Energy Conversion and Management 143 (2017) 188–202

Contents lists available at ScienceDirect

Energy Conversion and Management

journal homepage: www.elsevier.com/locate/enconman

Thermal resistance analysis and optimization of photovoltaic-

thermoelectric hybrid system
Ershuai Yin a, Qiang Li a,⇑, Yimin Xuan a,b
a
MIIT Key Laboratory of Thermal Control of Electronic Equipment, School of Energy and Power Engineering, Nanjing University of Science & Technology, Nanjing 210094, China
b
School of Energy and Power, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China

a r t i c l e i n f o a b s t r a c t

Article history: The thermal resistance theory is introduced into the theoretical model of the photovoltaic-thermoelectric
Received 27 December 2016 (PV-TE) hybrid system. A detailed thermal resistance analysis is proposed to optimize the design of the
Received in revised form 27 February 2017 coupled system in terms of optimal total conversion efficiency. Systems using four types of photovoltaic
Accepted 1 April 2017
cells are investigated, including monocrystalline silicon photovoltaic cell, polycrystalline silicon photo-
Available online 8 April 2017
voltaic cell, amorphous silicon photovoltaic cell and polymer photovoltaic cell. Three cooling methods,
including natural cooling, forced air cooling and water cooling, are compared, which demonstrates a sig-
Keywords:
nificant superiority of water cooling for the concentrating photovoltaic-thermoelectric hybrid system.
Photovoltaic-thermoelectric hybrid system
Thermal resistance analysis
Influences of the optical concentrating ratio and velocity of water are studied together and the optimal
Solar energy values are revealed. The impacts of the thermal resistances of the contact surface, TE generator and
Optimization the upper heat loss thermal resistance on the property of the coupled system are investigated, respec-
Renewable energy tively. The results indicate that amorphous silicon PV cell and polymer PV cell are more appropriate
for the concentrating hybrid system. Enlarging the thermal resistance of the thermoelectric generator
can significantly increase the performance of the coupled system using amorphous silicon PV cell or poly-
mer PV cell.
Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction
Photovoltaic (PV) technology is one of the highly competitive
technologies to convert the solar energy into electric energy. How-
ever, the reported commercial PV modules have very low conver-
sion efficiency within the range of 12–18% [1] and decreasing as
the increase of temperature. A significant part of solar energy con-
verts into heat and is wasted [2]. Therefore, improving the photo-
electric conversion efficiency and combining PV cell with heat
recovery-utilization system, such as thermoelectric (TE) module,
are in demand for further utilization of solar energy. Thermoelec-
tric module (TE) can generate electricity based on the Seebeck
effect when the hot side and the cold side exist a temperature dif-
ference [3–9]. Thus, combining PV cell and TE devices can not only
increase the output power but also utilize the unwanted heat that
may weaken the performance of the PV cell.
One of combining approaches is directly attaching the TE
generator to the back of the PV cell. Numerous theoretical and
experimental researches in the hybrid system have been carried
out recently. Most of the researching results pointed that this
⇑ Corresponding author.
E-mail address: liqiang@njust.edu.cn (Q. Li).
combining method can improve the utilization of the solar spec-
trum [10–18]. Zhang et al. [11] studied the concentrating PV-TE
coupled system with a forced air cooling method. The coupled sys-
tem achieved an increase in efficiency of 1–30% comparing to the
pure photovoltaic system. Park et al. [13] provided a novel hybrid
approach named lossless coupling which matched the resistance of
TE modules with PV circuits. The coupled system achieved great
improvement in the efficiency of PV cell (
30% when the TE gener-
ator maintained a 15 °C temperature difference). Zhu et al. [14]
considered an optimal thermal management for the coupled sys-
tem. The new coupled system achieved a significant increase
(
25%) in conversion efficiency compared to the PV cells. Wang
et al. [15] designed a PV-TE system combining dye sensitized PV
device (DSSC), solar selective absorber (SSA) and TE generator
which demonstrated an obvious increasing on the total conversion
efficiency of the hybrid systems comparing to the pure PV system.
The overall efficiency was 13.8% for the DSSC-SSA-TE hybrid
device, 12.8% for the DSSC-TE coupled device whereas the effi-
ciency was 9.26% for the pure PV system.
However, there were also some different results that pointed
out the capability of the coupled system was worse than the pure
photovoltaic system [19,20]. BjØrk et al. [19] theoretically
investigated the hybrid systems with four types of PV cells and a

http://dx.doi.org/10.1016/j.enconman.2017.04.004
0196-8904/Ó 2017 Elsevier Ltd. All rights reserved.
 E. Yin et al. / Energy Conversion and Management 143 (2017) 188–202 189

Nomenclature

A area (m2) B Auger recombination coefficients (cm6/s)

R thermal resistance (K/W) IAM1.5(k) solar irradiance of AM1.5
kB Stefan–Boltzmann’sconstant (1.38  1023 J/K) c speed of light (3  108 m/s)
T temperature (K)
Tin inlet temperature of water (K) Greek symbols
Tout outlet temperature of water (K) e surface emissivity
Ta temperature of environment (K) a absorptivity
u velocity of air or water (m/s) a(k) absorption coefficient (m1)
k thermal conductivity (W m1 K1) s transmissivity
M number of thermoelements of TE module d thickness (m)
h convection heat transfer coefficient (W m2 K1) q density of water (kg/m3)
n number of fins or channels k wavelength of photon (lm)
Af surface area of a fin (m2) x dielectric constant (Fm3)
At total convection surface area of fin (m2) u electrical potential (V)
Wfin thickness of the rectangular fins (m) / electron and hole lifetimes (s)
Hc,fin corrected fin height (m) dfc thickness of the front cover of water cooling system (m)
D diameter of the pin fins or effective diameter of the c structure parameter of the TE legs (m)
water cooling channels (m) g0,fin overall surface efficiency of fin
L length of fin (m) gf,fin efficiency of a fin
H height of fin (m) r electrical resistance of the p/n semiconductors (X m)
m fin constant g total efficiency of PV-TE hybrid system
Nu Nusselt number l carrier mobility (cm2 V1 s1)
Re Reynolds number gPV photo-electric conversion efficiency of PV cell
Pr Prandtl number
Gz Graetz number Subscripts and superscripts
v viscosity of air or water (m2/s)
rad radiation heat loss
V flow rate of air (m3/s) PV photovoltaic cell
a height of channels (m) g glass cover
b width of channels (m)
conv convection heat loss
l length of channels (m) wind natural wind
P output power (W) EVA optical adhesive EVA
Q heat flux (W) ceram ceramic wafers of TE module
Qin all the solar energy getting into PV-TE hybrid system
Cu Cu material and Cu electrodes of TE module
(W) c connecting interface
s Seebeck coefficient of p-type semiconductor legs or n- ad thermal conductive adhesive
type semiconductor legs (V/K) TE TE module
sm Seebeck coefficient of TE module (V/K)
n n-type semiconductor legs of TE module
I output current of TE module (A) p n-type semiconductor legs of TE module
r loading resistance of TE module (X) cool cooling system
rTE internal resistance of TE module (X) b fin base
C optical concentrating ratio
fin pin fin and rectangle fin
G solar irradiance (W/m2) air air
f friction factor w water cooling system and water
Cp thermal capacity of water (J/kg K)
ch channels of water cooling system
q electron charge (1.6  1019 C) loss heat loss from the top side the glass cover
DP pressure drop of air (Pa) down all the components below the PV cell
e electron concentration (cm3) h hot side of TE module
t hole concentration (cm3)
c cold side of TE module
NA acceptor doping concentration (cm3) e electron
ND donor doping concentration (cm3) t hole
NC the effective density of electrons (cm3)
NV the effective density of holes (cm3)
J electron and hole current densities (A/cm2) Abbreviations
PV photovoltaic cell
Gr generation rate (cm3 s1)
Z recombination rate (cm3 s1) TE thermoelectric module
RSRH Shockley-Read-Hall recombination rate (cm3 s1) a-Si amorphous silicon
RAug Auger recombination rate (cm3 s1) c-Si monocrystalline silicon
p-Si polycrystalline silicon
Eg band-gap energy (eV)
ni intrinsic carrier concentration (cm3)

universal bismuth telluride TE module. For c-Si, CIGS and CdTe PV got a tiny increase only when the a-Si PV cell was employed.
cells, the hybrid system achieved a worse performance compares Vorobiev et al. [20] investigated the hybrid system with
to the pure PV system. The output power of the coupled system concentrator of solar radiation. For existing thermoelectric
190 E. Yin et al. / Energy Conversion and Management 143 (2017) 188–202
materials, the combined system had a worse performance than the
pure PV system.
Many reasons result in these different conclusions. One of the
most important reasons is that the calculating model of the cou-
pled system was simplified in some ways. Most of the previous
theoretical researches ignored the thermal resistances of different
components, which have great influence on the operating temper-
atures and the performance of PV cell and TE module. Only several
researchers have investigated part of the thermal resistances in the
hybrid system [11,21–23]. Zhang et al. [11] studied the impact of
thermal contact resistance (TCR). When a small TCR was set, the
hybrid system got a higher total efficiency comparing with the
pure PV system whereas the opposite result occurred when the
TCR is large. Wu et al. [21] discussed the capability of the coupled
systems with and without glass cover. They pointed that, when
maintaining a higher figure of merit and concentrating ratio,
enlarging the transmissivity of cover glass and glazed system
may be better than the unglazed coupled system. Covering the
hybrid system with a glass means increasing the heat loss thermal
resistance above the PV cell. The concept of the heat loss resistance
has not yet been proposed in their work, and the influence of the
heat loss resistance was without doubt not discussed in detail.
Several researchers studied the influence of the cooling system
[21–23]. They pointed that a good cooling system can increase
the total coupled system output power. Nevertheless, only one
kind of cooling method is investigated ignoring the possible trans-
formations caused by different cooling methods. Although these
researchers had considered the part of the thermal resistances
and studied their effects, the thermal resistances inside the PV-
TE hybrid system have not been comprehensively researched.
The investigation and optimization of hybrid system through the
thermal resistance analysis have not been presented before.
In this paper, a thermal resistance theory has been introduced
into the theoretical model of the system coupling PV cell and the
TE module. A detailed thermal resistance analysis is proposed to
optimize the design of PV-TE hybrid system in terms of optimal
total efficiency. Four types of PV cells are explored to find the bet-
ter one for coupling, including c-Si, p-Si, a-Si and polymer photo-
voltaic cells. Firstly, three cooling methods are compared for a
most appropriate one. The influences of concentrating ratio, veloc-
ity of water are investigated together because of the strong inter-
action, and the optimal values are revealed in the case of
different PV cells. Then, based on the optimal working states
obtained earlier, the impacts of thermal contact resistance, thermal
resistance of TE device and upper heat loss thermal resistance on
the performance of PV-TE hybrid system are explored, respectively.
Some criteria for selecting coupling devices and optimal design are
obtained from the thermal resistance analysis. Finally, a route and
some principles of designing the practical PV-TE hybrid system are
advised in the Conclusion.
2. The theoretical model
2.1. The geometry structure
As is illustrated in Fig. 1(a), the PV-TE hybrid system includes an
optical concentrating system, a glass cover, a PV cell, a TE module
and a cooling system. The glass cover is connected to the top of the
PV cell by EVA in order to protect the cell and prevent the upper
heat loss. The TE device is attached to the back of the PV cell by
thermal conductive adhesive. The TE module consists of two cera-
mic wafers, Cu electrodes and p/n semiconductor legs. A cooling
system is added to the cold side of the TE generator to remove
the thermal energy, which cannot be transformed by the TE gener-
ator. Fig. 1(b) shows three kinds of cooling methods that are inves-
tigated in this paper. For hybrid system, the concentrating sunlight
impinges on the PV cell through the glass cover. The PV cell con-
verts some of the solar energy into electricity and the remaining
part is transformed to thermal energy. Some of the thermal energy
are lost by conduction to the glass cover, and then convects and
radiates to the environment. Most of the thermal energy delivers
to the TE generator. Then, the TE module absorbs a small part of
the heat based on the Seebeck effect, and the remaining thermal
energy will be taken by the cooling system.
In this paper, the following assumptions were made:
(1) Heat loss only appears at the top of the front glass cover.
(2) The heat transfer is steady-state and one-dimensional.
(3) As solar cells are very thin, the temperature gradients inside
the solar cells are neglected in the theoretical model.
(4) The physical property of TE legs is independent of
temperature.
All the parameters of the devices in the hybrid system besides
the properties of the cooling systems can be found in Table 1.
2.2. Thermal resistance
Fig. 2 gives the thermal resistance network of the hybrid sys-
tem. Each device in the hybrid system contains an internal thermal
resistance. The connecting surfaces of PV cell, TE module and cool-
ing system have thermal contact resistances. The upper heat loss
from the glass cover has convection thermal resistance and radia-
tion thermal resistance.
The radiation thermal resistance Rrad representing the ability of
the radiation heat loss from the front glass cover is given as:
1
Rrad ¼ ð1Þ
hrad APV
where APV is the area of the PV cell, hrad is the radiation heat transfer
coefficient. hrad is given as [24]:
hrad ¼ eg kB ðT 2g þ T 2a ÞðT g þ T a Þ ð2Þ
where eg is the emissivity of the front glass cover, kB is the Stefan–
Boltzmann’s constant, Ta is the temperatures of the environment, Tg
is the temperature of the top of the glass cover.
Rconv is the convection thermal resistance which prevents the
convection heat loss from the upper side of the hybrid system,
and Rconv can be found at:
1
Rconv ¼ ð3Þ
hconv APV
hconv is the convection heat transfer coefficient [25]:
hconv ¼ 2:8 þ 3:0  uwind ð4Þ
where uwind is the velocity of natural wind.
Rg, REVA, Rceram and RCu are respectively the thermal resistance of
glass cover, EVA, ceramic wafer and Cu electrode. All of them can
be found at following equation [26]:
di
Ri ¼ ð5Þ
k i Ai
where d is thickness, k is thermal conductivity. The area of glass
cover, EVA is set to be equal to the PV cell.
Rc is the thermal contact resistance. In the coupled system, PV
cell and TE module as well as TE module and cooling system are
connected with thermal conductive adhesive. Mirmira et al. [27]
showed that the temperature or the pressure of the contact
interface had few impacts on the thermal contact resistance of
the connecting surface with adhesive, whereas adhesive thickness
 E. Yin et al. / Energy Conversion and Management 143 (2017) 188–202 191

Fig. 1. (a) Schematic diagrams of PV–TE hybrid system; (b) three kinds of cooling methods investigated in this paper.

and thermal conductivity were crucial. Therefore, Rc is expressed Rcool is the cooling system thermal resistance. The materials of
as: three cooling systems are all copper. There are separate equations
calculating the thermal resistance of three cooling systems. For
dad
Rc ¼ ð6Þ natural cooling and forced air cooling methods, the thermal resis-
kad Aad tance of the cooling system is defined as:
where dad, kad and Aad are the thickness, the thermal conductivity Rcool ¼ Rb þ Rfin ð9Þ
and the area of the thermal conductive adhesive.
All the semiconductor legs in the TE module are in parallel in where Rb = db/kCuAb is the thermal resistance of the fin base, db and
the direction of heat flow. The total thermal resistance of the semi- Ab are the thickness and the area of fin base respectively, kb is the
conductor legs is defined as the thermal resistance of TE module thermal conductivity of copper. Ab is considered to be equal to
RTE, which is given as [28]: ATE. The thermal resistance of the fins Rfin can be found at [29]:
1 1
RTE ¼ ð7Þ Rfin ¼ ð10Þ
M cðkp þ kn Þ g0;fin hair At
where M is the number of thermocouples, kp is the thermal conduc- where g0,fin is the overall surface efficiency and can be calculated by
tivity of p-type semiconductor legs, kn is the thermal conductivity of [26]:
n-type semiconductor legs, c is the structure parameter of the TE
legs: nfin Af
g0;fin ¼ 1  ð1  gf ;fin Þ ð11Þ
At
A
c ¼ p=n ð8Þ where nfin, Af, At = nfinAf + Ab are respectively the number of fins, the
Lp=n
surface area of a fin and the total convection surface area, gf,fin is the
where Ap/n is the cross-sectional, Lp/n is the length. The structure efficiency of a fin. For pin fins and rectangular straight fins, gf,fin can
parameter c of all TE legs are deemed to be equal in this paper. be calculated by:
192 E. Yin et al. / Energy Conversion and Management 143 (2017) 188–202

Table 1 tanhðmHc;fin Þ
Parameters of PV-TE hybrid system. gf ;fin ¼ ð12Þ
mHc;fin
Parameters Symbol Unit Value
where Hc,fin is the corrected fin height, it is given as
Thickness of glass cover dg mm 3
Hc,fin = Hfin + (Wfin/2) for the rectangular fins and Hc,fin = Hfin + (Dfin/4)
Transmissivity of glass cover sg – 0.95
Absorptivity of glass cover ag – 0.018
for the pin fins. Hfin is the height of the fins, bfin is the width of the
Emissivity of glass cover eg – 0.85 rectangular fins and Dfin is the diameter of the pin fins. m is the
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Thermal conductivity of glass cover kg W m1 K1 1.8 fin constant which can be calculated by m ¼ 2hair =kcu W fin for
Velocity of natural wind uwind m/s 2 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Area of PV cell APV mm2 30  30 the rectangular fins and m ¼ 4hair =kcu Dfin for the pin fins. hair is
Thickness of EVA dEVA mm 0.5 the convection heat transfer coefficient, and hair = 8 W/m2 K for
Thermal conductivity of EVA kEVA W m1 K1 0.35 natural cooling. hair can be calculated by the following equations
Thickness of thermal conductive dad mm 0.5 when the forced air cooling is employed [30]:
adhesive
Thermal conductivity of thermal kad W m1 K1 1.5 kair Nu
conductive adhesive hair ¼ ð13Þ
Lfin
Area of TE module (ceramic wafer) ATE mm2 30  30
Thickness of ceramic wafer dceram mm 0.8
Thermal conductivity of ceramic wafer kceram W m1 K1 31 Nu ¼ 0:664 Re1=2 Pr1=3 ð14Þ
Area of Cu electrode ACu mm2 3 N  Ap/n
Thickness of Cu electrode dCu mm 0.3
uair Lfin
Thermal conductivity of Cu kCu W m1 K1 398 Re ¼ ð15Þ
Number of thermoelements M – 127 v air
1.83  104
where Lfin is the length of the fins, uair is the velocity of air, vair is the
Seebeck coefficient of p-type or n-type sp/n V/K
semiconductor legs
Electrical resistivity of p-type or n-type rp/n Xm 7.23  106 viscosity of air.
semiconductor legs For water cooling system, Rcool is defined as:
Thermal conductivity of p-type or n- kp/n W m1 K1 1.82
type semiconductor legs dfc
Rcool ¼ þ Rw ð16Þ
Length of p-type or n-type Lp/n mm 1.6 kcu Aw
semiconductor legs
Cross-sectional area of p-type or n- Ap/n mm2 0.87 where dfc and Aw are the thickness and the area of the front cover,
type semiconductor legs Rw is the convective thermal resistance of the water cooling system.
Structure parameter of the TE legs c mm Ap/n Lp/n1 Rw is given as:
Temperature of environment Ta K 300
1
Rw ¼ ð17Þ
2hw ðach þ bch Þnch lch
where ach, bch, lch, nch is the height, the width, the length and the
number of channels of the water cooling system. As the channels
are very short, the flow is in the entrance region. Therefore, the con-
vection heat transfer coefficient hw can be calculated by [26,31]:
kw Nuw
hw ¼ ð18Þ
lch

Dch
Gz ¼ Rew Prw ð19Þ
lch

0:0668Gz
Nuw ¼ 3:66 þ ðRe < 2300Þ ð20Þ
1 þ 0:04Gz2=3
"  0:7 #
Dch
Nuw ¼ 0:023Re0:8
w Pr
0:4
1þ ðRe > 2300Þ ð21Þ
lch

where kw is the thermal conductivity of water, Gz is the Graetz

Fig. 2. The one-dimensional thermal resistance network of the hybrid system.
number, Dch = 2Ach/(ach + bch,) is the effective diameter of the
channels, Ach is the flow cross-sectional area of the channels,
Rew = uwDch/vw is the Reynolds number, uw and vw are respectively
the velocity and the viscosity of water. Prw is the Prandtl number
of water. The properties of three cooling methods investigated in
this paper are detailed in Table 2.
The heat absorbed by the PV cell has two different flow direc-
tions. A little part of the thermal energy conducts to the top of
the glass cover and then loses in the environment. The most part
of heat transfers to TE module and then the cooling system. In
order to investigate the influence of heat loss on the hybrid system,
the total thermal resistance above the PV cell is defined as the ther-
mal resistance of heat loss
1
Rloss ¼ Rg þ REVA þ ð22Þ
1=Rconv þ 1=Rrad
 E. Yin et al. / Energy Conversion and Management 143 (2017) 188–202 193

Table 2
Parameters of cooling systems.

Cooling system Parameters Symbol Unit Value

2
Area of fin base Ab mm 30  30
Thickness of fin base db mm 3
Free cooling Number of pin fins nfin – 100
Diameter of pin fins Dfin mm 1.5
Height of pin fins Hfin mm 15
Convection heat transfer coefficient hair W m2 K1 8
Forced air cooling Number of rectangular fins nfin – 5
Length of rectangular fins Lfin mm 30
Width of rectangular fins Wfin mm 2
Fin-to-fin spacing mm 4
Height of rectangular fins Hfin mm 60
Thermal conductivity of air kair W m1 K1 26.3  103
Viscosity of air vair m2/s 15.89  106
Water cooling Area of water cooling system Aw mm2 30  30
Height of water cooling system mm 10
Thickness of the front cove dfc mm 1
Number of channels nch mm 6
Length of channels lch mm 30
Height of channels ach mm 8
Width of channels bch mm 4
Distances between two channels mm 1
Thermal conductivity of water kw W m1 K1 0.599
Viscosity of water vw m2/s 1.006  106
Density of water qw kg/m3 998.2
Thermal capacity of water Cp J kg1 K1 4183
Prandtl number of water Prw – 7.02
Inlet temperature of water Tin K 293

The thermal resistances of the components under the PV cell rðxr/Þ ¼ qðt  e þ ND  NA Þ ð26Þ
can be considered to be in series in the direction of heat flow,
the total thermal resistance can be calculated as rJe ¼ qðGr  ZÞ ð27Þ

Rdown ¼ 2Rc þ 2Rceram þ 2RCu þ RTE þ Rcool ð23Þ rJt ¼ qðGr  ZÞ ð28Þ
When the natural cooling or forced air cooling system is
employed in the PV-TE hybrid system, the two different directions J e ¼ qle er/ þ kB T le re ð29Þ
of heat absorbed by the PV cell have a same destination. Therefore,
the thermal resistance of heat loss and the total resistance under J t ¼ qlp t r/  kB T lt rt ð30Þ
the PV cell are thought to be in parallel. The total resistance of
where x is the dielectric constant, u is the electrical potential, q is
the hybrid system is given as
the electron charge, e and t are the electron and hole concentration,
1 1 1 NA and ND are the donor and acceptor doping concentration, Je and Jt
¼ þ ð24Þ are the electron and hole current density, le is the electron mobility
R Rloss Rdown
and lt is the hole mobility, Z is the recombination rate. The recom-
Consequently, analogizing to shunt principle of parallel circuit, bination rate includes Shockley-Read-Hall recombination rate
the lost heat Qloss and the heat transferring to the TE module Qh has (RSRH) and Auger recombination rate (RAug) and RSRH can be calcu-
a following relationship: lated by [32]:
Qh Rloss Rloss et  ni
2
¼ ¼ ð25Þ RSRH ¼ ð31Þ
Q loss Rdown 2Rc þ 2Rceram þ 2RCu þ RTE þ Rcool /t ðe þ niÞ þ /e ðt þ niÞ
This means that lessening the thermal resistance of any of the
pffiffiffiffiffiffiffiffiffiffiffiffi Eg
components under the PV module can increase the heat transfer- ni ¼ Nc NV e 2kT ð32Þ
ring to the TE module (this may prompt the TE generator to obtain
larger temperature difference and maybe more power output) and where /e and /t are the electron and hole lifetimes, ni is the intrin-
reduce the heat loss because the total heat absorbed by the PV cell sic carrier concentration, NC and NV are the effective density of
is constant. Although the above analysis of the total resistance states for electrons and holes. RAug is given as:
ignores the part of heat translated by the TE module, the result is RAug ¼ ðBe e þ Bt tÞðet  ni Þ
2
ð33Þ
credible and useful because the energy conversion does not change
the direction of the heat flux. When a water cooling system is where Be is the electron Auger recombination coefficient, Bt is the
hired, the relationship in Eq. (25) is applicative because only the hole Auger recombination coefficient.
thermal resistance of the heat transfer from the cooling water Gr is the generation rate which can be calculated by [33]:
and the environment is ignored. Z hc=Eg
Grðx; tÞ ¼ CIAM1:5 ðkÞaðkÞeaðkÞx dk ð34Þ
kmin
2.3. PV cell model
where IAM1.5(k) is the solar irradiance spectrum of AM1.5, k is the
The semiconductor equations are introduced to simulate the wavelength, a(k) is the absorption coefficient, c is the velocity of
photoelectric effect of silicon PV cells [23,32]. light. In this paper, the material, the structure of the PV cell are
194 E. Yin et al. / Energy Conversion and Management 143 (2017) 188–202

determined first. Then, different intensities of sunlight and operat- T h  T c I2TE rTE
ing temperatures of cell are given. The current-voltage curve under Q h ¼ sm T h ITE þ  ð38Þ
RTE 2
various conditions will be calculated. Finally, the conversion effi-
ciency can be obtained from the current-voltage curve. The struc- T h  T c I2TE rTE
ture and material parameters of the silicon PV cells are referenced Q c ¼ sm T c ITE þ þ ð39Þ
RTE 2
from literature [11]. The efficiency of polymer PV cell has a different
calculation method. The model, structure and material parameters
of polymer PV cell are from literature [34]. 2.5. The efficiency of hybrid system
The impact of temperature on the capability of the PV cell is
shown in Fig. 3(a) when C = 1. The result reveals that temperature For the coupled system using natural cooling, the total conver-
has a negative effect on the performance of all three silicon photo- sion efficiency g is defined as:
voltaic cells. The polymer PV cell appears an obvious different with PPV þ PTE
the silicon PV cells. This means the polymer maybe have a superi- g¼ ð40Þ
Q in
ority for the coupled system. The efficiency of the polymer PV cell
first increases and then decreases when the temperature rises. The where Qin is the total solar energy absorbed by the hybrid system.
efficiency of c-Si PV cell is the highest among the three Si PV cells, Qin is given as:
but c-Si PV cell has a larger temperature coefficient of conversion Q in ¼ CGAPV ðsg þ ag Þ ð41Þ
efficiency than p-Si and a-Si PV cells. Fig. 3(b) shows the influence
of optical concentrating ratio when Tpv = 300 K. The conversion where C is the concentrating ratio, G is the solar irradiance which is
efficiency of three types of Si PV cells increases when a larger value set as 1000 W/m2, sg and ag are the transmissivity and the absorp-
of optical concentrating ratio is set, whereas the polymer PV cell tivity of the glass cover.
has an opposite performance. The hybrid system using forced air cooling has a fan which con-
sumes little electricity. The total conversion efficiency can be cal-
2.4. TE module model culated as:
PPV þ PTE  PFan
The output power of the TE device is given as [35]: g¼ ð42Þ
Q in
PTE ¼ sm ITE ðT h  T c Þ  I2TE r TE ð35Þ where the power of the fan PFan is calculated as:
where sm = M (sp  sn) is the Seebeck coefficient, p is the p-type legs PFan ¼ V air DP air ð43Þ
and n is the n-type legs. ITE is the current, h is the hot side and c is
where Vair and DPair are the flow rate and the pressure drop of air.
the cold side of the TE generator, rTE is the internal resistance of the
They can be calculated by [37]:
TE device which can be expressed as [28]:
V air ¼ nfin Dfin Hfin uair ð44Þ
Mðrp þ rn Þ
r TE ¼ ð36Þ
c 12mair Lfin uair
DPair ¼ ð45Þ
where r is the electrical resistivity. The loading resistance r is set as D2fin
equal to rTE. The output current of the TE module can be calculated
by [36]: For the hybrid system using water cooling, the total conversion
efficiency g is defined as:
sm ðT h  T c Þ
ITE ¼ ð37Þ PPV þ PTE  PPump
r TE þ r g¼ ð46Þ
Q in
For the TE generator, the heat flowing from the PV cell to the hot
side (Qh) and the heat transferring from the cold side to the cooling where Ppump is the power consumed of the water cooling system,
system (Qc) are given as [28]: and it can be calculated by [23]:

Fig. 3. (a) The impact of temperature on the efficiency of the PV cells when C = 1; (b) the influence of optical concentrating ratio when TPV = 300 K.
 E. Yin et al. / Energy Conversion and Management 143 (2017) 188–202 195

l u2
PPump ¼ f   w  qw  Ach  uw  nch ð47Þ
Dch 2
where the friction factor f can be calculated by [24,26]:
64
f ¼ ðReW < 2300Þ ð48Þ
Rew

2
f ¼ ð0:79 ln Rew  1:64Þ ðRew > 2300Þ ð49Þ

2.6. The energy balance equations

Based on the structure and thermal resistance network of the

coupled system, the energy balance equations of front glass cover,
PV cell, TE module and cooling system are considered in this paper.
According to the assumption, the heat loss only appears at the
upper surface of the glass cover. The energy balance equation of
the front glass cover considering the convective and radiative heat
loss is given as:
T pv  T g T g  T a T g  T a
ag CGApv þ   ¼0
Rev a þ Rg Rconv Rrad
The energy balance of the photovoltaic cell and hot side of ther-
moelectric module can be expressed as:
T PV  T g T PV  T h
sg ð1  gPV ÞCGAPV   ¼0
Rev a þ Rg Rc þ Rceram þ RCu
T PV  T h
Qh  ¼0
Rc þ Rceram þ RCu
For the hybrid system with a natural cooling or a forced air cool-
ing system, the energy balance is given as:
Tc  Ta
Qc  ¼0
RCu þ Rceram þ Rc þ Rcool
When a water cooling system is employed, the energy balance
of the cold side of the TE generator and the water can be expressed
as:
Tc  Tm
Qc ¼ ¼ nch uw Ach qw C p ðT out  T in Þ
RCu þ Rceram þ Rc þ Rcool
where Tin, Tout and Tm = (Tin + Tout)/2 are the inlet, outlet and average
temperatures of water, Cp is the thermal capacity of water.
3. Results and discussion
3.1. Comparison of three cooling methods
For the PV-TE hybrid system, the changes of the cooling meth-
ods mean different values of Rcool. For forced air cooling and water
cooling, different velocities of the fluid are taken into account in
this paper. Natural convection coefficient hair for natural cooling
is set to 8 W/m2 K. The velocity of the air uair for forced air cooling
is equal to 2 m/s. And the velocity of the water uw for water cooling
system is equal to 0.1 m/s. Fig. 4 shows the proportion of different
thermal resistances accounting for the total thermal resistance
below the PV cell, when three different cooling methods are used.
Since the upward heat flow and the downward heat flow are in the
opposite direction, only the thermal resistances below the PV cell
are analyzed here. The thermal resistances above the PV cell are
considered in the latter section. For the hybrid system with a nat-
ural cooling system, Rcool which accounts for 76% of the total is the
main thermal resistance. It means that lessening the cooling sys-
tem thermal resistance is the most efficient way to enhance the
heat flux to the TE generator and improve the total output power
ð50Þ
Fig. 4. The proportion of different thermal resistances accounting for the total
thermal resistance below the PV cell.
ð51Þ according to Eq. (25). For the hybrid system using forced air cooling
and water cooling, the total thermal resistance lessens because the
values of Rcool decreases. RTE becomes the main thermal resistance
ð52Þ and Rc is second to RTE. Therefore, the impacts of Rcool, RTE and Rc
should be studied in detailed. Nevertheless, Rceram and RCu are
always so small (less than 1.5%) that it is not necessary to consider
their effects.
Fig. 5 presents the influence of cooling condition on the temper-
ð53Þ atures (a c-Si PV cell is taken as an example) when the optical con-
centration ratio rises. The result shows that, comparing with the
natural cooling and forced air cooling, water cooling is more suit-
able for the coupled system. Because a lower temperature of the
PV module and a larger temperature difference of TE generator
are obtained when the water cooling is employed, especially for
ð54Þ large value of optical concentrating ratio. When the thermal resis-
tance of the cooling system decreases, the temperature difference
of the TE module increases is because the total thermal resistance
under the PV cell Rdown is lessened which result an increase of the
downward heat flux according to the Eq. (25). Natural cooling
should not be used in the hybrid system with optical concentrator
owing to its bad cooling performance (Tpv reached 455 K when C is
set as 13). Fig. 6 indicates the influence of the cooling condition on
the total conversion efficiency. For the coupled system using a Si
photovoltaic cell, the total efficiency with a water cooling method
is obviously larger than systems using natural cooling and forced
air cooling. Moreover, along with the rise of the optical concentrat-
ing ratio, the water cooling performs more significant superiority.
The coupled system with the polymer PV cell has a different result.
When the value of optical concentrating ratio is small (C < 5), the
hybrid system using a natural cooling method has the largest total
efficiency. Because, under the present condition, the polymer PV
cell has gradually increased performance along with the rising
temperature. When the optical concentrating ratio continues
increasing, the water cooling system has become the most applica-
ble. Therefore, a water cooling system and a large value of optical
concentration ratio should be employed for the system coupling a
polymer PV cell and a TE generator. All of the results reveal that
water cooling is the most suitable cooling method for the coupled
system, especially concentrating the sunlight. Consequently, in the
following sections, water cooling is chosen in the models for
calculation.
196 E. Yin et al. / Energy Conversion and Management 143 (2017) 188–202
Fig. 5. The temperature of the PV-TE hybrid system using a c-Si PV cell (a) The temperature of the PV cell (b) the temperatures of the TE module (c) the temperature
difference of the TE module.
3.2. The influence of optical concentrating ratio and velocity of water
Large optical concentrating ratio requires good cooling condi-
tion. The influences of concentrating ratio, velocity of water are
investigated together because of their strong interaction. Fig. 7
illustrates the thermal resistance of the water cooling system and
the power of the pump as a function of the velocity of water. The
thermal resistance first dramatically decreases and then slightly
declines with the rise of water velocity, whereas the power of
pump obviously increases owing to the increasing water velocity.
One can see that, there is a sudden jump in the water velocity of
0.43 m/s. This is because the flow pattern of cooling water changes
from laminar to turbulent. Comparing with the thermal resistance
of cooling system Rcool, uw is a superior parameter (clear and adjus-
table) to investigate the influence of the cooling system. Since a
larger water velocity means better performance and greater water
pump consumption, an optimal uw needs to be obtained.
Fig. 8 illustrates the impacts of the optical concentrating ratio
and the velocity of water on the total efficiency. When the optical
concentrating ratio rises, the total efficiency first rises and then
decreases for the hybrid systems with c-Si and p-Si PV cells. The
reason for the increase is that the performance of silicon PV cell
improves when the optical concentrating ratio increases. In addi-
tion, concentrating sunlight leads to a larger temperature differ-
ence, which can enhance the thermoelectric effect of the TE
generator. However, when the concentrating ratio continues to
increase, the high temperature caused by concentrating sunlight
weakens the performance of PV cell which cannot be offset by
the increased thermoelectric effect. Increasing the velocity of the
cooling water can improve the performance of the PV cell and
the TE module but also needs more pump power which is sub-
tracted from the total output power. Therefore, there must exist
an optimal velocity. The hybrid system with c-Si PV cell reaches
optimal working state when concentrating ratio is approximate 6
and the velocity of the water is around 0.23 m/s. The optimal work-
ing condition for p-Si PV cell coupled system is C = 13 and
uw = 0.35 m/s. The total efficiency of the coupled system using a-
Si PV cell increases with the value of the concentrating ratio
increasing. The reasons are that the a-Si PV cell has a lower photo-
electric conversion efficiency on one hand and a smaller tempera-
ture coefficient of efficiency on the other hand comparing with the
other two silicon PV cells. When the concentrating ratio is 40, the
optimal velocity of the water in the system coupling a-Si PV cell
and TE generator is approximate 0.53 m/s. Since the performance
of the polymer PV cell declines along the optical concentrating
ratio, the total efficiency of the coupled system first decreases.
 E. Yin et al. / Energy Conversion and Management 143 (2017) 188–202
Fig. 6. Total efficiency versus concentrating ratio for the PV-TE hybrid system using different PV cells and cooling methods (a) c-Si PV (b) p-Si PV (c) a-Si PV (d) Polymer PV.
Fig. 7. Thermal resistance of the water cooling system and the power of the pump
as a function of velocity of water.
When the concentrating ratio continues increasing, the total effi-
ciency begins to increases. The reason is that the TE generator
has a greater effect on the power output of the coupled system
and the most part of generated electricity comes from the TE mod-
197
ule. The optimal velocity of water is around 0.44 m/s for the system
coupling the polymer PV cell and TE generator when C = 40.
3.3. The influence of thermal contact resistance
Two main thermal contact resistances (TCR), which are the TCR
between PV cell and TE generator as well as the TCR between TE
generator and cooling system, are considered in this paper. In order
to study the influence of thermal contact resistance, four types of
PV cells are investigated and the value of TCR ranges from 0 to
1000 K mm2/W. The results show that the impacts of the TCR on
the hybrid systems are similar when different photovoltaic cells
are used. Consequently, the system coupling the a-Si PV and TE
generator is used as an example to study the thermal contact resis-
tance, which is presented in Fig. 9. The optimal optical concentrat-
ing ratio and velocity of water are used for calculating, where
C = 40, uw = 0.53 m/s. Fig. 9(a) shows that the temperature of the
PV cell and the two sides of the TE generator all increase when
the thermal contact resistance grows. In addition, the TE module
obtains a smaller temperature difference. The impact of Rc on the
system efficiency is revealed in Fig. 9(b). One can see that, the effi-
ciency of the PV cell drops by 20%, when Rc changes from 0 to
1000 K mm2/W. The efficiency of TE module has a small decrease
(from 3.8% to 3.5%) because of the slight decline of the temperature
difference. Since the thermal contact resistance weakens the per-
198 E. Yin et al. / Energy Conversion and Management 143 (2017) 188–202

Fig. 8. The impacts of concentrating ratio and velocity of water on the total efficiency (a) c-Si PV (b) p-Si PV (c) a-Si PV (d) Polymer PV.

Fig. 9. The influence of thermal contact resistance (a) on the temperature (b) on the efficiency.

formances of the photovoltaic cell and the thermoelectric genera-
tor, the total efficiency drops by 15% (from 7.8% to 6.6%), when
Rc changes from 0 to 1000 K mm2/W. Therefore, the thermal con-
tact resistance should be as small as possible for the coupled
system.
3.4. The impact of thermal resistance of TE generator
All the semiconductor legs in the TE module are in parallel in
the direction of heat flow and the total thermal resistance of semi-
conductor legs is defined as the thermal resistance of the thermo-
 E. Yin et al. / Energy Conversion and Management 143 (2017) 188–202
Fig. 10. The electric resistance and thermal resistance of TE generator as functions
of the structure parameter of semiconductors.
electric generator RTE. RTE is adjusted by changing the structure
parameter c, while keeps the physical properties of the semicon-
ductors constant. The value of RTE is set from 0 to 10 K/W, and
RTE as a function of structure parameter is shown in Fig. 10. The
resistance of the TE module also changed as the structure parame-
ter (Fig. 10). The loading resistance r is always set as equal to the
internal resistance of TE generator. The optimal optical concentrat-
ing ratio and velocity of water are used to find out the way of fur-
ther optimizing the coupled system, when C = 6, uw = 0.23 m/s for
c-Si PV cell, C = 13, uw = 0.35 m/s for p-Si PV cell, C = 40,
uw = 0.53 m/s for a-Si PV cell, C = 40, uw = 0.44 m/s for polymer PV
cell. Fig. 11 gives the influence of RTE on the temperatures and
the output powers, where a-Si PV-TE hybrid system is taken as
an example. When RTE changes from 0 to 10 K/W, the temperatures
of the a-Si photovoltaic cell and the hot side of the TE generator
dramatically increase, whereas the temperature of the cold side
almost keeps constant. Therefore, for the system coupling a-Si pho-
tovoltaic and TE module, enlarging RTE can improve the output
power of the TE module but degrade the electric energy generated
from the PV cell, which is illustrated in Fig. 11(b). The total output
Fig. 11. The influence of thermal resistance of TE module (a) on the temperatures (b) on the output power.
199
power also has an increase because the thermoelectric generator
has a larger change of power output than the PV cell.
The total efficiency of the PV-TE hybrid system and the percent-
ages of power outputs generated by PV cell and TE generator on the
total output power as functions of RTE are shown in Fig. 12. In order
to study the applicability of different photovoltaic modules for
coupling and find out the better structure of the TE device, the
impacts of RTE on the total efficiencies of the systems coupling four
types of photovoltaic cells and thermoelectric module are investi-
gated. The total efficiency drops by 7% for c-Si PV-TE hybrid system
and 3.3% for p-Si PV-TE hybrid system when RTE changes from 0 to
10 K/W. The reason is that, although the rise of RTE improves the
value and the percentage of power output power of the TE gener-
ator, the efficiency of PV cell significantly decreases because of the
increasing temperature of the PV cell as are shown in Fig. 12
(a) and (b). It means that c-Si and p-Si PV cells are not suitable
for combining with the TE module. In contrast, for the coupled sys-
tems with a-Si PV cell or polymer PV cell, the total efficiency
increases with RTE. The total efficiency increase from 5.1% to 9.1%
for a-Si PV-TE hybrid system and from 2.3% to 8.8% for polymer
PV-TE hybrid system when RTE changes from 0 to 10 K/W. The per-
centage of power output of TE generator on the total output power
will exceed the one of PV cell when RTE is >4.6 K/W for a-Si PV-TE
hybrid system and 2.5 K/W for polymer PV-TE coupled system as is
illustrated in Fig. 12(c) and (d). For the a-Si PV cell, the total effi-
ciency of the coupled system (RTE is set as 10 K/W) approximately
increases by 78% comparing with the pure PV system when oper-
ating under the same condition. The total efficiency of the hybrid
system increase by 283% compared to the pure PV system when
a polymer PV cell is used. Therefore, a-Si PV cell and polymer PV
cell are more suitable for coupling than c-Si and p-Si PV cells,
and the TE module having a larger thermal resistance is better
for the combination. However, it should be noted that, when
RTE > 8 K/W, the temperatures of PV cells will be >473 K which
may exceed the allowable operating temperatures of commercial
PV cells.
3.5. The influence of heat loss thermal resistance
The heat loss thermal resistance indicates the ability of prevent-
ing the heat loss caused by convection and radiation. A larger value
of thermal resistance of heat loss is more un-benefit for the cooling
200 E. Yin et al. / Energy Conversion and Management 143 (2017) 188–202
Fig. 12. The total efficiency of the PV-TE hybrid system and the percentages of power outputs generated by PV cell and TE generator on the total output power as functions of
thermal resistance of the TE module (a) c-Si PV cell (b) p-Si PV cell (c) a-Si PV cell (d) Polymer PV cell.
of PV cell because it will lead to a higher temperature and a worse
performance of silicon PV cell. However, more thermal energy has
to conduct to the TE generator. This may increase the output power
generated by the TE module based on the Seebeck effect. Therefore,
investigating the impact of Rloss is necessary. Fig. 13 gives the effi-
ciencies of the overall system and TE device as functions of the
thermal resistance of heat loss. Rloss is adjusted by changing the
thickness of the front glass and the EVA. The optical concentrating
ratio and the velocity of water are used which is the same with
Section 3.4. The results show that enlarging the value of Rloss can
increase the efficiency of TE generator for these four hybrid sys-
tems. For the c-Si photovoltaic cell, the total efficiency gets a slight
decrease when Rloss increases. The reason is that the decreasing
output power of c-Si PV cell exceeds the increase of the output
power of TE generator. Differing from the c-Si PV cell, the total effi-
ciency increases when the thermal resistance of heat loss rises for
the hybrid systems with p-Si, a-Si and polymer PV cells because
these three PV cells have smaller efficiency temperature coefficient
than the c-Si PV cell. When Rloss rises from 75 K/W to 1000 K/W, the
total efficiency increases by 1% for p-Si PV cell, 1.9% for a-Si PV cell
and 3.2% for polymer PV cell. Although increasing the thermal
resistance of heat loss can only slightly improve the capability of
the coupled system, it will be more effective when a better TE
module is utilized. Accordingly, adjusting Rloss is an optional way
to further optimize the coupled system.
4. Conclusion
The thermal resistance theory is introduced into the theoretical
model of the PV-TE hybrid system. Four kinds of PV cells are
explored and the result shows that c-Si PV and p-Si PV are not suit-
able for the PV-TE hybrid system because of the decline of the total
efficiency comparing with the pure PV system. On the contrary, the
total efficiency of a-Si PV-TE hybrid system (RTE is set as 10 K/W)
approximately increases by 78% when operating under the same
condition, and increase by 283% for polymer PV-TE system. Three
different kinds of cooling methods are studied. Among the cooling
methods, water cooling is more suitable for the hybrid system with
optical concentrator. Influences of the optical concentrating ratio
and the velocity of water (thermal resistance of the cooling sys-
tem) are investigated together and the optimal values are revealed.
Based on optimally working states obtained earlier, the impacts of
thermal contact resistance, thermal resistance of TE generator and
thermal resistance of heat loss on the performance of the hybrid
system are explored, respectively. Increasing the thermal contact
resistance strongly weakens the capability of the hybrid systems.
The total efficiency drops by 15% (from 7.8% to 6.6%) for the a-Si
PV-TE coupled system when Rc changes from 0 to 1000 K mm2/
W. The TE module having a larger thermal resistance is more suit-
able for attaching to the back of a-Si PV cell and polymer PV cell.
The total efficiency increase from 5.1% to 9.1% for a-Si PV-TE hybrid
system and from 2.3% to 8.8% for polymer PV-TE hybrid system
when RTE changes from 0 to 10 K/W. Moreover, increasing the heat
loss thermal resistance can slightly improve the performance of
the hybrid system with p-Si, a-Si and polymer PV cells. It is an
optional way to further optimize the coupled system.
Consequently, when designing a practical PV-TE hybrid system,
the following principles and order are advised:
Firstly, choose the coupling devices and the cooling method.
(1) Pick the a-Si PV cell, the polymer PV cell or the other PV cell
which has the same small efficiency temperature coefficient
with the a-Si PV.
 E. Yin et al. / Energy Conversion and Management 143 (2017) 188–202
Fig. 13. Total efficiency and efficiency of TE module as functions of thermal resistance of heat loss (a) c-Si PV cell (b) p-Si PV cell (c) a-Si PV cell (d) Polymer PV cell.
(2) Choose the TE module whose p/n semiconductor legs are
fine and long in order to obtain a small structure parameter
and a large thermal resistance. In this way, TE module will
get a considerable temperature difference and contribute
more power output in the PV-TE hybrid system.
(3) Employ a water cooling system when an optical concentra-
tor is used.
Secondly, bond the devices with thermal conduct adhesive. The
contact interfaces between the PV cell, the TE module and the cool-
ing system must be coated with thermal conduct adhesive or bet-
ter thermal conduct material in order to decrease the thermal
contact resistance.
Thirdly, install a glass cover in the front of the PV module with
the EVA optical film. On the one hand, the glass cover can protect
the PV cell from the rain, wind or dust. On the other hand, it can
increase the heat loss thermal resistance and prevent the heat loss
from the top side of the PV cell. Furthermore, when without affect-
ing the light transmission, the EVA optical film and the glass cover
should be as thick as possible.
Finally, regulate the optical concentrating ratio and the velocity
of the cooling water at the same time to achieve an optimal oper-
ating condition of the PV-TE hybrid system as is illustrated in
Section 3.2.
Acknowledgments
This work was supported by the National Natural Science Foun-
dation of China (Grant No. 51590903).
201
References
[1] Tyagi V, Rahim NA, Rahim N, Jeyraj A, Selvaraj L. Progress in solar PV
technology: research and achievement. Renew Sustain Energy Rev
2013;20:443–61.
[2] Nelson J. The physics of solar cells. London: Imperial college press; 2003.
[3] Wang Y, Dai C, Wang S. Theoretical analysis of a thermoelectric generator
using exhaust gas of vehicles as heat source. Appl Energy 2013;112:1171–80.
[4] Sahin AZ, Yilbas BS. The thermoelement as thermoelectric power generator:
effect of leg geometry on the efficiency and power generation. Energy Convers
Manage 2013;65:26–32.
[5] He W, Zhang G, Zhang X, Ji J, Li G, Zhao X. Recent development and application
of thermoelectric generator and cooler. Appl Energy 2015;143:1–25.
[6] Xiao J, Yang T, Li P, Zhai P, Zhang Q. Thermal design and management for
performance optimization of solar thermoelectric generator. Appl Energy
2012;93:33–8.
[7] Chen W-H, Liao C-Y, Hung C-I, Huang W-L. Experimental study on
thermoelectric modules for power generation at various operating
conditions. Energy 2012;45:874–81.
[8] Gou X, Xiao H, Yang S. Modeling, experimental study and optimization on low-
temperature waste heat thermoelectric generator system. Appl Energy
2010;87:3131–6.
[9] Elsheikh MH, Shnawah DA, Sabri MFM, Said SBM, Hassan MH, Bashir MBA,
et al. A review on thermoelectric renewable energy: principle parameters that
affect their performance. Renew Sustain Energy Rev 2014;30:337–55.
[10] Deng Y, Zhu W, Wang Y, Shi Y. Enhanced performance of solar-driven
photovoltaic–thermoelectric hybrid system in an integrated design. Sol Energy
2013;88:182–91.
[11] Zhang J, Xuan Y, Yang L. Performance estimation of photovoltaic–
thermoelectric hybrid systems. Energy 2014;78:895–903.
[12] Van Sark W. Feasibility of photovoltaic–thermoelectric hybrid modules. Appl
Energy 2011;88:2785–90.
[13] Park K-T, Shin S-M, Tazebay AS, Um H-D, Jung J-Y, Jee S-W, et al. Lossless
hybridization between photovoltaic and thermoelectric devices. Sci Rep
2013;3.
[14] Zhu W, Deng Y, Wang Y, Shen S, Gulfam R. High-performance photovoltaic-
thermoelectric hybrid power generation system with optimized thermal
management. Energy 2016;100:91–101.
202 E. Yin et al. / Energy Conversion and Management 143 (2017) 188–202
[15] Wang N, Han L, He H, Park N-H, Koumoto K. A novel high-performance
photovoltaic–thermoelectric hybrid device. Energy Environ Sci
2011;4:3676–9.
[16] Attivissimo F, Di Nisio A, Lanzolla AML, Paul M. Feasibility of a photovoltaic–
thermoelectric generator: performance analysis and simulation results. IEEE
Trans Instrum Meas 2015;64:1158–69.
[17] Fisac M, Villasevil FX, López AM. High-efficiency photovoltaic technology
including thermoelectric generation. J Power Sources 2014;252:264–9.
[18] Luque A, Martı́ A. Limiting efficiency of coupled thermal and photovoltaic
converters. Sol Energy Mater Sol Cells 1999;58:147–65.
[19] Bjørk R, Nielsen KK. The performance of a combined solar photovoltaic (PV)
and thermoelectric generator (TEG) system. Sol Energy 2015;120:187–94.
[20] Vorobiev Y, González-Hernández J, Kribus A. Analysis of potential conversion
efficiency of a solar hybrid system with high-temperature stage. J Sol Energy
Eng 2006;128:258–60.
[21] Wu Y-Y, Wu S-Y, Xiao L. Performance analysis of photovoltaic–thermoelectric
hybrid system with and without glass cover. Energy Convers Manage
2015;93:151–9.
[22] Chen H, Ji J, Wang Y, Sun W, Pei G, Yu Z. Thermal analysis of a high
concentration photovoltaic/thermal system. Sol Energy 2014;107:372–9.
[23] Cui T, Xuan Y, Li Q. Design of a novel concentrating photovoltaic–
thermoelectric system incorporated with phase change materials. Energy
Convers Manage 2016;112:49–60.
[24] John AD, William AB. Solar engineering of thermal processes. 4th ed. New
York: Wiley; 2013.
[25] Makki A, Omer S, Su Y, Sabir H. Numerical investigation of heat pipe-based
photovoltaic–thermoelectric generator (HP-PV/TEG) hybrid system. Energy
Convers Manage 2016;112:274–87.
[26] Bergman TL, Incropera FP, DeWitt DP, Lavine AS. Fundamentals of heat and
mass transfer. 7th ed. New York: John Wiley & Sons; 2011.
[27] Mirmira S, Marotta E, Fletcher L. Thermal contact conductance of adhesives for
microelectronic systems. J Thermophys Heat Transfer 1997;11:141–5.
[28] Gomez M, Reid R, Ohara B, Lee H. Influence of electrical current variance and
thermal resistances on optimum working conditions and geometry for
thermoelectric energy harvesting. J Appl Phys 2013;113:174908.
[29] Najafi H, Woodbury KA. Modeling and analysis of a combined photovoltaic-
thermoelectric power generation system. J Sol Energy Eng 2013;135:031013.
[30] Mills AF. Heat transfer. 2th ed. America, New Jersey: Prentice Hall; 1999.
[31] Kays W, Crawford M, Weigand B. Convective heat and mass
transfer. Columbus: Tata McGraw-Hill Education; 2012.
[32] Da Y, Xuan Y. Role of surface recombination in affecting the efficiency of
nanostructured thin-film solar cells. Opt Express 2013;21:A1065–77.
[33] Van Sark W, Meijerink A, Schropp R, Van Roosmalen J, Lysen E. Enhancing solar
cell efficiency by using spectral converters. Sol Energy Mater Sol Cells
2005;87:395–409.
[34] Koster L, Smits E, Mihailetchi V, Blom P. Device model for the operation of
polymer/fullerene bulk heterojunction solar cells. Phys Rev B 2005;72:085205.
[35] Khattab N, El Shenawy E. Optimal operation of thermoelectric cooler driven by
solar thermoelectric generator. Energy Convers Manage 2006;47:407–26.
[36] Lin J, Liao T, Lin B. Performance analysis and load matching of a photovoltaic–
thermoelectric hybrid system. Energy Convers Manage 2015;105:891–9.
[37] Chen W-H, Wang C-C, Hung C-I, Yang C-C, Juang R-C. Modeling and simulation
for the design of thermal-concentrated solar thermoelectric generator. Energy
2014;64:287–97.