Home Search Collections Journals About Contact us My IOPscience An LED solar simulator for student labs This content has been downloaded from IOPscience. Please scroll down to see the full text. 2017 Phys. Educ. 52 035002 (http://iopscience.iop.org/0031-9120/52/3/035002) View the table of contents for this issue, or go to the journal homepage for more Download details: IP Address: 164.76.7.55 This content was downloaded on 21/03/2017 at 19:54 Please note that terms and conditions apply. Paper Phys. Educ. 52 (2017) 035002 (5pp) iopscience.org/ped An LED solar simulator for student labs Manuel I González Departamento de Física, Universidad de Burgos, Avda. Cantabria E-90006, Burgos, Spain E-mail: miglez@ubu.es Abstract Measuring voltage–current and voltage–power curves of a photovoltaic module is a nice experiment for high school and undergraduate students. In labs where real sunlight is not available this experiment requires a solar simulator. A prototype of a simulator using LED lamps has been manufactured and tested, and a comparison with classical halogen simulators has been performed. It is found that LED light offers lower levels of irradiance, but much better performance in terms of module output for a given irradiance. 1. Introduction requisites concerning irradiance, spectral matching to natural sunlight, spatial uniformity and temporal stability. The requisites are detailed in documents such as [3, 4]. For a student lab, of course, the requisites are much looser, and a common practice is to use incandescent halogen lamps as the source [5, 6]. Tungsten-based incandescent lamps are widespread and cheap, but lack spectral quality. The color temperature of sunlight is about 6000 K, whereas the color temperature of a tungsten filament does not usually exceed 2800 K [7]; otherwise the filament lifetime is drastically shortened. According to Wien’s law the emission peak at 2800 K lies at about 1030 nm, i.e. at the near infrared zone of the electromagnetic spectrum. By Planck’s law, this implies that most photons emitted by the filament simply do not have enough energy to create electron-hole pairs in an ordinary silicon PV module. Therefore one may expect a very low conversion efficiency of the module when lit with an incandescent source, as will be seen below. The recent spread of light emitting diode (LED) systems offers a new approach to solar simulation. LED light is not governed by the Energy sources and energy conversion have always been a major concern for physics teachers at all levels. Recent development of renewable sources makes this interest even greater especially (but not only) for laboratory instructors. Nowadays, catalogues of laboratory equipment dealers [1, 2] include a variety of experiments about solar thermal, solar photovoltaic, wind energy and other renewable sources. As for photovoltaic (PV) energy, a simple but enlightening experiment consists of obtaining the characteristic curves of a PV module. This experiment, which will be explained in detail in section 4, explores the conversion of solar radiation into electricity and involves concepts such as irradiance, power output, conversion efficiency and others. To carry out this experiment in labs where the supply of sunlight is not ensured, a solar simulator is required. A solar simulator is an artificial light source whose physical characteristics are similar to those of sunlight. In fact, solar simulators are routinely used to perform standard tests on PV modules in certification facilities. These ‘serious’ simulators must fulfil a number of 1361-6552/17/035002+5$33.00 1 © 2017 IOP Publishing Ltd M I González laws of thermal emission and LED lamps can be found in various equivalent color temperatures ranging from 2500 K (warm white) to 6000 K (daylight). This work deals with the manufacture of a prototype for a solar simulator using commercially available LED lamps. Specifically, the relationship between irradiance and distance to the simulator will be measured and an estimate of spatial uniformity in irradiance will be obtained. The characteristic curves of a student PV module under three sources (sunlight, halogen and LED) will be measured and discussed. Finally the pros and cons of the prototype will be examined. 2. The prototype Figure 1. Prototype of LED solar simulator. An LED solar simulator suitable for student labs should satisfy the following requirements: The arrangement occupied an approximate area of 37 × 30 cm2, which was considered sufficient to light a PV module having 24 × 18 cm2. The lamps were all connected in parallel by means of standard wires located in the rear of the panel. Figure 1 shows a front view of the prototype. Figure 2 shows the light beams of a conventional halogen simulator and the proposed unit. The photo is strongly underexposed in order to reveal the different color temperatures, although the difference is even more remarkable under direct vision. • Color temperature of the lamps should be substantially higher than that of halogen lamps. Otherwise the spectral matching to sunlight will be poor. • The number of lamps and the directional characteristics of their light beams should provide an irradiance close to 1000 W m−2, which is typical of natural sunlight under good conditions [8]. This is also the irradiance under which standard tests are carried out on PV modules. • The irradiance on the plane of the PV module should have an acceptable degree of spatial uniformity. • The whole simulator should not be prohibitively expensive. 3. Irradiance and spatial uniformity The first comparison performed on both simulators referred to the irradiance on the PV module and its spatial uniformity. The test module was centered in front of each simulator, and oriented to normal incidence. The distance between each simulator and the PV module was then varied. For each distance the irradiance was measured at nine points of the module: the center, the four corners and the midpoint of each edge. Irradiances were measured with a student lab pyranometer. Average irradiances for each distance were computed as weighted averages, instead of simple ones. The assigned weights to the aforementioned locations were 1, 0.25 and 0.5, respectively. Figure 3 shows the results. A simple estimate of spatial uniformity was computed as the standard deviation of the nine measurements, and displayed in figure 3 as error bars. Inspection of figure 3 shows that for a given distance, the halogen simulator provides higher According to these criteria the selected lamps had 5 W power consumption, 220 V (the European AC Standard), 345 lm luminous flux (similar to the luminous output of a 40 W incandescent lamp), 4000 K color temperature, GU10 bulb socket and a built-in reflector cup providing 36° beam aperture. The unit cost was approximately €5. 42 lamps were mounted on a wooden panel, in a hexagonal arrangement with 5.5 cm distance between adjacent lamps. The packing density was very high, since the diameter of the reflecting cups was only slightly lower, namely 5 cm. In this way, the density of electrical power consumption was about 1900 W m−2 although, as will be shown below, the irradiance on the PV module was much lower. May 2017 2 Phys. Educ. 52 (2017) 035002 Average irradiance / W m–2 An LED solar simulator for student labs Figure 2. ‘Classic’ halogen (left) and LED (right) simulators. Ihal = 256 000·d –1.581 1000 500 0 ILED = 588·e–0.019·d 0 40 Distance/cm 80 Figure 3. Irradiance versus distance. Upper data set: halogen simulator; lower data set: LED simulator. Dashed lines: least squares fits. irradiance. This is simply due to its higher power consumption (500 W against 210 W). The degree of spatial uniformity was reasonably good for the LED simulator at all distances and for the halogen simulator at large distances. At relatively shorter distances (40 cm or less) the halogen simulator exhibits poor uniformity as a logical consequence of the quasi point-like character of its filament. The trend lines in figure 3 do not have any particular physical meaning. In fact, the best fits have been achieved with different types of functions. They must be interpreted as simple tools for irradiance versus distance calibration. setup (except for the pyranometer) during the test of the LED simulator. In order to compare the performances of a given PV module under different light sources three trials were conducted. The first trial was carried out under sunlight at 1010 W m−2, which is very close to the irradiance level in standard tests. The second and third trials used our LED and halogen simulators, respectively, at 500 W m−2, which is accessible to both simulators. The calibration curves shown in figure 3 were used to select the distance between module and simulators. Figure 5 depicts the results of these trials. To compare the experiment performed at 1010 W m−2 to those performed at 500 W m−2, the latter were corrected by using the well-known property [9] that the output current of a PV module is approximately proportional to the irradiance, whereas the output voltage does not depend significantly on irradiance. Therefore the main comparison should involve the continuous lines in figure 5. Once the corrections are made short circuit currents were nearly equal for sunlight and LED (325 and 315 mA, respectively) and much lower for halogen light (170 mA), whereas open circuit voltages were similar: 18.8 V (sun), 18.2 V (LED) and 17.9 V (halogen). A similar conclusion can be drawn if one computes the power output by multiplying volt­ age and current and plots the P–V (power output versus voltage) diagram, as depicted in figure 6. Under sunlight the maximum power was 4.2 W, 4. Characteristic curves and discussion The classical test of a PV module consists of measuring its I–V (current versus voltage) curve. In a student lab one can ignore the subtleties required in a certification facility, and the experiment can be carried out with very little equipment: the module itself, a rheostat, an ammeter and a voltmeter. A pyranometer is also required if one wants an accurate measurement of irradiance. The electrical circuit is shown in the upper right inset of figure 4. Different pairs (I, V ) can be obtained by varying the load on the rheostat, and the I–V curve is plotted by joining them. Intersections of the I–V curve with both axes define the open circuit voltage Voc and the short circuit current Isc (shown in the upper central inset of figure 4). These are readily measured by connecting the voltmeter and the ammeter, respectively, to the output ports of the PV module. Figure 4 shows the experimental May 2017 1500 3 Phys. Educ. 52 (2017) 035002 M I González Figure 4. Experimental setup for measuring the characteristic curves of a PV module. The upper right inset shows the electric connections, whereas the upper central inset depicts the typical aspect of an I–V curve, showing the open circuit voltage and short circuit current. close to the 4.5 W rated power of the module. After correction, LED light provided a similar maximum output, 3.9 W, whereas halogen light delivered at most 2.0 W. For a PV module it is customary to define the conversion efficiency as the quotient between the maximum electrical power output and the radiant power on the collection plane. According to this definition the efficiencies at 1000 W m−2 were 10.4% (rated), 9.6% (sunlight test), 9.0% (LED simulator test) and 4.6% (halogen simulator test). In any case, figures 5 and 6 and the efficiencies tell us the same: for a given irradiance level PV module output under LED light is similar to that under sunlight and twice as much as under halogen light. The poor performance of halogen light is clearly due to its richness in high wavelength, low energy photons. Sun, 1010 W m–2 300 Current/mA LED, corrected to 1000 W m–2 Halogen, corrected to 1000 W m–2 LED, 500 W m–2 100 Halogen, 500 W m–2 0 0 10 Voltage/ V 20 Figure 5. I–V curve for a PV module under different light sources. Black line: sunlight; blue: LED simulator; red: halogen simulator. Dashed lines: tests performed at 500 W m−2; continuous lines: tests performed or corrected to 1000 W m−2. The minute correction of sunlight curve from 1010 to 1000 W m−2 has been omitted. lamps, mounts and cabling amounted to some €300 ($330 at the time of writing). Moreover the manufacture required a few hours of do-it-yourself. • Low power consumption of common LED lamps makes it a challenge to reach the standard 1000 W m−2 irradiance. Our prototype required a very dense lamp packing to deliver 500 W m−2. • Similar levels of spatial uniformity can be expected from LED and halogen simulators. 5. Conclusions Data presented in section 4 have been obtained by lighting a particular PV module with a particular LED simulator and a particular halogen simulator, but a number of useful and more general conclusions can be drawn: • A halogen simulator is expected to be cheaper than an LED one. The cost of a high power halogen lamp and its mount is about €50, whereas our LED prototype, including May 2017 200 4 Phys. Educ. 52 (2017) 035002 An LED solar simulator for student labs Power output/ mW 4000 –2 , un S 2000 10 10 W c rre –2 m 0W m 00 o1 t ted [3] IEC 60904-9:2007 document, International Electrotechnical Commission Webstore (https://webstore.iec.ch/publication/3880) (Accessed: 13 March 2017) [4] ASTM E927—10(2015) document, Standard Specification for Solar Simulation for Photovoltaic Testing (www.astm.org/Standards/ E927.htm) (Accessed: 13 March 2017) [5] FREDERIKSEN—manufacturer of Physics Equipment for teaching (http://int.frederiksen. eu/shop/product/halogen-lamp--350-400-w) (Accessed: 13 March 2017) [6] PHYWE SYSTEME web page (www.phywe. com/en/catalog/product/view/id/11064/s/ filament-lamp-220v-120w-with-reflector/ category/16696/) (Accessed: 13 March 2017) [7] Pritchard D C 1998 Lighting 5th edn (Longman: Singapore) [8] Duffie J A and Beckman W A 2006 Solar Engineering of Thermal Processes 3rd edn (Hoboken, NJ: Wiley) [9] Messenger R and Ventre J 2000 Photovoltaic Systems Engineering (Boca Raton, FL: CRC Press) W 000 –2 m co to 1 D, ected –2 m LE , corr 0W en 0 g 5 o , Hal LED m–2 n, 500 W Haloge 0 0 10 Voltage / V 20 Figure 6. P–V curve for a PV module under different light sources. • As for spectral content, LED lamps with high color temperature are far superior to halogen lamps. Under an LED at 4000 K, the response of a PV module in terms of short circuit current, maximum power output and efficiency is close to that under sunlight. Received 16 December 2016, in final form 24 January 2017 Accepted for publication 9 February 2017 https://doi.org/10.1088/1361-6552/aa5f86 Manuel I González teaches physics and renewable energy to engineering students at the University of Burgos. His main research areas are solar cooling and new technologies applied to teaching physics. References [1] PHYWE webpage (www.phywe.com/en/physics/ age-16-19/energy/) (Accessed: 13 March 2017) [2] PASCO SCIENTIFIC webpage (www.pasco.com/ filters/topics/index.cfm?topics=02-065-000) (Accessed: 13 March 2017) May 2017 5 Phys. Educ. 52 (2017) 035002