PERFORMANCE OF A CONCENTRATING PHOTOVOLTAIC/THERMAL SOLAR COLLECTOR Joe S Coventry Centre for Sustainable Energy Systems, Australian National University, Canberra, 0200, Australia +612 6125 3976, +612 6125 0506, [email protected]

Abstract – The performance of a parabolic trough photovoltaic/thermal collector with a geometric concentration ratio of 37x is described. Measured results under typical operating conditions show thermal efficiency around 57% and electrical efficiency around 11%, therefore a combined efficiency of 68%. The impact of non-uniform illumination on the solar cells is investigated using purpose built equipment that moves a calibrated solar cell along the line of the receiver and measures short circuit current. The measured illumination flux profile along the length shows significant variation, despite the mirror shape error being less than 1mm for most of the mirror area. The impact of the illumination non-uniformities due to the shape error, receiver support post shading and gaps between the mirrors is shown to have a significant affect on the overall electrical performance. The flux profile transverse to the receiver length is also investigated. Peak flux intensities are shown to be around 100 suns. The impact on efficiency due to open circuit voltage reduction is discussed. 1. INTRODUCTION The Centre for Sustainable Energy Systems (CSES) at the Australian National University (ANU) has developed a photovoltaic/thermal (PV/T) collector with geometric concentration ratio of 37x. The so-called Combined Heat and Power Solar (CHAPS) collectors consist of glass-onmetal mirrors that focus light onto high efficiency monocrystalline silicon solar cells to generate electricity. Fluid flowing through a conduit at the back of the cells removes most of the remaining energy as heat, which can then be used for building heating and domestic hot water. The first commercial installation of single-axis tracking CHAPS technology is a 300m2 system providing electricity, and domestic and heating hot water for Bruce Hall, a residential college at ANU, with construction commencing in October 2003. The term ‘PV/T collector’ has been used in its loosest sense to describe semi-transparent PV panels such as dyesensitised solar panels used for daylighting, where heat and light is transferred directly into a building. However, in general, the term refers to a system with active heat collection through a heat transfer fluid, usually air or water, and sometimes both (Elazari, 2000, Tripanagnostopolous et al., 2001b). Most studies of PV/T collectors relate to flat plate collectors, with investigation of parameters such as absorber plate and tube dimensions (Bergene and Lovvik, 1995), fluid flow rates (Bhargava et al., 1991, Kalogirou, 2001, Prakash, 1994), tank size (Agarwal and Garg, 1994), PV cell packing factor (Cox and Raghuraman, 1985, Fujisawa and Tani, 2001), use of amorphous silicon (LESOPB/EPFL et al., 2000, Platz et al., 1997, Ricaud and Roubeau, 1994), and in the case of air collectors, multiple pass configurations (Hegazy, 2000, Sopian et al., 1996, 1997, Tripanagnostopolous et al., 2001a). The use of low concentration non-imaging optics with PV/T has also

received some attention (Brogren et al., 2000, 2001, Garg et al., 1991, 1994, 1999). Comparisons in performance of a combined PV/T collector with conventional PV and thermal systems show increased solar energy conversion and potential cost benefits (Fujisawa and Tani, 1997, 2001, Huang et al., 2001, Wolf, 1976). Detailed theoretical models of PV/T collectors have been developed, showing how thermal and electrical outputs are related (Bergene and Lovvik, 1995, Coventry, 2002, Florschuetz, 1979, Mattei et al., 1998, Morita et al., 2000, Prakash, 1994, Sandnes and Rekstad, 2000, Zondag et al., 2002). There has been less work in the field of medium to high concentration PV/T collectors, perhaps due to the demands of tracking the sun and achieving effective cooling of the cells. One concept that reduces the heat flux on the cells, and allows high temperature applications, is spectral beam splitting. Using either a bandpass or bandstop filter, radiation in the useful range for photovoltaic conversion is directed to the solar cells, and the remainder is directed to a thermal receiver. A PV/T system using spectral beam splitting was modeled using TRNSYS by Hamdy et al. (1988) and investigated by various others, as reviewed by Imenes and Mills (2002). Spectral beam splitting has been proposed for two concentrator systems in Australia, the Multi Tower Solar Array (Mills et al., 2002) and the Solar Systems SS20 dish collector (Lasich, 2001). Even when beam splitting is used, a significant amount of radiation incident on the cells is converted to heat, as only the bandgap energy is used in the photovoltaic conversion. Therefore far better heat transfer from the cells to the cooling fluid is required for a concentrating PV/T collector than a flat plate PV/T collector. The other major challenge in designing a concentrator PV system is to achieve an acceptable radiation flux

distribution. The CHAPS system is a linear concentrator, and tracks the sun on a single axis. Therefore the impact on flux profile due to gaps between mirrors and receiver support posts is not confined to a particular area, but shifts along the receiver as the incidence angle of the sun’s radiation changes as the sun moves across the sky. Non-uniformities are compounded by mirror shape error, which, though the errors are small in magnitude, has a significant effect on the flux profile, and the electrical performance of the array. The passively cooled EUCLIDESTM array in Tenerife (Sala et al., 2000) is the only other large scale PV concentrating parabolic trough system. The effect of optical imperfections on the array performance is discussed by Anton et al., (2000), and was significant enough to warrant trials of a secondary concentrator. 2. DESCRIPTION OF THE CHAPS SYSTEM The development of the CHAPS systems was preceded by PV trough technology development at the ANU since the mid-1990s, culminating in the commissioning of a 20kW two-axis tracking passively cooled PV trough array at Rockingham, Western Australia (Smeltink et al., 2000). Since then efforts have focused on actively cooled CHAPS systems designed for integration in buildings to supply hot water and electricity. The collectors are made up of 1.5m long mirror and receiver modules, which are connected end-to-end to form a row. The first CHAPS prototype is a single trough, 15m long, pictured in figure 1. The Bruce Hall system is made up of 8 troughs, each 24m long. The mirror, receiver and solar cell widths are 155cm, 8cm and 4cm respectively, which gives a geometric concentration ratio of 37x excluding the shading due to the receiver.

phosphorous doping beneath the fingers to reduce the contact resistance and c) a relatively large electroplated silver finger cross-section. Most cells are around 20% efficient at 25°C under 30 suns concentration.

Figure 2. ANU concentrator cell. The solar cells are bonded to an aluminium receiver with a thermally conductive, electrically insulating tape. They are connected in series and encapsulated with silicone and low-iron glass. Schottky bypass diodes are used to protect the cells against going into reverse bias due to partial shading of the receiver. The heat transfer fluid, which is water with anti-freeze and anti-corrosive additives, is pumped through the extruded aluminium receiver (figure 3) to cool the cells and collect thermal energy. The back and sides of the receiver are insulated with 20mm thick glasswool encased by a galvanised steel cover. Internal fins have been incorporated in the fluid conduit to increase the heat transfer surface in order to minimise the operating temperature difference between the cells and the fluid.

Figure 3. Cross-section of the ANU receiver.

Figure 1. Prototype CHAPS collector at ANU. The solar cells manufactured by CSES (figure 2), are monocrystalline silicon cells designed to have low internal series resistance, since the high current density under concentrated radiation significantly affects the fill factor of the cell. Low series resistance around 0.043 cm2 is achieved in the ANU concentrator cells by a) narrow spacing of the conductive fingers, which reduces the distance electrons travel through the silicon, b) heavy

The parabolic mirrors were developed at the ANU by Glen Johnston and Greg Burgess, and follow on from similar development of three-dimensional curved mirrors for dishes (Johnston et al., 2001). The glass-on-metal laminate (GOML) mirrors are composed of a silver backed mirror 1mm thick, laminated to a sheet metal substrate. The mirror is held in its parabolic shape by stamped tab ribs at either end of the mirror. The mirrors are around 92% reflective, which compares well with other reflective surfaces such as anodised aluminium at 81% (Brogren et al., 2000). The glass surface is highly scratch resistant when compared to some plastic film concentrators, and the mirrors have been subjected to a number of years of accelerated lifetime testing without significant deterioration.

The sun-tracking controller, designed at the ANU (Dennis, 2002), controls a linear actuator that is connected to a circular tracking wheel by cables. The tracking accuracy is set to ±0.2°. Because of the stamped tab rib design, there are small gaps between the mirrors averaging around 19 mm. To achieve good shape accuracy with a GOML mirror, it is desirable to have a continuous sheet across the width, and therefore the receivers are supported from outside the mirror. The support arms and mirror are mounted on a single square section beam designed to have a maximum of 0.5° twist along its length. 3. EXPERIMENTAL APPARATUS AND METHOD The thermal and electrical performance of a CHAPS receiver was measured using a custom built outdoor testing unit (figure 4).

A resistive load was held across the receiver at the maximum power point. The load was set at a defined voltage, and adjusted to the maxium power voltage by regular comparison with IV curves measured from the receiver. The voltage was measured across the receiver, and current measured across a current sense resistor. A dataTaker DT600 was used for the data logging. Data was logged every 10 seconds, and returned each minute. The trough was tracked manually to face directly at the sun at all times. Flux profile along the receiver was investigated using a purpose built device, known at ANU as the ‘skywalker’, consisting of a 40mm wide and 50mm long solar cell that moves along the focal line of a trough (figure 5). Short circuit current and position are logged using a National Instruments data aquisition card, and processed and displayed in LABVIEWTM.

Figure 4. Outdoor testing unit. The 125cm wide trough is narrower than on the prototype system, but a little longer than the receiver to ensure an even radiation flux distribution along the length. The test receiver is made up of 28 cells connected in series, one of which is bypassed due to failure upon installation. The cells average 16.8% efficiency at 65 degrees under 30 suns flux intensity. The cells were chosen to have similar maximum power point currents. Parameters measured include the direct beam radiation, measured with a Kipp & Zonen pyrheliometer, the ambient temperature, measured with a PT100 probe mounted in a Stevenson’s screen, the wind speed and direction, the inlet temperature of the water, measured with a PT100 probe inserted into the flow, the temperature difference across the collector, measured using a differential thermocouple arrangement, and the mass flow, measured with a calibrated turbine meter with ±1% accuracy. The mass flow was held near 40 ml/s for all the tests. Inlet temperature was held constant with either a mechanical tempering valve or temperature controlled booster heater. The data was sampled across 4 days, with wind speeds ranging from 0 – 0.6 m/s and the receiver angle varying from around 45º to 70º from horizontal.

Figure 5. The so-called ‘Skywalker’ device. Current-voltage (IV) curves for the receiver were measured using a custom-built capacative curve tracer (Keogh, 2003). 4. EFFICIENCY RESULTS The thermal output Qth and electrical output Qelec of the CHAPS collector are calculated from the measured data as follows: Qth = c p ⋅ m ⋅ (Tout − Tin )

(1)

Qelec = I mp ⋅ Vmp

(2)

where c p is the specific heat at the average fluid temperature, m is the mass flow, and Tout and Tin are the outlet and inlet temperatures respectively of the fluid. I mp and Vmp are the current and voltage of the receiver at the maximum power point. The thermal efficiency ηth and electrical efficiency ηelec are calculated based on the following definitions:

Qth ηth = Gd × Am

ηelec =

Qelec Gd × Am

(3) (4)

where Gd is the direct beam radiation and Am is the product of the mirror width of 125cm, and length of 150cm. Note that the receiver aperture is actually only 143.5cm long, so the efficiency measure includes losses due to hydraulic connections at the ends of the receiver, as well as losses due to the shading caused by the receiver itself.

of 57% and electrical efficiency of 11%. The efficiency curve for a typical flat plate collector is superimposed on figure 6, and shows a thermal efficiency of 62% under the same operating conditions. The overall energy conversion for the CHAPS collector is higher, and 11% electrical conversion is equivalent to a good quality flat plate PV system under operating conditions. 5. LONGITUDINAL FLUX PROFILE RESULTS The flux profile along focal line for the mirror on the test rig is compared to that of a typical mirror on the prototype CHAPS system (figure 7). In both cases the incident radiation is normal to the mirror.

Figure 6 shows the efficiency results for the receiver when it is operating both with and without an electrical load. Tm and Tamb are the mean fluid temperature and ambient temperature respectively.

Figure 7. Radiation flux profiles for the test rig and a CHAPS prototype mirror.

Figure 6. Efficiency curves for the CHAPS receiver. The electrical efficiency is plotted on the same scale in order to show the data points corresponding to the thermal efficiency data points. However, the electrical efficiency depends on absolute temperature rather than temperature difference`. Addition of the electrical efficiency data to the thermal efficiency data yields a combined efficiency trend very similar to that of the thermal efficiency when there is no electrical load. The thermal efficiency compares well with other technologies designed to operate at a similar temperature. Under typical operating conditions of, say, fluid temperature of 65ºC, ambient temperature of 25ºC and direct radiation of 1000 W/m2 (shown by the vertical line on figure 6), the CHAPS collector has thermal efficiency

Because the solar cells are connected in series, the current passing through each cell is the same. Therefore, if a single cell has lower radiation flux intensity, it will limit the current of the rest of the string, a major cause of receiver inefficiency. Figure 7 shows how important the mirror optical quality is to the overall performance of the system. For the mirror from the prototype system, the cell with minimum current is 12% lower than the average cell current, which will force all cells to operate away from their optimal point (neglecting the use of bypass diodes). The magnitude of the resulting power drop will depend on the quality and illumination of the other cells, but is in order of 7%. On the other hand, the mirror from the test rig has only a 2% drop in current due to the lowest illuminated cell, which results in negligible power drop. Both flux profiles in figure 7 come from GOML mirrors. The difference in uniformity between the two demonstrates the importantance of accurate mirror fabrication. The shape of a representative mirror panel from the prototype trough has been measured by Glen Johnston using the photogrammetric method developed in his PhD thesis (Johnston, 1998), with accuracy estimated to be 20-40 microns. Figure 8 shows the shape error of the trough, which is the difference between the measured z-coordinates and those calculated for a parabolic trough fitted to the data using a least squares technique.

lowest illumination. However, as the sun moves towards the normal, the relative influence of the support posts and mirror gap change. The mirror shape error compounds the problem at the end region, as the slope error is highest at the ends. The relative influence of the support posts, mirror gap and mirror shape error has been the subject of a separate study using ray tracing carried out using OptiCAD (Opticad Corporation, 2001).

Figure 8. Deviation from a perfect parabolic trough.

The magnitude of peaks and troughs in the illumination profile at the focus could be expected to be attenuated somewhat due to the single short-circuit current reading for the 50mm long solar cell effectively averaging the illumination. This was investigated experimentally by masking the solar cell at the focus of the ‘skywalker’ device, all bar a narrow slit of 1.5mm.

Maximum deviation from the ideal shape is in the order of 1 mm, and the majority of the mirror surface is within 0.5 mm. This demonstrates that even small deviations from the perfect shape cause significant non-uniformities in the flux at the focal line, and hence power losses. The reason for the shape irregularities in the prototype CHAPS mirrors has since been identified and rectified. In the case where radiation is incident upon the mirror at an angle away from the surface normal, the effect of the gap between mirrors and the receiver support arms becomes significant. This has been investigated for a range of incidence angles using the ‘skywalker’ device. The results are shown in summary in figure 9.

Figure 10. Comparison between ‘skywalker’ results with a 50mm long cell and a 1.5mm long cell. Figure 10 shows the close similarity between the full cell data and the masked cell data with a moving average over 50mm, and confirms the flux profile is the same for the two measurements. Significant variation in the illumination profile is observed, particularly in the region under the influence of the mirror gap and support post shading (seen close up in the bottom graph in figure 10). The peaks and troughs in illumination are deep, and it is therefore fortunate that the 50mm long cells used in the receiver do provide some illumination ‘smoothing’. 6. THE EFFECT ACROSS THE CELL Figure 9. Flux profile at the focal line for a range of angles of incidence. The double dip that can be on the right of the first flux profile is due mainly to the shading from the receiver supports and mirror gap. At larger angles of incidence, the effect of the shading from the receiver supports is spread over a wider section of the receiver, and the gap between mirrors is dominant in causing the region of

OF

NON-UNIFORMITIES

The illumination profile across a cell has been determined experimentally by Glen Johnston using videographic flux mapping techniques discussed in his PhD thesis (Johnston, 1998). A typical mean flux crosssection for the 15m CHAPS prototype is shown in figure 11. Given the irregularity of the longitudinal flux profile

(figure 10), the maximum flux intensity exceeds 100 suns in localised regions.

Another consequence of localised high illumination is that the temperature in these regions can be significantly higher than elsewhere in the cell. Heat flow studies using the finite element package Strand7 have shown the central illuminated region of the cell may be as much as 14ºC hotter than the edges of the cell (Coventry et al., 2002). Franklin and Coventry, (2002) incorporated temperature effects into their model to show there is a further reduction in open circuit voltage due to the temperature non-uniformity. Overall, in the absence of a secondary flux modifier, the efficiency losses due to the non-uniform illumination flux profile across the cell are unavoidable, and for good cells, are in the order of 5-10% relative.

Figure 11. Mean flux profile cross-section on the CHAPS prototype system at 3.8º angle of incidence. The effect of highly non-uniform light on solar cells is discussed in detail in Franklin and Coventry, (2002). To simulate the effect of non-uniform flux distribution across a cell for a linear concentrator, an IV curve was measured in a flashtester (Keogh, 2001) for a single solar cell under two scenarios: 30 suns concentration over the whole cell, and 90 suns concentration over the middle third of the cell. Tests were carried out with the solar cell held to a heat sink block at 25°C. The results (figure 12) show a reduction in open circuit voltage of 6.5 mV, and a softening of the IV curve, resulting in an efficiency drop from 20.6% for uniform illumination to 19.4% for centralised illumination. The magnitude of the voltage drop due to non-uniform illumination becomes more significant in relative terms when the open circuit voltage is reduced by temperature. 0 Uniform illumination profile

Current (A)

-5

Non-uniform illumination profile

-15

-20

0.5

0.55

0.6

0.65

Figure 13 shows the sensitivity of open circuit voltage, fill factor and efficiency to temperature, comparing a cell measured on the flashtester that has the non-uniform illumination profile, with a receiver on the test rig.

Figure 13. The effect of temperature on open circuit voltage (per cell), fill factor and efficiency. The dominant feature is a drop in open circuit voltage of 2.0mV/°C and 1.8mV/°C per cell for the cell and receiver respectively. The reason for the sensitivity of voltage to temperature is that the reverse saturation current J0 increases rapidly with temperature. The fill factor is also softened a little with temperature. The measured drop in efficiency were 0.39%/°C and 0.41%/°C relative for the cell and receiver respectively.

-10

-25 0.45

7. TEMPERATURE DEPENDENCE OF THE PV RECEIVER

0.7

0.75

Voltage(V)

Figure 12. IV curves for uniform and centralised illumination of a solar cell. A reduction in open circuit voltage will occur regardless of the cell quality. As discussed by Franklin and Coventry, (2002), cells under open circuit conditions that are non-uniformly illuminated have significant internal current flows, and hence losses in potential.

8. CONCLUSIONS Achieving consistent optics is probably the key challenge in achieving good performance from a concentrator PV system. It is shown that electrical performance is most sensitive to non-uniformities longitudinal to the trough, which may result from imperfect mirrors, gaps between mirrors and regions of shading. The thermal and electrical efficiencies of the CHAPS collector are shown under ideal conditions, and are comparable to those achieved by equivalent flat plate thermal and PV collectors. Once commercialised, it is expected that the CHAPS system will be significantly cheaper than side-

by-side PV and thermal solar systems with equivalent output.

Dennis, M., 2002: The ANU Sun Tracking Controller User Manual (available on request at the ANU), Australian National University.

9. ACKNOWLEDGEMENTS The work described in this paper has been supported by the Australian Cooperative Research Centre for Renewable Energy (ACRE). ACRE’s activities are funded by the Commonwealth’s Cooperative Research Centres Program. Joe Coventry has been supported by an ACRE Postgraduate Research Scholarship.

Elazari, A., 2000: Building Integrated Multi pv/t/a Solar System roof tile. ISES Europe Solar Conference, Copenhagen. Florschuetz, L. W., 1979: Extenssion of the HottelWhillier Model to the Analysis of Combined Photovoltaic/Thermal Flat Plate Collectors. Solar Energy, 22, 361-366.

10. REFERENCES

Agarwal, R. K. and Garg, H. P., 1994: Study of a photovoltaic thermal system - thermosyphonic solar water heater combined with solar cells. Energy Conversion & Management, 35, 605-620. Anton, I., Sala, G., Arboiro, J. C., Monedero, J., and Valera, P., 2000: Effect of the optical performance on the output power of the EUCLIDESTM array. 16th European Photovoltaics Solar Energy Conference, Glasgow, 22252228. Bergene, T. and Lovvik, O. M., 1995: Model calculations on a flat-plate solar heat collector with integrated solar cells. Solar Energy, 55, 453-462. Bhargava, A. K., Garg, H. P., and Agarwal, R. K., 1991: Study of a hybrid solar system - solar air heater combined with solar cells. Energy Conservation and Management, 31, 471-479.

Franklin, E. and Coventry, J., 2002: Effects of Highly Non-uniform Illumination Distribution on Electrical Performance of Solar Cells. ANZSES Solar Conference, Newcastle, Australia. Fujisawa, T. and Tani, T., 1997: Annual exergy evaluation on photovoltaic-thermal hybrid collector. Solar Energy Materials & Solar Cells, 47, 135-148. Fujisawa, T. and Tani, T., 2001: Optimum design for residential photovoltaic-thermal binary utilization system by minimizing auxiliary energy. Electrical Engineering in Japan (English translation of Denki Gakkai Ronbunshi), 137, 28-35. Garg, H. P. and Adhikari, R. S., 1999: Performance analysis of a hybrid photovoltaic/thermal (PV/T) collector with integrated CPC troughs. International Journal of Energy Research, 23, 1295-1304.

Brogren, M., 2001: Low-Concentrating Photovoltaic Systems with Parabolic Reflectors, Licentiate Thesis, Division of Solid State Physics, Uppsala University, 75.

Garg, H. P., Agarwal, R. K., and Bhargava, A. K., 1991: The effect of plane booster reflectors on the performance of a solar air heater with solar cels suitable for a solar dryer. Energy Conservation and Management, 32, 543554.

Brogren, M., Nostell, P., and Karlsson, B., 2000: Optical Efficiency of a PV-Thermal Hybrid CPC Module. Eurosun 2000 - ISES Europe Solar Conference, Copenhagen.

Garg, H. P., Agarwal, R. K., and Joshi, J. C., 1994: Experimental study on a hybrid photovoltaic thermal solar water heater and its performance predictions. Energy Conversion & Management, 35, 621-633.

Coventry, J., 2002: Simulation of a concentrating PV/thermal collector using TRNSYS. ANZSES Solar Energy Conference, Newcastle.

Hamdy, M. A., Luttmann, F., and Osborn, D., 1988: Model of a Spectrally Selective Decoupled Photovoltaic/Thermal Concentrating System. Applied Energy, 30, 209-225.

Coventry, J., Franklin, E., and Blakers, A., 2002: Thermal and electrical performance of a concentrating PV/Thermal collector: results from the ANU CHAPS collector. ANZSES Solar Energy Conference, Newcastle. Cox, C. H. and Raghuraman, P., 1985: Design Considerations for Flat-Plate-Photovoltaic/Thermal Collectors. Solar Energy, 35, 227-241.

Hegazy, A. A., 2000: Comparative study of the performances of four photovoltaic/thermal solar air collectors. Energy Conversion & Management, 41, 861881. Huang, B. J., Lin, T. H., Hung, W. C., and Sun, F. S., 2001: Performance evaluation of solar photovoltaic/thermal systems. Solar Energy, 70, 443-448.

Imenes, A. G. and Mills, D. R., 2002: Spectral Beam Splitting for Increased Conversion Efficiency in Solar Concentrating Systems. ANZSES Solar Energy Conference, Newcastle. Johnston, G., 1998: Focal Region Modelling and Characterisation of Paraboloidal Dish Solar Concentrators, PhD thesis, Australian National University. Johnston, G., Burgess, G., Lovegrove, K., and Luzzi, A., 2001: Economic Mass Producable Mirror Panels for Solar Concentrators. ISES Solar World Congress, Adelaide. Kalogirou, S. A., 2001: Use of TRNSYS for modelling and simulation of a hybrid pv-thermal solar system for Cyprus. Renewable Energy, 23, 247-260. Keogh, W., 2001: Accurate performance measurement of silicon solar cells, PhD, Department of Engineering, Australian National University, 187. Keogh, W., 2003: Capacitive I-V Curve Tracer Manual, Australian National University, 34 pp. Lasich, J. B., Apparatus for separating solar radiation into longer and shorter wavelength components. Australian Patent 731,495 29 Mar. 2001

Prakash, J., 1994: Transient analysis of a photovoltaic thermal solar collector for co-generation of electricity and hot air water. Energy Conversion & Management, 35, 967-972. Ricaud, A. and Roubeau, P., 1994: "Capthel", a 66% efficient hybrid solar module and the "Ecothel" cogeneration solar module. 24th IEEE Photovoltaic Specialists Conference 1994, Hawaii, IEEE, 1012-1015. Sala, G., Anton, I., Arboiro, J. C., Luque, A., Camblor, E., Mera, E., Gasson, M., Cendagorta, M., Valera, P., Friend, M. P., Monedero, J., Gonzalez, S., Dobon, F., and Luque, I., 2000: The 480 kW(p) EUCLIDESTMTHERMIE Power Plant: Installation, Set-Up and First Results. 16th European Photovoltaic Solar Energy Conference, Glasgow, James & James Ltd, 2072-2077. Sandnes, B. and Rekstad, J., 2000: A photovoltaic/thermal (PV/T) collector with a polymer absorber plate; experimental study and analytical model. ISES Europe Solar Congress, Copenhagen. Smeltink, J., Blakers, A., and Hiron, S., 2000: The ANU 20kW PV/Trough Concentrator. 16th European Photovoltaic Solar Energy Conference, Glasgow, 23902393.

LESO-PB/EPFL, Enecolo AG, and Ernst Schweizer AG, 2000: New generation of Hybrid Solar PV/T collectors, Swiss Federal Office of Energy, 1-55 pp.

Sopian, K., Liu, H. T., and Kakac, S., 1997: Research and development of hybrid photovoltaic thermal solar air heaters. International Journal of Global Energy Issues, 9, 382-392.

Mattei, M., Crostpfaro, C., and Louche, A., 1998: Modelling a hybrid PV/T collector. 2nd World Conference and Exhibition on Photovoltaic Solar Energy Conversion, Austria.

Sopian, K., Yigit, K. S., Liu, H. T., Kakac, S., and Veziroglu, T. N., 1996: Performance analysis of photovoltaic thermal air heaters. Energy Conversion & Management, 37, 1657-1670.

Mills, D. R., Schramek, P., Dey, C., Buie, D., Imenes, A. G., Haynes, B. S., and Morrison, G., 2002: Multi Tower Solar Array Project. ANZSES Solar Energy Conference, Newcastle.

Tripanagnostopolous, Y., Nousia, T., and Souliotis, M., 2001a: Test results of air cooled modified PV modules. 17th European Photovoltaic Solar Energy Conference, Munich.

Morita, Y., Fujisawa, T., and Tani, T., 2000: Moment Performance of Photovoltaic/Thermal Hybrid Panel (Numeric Analysis and Exergetic Evaluation). Electrical Engineering in Japan (English translation of Denki Gakkai Ronbunshi), 133, 81-87.

Tripanagnostopolous, Y., Tzavellas, D., Zoulia, I., and Choratou, M., 2001b: Hybrid PV/T systems with dual heat extraction operation. 17th European Photovoltaic Solar Energy Conference, Munich.

Opticad Corporation, 2001: OptiCAD Optical Anaysis Program - User's Guide, Version 7.0, Opticad Corporation. Platz, R., Fischer, D., Zufferey, M. A., Anna Selvan, J. A., Haller, A., and Shah, A., 1997: Hybrid Collectors Using Thin-Film Technology. 26th Photovoltaics Specialists Conference, Anaheim, CA, IEEE, 1293-1296.

Wolf, M., 1976: Performance Analyses of Combined Heating and Photovoltaic Power Systems for Residences. Energy Conversion, 16, 79-90. Zondag, H. A., De Vries, D. W., Van Helden, W. G. J., Van Zolingen, R. J. C., and Van Steenhoven, A. A., 2002: The thermal and electrical yield of a PV-thermal collector. Solar Energy, 72, 113-128.