Energy Conversion and Management 195 (2019) 1382–1401

Contents lists available at ScienceDirect

Energy Conversion and Management

journal homepage: [Link]/locate/enconman

Performance evaluation of a new design of concentrator photovoltaic and T

solar thermoelectric generator hybrid system
Ahmed Abdoa,d, Shinichi Ookawaraa,b, Mahmoud Ahmeda,c,
⁎

a
Department of Energy Resources Engineering, Egypt-Japan University of Science and Technology (E-JUST), Alexandria 21934, Egypt
b
Department of Chemical Science and Engineering, Tokyo Institute of Technology, Tokyo 152-8552, Japan
c
Department of Mechanical Engineering, Assiut University, Assiut 71516, Egypt
d
Department of Mechanical Engineering, Tanta University, Egypt

ARTICLE INFO ABSTRACT

Keywords: A new configuration of a photovoltaic module and a solar thermoelectric generator in a hybrid system integrated
Solar thermoelectric generator with a microchannel heat sink has been developed. The photovoltaic module and the solar thermoelectric
Concentrator photovoltaic cooling generator sandwich the microchannel heat sink, as they are exposed to concentrated solar radiation. This newly
Microchannel heat sink developed hybrid system is compared to that with a conventional configuration in which an ordinary thermo-
Solar energy
electric generator is attached directly to the rear surface of the concentrator photovoltaic module. To evaluate
the performance of the new system in comparison to the conventional one, an all-inclusive three-dimensional
thermo-fluid-thermoelectric model has been developed. This model is numerically simulated and is validated
with both experimental and numerical results. The findings of this study show that the new system generates
more output power and can be operated at higher concentration ratios with even lower average solar cell
temperatures than the conventional system. In the new configuration, the average solar cell temperature is
approximately 77 °C and the total electrical output power is approximately 3.2 kW/m2 at a solar concentration
ratio of 20 suns. However, the maximum possible working concentration ratio for the conventional design is 10
suns, where the solar cell temperature approaches the highest tolerable temperature of approximately 90 °C.
Moreover, the total electrical output power is approximately 1.2 kW/m2. Furthermore, it is found that at
CR_PV = 20, the rate of the produced thermal energy is approximately 12 kW/m2, for the conventional design
while the new design achieves 15, 22, and 30 kW/m2 for CR_STEG of 1, 10, and 20, respectively. These findings
have greatly aided in identifying a new design that achieves the highest performance under concentrated solar
irradiance.

1. Introduction domestic use in heating water and air, water desalination as well as
numerous other well-documented applications [3]. Direct methods in-
Currently, as global warming continues to increase the temperature clude photovoltaic (PV) systems and solar thermoelectric generators
of the atmosphere, it is considered the greatest geopolitical challenge (STEG), which are supplemented by additional thermal benefits by way
facing the world. Carbon emissions, the primary contributor to global of absorbing extra heat for a variety of domestic purposes [4,5]. Direct
warming, are generated when fossil fuels are burned to generate elec- conversion solar technologies can be classified into simple or hybrid
tricity [1]. For this reason, the majority of research institutions have systems and concentrated or non-concentrated systems. Concentrated
shifted their focus to harnessing renewable energy sources such as solar, systems increase the amount of electricity generated but decrease the
wind, and biomass. Solar energy, the main source of most renewable overall conversion efficiency. Hybrid systems aim to achieve a synergy
energy, has become an interesting practical field of research [2]. Solar by increasing the efficiency of concentrated systems and maximizing
energy can be either directly or indirectly converted into electricity. benefits of each system, exploiting the fact that both thermal and
Indirect methods include solar power stations, for instance, the para- electrical energy can be obtained.
bolic trough collector system or the dish collector system, both of which PV systems absorb and convert most of the visible incident radiation
utilize a working fluid to power turbines to generate electricity. to electricity. Subsequently, the remaining part is absorbed and con-
Moreover, solar energy can be converted into thermal energy viable for verted into thermal energy, which increases the solar cell temperature.

⁎
Corresponding author at: Department of Energy Resources Engineering, Egypt-Japan University of Science and Technology (E-JUST), Alexandria 21934, Egypt.
E-mail addresses: aminism@[Link], [Link]@[Link] (M. Ahmed).

[Link]
Received 5 February 2019; Received in revised form 28 April 2019; Accepted 29 April 2019
Available online 14 June 2019
0196-8904/ © 2019 Elsevier Ltd. All rights reserved.
A. Abdo, et al. Energy Conversion and Management 195 (2019) 1382–1401

Nomenclature ab Absorber coated aluminum substrate

Back
A Area (mm2) co, i Coolant fluid at inlet
C Coolant specific heat (J/kg.K) co, e Coolant fluid at exit
D Hydraulic diameter (mm) cop Copper
e Error (dimensionless) con Convection losses
G Net concentration solar sell (W/m2) c, c Cold ceramic
h Hours, micro-channel height (mm), heat transfer coeffi- c, h Hot ceramic
cient by convection (W/m2.k), e, b Lower layer of EVA below silicon wafer
H Height (mm) e, t Upper layer of EVA above silicon wafer
L Length (mm) Fr Fresnel lens
.
m Mass flow rate (kg/s) gc Glass cover
E Power (W) l Layer
P Pressure (Pa) leg Thermoelectric leg
Q Rate of heat transfer (W) mch Microchannel
Re Reynolds number mod Module
Kn Knudsen number mfp Molecular mean free path
T Temperature (°C) Net Net
t Thickness (mm) rd Radiation losses
V Velocity (m/s) sk Sky
W Width (mm) sw Silicon wafer
X Length in x direction (mm) th Thermal
Y Length in y direction (mm) te Tedlar layer
Z Length in z direction (mm) tot Total
ZT Dimensionless Figure of merit wl Wall
x Coordinate in x direction
Greek symbols y Coordinate in y direction
z Coordinate in z direction
Absorptivity, Seebeck coefficient
Transitivity Abbreviations
Emissivity
Fluid mass density (kg/m3), electrical resistivity PV Photovoltaic
µ Fluid absolute viscosity (Pa.s) TE Thermoelectric
Stephan-Boltzmann constant 5.67 × 10−8 (W.m−2·K−4) TEC Thermoelectric cooler
Solar cell temperature coefficient TEG Thermoelectric generator
Solar cell temperature coefficient STEG Solar thermoelectric generators
Thermal conductivity (W/m.k) PCM Phase change material
HS Heat sink
Subscript PVF Polyvinyl fluoride

am Ambient

In STEG, (i.e., TEG and absorber), solar irradiation is initially absorbed splitting technique, the shorter wavelengths were transformed into
by a surface with high absorptivity (absorber) such as copper or alu- electricity by the PV system, while the longer wavelengths were used to
minum with a selective coating and is then transferred through the TEG generate electricity by the TEG system. The conclusion is that opti-
material as sensible thermal energy. Subsequently, part of this energy is mizing the system configuration is the main factor for high overall ef-
converted directly into electricity through the Seebeck effect; the con- ficiency.
verted amount depends on the materials and the temperature differ- Temperature is the key parameter; it has different and opposite ef-
ences between the hot and cold surfaces of the TEG. However, as fects on the two systems and influences the conversion efficiencies of
electrical energy is considered more important than thermal energy, both the PV and TEGs. By increasing the cell temperature, the con-
considerable efforts have been made to design a system with a rela- version efficiency decreases [7]. It was documented by Skoplaki and
tively high electrical conversion efficiency. Kraemer et al. [4] observed Palyvos [8] that the solar cell efficiency reduces by approximately
that solar radiation contains both visible and infrared wavelengths. 0.5 % per one degree Celsius rise in temperature in the silicon layer
Visible wavelengths are suitable for a PV band gap, while infrared relative to the nominal working temperature. Therefore, several cooling
wavelengths are appropriate for the promotion of high temperature methods are employed to avoid high-temperature effects on the solar
differences in thermoelectric material; thus, increasing the conversion cell performance such as passive cooling reported by Emam and Ahmed
efficiency. Accordingly, Xing et al. [5] sought to split solar radiation [9–11], Zarma et al. [12], Rabie et al. [13], and active cooling as re-
into the visible range for PV system purposes and the infrared range for ported by Ahmed and Radwan [14], Radwan et al. [15,16].
thermoelectric materials. A model was developed for spectrum splitting In contrast, for STEGs, increasing the TEG surface temperature with
in a PV–thermoelectric generator (TEG) hybrid system under different respect to the cold side enhances the conversion efficiency as reported
values of solar irradiance (550–770 W/m2). It was documented that the by Li et al. [17]. Sundarraj et al. [18] assessed the cooling methods of
PV–TEG hybrid system performed better than the conventional PV STEGs and indicated that increasing the temperature difference be-
system. Liang et al. [6] integrated PV and thermoelectric technologies tween the hot and cold surface of the STEG significantly improves its
to benefit from a wide spectral range of solar energy. By utilizing the conversion efficiency. The thermal energy absorbed by water can be

1383
A. Abdo, et al. Energy Conversion and Management 195 (2019) 1382–1401

utilized for domestic purposes. Therefore, a hybrid PV–TEG system is

proposed, to improve the total efficiency. Yin et al. [19] attempted to
recover part of the thermal energy inside the PV material by using
thermoelectric materials as a heat absorber for the PV. Wu et al. [20]
presented a comparative study between two solar PV–TEG and PV–TEC
(thermoelectric cooler) hybrid systems, in which they calculated the
temperature of the PV cell, the conversion efficiency, and the output
power under different conditions. The results indicated that the
PV–TEG performs better than PV–TEC in terms of power and efficiency,
because while the TEC consumes part of the generated electricity, the
TEG recovers part of the heat and converting it into electricity. Thus, in
addition to producing electricity, the TEG absorbs heat from the PV to
decrease its temperature. Cui et al. presented an experimental [21] and
theoretical [22] study for a PV–TEG solar hybrid system with a phase
change material (PCM) heat sink. A series of structural factors were
investigated to determine the optimal parameters for the PV–PCM–TEG
system and the impact of these parameters on system performance. It
was deduced that the performance of the studied hybrid system was
higher than single PV cells and PV–TEG without PCM.
Hajji et al. [23] studied the performance of thermoelectric and PV
coupling. In their design, a concentrator was separately located be-
tween thermoelectric and PV systems. They used different values of
solar concentration up to 2000 W/m2 and variable cooling capacity.
They concluded that an indirect coupling was a favorable alternative to
maximize the output energy. Rodríguez et al. [24] examined a hybrid
system that included a silicon PV module, a bismuth telluride TEG, and
a concentrating Fresnel lens. The Fresnel lens was used for the TEG,
whereas the PV module absorbed non-concentrated light. Mohsenzadeh
et al. [25] experimentally studied a new configuration for concentrating
PV–TEG hybrid systems cooled with channels. Their results showed that
the total maximum electrical efficiency of the hybrid system reached
5.76%. Li et al. [17] developed hybrid systems consisting of double
TEGs and established that the conversion efficiency was enhanced by
increasing the concentration ratio (CR).
Based on a survey of the literature, the objective of the current work
is to evaluate the performance of a new design of concentrator PV and
STEG hybrid system. Accordingly, the originality of this study is char-
acterized by two main concepts. First, the development of a novel
concentrator PV and STEG hybrid system combined with a micro-
channel heat sink placed between both units. Second, both the PV
module and the STEG are exposed to concentrated solar radiation
ranging from 1 to 20 suns. To assess the performance of the new system,
a comprehensive three-dimensional thermo-fluid-thermoelectric model
is developed. Furthermore, the performance of the newly developed
system is compared with two existing systems, the first one being a
conventional system in which the thermoelectric generator is attached
directly to the rear surface of the concentrator PV module. A cold
ceramic plate, followed by a microchannel, is also attached to the
configuration. The second system only involves the concentrator PV
module with a microchannel heat sink. The average solar celltem-
perature, the total electrical power, and the rate of thermal energy is
estimated for all suggestedconfigurations and the most effective one is
recommended for applications.

2. Physical model

In the current study, two distinct configurations of a concentrator

hybrid PV–TEG system integrated with a microchannel heat sink are
proposed. The conventional configuration includes a linear Fresnel lens
concentrator, PV module, microchannel heat sink, and a TEG system.
The TEG is connected directly to the rear surface of the PV system so as
to recover part of the thermal energy extracted from the PV system. The
microchannel heat sink is integrated with the cold surface of the TEG to
absorb the remaining thermal energy from the entire system, while the
lower wall of the microchannel is exposed to ambient conditions, as (caption on next page)
shown in Fig. 1. The newly developed design comprises a microchannel

1384
A. Abdo, et al. Energy Conversion and Management 195 (2019) 1382–1401

Fig. 1. Schematic drawing of the conventional design: 1-a. Linear Fresnel lens
concentrated PV-TEG system with microchannel heat sink. 1-b. Computational
domains of PV combined with microchannel heat sink and TEG. 1-c. Assembly
layers.

heat sink between the PV and STEG modules wherein it is attached

directly to the back surface of the PV to promote efficient cooling. The
other surface of the microchannel is attached to the cold surface of the
STEG. Both the PV and the STEG modules are subjected to concentrated
irradiance via Fresnel lenses, as shown in Fig. 2. The STEG unit contains
a TEG and a solar absorber; hence, it referred to as a solar thermo-
electric generator (STEG), as documented in two previous studies
[26,27]. In the present work, the STEG is a combination of a TEG and a
2.0 mm aluminum sheet coated with a high solar absorptive material.
An aluminum absorber plate is selected owing to its high thermal
conductivity and low cost [26].
To simultaneously have different CRs for the PV and the STEG
modules, two Fresnel lenses are used, as presented in Fig. 2. Each lens
has the dimension LFr , WFr and its height can be varied using a screw
mechanism in order to adjust the CR for each one. Two mirrors are used
to redirect the concentrated solar irradiance toward the glass cover of
the PV module and the absorber plate of the STEG. The values of the CR
can be varied separately for the PV module and the STEG by using the
screw mechanism to adjust the height HFr of the Fresnel lens to obtain
the required CR for each system, as shown in Fig. 2-a. The CR is varied
up to 20 for both designs. The PV module absorbs visible light from the
received concentrated irradiance and converts it into electricity. Fur-
thermore, the STEG absorbs most of the incident concentrated solar
radiation (90%) and converts it first into thermal energy, and thereafter
into electricity, based on its conversion efficiency. Finally, the micro-
channel heat sink absorbs excess thermal energy from both the PV
module and the STEG.
The developed system has three main components PV, STEG, and
microchannel heat sink. The microchannel heat sink is the sole con-
nection between the PV and STEG. It is used to absorb the heat from
each PV and STEG device. The side surfaces of the whole system are
isolated from the adjacent vertical wall. The vertical wall acts as a
middle partition and does not penetrate the system, only built around
the perimeter. This vertical wall has two main functions; it separates
the two radiation zones, the first for the PV side and the second zone for
the STEG side, and enhances the concentrated radiation below each
Fresnel lens in each radiation zone from the reflective surface on the
vertical wall.
The examined PV solar cell is a polycrystalline type, as stated by
Sandia Laboratory [28], with the typical dimensions of a silicon wafer
of 31.25 mm by 31.25 mm. According to the manufacturing technology,
the PV cell has several layers. It has a glass layer with high transmis-
sivity (from 0.92 to 0.95), as documented by Hedayatizadeh et al. [29]
and Zhou et al. [30]. A silicon wafer with low absorptivity and re-
flectivity is used. The silicon can reflect up to 35% of the received ra-
diation. This silicon wafer is imbedded in the transparent ethylene vinyl
acetate (EVA) layer for preservation, as the EVA is an electrical in-
sulator and water-resistant. Its total thickness is approximately 1.0 mm;
this is divided into two parts: 0.5 mm above the silicon wafer and an-
other 0.5 mm below it. The final layer is composed of a Tedlar polymer,
a photo-stable layer made of polyvinyl fluoride (PVF) [31].
For the conventional system, the thermoelectric TEG module di-
mensions WTEG, LTEG are 31.5 mm and consist of 98 thermo-couples (i.e.
couples) joined in series electrically and in parallel thermally. Each
couple has P-N-junction legs. The leg dimensions are denoted as
Wleg , Lleg , and tleg representing the length, width, and height, respec-
tively. The layers of the TEG are defined as: ceramic plates, copper
strips, and legs. Legs are made from bismuth telluride that hosts the
positive and negative charge carriers. Strips are made from copper and
Fig. 2. Schematic drawing of the developed design: 2-a. Linear Fresnel lens
are used to transfer current generated through the TEG to and from the
concentrated PV-STEG system with microchannel heat sink. 2-b. Computational
domains of PV combined with microchannel heat sink and STEG. 2-c. Assembly
1385
A. Abdo, et al. Energy Conversion and Management 195 (2019) 1382–1401

external load. The hot and cold ceramic plates are made of aluminum 31.1 × 31.5 × 0.1 mm3 . Generally, all cases are studied under an ambient
oxide [32]. The function of the hot ceramic plate is to absorb thermal temperature of 30 °C and a wind speed of 1.0 m / s . Moreover, in the
energy from the Tedlar in the conventional design and from the solar current study, a value of solar concentration G = 1000 W/m2 is con-
absorber plate in the newly developed design. The cold ceramic plate is sidered as a CR of 1.
integral to releasing the extra rejected thermal energy and sending it to
the coolant through the microchannel heat sink, whereby it could be 3. Theoretical analysis
utilized for domestic purposes. The thermo-physical properties of
copper and bismuth telluride are temperature dependent, according to The newly developed hybrid system comprises distinct layers of the
the polynomial equation presented by Shen et al. [33]. The detailed PV and STEG modules along with the microchannel heat sink.
properties of the used materials are summarized in Table 2. Optical Therefore, a three-dimensional thermo-fluid-thermoelectric model is
properties of the PV layers and solar absorber plate are given in Table 3. developed on the basis of the following assumptions:
A microchannel heat sink made from aluminum with the dimen-
sions of Wmch, Lmch is used in both systems. The total thickness equals a. The PV cell material properties are temperature independent and
2twl + hmch = 0.5 mm . The water flows through a microchannel with isotropic [35].
depth equaling hmch = 0.1 mm and a length of Lmch = 31.5 mm . Heat is b. The TEG material properties are varying with temperature [33].
collected from the back surface of the TEG in the conventional design c. The fluid properties are temperature dependent [36].
and from both the Tedlar and the cold ceramic plates in the newly d. The flow in the microchannel is laminar, incompressible, and fully
developed design. The main objective of the microchannel is to absorb developed.
extra thermal energy from the entire system through the water flowing e. The thermal contact resistances within each layer of the system are
inside it, with the maximum Reynolds number not exceeding 100, so neglected [34].
as to avoid any air gaps or slippage through the microchannel.
Additionally, this reduces thetemperature of the cold ceramic plate in 3.1. PV layers
both configurations and increases the temperature difference across the
TEG in order to enhance its conversion efficiency. Further details of the Heat transfer through the PV layers can be mathematically re-
main parts of both studied configurations are shown in Figs. 1 and 2. presented by the heat conduction equation as follows:
In the new design of concentrated PV/STEG hybrid system, there
Tl
are two energy streams that meet in the microchannel heat sink. First, l p, l = . ql + Ql = .( l Tl ) + Ql = 0, and l = 1, 2 6
solar radiation is received by the PV module, where part of it is re- t
flected and another part is absorbed by the PV layers according to their (1)
transmissivity and absorptivity. The silicon layer transforms part of the where l , Cp, l, l, Ql are density, specific heat at constant pressure,
absorbed energy into electrical energy as stated in previous work [34]. thermal conductivity, and the heat absorbed per unit volume of layer l,
The remaining energy in the PV layers is transferred as a thermal en- respectively. As previously explained, the polycrystalline silicon cell is
ergy to the microchannel heat sink by conduction, depending on tem- composed of glass, EVA, silicon wafer, and (Tedlar Polyester Tedlar)
perature difference, thermal conductivity, and the thickness of the PV TPT layers. Because the glass and EVA layers have considerable trans-
layers and the microchannel wall. Second, solar radiation is received by missivity, the heat absorbed (Ql ) for these layers can be determined as
the high-absorptivity coating surface of the solar absorber. This coating follows:
can absorb the majority of incident solar radiation in the form of
gc (G ) Agc
thermal energy as previously reported in [26,27]. Subsequently, this Qgc =
thermal energy is transferred by conduction from the solar absorber to gc (2)
the hot ceramic plate, depending on the temperature difference,
e. gc . (G ). At . e
thermal conductivity, and the thickness of the plate. Heat is transferred Qt , e =
from the absorber plate to the hot ceramic plate. Thereafter, from the top, e (3)
hot ceramic plate and through the copper strips, the heat reaches the (1 sw ). sw . gc . e. (G ). Asw
surfaces of the P-N legs. Depending on the temperature difference be- Qsw =
sw (4)
tween the hot surface and the cold surface, and both the thermal and
electrical properties of the P-N legs, an electrical energy is generated in where Qgc , Qt , e , Qsw are heat generation per unit volume for the glass,
the P-N legs. Such electrical energy flows through the copper strips to
reach outside the TEG. The remaining part of the thermal energy is Table 1
absorbed by the microchannel heat sink. The dimensions of the main studied parts.
For the conventional design, the typical dimensions of the compu- Part name Width, mm Length, mm Thickness, mm Area, mm2
tational domain Wmod, Lmod , tmod are 31.5 mm × 31.5 mm × 11.9 mm
for the conventional configuration and 31.5 mm × 31.5 mm × 13.9 mm Fresnel lens WFr = 150 LFr = 150 tFr = 150 AFr = 22500
PV tPV = 4.5 APV = 992.25
for the newly developed configuration. The computational domain in-
Glass cover Wgc = 31.5 Lgc = 31.5 tgc = 3
cludes the solid domain and fluid domain. The solid domain includes
Silicon wafer Wsw = 31.5 Lsw = 31.5 tsw = 0.2
both PV and TEG layers and microchannel aluminum walls, as shown in LVA layer We = 31.5 Le = 31.5 te = 0.5
the conventional configuration. In the newly developed configuration, Tedlar layer Wte = 31.5 Lte = 31.5 tte = 0.3
the solid domain includes both PV and STEG layers, which contain the TEG tTEG = 6.9 ATEG = 992.25
ceramic plate
same layer of TEG in addition to the absorber plate, and microchannel Wc = 31.5 Lc = 31.5 tc = 0.8
P/N leg Xleg = 1.5 Lleg = 1.5 tleg = 5
aluminum walls. The dimensions of the PV and TEG layers are sum-
Copper strips ts = 0.15
marized in Table 1. The TEG layers include two ceramic plates, each with Solar absorber Wab = 31.5 Lab = 31.5 tab = 2 A ab = 992.25
the dimensions Wc , Lc , tc = 31.5 × 31.5 × 0.8 mm3 , copper strips with a Microchannel tMHS = 0.5
thickness of 0.15 mm, and 14 × 14 = 196 legs with the dimensions of channel Wmch = 31.5 Lmch = 31.5 tmch = 0.1
Wl , Ll , tl = 1.5 × 1.5 × 5 mm3 . In the new configuration, the aluminum walls tw = 0.2
Conventional Wmod = 31.5 Lmod = 31.5 tmod = 11.9 Amod = 992.25
absorber plate of dimensions Wab, Lab , tab = 31.5 × 31.5 × 2 mm3 is
module
added to the TEG. The fluid computation domain for both studied New module Wmod = 31.5 Lmod = 31.5 tmod = 13.9 Amod = 992.25
designs has the dimensions of (Wmch 2tw1), Lmch , hmch =

1386
A. Abdo, et al. Energy Conversion and Management 195 (2019) 1382–1401

Table 2 Ef = T + J (13)
Physical properties of materials used in the present study.
The generated power per unit volume Q can be calculated from the
Parameter Density (kg/ Specific heat Thermal conductivity
m3) (J/kg.K) (W/m.K) following equation:

Glass 3000 500 2 Q = Ef . J (14)

Encapsulation (EVA) 960 2090 0.35
Silicon 2330 677 148 By substituting Eqs. (12), (13), and (14) into Eq. (11), the tem-
Tedlar 1200 1250 0.2 perature distribution in the interior of the TEG is then given by:
Aluminum 2719 871 202.4
Ceramic 3570 837 35 d
.( T) + J . J T T. J = 0
Copper 8933 400 385 dT (15)
Bismuth telluride 7530 544 Shen [33] d
Bi2 Te3 where is defined as the Thomson coefficient and , , J , and are
T dT
the solid thermal conductivity, Seebeck coefficient, current density, and
electrical resistivity, respectively. For the maximum electrical power
the top EVA layer, and the silicon wafer layer, respectively; gc , gc , sw , produced by the TEG in the STEG, Rin = Rload . Hence,
and e are glass absorptivity, glass transmissivity, silicon absorptivity, 2
Vopen
and EVA absorptivity, respectively; gc , top, e , sw are the volume of ETEG = . Rload
glass, top EVA, and silicon wafer, respectively and G is the net con- Rin + Rload (16)
centrated radiation. The total heat absorbed by the PV material, which ETEG
is converted into thermal energy, is expressed as follows: =
STEG
Qab (17)
QPV = Qgc + Qt , e + Qsw (5) where Vopen, Rin , and Rload are the open voltage for the TEG electrical
circuit, internal resistance of the TEG device, and load resistance, re-
EPV = sw . gc . t , e . (G ) (6)
spectively.
sw = PV = ref (1 ref . (Tsw Tref )) (7)
3.3. Microchannel heat sink
where ref and ref are the efficiency of the solar cell and the tem-
perature coefficient at a standard temperature of Tref = 25 °C, respec- For microchannel walls, heat transfer by conduction is expressed by
tively, as previously reported [29]. Subsequently, the heat transfer rate the following equation:
through the interface between the solar cell layers can be calculated as:
.( mch Tmch ) = 0 (18)
q1 = q2 = 1 T1 = 2 T2 (8)
The governing equations of steady state fluid flow are written as
where subscript 1 and 2 are related to the layer from and to where the follows:
heat transfers, respectively. Conservation of mass:

3.2. STEG – layers

. ( V) = 0 (19)
Conservation of momentum:
The following equation is used to calculate the solar energy ab-
sorbed by STEG: V. ( V) = P+ (µ . V) (20)
Qab = ab. G (9) Conservation of energy:

where ab is the absorptivity of the aluminum absorber sheet, which V. ( Cco Tco) = .( co Tco) (21)
should be as high as possible to absorb most of the solar energy and
convert it into thermal energy. Varying of the thermo-physical properties of the water versus
The governing equations of heat transfer and electrical current can temperature is determined as reported in Jayakumar [38].
be formulated in vector form as follows [37]:
3.4. System performance characterization
.J =0 (10)
The electrical and thermal efficiency of the system can be de-
. q + Q = 0 (11)where q is the reversible heat transfer by the termined using the following equations:
Peltier effect and is expressed using Fourier’s law as follows:
EPV + ETEG
=
q = TJ T (12)
tot , ele
Qtot (22)

The electric field density Ef at any location is determined from the Qth
=
combination of the Seebeck and Ohm effects as follows:
tot , th
Qtot (23)

Table 3
Optical properties of materials used in the present study.
Part name Reflectivity Transmissivity ( ) Absorptivity ( ) Emissivity ( )

Glass 0.04 0.92 0.04 0.85

Silicon cell 0.08 0.02 0.9 –
EVA 0.02 0.9 0.08 0.9
Tedlar 0.86 0.012 0.128 0.9
Ceramic – – 0.92
Solar absorber [26] – – 0.9 0.15

1387
A. Abdo, et al. Energy Conversion and Management 195 (2019) 1382–1401

where the useful thermal heat Qth absorbed by water is determined as [Link]. TEG – layers. For the interface surfaces of the TEG layers that
follows: are parallel to the upper surface at 0 x Wmod, 0 y Lmod , the
. heat transfer by diffusion is the same for the following layers:
Qth = Qtot EPV ETEG Ltot = mco . Cp, co. (Tco, o Tco, i ) (24)
Interface between the hot ceramic and the hot copper strips at
Losses through radiation and convection Ltot ,1, Ltot ,2 for the con- z = tmch + tc + 2ts + tleg ,
ventional design and new design, respectively, are calculated from the Interface between the hot copper strips and TEG legs at
following equations: z = tmch + tc + ts + tleg ,
Interface between the TEG legs and the cold copper strips at
Ltot ,1 = hgc . (Tgc Tam) + gc . . (Tgc4 Tsk4 ) + h wl . (Twl Tam) (25) z = tmch + tc + ts ,
Ltot,2 = hgc . (Tgc Tam) + 4 4 4
Tsk4 ) Interface between the cold copper strips and the cold ceramic at
gc . . (Tgc Tsk ) + hab. (Tab Tam ) + ab. . (Tab
z = tmch + tc ,
(26) Interface between the cold ceramic and the upper wall of the mi-
where Tsk is the sky temperature and is determined by knowing the crochannel atz = tmch
ambient temperature as previously reported [39]; hab, hgc , h wl are
convection heat transfer coefficients from the solar absorber, the glass [Link]. Microchannel heat sink. At the channel inlet
cover, and the back surface of the microchannel heat sink in the con- twl x Wmod w , y = 0, twl z hmch + twl
ventional design, respectively, and are estimated from the equations
u = 0, v = constant , w = 0, Tco, i = Tam = 30°C
previously reported [40]; , gc, ab are the Stephan–Boltzmann con-
stant, and the emissivities of the glass cover and solar absorber, re- At the channel outlet twl x Wmod twl, y =Lmod , twl z hmch + twl
spectively. The total input energy for the conventional design and the pout = 0 (atmospheric pressure ) , and k z = 0 due to heat transfer by
T

new design Qtot ,1, and Qtot ,2 , respectively, is estimated from the fol- convection between parallel layers of laminar flow.
lowing equations: For the lower aluminum surface 0 x Wmod, 0 y Lmod , z =0,
the heat transfer by convection between the surface and the sur-
Qtot ,1 = ( gc + gc . e + gc . e ). (G ) (27)
roundings is given by:
Qtot ,2 = ( gc + gc . e + gc . e ). (G ) + ab. (G ) (28) T
k wl = h wl (Twl Tam)
z
3.5. Boundary conditions
3.5.2. Newly developed photovoltaic/solar thermoelectric generator hybrid
The conditions at the boundaries must be identified to perform a system
numerical simulation. For the conventional design, the thermal [Link]. PV layers. Similarly, boundary conditions of the conventional
boundary condition for the outside surface of the glass layer of the PV is design can also be applied to the newly developed design.
the heat loss by radiation and convection. The radiation heat loss is the
loss between the surface of the glass and the sky. The loss by convection [Link]. STEG – layers. For the upper surface of the aluminum
is due to wind effect. All side surfaces of the computational domain for absorber at 0 x Wmod, 0 y Lmod , z = 0 , there is heat
the PV module and the attached TEG are assumed to be insulated. The transfer by radiation and convention between the glass surface
lower surface of the computational domain is subjected to heat loss by and the surroundings according to the condition kab
convection. Second, for the new design, both the PV module and STEG T
= hab (Tab Tam) + ab (Tab4
Tsk4 )
z
have outer surfaces that are exposed to the surroundings and heat is For the other parts, the boundary conditions of the conventional
dissipated through radiation and convention from both sides. design are applied to the newly developed design.
The velocity in y-direction (v) through the wide microchannel is
assumed to be uniform and the other components are assumed to be [Link]. Microchannel heat sink. Similarly, the boundary conditions of
zero. The inlet temperature is uniform and equal to the ambient tem- the conventional design are applied to the newly developed design.
perature [41].
4. Numerical solution
3.5.1. Conventional photovoltaic/thermoelectric generator hybrid system
[Link]. PV – layers. For the upper surface of the glass at The governing equations of the PV layers, the microchannel heat
0 x Wmod, 0 y Lmod , z = tmch + tTEG + tPV , there is heat transfer sink, and the TEG are numerically solved to determine the solar cell and
by radiation and convention between the glass surface and TEG temperatures. Accordingly, the thermal and electrical power and
the surroundings according to the following equation: other performance parameters can be determined. The CFD ANSYS
T 4
kgc z = hgc . (Tgc Tam) + gc. . (Tgc Tsk4 ) FLUENT 17.2 code was utilized to numerically solve the comprehensive
For the interface between the PV layers that are parallel to the upper model. The mechanical APDL solver was used to calculate the ther-
surface at 0 x Wmod, 0 y Lmod , heat transfer by diffusion is given moelectric model for the TEG and was coupled with FLUENT to solve
as: the thermo-fluid models. Details pertaining to the numerical scheme
Glass cover, top EVA interface at z = tmch + tTEG + tte + 2te + tsw , and solution procedures for TEG have been documented [42] and are
Top EVA, silicon wafer interface at z = tmch + tTEG + tte + te + tsw , presented in the flow chart in Fig. 3. Parallel computing was performed
Silicon wafer, Lower EVA interface at z = tmch + tTEG + tte + te , using a Dell Precision T7500 workstation.
Lower EVA, Tedlar interface at z = tmch + tTEG + tte ,
and Tedlar rare surface at z = tmch + tTEG . 4.1. Mesh independent test
For all side surfaces of the PV layers (glass, EVA, Tedlar) at
x = 0, x = Wmod, 0 y Lmod , y = 0, y = Lmod , 0 x Wmod for all The grid independent test is conducted for all components, in-
z values. There is no heat transfer in the x and y direction according to cluding the PV module, the microchannel heat sink, and the TEG. The
T
the conditions: ki x = 0 and kl y = 0 due to the adiabatic boundary
T
computational field includes the solid domain and the fluid domain. In
condition assumptions, where ki is the thermal conductivity of layer i the fluid domain, the cells are hexahedrons with high density at the
that represents the absorber, hot ceramic layer, strips, legs, and cold boundary layer region adjacent to the microchannel walls. Similarly,
ceramic layer. for the solid domain, hexahedron cells are adopted. However, for the

1388
A. Abdo, et al. Energy Conversion and Management 195 (2019) 1382–1401

available numerical and experimental data were obtained from studies

by Joshi et al. [43], Slimani et al. [44], Radwan and Ahmed [35], Shen
et al. [33], and Mahmoudinezhad [45].

4.2.1. Photovoltaic model validation

The predicted average temperature of the PV model is compared
with that measured by Joshi et al. [43] and numerically predicted by
Slimani [44] as shown in Fig. 5-a. Comparisons show a good accordance
between the current predicted results and those measured by Joshi
et al. [43], with the maximum difference in average temperature of
about 3.4 % and about 4 % compared with the numerical estimated
values by Slimani [44].
Fig. 5-b presents the comparison between the estimated outlet air
temperature, the measurements [43], and the theoretical results [44].
The maximum error is approximately 7.0 % for the experimental mea-
surements and 2.3 % for the numerically calculated values. Fig. 5-c
shows the correlation between the predicted power generated and the
measurements [43] and the theoretical results [44]. The error was
found to be approximately 8 % for the experimental measurements and
3.3 % for the numerically calculated values. Moreover, the predicted
maximum cell temperatures are compared with those calculated nu-
merically by Radwan and Ahmed [35], as shown in Fig. 5-d. A reliable
consistency is observed with the highest deviation of about 1.3 %.

4.2.2. Thermoelectric generator model validation

The predicted maximum generated power and the conversion effi-
ciency of the STEG model at different hot ceramic temperatures for
bismuth telluride-based thermoelectric material are compared with
those numerically estimated by Shen et al. [33], as shown in Figs. 6-a
and b. Comparisons indicate no significant difference with the highest
deviation of around 3.0 % in power and 2.0 % in conversion efficiency.
This is mainly attributed to the assumptions made when developing the
mathematical model of the TEG system.

4.2.3. Photovoltaic–solar thermoelectric generator model validation

Lastly, for the integrated PV–STEG system, comparisons between
both the solar cell average temperature and the hot ceramic tempera-
ture and the measurements of Mahmoudinezhad [45] are performed as
detailed in Figs. 7-a and b. Reasonable agreement was obtained with a
maximum error of 8 % in the solar cell average temperature and about

Fig. 3. Flowchart of numerical simulations.

ceramic plate, tetrahedron cells are used. The number of grids utilized
to estimate the average hot ceramic temperature and cell temperature is
shown in Figs. 4-a and b. In addition, further detail regarding the error
associated with solar cell temperature, pressure drop, and cold ceramic
temperature are summarized in Table 4. The number of selected grids is
2 × 107 where the maximum errors do not exceed 0.4%.

4.2. Model validation

To validate the developed models for concentrator PV/TEG/heat Fig. 4-a. Mesh independent test for the TEG system. CR_PV = CR_STEG = 10
sink, several sets of numerical and experimental results are used. The suns, Tco,i = Tam = 30 °C.

1389
A. Abdo, et al. Energy Conversion and Management 195 (2019) 1382–1401

Fig. 4-b. Mesh independent test for the PV system. CR_PV = CR_STEG = 10
suns, Tco,i = Tam = 30 °C. Fig. 5-a. Comparison between the predicted average solar cell temperature and
both experimental [43] and numerical data [44].

7 % for the average hot ceramic plate temperature.

5. Results and discussion

The current section has three subsections. The first part serves to
present the performance characteristics of the conventional con-
centrated PV/TEG-microchannel heat sink. These characteristics in-
clude the variation of average cell temperature, system output power,
and electrical efficiency at different values of CRs of the PV system. In
the same manner, the second part assesses the performance of the
newly developed concentrator PV and STEG hybrid system combined
with the microchannel heat sink in which the solar CR is varied for the
PV module and the STEG. Lastly, detailed comparisons between the
performance parameters of the newly developed hybrid system and
both the conventional PV/TEG and the stand-alone PV system are
performed.

5.1. Performance of the conventional concentrator PV/TEG hybrid system

In such design, the TEG works as a heat sink for the PV system, as
the hot ceramic plate surface is in contact with the PV back surface,
which is attached to the Tedlar layer. Furthermore, the microchannel
functions as a heat sink for the TEG, as the surface of the cold ceramic
plate is attached to the upper aluminum wall of the microchannel. Here, Fig. 5-b. Comparison between the calculated outlet air temperature and both
two sources of electrical output power are available: one from the PV experimental [43] and numerical data [44].
system; the other from the TEG device in which the thermal energy
from the PV back surface is used by the TEG and transformed into 5.1.1. Average solar cell temperature
electrical energy. The remaining part is used as thermal energy by way Fig. 8-a shows the variation of the average cell temperature with
of the microchannel heat sink. Because temperature is the key para- Reynolds number (Re) at distinct values of CR up to 20. As outlined,
meter, the variation in the average cell temperatures under distinct increasing the Re up to 20 results in a markedly reduction of the
operating conditions is explained along with the electrical power and average cell temperature at all values of CRs. Increasing the Re beyond
electrical efficiency.

Table 4
Variations of tested parameters at different mesh numbers.
Parameter Maximum value Minimum value Difference Error %

Silicon Wafer (°C ) 51.71278 51.93882 0.23 0.4%

Cold Ceramic Plate (°C ) 31.61925 31.6223 0.0031 0.01%
Pressure drop (Pa) 18910.46 18849.22 61.24 0.3%

1390
A. Abdo, et al. Energy Conversion and Management 195 (2019) 1382–1401

Fig. 6-a. Comparison between the calculated maximum power and numerical
Fig. 5-c. Comparison between the calculated power and both experimental results of [33] at Tc .c = Tam = 27 °C, and R = RL.
[43] and numerical data [44].

Fig. 6-b. Comparison between the calculated conversion efficiency and nu-
merical results of [33] at Tc .c = Tam = 27 °C , and R = RL.

Fig. 5-d. Comparison between the calculated maximum solar cell temperature eventually damage the solar cell. The maximum CR_PV in this case must
and numerical results [35]. CR_PV = 20 suns, Tam = Tco,i = 30 °C. be less than 10 suns to avoid a higher temperature than the maximum
allowable value.
Fig. 8-b presents the variations of the average temperature of the
that value causes a minor reduction in the average cell temperature.
hot ceramic plate versus Re at distinct values of PV_ CR varying from 1
Several explanations have been reported seeking to determine the
to 20 suns. Over the whole span of PV CRs, the average temperature of
reason for such a trend. This trend has been reported owing to the
the hot ceramic plate reduces with increasing Re up to 20. Further in-
decrease in contact time between the upper wall and the coolant, as a
crease of the Re beyond 20 causes the average temperature of the hot
result of higher velocities associated and interrelated with higher Re.
ceramic plate to slightly decrease. A similar pattern was previously
Further details can be found in a previous study [7]. Despite this, at any
revealed for the variation in cell temperature. Additionally, it is ob-
value of Re, the average cell temperature rises as the CR increases
served that the solar radiation striking the PV surface has a considerable
owing to a rise in the incident solar irradiance. For example, at
effect on the hot ceramic plate temperature at all ranges of Re. For
Re = 50, the average solar cell temperature rises from 35 °C to 147 °C
instance, at Re = 50, the increase in the PV solar CR (CR_PV) from 1 to
as CR_PV varies from 1 to 20. However, the maximum tolerable solar
20 suns prompts an increase in the hot ceramic plate temperature to
cell temperature is approximately 90 °C [35] as observed at
approximately 110 °C.
CR_PV = 10 and increasing the temperature beyond this value will

1391
A. Abdo, et al. Energy Conversion and Management 195 (2019) 1382–1401

Fig. 7-a. Comparison between the calculated average cell temperature and
Fig. 8-a. Average solar cell temperature variation versus Re at distinct values of
previous measurements [45].
CR_PV.

Fig. 7-b. Comparison between the calculated average hot ceramic temperature Fig. 8-b. Average hot ceramic surface temperature variation versus Re at dis-
and previous measurements [45]. tinct values of CR_PV.

5.1.2. System output power solar irradiance. Further increase in CR beyond 15 slightly enhances the
Fig. 9-a presents the variation of the electrical power generated output power. This is because the cell temperature is so high at those
from the PV system versus the values of Re at different PV solar CRs. As values of CR such that it declines the electrical efficiency of the solar
seen in the figure, at Re less than 20, a minor increase in electrical cell system and the output electrical power.
power with increasing Re is observed at CR = 1 and 5. By increasing CR Fig. 9-b presents the variation of electrical power generated from
beyond 5, a noticeable increase in electrical power is seen. The more the TEG system versus the Re at varying CR ranging from 1 to 20. As
the CR, the higher the increase in electrical power. This is owing to the shown in the figure, increasing the CR noticeably increases the output
following, based on Fig. 8-a, it is recorded that at higher CRs and low power due to the rise in the temperature of the hot ceramic plate, which
Re, the temperature is remarkably reduced with increasing Re. Conse- thereby boosts the temperature difference across the TEG. This coin-
quently, it boosts the electrical efficiency and in return raises the output cides with the increase in the conversion efficiency of the TEG, and
electrical power. This is because of the reduction in cell temperature. consequently the output electrical power. In addition, at any value of
Similarly, at higher Re, the effect of increasing Re slightly enhances the CR and at Re lower than 20, increasing the Re enhances the output
output electrical power. At all values of Re, rising CR from 1 up to 15 electrical power. This is ascribed to the reduction in the temperature of
remarkably increases the output power owing to the increase in the the cold ceramic plate, which helps reduce the difference in

1392
A. Abdo, et al. Energy Conversion and Management 195 (2019) 1382–1401

5.1.3. System efficiency

Varying the PV electrical efficiency versus the Re of the micro-
channel flow at distinct values of CR is shown in Fig. 10-a. Based on the
figure, the electrical efficiency is noticeably reduced when the CR va-
lues are increased. This is attributed to the increase in cell temperature
that occurs as the CR increases. At a Re of 50 and a CR of 1, the elec-
trical efficiency is about 18 %. A further increase in CR up to 20 reduces
the electrical efficiency to around 8 %. The effect of raising the Re on
the electrical efficiency depends on the CR and the Re. At a Re of less
than 20, increasing the Re remarkably improves the electrical efficiency
at higher values of CR, while this effect is reduced at low values of CR.
At a Re greater than 20, increasing the Re slightly boosts the electrical
efficiency, and this trend is marked for all values of CR. For the TEG, an
opposite trend is observed, in that rising the values of CR tends to in-
crease the TEG electrical efficiency, as indicated in Fig. 10-b. This is
because raising the value of CR increases the hot side temperature of
the TEG. Subsequently, the temperature difference across the TEG
(between the hot and cold sides) enhances the electrical efficiency of
the TEG device. The maximum TEG conversion efficiency was around
3 % at CR = 20 and the lowest was around 0.12 % at CR = 1. Fig. 10-c
shows the variation of the total electrical efficiency of the PV/TEG
hybrid system, with the Re at CRs varying from 1 to 20. As shown in the
figure, at a Re less than 20, rising the Re slightly boosts the total
Fig. 9-a. Variation of PV electrical power versus Re at distinct values of CR_PV. electrical efficiency at low CRs. The effect of increasing the Re is sub-
stantial at high CR. At a Re greater than 20, increasing the Re has in-
significant effects on the efficiency for all investigated values of CR.
Contrarily, at any value of Re, increasing the CR noticeably reduces the
total electrical efficiency.
In conclusion, based on Figs. 8, 9, and 10, the highest tolerable
working temperature for a solar cell is about 90 °C , and any additional
increase in cell temperature not only reduces the electrical efficiency
significantly, but also potentially damages the solar cell. Thus, the
maximum acceptable PV solar CR should not exceed 10 suns at Re
higher than 40 (Re > 40). In compliance with this, the maximum
available total generated electrical power is about 1.25 kW/m2 and the
electrical efficiency of the system is approximately 15 %.

5.2. Performance of the new concentrator PV/STEG hybrid system

In this section, the performance of the newly developed PV/STEG

hybrid system is investigated when subjected to concentrated

Fig. 9-b. Variation of TEG electrical power versus Re at distinct values of

CR_PV.

temperature across the TEG to a little extent. The higher difference in

temperature between the cold and hot sides of the TEG amplifies its
conversion efficiency and the output electrical power.
Fig. 9-c outlines the variation of the total electrical power from the
PV module and the TEG system per unit area versus the Re under dif-
ferent values of CR. At low CRs of 1 and 5, increasing the Re has no
remarkable effect on the total electrical power. At a CR of 10 and
higher, increasing the Re up to 30 improves the total electrical power.
Any additional increase in the Re beyond this value slightly increases
the total electrical power. Furthermore, at any value of Re, rising the CR
markedly raises the total electrical power. This is because increasing the
CR enhances the temperature difference across the TEG, which boosts
the power generated from the TEG. A similar effect is observed wherein
the PV output electrical power also increases as the CR increases. Fig. 9-c. Variation of total electrical power versus Re at distinct values of
CR_PV.

1393
A. Abdo, et al. Energy Conversion and Management 195 (2019) 1382–1401

Fig. 10-c. Variation of total electrical efficiency versus Re at distinct values of

Fig. 10-a. Variation of PV electrical efficiency versus Re at distinct values of CR_PV.
CR_PV.

Fig. 11-a. Average cell temperature variation versus Re at distinct values of

Fig. 10-b. Variation of TEG electrical efficiency versus Re at distinct values of CR_PV and CR_STEG = 10 and 20.
CR_PV.
value of CR_PV. Yet, at any value of Re, raising the CR_PV causes a
irradiance from both sides (PV and STEG sides). Two reflecting surfaces significant increase to the average cell temperature. For example, at
are used to redirect solar radiation toward the surfaces of the PV Re = 50, the average cell temperature is approximately 33 °C at
module and the STEG simultaneously. Both the PV module and the CR_PV = 1; it increases to 75 °C at CR_PV = 20. The same trend can be
STEG are cooled by the same microchannel heat sink to reduce the observed by increasing CR_STEG to 20, as seen in Fig. 11-a. At Re = 50,
pumping power demand. The variations of PV average temperatures the average solar cell temperature varies from 35 °C at CR_PV = 1, up to
under different operating conditions are presented, along with elec- 77 °C at CR_PV = 20. To conclude, raising CR_STEG from 10 to 20 has a
trical power and efficiencies. slight effect on the average cell temperature, whereas at CR_PV = 20
and Re = 50, it varies from 75 °C to 77 °C as the CR_STEG increases
from 10 to 20.
5.2.1. Average solar cell temperature By keeping the solar cell CR_PV constant at 10 and varying the CR of
Fig. 11-a shows the variations of the average temperature of the the STEG from 1 to 20, Fig. 11-b shows the variation of the average cell
solar cell with Re at distinct values of CR_PV, varying from 1 to 20, temperature. Increasing the Re reduces the average cell temperature;
while the CR of the STEG (CR_STEG) surface is kept constant at 10. As this is seen for all values of CR_STEG. In contrast, at a low value of Re,
evidenced, rising the Re markedly reduces the cell temperature at any

1394
A. Abdo, et al. Energy Conversion and Management 195 (2019) 1382–1401

increasing the CR_STEG causes a small increase in the average solar cell at a constant Re of 50. As seen in the figure, the total electrical power
temperature and with further increase in Re, the effect of CR_STEG is moderately rises with the boost in CR_STEG, while it markedly in-
weakened. It is found that an increase in CR_STEG from 1 to 20 suns creases with increasing PV CR, as shown in Fig. 13-f. The reason for this
causes a boost in the average solar cell temperature of approximately is that increasing the CR_PV causes a greater increase in electrical
13.5 °C at Re = 10 and 2 °C at Re = 50. A further increase in CR_PV to power than that obtained from the STEG by increasing the CR_STEG. It
20 results in the same trend, as shown in Fig. 11-b. As CR_STEG in- is found that the increase in total electrical power as the CR_PV varied
creases from 1 to 20 suns, the average cell temperature increases by from 1 to 20 ranged from 0.16 kW/m2 to 2.58 kW/m2. However, in-
15 °C at Re = 10 and 2 °C at Re = 50. Based on Fig. 11-b, the average creasing the CR_STEG from 1 to 20 results in increase in the total
solar cell temperature is significantly increased from around 55 °C to electrical power from 0.16 kW/m2 to 0.6 kW/m2.
75 °C, as the CR_PV varies from 10 to 20 suns.
Based on Fig. 11, one can conclude that increasing CR_STEG has an
insignificant influence on the average cell temperature. As previously 5.2.2. System efficiency
mentioned, at a CR_STEG of 10, the average cell temperature is about Fig. 14-a shows the variations in the electrical efficiency of the solar
75 °C at a CR_PV of 20. Increasing the CR_STEG to 20 results in in- cell with CR_STEG at distinct values of PV solar CRs ranging from 1 to
creasing the cell temperature to 77 °C, at the same CR_PV and at a 20, while the Reynolds number is kept constant at 50. The reason is that
constant value of Re. However, increasing the CR_PV significantly raises at Re = 50, the solar cell temperature reduces to the lowest value. With
the solar cell temperature. As the CR_PV varies from 10 to 20, the cell an additional increase in the Re beyond that value, there is no observed
temperature increases from 52 °C to 77 °C, while the Re is maintained at reduction in solar cell temperature. Furthermore, the coolant flow rate
a constant value of 50. increases and in turn increases the power consumption.
Further clarification of the average solar cell temperature variation In this figure, a very slight reduction in the electrical efficiency of
at distinct values of CRs of both STEG and PV systems at a Re of 50 is the PV module is observed with rising CR_STEG. However, the solar cell
presented in Figs. 12-a and b. The average cell temperature variation electrical efficiency significantly reduces with rising CR_PV, as a con-
versus CR_STEG at distinct values of CR_PV ranged between 1 and 20, sequence of the associated rise in temperature. This trend can be ob-
and is shown in Fig. 12-a. As shown in the figure, for all values of PV served at any value of CR_STEG. At CR_PV = 1, the solar cell electrical
CRs, the average solar cell temperature slightly rises with increasing efficiency is about 19.3 %, and by increasing the CR_PV to 20, the
STEG CR. The maximum rise in the average solar cell temperature is electrical efficiency of the solar cell reduces to about 15.5 %. Similar
around 2 °C, as the CR_ STEG varies from 1 to 20. This may be owing to results are observed for the variations in PV electrical efficiency with
positioning of the microchannel heat sink between the PV module and both the CR_PV and CR_STEG, as seen in Fig. 14-b.
the STEG, and its ability to absorb most of the thermal energy coming The variation of STEG efficiency at varying CR_PV and CR_STEG are
from the low-temperature surface of the STEG. Furthermore, the low shown in Figs. 14-c and d. The STEG efficiency is greatly enhanced by
conductivity of the STEG legs reduces the heat transfer from the hot increasing the CR. It varies from 0.5 % to 3.5 % as the CR_STEG rises
ceramic surface to the cold ceramic surface. Distinctly, at any value of from 1 to 20. Nevertheless, increasing the CR_PV yields insignificant
CR_STEG, increasing the values of CR_PV significantly increases the effects on the TEG efficiency. This may be due to the following reasons:
average solar cell temperature. The rise in temperature is almost the temperature difference across the STEG is the driving force behind
identical for all values of CR_STEG, and is equal to 43 °C, as the PV CR the recovery of more electrical energy from the absorbed thermal en-
varies from 1 to 20. At the maximum value for the CR for both the PV ergy flowing through the STEG. Solar radiation promotes the rise in the
module and the STEG, the average solar cell temperature is about hot side temperature of the STEG, while the PV solar radiation pro-
76.6 °C. A further explanation on the effect of raising the PV CR on the motes the rise in the cold side temperature of the STEG. Consequently,
average cell temperature at a constant Re of 50 is illustrated in Fig. 12- enhancement in the electrical efficiency of the STEG is dominated by
b. As seen in the figure, a marked rise of average solar cell temperature the intensity of the solar radiation that strikes the STEG surface.
is noted as the PV_ CR increases. However, there is a slight increase in The effects of varying the CR of both the PV module and the STEG
the average cell temperature with an increase in the CR_ STEG, and this
increase is almost the same for any value of the CR_ PV. By factoring in
the observations made in Fig. 12, a similar conclusion to that made
previously can be drawn, wherein increasing the CR_STEG has an in-
significant effect on the average solar cell temperature. Contrarily, a
marked increase in the average cell temperature is observed as the
CR_PV increases.

[Link]. System output power. The PV electrical power variation with

CR_STEG at distinct values of CR _PV and a constant Re of 50 is shown
in Fig. 13-a. The PV electrical power is approximately independent of
the CR of the STEG, and it markedly rises with an increase in the CR_PV.
The same trend is observed in Fig. 13-b.
Fig. 13-c presents the STEG electrical power variation with the CR of
the STEG at different values of CR_PV, while the Re remains constant at
50. As evidenced in this example, the STEG electrical power markedly
increases along with the increase in the CR of the STEG. This is owing to
the rise in the hot ceramic plate temperature due to the increase in
CR_STEG compared with the temperature of the cold ceramic plate.
Accordingly, the difference in temperature between the hot and cold
surfaces of the TEG becomes higher, augmenting the electrical power
generated. Despite this, the STEGelectrical power is independent of the
CR_PV. This same conclusion can be ascertained from Fig. 13-d
Fig. 13-e shows the total electrical power variation of both the PV Fig. 11-b. Average cell temperature variation versus Re at distinct values of
and the STEG versus the CR of the STEG at distinct values of PV CRs and CR_STEG and CR_PV = 10 and 20.

1395
A. Abdo, et al. Energy Conversion and Management 195 (2019) 1382–1401

Fig. 13-a. Variation of the PV electrical power versus CR_STEG at distinct va-
lues of CR_PV.
Fig. 12-a. Average solar cell temperature variation versus CR_STEG at distinct
values of CR_PV.

Fig. 13-b. Variation of the PV electrical power versus CR_PV at distinct values
of CR_ STEG.

Fig. 12-b. Average solar cell temperature variation versus CR_PV at distinct
values of CR_STEG. CR_STEG, increasing the CR_PV significantly increases the total elec-
trical efficiency, as clearly shown in Fig. 15-a. This is attributed to the
remarkable rise in the system electrical output power, which is ob-
on the total electrical efficiency are presented in Figs. 15a and b. In served as the CR_PV increases.
these figures, the total electrical efficiency is illustrated as the total Similarly, Fig. 15-b presents the variation of total electrical effi-
electrical power of both the PV and STEG system, divided by the total ciency with varying PV CRs. Based on the figure, the total electrical
solar power incident on both systems, as shown in Eq. (22). At efficiency significantly increases with increasing CR_PV at all values of
CR_PV = 1, the total electrical efficiency reduces with the rise in the CR_STEG, except at CR_STEG = 1. The efficiency increases till
CR_STEG up to 8 and with further increase in CR_STEG, the total CR_PV = 8, and then slightly decreases with the increase in the CR_PV.
electrical efficiency increases, as shown in Fig. 15-a. However, for a The maximum obtained total electrical efficiency is around 16 % at
CR_PV ranging between 5 and 20, the total electrical efficiency de- CR_PV = 10 and CR_STEG = 1.
creases with increasing CR_STEG. The reason is most likely owing to the
slight increase in the electrical output power from the STEG system
compared with the incident solar power, where it varies from 5.3. Performance assessment of the new PV/STEG hybrid system
0.0016 kW/m2 to 0.6 kW/m2, as the incident irradiance on the surface
of the STEG varies from 1 to 20 kW/m2. Additionally, at any value of The predicted performance characteristics of the new PV/STEG
system, and both the conventional PV/TEG as well as the stand-alone

1396
A. Abdo, et al. Energy Conversion and Management 195 (2019) 1382–1401

Fig. 13-c. Variation of STEG electrical power versus CR_STEG at distinct values Fig. 13-e. Variation of total electrical power versus CR_STEG at distinct values
of CR_PV. of CR_PV.

Fig. 13-d. Variation of STEG electrical power versus CR_PV at distinct values of Fig. 13-f. Variation of total generated power versus CR_PV at distinct values of
CR_STEG. CR_STEG.

PV system are compared, as seen in Figs. 16-a, b, and c. All systems are heat sink. In consideration of this, such a system cannot be used at
integrated with microchannel heat sinks. Fig. 16-a presents the average higher CRs because of the associated high temperatures, because the
cell temperature variation versus the CR_PV for the three systems; the maximum tolerable temperature is approximately 90 °C. Therefore, the
average cell temperature for the conventional PV_TEG-microchannel maximum operating CR_PV must not exceed 10 suns. Conversely, the
rises from 35 °C to approximately 147°C , as the CR_PV increases from 1 new design is more resilient than the conventional design, as it can
to 20. However, for a stand-alone PV system combined with a micro- operate at CR_PV = 20 with a maximum solar cell temperature of 77 °C.
channel heat sink, the average cell temperature increases from 35°C up Fig. 16-b shows the electrical power output variation versus CR_PV
to 75 °C with the same increase in CR_PV. For the new PV/STEG hybrid for the three different designs. According to the figure, the total elec-
system, by varying the CR_PV from 1 to 20, the average cell tempera- trical power output for the conventional design increases to approxi-
ture rises from 35°C up to 75°C , 76°C and 77°C for CR_STEG of 1, 10, mately 1.9 kW/m2, as the CR_PV increases to 20. When considering the
and 20, respectively. On the basis of these results, it is evident that the PV system alone, the electrical output power increases to about 2.6 kW/
average cell temperature of the conventional PV/TEG is the highest in m2, under the same CR_PV. For the new design, the total electrical
comparison with those of other systems. The reason is attributed to the power output under the same CR_PV is approximately 2.6, 2.7, and
high thermal resistance of the heat transfer associated with the ex- 3.2 kW/m2, in the case of a CR_STEG = 1, 10, and 20, respectively.
istence of the TEG unit between the PV module and the microchannel Based on these results, using the conventional PV/TEG system produces

1397
A. Abdo, et al. Energy Conversion and Management 195 (2019) 1382–1401

Fig. 14-a. Variation of PV electrical efficiency versus CR_STEG at distinct va- Fig. 14-c. Variation of STEG electrical efficiency versus CR_STEG at distinct
lues of CR_PV. values of CR_PV.

Fig. 14-b. Variation of PV electrical efficiency versus CR_PV at distinct values

of CR_STEG. Fig. 14-d. Variation of STEG electrical efficiency versus CR_PV at distinct va-
lues of CR_STEG.

the least total electrical power in comparison to the other designs.

Additionally, if the value of CR_PV is selected to be 10 suns so as to thermal energy is approximately 12 kW/m2, and this is the same for
avoid higher solar cell temperatures compared to the maximum al- both the stand-alone PV system and the conventional design wherein
lowable temperature, the total output of electrical power will be ap- the PV module is integrated with the TEG. The new design achieves the
proximately 1.28 kW/m2. Therefore, the new design helps increase the highest rate of thermal energy of 15, 22, and 30 kW/m2 for CR_STEG of
amount of power generated by increasing the amount of solar energy 1, 10, and 20, respectively, at the same CR_PV. This is as a result of the
absorbed by the STEG. thermal heat absorbed in the new design due to the positioning of the
Fig. 16-c presents the rate of thermal energy variation versus CR_PV STEG with respect to the microchannel. The position helps to increase
for the three different designs. As evidenced, increasing the CR_PV re- the surface area, absorb solar energy, and concurrently optimize the
sults in a boost in the thermal energy output rate for each of the studied heat transfer surface area by convection between the solid layers and
systems. It is found that at CR_PV = 20, the rate of the produced fluid flowing through the microchannel.

1398
A. Abdo, et al. Energy Conversion and Management 195 (2019) 1382–1401

Fig. 16-a. Variation of average solar cell temperature versus CR_PV for all
Fig. 15-a. The total electrical efficiency variation versus CR_STEG at distinct
studied designs at Re = 50.
values of CR_PV.

Fig. 15-b. Variation of total electrical efficiency versus CR_PV at distinct values Fig. 16-b. Variation of total electrical power versus CR_PV for all studied de-
of CR_STEG. signs at Re = 50.

6. Conclusion m2 being achieved. Integrating the conventional TEG directly allows

the thermal energy from the PV back surface to be transformed into
A new configuration of a PV module and a STEG in a hybrid system electrical energy by the TEG. It significantly increases the solar cell
integrated with a microchannel heat sink has been developed. Both the temperature to about 90 °C, even at a lower CR of 10, with an electrical
PV module and the STEG are exposed to concentrated solar radiation. output power of 1.28 kW/m2. Furthermore, the newly developed
The newly developed system performance is compared with a con- system offers a much higher rate of thermal energy yield when com-
ventional configuration. In the light of this study, the findings can be pared to a conventional one.
stated as follows: the newly developed system can work safely at higher
CRs compared to the conventional one, and a higher rate of perfor-
mance can thus obtain. The average solar cell temperature is main- Conflict of interest
tained at values as low as 77 °C when both the PV module and the STEG
are subjected to a CR of 20, with an electrical power output of 3.2 kW/ There is no conflict of interest.

1399
A. Abdo, et al. Energy Conversion and Management 195 (2019) 1382–1401

[14] Ahmed M, Radwan A. Performance evaluation of new modified low-concentrator

polycrystalline silicon photovoltaic/thermal systems. Energy Convers Manage
2017;149:593–607. [Link]
[15] Radwan A, Ookawara S, Ahmed M. Thermal management of concentrator photo-
voltaic systems using two-phase flow boiling in double-layer microchannel heat
sinks. Appl Energy 2019;241:404–19. [Link]
03.017.
[16] Radwan A, Ookawara S, Mori S, Ahmed M. Uniform cooling for concentrator
photovoltaic cells and electronic chips by forced convective boiling in 3D-printed
monolithic double-layer microchannel heat sink. Energy Convers Manage
2018;166:356–71. [Link]
[17] Li P, Cai L, Zhai P, Tang X, Zhang Q, Niino M. Design of a concentration solar
thermoelectric generator. J Electron Mater 2010;39:1522–30. [Link]
1007/s11664-010-1279-0.
[18] Sundarraj P, Maity D, Roy S, Taylor RA. Recent advances in thermoelectric mate-
rials and solar thermoelectric generators – a critical review. RSC Adv
2014;4:46860–74. [Link]
[19] Yin E, Li Q, Xuan Y. Thermal resistance analysis and optimization of photovoltaic-
thermoelectric hybrid system. Energy Convers Manage 2017;143:188–202. https://
[Link]/10.1016/[Link].2017.04.004.
[20] Wu SY, Zhang YC, Xiao L, Shen ZG. Performance comparison investigation on solar
photovoltaic-thermoelectric generation and solar photovoltaic-thermoelectric
cooling hybrid systems under different conditions. Int J Sustain Energy
2018;37:533–48. [Link]
[21] Cui T, Xuan Y, Yin E, Li Q, Li D. Experimental investigation on potential of a
concentrated photovoltaic-thermoelectric system with phase change materials.
Energy 2017;122:94–102. [Link]
[22] Cui T, Xuan Y, Li Q. Design of a novel concentrating photovoltaic – thermoelectric
system incorporated with phase change materials 2016;112:49–60. doi: 10.1016/j.
enconman.2016.01.008.
Fig. 16-c. Variation of rate of thermal energy versus CR_PV for all studied
[23] Hajji M, Labrim H, Benaissa M, Laazizi A, Ez-Zahraouy H, Ntsoenzok E, et al.
designs at Re = 50. Photovoltaic and thermoelectric indirect coupling for maximum solar energy ex-
ploitation. Energy Convers Manage 2017;136:184–91. [Link]
enconman.2016.12.088.
Acknowledgment [24] Willars-Rodríguez FJ, Chávez-Urbiola EA, Vorobiev P, Vorobiev YV. Investigation
of solar hybrid system with concentrating Fresnel lens, photovoltaic and thermo-
The first author would like to thank the Egyptian Ministry of Higher electric generators. Int J Energy Res 2017;41:377–88. [Link]
3614.
Education as well as the Egypt-Japan University of Science and [25] Mohsenzadeh M, Shafii MB, Jafari mosleh H. A novel concentrating photovoltaic/
Technology, for offering the financial support and computational tools thermal solar system combined with thermoelectric module in an integrated design.
to conduct this research. Renew Energy 2017;113:822–34. [Link]
[26] Sudharshan KY, Kumar VP, Barshilia HC. Performance evaluation of a thermally
concentrated solar thermo-electric generator without optical concentration. Sol
References Energy Mater Sol Cells 2016;157:93–100. [Link]
05.033.
[27] Candadai AA, Kumar VP, Barshilia HC. Performance evaluation of a natural con-
[1] Rahman FA, Aziz MMA, Saidur R, Bakar WAWA, Hainin MR, Putrajaya R, et al.
vective-cooled concentration solar thermoelectric generator coupled with a spec-
Pollution to solution: capture and sequestration of carbon dioxide (CO2) and its
trally selective high temperature absorber coating. Sol Energy Mater Sol Cells
utilization as a renewable energy source for a sustainable future. Renew Sustain
2016;145:333–41. [Link]
Energy Rev 2017;71:112–26. [Link]
[28] King DL, Boyson WE, Kratochvill JA. SANDIA REPORT Photovoltaic Array
[2] Babu C, Ponnambalam P. The role of thermoelectric generators in the hybrid PV/T
Performance Model 2004.
systems: a review. Energy Convers Manage 2017;151:368–85. [Link]
[29] Hedayatizadeh M, Ajabshirchi Y, Sarhaddi F, Safavinejad A, Farahat S, Chaji H.
1016/[Link].2017.08.060.
Thermal and electrical assessment of an integrated solar photovoltaic thermal (PV/
[3] Kalogirou S. Solar energy engineering : processes and systems. Elsevier/Academic
T) water collector equipped with a compound parabolic concentrator (CPC). Int J
Press 2009.
Green Energy 2013;10:494–522. [Link]
[4] Kraemer D, Hu L, Muto A, Chen X, Chen G, Chiesa M. Photovoltaic-thermoelectric
[30] Zhou J, Yi Q, Wang Y, Ye Z. Temperature distribution of photovoltaic module based
hybrid systems: a general optimization methodology. Appl Phys Lett 2008;92:23–5.
on finite element simulation. Sol Energy 2015;111:97–103. [Link]
[Link]
1016/[Link].2014.10.040.
[5] Ju X, Wang Z, Flamant G, Li P, Zhao W. Numerical analysis and optimization of a
[31] Armstrong S, Hurley WG. A thermal model for photovoltaic panels under varying
spectrum splitting concentration photovoltaic-thermoelectric hybrid system. Sol
atmospheric conditions. Appl Therm Eng 2010;30:1488–95. [Link]
Energy 2012;86:1941–54. [Link]
1016/[Link].2010.03.012.
[6] Li Y, Witharana S, Cao H, Lasfargues M, Huang Y, Ding Y. Wide spectrum solar
[32] Yaws CL. Chemical properties handbook : physical, thermodynamic, environmental,
energy harvesting through an integrated photovoltaic and thermoelectric system.
transport, safety, and health related properties for organic and inorganic chemicals/
Particuology 2014;15:39–44. [Link]
Carl L. Yaws. – Version details – Trove. 1999th ed. New York: McGraw-Hill; n.d.
[7] Radwan A, Ookawara S, Ahmed M. Analysis and simulation of concentrating pho-
[33] Shen ZG, Wu SY, Xiao L, Yin G. Theoretical modeling of thermoelectric generator
tovoltaic systems with a microchannel heat sink. Sol Energy 2016;136:35–48.
with particular emphasis on the effect of side surface heat transfer. Energy
[Link]
2016;95:367–79. [Link]
[8] Skoplaki E, Palyvos JA. On the temperature dependence of photovoltaic module
[34] Radwan A, Emam M, Ahmed M. Comparative study of active and passive cooling
electrical performance: a review of efficiency/power correlations. Sol Energy
techniques for concentrated photovoltaic systems. Exergetic Energy Environ
2009;83:614–24. [Link]
Dimens 2018:475–505. [Link]
[9] Emam M, Ookawara S, Ahmed M. Performance study and analysis of an inclined
[35] Radwan A, Ahmed M. The influence of microchannel heat sink configurations on
concentrated photovoltaic-phase change material system. Sol Energy
the performance of low concentrator photovoltaic systems. Appl Energy
2017;150:229–45. [Link]
2017;206:594–611. [Link]
[10] Emam M, Ahmed M. Cooling concentrator photovoltaic systems using various
[36] Radwan A, Ahmed M. Thermal management of concentrator photovoltaic systems
configurations of phase-change material heat sinks. Energy Convers Manage
using microchannel heat sink with nanofluids. Sol Energy 2018;171:229–46.
2018;158:298–314. [Link]
[Link]
[11] Emam M, Ahmed M. Performance analysis of a new concentrator photovoltaic
[37] Lee H, editor. Thermoelectrics: design and materials. Chichester, UK: John Wiley &
system integrated with phase change material and water jacket. Sol Energy
Sons, Ltd; 2016. doi: 10.1002/9781118848944.
2018;173:1158–72. [Link]
[38] Jayakumar JS, Mahajani SM, Mandal JC, Vijayan PK, Bhoi R. Experimental and CFD
[12] Zarma I, Ahmed M, Ookawara S. Enhancing the performance of concentrator
estimation of heat transfer in helically coiled heat exchangers. Chem Eng Res Des
photovoltaic systems using nanoparticle-phase change material heat sinks. Energy
2008;86:221–32. [Link]
Convers Manage 2019;179:229–42. [Link]
[39] Rejeb O, Dhaou H, Jemni A. A numerical investigation of a photovoltaic thermal
10.055.
(PV/T) collector. Renew Energy 2015;77:43–50. [Link]
[13] Rabie R, Emam M, Ookawara S, Ahmed M. Thermal management of concentrator
2014.12.012.
photovoltaic systems using new configurations of phase change material heat sinks.
[40] Churchill SW, Chu HHS. Correlating Equations for Laminar and Turbulent Free
Sol Energy 2019;183:632–52. [Link]
Convection from a Vertical Plate vol. 18. Pergamon Press; 1975.

1400
A. Abdo, et al. Energy Conversion and Management 195 (2019) 1382–1401

[41] Dehghan M, Daneshipour M, Valipour MS, Rafee R, Saedodin S. Enhancing heat [44] Slimani MEA, Amirat M, Kurucz I, Bahria S, Hamidat A, Chaouch WB. A detailed
transfer in microchannel heat sinks using converging flow passages. Energy Convers thermal-electrical model of three photovoltaic/thermal (PV/T) hybrid air collectors
Manage 2015;92:244–50. [Link] and photovoltaic (PV) module: comparative study under Algiers climatic condi-
[42] Brito FP, Figueiredo L, Rocha LA, Cruz AP. Analysis of the Effect of Module tions. Energy Convers Manage 2017;133:458–76. [Link]
Thickness Reduction on Thermoelectric Generator Output 2016;45:1711–29. doi: enconman.2016.10.066.
10.1007/s11664-015-4182-x. [45] Mahmoudinezhad S, Rezania A, Cotfas DT, Cotfas PA, Rosendahl LA. Experimental
[43] Joshi AS, Tiwari A, Tiwari GN, Dincer I, Reddy BV. Performance evaluation of a and numerical investigation of hybrid concentrated photovoltaic – thermoelectric
hybrid photovoltaic thermal (PV/T) (glass-to-glass) system. Int J Therm Sci module under low solar concentration. Energy 2018;159:1123–31. [Link]
2009;48:154–64. [Link] 10.1016/[Link].2018.06.181.

1401