Fiszera 14, 80-231 Gdansk, Poland TESTS AND THEORETICAL ANALYSIS OF A PVT HYBRID COLLECTOR OPERATING UNDER VARIOUS INSOLATION CONDITIONS

The main goal of the study was to investigate the relationship between the orientation of the PVT (PhotoVoltaic Thermal) collector and the thermal and electric power generated. Extensive research was performed to find optimal tilt angles for hybrid solar thermal collectors, which combine photovoltaic as well as thermal collection in a single unit, known as PVT (PhotoVoltaic Thermal) modules for an office building with working hours between 7.00 and 16.00. The comprehensive study included field measurements of the modules in central Poland and tests under AM (air mass) 1.5 conditions in a certified laboratory KEZO (Centre for Energy Conversion and Renewable Resources) Polish Academy of Sciences in Jablonna. Furthermore, a PVT system was investigated using the simulation method based on the dedicated Polysun software. The PV characteristics and efficiency of the PV module and the relation between power or efficiency of the PVT module and incidence angle of solar-irradiance were studied. Optimal work conditions for commercial PVT modules were ascertained. In addition, it was found that the maximum efficiencies of PV module (ηPV), solar thermal-collector (ηc) and PVT hybrid collector (ηPVT) registered under field conditions were higher than the ones measured under laboratory conditions.


Introduction
Energy plays a vital role in daily human needs. Worldwide, energy consumption has shown rapid growth, which is the important global energy problem. PV modules generate clean electricity, reducing air pollution and having only little environmental impact. Now, typical efficiency of PV modules is in the range 15%÷20% [1]. The conversion of solar energy into electric power depends on the location, orientation, and type of photovoltaic (PV) system. PVT hybrid collector works with even higher efficiency than sum of the solar thermal collector and PV module efficiencies [2] and simultaneously produces electrical and thermal power. The thermal efficiency of the flat-plate thermal collector is 40-60%. Therefore, the estimated efficiency of the PVT collector is the range from 60 to 80% [3]. In 1979 the first PVT model was demonstrated in Ref. [4]. Until now the PVT systems have been studied from various aspects: system design, performance analysis, simulation models, field or indoor experimental validation, economic and cost-effective analysis [2,[5][6][7][8]. Coventry [9] has developed and tested a PVT with a geometric concentration ratio of 37. Chow et al. [10] developed a prototype thermosyphon PVT collector made of aluminum alloy flat box. Dupeyrat et al. [11] tested the design of a prototype single glazed flat plate PVT collector focusing on the heat transfer between PV cells and fluid [3].
Nevertheless PVT hybrid collectors presents numerous advantages, specific work conditions and often contradictory requirements of both technologies cause extensive need for further technology investigation. Considering above facts, this paper aims on to studies of the relationship between orientation of PVT collectors and their thermal and electric power generation. Extensive research was performed to determine the optimal working angle for practical PVT installations. The parameters of the PVT system were measured in field and laboratory conditions. The results were confirmed using specialized computer program Polysun. Acquired results are useful for the design and construction of functional PVT systems.

Description of experiments
The research subject was a commercially available photovoltaic thermal hybrid solar collector manufactured by Volther Power Therm measuring 1640 x 870 x 105 mm. Fig. 1 presents the operation parameters of the collector for STC (Standard Test Conditions) [12]. Initial studies were performed under field conditions. Field experiments were conducted at the Technology Transfer Centre for Renewable Energy in Lodz (Φ = 51° 44' N, φ = 19° 21' E). The PVT hybrid collectors were located on the roof of the building and were inclined at 30º from the horizontal axis in a southern orientation: (E = +90°, S = 0°, W = -90°). Figure 2 presents a schematic diagram of the installation.
Research was done during the autumnal equinox at the beginning of October, during daytime between 8.00 and 16.00. It is assumed that the results obtained at the equinox will correspond well to the annual average values. The weather station on the roof of the building allowed determination of the meteorological conditions. Later, PVT tests were conducted under laboratory conditions. Clean water was used as a solar working fluid. The photovoltaic characteristics of the PVT module were measured under STC conditions using a class AAA steadystate solar simulator (LA150200, Eternal Sun) equipped with an AM 1.5 G filter. The light intensity was determined by a silicon reference cell (ReRa Solutions) and corrected to 1000 W/m 2 . I-V curves were recorded on IV-System Pro (ReRa Solutions) and a Tracer IV-curve software (ReRa Solutions). The IV-System Pro consist of an EL 9000A (Elektro Automatik) electronic DC load connected with two 34410A (Agilent) multimeters, for PV panels measurements, and a 2401 SourceMeter Keithley (uncertainty ±0.012%), for I-V measurements of a reference cell. During measurements the temperature difference between the inlet fluid and the ambient air was set to: ΔT = 5 K and ΔT = 30 K. Temperatures were measured by PT100 (class A, uncertainty ±0,15 °C) temperature sensors connected to a 34970A (Agilent) data acquisition control unit. Afterwards, solar thermal collector tests were carried out. During measurements, in order to better reflect the natural conditions, the solar simulator was operated in ST1000 mode wherein infrared radiation intensity is higher. Therefore, the light intensity was determined using an IR02 pyrgeometer Hukseflux (uncertainty ±15 °C) and a SR11 pyranometer Hukseflux (uncertainty ±8%) connected to a 34970A (Agilent) data acquisition control unit. The mass flow rate of fluid through the solar thermal collector was set to 1.3 kg/min and was determined by an Optimass 6400 C Krohne (uncertainty ±0.1%) mass flow meter.

Model description
The measured energy characteristics and the PVT system efficiency were verified using the Polysun software from the company VELA SOLARIS [14]. This part is mostly for comparative and demonstration use rather than for reliably justified conclusions, as the applied meteorological conditions do not refer to the specific test localization (due to known limitation regarding the license of this software version). The program provides dynamic simulations of thermal, photovoltaic and PVT systems and enables their optimization [16]. The parameters of the PVT system used in our simulations are listed down in Table 2. The setup of the solar PVT system used in this research is depicted in Figure 4. On the left side of the picture is the PVT collector, and in the center is the APU (auxiliary power unit). Both components are connected to the heat exchangers. Number of PVT modules was dedicated by program Polysun for 4 residents.

Time [h]
nc Pc η c P c

PVT collector
Storage tank Boiler oil system. Figure 5a summarizes the data related to the temperatures of ambient air (Ta), of the cooling water at the collector outlet (Tf) and of water in the storage tank (Ts). The daytime operation of a solar thermal collector can be divided into two periods. First, between 8.00 and 10.00, the cooling water temperature of the solar thermal collector, Tf, rises from 13 to 42ºC. At that moment, a circulation pump is activated leading to a heat exchange between water from the heat collector and water in the storage tank (see red curve in Fig. 5a). The temperature of the water in the storage tank, Ts, slowly increases but the temperature of the water in the collector holds steady for one hour (see plateau on Tf curve). Later, between 11.00 and 14.00 the water temperature in the collector increases quasi linearly. Although the temperature Tf decreases after 14.00 due to a strong decrease of solar irradiance and heat exchange with colder water in storage tank, the heat energy collected at the storage tank increases even after 16.00. At 16.00 the controller turns off the circulation pump to prevent possible heat losses from the water in the storage tank. The temperature difference of water collected in the solar collector and water in the storage tank approaches "zero" and later goes negative. The thermal heat energy accumulated in the storage tank (Q) was ~ 10 MJ. It can be calculated using the measured temperatures and model described in Ref. [16] given by the following equation: (2) where V is the fluid capacity flow rate during the day (m 3 ), ρ is density (kg/m 3 ), kc is the heat capacity of water J/(kg·K), Tz is the inlet fluid temperature to storage tank (K), Tp is the outlet fluid temperature to the storage tank (K). Figure 6b shows the instantaneous power (Pc) and efficiency (ηc) of the thermal collector calculated using the model [17]. The power Pc is given by the following equation: where Ac is total collector aperture area (m 2 ), βτ is transmittance-absorptance product, k is solar-collector heattransfer loss-coefficient (W/m 2 ·ºC), Tab is average temperature of the absorbing surface (ºC), Ta is ambient temperature (ºC), Irr is instantaneous solar irradiance (W/m 2 ).
The PV electric performance was also studied. Figure 6a shows how the module electric parameters (Imp, Vmp) at maximum power point changes during the chosen day. The interrelation between the instantaneous solar irradiance (Irr) and instantaneous, PV power at MPP (Pmax) and efficiency (ηPV) is shown in Figure 7b.  Figure 6a shows that the Imp current strongly depends on irradiance (compare with Irr at Fig. 6b), while voltage varies in a small range 14÷21 V. The black lines represent error bars. The PV current in the range 0.8÷ 3.9 A was measured. As solar irradiance increases three times, the current grows ~400%. Figure 7b shows that as the solar irradiance increases, the values of Pmax and ηPV also increase. The instantaneous efficiency was found in the range from about 8% (morning hours, Se = 10.73º) to 14% (irradiation about 600 W/m 2 , Se = 33.14º). The amount of power generated and voltage by PVT system was consistent with the solar irradiance. The maximum values were achieved at 14.00. The electric energy production-density, 475 Wh/m 2 , was determined using formula in Ref. [18]. The instantaneous maximum output-power MPP, 143 W (Se = 33.14º) was about 28% lower than the nominal power of the PV module (200 W).

Irr
Pmax ηPV I rr P max η PV

PVT tests under laboratory conditions
The relation of PVT module performance and solar-simulator altitude-angle were investigated under laboratory conditions. First, the PV module and later thermal collector performance was studied.

Impact of solar-simulator altitude-angles on performance of PV module and thermal collector
The photovoltaic electric characteristics were measured for different solar-simulator altitude-angles α and temperature difference ΔT = 5ºC and ΔT = 30ºC (between the temperature of water at the inlet to thermal collector and the ambient air). The light intensity was adjusted to 1000 W/m 2 . The altitude angle was varied between 30 º and 90º (for 90º the solar simulator was situated parallel to the PVT module).   Then, the PV module and the solar thermal collector were also studied for ΔT = 10ºC. Figure 9c shows that the PV power at MPP and efficiency increases when the solar altitude angle increases, from 104 to 115 W and 10 to 15%, respectively.
The performance of the thermal part of PVT and its dependence on solar-altitude angles was also investigated. The relation between instantaneous thermal power of solar collector (Pc) calculated based on eq. (3), (as well as efficiency) and the solar altitude angle was studied (see results in Table 3 and Figure 9). Table 3 shows solarsimulator irradiance and measured difference between cooling fluid temperature at the collector outlet and at the inlet (Toutlet -Tinlet) as function of the solar-altitude angle. As the solar-altitude angle increases from 45-75° the temperature difference, Toutlet -Tinlet , falls 3 times (from 0.64 to 0.21 K). As the altitude angle of the solar simulator increases, solar irradiance and the solar collector output-power also increase.
The efficiency (in contrast to power) of the thermal collector varies differently in relation to the solar altitude angle; it decreases slightly and then increases with the solar altitude angle following the tendency of Toutlet -Tinlet. Moreover, the efficiency drops only ~6%, whilst power increases by 75% of initial value (Fig. 9). This is important for electric-thermal power balancing under real installation conditions.

Comparison of results
The efficiency of PVT (ηPVT) collectors under both field (during the autumnal equinox) and laboratory conditions as a function of the solar-altitude angle was calculated using equation [19]: In the case of the field experiment, the values of efficiency for the thermal collector (ηc) (see Figure 10a) were deduced from data presented in Figure 6b, where the solar-altitude angle as a function of time was determined using the data shown in Figure 3 for a module positioned under 30° angle (Sa30º). The same was done for efficiency of PV module (ηPV) under field conditions, but this time using the data presented in Figure  6b. The rather erratic character of the field experiment data should be related to changing weather conditions, such as wind and air temperature. The variation of the solar altitude angle is limited to range 40-65º achievable for the chosen position of the solar panel during the autumnal equinox.
In the case of the laboratory experiment, the efficiency of the thermal collector (presented in Figure 10b) was assumed from the data presented in Figure 10c. The values of PV module efficiency were derived from the data presented in Figure 10, as ~ΔT = 10ºC was observed during the field experiment. The calculated maximum values of ηPV, ηc and ηPVT efficiencies and optimal (i.e. corresponding to the maximum value) solar altitude angle, α, for a PVT collector are shown at Table 4. It should be underlined that the values of PVT efficiency do not vary much under laboratory conditions and are close to the maximal values under field conditions. Generally good agreement between the maximal values of efficiencies can be observed, however significant difference in the optimal tilt angle needs to be investigated and addressed in the future. The optimal value of the solar altitude angle for ηc is the smallest considered in this study, both under field and laboratory conditions. In both cases, the PV efficiency grows with α. We recalculated altitude angle to the tilt angle (it is the angle between receiver and horizontal position) to compare results in literature. In the literature there are only tilt angle of receivers. The results are in agreement with the earlier reported works [20]. Calculated values of PVT collector efficiency (ηPVT) were compared with the values obtained from the computer program Polysun for the period during the autumnal equinox at 12.00 for solar altitude angles in the range 45 -90º. The PVT module efficiency reaches its maximum at α = 90º; the results under laboratory conditions show a similar tendency, although this is screened by oscillations due to weather conditions (e.g. wind and rain). This fact indicates the need for longer-term outdoor observation, necessary for temporary disturbances elimination Nevertheless one may confirm that, the simulated values do not vary much in the range of the considered solar altitude angle, α.

Conclusions
In this study, a PVT collector under field and laboratory conditions was examined. Furthermore, a PVT system was simulated using the demonstration version of Polysun software. Through the research, the relationship between orientation of the PVT collector and thermal and electric power generated was investigated. The maximum values of efficiencies for PV module (ηPV), solar thermal-collector (ηc) and PVT hybrid collector (ηPVT) under different conditions were determined. It was found that the efficiency of the PV module increases with the altitude angle both in laboratory and field experiments. In the case of field experiments, the range of altitude angle was limited (40º;65º). The efficiency of the thermal collector varies slightly in the considered range of altitudes angles and peaks at the lowest considered angle. However, the efficiency of the PV module is more significant overall. Furthermore maximum values of efficiencies for PV, PVT and thermal collector corresponds at the satisfactory level in experiments performed under field and laboratory conditions. The results are in agreement with the earlier reported work on the relation of the tilt angle and efficiencies of PVT modules [21][22][23]. The simulation results using the Polysun software agree reasonably well with experimental data received under laboratory conditions. Greek Letters α angle of incidence, º βτ transmittance-absorptance product, εsys solar yield, % ΔT difference between collector inlet fluid temperature and the ambient air temperature, K ηc efficiency of solar thermal collector, % ηPV efficiency of the PV module, % ηPVT efficiency of PVT hybrid collector, % As stated in the Introduction, increase in PV module temperature results in a significant efficiency drop. In order to check the impact of temperature on performance of the studied PV, experiments were carried out using a constant light intensity source of 1000 W/m 2 with cell temperature in the range 20-60ºC. From Figs 12a and 12b it is clearly seen that the temperature of the PV module significantly influences the current-voltage characteristic and module efficiency. It was calculated that the forward voltage of the solar cell decreases ~2 mV/K, with temperature increase. The results are in agreement with the earlier data for silicon solar cells [24][25][26][27][28][29].  This experiment was conducted at the Department of Semiconductor and Optoelectronic Devices, Lodz University of Technology [30]. Average surface temperature of the solar hybrid system working without liquid flow is ~82 ºC and average surface temperature the same solar hybrid system after 15 minutes of liquid flow is ~42 ºC.