Energy Reports xxx (xxxx) xxx

Contents lists available at ScienceDirect

Energy Reports
journal homepage: www.elsevier.com/locate/egyr

Numerical model and parametric analysis of a liquid based hybrid

photovoltaic thermal (PVT) collector
∗
Madalina Barbu a,b , Monica Siroux b , , George Darie a
a
Department of Energy Generation and Use, Faculty of Power Engineering, University POLITEHNICA of Bucharest, Bucharest, Romania
b
INSA Strasbourg ICUBE, University of Strasbourg, Strasbourg, France

article info a b s t r a c t

Article history: A hybrid photovoltaic thermal (PVT) panel is a module in which the photovoltaic (PV) layer is not
Received 18 December 2020 only producing electricity, but also operates as a thermal absorber. As a result, thermal and electrical
Received in revised form 24 June 2021 energy are being produced simultaneously, operating as a micro-cogeneration equipment. As in any
Accepted 19 July 2021
cogeneration system, there is tight link between the electrical and thermal performance and it is
Available online xxxx
dependent on multiple parameters: climate conditions, thermo-physical, geometrical and electrical
Keywords: properties. This paper investigates the effect of the variation of several of these parameters on the
Renewable energy electrical and thermal performance, as well as on the global output. In order to achieve this, a dynamic
Solar energy numerical model is proposed, which simulates the heat exchange between the layers of the PVT
PVT panel. The model was applied to two different climatic conditions: Bucharest, Romania and Strasbourg,
Micro-cogeneration
France, in order to assess and compare their behavior and performance. The simulation computes
Heat transfer
the temperature of each layer at any particular time, and a slightly higher outlet temperature of the
Dynamic simulation
working fluid can be observed in Bucharest during the summer, and in Strasbourg during the winter.
The model is validated against data from the literature and can be applied to any climatic conditions
and adapted for multiple geometrical and thermo-physical configuration. Next, a one-factor-at-a-time
parametric analysis is carried out in order to assess the impact of various parameters on the electrical,
thermal and global efficiency. The results showed in most cases a compromise between the electrical
and thermal performance: in terms of wind speed and insulation, the thermal benefits of low wind
and high insulation overcome the decrease in electrical efficiency. The packing factor was found to
be optimum when maximized, as the electrical benefits are more significant than the thermal loss.
The width of the channels in the heat exchanger should also be maximized as far as technologically
possible for best performance.
© 2021 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction higher overall conversion efficiency of the incident solar radiation

compared to individual PV or ST panels.
Solar energy is among the most promising renewable sources, PVT collectors are a type of micro-cogeneration technology
playing an essential role in the energy mix of the future. Energy that can be very efficiently integrated into domestic households
can be extracted from the sun in the form of heat or electricity, and can achieve decentralized production of clean heat and en-
by means of the established solar technologies currently available ergy for the end consumer. Furthermore, PVT systems can obtain
on the market: solar thermal (ST) panels and photovoltaic panels an overall better payback of the initial investment compared to
(PV). Several comprehensive reviews on the latest solar technolo-
individual side-by-side PV and ST panels (Ramos et al., 2017;
gies has been carried out (Ahmadi et al., 2018; Khatibi et al., 2019)
Barbu et al., 2019c). Along with the financial benefits, PVT sys-
comparing and contrasting some of the technologies in terms of
tems have several additional advantages. Firstly, by extracting the
economic feasibility and CO2 emissions.
heat losses that occur during the photovoltaic conversion through
Hybrid photovoltaic–thermal (PVT) collectors combine and
maximize the benefits of these two types of energy extraction the processes of thermalization and non-absorption, the overall
technologies into one single equipment, and can produce si- system efficiency increases. Also, the electrical efficiency of the
multaneously electrical and thermal energy. This allows for a constituent PV cells is improved as a result of a lower operating
temperature. Normally, the PV cells decrease in efficiency when
∗ Corresponding author. operating in high temperatures. Most of them lose about 0.3–
◦
E-mail addresses: madalina.manole0401@upb.ro (M. Barbu), 0.9% for each C above their standard operating temperature of
◦
monica.siroux@insa-strasbourg.fr (M. Siroux), geo@energy.pub.ro (G. Darie). 25 C (Dubey et al., 2013). Thus, by cooling the cells through the

https://doi.org/10.1016/j.egyr.2021.07.058
2352-4847/© 2021 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Please cite this article as: M. Barbu, M. Siroux and G. Darie, Numerical model and parametric analysis of a liquid based hybrid photovoltaic thermal (PVT) collector. Energy
Reports (2021), https://doi.org/10.1016/j.egyr.2021.07.058.
M. Barbu, M. Siroux and G. Darie Energy Reports xxx (xxxx) xxx

heat collector, the PV cells in PVT systems can have a 4%–12% bet- the least squares support vector machines method performed
ter performance compared to the stand-alone PV module (Ramos better than the other two, and is quite suitable for cases when
et al., 2017). By combining the electrical and thermal components, experimental measurements are difficult or costly (Ahmadi et al.,
the global efficiency of the system is improved, since the global 2020; Zamen et al., 2019).
efficiency ηGL can be written the sum between the electrical Another parametric study on a solar cooling system was done
efficiency ηEL and the thermal efficiency ηTH . for the weather conditions of Athens, Greece (Tzinnis et al., 2016),
This solution is particularly interesting and cost-efficient for and after optimization it achieved a solar coverage ratio up to
small scale applications, where heat and electricity are required 89.85%. An interesting research (Nhut et al., 2020) performed a
simultaneously, and roof and facade space are limited, espe- parametric study of a solar-assisted house heating system cou-
cially in urban locations. In addition, several studies (Barbu et al., pled with a seasonal underground thermal energy storage tank,
2019c; Herrando and Markides, 2016) investigate the economic evaluating the annual energy production as a function of some of
benefits of the PVT systems compared to individual side-by- the main system parameters, such as the insulation thickness of
side PV and solar-thermal panels, showing that the PVT panels the tank and the size of the tank.
may achieve an improved payback period and return on invest- Another parametric study was carried out on a grid-connected
ment. The economic benefits are, however, highly sensitive to PV-battery systems investigating the effects of the weather, price
the energy market price fluctuations and the government in- of electricity, feed-in tariff, system cost, and electrical specifica-
centives available for renewable energy (Feed in Tariffs, Green tions from a technical–economical point of view (Khalilpour and
Certificates). Vassallo, 2016). Another study (Arsalis and Alexandrou, 2015) in-
The PVT panels have been first introduced in the 1970s’ (Kern vestigates a solar-heating-and-cooling system for a single-family
and Russell, 1978) and have been slowly developing ever since, household in terms of cost savings.
with increased research interest in the last decade. However, the This paper proposes an explicit dynamic numerical model of
PVT technology is not yet well-established in neither academic a particular configuration of PVT panel (water-based roll-bond
research field or on the commercial market. In recent years, collector with Mono-Crystalline Silicon PV cells). The thermal
the research on PVTs is focused on technology development, model is able to evaluate the temperatures at each layer inside
modeling, HVAC integration, and optimization. On one hand, re- the collector, taking into account all the heat transfer processes
search can focus on the development of PVT collectors in order between the layers or with the environment. Based on the output
to improve the design solutions and to increase the efficiency of temperatures, the system performance indicators can then be
solar energy conversion (Papoutsis et al., 2017; Bombarda et al., assessed for various operational and constructive parameters. As
2016; Guo and Zheng, 2017). On the other hand, the research a case study, the model is applied to the climatic conditions of
can investigate the integration of PVT technologies in buildings Bucharest, Romania and Strasbourg, France. The novelty of this
to assess the operative performances and actual energy-economy paper stands in the easy to follow and replicate methodology for
benefits (Chow et al., 2005; Prabhakant and Tiwari, 2010; Agrawal simulating the operation of the panels, along with the comparison
and Tiwari, 2010). This study falls in the first category, aiming to of its performance in two climatic conditions. The one-factor-at-
develop an explicit PVT model that can assess the suitability of a-time (OFAT) analysis carried out based on the numerical model
various design parameters in any ambient conditions, which can reveals some useful design information for improving the perfor-
consequently lead to improving performance of the panel. mance of the collectors. Other aforementioned parametric studies
Various configurations of PVT have been developed, and their have been carried out previously (Chow et al., 2005; Aste et al.,
performance varies according to multiple parameters: optical, 2015; Han et al., 2006), but the novelty of this research stands
geometrical and thermo-physical properties, type of fluid, wind in the application of the model in various climatic conditions,
speed and direction, solar radiation and ambient temperature. besides from the OFAT analysis, and the analysis of not only the
A number of models have been identified in the literature for electrical and thermal performance, but also of the global and
various configurations and climatic conditions. Rejeb et al. (2015) primary energy saving (PES) efficiencies.
develops a sheet-and-tube PVT model and simulates its perfor-
mance in Monastir, Tunisia. Aste et al. (2015) proposes a model 2. Description of PVT collector
to evaluate the daily and annual yield of a thin film roll-bond PVT
collector in Milan, Italy. Chow et al. (2005) develops a seven-node The PVT collector can take various configurations: liquid, air or
model and simulates a sheet and tube panel. dual liquid–air based, flat or concentrated. The thermal collector
A study by Brottier (2019) follows 28 PVT panels installations can also have multiple geometries (harp, direct flow, fractal) and
in France, Portugal and Switzerland. The instantaneous response manufacturing methods (sheet and tube, roll bond, rectangular
of a PVT collector to various climatic conditions is analyzed in a channels). A typical PVT system is obtained by coupling a solar
paper by Armstrong and Hurley (2010). Some parametric studies thermal collector at the back of a PV panel. A comprehensive
have been carried out for establishing the influence of mass flow review is carried out by Aste et al. (2014), which summarizes the
rate (Chow et al., 2005), thickness of insulation, wind speed and various geometrical configurations, thermo-physical properties
tube spacing (Rejeb et al., 2015). A box-channel model was devel- and manufacturing solutions for PVT systems. One of the most
oped and the influence of the channel dimensions was assessed common configurations, which was also used in this study, is
in a study by Han et al. (2006). Most of these models are based described below.
on the same theoretical background, starting from the classical The PVT chosen panel is liquid based, with a direct flow
numerical model developed by Florschuetz (1979), based on the geometry of the collector and roll bond channels (Fig. 1). The
Hottel–Whillier–Bliss equations with various improvements, and liquid-based exchanger was chosen due to its better heat transfer
the methodology of calculating the temperatures in the PVT lay- capacity compared to the air-based systems, in spite of its higher
ers is based on the exchange of energy between the layers and level of complexity. In addition, the liquid systems can be utilized
with the exterior. directly for preheating domestic hot water (DHW). The direct
A novel approach to simulating the operation of PVT panels is flow channel was chosen for their simplicity and well proved ex-
by using machine learning methods of artificial neural networks, perimental performance. The roll bond fabrication method is easy
least squares support vector machines and neuro-fuzzy. By com- to achieve, very flexible in terms of geometry and relatively cheap
paring the performance of the three methods it was found that to produce (Aste et al., 2014). The PV cell is Mono Crystalline
2
M. Barbu, M. Siroux and G. Darie Energy Reports xxx (xxxx) xxx

Fig. 1. Cross section of the PVT collector (Barbu et al., 2019a).

Silicon, with a reference electrical efficiency (ηSTC ) of 15% at the

standard testing conditions (STC) - operating temperature (Tref ) of
25 ◦ C, solar radiation of 1000 W/m2 and wind speed 0 m/s. The
temperature coefficient (βT ) is 0.5%/K. The relationship between
the instantaneous electrical efficiency and the reference electrical
efficiency is given below.
( )
ηEL = ηSTC [1 − βT Tc − Tref ] (1)
where Tc is the temperature of the cell.
The collector fluid is a 30/70% glycol–water mix due to the
fact that it provides better heat transfer parameters than plain
water, and it has a lower freezing point, which is important
in continental climates where winters are defined by negative
temperatures. The size of the panel is 1 m x 2 m, and the diameter
of the channel is 0.008 m, with a thickness of insulation of 5 cm.
Table 1 summarizes the main PVT collector properties.
The cross section of the panel is shown in Fig. 1, with all the
composing layers and the corresponding subscripts – glass (g), Air Fig. 2. Schematic heat transfer processes in the layer of a PVT panel.. (For
gap, PV layer (p), thermal absorber (a), insulation, fluid (f). The PV interpretation of the references to color in this figure legend, the reader is
referred to the web version of this article.)
layer is made of PV glass, ethylene vinyl acetate (EVA) film, Silicon
cell, Tedlar layer and adhesive.
The assumptions made for developing the model are:
to low temperatures. Convection is the heat transfer in which
• edges are well insulated; no heat is lost at the edges; energy moves within a fluid or gas through flow of matter from
• the layers are very thin thus the vertical heat conduction can higher to lower temperature regions. It can be natural, occurring
be assumed negligible; due to difference in densities, or forced by an external physical
• the optical properties of materials remain constant along the movement, such as a pump or the wind.
area; The heat transfer processes that take place inside a PVT panel
• no dust effects or partial shading are taken into account; are shown in Fig. 2; radiation with a yellow arrow, convection
• the flow rate of the fluid is constant, ṁ = 0.05 kg/s.
in green and conduction in red. The glass layer receives direct
• the initial fluid inlet temperature at time t = 0 is Ti0 =
and indirect solar radiation from the sun, a fraction of which is
15 ◦ C. At every consecutive time step, the new Tfi is the
reflected, another part absorbed and the rest transmitted to the
average value between Ti0 (assumed to be the cold-water
inlet main temperature) and Tfo from the previous time step PV layer. It also receives reflected radiation and radiative losses
(the temperature of the fluid at the outlet of the panel). from the PV layer. It loses some heat to the environment through
forced convection caused by the wind. The PV layer receives a
The layers analyzed in this model are: glass, PV, thermal absorber fraction of the solar radiation, of which some is reflected back to
and fluid. Some models (Chow et al., 2005; Pierrick et al., 2015) the glass layer and some is absorbed. The PV layer also loses heat
also analyze the behavior of the air gap layer and the insulation through radiation to the glass. In addition, natural (free) convec-
layer, but those temperatures were found not to be relevant for
tive heat transfer between the glass and the PV takes place inside
the purpose of this study, and thus the model is simplified to only
the air gap. Out of the incident radiation, a fraction is converted
compute the four layers of interest.
to electricity through the photovoltaic effect. The absorber layer
The thermal energy circulates through the layers of the PVT
exchanges heat with the PV layer, with the insulation and with
panel and to/from the exterior through three types of heat trans-
fer processes: radiation, convection and conduction. Radiation is the fluid through conduction, and the insulation loses heat to
the process of heat transfer without any physical contact and the environment through forced convection from the wind. The
it does not heat the transfer medium. Any physical body of a heat transfer in the fluid occurs also through forced convection
temperature greater than 0 K emits a level of radiation through generated by the pumping system.
electromagnetic waves. When radiation reaches a body, it can be
either be absorbed, reflected or transmitted. Conduction is the
3. Implementation of numerical model
movement of heat through a solid material or at the interface
of two materials/layers that are in contact due to molecular
collisions. Conduction that occurs on the solar panel is caused by The numerical model was developed in a previous paper
the thermal gradients and its direction is from high temperature (Barbu et al., 2019b) and is presented in the next section.
3
M. Barbu, M. Siroux and G. Darie Energy Reports xxx (xxxx) xxx

Table 1
PVT properties (Rejeb et al., 2015; Aste et al., 2015; Pierrick et al., 2015).
Property Glass Air gap PV Thermal Fluid Insulation Unit
absorber
Emissivity (ε ) 0.9 – 0.96 – – – –
Absorbance (α ) 0.1 – 0.9 – – – –
Transmittance (τ ) 0.93 – – – – – –
Thickness (H) 0.004 0.02 0.006 0.001 – 0.04 m
Density (ρ ) 2200 – 2330 2699 1050 16 kg/m3
a
Specific heat (c) 670 – 900 800 1120 J/(kg K)
a a
Thermal 1.1 140 237 0.035 W/(m K)
conductivity (k)
a
Depends on the temperature.

3.1. Heat balance K−4 ), the emissivity of glass (εg ) and the equivalent radiative
temperature of the sky (Tsky ). There are multiple equations avail-
The numerical model is based on the heat transfer processes able for defining the sky temperature, most commonly it can be
described in the previous section. A heat balance equation is expressed as a linear function of the ambient temperature (Te )
developed for each of the significant layers of the system. This and the sky cloud coverage in octaves (N) (Aste et al., 2015). If
is based on the equation for the variation of internal energy in a clear sky conditions are assumed or there is no data available on
physical body: the cloud coverage, Eq. (6) can be further simplified to Eq. (7),
dT dU with less that 1% effect on the thermal and electrical output of
Mc = (2) the system (Guarracino et al., 2016; Touafek et al., 2014).
dt dt
where M [kg] is the mass, c [J/(kg K)] is the specific heat capacity, hg −e,RD = εg σ (Tg 2 + Tsky 2 )(Tg + Tsky ) (5)
T [K] the temperature, dt [s] is the chosen time step and dU [J] is Tsky = 0.0552Te 1.5
+ 2.652N (6)
the change in internal energy of the layer. 1.5
Tsky = 0.0552Te (7)
3.2. Glass layer (g) The thermal coefficient of the gap (hgap ) takes into account the
conduction in the top layers of the PV sandwich (PV glass and
For the thermal balance of the glass layer ‘g’ we considered EVA) and the convective heat transfer in the air gap.
the convective and radiative losses from the glass to the envi-
ronment (Qg−e,CV and Qg−e,RD ), the convective and radiative heat 1 1 HPVg HEVA
= + + (8)
transfer from the glass to the PV layer (Qg−p,CV and Qg−p,RD ) and hgap hg −p,CV kPVg kEVA
the heat absorbed by the glass (Qg ). Nuair kair
hg −p,CV = (9)
dTg Hgap
Mg cg = Qg −e,CV + Qg −e,RD + Qg −p,CV + Qg −p,RD + Qg
dt ( ) ( ) where H is the height of the layer (measured at the middle of the
= hg −e,CV Ag Te − Tg + hg −e,RD Ag Tsky − Tg layer) [m], k is the thermal conductivity [W/(m K)] and Nuair is
( ) ( the Nusselt number for air [–].
+ hgap Ag Tp − Tg + hg −p,RD Ag Tp
) The Nusselt number correlation for inclined plates, tilt angles
−Tg + Ag αg Girr (3) from 0 to 60, is given by Hollands et al. (1976). The radiative
coefficient between the glass and the PV (hg −p,RD ) is expressed
where hg −e,CV is the coefficient of forced convection between
below:
the glass and environment [W/(m2 K )], Ag is the area of glass
[m2 ], Te is the temperature of the environment [K], Tg is the 1
hg −p,RD = 1 1
σ (Tg2 + Tp2 )(Tg + Tp ) (10)
temperature of the glass layer [K], hg −e,RD is the coefficient of ε
+ ε
− 1
g p
radiative heat loss between the glass and environment [W/(m2
K)], Tsky is the equivalent radiative temperature of the sky [K], where εp is the emissivity of the PV layer [–] and εg is the
hgap is the thermal coefficient of the gap [W/(m2 K)], Tp is the emissivity of the glass [–].
temperature pf the PV layer [K], hg −p,RD is the coefficient of
3.3. PV layer (p)
radiative heat loss between the glass and the PV layer [W/(m2
K)], αg is the absorbance of the glass [–] and Girr is the incident
The PV layer ‘p’ includes glass, two layers of EVA, cell and Ted-
irradiation on the absorber [W/m2 ].
lar. The thermal balance of the PV layer considers the convective
Multiple reviews were carried out on the coefficient of forced
and radiative heat transfer to the PV layer (Qp−g,CV and Qp−g,RD ),
convection due to wind (hg −e,CV ) (Rejeb et al., 2015; Aste et al.,
the conductive heat transfer to the absorber layer (Qp−a,CD ), the
2014), which can be expressed through various empirical correla-
heat absorbed by the PV layer (Qp ) and electricity production (E).
tions. The correlation proposed by McAdams (1954) was chosen
for this study. It applies to winds speeds (vw ) from 0 to 10 m/s, dTp
Mp cp = Qp−g ,CV + Qp−g ,RD + Qp−a,CD + Qp − E
gives good agreement between experimental and simulation data, dt (
and has been widely used in the literature for numerical PVT
) ( )
= hgap Ap Tg − Tp + hp−g ,RD Ap Tg − Tp
models (Rejeb et al., 2015; Lämmle, 2018). ( )
+ hp−a,CD Ap Ta − Tp + Ap (ατ )p Girr (1
m
⎧
⎨ 5.7 + 3.8vw
⎪ for vw < 5 − ηEL(T ) ) (11)
hg −e,CV = s (4)
m 2
⎩6.47 + vw0.78
⎪ for vw ≥ 5 where Ap is the area of the PV layer [m ], hp−a,CD is the conduction
s coefficient from the PV layer to the absorber [W/(m2 K)], Ta is the
The coefficient of radiative heat loss (hg −e,RD ) is calculated temperature of the thermal absorber layer [K], and ηEL(T ) is the
by using Stefan–Boltzmann constant (σ = 5.67×10−8 W m−2 time dependent electrical efficiency of the cell [–].
4
M. Barbu, M. Siroux and G. Darie Energy Reports xxx (xxxx) xxx

The heat transfer coefficient in the air gap and the radia- 3.6. System performance
tion coefficient are calculated in Eqs. (8)–(10). The conduction
coefficient through the layers of EVA, Tedlar and adhesive is: The energy performance of the system can be computed in
terms of thermal, electrical and overall efficiencies (ηTH , ηEL ,
1 HEVA HTed Hadh
= + + (12) ηoverall ). The electrical efficiency is calculated as the ratio between
hp−a,CD kEVA kTed kadh the power generated by the system (PEL ) and the amount of solar
radiation incident on the surface of the collector (AxG irr ):
The conversion efficiency of the cell decreases with the in-
crease in the operating temperature Tp and depends on the tem- PEL
ηEL = (18)
perature coefficient βPV . AGirr

ηEL(T ) = ηSTC [1 − βPV (Tp − Tref )] (13) The standard equation for thermal efficiency is represented by
the ratio between the amount of thermal energy generated by the
where ηSTC is the reference electrical efficiency measured at the system (QTH ) and the solar radiation incident on the surface of the
standard operating temperature Tref of 25 ◦ C [–]. collector (AxG irr ):
QTH ṁcf (Tfo − Tfi )
ηTH = = (19)
3.4. Thermal absorber (a) AGirr AGirr
where ṁ is the mass flow rate of the fluid and cf is the specific
The thermal balance of the absorber layer ‘a’ considers the heat capacity.
conductive heat transfer to the PV layer (Qa−p,CD ), the heat trans- Another definition for the thermal efficiency was proposed
fer to the fluid (Qa−f ) and the heat loss to the exterior through by Bombarda et al. (2016), where distinction is made between
the insulation (Qa−e,CD ). the electrical and thermal power of the system, and thus the
electrical power production is subtracted from the total incident
dTa irradiation. This definition allows for a more accurate comparison
Ma ca = Qa−p,CD + Qa−f + Qa−e,CD
dt between the thermal performance of a PVT and stand-alone solar
thermal system.
( ) ( )
= ha−p,CD Aa Tp − Ta + ha−f Ach Tf − Ta + ha−e,CD Aa (Te − Ta )
(14) QTH ṁcf (Tfo − Tfi ) ηTH
ηTH ∗ = = = (20)
AGirr − PEL AGirr − PEL 1 − ηEL
where ha−p,CD is the heat transfer coefficient from the absorber
to the PV layer [W/(m2 K)], Aa is the area of the thermal absorber From the point of view of the first law of thermodynamics, the
[m2 ], Ach is the area of the channel [m2 ], ha−f is the heat transfer global efficiency of the system (ηGL ) can be calculated as the sum
coefficient of the fluid [W/(m2 K)], Tf is the temperature of the of the thermal and electrical efficiencies. This is also known as
fluid [K], ha−e,CD is the conductive heat loss coefficient to the the first law efficiency:
environment [W/(m2 K)]. ηGL = ηEL + ηTH (21)
The heat transfer coefficient to the PV layer (ha−p,CD ) is the
same as in Eq (12). The heat transfer coefficient of the fluid Another approach is to consider the difference in grade be-
depends on the type of flow (laminar or turbulent) (Incropera, tween 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
2011).
⎧ a kWel cannot be directly compared to a kWth , some authors (Aste
kf et al., 2015; Huang et al., 2001) propose the primary energy
⎪4.36 , for Re < 2300 [−]
⎪
⎪
⎪
⎪ D H
saving efficiency (ηPES ) as a more accurate way of assessing the
system performance:
⎪
kf
⎨
ha−f = 0.023 Re0.8 Pr 0.4 for Re ≥ 2300 [−] (15) ηEL
⎪
⎪ D H ηPES = ηTH + (22)
⎪
⎪
⎪ kf kg ηTpower
⎩ 2 ,
⎪ for ṁ = 0
DH s where ηTpower is the average efficiency of producing electrical
where kf is the thermal conductivity of the fluid [W/(m K)], DH is power, and it depends on the country of reference. On average, it
can be assumed to be 0.4 (Aste et al., 2015), and this value was
the hydraulic diameter [m], Re is the Reynolds number [–], Pr is
used for the calculations. For Romania, the average efficiency of
the Prandtl number [–] and ṁ is the mass flow rate [kg/s].
producing electrical power in 2016 was 0.532, while in France it
was 0.601 (European Environment Agency, 2018).
3.5. Fluid (f)
3.7. MatLAB implementation
The energy balance equation for the fluid ‘f’ considers the
By rearranging the energy balance equations (3), (11), (14)
thermal energy coming from the absorber (Qf−a ) and the heat and (16), a system of coupled equations is obtained. The system
accumulated by the fluid (Qf ). of equations is then solved in MatLAB with the ordinary ‘fsolve’
dTf π DL ( ) ( ) function. The coefficients C and D are defined by rearranging the
Mf cf = Qf −a + Qf = ha−f Ta − Tf + ṁcf Tfi − Tfo (16) free terms of the energy balance equations above.
dt 2 ⎧
The heat transfer coefficient of the fluid is calculated in Eq. ⎪ C1 Tg + C2 Tp − D1 = 0
C3 Tg + C4 Tp + C5 Ta − D2 = 0
⎨
(15), D is the diameter of the pipe and L the length [m], ṁ [kg/s] (23)
is the mass flow rate of the fluid in the channels and the inlet and ⎩ C6 Tp + C7 Ta + C8 Tf − D3 = 0
⎪
C9 Ta + C10 Tf − D1 = 0
outlet temperatures are represented by Tfi and Tfo respectively
[K]. where
Mg cg
Tf = 0.5Tfi + 0.5Tfo (17) C1 = + hg −e,CV Ag + hg −e,RD Ag + hgap Ag + hg −p,RD Ag (24)
dt
5
M. Barbu, M. Siroux and G. Darie Energy Reports xxx (xxxx) xxx

Table 2
Parameters of validation model (Bhattarai et al., 2012).
Collector size 2 m2 Tube diameter 9 mm
Cell efficiency 17.3% Mass flow rate 0.002 kg/m2 s
Packing factor 80% Solar radiation 1000 W/m2
Temperature coefficient −0.405%/K Wind speed 1 m/s
No. of tubes 10 Ambient temperature 30 ◦ C
Tube spacing 10 cm Tilt angle 30◦

C2 = −hgap Ag − hg −p,RD Ag (25)

Mg cg
D1 = Tg + hg −e,CV Ag Te + hg −e,RD Ag Tsky + Ag αg Girr (26)
dt
C3 = −hgap Ap − hp−g ,RD Ap (27)
Mp cp
C4 = + hgap Ap + hp−g ,RD Ap (28)
dt
C5 = −hp−a,CD Ap (29)
Mp cp
D2 = Tp + Ap (ατ )p Girr (1 − ηEL(T ) ) (30)
dt
Fig. 3. Comparison of simulation and experimental results of thermal and
C6 = −ha−p,CD Aa (31)
electrical efficiencies as a function of the reduced temperature.
Ma ca
C7 = + ha−p,CD Aa + ha−f Ach + ha−e,CD Aa (32)
dt
C8 = −ha−f Ach (33) 4. Results and discussion
Ma ca
D3 = Ta + ha−e,CD Aa Te (34) 4.1. Case studies
dt
π DL
C9 = −ha−f (35) The developed numerical model was applied for a PVT panel
2
in two different climate conditions: Bucharest and Strasbourg,
Mf cf π DL
C10 = + ha−f + 2ṁcf (36) which enables a comparison of the performance of the PVT sys-
dt 2 tem in both regions. The climate of Romania is temperate con-
Mf cf
D4 = ˙ f Tfi
Tf + 2mc (37) tinental, with hot and sunny summers with occasional showers
dt and thunderstorms and cold, cloudy winters and frequent snow
The fsolve function solves the equations for each individual and fog. In the southern region, where Bucharest is located, the
timestep and gives a vector solution for [Tp Tg Ta Tf ]. The results winters are milder and the summers hotter. France’s climate is
are plotted on a graph and shown in Section 4. For the parametric temperate, with four specific regions. Alsace, where Strasbourg
study, a ‘for’ function is introduced to vary the variables in a is located, is characterized by a continental climate which har-
chosen range, while keeping the rest constant. bors cold winters, hot summers, and strong winds. The average
summer months temperature in Bucharest is 29.1 ◦ C, while in
3.8. Model validation Strasburg it is 24.2 ◦ C. For the winter months, Bucharest has an
average of 4 ◦ C, and Strasbourg 5.2 ◦ C. The monthly summer
The developed model was validated by comparing the per- sun hours are 278.8 and 222.0 for Bucharest and Strasbourg
formance results to experimental data found in the literature. respectively, while the winter sun hours are 72.6 and 54.5 (WMO,
In order to achieve this, the PVT model described by McAdams 2019).
(1954) is recreated using this model, with the same configuration The variation of the weather during the day was used as an
and operating conditions used in the experiment. The variation of input into the numerical model in order to assess in real time the
the thermal and electrical efficiency is plotted against the reduced temperatures inside the panel. The input is: solar irradiation (G
temperature for both the simulated and experimental model. [W/m2 ]), ambient temperature (Tamb [◦ C]) and wind speed (W
The characteristics of the experimental model are summarized in [m/s]). The yearly profile with hourly resolution for each param-
Table 2. eter was obtained from Meteonorm Database for both Bucharest,
The system performance can be assessed by computing the Romania and Strasbourg, France. The meteorological conditions
electrical and thermal efficiencies, according to Eqs. (18)–(19). of a typical day are presented in the figures below. Fig. 4a and b
These are plotted against the reduced temperature, defined in show the daily weather conditions for Strasbourg and Bucharest
Eq. (38) below, to account for the fact that the thermal efficiency respectively from 8:00 to 20:00, with an hourly time resolution,
is, in fact, dependent on ambient temperature, inlet temperature while Fig. 5a and b show the winter weather conditions. Given
and global irradiation. the fact that the simulations were run for one day, it is important
to note the sensitivity of the chosen data on the results of the
Tfi − Tamb
T∗ = (38) numerical simulations. This means that the specific day which
Girr was chosen from the Meteonorm yearly database is of crucial
Fig. 3 shows a good resemblance between the reference ex- importance and given the significant day to day variation of
perimental data and the results obtained by the numerical model the weather, a random day can be misrepresenting the seasonal
developed in this paper. It shows that under different opera- climate. In this case, the same data was chosen for both cities,
tion conditions and constructive parameters of the collectors, making sure no extreme condition occur and the two weather
the result of the experimental study from the literature can be conditions are comparable at the given time. The representative
recreated with the present numerical model. winter day is 3rd of January and for the summer is 12th of July.
6
M. Barbu, M. Siroux and G. Darie Energy Reports xxx (xxxx) xxx

Fig. 4. Temperature, wind speed and solar radiation for a summer day in (a) Strasbourg (b) Bucharest.

Fig. 5. Temperature, wind speed and solar radiation for a winter day in (a) Strasbourg (b) Bucharest.

Daily simulations were carried out for a typical summer and Table 3
Range of variation of parameters for OFAT parametric analysis.
winter day in Bucharest, Romania and Strasbourg, France, from
Wind speed 0–10 m/s
8:00 to 20:00, with the weather data including ambient tem- Packing factor 75%–95%
perature, wind speed and irradiation, discussed in the previous Thickness of insulation 1–10 cm
Width (thickness) of tube 5–10 mm
section. Given these conditions, the time dependent temperature
variation of the layers in the panel is shown below. Fig. 6a and
b show the temperature evolution of the layers, as well as the
ambient temperature, during a summer day in Strasbourg and 4.2. Parametric analysis
Bucharest respectively. Fig. 7a and b show the temperature evo-
Next, a OFAT analysis was carried out for some of the main
lution of the layers, as well as the ambient temperature, during a
constructive parameters of the panel, in order to evaluate their
winter day in Strasbourg and Bucharest respectively. As expected, impact on the system efficiency. This is achieved by varying the
the layers with the higher thermal conductivity reach the high- parameters according to Table 3, while keeping the rest of them
est temperatures. In all cases, the PV layer is the most heated, to the default values reported in Table 1. The reference electrical
reaching 52 ◦ C for a summer day in Bucharest. The next layer efficiency of the panel is 15%, while the thermal efficiency is
is the thermal absorber, followed by the fluid. The glass layer dependent on ambient temperature, inlet temperature and global
irradiation.
is always the coolest. The temperature profile of all the layers
The effect of the two main meteorological conditions, ambient
follows closely the profile of the ambient temperature and global temperature and solar radiation, on the PVT behavior are easily
irradiation. The effect of the wind speed is not easily noticeable observable in the case studies above. However, the influence of
in the simulations. It can also be observed that the working the third component, wind speed, is not so obvious, therefore this
fluid reaches slightly higher temperatures in Bucharest during the relationship is further analyzed by varying its value from 0 to 10
summer and in Strasbourg during the winter, due to their specific m/s and assessing the system performance. Increasing the wind
weather conditions. As a result, assuming that the system would speed results in a higher forced convection coefficient, according
to Eq. (4). The results in Fig. 8 indicate a 2% decrease in the
have a water storage tank, it can be expected that the system
thermal efficiency, which can be explained by an increased forced
would have a better thermal efficiency in Bucharest compared to convection heat loss from the glass to the ambient. On the other
Strasbourg during the summer, and higher in Strasbourg during hand, a 0.1% increase in the electrical efficiency can be observed.
the winter. A higher forced convection leads to a better cooling of the PV
7
M. Barbu, M. Siroux and G. Darie Energy Reports xxx (xxxx) xxx

Fig. 6. Time dependent temperature evolution of PVT layers for a summer day in (a) Strasbourg (b) Bucharest.

Fig. 7. Time dependent temperature evolution of PVT layers for a winter day in (a) Strasbourg (b) Bucharest.

cell and a lower operating temperature, which leads to a slightly packing factors of 75% up to 95% for the highest performing
improved electrical output. However, the effect of high wind ones. It can be observed that the increased packing factor has
speeds on the decrease in thermal performance is much more a significant positive effect on the electrical performance, with
significant than the beneficial increase in electrical performance. more than 2% improvement, due to a larger surface area of the
Overall, the global efficiency decreases with higher wind speeds. panel being covered with PV cells, and implicitly, a larger amount
It is also noticeable that the slope of electrical efficiency decreases of incident solar radiation being converted to electricity. The
gradually and tends towards flattening, which indicates that the 20% decrease in thermal efficiency with higher packing factors
operating temperature is close to reaching the STC value of 25 ◦ C, is explained by the fact that a higher fraction of the solar energy
after which cooling no longer has an effect on the electrical is converted to electricity, and thus a smaller fraction remains to
output. be converted into heat. Thus, the variation of the two efficiencies
Apart from the convective losses caused by the wind, heat appears equivalent; However, depending on the consumer profile,
losses also take place at the back of the panel through the layer it might be beneficial to choose a high packing factor for the PVT
of insulation, and their magnitude is directly influenced by the
design, as the electrical energy is more valuable both economi-
thickness of the layer. The effect of the insulation losses on the
cally and exergetically than the thermal energy. A high packing
system performance can be assessed by a parametric variation of
factor typically involves increased costs of the PV cells, thus a
the insulation thickness, from 0.01 to 0.1 m. Fig. 9 shows a steep
cost–benefit analysis should be carried out for optimum design.
effect on both efficiencies at first, followed by a flattened curve,
which indicates the fact that after a certain thickness, the increase Overall, the global efficiency increases with a higher packing
in insulation has an insignificant effect on the panel performance. factor.
In the case of the electrical performance, this occurs at 0.05 m Finally, the width of the absorber channels is varied from 5 to
of insulation, where the PV cells reach the STC temperature. For 10 mm for a constant flow rate of 0.05 kg/s. A wider diameter
the thermal performance, the curve is not completely flat even channel results in a larger cross section area and thus a slower
at 0.1%, but the slope decreases gradually. Again, as in the case flow rate of the fluid in the channels, but also a larger area
of wind speed, the effect on the thermal performance is much of heat transfer between the absorber and the fluid. A better
more significant than on the electrical one, with a variation of performance of both components is observed with the increase
3.5% compared to 0.1%. Overall, the global efficiency increases in channel diameter, shown in Fig. 11. The electrical efficiency
with thicker insulation. increases by 0.6% due to better cooling of the PV, while the
The effect of the packing factor of the PV cell on the system significant increase of 20% in the thermal efficiency is explained
efficiency is illustrated in Fig. 10. Typical commercial cells have by a better heat transfer between the absorber and the fluid,
8
M. Barbu, M. Siroux and G. Darie Energy Reports xxx (xxxx) xxx

Fig. 8. Variation of thermal and electrical performance as a function of wind speed.

Fig. 9. Variation of thermal and electrical performance as a function of the thickness of insulation.

Fig. 10. Variation of thermal and electrical performance as a function of the packing factor.

and thus an improved thermal energy output. However, there absorbers that limit the width to 8–10 mm depending on manu-

are technical constrains on the fabrication methods of roll bond facturing method, materials and distance between the channels.
9
M. Barbu, M. Siroux and G. Darie Energy Reports xxx (xxxx) xxx

Fig. 11. Variation of thermal and electrical performance as a function of tube width.

Overall, the global and PES efficiencies increase with a higher tube Table 4
thickness. Efficiency maximizing solutions.

The results of the sensitivity analysis mostly indicate that, as Electrical-driven Thermal-driven Compromise
solution solution solution
in any cogeneration system, there is a compromise between the
production of electrical and thermal power of the system. The Wind speed (m/s) 10 0 10
Insulation thickness (cm) 1 10 10
design can lean towards improving one of the performances of Packing factor (%) 95 75 95
the system. In most cases, the electrical performance will be more Tube width (mm) 30 30 30
important, due to a higher value of electrical energy compared to
thermal energy.
As a result of this analysis, a number of technical recom-
mendations can be formulated for the efficient operation of PVT • thermal-driven – maximizing the thermal efficiency
collectors. First of all, the roof installation of the panels is im- • compromise solution – maximizing the global efficiency
portant, not only in terms of South-facing positioning (for the
The results are summarized in Table 4. In the case of most param-
Northern hemisphere) and tilting it to a suitable angle according eters, the compromise solution coincides with the thermal-driven
to the latitude of the location, but also in terms of wind protec- solution, except for the packing factor.
tion. In some cases, space will be limited and there is no flexibility
with the mounting, which can lead to high winds and a decrease 5. Conclusion and perspectives
in global efficiency of the collector. Where there is flexibility
in the location, a rose wind is a useful tool for ensuring there This study proposed an explicit numerical model of a PVT
is wind protection for the system of collectors, thus improving collector based on the heat balance equations between each layer.
the overall efficiency. Wind protection is not only beneficial for The model has been validated by numerically recreating a panel
the efficiency, but also decreases the amount of dust collected that has been experimentally studied in the literature, and the
and reduces the maintenance costs. In terms of insulation, it was results show good accordance. Next, the model was applied in
observed that the benefits of increasing its width are significant. two case studies for the climates of Bucharest, Romania, and
However, this can add to the cost of production and can increase Strasbourg, France. The input consists of ambient temperature,
the bulkiness and weight of the equipment. Another technical wind speed and global irradiance, which were obtained for both
solution is investigating alternative materials for the thermal locations from the Meteonorm Database. Simulations were run
insulation, with lower coefficient of thermal loss, so that both the for representative summer and winter days, from 8:00 to 20:00.
thickness and thermal losses can be kept to a minimum. In terms The results show the temperature of each layer at any particular
of choosing a packing factor for the PV layer, it is important to time, and it can be observed a slightly higher temperature of the
keep in mind the end consumer and the applications. Some man- working fluid in Bucharest during the summer, and in Strasbourg
ufacturers have even started to produce two types of collectors, during the winter.
thermally driven and electrically driven, depending on the needs The OFAT parametric analyses carried out in this study showed
of the end user (ENDEF Solar Solutions, 2021). The width of the in most cases a compromise between the electrical and thermal
channels can be maximized but keeping in mind that there will be performance of the PVT system. In terms of wind speed and
an increase in the pumping power for maintaining the required insulation, it was observed that the thermal benefits of low wind
flow rate. and high insulation overcome the decrease in electrical efficiency.
For optimum design, the insulation thickness should be maxi-
4.3. Performance maximization mized and the wind speed minimized, which can be achieved by
strategic placement of the panels in areas protected from wind.
The PVT panel can be optimized in order to maximize the The packing factor was found to be optimum when maximized,
production of either thermal or electrical energy, depending on as the electrical benefits are more significant than the thermal
the scope and building applications: ones. The width of the channels in the heat exchanger should
also be maximized as far as technologically possible for best
• electrical-driven – maximizing the electrical efficiency performance.
10
M. Barbu, M. Siroux and G. Darie Energy Reports xxx (xxxx) xxx

Future work will be focused on setting up an experimental sky Sky

stand for a more direct validation of the model and of the analysis, Ted Tedlar
and the dynamic coupling with residential consumers. Also, an TH Thermal
a Thermal absorber
algorithm for optimizing the design can be developed, taking into
account the exergy of the system, the global performance, and the
Declaration of competing interest
cost benefits.
The authors declare that they have no known competing finan-
Nomenclature cial interests or personal relationships that could have appeared
to influence the work reported in this paper.
α Absorbance [–]
α Absorptivity [–] References
A Area [m2 ] Agrawal, B., Tiwari, G., 2010. Life cycle cost assessment of building integrated
ρ Density [kg/m3 ] photovoltaic thermal (BIPVT) systems. Energ. Build. 42, 1472–1481.
D Diameter [m] Ahmadi, M.H., Baghban, A., Sadeghzadeh, M., Zamen, M., Mosavi, A., Shamshir-
DHW Domestic hot water band, S., Kumar, K., Mohammadi-Khanaposhtani, M., 2020. Evaluation of
electrical efficiency of photovoltaic thermal solar collector. Eng. App. Comp.
η Efficiency [%]
Fl. Mech. 14 (1), 545–565.
E Electrical energy [J] Ahmadi, M.H., Ghazvini, M., Sadeghzadeh, M., et al., 2018. Solar power tech-
ṁ Flow rate [l/h] nology for electricity generation: A critical review. Energy Sci. Eng. 6,
Q Heat [J] 340–361.
Armstrong, S., Hurley, W.G., 2010. A thermal model for photovoltaic panels under
h Heat transfer coefficient [W/(m2 K)]
varying atmospheric conditions. Appl. Therm. Eng. 30 (1488), 11–12–1495.
HVAC Heat, Ventilation, Air Conditioning Arsalis, A., Alexandrou, A., 2015. Parametric study and cost analysis of a solar-
U Internal energy [J] heating-and-cooling system for detached single-family households in hot
L Length [m] climates. Sol. Ene 117, 59–73.
M Mass [kg] Aste, N., Del Pero, C., Leonforte, F., 2014. Water flat plate PV-thermal collectors:
A review. Sol. Energy 102, 98–115.
Nu Nusselt number [–] Aste, N., Leonforte, F., Del Pero, C., 2015. Design, modeling and performance
OFAT One factor at a time monitoring of a photovoltaic-thermal (PVT) water collector. Sol. Energy 112,
PVT Photovoltaic thermal 85–99.
PV Photovoltaic Barbu, M., Darie, G., Siroux, M., 2019a. Analysis of a residential photovoltaic-
thermal (PVT) system in two similar climate conditions. Energies 12, 1–18.
P Power [W]
http://dx.doi.org/10.3390/en1219359.
Pr Prandtl number [–] Barbu, M., Mey-Cloutier, S., Siroux, M., Darie, G., 2019b. Dynamic modelling and
PES Primary energy saving sensibility analysis of a hybrid photovoltaic-thermal (PVT) system. In: 7th
Eur. Conf. Ren. Energy Sys. 10, Madrid, Spain.
G Radiation [W/m2 ]
Barbu, M., Patrascu, R., Darie, G., Tutica, D., 2019c. A technical-economical
Re Reynolds number [–] analysis of the implementation of hybrid solar energy systems in small
N Sky cloud coverage in octaves energy prosumer applications. Quality - Access Success 20 (169), 134–138.
ST Solar thermal Bhattarai, S., Oh, J.H., Euh, S.H., Krishna Kafle, G., Hyun Kim, D., 2012. Simulation
c Specific heat [J/(kg K)] and model validation of sheet and tube type photovoltaic thermal solar
system and conventional solar collecting system in transient states. Sol.
v Speed [m/s] Energy Mater. Sol. Cells 103, 184–193.
STC Standard test conditions Bombarda, P., Di Marcoberardino, G., Lucchini, A., Leva, S., Manzolini, G.,
σ Stefan–Boltzmann constant [W m−2 K−4 ] Molinaroli, L., Pedranzini, F., Simonetti, R., 2016. Thermal and electric
T Temperature [◦ C] performances of roll-bond flat plate applied to conventional PV modules for
heat recovery. Appl. Therm. Eng. 105, 304–313.
βT Temperature coefficient [%/K] Brottier, L., 2019. Optimisation Biénergie D’un Panneau Solaire Multifonctionnel
t Time [s] : Du Capteur Aux Installations in Situ ((PhD) Université Paris-Saclay). Paris.
H Thickness [m] Chow, T., Chan, A., Fong, K., et al., 2005. Energy performance of a solar hybrid
k Thermal conductivity [W/(m K)] collector system in a multistory apartment building. Proc. IMechE, Part A: J.
Power Energy 219 (1), 1–11.
τ Transmittance [–]
Dubey, S., Sarvaiya, J.N., Seshadri, B., 2013. Temperature dependent photovoltaic
Subscripts (PV) efficiency and its effect on PV production in the world – a review.
adh Adhesive Energy Procedia 33, 311–321.
air Air ENDEF Solar Solutions, 2021. Hybrid solar collectors. [online] Available at: https:
c Cell //endef.com/en/hybrid-solar-panel/.
European Environment Agency, 2018. Efficiency of conventional thermal
CD Conduction electricity and heat production in europe. Available online: https:
CV Convection //www.eea.europa.eu/data-and-maps/indicators/efficiency-of-conventional-
EL Electrical thermal-electricity-generation-4/assessment-2.
e Environment Florschuetz, L.W., 1979. Extension of the Hottel-Whillier model to the analysis of
combined photovoltaic/thermal flat plate collectors. Sol. Energy 22, 361–366.
EVA EVA
Guarracino, I., Mellor, A., Ekins-Daukes, N.J., Markides, C.N., 2016. Dy-
f Fluid namic coupled thermal-and-electrical modelling of sheet-and-tube hybrid
gap Gap photovoltaic/thermal (PVT) collectors. Appl. Therm. Eng. 101, 778–795.
g Glass Guo, J., Zheng, L., 2017. Numerically study on a new hybrid photovoltaic thermal
i Inlet (PVT) collectors with natural circulation. Appl. Sol. Energy 53 (4), 316–321.
Han, J., Chow, T., Han, C., 2006. Effect of flow channel dimensions on the
irr Irradiation performance of a box-frame photovoltaic/thermal collector. Proc. IMechE,
o Outlet Part A: J. Power Energy 220 (7), 681–688.
p PV layer Herrando, M., Markides, C., 2016. Hybrid PV and solar-thermal systems for do-
RD Radiation mestic heat and power provision in the UK: Techno-economic considerations.
Appl. Energy 161, 512–532.
ref Reference
Hollands, K.G.T., Unny, T.E., Raithby, G.D., Konicek, L., 1976. Free convective heat
transfer across inclined air layers. J. Heat Transfer 98 (2), 189.

11
M. Barbu, M. Siroux and G. Darie Energy Reports xxx (xxxx) xxx

Huang, B., Lin, T., Hung, W., Sun, F., 2001. Performance evaluation of solar Prabhakant, Agrawal B., Tiwari, G., 2010. Return on capital and earned carbon
photovoltaic/thermal systems. Sol. Energy 70 (5), 443–448. credit by hybrid solar Photovoltaic—wind turbine generators. Appl. Sol.
Incropera, F., 2011. Fundamentals of Heat and Mass Transfer. John Wiley and Energy 46 (1), 33–45.
Sons. Ramos, A., Guarracino, I., Mellor, A., Alonso-Alvarez, D., Childs, P., Ekins-
Kern, E., Russell, M., 1978. Combined photovoltaic and thermal hybrid collector Daukes, N.J., Markides, C.N., 2017. Solar-Thermal and Hybrid Photovoltaic-
systems. In: IEEE Photovolt. Spec. Conf., Vol. 1. p. 1153. Thermal Systems for Renewable Heating. Grantham Institute Briefing paper
Khalilpour, K., Vassallo, A., 2016. Technoeconomic parametric analysis of No 22, Imperial College London..
PV-battery systems. Ren. En. 97 75, 7–768. Rejeb, O., Dhaou, H., Jemni, A., 2015. Parameters effect analysis of a photovoltaic
Khatibi, A., Razi Astaraei, F., Ahmadi, M.H., 2019. Generation and combination thermal collector: Case study for climatic conditions of monastir. Tunisia.
of the solar cells: A current model review. Energy Sci. Eng. 00, 1–18. Energy Convers. Manag. 89, 409–419.
Lämmle, M., 2018. Thermal Management of PVT Collectors - Development and Touafek, K., Khelifa, A., Adouane, M., 2014. Theoretical and experimental study
Modelling of Highly Efficient Glazed, Flat Plate PVT Collectors with Low-E of sheet and tubes hybrid PVT collector. Energy Convers. Manag. 80, 71–77.
Coatings and Overheating Protection ((PhD) Fraunhofer Institute ISE). Tzinnis, S., Bellos, E., Tzivanidis, C., 2016. Parametric analysis of a solar cool-
McAdams, W.H., 1954. Heat Transmission, third ed. McGraw-Hill, New York. ing system designed for athens climate. In: EinB2016–5th International
Nhut, L., Raza, W., Park, Y., 2020. A parametric study of a solar-assisted house Conference ENERGY in BUILDINGS 2016.
heating system with a seasonal underground thermal energy storage tank. WMO, 2019. WMO - world meteorological organization. In: Country Profile
Sustainability 12 (8686). Database.
Papoutsis, E., Koronaki, I., Papaefthimiou, V., 2017. Numerical simulation Zamen, M., Baghban, A., Pourkiaei, S.M., Ahmadi, M.H., 2019. Optimization meth-
and parametric study of different types of solar cooling systems under ods using artificial intelligence algorithms to estimate thermal efficiency of
mediterranean climatic conditions. Energ. Build. 138, 601–611. PV/T system. Energy Sci. Eng. 00, 1–14.
Pierrick, H., Christophe, M., Leon, G. others, 2015. Dynamic numerical model of
a high efficiency PV-T collector integrated into a domestic hot water system.
Sol Energy 111, 68–81.

12