This article appeared in a journal published by Elsevier.

The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
 Author's personal copy

Available online at www.sciencedirect.com

Solar Energy 85 (2011) 139–153

www.elsevier.com/locate/solener

A prototype photovoltaic/thermal system integrated

with transpired collector
Andreas K. Athienitis ⇑, James Bambara, Brendan O’Neill, Jonathan Faille
Dept. of Building, Civil and Environmental Engineering, Concordia University, 1455 Maisonneuve W., Montreal, Quebec, Canada H3G 1M8

Received 30 July 2009; received in revised form 1 March 2010; accepted 18 October 2010
Available online 4 December 2010

Communicated by: Associate Editor Harvey Bryan

Abstract

Building-integrated photovoltaic/thermal (BIPV/T) systems may be utilized to produce useful heat while simultaneously generating
electricity from the same building envelope surface. A well known highly efficient collector is the open-loop unglazed transpired collector
(UTC) which consists of dark porous cladding through which outdoor air is drawn and heated by absorbed solar radiation. Commer-
cially available photovoltaic systems typically produce electricity with efficiencies up to about 18%. Thus, it is beneficial to obtain much
of the normally wasted heat from the systems, possibly by combining UTC with photovoltaics. Combination of BIPV/T and UTC
systems for building facades is considered in this paper – specifically, the design of a prototype facßade-integrated photovoltaic/thermal
system with transpired collector (BIPV/T). A full scale prototype is constructed with 70% of UTC area covered with PV modules spe-
cially designed to enhance heat recovery and compared to a UTC of the same area under outdoor sunny conditions with low wind. The
orientation of the corrugations in the UTC is horizontal and the black-framed modules are attached so as to facilitate flow into the UTC
plenum. While the overall combined thermal efficiency of the UTC is higher than that of the BIPV/T system, the value of the generated
energy – assuming that electricity is at least four times more valuable than heat – is between 7% and 17% higher. Also, the electricity is
always useful while the heat is usually utilized only in the heating season. The BIPV/T concept is applied to a full scale office building
demonstration project in Montreal, Canada. The ratio of photovoltaic area coverage of the UTC may be selected based on the fresh air
heating needs of the building, the value of the electricity generated and the available building surfaces.
Ó 2010 Elsevier Ltd. All rights reserved.

Keywords: Building-integrated photovoltaic/thermal system; Unglazed transpired collector; Air preheating; Equivalent efficiency

1. Introduction (essentially a perforated dark cladding with small pores

and an insulated plenum/cavity behind the cladding) the
Building facßades that have good solar orientation are UTC is a highly efficient collector system, particularly for
generally well suited for integrated unglazed transpired fresh air preheating since all heat collected is useful.
collectors (UTC) for ventilation air preheating in residen- Dymond and Kutscher (1997) state that the pores generally
tial and commercial buildings. For mid-latitudes, where cover 0.5–2% (known as porosity) of the collector area. Air
heating is generally required in the winter, the optimal ori- is drawn through the porous cladding with the aid of a fan
entation for a vertical surface is an azimuth angle up to and thermal efficiencies of 60% or more are typically
about ±35° from due south in the northern hemisphere. achieved.
Kutscher (1994) states that, although very simple, Heat transfer in a UTC has been modelled assuming uni-
form suction by Kutscher et al. (1993) who developed a
⇑ Corresponding author. Tel.: +1 514 848 2424x8791; fax: +1 514 848 thermal model to predict UTC performance. Dymond
7965. and Kutscher (1997) also developed a flow distribution
E-mail address: [email protected] (A.K. Athienitis). and design model for a UTC. Van Decker et al. (2001)

0038-092X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.solener.2010.10.008
 Author's personal copy

140 A.K. Athienitis et al. / Solar Energy 85 (2011) 139–153

Nomenclature

ABIPV/T total surface area of the BIPV/T system (m2) hr1 radiative heat transfer coefficient between the
APV area of the BIPV/T system covered by PV mod- back side surface of the PV module and the
ules (m2) UTC surface of the BIPV/T system that is cov-
AUTC area of the UTC system (m2) ered by the PV (W/m2/°C)
AUTC_exposed exposed UTC area of the BIPV/T system hr2 radiative heat transfer coefficient between the
(m2) UTC surface of the BIPV/T system (that is cov-
AUTC_PV UTC area of the BIPV/T system that is cov- ered by the PV) and the plenum insulated sur-
ered by attached PV modules (m2) face (W/m2/°C)
cp specific heat capacity of air (J/kg/°C) khole thermal conductivity of air evaluated at the
cp_avg specific heat of air evaluated at the average UTC UTC perforation temperature (W/m/°C)
surface temperature and ambient air tempera- mBIPV/T mass flow rate of air through the BIPV/T system
ture (J/kg/°C) (kg/s)
cp_BIPV/T specific heat of air evaluated at the average of mPV mass flow rate of air through the portion of the
the BIPV/T outlet air temperature and ambi- BIPV/T system that is covered by the PV mod-
ent air temperature (J/kg/°C) ules (kg/s)
cp_hole specific heat of air evaluated at the average of msuction mass flow rate of air through one square meter
the UTC pore outlet air temperature and ambi- of transpired collector surface area (kg/h/m2)
ent air temperature (J/kg/°C) mUTC_exposed mass flow rate of air across the exposed
cp_PV specific heat of air for the BIPV/T system UTC area of the BIPV/T system (kg/s)
portion that is covered by the PV modules eval- mUTC mass flow rate of air across the UTC system
uated for the average of outlet air and ambient (kg/s)
air temperatures (J/kg/°C) Nuhole Nusselt number for heat transfer occurring in
cp_UTC specific heat of air evaluated at the average of the front, the sides and the back of the UTC per-
UTC outlet air temperature and ambient air foration.
temperature (J/kg/°C) p UTC perforation pitch (m)
d UTC perforation (pore) diameter (m) QBIPV/T total useful heat transfer from the BIPV/T sys-
EBIPV/T electrical energy produced by the BIPV/T sys- tem
tem (W) QPV total useful heat transfer from the portion of
Thx heat exchange effectiveness of the UTC BIPV/T system that is covered by PV modules
EPV electrical energy produced by the PV modules and which enters the plenum behind the tran-
per unit area of PV module (W/m2) spired collector plate (W)
G total solar irradiance on vertical facade QUTC_exposed total heat transfer from the exposed UTC
(W/m2) area of the BIPV/T system (W)
hc1 convective heat transfer coefficient between the QUTC heat transfer from the UTC system (W)
surface of the PV module and the PV backside Rehole Reynolds number based on the hole(pore) diam-
cavity air (W/m2/°C) eter
hc2 convective heat transfer coefficient between the T temperature used in the air property correla-
PV backside cavity air and the UTC surface of tions (°C)
the BIPV/T system that is covered by the PV Tcavity average air temperature behind the PV modules
(W/m2/°C) (°C)
hc3 convective heat transfer coefficient between the Tins plenum insulated surface temperature (°C)
UTC surface and the air entering the UTC ple- To ambient air temperature (°C)
num of the BIPV/T system (for the section cov- Toutlet_BIPV/T outlet air temperature of the BIPV/T sys-
ered by the PV) (W/m2/°C) tem (°C)
hc4 convective heat transfer coefficient between the Toutlet_PV outlet air temperature of the BIPV/T system
air flowing in the UTC plenum of the BIPV/T that is covered by the PV modules for a con-
system (for section covered by the PV) and the trol volume corresponding to one module (°C)
plenum insulated surface (W/m2/°C) Toutlet_UTC outlet air temperature of the UTC system
hi combined radiative and convective PV module (°C)
backside heat transfer coefficient (W/m2/°C) TPV average PV temperature (°C)
ho combined radiative and convective exterior heat Tstc PV modules cell temperature at standard test
transfer coefficient (W/m2/°C) conditions
 Author's personal copy

A.K. Athienitis et al. / Solar Energy 85 (2011) 139–153 141

TUTC_PV average temperature of the UTC surface of the gthermal_BIPV/T equivalent thermal efficiency of the
BIPV/T system for the section that is covered by BIPV/T system
PV modules (°C) mhole kinematic viscosity of air evaluated at the aver-
Vsuction suction velocity (m/s) age UTC perforation outlet air and ambient air
Vwind wind velocity (m/s) temperature (m2/s)
n UTC porosity (%)
Greek symbols q density of air (kg/m3)
aPV average PV module solar absorptance qhole density of air in the UTC holes (evaluated at the
bPV PV module temperature coefficient average of the UTC outlet air temperature and
gstc PV module efficiency at standard conditions ambient air temperature) (kg/m3)
gUTC thermal efficiency of UTC system
gBIPV/T combined thermal and electrical efficiency of the
BIPV/T system

developed correlations for the effectiveness of different UTC 2. Useful heat generation from the PV. In BIPV/T sys-
plates for a number of geometrical patterns of the pores and tems, a cooling fluid, such as water or air, is used to
different porosities and flow rates; the effectiveness is sepa- extract heat from the PV in an open-loop or closed-
rated into three parts: heat transfer from the front of the loop configuration. For example, in an open loop air
plate, in the hole, and at the back of the plate. The UTC system, outdoor air is passed under envelope-integrated
attains its high efficiency by reducing convective heat losses. PV panels, cooling them and recovering useful heat
Both the velocity and thermal boundary layers asymptoti- that would otherwise be lost to the outdoor environ-
cally reach values which remain practically constant over ment. Since PV typically has an electrical efficiency of
most of a large collector surface. Nevertheless, high wind 6–18% – with the remainder 80% or so of incident solar
speeds (higher than about 2 m/s) reduce the efficiency of radiation lost largely as heat, and since it also produces
the UTC by increasing turbulence and convection losses, more electricity when cooled, there is a dual benefit in
particularly for low suction velocities. Gunnewiek et al. cooling the PV modules – increased electricity produc-
(2002) studied the effect of wind flow on UTC performance tion and generation of useful heat. This heat can be
and, to avoid flow reversal, recommend a suction velocity of used for space, ventilation or domestic hot water
0.017 m/s for long buildings and wind normal to the UTC, (DHW) heating either by direct means or through a
0.026 m/s for cubical buildings with normal wind, and heat pump.
0.039 m/s for wind at 45° to UTC on a cubical building.
Another technology that may potentially be combined Various researchers (Bazilian et al., 2001; Charron and
with UTC to produce both useful heat and electricity is a Athienitis, 2006; Guiavarch and Peuportier, 2006; Liao et
building-integrated photovoltaic system. Photovoltaic al., 2007) have studied the energy performance of actively
modules may be integrated as construction elements into cooled PV panels integrated into facßades and roofs.
roofs, walls and windows of buildings. Integration Correlations for convective heat transfer coefficients in
improves the cost effectiveness of PV by providing, in addi- photovoltaic/thermal systems have been developed by Liao
tion to electricity production, one or more of the following et al. (2007) through experiments with a particle image
functions that also define the type of BIPV system: velocimetry system and with computational fluid dynamics
simulations. Charron and Athienitis (2006) performed a
1. Building envelope cladding or outer layer of roof (BIPV) theoretical study of ventilated double-facades with inte-
replacing, for example, roof tiles or asphalt shingles. grated PV to study the effect of different design parameters
Architectural and aesthetic integration is a major require- on thermo-electrical system performance.
ment in this type of BIPV system. BIPV has a significant BIPV/T systems may potentially be made by combining
advantage over modular systems that are an add-on to the UTC with solar cells directly on the collector cladding.
building and often require penetrations of the envelope to Delisle and Collins (2007) investigated such a system
be attached. BIPV orientation may not always be optimal through modelling and a small scale prototype. They
due to other constraints in building design. However, report that the electricity produced may be significantly
building integration of photovoltaic modules results in higher than the reduction in useful thermal energy. How-
a much higher overall performance and enhanced dura- ever, significant technical challenges remain in the develop-
bility of the building envelope when compared with ment of UTC with PV cells directly integrated on it (the
attached PV systems. For example, an IEA report cells would typically cover 20–60% of the available area)
(2002) finds that PV systems can last 50 years while cur- primarily due to the large temperature differentials and
tain wall components such as sealed glazing units may the competing thermal and electrical performance require-
need replacement every 25 years or so. ments. Another option is to integrate specially designed
 Author's personal copy

142 A.K. Athienitis et al. / Solar Energy 85 (2011) 139–153

photovoltaic modules onto the UTC using a suitable 2. Design concept and prototype for BIPV/T system
attachment system. In this option, there will be some
uncovered UTC and a larger covered part with PV panels. 2.1. Design concept
This is the option investigated in this paper through the
construction of a prototype and comparison with UTC The combination of UTC and attached PV panels on
under the same conditions. Finally, a full scale demonstra- facades is promising for building applications where there
tion project of the concept is briefly presented. is a significant need to heat ventilation air in winter. Since

Fig. 1. (a) Concept schematic for BIPV/T system and (b) detail showing attachment of PV modules and airflow paths around the bottom frame of a PV
module and into the transpired collector.
 Author's personal copy

A.K. Athienitis et al. / Solar Energy 85 (2011) 139–153 143

most commercial buildings use heat recovery ventilators to cladding (Fig. 1b). Buoyancy forces due to natural con-
extract as much useful heat as possible from exhaust air, vection act in the opposite direction of mechanically dri-
any excess heat from a BIPV/T system may be utilized ven air flow, thus the attached upper edge encourages air
for space heating or possibly to heat water through a heat flow into the plenum.
pump or air-to-water heat exchanger. A potential advan- 3. The vertical spacing between the modules was chosen
tage of a BIPV/T combination versus simple UTC for a based on the above criterion.
facade is that much of the solar heat generated in summer 4. The value of the electricity produced is assumed here to be
may not be required, so the production of electricity may about four times the value of the heat based on the con-
be an added bonus. This is the case for a new building – cept that with one unit of electricity as input to an air
the John Molson School of Business (JMSB) building – source heat pump, one can produce on average four units
built recently in Montreal at Concordia University. The of heat. In general, the ratio between the relative value of
present study served as a basis for the design of the electricity and heat ranges from 1 for a simple first law
BIPV/T demonstration project of this building and is approach to 17 when a typical exergy analysis is consid-
briefly introduced at the end of the paper. ered (Coventry and Lovegrove (2003)). Others prefer to
The basic concept for the BIPV/T system with tran- use the concept of energy savings efficiency which
spired collector is represented in Fig. 1a and b where the increases the value of BIPV/T electricity based on the elec-
prototype is shown together with a cross section showing tric power generation efficiency of a conventional power
the airflow paths for the heated air. The UTC area without plant (Chow, 2010).Wall (1990) states that electricity is
PV attached is designated as AUTC_exposed while the area of approximately five times as valuable as heat for ordinary
PV is APV. room temperatures. Based on such a concept, we can opti-
This system will generate heat and electricity at a ratio mize the area covered by photovoltaic panels APV versus
that depends on the following factors: the uncovered UTC area AUTC_exposed (the assumed factor
four can change depending on the application).
1. The electrical conversion efficiency of the PV panels and 5. The fan energy consumption needs also to be considered
their temperature coefficient which determines how elec- in designing the system, but this is generally much less (less
trical efficiency is reduced by rising panel temperature. than 5%) than the energy recovered. The outlet air tem-
2. The panel dimensions, framing solar absorptance and perature produced by the system is also important in
thermal properties; these affect the energy balance of deciding flow rate in addition to thermal efficiency, as well
the panels. as electrical efficiency increase due to PV cooling.
3. The array design – non-uniform array temperatures,
caused by vertical stratification, may affect electrical
production depending on the PV array series and paral- 2.2. Prototype development and outdoor experiment
lel interconnections.
4. The porosity of the UTC cladding and how the corruga- A prototype was designed based on the above concept,
tions are oriented – vertical or horizontal. and an experimental facility was built to test and monitor
5. The incidence angle of the solar radiation since at high the performance of the BIPV/T system. Located on the
sun the PV modules will partly shade the area without roof of a building in downtown Montreal at Concordia
PV. University (latitude 45°N), the south-facing facßade of the
6. The flow rate behind the UTC cladding, which generally facility consists of equal areas of UTC and BIPV/T systems
increases heat extraction; however, as it is raised, fric- operating side by side. Since the only difference between the
tion losses and energy consumed by the fan also two setups is the addition of PV modules on the BIPV/T
increase. side, their influence can be directly evaluated. Fig. 2 shows
7. The wind speed which generally affects the exterior con- a schematic of two facßade test sections and the airflow sys-
vective heat loss coefficient. tem. Air is drawn through each test section with a variable
speed fan and airflow is measured by means of a laminar
The design of the system adopted the following impor- flow element with a maximum error of 0.31% at the peak
tant concepts: flow rates.
A Li-Cor weather station was used to measure wind speed
1. The transpired collector cladding was placed with the and direction, as well as ambient air temperature at about
webs horizontal as shown in Fig. 1b, and the PV mod- 10 m high from the local ground (rooftop). Total vertical
ules were sized to minimize the vertical height to hori- irradiance was measured with a Li-Cor Li-200SA pyranom-
zontal length aspect ratio to improve heat removal eter. The local ambient air temperature was measured at one
from the modules. third of the facßade height and at two thirds with shielded
2. The upper edge of the PV modules was mounted directly thermocouples. The temperature distribution along the
against the UTC cladding. This was done to minimize height in various locations was measured with calibrated
the escape of heated air through a potential gap between T-type thermocouples having an accuracy of ±0.3 °C. Ther-
the top frame of the panel and the transpired collector mocouples were placed throughout each facßade test section
 Author's personal copy

144 A.K. Athienitis et al. / Solar Energy 85 (2011) 139–153

Fig. 2. Schematic of experimental setup: (a) UTC and BIPV/T test sections and (b) airflow system schematic.

along the central vertical axis. In the BIPV/T section, three The experiment focused on the performance of the two
of the five modules (first, middle and top) were fitted with systems under clear low wind (velocity less than 2 m/s)
six thermocouples each; three distributed along the central quasi-steady state conditions within 1 h from solar noon.
vertical axis, and three along the outer vertical axis. As the A variable speed controller was used to implement four
photovoltaic panels are centrally located with respect to flow rates in the experiment: 50, 80, 115, 150 kg/h/m2 (mass
the test facade, symmetric temperature distribution about flow rate of air per square meter of collector). A 30-min
the vertical central axis can be assumed (this was confirmed interval between flow rates was chosen to ensure that
through infrared thermography). quasi-steady state conditions were achieved (the time
 Author's personal copy

A.K. Athienitis et al. / Solar Energy 85 (2011) 139–153 145

constant of each facßade section is about 3 min). The system 2.4. System modelling and analysis
was found to be balanced without any need to adjust the
balancing dampers, which shows that the addition of the The BIPV/T system was separated into two areas for
PV modules in the BIPV/T prototype did not affect the analysis as shown in Fig. 2a. The first part is the fully
pressure drop in the system (the flow rate was equal in both exposed UTC and the second part is the UTC part that
test sections). Electricity produced by the PV modules is covered by PV panels. Due to the small size of the exper-
(operating at their maximum power point) was measured imental setup, the suction across the UTC was assumed to
and stored in a battery through a charge controller. be uniform. The parameters that are of interest include the
outlet air temperatures and the PV and UTC surface tem-
2.3. Photovoltaic module design considerations peratures for both the exposed and covered portions. These
parameters are greatly influenced by the mass flow rate of
The PV modules selected in this experiment have a num- air per unit transpired collector surface area, msuction, which
ber of custom designed features to improve total absorbed is generally chosen at the early stages of design. It should
solar radiation that can be converted to useful heat. First, be noted that in the design of such BIPV/T systems, which
efforts to increase the effective solar absorptance of the generate both electricity and useful heat, the selection of
panel (including the area between cells) were made by the ratio of the two areas is an important decision.
selecting black PV module backsheet. In addition, the alu- Based on the areas of exposed and covered UTC por-
minum frame was selected to be black. Such modifications tions of the BIPV/T system (Fig. 2a) and the assumption
yield an area-weighted (including framing) average module of uniform suction, the total mass flow rate from each area
normal solar absorptance of 92% as compared to 85% for is given by:
their traditional lighter coloured counterpart. As a conse-
mUTC exposed ¼ ðAUTC exposed  msuction Þ=3600 ð1Þ
quence, more thermal energy can be recovered from these
panels for ventilation air preheating. Another important mPV ¼ ðAUTC PV  msuction Þ=3600 ð2Þ
design parameter was the sizing of the photovoltaic mod-
ule. Although greatly dictated by the cell sizes available where the total mass flow rate of air for the UTC and
in the industry, a long narrow rectangular module was cho- BIPV/T systems is, respectively:
sen to reduce vertical temperature stratification of the air in mUTC ¼ ðAUTC  msuction Þ=3600 ð3Þ
the cavity between the PV panel and the UTC. This reduces
the PV operating temperature and facilitates flow of air mBIPV=T ¼ ðABIPV=T  msuction Þ=3600 ð4Þ
from behind the PV into the UTC. A custom designed
where
70 W polycrystalline module containing two rows of nine
solar cells and measuring 1465 mm  359 mm  38 mm ABIPV=T ¼ AUTC exposed þ AUTC PV ð5Þ
was used for this application.
The vertical and horizontal spacing between the panels The procedure to analyze the system is presented below
is both an application decision and a spatial configuration by using the experimental results.
constraint. For the prototype system, a major objective was The electricity produced by the PV modules (per unit
to facilitate optimization of the value of the generated area) as a function of their temperature TPV and total solar
energy, assuming that electricity is about four times as irradiance G is given by Skoplaki and Palyvos (2009) and
valuable as heat. Thus, a major purpose of the UTC is to Whitaker et al. (1991)
recover the heat produced by the PV modules while serving EPV ¼ G  ½gstc  ð1  bPV  ðT PV  T stc Þ ð6Þ
as a support system. In an effort to match the corrugations
of commercially available UTC, a vertical spacing of where bPV = 0.0046 (PV module temperature coefficient),
90 mm between the PV modules was selected. gstc = 0.125 (PV module efficiency at standard test condi-
The transpired collector used in the prototype is black tions (STC)), Tstc = 25 °C (PV module cell temperature at
galvanized steel (26 gauge) with a porosity of 0.6%. The STC).
plenum depth (distance between UTC and insulation) The total electricity produced by the BIPV/T system is:
was 0.15 m. The transpired collector sheets were installed EBIPV=T ¼ APV  EPV ð7Þ
with the corrugations running horizontally to facilitate
closing of a gap between the upper frame of the PV panel The thermal energy produced by the UTC and BIPV/T
(Fig. 1b) and the UTC, so as to reduce heat losses by nat- systems can be calculated as follows:
ural convection, while inducing turbulence behind the pan-
els and increasing mixing in comparison to vertically QUTC ¼ mUTC  cp UTC  ðT outlet UTC  T oÞ ð8Þ
oriented corrugations. Thus, airflow behind the PV mod- QBIPV=T ¼ mBIPV=T  cp BIPV=T  ðT outlet BIPV=T  T oÞ ð9Þ
ules is possible through the bottom and the sides. Finally,
a portion of the UTC was left uncovered at the top to pro- where the specific heat of air is given by (ASHRAE, 2009).
mote buoyancy driven cooling of both the plenum and the
modules in summer. cp ¼ 1005:7  0:066  ðT  27Þ ð10Þ
 Author's personal copy

146 A.K. Athienitis et al. / Solar Energy 85 (2011) 139–153

Since both UTC and BIPV/T systems are operating side

by side, the thermal energy produced by the exposed UTC
portion of the BIPV/T system can be estimated from the
results of the adjacent UTC system which are given by
the following ratio:
AUTC exposed
QUTC exposed ¼  QUTC ð11Þ
AUTC
Knowing the total thermal energy production of the
UTC system and the proportional contribution of the
uncovered UTC portion in the BIPV/T system, we can
determine the useful heat extracted from the PV modules
(QPV) by
QPV ¼ QBIPV=T  QUTC exposed ð12Þ

The thermal efficiencies of the UTC and BIPV/T sys-

tems are given by
QUTC
gUTC ¼ ð13Þ
AUTC  G
QBIPV=T
gthermal BIPV=T ¼ ð14Þ
ABIPV=T  G
The combined thermal and electrical efficiency of the
BIPV/T prototype is given by
QBIPV=T þ EBIPV=T
gBIPV=T ¼ ð15Þ
ABIPV=T  G
In the above equation, we may multiply E by a factor
COP (heat pump coefficient of performance) representing
conversion of electricity to heat to compute an equivalent
BIPV/T thermal efficiency as follows:
QBIPV=T þ COP  EBIPV=T
gBIPV=T thermal equiv ¼ ð16Þ
ABIPV=T  G
Fig. 3a provides a thermal network representation of the
heat transfer from the BIPV/T system portion covered by
PV. Thermal conductances are shown, including that asso-
ciated with air flow from outside, then behind the panels
and into the plenum behind the transpired collector
(mPVcp_avg). It is assumed that the PV module has negligi-
ble thermal resistance and no significant thermal capacity.
Liao et al. (2007) demonstrate that this assumption is
acceptable for BIPV/T systems with no added thermal
storage.
Experimental data (see Table 1) reveal that the temper- Fig. 3. (a) Thermal network of the heat transfer from the BIPV/T portion
covered by PV (thermal conductance values are shown); and (b) simplified
ature of the air in the cavity behind the module is almost
thermal network assuming that the cavity temperature is very close to
identical to the UTC surface temperature behind the PV. TUTC_PV.
At the lowest flow rate, the UTC cladding temperature is
found to be slightly higher than the air in the cavity
because of the radiated heat from the relatively hot PV surface. Thus, the simplified network in Fig. 3b can repre-
back side. As shown in Table 1, the temperature difference sent for design purposes the effective thermal conductance
between the air in the cavity and the UTC cladding is usu- from the PV to the air that flows into the plenum at
ally less than 0.5 °C. Therefore, very efficient heat transfer approximately a temperature equal to the transpired collec-
occurs between the cavity air and the UTC surface, signify- tor cladding temperature. Thus, the heat transfer coefficient
ing that hc2 is very high, eliminating the need to model a hc2 can be eliminated, resulting in the following equivalent
resistance between the cavity temperature and the UTC PV backside HTC (Fig. 3b).
 Author's personal copy

A.K. Athienitis et al. / Solar Energy 85 (2011) 139–153 147

Table 1 monitored under various flow rates. Such a setup allowed

Temperatures of the UTC surface covered by PV modules and of the air for performance comparisons under the same conditions.
cavity behind the PV module at various flow rates on April 24th, 2008
between noon and 13:00 pm (for BIPV/T system).
However, the analysis of the BIPV/T system by itself is
challenging due to the ever-changing environmental condi-
Flow rate (kg/h/m2) Temperature (°C)
tions such as wind speed, solar radiation and ambient air
UTC surface covered Air space between PV temperature. In addition, the flow rates were increased
by PV module module and UTC
every 30 min (see Table 2) in order to maximize the data
50 30.7 30.1 collected on a clear day. The time constant of the two facß-
80 29.6 29.1
115 29.6 29.6
ade systems was found to be about 3 min, so a quasi-steady
150 28.5 28.5 state was established for each flow rate. Therefore, obtain-
ing many data points that have close environmental condi-
tions is rare and normalizing techniques must be used to
Table 2 analyze the system. Unless specified otherwise, the results
Tested mass flow rates and the associated suction velocities, maximum reported in this paper are the average of the 30 min of data
plenum velocities and heat exchange effectiveness of the hole (see
collected (data was recorded every minute) for each flow
Appendix A for equations used).
rate (excluding initial transient of about 10 min), between
Mass flow rate (kg/h/m2) 50 80 115 150
1130 h and 1330 h local time. Table 3 presents the dates
Suction velocity (m/s) 0.011 0.019 0.026 0.033
Maximum plenum velocity (m/s) 0.172 0.289 0.409 0.517 and average environmental conditions that were selected
Hole heat exchange effectiveness (%) 75.5 64.9 57.6 52.8 for analysis.
The transpired solar collector is an open loop system in
which the heat collected is proportional to the rise in tem-
hi ¼ hr þ hc1 ð17Þ perature relative to ambient. Therefore, accurately measur-
ing the temperature of the air entering this type of collector
where hr1 is the radiative heat transfer coefficient and hc1 is
is crucial to predicting the system efficiency. However, this
the convective heat transfer coefficient.
can be challenging, especially when the ambient tempera-
The network to the right of TUTC_PV (Fig. 3a) has been
ture is subjected to local heating by the ground or
replaced, in the simplified network (Fig. 3b), with the ther-
neighbouring structures. This was the case for this experi-
mal energy removed from the PV modules (QPV) by the
mental setup, which was located on a conventional roof,
flowing air which has been calculated using Eq. (12).
Using the simplified thermal network, the effective PV
backside HTC (hi) can be estimated by writing an energy
balance at the node corresponding to TUTC_PV.
QPV  mPV  cp avg ðT o  T UTC PV Þ
hi ¼ ð18Þ
APV ðT PV  T UTC PV Þ
The calculated value of the effective PV backside HTC is
presented later in this paper. Although this calculation of hi
is not essential in the analysis of the BIPV/T prototype per-
formance, it provides valuable insight into the effect of flow
rate on heat extraction from the PV panels.

3. Experimental results

As explained above, two identical size systems, a UTC Fig. 4. Comparison of UTC efficiency using two different ambient air
and a BIPV/T system, were installed side by side and temperature sensor locations and the manufacturer’s published data.

Table 3
Dates and average environmental conditions selected for analysis.
Date Local time Mass flow rate per unit surface Incident total solar Wind speed Ambient
area (kg/h/m2) irradiance (W/m2) (m/s) temperature (°C)
April 15th (2009) 1:03 PM–1:28 PM 50 585 2.0 16.6
April 16th (2009) 11:32 AM–11:57 AM 80 570 1.7 15.8
April 16th (2009) 12:05 PM–12:30 PM 115 588 1.0 16.8
April 16th (2009) 12:37 PM–1:02 PM 150 587 1.9 16.3
July 15th (2009) 12:20 PM–12:45 PM 0 430 0.4 25.6
February 9th (2010) 11:27 AM–11:52 AM 50 930 0.5 4.2
February 11th (2010) 12:42 PM–1:07 PM 150 877 0.8 4.4
February 7th (2010) 11:33 AM–11:58 AM 0 926 1.5 2.7
 Author's personal copy

148 A.K. Athienitis et al. / Solar Energy 85 (2011) 139–153

where the ambient air temperature was found to be up to conditions and have an accuracy of ±3% based on the
6 °C warmer at the base of the BIPV/T installation than accuracy of the sensors used in the measurements.
at the weather station located about 10 m above. The effect The concept of equivalent thermal efficiency (Eq. (16))
of both of these ambient temperature sensors on the calcu- was introduced to compare the BIPV/T and UTC systems
lated UTC efficiency is demonstrated in Fig. 4. An effi- based on the quality of energy produced. This approach
ciency curve from the UTC manufacturer’s published assumes that one unit of electricity produced by the PV
data (SOLARWALL, 2009) is shown as a reference. For modules can be used to produce four units of thermal
the experimental results presented further in this paper, energy through a heat pump with a coefficient of perfor-
the ambient air temperature recorded at the BIPV/T base, mance (COP) equal to four – a reasonable estimate for
near the roof surface, was selected. This is the most conser- commercially available air source heat pumps (ASHRAE,
vative approach resulting in the lowest calculated efficiency 2009) at source temperatures above 0 °C. Fig. 6 compares
since it is the warmest ambient air temperature representing the measured UTC thermal efficiency, combined BIPV/T
air actually entering the collector. This does not affect the
main objective of the experiment which is relative compar-
ison of UTC and BIPV/T systems.
A major design goal of the BIPV/T system is to maxi-
mize the heat recovered from the PV modules using the
techniques discussed earlier in this paper. Two types of
PV modules, covering approximately the same area, were
tested under four flow rates and the total thermal energy
recovery of the system was calculated (Fig. 5). The large
PV modules have an aluminum frame with a white module
backsheet, whereas their smaller counterparts made for this
project have a black frame and backsheet. The experiments
were conducted between 12:00 h and 13:00 h on April 17th
2008 for the large modules and April 16th 2009 with the
small modules, under similar environmental conditions.
Fig. 5 shows that a 33% increase in thermal energy at the
high flow rate is achieved by using the small darker mod-
ules (PV covered UTC area is decreased by 2.5% and the Fig. 6. UTC thermal efficiency, combined BIPV/T efficiency (thermal plus
overall solar absorptance of the PV modules is increased electrical) and equivalent thermal BIPV/T efficiency (Eq. (16)) as a
by 7%).The results presented are taken under low wind function of mass flow rate.

BIP
PV/T WIT
TH L
LAR
RGE M
MOD
DUL
LES BIPV
V/T WIT
TH SM
MAL
LL C
CUST
TOM
M MO
ODU
ULES
S

Maass Floow B
BIPPV/T
T B
BIPV V/T
T
Raate Therrmaal E
T Enerrgyy [W
W] Thherrmaal Enerrgy [W
W]
[kgg/hrr/m
m²] Largge M
(L Moddulees) (Smmalll M
Moddulees)
1115 8660 11337
1550 8666 11551
Fig. 5. Comparison of the BIPV/T system thermal energy production using PV modules with an aluminum frame and white backsheet (left) to that with
specially designed PV modules with a black frame and back sheet (right).
 Author's personal copy

A.K. Athienitis et al. / Solar Energy 85 (2011) 139–153 149

efficiency (thermal plus electrical; Eq. (15)) and equivalent

thermal BIPV/T efficiency (Eq. (16)) as a function of mass
flow rate. Note that all estimates of heat recovery are con-
servative as they are based on the inlet air temperature at
the bottom of the two collectors as opposed to ambient
temperature. However, this is not important as the main
objective here is to compare UTC with BIPV/T system.
The combined efficiency of the BIPV/T system varies from
about 24–36% at the low and high flow rates respectively.
Following the trend of the UTC system, the BIPV/T sys- Fig. 7. Normalized parameter (Toutlet  To)/G as a function of mass flow
tem’s combined efficiency increases the most (6%) between rate.
flow rates of 50 kg/h/m2 and 80 kg/h/m2. The efficiency
gain beyond 80 kg/h/m2 diminishes and is negligible
between 115 kg/h/m2 and 150 kg/h/m2. This result will of T UTC ¼ T o þ G  ð1:157  104  msuction þ 3:094  102 Þ
course change with porosity and other UTC parameters. ð19Þ
Fig. 6 shows that the BIPV/T equivalent thermal effi-
ciency varies from about 50–63% at the low and high flow T BIPV=T ¼ T o þ G  ð4:687  105  msuction þ 1:381  102 Þ
rates, respectively. The BIPV/T equivalent thermal effi- ð20Þ
ciency is about 17% versus 7% higher than that of the
UTC at the low and high flow rate, respectively. This is Such linear relationships may be obtained from a few
because the electrical energy produced at all flow rates is measurements for any system during commissioning and
relatively constant, and in fact benefiting due to the used to optimally control the air flow rate.
increased heat removal, which in turn, when converted
into a thermal energy equivalent, represents a more 3.2. Increase in electricity production caused by cooling the
significant portion than the actual thermal energy at low PV
flow rates.
One could argue that the assumed average coefficient of The airflow behind the PV modules not only recovers
performance of four would change with source tempera- thermal energy but also increases the electrical efficiency
ture and also that there is a need to include fan energy of the modules due to lower operating temperatures. In
consumption to fully compare the two systems. The effect the winter season, the heat from the panels preheats the
of factor for conversion from electricity to heat will be incoming ventilation air. However, the heat may not be
studied further in a follow-up investigation with a full scale required in summer unless absorption or desiccant cooling
system. The fan energy consumption is equal for the UTC or domestic hot water preheating systems are implemented.
and BIPV/T systems, so it does not affect the relative com- Nevertheless, in some cases, the PV power production
parison. Finally, it should be noted that an annual energy increase due to cooling may equal or exceed the energy
analysis is required, taking into account whether the heat spent to operate the fan.
produced is useful throughout the year or only in the heat- The added benefit of cooling the panels is investigated
ing season. for the experimental BIPV/T system. The pressure drop
across the two experimental systems was measured for
3.1. Prediction of outlet air temperature for control of the tested range of flow rates and, mechanical and electrical
airflow efficiencies of the fan were assumed to be 50% and 85%,
respectively. In addition, the PV module’s temperature
It is well known that outlet (supply) air temperature coefficient and efficiency at standard conditions were exper-
from the UTC and BIPV/T systems is reduced with imentally determined to be 0.46% and 12.5%, respectively,
increasing flow rate, while thermal efficiency increases. from I–V curves generated by a photovoltaic I–V curve tra-
Depending on the application, there will be an optimal flow cer. Average PV module temperatures were recorded for
rate and air outlet temperature. both the no flow and high flow rate conditions and the
The outlet temperature of the experimental systems is results were normalized for solar radiation and ambient
mainly influenced by the wind speed, the total incident air temperatures. This normalized parameter (TPV  To)/
solar irradiance and the ambient air temperature. Normal- G was experimentally determined to be 0.033 °C m2/W
ization of the results for ambient air temperature and solar for the no flow conditions and for the tested flow rates
irradiation can be achieved by considering only low wind (see Fig. 8; both summer and winter data yield approxi-
speeds and studying the parameter (Toutlet  To)/G as a mately the same result). From this, the temperature of
function of the mass flow rate (Fig 7). The system outlet PV modules can be calculated at any chosen solar irradi-
temperature can be estimated using the following equations ance and compared to the no flow condition to determine
which were derived from a linear curve fit of the measured the increase in electricity production using Eq. (6). Fig. 9
data. shows that, at a flow rate of 150 kg/h/m2, the PV modules
 Author's personal copy

150 A.K. Athienitis et al. / Solar Energy 85 (2011) 139–153

to the HVAC system. However, it demonstrates the low

power increase that would be required when adding such
a BIPV/T design to the fresh air delivery system of an
HVAC installation.

3.3. PV backside heat transfer coefficient

As discussed earlier and illustrated in Fig. 3b, the effec-

tive PV module backside HTC, hi, consists of a convective
HTC in parallel with a radiative HTC. This parameter is
important for the optimal design of BIPV/T systems.
Fig. 10 shows how hi increases with mass flow rate. As
Fig. 8. Normalized parameter (TPV  To)/G as a function of mass flow rate. expected, hi increases with flow rate, leading to higher ther-
mal efficiency. As expected, the rate of increase of hi falls at
are cooled by 8.1 °C and 4.5 °C at solar irradiances of the highest flow rates. The effect of wind on ho and hi
900 W/m2 and 500 W/m2, respectively, resulting in an requires further study, particularly in full scale systems.
increase in electricity production of 3.8 W/m2 and 2.1 W/ Once a correlation for hi and ho is obtained as a function
m2 (a 90% efficient inverter is assumed). of wind conditions, it may be used to optimally select the
Fig. 9 presents the calculated fan energy consumption fraction of UTC covered by PV for given porosity, PV effi-
and the increase in PV electricity production as a function ciency and utilization of thermal energy produced.
of flow rate for two solar irradiance levels. The intersection
of the curves indicates the flow rate beyond which the fan 4. A BIPV/T demonstration project and discussion
energy is no longer recovered by the increase in PV power
production due to active cooling. The results show that the Building integration of the photovoltaic/thermal con-
energy spent to operate the fan under high solar irradiation cept presented in this paper can be most optimally done
conditions (900 W/m2) is exceeded by the resulting increase at the early design stage when the form of the building is
in PV electrical production at all flow rates. However, a being decided along with the location of the mechanical
breakeven mass flow rate exists (approximately 130 kg/h/ room with its ventilation and electrical service systems.
m2) at the lower solar irradiation conditions (500 W/m2), The two essential elements for an optimal design are
beyond which the energy consumed by the fan is no longer optimal orientation, proximity to the ventilation units
recovered by the increase in PV electricity production and integration with any other heat recovery and ventila-
alone. tion devices planned. However, retrofit applications are
This curve only presents the fan energy which would also possible with a new active facßade constructed over
draw air across the BIPV/T system, and not its delivery an existing one that may need to be renovated.

Fig. 9. Comparison of photovoltaic electricity production increase due to cooling with fan energy consumption (all parameters are per square meter of the
total area) for two solar irradiance levels (500 and 900 W/m2).
 Author's personal copy

A.K. Athienitis et al. / Solar Energy 85 (2011) 139–153 151

20

17
hi [W/m²/ C]

14

11

5
40 60 80 100 120 140 160

Mass Flow rate [kg/hr/m²]

Fig. 10. Effective heat transfer coefficient hi between a PV module (back

side) and UTC surface of the BIPV/T system (that is covered by the PV) as
a function of mass flow rate.

The ideas and results of the design concept presented in

this paper were used as a basis for the design of a BIPV/T
system for a new office building – that of the John Molson
School of Business (JMSB) at Concordia University
located in Montreal, latitude 45°N; Montreal has a cold cli-
mate with sunny winters that make this an appropriate
application for fresh air preheating and simultaneous
generation of electricity from one building surface. For
example, peak values of total incident solar irradiance
approaching 1100 W/m2 have been frequently measured
in winter for a near-south-facing facßade under clear cold
sky conditions. This first of its kind demonstration project
in a high visibility location was a unique opportunity for
such an application since the area utilized is the facßade of
the roof-top mechanical room of the 16-storey high build-
ing with a surface azimuth angle of 30°. This facßade will Fig. 11. (a) BIPV/T demonstration project in a new university building in
never be significantly shaded because municipal regulations Montreal (top floors), (b) photo showing clamps for attaching PV panels
do not allow higher buildings in this area of downtown to UTC (they were covered with black caps) and (c) location of the five
Montreal. inverter arrays on the BIPV/T system (array 5 may be connected to a test
This active BIPV/T system replaces the building enve- inverter).
lope of the near south-facing facßade of the mechanical
room, where its proximity to the fresh air intake reduces is under commissioning and its detailed performance will
the need to build long ducts from the facßade to the HVAC be presented in a follow-up study.
system. This 288 m2 installation utilizes 384 building-inte- Ideally, thermal energy produced in the summer should
grated PV modules (the same modules used in the proto- also be utilized for appropriate applications such as heating
type) connected to five inverters (Fig. 11). As in the water or for solar cooling systems. The BIPV/T system for
prototype, the UTC area covered by PV is about 70%. Spe- the JMSB building may also provide a model for retrofit
cial clamps were used for this project to attach the PV pan- applications where a new BIPV/T facßade can be con-
els so as to hinder solar-heated air from escaping from a structed over the existing facßade, thus reducing installation
gap between a module top frame and the UTC (see costs.
Fig. 1a and b). There are three outlets for the heated air
connected to an HVAC manifold and the top section of
the UTC is not covered with PV so that it can safely over- 5. Conclusion
heat and vent through natural ventilation the collected
summer heat (if not used). The system has a peak power The combination of unglazed transpired collector with
output of 24.5 kW, while also preheating up to about attached photovoltaic panels on facades is promising for
7.5 m3/s of fresh air for the building occupants. The project building applications where there is a significant need to
 Author's personal copy

152 A.K. Athienitis et al. / Solar Energy 85 (2011) 139–153

heat ventilation air in winter. A building-integrated photo- modules. Luis Candanedo, Jose Candanedo and Sam Sad-
voltaic/thermal system and unglazed transpired collector ighi also contributed to the experiments.
combination may be utilized to produce useful heat while
generating simultaneously electricity from the same sur- Appendix A. (equations used in Table 2 effectiveness
face. Based on this idea, this paper presents a design con- calculation)
cept and a prototype facßade-integrated photovoltaic/
thermal system with transpired collector (BIPV/T) tested Correlations developed by Dymond and Kutscher
in an outdoor experimental facility. The experimental pro- (1997) may be used to calculate the thermal energy transfer
totype was constructed with 70% of UTC area covered to the air flowing through the UTC.
with black-framed PV modules specially designed to The suction velocity is given by:
enhance solar energy absorption and heat recovery and msuction
was compared side by side with a UTC of the same area V suction ¼ ðA:1Þ
q  3600  hrs
under outdoor sunny conditions with low wind.
The orientation of the corrugations in the UTC is hori- where the density of air is calculated by ASHRAE (2009).
zontal and the PV modules are attached so as to hinder air q ¼ 1:1774  0:00359  ðT  27Þ ðA:2Þ
from escaping from the cavity formed by the panels and the
UTC, and instead flow into the UTC plenum. While ther- The Reynolds number based on the hole diameter is given
mal efficiency of the UTC system is higher than the BIPV/T by:
combined thermal plus electrical efficiency, the equivalent V suction  d
thermal efficiency of the BIPV/T system (assuming that Rehole ¼ ðA:3Þ
mhole  n
electricity can be converted to four times as much heat) is
7–17% higher. The assumption that electricity is worth The following empirical correlation was determined by
about four times the useful heat produced is based on typ- Dymond and Kutscher (1997) to predict the Nusselt num-
ical coefficients of performance of commercially available ber for heat transfer occurring on the front, the sides and
air source heat pumps which are at least equal to four the back of the UTC and is valid for porosities of 0.5–
for the range of temperatures considered. Connection of 5% and Rehole in the range 100–2000.
such heat pumps to solar-heated air as the heat source will "  0:48 #
p 1:21 V wind
enable expanded use of such BIPV/T systems over a wider Nuhole ¼ 2:75   Re0:43 þ 0:011  n  Rehole 
hole
range of outdoor temperatures for space heating in addi- d V suction
tion to fresh air preheating. ðA:4Þ
Finally, the concept was applied to a full scale office
building demonstration project in Montreal; this project For the present experiment, the hole diameter (d) is
is a near optimal application in an urban location due to 1.8 mm and the pitch (p) is 20.6 mm.
the fact that the mechanical room with the HVAC system In the BIPV/T system, for the area covered by PV, the
was located directly behind the BIPV/T facßade and the ori- velocity Vwind may be assumed to be zero.
entation was in the optimal range. The concept may be uti- The heat exchange effectiveness of the UTC is given by
lized both for retrofit and new buildings. Detailed design Dymond and Kutscher (1997) as follows:
and performance considerations of this project and similar  
ðn  1Þ  k hole  Nuhole
full scale systems will be reported in a future study that will Ehx ¼ 1  exp ðA:5Þ
d  qhole  cp hole  V suction
also consider the various parameters that need to be con-
sidered for such a system; these include instantaneous The more detailed relationships developed by Van
price/value of heat and electricity, need for fresh air pre- Decker et al. (2001) may also be utilized for higher accuracy.
heating, potential use of heated air outside the heating sea-
son and comprehensive optimization considerations. References

Acknowledgements ASHRAE, 2009. ASHRAE Handbook of Fundamentals. American Society

of Heating, Refrigerating and Air-Conditioning Engineers, Atlanta,
The authors acknowledge the financial support of Nat- GA, USA.
Bazilian, M.D., Leenders, F., Van Der Ree, B., Prasad, D., 2001. Photo-
ural Resources Canada through its CanmetENERGY Lab-
voltaic cogeneration in the built environment. Solar Energy 71 (1), 57–
oratory (Technology Early Action Measures Program) and 69.
the Natural Sciences and Engineering Research Council of Charron, R., Athienitis, A.K., 2006. Optimization of the performance of
Canada through the Solar Buildings Research Network. double-facades with integrated photovoltaic panels and motorized
Josef Ayoub of CanmetENERGY was instrumental in blinds. Solar Energy 80 (5), 482–491.
Chow, T.T., 2010. A review of photovoltaic/thermal hybrid solar
bringing together the partners for the demonstration pro-
technology. Applied Energy 87 (2), 365–379.
ject. Major industry partners in the demonstration project Coventry, J.S., Lovegrove, K., 2003. Development of an approach to
include Conserval who produced the UTC system and Day compare the ‘value’ of electric and thermal output from a domestic
4 Energy Inc. who produced and supplied the custom PV PV/thermal system. Solar Energy 75 (1), 63–72.
 Author's personal copy

A.K. Athienitis et al. / Solar Energy 85 (2011) 139–153 153

Delisle, V., Collins, M.R., 2007. PVT unglazed corrugated transpired solar Liao, L., Athienitis, A.K., Candanedo, L., Park, K.W., 2007. Numerical
collector model. Canadian Solar Buildings Conference, June 2007, and experimental study of heat transfer in a BIPV–thermal system.
Calgary, Canada. ASME Journal of Solar Engineering 129 (4), 423–430.
Dymond, C., Kutscher, C.F., 1997. Development of a flow distribution Skoplaki, E., Palyvos, J.A., 2009. On the temperature dependence of
and design model for transpired solar collectors. Solar Energy 60 (5), photovoltaic module electrical performance. A review of efficiency/
291–300. power correlations. Solar Energy 83 (5), 614–624.
Guiavarch, A., Peuportier, B., 2006. Photovoltaic collectors efficiency SOLARWALL, 2009. The SOLARWALL Solar Heating System. Conser-
according to their integration in buildings. Solar Energy 80 (1), 65–77. val Engineering Inc., Canada. <http://www.solarwall.com/> (accessed
Gunnewiek, L.H., Brundett, E., Hollands, K.G.T., 2002. Effect of wind in July 2009).
flow distribution in unglazed transpired plate collectors. Solar Energy Van Decker, G.W.E., Hollands, K.G.T., Brunger, A.P., 2001. Heat-
72 (4), 317–325. exchange relations for unglazed transpired solar collectors with
IEA, 2002. Reliability study of grid connected PV systems. In: IEA PV circular holes on a square or triangular pitch. Solar Energy 71 (1),
Power Systems Program Report No. PVPS T7-08. Task 7. 33–45.
Kutscher, C.F., 1994. Heat exchanger effectiveness and pressure drop for Wall, G., 1990. Exergy conversion in the Japanese society. Energy 15 (5),
air flow through perforated plates with and without crosswind. Journal 435–444.
of Heat Transfer 116, 391–399. Whitaker, C.M., Townsend, T.U., Wenger, H.J., Iliceto, A., Chimento,
Kutscher, C.F., Christensen, C., Barker, G., 1993. Unglazed transpired G., Paletta, F., 1991. Effects of irradiance and other factors on PV
solar collectors: heat loss theory. ASME Journal of Solar Engineering temperature coefficients. In: Conference Record of the 22nd IEEE
115 (3), 182–188. Photovoltaic Specialists Conference, vol. 1, pp. 608–613.