ISSN 0003-701X, Applied Solar Energy, 2016, Vol. 52, No. 2, pp. 86–92. © Allerton Press, Inc., 2016.

DIRECT CONVERSION OF SOLAR ENERGY

TO ELECTRIC ENERGY

Solar Panel Cooling System with Hollow Fibres1

Michaela Schmidta, Ilya Astrouskib, *, Marcus Reppicha, and Miroslav Raudenskyb
a
Faculty of Mechanical and Process Engineering, Augsburg University of Applied Sciences, 86161 Augsburg, Germany
b
Heat Transfer and Fluid Flow Laboratory, Faculty of Mechanical Engineering,
Brno University of Technology, Brno 616 69, Czech Republic
*e-mail: astrouski@LPTaP.fme.vutbr.cz
Received February 1, 2016

Abstract—Photovoltaic (PV) panel overheating drastically reduces their efficiency and lifespan. Overheating
also has the potential to form electric arcs which can melt metal fixtures and burn away the module’s insulat-
ing materials. Due to these phenomena, the introduction of water-cooling or, more generally, liquid-cooling
systems inside the PV panel appears reasonable. Hollow fibre cooling system consisting of plastic tubes of a
small diameter (less than 1 mm) was tested as a simple solution. Fibres can be placed in contact with the back
surface of a PV panel and coolant flowing through provides efficient and uniform cooling. Heat removed
from the panel may be used for domestic or industrial needs or released to the surroundings. Hollow fibres
have very thin walls (about 0.1 mm) to transfer heat easily, and the system is light, compact and resistant to
corrosion. Experiments confirmed that such system can cool the PV panel, removing up to 1 kW of heat and
lowering the module temperature from 90°C to about 50°C. It was observed that cooling improves the electric
efficiency of PV cells by about 50%.

DOI: 10.3103/S0003701X16020213

INTRODUCTION Results showed that a PVT’s output has between 6.5

With the aim of providing more and more of the
world’s energy supply with renewable energy, solar
photovoltaics (PVs) are an important technology, and
are characterized by a rapidly growing market. In 2014,
40 GW of solar PV systems were installed for a total
global capacity of about 177 GW [1]. Solar PVs have
begun to play a substantial role in electricity genera-
tion in some countries, including Italy, Greece and
Germany. The desire to increase the effectiveness of
PV panels has led to the development of photovoltaic
thermal hybrid solar collectors, sometimes known as
PVT systems. These systems combine photovoltaic
cells, which convert electromagnetic radiation into
electricity, with a solar thermal collector, which cap-
tures the remaining energy and removes waste heat
from the PV module. The main advantage of this
design is that coupling these two devices (PV system
and solar collector) decreases the temperature of PV
cells, thus increasing their electrical efficiency and
operational life. The rise of efficiency due to cooling is
the most pronounced for widely-used crystalline sili-
con PV cells.
An exergy analysis of solar thermal, photovoltaic,
and hybrid photovoltaic thermal systems conducted by
[2] confirmed the high efficiency of PVT systems
compared to classical PV panels and solar collectors.
1 The article is published in the original.
and 8.4% more exergy when compared to purely PV
systems and 4 times as much exergy as a pure solar
thermal system.
An overview of the field since the beginning of the
1970s was performed by [3]. The authors reported that
PVT products have a much shorter economic payback
period than their PV counterpart and that PVT’s
renewable energy is expected to become competitive
with conventional power generation. Extended review
on PVT designs and theoretical and experimental
works was done by [4]. The review classifies and sys-
temizes existing designs according to the working fluid
type and PVT purpose. It was shown that different
PVT designs should be chosen depending on location
and meteorological factors [4].
Methods of cooling PV panels and different PVT
designs have been studied in several works. It was
designed a solar PV panel cooling system based on a
thermosyphon heat pipe [5]. It was found that the
introduction of cooling prevents photovoltaic panels
from overheating to over 46°C over 4 hours, substan-
tially increasing electrical output. It was investigated
[6] the cooling potential of both water and air as cool-
ing agents for PVT panels. Water-based cooling was
found to be more efficient but also more expensive. It
was developed a water spraying system and mathemat-
ical model for cooling PV panels [7]. It was shown that
the maximum allowable temperature, 45°C, is a com-

86
 SOLAR PANEL COOLING SYSTEM WITH HOLLOW FIBRES
(a)
Fig. 1. Polymeric hollow fiber heat exchanger (a), PVT module with inbuilt heat exchanger (b).
promise temperature between the output energy from
the PV panels and the energy needed for cooling. Dif-
ferent designs of PVT were analyzed in review by [8].
The paper presents the state-of-the-art on flat plate
PVT collector classification, design and performance
evaluations of water, air and combinations of water
and/or air based systems. The author concluded that
PVT solar collectors are a promising alternative system
for low-energy applications and suggested several
points for future investigations. It was investigated [9]
PV panel cooling by parallel arrays of ducts with an
inlet/outlet attached to the back of the PV panel. It
was found that without active cooling the temperature
of module was high and solar cells could only achieve
an efficiency of 8–9%.
Under active cooling conditions, the temperature
dropped significantly, leading to an increase in the
efficiency of solar cells to between 12 and 14%. It was
studied output parameters of PVT converters with sil-
icon solar cell [10]. The described PVT system can be
used to obtain hot water up to 60°C, providing 25%
more electrical energy than equivalent PV module
without cooling. It was designed and studied experi-
mally thermal performance of water-cooled hybrid
PVT module [11]. It was found that total PVT effi-
ciency of 80% could be achieved.
The use of polymeric hollow fibres for heat
exchange applications was first proposed almost
10 years ago and several papers describing the hydrau-
lic and thermal response of hollow fibres were pub-
lished [12, 13]. Hollow fibres are polymeric tubes with
small diameters (about 1 mm) that can be produced by
an extrusion process using various plastics. The idea to
use hollow fibres to cool PV modules is similar to the
application of microcapillary films [14]. It was shown
that such devices can effectively remove heat from a
heated surface and can be used for cooling electronic
equipment. These types of polymeric films were stud-
ied for solar collector applications as well [15]. It was
reported that it is possible to fabricate all-plastic, low
APPLIED SOLAR ENERGY Vol. 52 No. 2 2016
87
(b)
cost and low weight thermal collectors with all the
advantages associated with a product made from one
material (plastic).
In the same manner, a PVT cooling system from
polypropylene hollow fibres was designed, purpose-
built and tested in this work. To determine any power
increase, measurements were done with an unmodi-
fied photovoltaic module and compared with mea-
surements of the same module after linking it with the
polymeric hollow fiber heat exchanger.
EXPERIMENTAL DETAILS
The PVT cooling system (Fig. 1) was made of poly-
propylene fibres [16] which were uniformly distributed
and the ends of the fibres were connected to PP
inlet/outlet collectors. The system is light, compact,
non-corrosive and has a small internal volume, mak-
ing it suitable for thermal control. Table 1 presents the
main parameters of the cooling system (for more
details see [17]). The system was tested for leaks by
pressurizing with water up to a 3-bar gauge pressure.
Moreover, coloured water (with dye added) was used
to determine which fibres water could flow through.
A monocrystalline silicon PV module was used to
test the cooling system. The module was a donation
the University of Augsburg, so the exact origin and age
of the module were unknown. Information about the
module was taken from the label on the back of the
module: STORM, model ND1325X674. The corre-
sponding data are summarized in Table 2.
The cooling system was placed inside the module
as shown in Fig. 1. Contact between the fibres and the
back side of the PV module was created by pressure
from the insulation layer. To estimate the fibre heat
transfer area in contact with the back surface of the
PV, the following procedure was performed: fibres
were painted with ink and the cooling system was
placed inside the PV module. Next, contact pressure
was introduced and the module was mounted. Finally,
88 MICHAELA SCHMIDT et al.

Table 1. Heat exchanger characteristics

Parameter Value Unit

Length of collectors with fibres 1000 mm

Width of the system/Fibre length 600/570 mm
Inlet/outlet collector tube diameter 1/2 inch
Full number of fibres/Number of functional fibres 472/425 –
Fibre outer/inner diameter 0.83/0.63 mm
Heat transfer area based on full external surface of fibres 0.63 m2
Heat transfer area in touch with PV back surface 0.078 m2

the module was dismounted and the cooling system resistance values of 15.4 and 48 Ohms. This test
was removed. The fibres' footprints were measured allowed the team to estimate the trend pattern of the
and used to calculate the heat transfer area. module temperature during normal operation and to
Figure 2 shows the principal layout of the experi- evaluate all parameters in the plateau region (once the
mental stand. It consists of a set of lamps: 32 mercury temperature had reached a stable value).
vapour lamps and 27 halogen lamps, providing radia- Experiments with cooling were repeated for three
tion in the solar spectrum range; cooling circuit equip- temperatures of cooling water. The distance from
ment and measurement equipment. The radiation lamps to module and their intensity were constant,
produced by the lamps had some differences from day creating constant irradiation. Pump settings were con-
to day and point to point. The average intensity of the stant for all tests as well.
lamps (24 measurements) was 802 W/m2, taking 716 W
for the entire surface of the module.
Stand equipment was used to collect the electrical REDUCTION DATA
and thermal characteristics of the module [17]. In par- The electrical power and the heat transfer of cool-
ticular, the following parameters were recorded: ing system rate can to be calculated based on experi-
• ambient temperature; mental data (Table 1). The electrical power P can be
• radiation intensity (at the edge of the test stand); calculated as:
• current and voltage; P = U ⋅ I. (1)
• three surface temperatures on the top of the Considering a stationary regime with no heat
module; sources inside the cooling system, constant water den-
• four surface temperatures on the back of the sity and heat capacity [14] and absence of heat losses
module; to surrounding, the heat transfer rate can be deter-
• pressures, temperatures and the flow rate of mined:
cooling water.
This data was used to calculate performance, mean Q = m ⋅ c p ⋅ Δϑ, (2)
temperature on the top of the module, mean tempera-
ture on the back of the module and average module ρ = m , (3)
temperature. The positions of the temperature sensors V
were between the individual solar cells.
Three different types of tests were performed to Table 2. PV module characteristics
compare the cooled and uncooled module: the Maxi- Parameter Value Unit
mum-Power-Point test, the resistance test and the
long-term test. To find the Maximum-Power-Point Rated maximum power 110 W
the resistance was changed continuously using the Rated voltage 34.2 V
rheostats from a minimum value to an infinite one. In Rated current 3.22 A
the Change-of-Resistance test, the series of defined
resistance values were approached. At intervals of Open-circuit voltage 42.8 V
approximately 3 min, the resistance was changed using Short-circuit current 3.40 A
the rheostats. The resistance values were 0.55, 8.8, 11, Maximum system voltage 1000 V
12.1, 15.4, 48 and ∞ Ohms. During the Long-term Length 1.32 m
test, all parameters were recorded for three hours for a
specific resistance value. This test was performed with Width 0.67 m

APPLIED SOLAR ENERGY Vol. 52 No. 2 2016

 SOLAR PANEL COOLING SYSTEM WITH HOLLOW FIBRES
Ambient air
temperature
Pyranometer
Anemometer
Manometer
Fuild temperature Surface temperature
Volumetric flow rate meter
Pump
Rising device
Fig. 2. Principal scheme of the measurement stand (hybrid module).
Q = ρ ⋅ V ⋅ c p ⋅ Δϑ.
The energy of pumping P can be simply calculated as:
P = Q × Δp.
RESULTS AND DISCUSSION
Maximum-Power-Point Test
In the Maximum-Power-Point test (Table 3) the
temperatures measured in different locations on the
PV panel were slightly different. The temperature vari-
ation was caused by the non-homogeneity of the cool-
ing and radiation of lamps. For example, in the area of
the panel electric box, cooling was lower because
fibres were in poor contact with the surface. The more
the module was cooled, the more pronounced this
effect was.
As was expected, the temperature difference
between the inlet and outlet was higher for lower cool-
ing. This was due to the greater driving temperature
APPLIED SOLAR ENERGY Vol. 52 No. 2
89
Amperemeter and
voltmeter
A V
Resistance
Manometer
Fuild temperature
Plate heat exchanger
Volumetric flow rate meter
Cooling
thermostat
(4) difference between the coolant and PV panel. The sur-
face temperatures on top of the panel were slightly
lower than those on the bottom (the maximum differ-
ence was of 2.6°C). Although the cooling effect was
(5)
high enough in all cases and caused a significant
reduction of the average panel temperature, the tem-
perature of the coolant water had a small influence on
cooling.
Figure 3 shows that cooling leads to a significantly
greater peak of electrical power output. By cooling,
the maximum power point is shifted to the right,
reaching the maximum only at a higher voltage. At low
voltages, cooling does not influence the resulting out-
put or has only a minor influence on it.
Whether the module is cooled with water at tem-
perature of 10 or 15°C, the cooling effect is irrelevant.
Without cooling, a maximum power of 31.2 W was
achieved. This value significantly differs from the
manufacturer’s specified nominal output of 110 W
because the aging of the module decreases its effi-
2016
90 MICHAELA SCHMIDT et al.

Table 3. Results of the Maximum-Power-Point test

Parameter and unit Cooling regime 1 Cooling regime 2 Cooling regime 3 No cooling
Heat flow, W 1011 895 776 –
Water inlet temperature, °C 8.6 12.5 17.0 –
Water outlet temperature, °C 12.1 15.3 19.3 –
Water flow rate, l/min 4.04 4.53 4.74 –
Water pressure drop, bar 0.42 0.41 0.38 –
Temperature of the module top, °C 50.2 52.0 55.5 90.8
Temperature under the module, °C 52.6 52.6 55.6 89.7
Maximum electrical power, W 44.8 38.1 38.6 31.2

ciency. However an increase in the maximum power
output by up to 44% to 44.8 W can be achieved by
cooling.
Change-of-Resistance Test and Long-term Test
The results of the Change-of-Resistance test and
Long-term test had similar trends as the Maximum-
Power-Point test. To evaluate the influence of resis-
tance, six resistances were tested. The inlet/outlet
coolant temperatures of the resistance tests were close
to the corresponding temperatures at the Maximum-
Power-Point test. The results of the resistance test
showed similar trends as the results of the Maximum-
Power-Point test. The effect of cooling is pronounced
only for 48 Ohms of resistance, and it was confirmed
that the cooling affects output only for resistances near
the maximum power point.
Firstly, the long-term test was carried out at a resis-
tance of 15.4 Ohms and then it was repeated at a resis-
tance value of 48 Ohms. For the 15.4-Ohm resistance
there was no significant influence of cooling on volt-
age, current and power. These parameters were stable
during the three hour test even when the temperature
increased from 45 to 90°C. This result agreed with the
results of the resistance test.
The opposite was observed for the long-time test
with a resistance of 48 Ohms (Figs. 4, 5). Table 4
shows the results in the plateau region. Although the
Power P, W
50
Cooling regime 1
40 No cooling
30
20
10
inlet/outlet coolant temperatures and module tem-
peratures were very similar to the 15.4 Ohms resis-
tance, the influence of cooling on electrical perfor-
mance was pronounced. Substituting the performance
of the uncooled module equal to 100%, we achieved a
performance increase of over 50%. During the three
hour experiment the temperature of the module
mostly achieved its asymptotic value of 95°C. The sig-
nificant improvement in the performance due to the
cooling is again apparent.
Pressure Drops and Energy Required for Pumping
Cooling Water
The implementation of forced circulation of water
(or more generally, cooling fluid) for PVT cooling is
simple and reasonable but its economical expediency
should be confirmed. It is obvious that the energy
profit from the cooling should be higher than the elec-
trical energy required for pumping.
The pressure drop and flow rate data from Table 2
seem unexpected (lower pressure drop for higher flow
rate). This discrepancy can be explained by the change
of the cooling intensity and the temperature regime.
The fibres’ inner diameter is larger for higher tempera-
tures (cooling regime 1) due to polypropylene thermal
expansion and higher mechanical deformation by
inner pressure because polypropylene loses its stiffness
with increasing temperature. Moreover, water viscos-
ity rises with a decrease in temperature so it is smaller
for a higher water flow rate (cooling regime 2 and 3).
Module mean temperature tmodule, °C
100
80
60
40
20 Cooling regime 1 No cooling

0 5 10 15 20 25 30 35 40 0 2000 4000 6000 8000 10 000

Voltage U, V Time, s

Fig. 3. Maximum-Power-Point test: power vs the voltage. Fig. 4. Long-term test: module temperature over the time.

APPLIED SOLAR ENERGY Vol. 52 No. 2 2016

 SOLAR PANEL COOLING SYSTEM WITH HOLLOW FIBRES 91

Electrical power, W The advantage of the cooled module would be

40
30
20
10 Cooling regime 1 No cooling
0 2000 4000 6000 8000 10 000
Time, s
Fig. 5. Long-term test: electrical power over the time.
Thus, the combination of these two factors causes a
reduction of the pressure drop to flow rate ratio for
constant pump adjustment.
In our experiment, a Grundfoss Type UPS 25–80
[18] centrifugal pump installed as circulation pump
was operated in the low efficiency range. The mean
water flow rate was 4.5 l/min and mean pressure drop
in hollow fibres was 0.4 bar (Table 1). The mechanical
energy of pumping can be easily calculated based on
flow rate and pressure drops, and was 3 W. Consider-
ing efficiencies of the pump, transmission and electric
motor, the energy required for pumping is 5 W. This
value shows that the amount of electricity saved due to
the cooling is three times higher than the electricity
required for water pumping.
Conclusions and Outlook
Various tests showed that a hollow fibre system
could work as an integrated heat exchanger of PVT
system. It allows for the removal of to 1 kW of heat and
reduces the module temperature from 90°C to about
50°C. It was observed that cooling of the PV module
improves its efficiency. The mean efficiency increase
is about 50% and it is expected that it can be even more
for high-efficiency modules. The investigation of new
commercially available photovoltaic modules could
show how high the appropriate additional income
would be. The energy saved do to the cooling is signifi-
cantly higher than the energy for pumping the cooling
fluid through the module.
more pronounced in hot climate areas. However, the
system can provide useful features in cold climates as
well. In addition to the thermal energy production, the
system can be used, for example, to melt the snow on
the module surface. These efforts are particularly
advantageous in a limited space because the electrical
power generated per square meter increases and ther-
mal energy can be produced without the installation of
additional devices. Another relevant factor is that the
total weight of the system stays low. This is important
not only for the roof load, but also for the manageabil-
ity. From this point of view the use of polymeric mate-
rial is a significant advantage.
On the other hand, the cooling system increases
the cost and requires energy to operate the pump in the
coolant circuit. Moreover, a suitable material (poly-
amide 66) should be chosen because the temperature
resistance and durability of polypropylene is limited
and overheating of the module over 100°C could dam-
age the system.
REFERENCES
1. Renewable Energy Policy Network for the 21st Cen-
tury, Ed., Renewables 2015 Global Status Report. Paris
2015. Retrived online 4.8.2015 http://www.
ren21.net/status-of-renewables/global-status-report/
2. Pathak, M.J.M., Sanders, P.G., and Pearce, J.M.,
Optimizing limited solar roof access by exergy analysis
of solar thermal, photovoltaic, hybrid photovoltaic
thermal systems, Applied Energy, 2014, vol. 120,
pp. 115–124.
3. Chow, T.T., A review of photovoltaic/thermal hybrid
solar technology, Applied Energy, 2010, vol. 87, no. 2,
pp. 365–379.
4. Avezov, R.R., Akhatov, J.S., and Avezova, N.R., A
review on photovoltaic thermal (PV–T) air and water
collectors, Appl. Sol. Energy, 2011, vol. 47, pp. 169–183.
5. Akbarzadeh, A. and Wadowski, T., Heat pipe-based
cooling systems for photovoltaic cells under concen-
trated solar radiation, Appl. Therm. Eng., 1996, vol. 16,
pp. 81–87.
6. Chaniotakis, E., Modelling and analysis of water
cooled photovoltaics, M.Sc. Thesis, Faculty of Energy

Table 4. Results of the Long-term test with a resistance of 48 Ω

Parameter and unit Cooling regime 1 Cooling regime 2 Cooling regime 3 No cooling
Temperature of the module top, °C 51.9 55.1 57.7 90.5
Temperature under the module, °C 50.8 53.4 56.3 94.2
Water inlet temperature, °C 8.2 12.6 17.0 –
Water outlet temperature, °C 10.4 14.6 18.9 –
Water flow rate, l/min 4.60 4.81 5.00 –
Water pressure drop, bar 0.43 0.40 0.38 –
Maximum electrical power, W 27.4 26.7 26.1 18.2
Heat flow, W 707 691 662 –

APPLIED SOLAR ENERGY Vol. 52 No. 2 2016

92 MICHAELA SCHMIDT et al.
System and Environment, Department of Mechanical
Engineering, University of Strathclyde, Glasgow, Scot-
land, 2001.
7. Moharram, K.A., Abd-Elhady, M.S., Kandil, H.A.,
and El-Sherif H., Enhancing the performance of pho-
tovoltaic panels by water cooling, Ain Shams Engineer-
ing Journal, 2013, vol. 4, pp. 869–877.
8. Ibrahim, A., Othman, M., Ruslan, M.H., Mat, S., and
Sopian, K., Recent advances in flat plate photovol-
taic/thermal (PV/T) solar collectors, Renew Sustain
Energy Rev., 2011, vol. 15, pp. 352–365.
9. Teo, H.G., Lee, P.S., and Hawlader, M.N.A., An
active cooling system for photovoltaic modules, Applied
Energy, 2011, vol. 90, pp. 309–315.
10. Mirzabaev, M., Mirzabaev, A.M., Kononerov, V.P.,
et al. Photothermal converters with silicon solar cells,
Appl. Sol. Energy, 2007, vol. 43, pp. 75–77.
11. Touafek, K., Haddadi, M., and Malek, A., Experimen-
tal study on a new hybrid photovoltaic thermal collec-
tor, Appl. Sol. Energy, 2009, vol. 45, pp. 181–186.
12. Zarkadas, D.M. and Sirkar, K.K., Polymeric hollow
fiber heat exchangers: An alternative for lower tempera-
ture applications, Ind. Eng. Chem. Res., 2004, vol. 43,
pp. 8093–8106.
13. Astrouski, I. and Raudensky, M., The study of poly-
meric hollow fibre heat exchangers, Proc. of the 18th
Inter. Conf. Engineering Mechanics, Svratka, 2012,
pp. 47–57.
14. Hallmark, B., Hornung, C.H., Broady, D., et al. The
application of plastic microcapillary films for fast tran-
sient micro-heat exchange, Intern. Journal of Heat and
MassTransfer., 2008, vol. 51, pp. 5344–5358.
15. Dorfling, C., Hornung, C.H., Hallmark, B., et al., The
experimental response and modelling of a solar heat
collector fabricated from plastic microcapillary films,
Sol. Energy Mater. Sol. Cells, 2010, vol. 94, p. 1207e21.
16. Zena Membranes s.r.o. Retrived online 4.8.2015
http://www.zena-membranes.cz.
17. Schmidt, M., Leistungsoptimierung eines Photovoltaik-
moduls durch Kühlung mittels Kunststoffhohlfaser-
wärmeübertrager, Masterarbeit, Fakultät für
Maschinenbau und Verfahrenstechnik, Hochschule
Augsburg, Germany, 2014.
18. Grundfoss Type UPS 25-80 manual. Retrived online
4.8.2015 http://www.pumppower.com.au/wp-con-
tent/uploads/2013/07/ups25-80-N-180.pdf.
APPLIED SOLAR ENERGY Vol. 52 No. 2 2016