Analysis of a Residential Photovoltaic-Thermal (PVT) System in Two Similar Climate Conditions

: Photovoltaic-thermal (PVT) panels combine solar thermal and photovoltaic technologies and generate simultaneously both heat and electricity. This paper looks at the potential of integrating these systems into small domestic prosumer households for the climates of Bucharest, Romania, and Strasbourg, France. First, some brief background information on PVT systems and the concept of prosumers is introduced, highlighting their features as well as the solar energy market setting in Romania and France. Next, a PVT system is proposed for a given household consumer in Strasbourg and Bucharest with the variable weather conditions corresponding to the two locations. The PVT system and the coupled consumer are modelled in TRNSYS (v17, Thermal Energy System Specialists, Madison, USA). A performance analysis is carried out in order to establish the daily instantaneous energy output and the annual energy production. The results indicate a 10–12% better performance in Bucharest compared to Strasbourg due to slightly better weather conditions. The system e ﬃ ciency was assessed through various methods (ﬁrst law e ﬃ ciency and primary energy saving). Depending on the method used, the location and time of year, the results vary from 15% for the ﬁrst law e ﬃ ciency to 90% for the primary energy saving e ﬃ ciency. The most suitable e ﬃ ciency assessment method for this study was found to be the primary energy saving method, as it takes into account the regional di ﬀ erences in energy production. This study concludes that the Romanian PVT market has a good potential for adopting the technology, especially since it is currently less mature than in France.


Introduction
Solar energy is the most wide-spread type of renewable energy, due to its multiple benefits: It is easily available in many locations, has low operational costs and can be adapted to small, decentralized systems. In theory, there is sufficient solar radiation reaching the Earth to meet 10,000 times the current energy needs of the world [1].
The established solar technologies currently available on the market can be broadly classified into photovoltaic (PV) panels and passive solar-thermal panels (ST). Hybrid photovoltaic-thermal systems (PVT) combine these two technologies and maximize their benefits into one single piece of equipment, producing both electricity and useful heat (Figure 1). PVT systems are a type of micro-cogeneration technology that can be very efficiently integrated into individual households. This can achieve decentralized production of clean heat and energy, maximizing the South facing available space and obtaining an overall better payback of the investment, compared to separate side-by-side PV and ST systems [2].

44
The current social trend of environmental responsibility, combined with the various financial 45 incentives available for renewable energy (Feed-in Tariffs, Green Certificates), lead to an increasing 46 number of individual consumers that wish to produce and utilize their own energy ('prosumers').
On this market there is a great potential for introducing the PVT technology in order to meet their

•
Independence from the regulated grid and safety against fluctuating energy prices;

81
• Energy losses and transport costs are reduced;

82
• Financial benefits from subsidies such as green certificates and 'Feed in Tariffs'.

83
The PVT systems were investigated and proved efficient for hot tropical climates, in both 84 experimental numerical studies [8,9]. Currently, in Europe there is also an active interest in the The current social trend of environmental responsibility, combined with the various financial incentives available for renewable energy (Feed-in Tariffs, Green Certificates), lead to an increasing number of individual consumers that wish to produce and utilize their own energy ('prosumers'). On this market there is a great potential for introducing the PVT technology in order to meet their energy demand with clean and on-site produced heat and power.
Residential PVT systems and their applications in various climate conditions have been of research interest in the past years, with multiple reviews and state of the art papers on this topic [3,4]. A study has been carried out to assess the performance of building integrated PVT (BIPVT) compared to simple building integrated PV (BIPV) in the context of New Delhi, India [5], and also to analyze various types of PV cells: Mono-crystalline and amorphous silicon. It was found that although the first was more efficient, the latter is more economical for the regional context. Another study [6] compares the performance of PVT panels in Athens, Munich and Dundee, and, as expected, Athens offers the most promising conditions. Similar to Romania, the PVT market in Greece is still under development and offers great opportunity for development. An experimental study [7] also carried out for India, showed that PVT system performed better than the simple PV and solar panels, which reinforces the interest for further developing the PVT technology.
This paper uses the software TRNSYS to simulate and analyze the energy output and performance of PVT panels in a small sized prosumer household, comparing two different climate conditions: Bucharest, Romania, and Strasbourg, France. The continental temperate climates are not the ideal locations for PVT systems, as it can be noticed from the literature, and more arid and hot areas are preferred. However, there can be a potential growing market even in colder locations. An interesting comparison can be carried out between France, where the market is developed and there are several manufacturers and distributers of PVT products available, and Romania, where the market is under-developed. Another novelty of this study is related to the explicit TRNSYS model which is analyzed by two performance indicators: First law efficiency and primary energy saving efficiency.

Residential Prosumers and the Energy Market
A particularly important growing market for renewable energy is the individual self-consumers (households or businesses). Self-consumption represents the production and utilization of energy on-site, which is the most efficient way of using renewable sources. Decentralized systems have the advantage of generating power where it is needed, and are more efficient in terms of transportation losses, since the consumer is close to the energy source. They are also more flexible, offer independent access in remote places and can benefit from localized smart micro-grids. An emerging term for these self-producers is 'prosumers', which is a combination of 'producer' and 'consumer'. More and more consumers are inclined to produce their own clean energy on-site by using renewable sources, driven by various motivations: • Independence from the regulated grid and safety against fluctuating energy prices; • Energy losses and transport costs are reduced; • Financial benefits from subsidies such as green certificates and 'Feed in Tariffs'.
energy efficiency of 4-12% higher compared to the same cell in an individual PV panel [18], due 112 to the fact that the cell is cooled through the extraction of the thermal energy.

113
• A significant amount of the Sun's spectral emissivity on a PV panel is lost in the form of heat 114 (either by non-absorption or thermalization) ( Figure 3). This heat can be recovered and used for 115 domestic hot water (DHW) or heating applications, which leads to increasing the spectrum of 116 operation of the panels and improving the rate of solar energy conversion.

117
• The global efficiency of the system is increased by combining the electrical and thermal 118 components, since the global efficiency can be described as the sum between the electrical and 119 thermal efficiency of the system, as shown in Equation (1). Furthermore, PVT reaches a higher 120 reduction of the non-renewable primary energy as well as a higher production of solar thermal 121 energy compared to the two separate systems (PV and ST) [19].

PVT Systems
Hybrid PVT systems benefit from a number of performance advantages compared to their individual solar thermal and photovoltaic counterparts:

•
The electrical efficiency of PV cells decreases with a temperature coefficient of 0.3-0.9%/ • C above the standard test condition (STC) of 25 • C. Thus, the PV cells in hybrid PVT systems show an energy efficiency of 4-12% higher compared to the same cell in an individual PV panel [18], due to the fact that the cell is cooled through the extraction of the thermal energy. • A significant amount of the Sun's spectral emissivity on a PV panel is lost in the form of heat (either by non-absorption or thermalization) ( Figure 3). This heat can be recovered and used for domestic hot water (DHW) or heating applications, which leads to increasing the spectrum of operation of the panels and improving the rate of solar energy conversion.

•
The global efficiency of the system is increased by combining the electrical and thermal components, since the global efficiency can be described as the sum between the electrical and thermal efficiency of the system, as shown in Equation (1). Furthermore, PVT reaches a higher reduction of the non-renewable primary energy as well as a higher production of solar thermal energy compared to the two separate systems (PV and ST) [19].
• In urban locations, where South facing roof and façade space are limited, this solution is ideal for fulfilling or partially contributing towards the heating and electricity needs of the consumer. In addition, the costs associated with transport, installation, and maintenance are significantly reduced by implementing a single piece of equipment instead of two separate ones. In addition, the costs associated with transport, installation, and maintenance are significantly represents the excess energy that is lost as heat and the red line represents the band-gap of Si.

135
configuration investigated in this study is illustrated in Figure 4.

136
The model developed for this study is a water-based PVT system, with a direct flow geometry 137 of the collector (Figure 4a), roll bond channels (Figure 4b), and mono-c-Si PV cells. The water-based 138 exchanger was chosen due to its better heat transfer capacity compared to air-based systems, and in 139 order to fulfill the domestic hot water (DHW) needs of a prosumer household. The direct flow 140 geometry of the channels provides optimal experimental performance and geometrical simplicity [4].

141
The roll bond channels are thermally efficient, easy to produce, relatively inexpensive and very 142 flexible in terms of geometry [4]. The c-Si PV cells are the most widely used on the market, with the 143 mono-crystalline cells providing better efficiency than the poly-crystalline cells.   Energy flow diagram of a PVT system: The blue area represents the energy that can be extracted as electricity, the grey area represents the photons with a low level of energy that are unable to excite the electrons to the conduction band and are lost through non-absorption, the green area represents the excess energy that is lost as heat and the red line represents the band-gap of Si.
The PVT system can be configured in multiple ways: Flat or concentrated, with a liquid, air or dual liquid-air working fluid. The geometry of the thermal collector can also take multiple forms (harp, spiral, direct flow, fractal) and can be designed by various manufacturing methods (sheet and tube, roll bond, rectangular channels). Aste et al. [4] carries out a comprehensive review on the various geometrical, thermo-physical, and manufacturing solutions for PVT systems. The configuration investigated in this study is illustrated in Figure 4. In addition, the costs associated with transport, installation, and maintenance are significantly 125 reduced by implementing a single piece of equipment instead of two separate ones.

136
The model developed for this study is a water-based PVT system, with a direct flow geometry geometry of the channels provides optimal experimental performance and geometrical simplicity [4].

141
The roll bond channels are thermally efficient, easy to produce, relatively inexpensive and very 142 flexible in terms of geometry [4]. The c-Si PV cells are the most widely used on the market, with the 143 mono-crystalline cells providing better efficiency than the poly-crystalline cells.   The model developed for this study is a water-based PVT system, with a direct flow geometry of the collector (Figure 4a), roll bond channels (Figure 4b), and mono-c-Si PV cells. The water-based exchanger was chosen due to its better heat transfer capacity compared to air-based systems, and in order to fulfill the domestic hot water (DHW) needs of a prosumer household. The direct flow geometry of the channels provides optimal experimental performance and geometrical simplicity [4]. The roll bond channels are thermally efficient, easy to produce, relatively inexpensive and very flexible in terms of geometry [4]. The c-Si PV cells are the most widely used on the market, with the mono-crystalline cells providing better efficiency than the poly-crystalline cells.

PVT System
The PVT system modelled in this paper is illustrated in Figure 5. The system comprises of a PVT collector connected to a water storage tank through a heat exchanger, a circulation pump, connection to the household water main and an inverter/regulator with a battery bank. The input in the system is the meteorological data: Solar irradiation (G (W/m 2 )), ambient temperature (T amb ( • C)) and wind speed (W (m/s)). The temperature at the outlet of the collector is T out_coll , the temperature of the tank is T tank , the temperature of the water main that feeds the tanks is T main , and the temperature at the inlet of the collector is T in_coll , all in • C. The PVT acts as a pre-heater when the temperature reached by the tank is not sufficient for providing domestic hot water.

157
The PVT panel is assumed to be a flat plate collector, with mono-c-Si PV cells and a direct flow 158 roll bond thermal absorber, with the technical characteristics summarized in Table 1.  The PVT panel is assumed to be a flat plate collector, with mono-c-Si PV cells and a direct flow roll bond thermal absorber, with the technical characteristics summarized in Table 1.

Meteorological Conditions
The output of a PVT system is dependent upon three meteorological parameters: Ambient temperature, irradiation, and wind speed. The solar irradiation has the most significant impact on the performance of a PVT system, as it determines the electrical output, but also influences the temperature of the materials through radiative heat transfer. The temperatures in the PVT layers are also dependent on the ambient temperature, through convective and conductive heat transfer processes in the solid and liquid layers of the system, but also on the speed of the wind which determines the convective heat losses at the surface and at the back of the panel. The yearly profile of each parameter is obtained from Meteonorm Database.
The total radiation comprises of three components: Direct beam radiation, diffuse radiation, and reflective radiation. Direct radiation is the direct incident beam radiation that hits the surface of the panel, the diffuse radiation refers to the radiation that has been scattered through the molecules of the atmosphere, and reflective radiation (also referred to as albedo) is the radiation reflected by the surroundings.
The meteorological conditions of both Strasbourg, France, and Bucharest, Romania are used in parallel for the simulations. This allows a comparison of the performance of the PVT system in both regions. The climate of Romania can be described as temperate continental, with cold, cloudy winters and frequent snow and fog, and sunny summers with frequent showers and thunderstorms. In the southern region, where Bucharest is located, the winters are milder and the summers hotter. France's climate is temperate, with four specific regions. Alsace, where Strasbourg is located, is characterized by a continental climate which harbors cold winters and hot summers. Although, in broad terms, both cities have a similar climate (hot summers and cold winters), there are some small differences in the average values of the meteorological parameters, which are summarized in Table 2. Figures 6 and 7 illustrate a sample of the daily total radiation for a typical summer and winter week, respectively, in Strasbourg and Bucharest. The two cities have a similar average rage of values during the summer days, of 700-900 W/m 2 . The daily curve of ambient temperature for the two cities is shown in Figures 8 and 9. It can be observed that summer temperatures are slightly higher in Bucharest compared to Strasbourg, while the Romanian winters are colder. Figures 10 and 11 show the variation of wind speed during the summer and winter periods, respectively. On average, the wind is stronger in Strasbourg, with only a few higher peaks in Bucharest. Overall, the wind and temperature data suggest that the PVT will perform better in Bucharest, while the solar radiation is more or less equivalent. The total radiation comprises of three components: Direct beam radiation, diffuse radiation, and 170 reflective radiation. Direct radiation is the direct incident beam radiation that hits the surface of the panel, the diffuse radiation refers to the radiation that has been scattered through the molecules of the atmosphere, and reflective radiation (also referred to as albedo) is the radiation reflected by the surroundings.

205
The profile used for this simulation was created using the DHW-calc software (v1, University of profiles for European countries for any required period of time and timestep [22,23]. This software 208 has been successfully used and validated in multiple research papers [21,24,25]. The user can define function of TRNSYS, shown in Figure 13.

214
The temperature level for DHW was set to 45 °C, while the cold water from the main is assumed

DHW and Electricity Demand
In order to carry out an accurate simulation, a realistic daily time dependent profile of the household DHW demand is required. The daily profile of DHW is dependent upon various parameters: The number of inhabitants, appliances, ambient conditions, seasonal variations, the average daily hot water volume, and the yearly demand [21].
The profile used for this simulation was created using the DHW-calc software (v1, University of Kassel, Kassel, Germany) developed for the IEA-SHC Task 26, which can generate realistic DHW profiles for European countries for any required period of time and timestep [22,23]. This software has been successfully used and validated in multiple research papers [21,24,25]. The user can define a number of parameters, such as the total mean daily water volume, probability distribution, flow rates, holidays, etc. Figure 12 shows an example of a three-day profile with a 1 h time step resolution for 200l/day average consumption, which was used for this simulation. The profile generated by DHW-calc is different for weekdays and for weekends, and the profile is included in the DHW Load function of TRNSYS, shown in Figure 13.
The temperature level for DHW was set to 45 • C, while the cold water from the main is assumed constant across the year at 10 • C.     An electrical demand curve was generated using a forcing function in TRNSYS and considering weekday and weekend variations, as well as holidays, and the system includes a battery bank. The electricity demand also includes the power consumption of the pump for circulating the fluid. The excess of produced electricity is exported to the grid when the battery is full, and when the battery load is less than the demand, electricity is bought back from the grid.
The same DHW and electricity demand profiles were used for both simulations, in Strasbourg and Romania. The DHW profile generated with DHW-calc has been developed to be applicable for any European country, and the electricity demand was also assumed to be similar. This way, a comparison based solely on the input meteorological data can be achieved.

TRNSYS Model
TRNSYS (transient system simulation tool) is a flexible software that can simulate the behavior of dynamic transient systems. A TRNSYS model was developed to replicate the system in Figure 13, and it includes: Weather Data (Type 15), PVT collector (Type 50), Stratified Storage Tank with a capacity of 100l (Type 4), Single Speed Pump (Type 114), Inverter/Charge controller (Type 48), a Battery Bank (Type 47) and Controllers (Type 2). The Type 15 component is a tool that reads weather data either from the built-in database or from a user provided spreadsheet, and processes it to provide the input weather data for the simulation (temperature, radiation, wind speed). The storage tank (Type 4) models a tank that can be subjected to thermal stratification, by establishing the number of thermal nodes N. If N is equal to one, the tank is modeled as fully-mixed with no stratification, which was assumed for the purpose of this study. Type 114 is the component for a constant speed pump that maintains a constant fluid outlet mass flow rate. The pump control strategy modelled by a Type 2 component is based on the temperature of the liquid in the storage tank and the DHW demand. The collector pump circulates water from the tank when the energy can be collected, and the tank is emptied by the DHW load and refilled when it reaches a low volume. Type 48 models both the regulator of the system, which distributes DC power from the solar cell array to and from a battery, and the inverter converts the DC power to AC. The electrical controller decides whether to use the electrical load for charging the battery or for meeting the demand. Type 47 models the battery bank of the system, which is based on a lead-acid type of battery and can output data on the variation of the state of charge over time and the rate of charge or discharge.
The simulation assumes that no heat losses occur on the pipes, the optical properties of materials are constant and no surroundings partial shading or dust is considered. The electrical flow is shown with purple arrows, the liquid flow in blue and the dotted arrows connect the control circuit and energy demand. The PVT component has eight operational modes, which vary the method of calculation for the thermal loss coefficient and transmittance-absorptance factor. This model used mode four, which calculates the parameters according to definitions of Klein (1976) [26].
The numerical model behind the PVT collector (Type 50) has been developed based on the work of Florschuetz [27], using the Hottel-Whillier-Bliss equations. Starting from the steady state energy balance of a system: where T ( • C) is the temperature of the element dx, k is the thermal conductivity (W/mK), δ is the thickness of the element (m), U L is the thermal loss coefficient (W/(m 2 ·K)), T a is the ambient temperature ( • C), S is the absorbed solar energy (W) and q e is the electrical output (W), the paper defines the useful thermal output of the PVT collector as: where A c is the area of the collector (m 2 ), F R is the heat removal factor (-) and T f,i is the inlet temperature of the fluid ( • C). The electrical output of the system is defined as: where η a is the efficiency of the cell at ambient temperature (%), α is the effective absorbance of the cell (-), η r is the reference cell efficiency (%) and β r is the temperature coefficient of the cell (%/K). The temperature at the outlet of the collector can be approximated as: where V flow is the volumetric flow in the collector (m 3 /s), ρ is the fluid density (kg/m 3 ) and c f is the heat capacity of the collector fluid (J/(kg·K)).
This model is a simplified version of a complete numerical dynamic analysis that can be performed on a PVT based on the heat transfer in each layer of the panel [20,21], where a system of ordinary differential equations is solved simultaneously to find the temperatures at any time of the simulation. The detailed models are used for in-depth layer analysis and geometrical, optical or thermo-physical optimizations and sizing. For the purpose of a quantitative analysis of overall daily, weekly and annual production of energy, the TRNSYS Type 50 model was validated and proved accurate.
The PVT component requires a set of input parameters that depend on the specific PVT that is to be modelled. These can be found in the manufacturer's specifications and refer to: Collector fin efficiency, plate absorbance, number of glass covers, optical efficiency, collector plate emittance, loss coefficient, temperature coefficient, STC temperature, and packing factor. Thus, the module can be customized to replicate a product available on the market.

273
The PVT component requires a set of input parameters that depend on the specific PVT that is 274 to be modelled. These can be found in the manufacturer's specifications and refer to: Collector fin 275 efficiency, plate absorbance, number of glass covers, optical efficiency, collector plate emittance, loss 276 coefficient, temperature coefficient, STC temperature, and packing factor. Thus, the module can be 277 customized to replicate a product available on the market.
where ̇ is the mass flow rate of the fluid (kg/s) and cf is the specific heat capacity (J/(kg·K)), which 288 in this case for Glycol water 40%/60% is 3600 J/(kg·K).
289 Figure 13. Components of the transient system simulation tool (TRNSY) model.

System Performance
The energy performance of the system can be computed in terms of thermal, electrical, and overall efficiencies (η TH , η EL , η overall ). The electrical efficiency is calculated as the ratio between the power generated by the system (P EL ) and the amount of solar radiation incident on the surface of the collector (AxG irr ): The standard equation for thermal efficiency is represented by the ratio between the amount of thermal energy generated by the system (Q TH ) and the solar radiation incident on the surface of the collector (AxG irr ): where . m is the mass flow rate of the fluid (kg/s) and c f is the specific heat capacity (J/(kg·K)), which in this case for Glycol water 40%/60% is 3600 J/(kg·K).
Another definition for the thermal efficiency was proposed by Bombarda et al. [28]. Here, a distinction is made between the electrical and thermal power of the system, and thus the electrical power production is subtracted from the total incident irradiation. This definition allows for a more accurate comparison between the thermal performance of a PVT and stand-alone solar thermal system.
From the point of view of the first law of thermodynamics, the global efficiency of the system (η GL ) can be calculated as the sum of the thermal and electrical efficiencies. This is also known as the first law efficiency: Another approach is to consider the difference in grade between the two types of energy, i.e., electrical energy is a higher form of energy compared to the thermal. Thus, due to the fact that a kW el cannot be directly compared to a kW th , some authors [19,29] propose the primary energy saving efficiency (η PES ) as a more accurate way of assessing the performance of the system: where η Tpower is the average efficiency of producing electrical power, and it depends on the country of reference. On average, it can be assumed to be 0.4 [19]. For Romania, the average efficiency of producing electrical power in 2016 was 0.532, while in France it was 0.601 [30]. These values were used for computing the primary energy saving efficiency in Section 5.

Simulation Results
The detailed simulations were run for the two significant one-week periods of summer and winter (7 Jan-13 Jan and 20 Jul-26 Jul) and for both the meteorological conditions of Bucharest and Strasbourg, with a 15-minute time step. In addition, an annual simulation was performed to determine the weekly electrical and thermal energy production for the duration of a year.
The electrical power output of the system for summer and winter is illustrated in Figures 14 and 15, respectively. The variation in the output follows closely the variation in solar radiation. The ambient temperature also has an effect on the power output: As the operating temperature increases, the efficiency of the PV cells drops. Another definition for the thermal efficiency was proposed by Bombarda et al. [28]. Here, a 290 distinction is made between the electrical and thermal power of the system, and thus the electrical 291 power production is subtracted from the total incident irradiation. This definition allows for a more 292 accurate comparison between the thermal performance of a PVT and stand-alone solar thermal 293 system.
From the point of view of the first law of thermodynamics, the global efficiency of the system 295 (ηGL) can be calculated as the sum of the thermal and electrical efficiencies. This is also known as the 296 first law efficiency: Another approach is to consider the difference in grade between the two types of energy, i.e., 298 electrical energy is a higher form of energy compared to the thermal. Thus, due to the fact that a kWel 299 cannot be directly compared to a kWth, some authors [19,29] propose the primary energy saving 300 efficiency (ηPES) as a more accurate way of assessing the performance of the system: where ηTpower is the average efficiency of producing electrical power, and it depends on the country of 302 reference. On average, it can be assumed to be 0.4 [19]. For Romania, the average efficiency of 303 producing electrical power in 2016 was 0.532, while in France it was 0.601 [30]. These values were 304 used for computing the primary energy saving efficiency in Section 5.

306
The detailed simulations were run for the two significant one-week periods of summer and

310
The electrical power output of the system for summer and winter is illustrated in Figures 14 and   311 15, respectively. The variation in the output follows closely the variation in solar radiation. The 312 ambient temperature also has an effect on the power output: As the operating temperature increases,     Tout_coll_Buch Tout_coll_Strasb T_tank_Stras T_tank_Buch 07-Jan 08-Jan 09-Jan 10-Jan 11-Jan 12-Jan 13-Jan 14-Jan

322
Over a year, the weekly average temperature and solar radiation for Bucharest and Strasbourg 323 are illustrated in Figure 18. The solar radiation was averaged for the non-zero values, from sunrise to 324 sunset. Given these conditions, the thermal and electrical energy outputs are illustrated in Figure 19.

325
Overall, the annual energy produced by the system in the two cities is summarized in Table 3. Tout_coll_Buch Tout_coll_Strasb T_tank_Stras T_tank_Buch 07-Jan 08-Jan 09-Jan 10-Jan 11-Jan 12-Jan 13-Jan 14-Jan

322
Over a year, the weekly average temperature and solar radiation for Bucharest and Strasbourg 323 are illustrated in Figure 18. The solar radiation was averaged for the non-zero values, from sunrise to 324 sunset. Given these conditions, the thermal and electrical energy outputs are illustrated in Figure 19.

325
Overall, the annual energy produced by the system in the two cities is summarized in Table 3. Over a year, the weekly average temperature and solar radiation for Bucharest and Strasbourg are illustrated in Figure 18. The solar radiation was averaged for the non-zero values, from sunrise to sunset. Given these conditions, the thermal and electrical energy outputs are illustrated in Figure 19. Overall, the annual energy produced by the system in the two cities is summarized in Table 3. Assuming an annual energy consumption of 2000 kWh and a thermal energy for DHW of 6000 kWh TH , the table also shows the percentage of coverage for the household. It can be observed that the system performs better under the meteorological conditions of Bucharest, where it is capable of covering more than 50% of the annual consumption.
The results indicate that a 6 m 2 system would operate well in both climates, but with a 10-12% better performance in Bucharest. The climatic conditions in Bucharest are more suitable (higher radiation and temperature) during the summer but worse during the winter. The overall effect shows a slight advantage of PVT systems in Bucharest compared to Strasbourg.

330
The results indicate that a 6 m 2 system would operate well in both climates, but with a 10-12% 331 better performance in Bucharest. The climatic conditions in Bucharest are more suitable (higher 332 radiation and temperature) during the summer but worse during the winter. The overall effect shows 333 a slight advantage of PVT systems in Bucharest compared to Strasbourg.

338
Another investigated topic is the efficiency of the system in terms of electrical, thermal, and 339 global performance. As discussed in Section 4.4, the system performance can be described in multiple

330
The results indicate that a 6 m 2 system would operate well in both climates, but with a 10-12% 331 better performance in Bucharest. The climatic conditions in Bucharest are more suitable (higher 332 radiation and temperature) during the summer but worse during the winter. The overall effect shows 333 a slight advantage of PVT systems in Bucharest compared to Strasbourg.

338
Another investigated topic is the efficiency of the system in terms of electrical, thermal, and 339 global performance. As discussed in Section 4.4, the system performance can be described in multiple  Another investigated topic is the efficiency of the system in terms of electrical, thermal, and global performance. As discussed in Section 4.5, the system performance can be described in multiple ways: First law efficiency and primary energy saving efficiency. The thermal efficiency was computed according to Equation (8), by subtracting the electrical power from the total incident irradiation. The weekly efficiency of the system is calculated by taking into account the total weekly energy production (shown in Figure 19) and the total energy received from the sun to the surface of the panel during that week. Figure 20 illustrates the weekly electrical and modified thermal efficiency of the system. It can be observed that the electrical efficiency in both cases slightly drops during the summer, due to the increase in the operating temperature of the cells, while the thermal efficiency is at its maximum during the summer weeks, reaching 64% in Bucharest and 55% in Strasbourg. In addition, Table 4 summarizes the monthly energy production and efficiencies for both locations, for a better comparison. The monthly values of the energy production are calculated by adding the weekly production illustrated in Figure 19, and the efficiency values are obtained by averaging the weekly values shown in Figure 20. irradiation. The weekly efficiency of the system is calculated by taking into account the total weekly 343 energy production (shown in Figure 19) and the total energy received from the sun to the surface of 344 the panel during that week.
345 Figure 20 illustrates the weekly electrical and modified thermal efficiency of the system. It can 346 be observed that the electrical efficiency in both cases slightly drops during the summer, due to the 347 increase in the operating temperature of the cells, while the thermal efficiency is at its maximum 348 during the summer weeks, reaching 64% in Bucharest and 55% in Strasbourg. In addition, Table 4 349 summarizes the monthly energy production and efficiencies for both locations, for a better    All the global performance indicators described in Section 4.5 were computed and are described in Figure 21. The first law efficiency is simply the sum of the thermal and electrical efficiencies illustrated in Figure 20. Overall, a better performance in Bucharest can be easily observed. The efficiency calculated by the PES method is higher than the first law efficiency, reaching 90% for operation in Bucharest. Due to the fact that the average efficiency of producing electrical power in Bucharest is lower than in Strasbourg (0.532 compared to 0.601), the difference between the two cities in terms of PES increases further compared to the first law efficiency, which can be observed in the graph. The overall trend follows closely the trend of the thermal efficiency in Figure 20, with high values during the summer and low during the winter. The inverse variation of the electrical efficiency is insignificant and does not affect the general trend.  6. Discussion

369
The system presented in this paper couples a typical residential household with a PVT system

389
Energy Saving efficiency takes into account the regional differences in energy production, and can be 390 more useful to compare the efficiency of the system in two different locations, as is the case of this

Discussion
The system presented in this paper couples a typical residential household with a PVT system that generates both heat and electricity, in two similar climate conditions. According to the meteorological data, the two cities, Bucharest and Strasbourg, have similar overall weather patterns with hot summers and cold winters. However, there are some small dissimilarities in terms of wind speed, average ambient temperature, and total irradiation which would indicate that Bucharest is a more suitable location for achieving the best PVT performance. The simulations were run for representative seven-day summer and winter periods to assess the electrical power, temperature in the storage tank, and the temperature of the working fluid, and also for a yearlong period to assess the global annual performance. As expected, the electrical performance flows closely the global irradiance with little influence from the ambient temperature and wind speed. The tank is coupled with a DHW consumer, thus the temperature does not reach high values, even though the fluid at the outlet at the collector can reach up to 43 • C.
Overall, the yearly production of electrical and thermal energy, both supported by back-up solutions (battery bank and auxiliary heater respectively), indicates that the energy requirements of an average family household could be half fulfilled over the course of one year, with a slight advantage in the Bucharest climate compared to Strasbourg. This was expected due to the slightly improved weather conditions. In terms of system performance, there are various ways of assessing it and it is debatable which is the most accurate and suitable method. The first law efficiency is the simplest and widely used method, where there is no distinction made between the electrical and thermal power. The Primary Energy Saving efficiency takes into account the regional differences in energy production, and can be more useful to compare the efficiency of the system in two different locations, as is the case of this study.
Some of the limitations of this simulation and its results are: The simplicity of the PVT model compared to a full finite element analysis, the fact that all the components are pre-defined in the TRNSYS software and thus a full customization is difficult, the consumer profiles were obtained by means of forcing functions and statistical methods and can vary significantly according to the individual end user and the location of the system.

Conclusions and Perspectives
This paper has performed a dynamic simulation of a PVT system for a small size domestic household using the software TRNSYS. The system includes both thermal and electrical storage, and is coupled to the simulated electrical and DHW demand of the household. The results show that, with a battery storage system, the household demand can be 58% covered in Bucharest and 48% in Strasbourg. The hot water heating system, provided that an auxiliary heater is utilized, can also be covered 51% and 41%, respectively.
Overall, it appears that small scale PVT is a promising solution for maximal harvesting of the solar energy. Residential prosumers are the target market of this technology, which is quickly developing in the Western European countries, but is lagging behind in the East. More public awareness and demonstrative projects providing proof of concept are required to further push this technology on the mainstream market, next to conventional PV and solar thermal.
An analysis of the costs, profitability, and return period of the system would be an interesting direction for further research, including consideration of the CO 2 emission savings, from the cradle to the grave of the panel. This can be proved challenging due to the fluctuating market prices of energy, a non-homogeneity in the quoted prices of PVT on the market and the difficulty in accurately accounting for the true costs of CO 2 emissions. Another issue to be further investigated is the coupling of the PVT panels with heat pumps for a more efficient utilization of the thermal energy and assessing the cost-benefits compared to those of a simple PVT system. This analysis can be carried out for a number of different consumer profiles, not only residential households, but also office buildings, schools and hospitals, and the system optimized to fit the demand curve. Finally, an energetic study of the PVT system would also be an interesting research direction.

Conflicts of Interest:
The authors declare no conflict of interest.