J. Ind. Eng. Chem., Vol. 13, No. 7, (2007) 1122-1130 A Study on the Design of the Solar Lamp Bank of Metal Halide-Infrared Lamps Combination Method Sang Hwa Baek, Jeong Yong Lee, So Jin Park*,, Sung Dae Kim**, and Han Seo Ko** Defense Systems Test Center, Agency for Defense Development, Chungnam 357-900, Korea *Department of Chemical Engineering, Chungnam National University, Daejeon 305-764, Korea **Department of Mechanical Engineering, Sungkyunkwan University, Suwon 446-746, Korea Received July 18, 2007; Accepted September 28, 2007 Abstract: A solar lamp bank, a very important part of a solar simulator, has been developed by combining commercial metal halide lamps and infrared lamps. A program for a design of a lamp bank based on a point light source theory has also been developed in this study. The reliability of the program for lamp bank design has been verified by radiance experiments with a lamp for horizontal distances. The perfect solar lamp bank was developed according to the test guidelines. A ceiling board shape put next to the lamp bank has been designed to promote cooling efficiency of the lamp and to reduce the temperature and air velocity deviations in the chamber. It has been confirmed that the partially closed type ceiling board was the best when compared to several other types of ceiling boards. Keywords: lamp bank, radiance, ceiling board, uniformity Introduction 1) The solar lamp bank is the core part of the solar simulator that radiates simulated sunshine. For the environmental test, simulated sunshine caused by the solar lamp should be as similar to real sunshine as possible. This solar lamp bank was invented with the basis of the maximum radiation of the summer season according to the guidelines of the solar radiation test [1,2]. The existing lamp bank of the car and air handling industrial circles use only an infrared lamp, and the existing lamp bank of the clothing test institution use only an ultra violet lamp shinning from a different direction. This is not a real solar simulator because this lamp radiates only ultraviolet or infrared rays. The only solar lamp bank is an ark-xenon lamp. However, it has a limited radiation area, is expensive, and reproduces badly. In order to solve these problems, we developed the solar lamp bank with mercury and a halogen lamp during 1997~2002. But it had a couple of faults : the lamp went off during To whom all correspondence should be addressed. (e-mail: sjpark@cnu.ac.kr) the test, and the radiation was unstable. It was designed after considering the estimated lamp position, the radiation value of each lamp, the radiation value of the entire lamp bank, and the field uniformity using metal halide and the infrared lamp. The Theoretical Background (1) The radiation values of the radiation area correspond to the lighting theory of the equation (1) with a source of light [3,4]. I is intensity of light, P is the distance between the lamp and the target area and θ is the entering angle of the lamp light. The radiation values of the radiation area are decided by the angle the lamp is shining from since the temperature is different according to the altitude of sun. The lamp bank was designed as both a plane and a semicircle. The lamp bank can be a hemisphere so the lamp position was calculated on an oval bank. One point on the oval bank is expressed by equation (2).
A Study on the Design of the Solar Lamp Bank of Metal Halide-Infrared Lamps Combination Method 1123 Figure 1. Incidence angle between light (lamp) and target area. (2) Figure 2. Relation between metal halide lamp radiance and horizontal distance. For the arrangement of the lamp bank on an oval solid using a rectangular coordinate, there are two restricted conditions. The first is that the core coordinate is (0, 0, 0), a section of the lamp bank is a circle if a is equal to b while the core coordinate is (0, 0, z 1 ) if c is equal to H 0. In other words, an axis of z passes the core of the lamp bank. H 0 is the distance between the core of the item to be tested and the core of the lamp bank. The second is that the core coordinates are (0, 0, z 2 ) of the tested item. z coordinates can be calculated with initial condition values of (0, 0, 0) from equation (2). (3) Experimental The Experimental Equipment and the Method After a pyranometer and a light measuring instrument are located in the middle of the target area, the lamp bank is installedinto a 3 meter spot over the target area, and the intensity of radiation is measured by a pyranometer when the lamp is migrated up and down [5-7]. We can know the intensity of the radiation according to the distance [8,9]. The effects of piling one on another and being turned off can not be known because of the character of the lamp, the light from the face of the wall, the control temperature of the heat, and the lamp socket base, etc. Therefore, we gather information about a single lamp and compare it with the value to be estimated by a design program. Figure 3. Relation between infrared lamp radiance and horizontal distance. Modeling Verification Figure 2 shows the correlation between metal halide lamp radiance and the horizontal distance. The radiance decreases to 1.5~0.3 W/m 2 because of the horizontal distance migration from the core of the radiation area. The value to be simulated is in accordance with the experimental value. Figure 3 shows the correlation between the infrared lamp radiance and the horizontal distance radiance. The radiance decreases to 35~8 W/m 2 because of the horizontal distance migration from core of the radiance area. The value to be simulated agrees with the experimental value.
1124 Sang Hwa Baek, Jeong Yong Lee, So Jin Park, Sung Dae Kim, and Han Seo Ko Table 1. Spectral Energy Distribution and Permitted Tolerance Characteristic Bandwidth Spectral Region Ultraviolet Visible Infrared 280~ 400 nm 400~ 780 nm 780~ 3000 nm Sum 280~ 3000 nm Irradiance 68 W/m 2 560 W/m 2 492 W/m 2 1120 W/m 2 Tolerance ± 35 % ± 10 % ± 20 % ± 10 % Figure 4. Side picture of solar simulator. Results and Discussion The existing solar lamp bank went out. The origin is three. There are three reasons. The first reason is that most of the halogen type lamps have from five to fifteen times more radiation energy than a visible ray even if the commercial halogen lamp has to radiate 560 W/m 2 on a wavelength domain of 400~780 nm and 492 W/m 2 on a wavelength domain of 780~3000 nm [10]. So the energy of an infrared wavelength had to be decreased because of the infrared filter, however it circulated in the test chamber, and caused the surface temperature of the lamp to rise which caused the lamp to go off. The cooler had to be installed in order to cool the energy to be filtered. The second reason is the problem of the lamp s array. The lamp array decides if the durability is shortened or lengthened. The existing solar lamp bank chose the way that the infrared lamp was concentrated. This way caused radiant energy to concentrate on the center, so was the source that caused the lamp lights to go off. This invented lamp bank was designed to reduce the possibility that the lamp would go off and to satisfy the demand of the test regulation with the radiation value according to the wavelength domain. The last reason is to undo the remainder of the test chamber ceiling. The air from the air handling unit makes the temperature of the lamp descend. But much of the cooling air in the ceiling of the existing test chamber didn t bump against the lamp. If most of air can be sent to the lamp to make the temperature of the lamp descend, the durability of the lamp can be prolonged. Because the remainder of this time test chamber ceiling was designed to fill up the plate of the hole, much of the cooling air would bump against the lamp. The most suitable hole size of the plate has to Figure 5. Spectral radiance of the metal halide ultraviolet lamp (150 W). be saught to satisfy the demand air velocity principle. In Figure 4 handls the air movement from A (the evaporator) to B (the fan), from B to C (the ceiling of the test chamber), and from C to D (the lamp bank) or E (the ceiling board). Because of the make of the E part-closing shape, the cooling air from the air handling unit can be concentrated on the lamp bank. The design of the ceiling plate around the lamp bank was made considering the cooling efficiency and the temperature stability of the lamp. The temperature sensor is located under the table. The temperature distribution in the test chamber should be uniform. The Layout of the Lamp Bank The lamp bank is D part of Figure 4. We had to choose lamps which would be installed in the lamp bank prior to designing the lamp bank. The lamp bank has to satisfy the requirement of Table 1 the spectral energy distribution and permitted tolerance [1,2,10]. The ultraviolet radiance covers 6.1 % of the total maximum radiance. The visibleradiance covers 50 %. The infrared radiance covers 6.1 %. There is nothing to satisfy Table 1 the spectral energy distribution and the
A Study on the Design of the Solar Lamp Bank of Metal Halide-Infrared Lamps Combination Method 1125 Figure 6. Spectral irradiance of the metal halide visible lamp (150 W). Figure 8. Lay-out of the infrared lamps (60 W) concentration method ( infrared lamp, metal halide visible lamp, metal halide UV lamp). has the capability to reduce the temperature of the lamps surface slightly by using the metalhalide visible lamps. The Middle Concentration Design of a 60 W Infrared Lamp Using the point light source theory (1), the lamp bank in Figure 8 emits the ultraviolet energy of 75.26 W/m 2, the visible energy of 594.9 W/m 2 and infrared of 556.9 W/m 2. But there are the many unused lamp sockets in the lamp bank, which have the demerits to reduce the cooling effect of the lamps. Because infrared lamps and visiblelamps concentrate on the middle of the lamp bank, it is very difficult to distribute the heat of the lamp bank. This type also has a small target area. This lamp bank requires ultraviolet lamps of 12 numbers, visible lamps of 32 numbers and infrared lamps of 12 numbers. Figure 7. Spectral irradiance of the infrared lamp (40 W). permitted tolerance among the commercial lamps in Korea. Therefore, considering Table 1, we have to choose each lamp [11]. Figures 5, 6 and 7 are the spectral radiance graphs of each lamp. The metal halide visible lamp has more spectral energy of the visible wavelength than the spectral energy of the infrared wavelength. Halogen lamps which we used before relatively include more infrared energy than visible energy. So we cut down the infrared energy with infrared absorption filters. But this method had a reverse effect : the infrared energy cut down by infrared absorption filters raised the temperature of the lamps surfaces again. This reverse effect shortened the life time of the lamps. But the lamp bank developed at this time The Middle Concentration Design of a 25 W Infrared Lamp Using the point light source theory (1), the lamp bank of Figure 9 emits the ultraviolet energy of 84.37 W/m 2, the visible energy of 623.7 W/m 2 and the infrared of 598.6 W/m 2. But there are a few unused lamp sockets in the lamp bank, which just have demerits to reduce the cooling effect of the lamps. Because infrared lamps concentrate on the middle of the lamp bank, it is very difficult to distribute the heat resulted in the lamp bank. This lamp bank requires ultraviolet lamps of 12 numbers, visible lamps of 38 numbers and infrared lamps of 33 numbers. The Combination Layout of a 60 Watt Infrared Lamp and a Visible Lamp Using the point light source theory (1), the lamp bank of Figure 10 emits the ultraviolet energy of 110.93 W/m 2,
1126 Sang Hwa Baek, Jeong Yong Lee, So Jin Park, Sung Dae Kim, and Han Seo Ko Figure 9. Lay-out of the infrared lamps (25 W) concentration method ( infrared lamp, metal halide visible lamp, metal halide UV lamp). Figure 11. Lay-out of the infrared lamps (40 W) combination method ( infrared lamp, metal halide visible lamp, metal halide UV lamp). respective lamps radiance values satisfy those demanded by test guidelines. Finally, we made sure that this method is adequate for a solar simulator. This lamp bank requires ultraviolet lamps of 12 numbers, visible lamps of 37 numbers and infrared lamps of 50 numbers. Figure 10. Lay-out of the infrared lamps (60 W) combination method ( infrared lamp, metal halide visible lamp, metal halide UV lamp). the visible energy of 637.45 W/m 2 and the infrared rays of 611.94 W/m 2. But the merits of this method are that unused lamp sockets are not on the lamp bank, the respective lamps radiance values don t satisfy those demanded by the test guidelines. Therefore, we have to reduce the wattage of infrared lamps. The lamps combination method is satisfied. This lamp bank requires ultraviolet lamps of 16 numbers, visible lamps of 33 numbers and infrared lamps of 50 numbers. The Combination Layout of a 40 Watt Infrared Lamp and a Visible Lamp The lamp bank of Figure 11 emits the ultraviolet energy of 88.99 W/m 2, the visible energy of 658.57 W/m 2 and infrared of 550.34 W/m 2. The merits of these methods that unused lamp sockets are not on the lamp bank. The The Ceiling Board Design The ceiling Board is E part of Figure 4. The ceiling board should be designed. Considering the lamps cooling effect and temperature uniformity in the solar simulator. The air velocity coming out of the air handling unit is high to make the cooling effect of the lamp bank good, but the air velocity in the solar simulator is not as high to make the temperature uniformity good. Therefore, the most suitable air velocity should be found outby calculations and experiments, because it is impossible to measure the cooling effects of the lamps installed in the lamp bank. We should analyze the lamps cooling effect with modeling and simulation software we develop. The temperature uniformity was examined with a temperature and velocity sensor for a definite period time. The main points of the temperature uniformity are the average temperature, the standard deviation of the temperature, the average velocity, and the standard deviation of the velocity of the conditioned air in the solar simulator. The Analysis of the Ceiling Board Design The fully closed type ( A type) is a type of flat panel without holes. This type only contains the conditioned air flow in the direction of the lamp bank, so the cooling effect of the lamps is the better than with other types. But this type may make the air velocity very fast in the chamber, and so it makes the temperature condition in the chamber unstable. The merit of the fully opened type ( B type) is that the temperature condition in the
A Study on the Design of the Solar Lamp Bank of Metal Halide-Infrared Lamps Combination Method 1127 Figure 12. The air flow at partially closed type ( C type). chamber is stable. But because the air from the air conditioning unit is uniformly distributed to the lamp bank and the ceiling board, the cooling effect of the lamp bank is bad in comparison with other types. Finally, the special features of the partially closed type ( C type) are evenly distributed between the fully closed type ( A type) and the fully opened type ( B type). The environmental test procedures demand that the air velocity in the chamber is between 0.25 and 1.5 m/s, the air temperature deviation in the chamber is ±1.0 o C. The ceiling board should be designed in terms of satisfying these conditions. In advance, we put M&S around the air flow in the chamber to design the ceiling board. Following the Figures 12, 13 and 14, The air flow of the partially closed type ( C type) is generally uniform in the chamber. The air flow of the fully closed type ( A type) is focusing on the other side of the fan. The air flow of the fully opened type ( B type) generally distributes to all positions in the chamber. The Result of the Ceiling Board Design The air flow out of the air handling unit is not turbulent (5 10 5 < Re < 10 7 ) but laminar [12,13]. With the fully closed type ( A type), the Reynolds number of the air through the lamp bank is 3755. With the fully opened type ( B type), the Reynolds number of the air through the lamp bank is 1140. Therefore, because the range of the Reynolds number is 1140 < Re < 3755, the air flow out of the air conditioning unit is just laminar. We assume that the air flow out of the air conditioning unit is steady, incompressible with constant fluid properties and negligible viscous dissipation. The hydrodynamic solution may be solved by the method of Blasius [13,14]. Finally, the local convection coefficient is expressed with the function with Reynolds number and the Prandtl Figure 13. The air flow at fully closed type ( A type). Figure 14. The air flow at fully opened type ( B type). number [14]. Closely examining the cooling effect of each lamp, we didn t take advantage of the average convection coefficient, but the local convection coefficient. (4) The temperature of the air out of the air conditioning unit is 322 K. The density ρ is 1.0782 kg/m 3, the viscosity µ is 196.4 10-7 N s/m 2, the heat transfer coefficient K is 28.15 10-3 W/m K, the Prandtl number is 0.7035, the air velocity is V 1 is 1.5 m/s, the surface area is A 1 is 0.87 m 2. We can see the heat properties of the air flow in the lamp bank and the air flow out the lamp bank by using Bernoulli s equation [15,16]. With the fully closed type ( A type), because the conditioned air flows in the direction the lamp bank only, the area of air out-flow is small and the air velocity
1128 Sang Hwa Baek, Jeong Yong Lee, So Jin Park, Sung Dae Kim, and Han Seo Ko Table 2. Lamp Cooling Efficiency Types V 2 (m/s) A 2 (m 2 ) Re (ρvx/µ) x (m) A 0.570 2.28 3,755 0.12 4.244 B 0.173 7.54 1,140 0.12 2.388 C 0.363 3.60 2,388 0.12 3.385 h T s (K) A x (m 2 ) Q (W) M type 363 7.68 0.04411 I type 383 11.42 M type 363 4.32 0.04411 I type 383 6.43 M type 363 6.12 0.04411 I type 383 9.11 Figure 15. Comparison of lamp cooling efficiency according to the shape of ceiling board. is fast. With the fully opened type ( B type), the area of air out-flow is big and the air velocity is slow. With the partially closed type ( C type), the summation of a ø5 mm hole area is the total area of the holes in which the air passes through the lamp bank. The lamp length, including the socket base is 0.12 m. The temperature of metal halide lamp is 363 K during the test. The temperature of the infrared lamp is 383 K. We calculated the cooling effects of a lamp according to the ceiling board types [17-19]. As shown in Table 2 and Figure 15, we examined that the cooling effect of a lamp is the highest with the fully closed type ( A type) and the lowest with the fully opened type ( B type). If we consider the metal halide lamp as a reference (100 %) with the fully closed type ( A type), the cooling effect of the metal halide lamp with the fully opened type ( B type) is 56.3 %. And if we consider that with the partially closed type ( C type) it is 79.8 %. The next step we should determine is the shape of the most suitable ceiling board. We should consider the relationship between the cooling effect and the temperature distribution/air velocity in the chamber. Figure 16. Temperature deviations according to the shapes of ceiling board. When the maximum radiance was maintained stably, the temperature and air velocity were measured 20 times every 5 min. The temperature distribution and air velocity of each shape is given in Table 3. As shown in Table 3, Figures 16 and 17, the temperature deviation of A type is 0.95, which is a high level similar to the limitation value. The deviation of air velocity also is high as compared with other types. The test procedure demands that the temperature deviation should be within ±1 o C. The temperature deviation of B type is 0.33 and the deviation of air velocity also is low, but the cooling effect is very low. The C type has low temperature deviation and velocity deviation. The C type also has a good cooling effect as compared with the B type. Hence, finally, we designed the partially closed type (the C type, 5 mm holes). Conclusion As all the domestic solar lamp banks depend on import, we are sure that the technology of the lamp bank design
A Study on the Design of the Solar Lamp Bank of Metal Halide-Infrared Lamps Combination Method 1129 Figure 17. Air velocity deviations according to the shapes of ceiling board. Table 3. Comparison of Air Temperature and Velocity in the Chamber Types Q (W) T avg (K) T dev (K) V avg (m/s) V dev (m/s) A 11.42 321.9 0.95 0.36 0.07 B 6.43 322.4 0.33 0.12 0.04 C 9.11 322.2 0.50 0.21 0.04 is not at all in the country. Because Lamp banks imported only consist of infrared lamps, we can t call those solar lamp banks for certain. The costly arc-xenon lamp banks which advanced countries have been adopting at present and infrared lamp banks in the country could be replaced by the solar lamp bank developed at this time. We can cut down the phenomenon of the lamps off which took place during the test with the lamp bank of the ultraviolet-halogen combination method we developed from 1997 to 2002. We can also complement the demerits of difficulty of heat distribution and excess of the infrared energy [20,21]. At the beginning of the research, we applied the point light source theory to the lamp bank design technology. Because we deeply perceived the problems of the heat distribution and radiance uniformity on the target area which the lamp bank of the past infrared lamps concentration method includes, we designed the lamp bank of the combination method of metal halide-infrared lamps. This lamp bank requires the ultraviolet lamp of 12 numbers, visible lamp of 37 numbers and infrared lamp of 50 numbers. Finally, to raise the cooling effect of the lamps installed the lamp bank, we designed the ceiling board. When we designed the ceiling board, we researched the relation between the heat cooled by forced convection and the temperature/air velocity in the chamber. The most suitable ceiling board is the partially closed type. With the partially closed type, the cooled heat value of an infrared lamp is 9.11 W, and that of a metal halide lamp is 6.12 W. The temperature deviation in the chamber is 0.50 K. The deviation of the air velocity in the chamber is 0.04 m/s. The fully closed type has a relatively good cooling effect, but the temperature deviation in the chamber is close to the test requirement ±1 o C. Also, being compared with other types, this type was examined and we found that the deviation of the air velocity in the chamber is very high. With the fully opened type ( B type), the temperature deviation in the chamber is 0.33 o C. The deviation of the air velocity in the chamber is 0.04 m/s. The deviation of the temperature and the air velocity is good, but the cooling effect of the lamps is very low in the comparison with other types [21-23]. So as this type can make the life time of the lamps short, we didn t choose this type. References 1. DTC, MIL-STD-810F (Method 505.4), Department of Defense (2000). 2. US ARMY Test and Evaluation Command, ITOP 4-2-826 (Solar Radiation Tests), US Army Aberdeen Proving Ground (1983). 3. C. M. Lee, K. M. Park, and J. H. Do, Electricity Application, pp. 11~19, Taeyoung, Korea (2006). 4. K. S. Lee, K. Y. Jung, and D. H. Lee, Electricity Application, pp. 26~29, Taeyoung, Korea (2007). 5. F. Grum and R. J. Becherer, Optical Radiation, pp. 235~270, Academic Press, New York (1983). 6. R. Danial Overheim and David L. Waguer, Light and Color, pp. 110~115, John Wiley & Sons, New York (1982). 7. Hecht, Optics, Addison-Wesely, p. 53, New York (1987). 8. W. Budde, Optical Radiation Measurements, pp. 10~45, Academic Press, New York (1983). 9. F. Grum and R. J. Becherer, Optical Radiation, pp. 15~63, Academic Press, New York (1983). 10. S. H. Baek, J. Korean Ind. Eng. Chem., 10, 1147 (1999). 11. J. F. Waymouth, Very High Output Lamp, The Sylvania Technologist, pp. 102~110 (1961). 12. Donald R. Pitts, Theory and Problems of Heat Transfer, Mcgraw-Hill, pp. 144~183 (1997). 13. R. Byron Bird, Transport Phenomena, John Wiley & Sons, pp. 291~297, New York (1967). 14. Frank P. Incropera, Fundamentals of Heat and Mass Transfer, John Wiley & Sons, pp. 353~378, pp. 428~
1130 Sang Hwa Baek, Jeong Yong Lee, So Jin Park, Sung Dae Kim, and Han Seo Ko 444, New York (1996). 15. I. S. Kim, Ph. D. Thesis, Convergence Aceleration by Multigrid Methode on Hypersonic Flow Analysis through Ramp, Seoul National University (2002). 16. Korean Institute of Chemical Engineers, Energy Engineering, Kyobo-book center, p. 603, Korea (1996). 17. H. J. Park, Ph.D. Thesis, Study on the Effect of the Ambient Temperature on the Luminous Flux of Bulb Type Flourescent Lamp, Yonsei University (1989). 18. G. H. Jo, Ph.D. Thesis, Development of Hot Restarting Ignition Circuits to Expand HID Lamp Life, Kangwon National University (2004). 19. K. Y. Kim, Thesis, The Characteristic Evaluation of Thermal Fatigue for Tungsten Filament, Kyung Hee University (2002). 20. D. A. Spera, What is thermal fatigue?, ASTM STP 612, pp. 3~9 (1976). 21. L. F. coffin and R. P. Wesley, An Apparatus for the Study of the Effects of Cyclic Thermal Stress on Ductile Metals, Trans. of ASME, 76, 923 (1954). 22. A Nitta and K. Kuwbara, Thermal-Mechanical Fatigue Failure and Life Prediction, High Temperature Creep-Fatigue, Current Japanese Materials Research, 3, pp. 203~222, New York (1988). 23. David D. Van Horn, Mechanical and physical bases for incandescent lamp exponents, illuminating engineering, 6, pp. 196~202 (1965).