal position of the sample in the black body cavity and to verify the hypothesis of stationary regime within the cavity: the thermal equilibrium of the sample was verified for different values of cavity temperature of the black body. The authors have provided a semi-empirical formula, obtained experimentally at CIRA’s laboratory for three different pyrometer models, that allows knowing the actual emissivity value of the sample surface whatever emissivity value set on the pyrometer. The agreement between the experimental emissivity determined and the emissivity predicted by semi-empirical formulas has been verified. The two methods, semi-empirical and experimental, for the emissivity evaluation, show deviations at low temperature, in particular the largest percentage e error lower than 3%. References
Fig. 11. Internal temperature profiles of the sample.
Table 8 Comparison between temperature simulated and measured. TBB/°C
K-TC/°C
Tsimul/°C
500 600 700 800 900
462 559 657 754 854
467 564 661 761 859
[1] C. Purpura, E. Trifoni, P. Barrera, G. De Gaetano, Experimental investigation of the emissivity of the expert flap in scirocco plasma wind tunnel tests, in: Proceedings of the 63rd International Astronautical Conference, 1–5 October, Naples, 2012. [2] G. Teodorescu, P.D. Jones, R.A. Overfelt, B. Guo, Normal emissivity of high-purity nickel at temperatures between 1440 and 1605 K, J. Phys. Chem. Solids 69 (1) (2008) 133–138. [3] P. Hagqvist, F. Sikstrom, A.K. Christiansson, Emissivity estimation for high temperature radiation pyrometry on Ti–6Al–4V, J. Int. Meas. Confed. 46 (2) (2013) 871–880. [4] D. Shi, F. Zou, S. Wang, Z. Zhu, J. Sun, B. Wang, Spectral emissivity modelling of steel 201 during the growth of oxidation film over the temperature range from 800 to 1100 K in air, Infrared Phys. Technol. 67 (2014) 42–48. [5] Fluent, User’s Guide, Fluent Inc, 2005. [6] J.H. Thomas, N.A. Amanie, Experimental and Numerical Characterization of a Steady-State Cylindrical Blackbody Cavity at 1100 °C, NASA, 2000 (August).
Contents lists available at ScienceDirect
Measurement journal homepage: www.elsevier.com/locate/measurement
Methodology for spectral emissivity measurement by means of single color pyrometer Carlo Purpura a, Eduardo Trifoni a, Marilena Musto b,⇑, Giuseppe Rotondo b, Roberto della Ragione b a b
CIRA Italian Aerospace Research Centre, Via Maiorise, 81043 Capua, Italy Dipartimento di Ingegneria Industriale, Università degli Studi di Napoli Federico II, P.le Tecchio, 80, 80125 Napoli, Italy
a r t i c l e
i n f o
Article history: Received 1 September 2015 Received in revised form 8 January 2016 Accepted 11 January 2016 Available online 14 January 2016 Keywords: Emissivity CFD Temperature measurement Single color pyrometer Semi-empirical formula
a b s t r a c t The application of non-intrusive optical devices, such as infrared pyrometers able to measure the temperature of surfaces, makes possible the evaluation of emissivity curve of the tested materials at different temperature values. In this paper the authors propose a methodology for the spectral emissivity measurement by means of a single color pyrometer providing a semi-empirical formula, obtained experimentally at CIRA’s laboratory. The semi-empirical formula allows to know the actual emissivity value of the sample’s surface for whatever emissivity value set up on the pyrometer. The agreement between the experimental emissivity and the emissivity predicted by semi-empirical formula was verified. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction The measurement of the temperature of the material’s surface solicited by thermal loads, has always been the target of experiments developed in laboratory. The development of optical devices operating in the infrared region of the spectrum has given the opportunity to measure temperatures up to several thousands of kelvin. In fact the use of non-contact devices such as pyrometers or thermocameras has given the possibility to measure temperatures of the material’s surface without the insertion of temperature sensors in the samples. The only problem occurring with this type of application is the knowledge of the material’s surface characteristics, in particular the emissivity. Purpura et al. [1] have performed a technique to determine the experimental emissivity of a material during a test in the PWT-SCIROCCO by comparing the temperature values obtained by a thermocamera operating with the ⇑ Corresponding author. Tel.: +39 081 7682290; fax: +39 0812390364. E-mail address: [email protected] (M. Musto). http://dx.doi.org/10.1016/j.measurement.2016.01.018 0263-2241/Ó 2016 Elsevier Ltd. All rights reserved.
known emissivity and by a dual color pyrometer. They have found an increasing of about the 10% of the material’s emissivity during the development of the test campaign. The use of a dual color pyrometer to determine the temperature of the sample was performed also by Teodorescu et al. [2], to measure the spectral emissivity of nickel by a Fourier transform infrared spectrometer. But in that analysis the sample was located in a vacuum chamber, so it was not considered the oxidation influence on emissivity. The introduction of dual color optical devices has allowed the temperature measurement by-passing the knowledge of the material’s emissivity and considering the material as a gray-body in the two operating wavelengths k. The use of thermocouples to calibrate a pyrometer to measure the temperature of a sample was performed by Hagqvist et al. [3], which determined that the temperature uncertainty was less than 2.5%. In that analysis the sample was located in air at ambient pressure so the effect of oxidation phenomena was considered. The use of thermocouples, and the effect of surface oxidation, were described also by Shi et al. [4] to affect the spectral emissivity.
404
C. Purpura et al. / Measurement 82 (2016) 403–409
To evaluate the spectral emissivity values, the authors were performed experiments in air at ambient pressure on metallic sample using optical single color pyrometers, at different wavelengths and temperatures. In particular, a sample of OFHC Copper was installed on a quartz beam and then inserted inside the graphite spherical cavity of a black body in the temperature range from 400 °C to 900 °C. A K-type thermocouple was installed in the sample to measure its temperature. This temperature was assumed as the temperature of the entire sample in the hypothesis of uniform temperature distribution. The hypothesis of uniform temperature distribution was assessed by a preliminary CFD analysis in steady state condition. By combining the temperatures measured by the pyrometers and by the thermocouple, the spectral emissivity of the sample’s surface as function of temperature at three different wavelengths was determined, maintaining fixed the emission angle. The authors have been provided a semi-empirical formula, obtained experimentally at CIRA’s laboratory for three different of pyrometer’s models, that allows to know the value of the emissivity ek;th of the sample’s surface for whatever emissivity value ek;SC set on the pyrometer. Furthermore, the ek;th was compared with the experimental emissivity value, ek;exp , that was evaluated directly by changing the emissivity value on the pyrometer until the temperature provided by pyrometer was equal to the ones provided by the thermocouple. The agreement between the experimental emissivity determined and the emissivity predicted by semi-empirical formula was verified.
2. Equipment used for the investigation To carry out the investigation, a blackbody with spherical cavity was used. As shown in Fig. 1, the sample was installed on a quartz holder with an inner pipe for the passage of a K-type thermocouple to measure the temperature of the sample. The blackbody is a MIKRON M305 with a temperature range of 100–1000 °C and error equal to ±0.2% RDG + 1 °C. The pyrometer system is consisting of one IMPAC device and two MIKRON ones. The IMPAC pyrometer is
the IGAR-12LO with focusable optics MB10, operating at k1 = 1.52 lm and k2 = 1.64 lm, at a temperature range 300–1000 °C, error equal to ±0.5% RDG + 1 °C. The pyrometer is controlled by a remote PC and operates both in single and dual color modes. The MIKRON pyrometers are both the M67S model operating in single color mode only, at a temperature range 400–800 °C with k = 1–1.16 lm, while the other is operating at temperature range 600–1000 °C with k = 0.78–1.06 lm; the error is equal to ±0.5% FS, or ±1 °C whichever is greater. The sample is of OFHC copper, with nominal surface emissivity equal to 0.80. In Fig. 2 (on the right side) it is possible to observe the presence of a little hole, £ 3 mm and 3 mm in depth located on the lower flat face, in order to allow the insertion of the K-type thermocouple for the measurement of its temperature. The K-type thermocouple measures in the range of temperature 200 to 1100 °C with an error of ±2.5 °C or 0.0075 ⁄ T (°C) and its wire is protected by insulation material. The end of the thermocouple is the sensor and it is at direct contact with the inner material of the sample. 3. Mathematical approach and experimental setup 3.1. Mathematical approach An infrared thermometer (IRT) consists of two parts, the optical system and the detector. The output of the detector may be different depending on the wavelength, and it is proportional to the amount radiated by the target at the specific wavelength. The calibration function of the thermometer, that is the thermometer output U, varies with the temperature and wavelength. U(T) is the integral of the Plank’s law that in the Wien’s hypothesis, can be written as:
Z UðTÞ ¼ kC 1
k2
k1
C2
k5 e kT dk
ð1Þ
with C1 the first radiation constant 3:7415 1016 W=m2 , C2 the second radiation constant 1:43879 102 m K, and k is a constant depending on the construction of the thermometer. At a single temperature or over a narrow range of temperatures, the calibration function may be expressed as: C2
UðTÞ ¼ ek kT LT
ð2Þ
where T is the temperature of the target with emissivity ek, and L = L(T) is a characteristic of the thermometer defined during the calibration. Now, let assume as ‘‘a” the ratio:
a¼
Fig. 1. Sample installed on the support.
C2 UðTÞ ¼ ek T LT k
ð3Þ
which is constant when the target is at a fixed temperature. Hence, by setting a blackbody at a fixed temperature TBB, and pointing the optical thermometer (pyrometer) at the center of the opening of the black body cavity, it is possible to measure the corresponding temperature at the given emissivity e = 0.995. So, the a(T) and L(T) functions
405
C. Purpura et al. / Measurement 82 (2016) 403–409
Fig. 2. The sample of standard materials used for the investigation and the hole for the installation of the K-thermocouple.
were obtained and a correlation function (Eq. (4)) was found:
ek ¼
a
ð4Þ
C
LT 2 T pyropyro
The Eq. (4) allows to obtain the emissivity ek as function of the temperature measured by the pyrometer (Tpyro). In order to evaluate the ek value in the temperature range considered, it was assumed L ¼ L that is the average value between two successive calibration points. So, being ‘‘a” constant at fixed temperature, the following equation can be written:
C2
a ¼ ek;th T ref ref LT
a ¼ ek;SC T SC
ð5Þ
C2
ð6Þ
LT SC
from which, being ‘‘a” constant, for each pyrometer considered can be written:
C2
C2
ek;th T ref ref ¼ ek;SC T SCLT SC LT
ð7Þ
where Tref is the temperature of reference measured by means a thermocouple installed inside the sample and ek,th is its actual emissivity. The semi-empirical formula (7), obtained experimentally at CIRA’s laboratory for three different pyrometer models, allows to know the actual emissivity value of the sample surface for whatever emissivity value ek;SC was set on the pyrometer. The semiempirical formula was obtained for the single-color pyrometer MIKRON M67/S 400–800 °C. Such method may be applied for every single color pyrometer. 3.2. Experimental setup To perform the investigation, the pyrometers system was installed on a tripod with micro heads for a fine pointing. The pyrometers are simultaneously pointing the sample at the same point as shown in Fig. 3. The sample is located inside the spherical cavity of the blackbody with the K-thermocouple installed. In this way, the convective and radiative fluxes coming from the black body, will heat the specimen. Then the Kthermocouple shall indicate the temperature (Tref/K) of the sample and each single color pyrometer shall indicate the temperature of sample’s surface (TSC/K). When the tem-
Fig. 3. Experimental setup of pyrometers, sample support system and blackbody.
perature becomes stable and the steady state condition is achieved, it is possible to match the temperature provided by the pyrometers with that of the thermocouple by tuning the emissivity value of each single color pyrometer. It is also possible to evaluate the proper emissivity by means pyrometer by applying the following semiempirical formula (8):
ek;th ¼ ek;SC
C2 LT SC
T SC
ð8Þ
C2 LT ref
T ref
where C2 is equal to 1.439 104 lm K, and L/lm is a constant different for every pyrometer and provided in Table 1. Table 1 Constant L (lm) of the semi-empirical formula. Pyrometer model
L (lm)
MIKRON M67S SC 400–800 °C MIKRON M67S SC 600–1000 °C IMPAC IGAR12-LO 300–1000 °C
8.75 5.66 9.15
C. Purpura et al. / Measurement 82 (2016) 403–409 Table 2 Comparison between the ek,exp and the ek,th emissivity values concerning the IGAR12-LO at k = 1.52 lm, [300, 1000 °C].
4. Test results: emissivity variation with the temperature To validate the proposed formula, some tests were carried out on a sample of OFHC copper material. The investigation was carried out for different temperatures for three different range of wavelengths. The sample was mounted on a homemade support and introduced inside the cavity of a blackbody. A K-thermocouple installed inside the sample measured the inner temperature of the sample at same time a pyrometers system measured the surface temperature of the sample. After the achievement the thermal equilibrium condition, the temperature of the sample was measured in order to characterize the material surface emissivity. The experimental emissivity value ek,exp was evaluated directly changing the emissivity value on pyrometer until the temperature provided by pyrometer matching the ones provided by thermocouple, instead the experimental/theoretical emissivity value ek,th provided by means formula (8), in which the temperature TSC was obtained setting pyrometers with emissivity value of 0.80. Fig. 4 shows the temperature values obtained: (i) at fixed value of emissivity chosen in literature (e = 0.80), (ii) experimentally; (iii) by means the semi-empirical formula. Furthermore, in Fig. 4, the experimental emissivity values, obtained by means the pyrometer ‘‘IMPAC IGAR12-LO 300–1000 °C” operating at the wavelength of k = 1.52 lm, are reported. In Table 2 the ek,exp and the emissivity ek,th values concerning the IGAR12-LO, are summarized at different temperature values. It is possible to note a good agreement between the semi-empirical and experimental emissivity values. In particular the largest percentage emissivity error is lower than 3% in correspondence of the temperature of the surface of 657 °C. In Fig. 5, the results of measurement performed by using the MIKRON pyrometer M67S in the range 400– 800 °C operating at wavelength between 1 and 1.6 lm
Single-Colour Temperature/ °C
1000 950
ελ,th=0.81 ελ,exp=0.81
900 850 800
ελ,th=0.79 ελ,exp=0.79
ελ,th=0.75 ελ,exp=0.74
750 700 650 600 550 500
ελ,th=0.62 ελ,exp=0.61
ε=0.80
ελ,th=0.66 ελ,exp=0.65
=
450
500
550
600
650
IGAR12-LO, SC (°C), fixed e = 0.80
ek,exp
ek,th
462 559 657 754 854
447 545 652 753 855
0.61 0.65 0.74 0.79 0.80
0.62 0.66 0.76 0.79 0.81
900 850
ελ,th=0.85 ελ,exp=0.85
800 750
ελ,th=0.77 ελ,exp=0.77
700 650 600 550
ελ,th=0.58 ελ,exp=0.60
500
ε=0.80
ελ,th=0.67 ελ,exp=0.67
450
ελ,exp ελ,th
400 400
450
500
550
600
650
700
750
800
K-Thermocouple/ °C Fig. 5. MIKRON M67S pyrometer at k = 1–1.6 lm [400, 800 °C]: temperature measurement comparison between the K-type thermocouple and the SC pyrometer by setting the proper material surface emissivity of the sample.
are shown. Fig. 5 shows the temperature values obtained: (i) at fixed value of emissivity chosen in literature (e = 0.80), (ii) experimentally; (iii) by means the semiempirical formula. In Table 3 the ek,exp and the emissivity ek,th values concerning the pyrometer M67S in the range 400–800 °C, are summarized at different temperature values. The two methods, semi-empirical and experimental, for the emissivity evaluation, show deviations at low temperature, in particular the largest percentage emissivity error is in correspondence of the temperature of the surface of 462 °C, where it is lower than 3%. It is possible to observe that the temperature profile, obtained with the fixed emissivity value at 0.80, exhibits a large deviation at low temperature of the surface and the deviation decreases as temperature increases. The application of the single color pyrometer MIKRON M67S 600–1000 °C shows a good agreement in the measurement of the temperature with the theoretical and experimental emissivity, as shown in Fig. 6. In Table 4
ελ,exp ελ,th
450 400 400
K-TC (°C)
Single-Colour Temperature/ °C
406
700
750
800
850
900
K-Thermocouple/ °C Fig. 4. IGAR 12-LO pyrometer at k = 1.52 lm, [300, 1000 °C]: temperature measurement comparison between the K-type thermocouple and the SC pyrometer by setting the proper material surface emissivity of the sample.
Table 3 Comparison between the ek,exp and the ek,th emissivity values concerning the M67S SC pyrometer at k = 1–1.6 lm [400–800 °C]. K-TC (°C)
M67S 400–800 (°C), fixed e = 0.80
ek,exp
ek,th
462 559 657 754
446 547 656 761
0.60 0.67 0.77 0.85
0.58 0.67 0.77 0.85
407
C. Purpura et al. / Measurement 82 (2016) 403–409
Single-Colour Temperature/ °C
900
of the sample in the black body cavity to obtain a uniform temperature distribution in the copper sample [6]. The grid used for the sample is composed by 53,709 tetrahedral elements with a spacing of 1 mm, while the surface has a spacing of 0.5 mm. The grid of the air inside the blackbody is composed by 89,767 tetrahedral elements with spacing, on the wall, of 5 mm. A parallelepiped air box, with dimension of 8 7 16.5 cm3, simulates the environmental. It was been assumed, for the quartz holder:
850
ελ,th=0.84 ελ,exp=0.84
ελ,th=0.85
800
ελ,exp=0.85
750 700
ελ,th=0.80 ελ,exp=0.80
ε=0.80 ελ,exp
650 600 600
ελ,th 650
700
750
800
850
900
– – – –
constant emissivity equal to 0.93; density q = 2200 kg/m3; specific heat at constant pressure cp = 670 J/(kg K); conductivity, k = 1.4 W/m K.
K-Thermocouple/ °C Fig. 6. MIKRON M67S pyrometer at k = 0.78–1.06 lm [600, 1000 °C]: temperature measurement comparison between the K-type thermocouple and the SC pyrometer by setting the proper material surface emissivity of the sample.
Table 4 Comparison between the ek,exp and the ek,th emissivity values concerning the MIKRON M67S pyrometer at k = 0.78–1.06 lm [600–1000 °C]. K-TC (°C)
M67S 600–1000 (°C) fixed e = 0.80
ek,exp
ek,th
657 754 854
659 758 858
0.80 0.85 0.84
0.80 0.85 0.84
Table 5
ek,exp variation of the material sample with k at fixed temperature. K-TC (°C)
k (lm)
ek,exp
462.3
1.30 1.52
0.60 0.61
559.2
1.30 1.52
0.67 0.65
657
0.92 1.30 1.52
0.80 0.77 0.74
754
0.92 1.30 1.52
0.85 0.85 0.79
854
0.92 1.52
0.84 0.80
the ek,exp and ek,th values concerning the pyrometer M67S in the range 600–1000 °C, are summarized at different temperature values. It is possible to note that the percentage emissivity error tends to zero in correspondence of temperature values larger than 650 °C. Hence, by means of the measurement performed on the copper sample, it is possible to realize the emissivity characterization as a function of the pyrometer wavelength. In the following Table 5 the experimental emissivity values obtained by the analysis are reported. 5. Numerical analysis A preliminary CFD analysis (by means commercial code FluentÒ [5]) was employed to evaluate the optimal position
Turbulent and radiative models – The k–x Standard model was used to predict the turbulence, and the motion of the air near the sample [5]. – For the radiative heat exchange has been used the Discrete–Ordinate model [5]. Boundary conditions – The cavity walls were modeled as wall at constant temperature, with an emissivity of 0.995. – Temperature of the black body cavity (TBB) is equal to 500 °C. – For the air box, the upper surface was modeled as ‘pressure-outlet’ while the others were modeled as ‘pressure-inlet’. – To the surface of the sample, the emissivity has been set as the result of experimental investigation of Table 5. – The air was assumed as ideal-gas. Mesh analysis – Fluid dynamic mesh analysis: Table 6 shows the results obtained for different mesh spacing. It was chosen the type A mesh characterized by 5.0 mm air grid spacing and 1.0 mm sample grid spacing. – Radiative mesh analysis: discretization refinement was performed to verify the discretization independent results. In Table 7 the discretization used for the
Table 6 Temperature of the sample for different mesh spacing. Mesh
Air grid spacing (mm)
Sample grid spacing (mm)
T (°C)
A B C
10.0 5.0 4.0
1.0 1.0 0.5
475 467 467
Table 7 Temperature of the sample for different discretization. Theta Division
Phi Division
Theta Pixels
Phi Pixels
T (°C)
2 6 8
2 6 8
1 3 5
1 3 5
478 467 467
408
C. Purpura et al. / Measurement 82 (2016) 403–409
simulation was reported. Mesh with Theta Division = Phi Division = 6 and Theta Pixels = Phi Pixels = 3 was chosen. In order to verify that the quartz holder doesn’t influence significantly the temperature’s field in the area where it’s placed the sample, a numerical analysis was carried out. The numerical analysis shows that the presence of the holder influences only the lower part of the cavity, while in the zone around the simple there is a maximum temperature difference of 7.0 °C (see Fig. 7). In particular in Fig. 7, the comparison between the examined cases with and without the quartz holder in terms of temperature fields in the cavity, was reported. The computational domain and its grid model are shown in Fig. 8.
Fig. 8. Grid model.
6. Numerical results Figs. 9 and 10 show, respectively, the temperature contour on the symmetry section when the temperature of the cavity (TBB) is equal to 500 °C and the temperature contours on the front and the symmetry side of the sample. It is possible to note that there are not significant temperature gradients in the measurement spot of pyrometers. In Fig. 11, temperature profiles on internal line of the sample, in function of the x-coordinate of the grid in the thickness of the sample, are shown. It is possible to note that the temperature value, in the zone closed to the sample, varies of 0.20 °C at most. For the other temperature, since that the experimental spectral emissivity values are more different, the model used for the radiative heat transfer include three gray bands. The entire thermal spectrum of wavelength (0.3–1000 lm) was divided in three or two gray bands. For each band the emissivity value was imposed equal to the experimental one.
Fig. 9. Temperature contours of the symmetry section for TBB = 500 °C.
In Table 8, a comparison between the numerical and experimental temperature values (Tsimul and K-TC, respectively) inside the sample are reported.
Fig. 7. Cavity temperature fields and temperature profiles along position from the center to the exit of the black body with and without the quartz holder.
C. Purpura et al. / Measurement 82 (2016) 403–409
409
The comparison shows that Tsimul values are always greater than the K-TC values; this is due to the assumption of the gray bands, in fact the copper exhibits an emissivity that reduces as the wavelength increases, while in the simulation the emissivity was imposed constant for the entire wavelength in the gray-bands. For all the TBB values considered, the temperature variations inside the sample are less than 0.5 °C, so the hypothesis of thermal equilibrium is numerically verified. 7. Conclusions
Fig. 10. Temperature on the front and symmetry face of the sample.
An experimental investigation has been performed to establish a methodology for the determination of the emissivity curve of a surface of material. Different optical pyrometers in the low infrared region have been used. A preliminary CFD analysis was carried out to obtain the optimal position of the sample in the black body cavity and to verify the hypothesis of stationary regime within the cavity: the thermal equilibrium of the sample was verified for different values of cavity temperature of the black body. The authors have provided a semi-empirical formula, obtained experimentally at CIRA’s laboratory for three different pyrometer models, that allows knowing the actual emissivity value of the sample surface whatever emissivity value set on the pyrometer. The agreement between the experimental emissivity determined and the emissivity predicted by semi-empirical formulas has been verified. The two methods, semi-empirical and experimental, for the emissivity evaluation, show deviations at low temperature, in particular the largest percentage e error lower than 3%. References
Fig. 11. Internal temperature profiles of the sample.
Table 8 Comparison between temperature simulated and measured. TBB/°C
K-TC/°C
Tsimul/°C
500 600 700 800 900
462 559 657 754 854
467 564 661 761 859
[1] C. Purpura, E. Trifoni, P. Barrera, G. De Gaetano, Experimental investigation of the emissivity of the expert flap in scirocco plasma wind tunnel tests, in: Proceedings of the 63rd International Astronautical Conference, 1–5 October, Naples, 2012. [2] G. Teodorescu, P.D. Jones, R.A. Overfelt, B. Guo, Normal emissivity of high-purity nickel at temperatures between 1440 and 1605 K, J. Phys. Chem. Solids 69 (1) (2008) 133–138. [3] P. Hagqvist, F. Sikstrom, A.K. Christiansson, Emissivity estimation for high temperature radiation pyrometry on Ti–6Al–4V, J. Int. Meas. Confed. 46 (2) (2013) 871–880. [4] D. Shi, F. Zou, S. Wang, Z. Zhu, J. Sun, B. Wang, Spectral emissivity modelling of steel 201 during the growth of oxidation film over the temperature range from 800 to 1100 K in air, Infrared Phys. Technol. 67 (2014) 42–48. [5] Fluent, User’s Guide, Fluent Inc, 2005. [6] J.H. Thomas, N.A. Amanie, Experimental and Numerical Characterization of a Steady-State Cylindrical Blackbody Cavity at 1100 °C, NASA, 2000 (August).