You are on page 1of 10

Solar Energy 79 (2005) 638–647

www.elsevier.com/locate/solener

Evaluation of CPC-collector designs for stand-alone,


roof- or wall installation
M. Adsten a, A. Helgesson b, B. Karlsson b,*

a
The Ångström Laboratory, Uppsala University, Box 534, S-751 21 Uppsala, Sweden
b
Vattenfall Utveckling AB, 814 26 Älvkarleby, Sweden

Received 25 November 2004; received in revised form 2 December 2004; accepted 26 April 2005
Available online 29 August 2005

Communicated by: Associate Editor Brian Norton

Abstract

An asymmetrically truncated non-tracking compound parabolic concentrator type collector design concept has been
developed. The collector type has a bi-facial absorber and is optimised for northern latitudes. The concept is based on a
general reflector form that is truncated to fit different installation conditions. In this paper collectors for stand-alone,
roof and wall mounting are studied. Prototypes of six different collectors have been built and outdoor tested. The eval-
uation gave high annual energy outputs for a roof mounted collector, 925 MJ/m2, and a stand-alone collector with Tef-
lon, 781 MJ/m2, at an operating temperature of Top = 75 °C. A special design for roofs facing east or west was also
investigated and gave an annual energy output of 349 (east) and 436 (west) MJ/m2 at Top = 75 °C. If a high solar frac-
tion over the year is the objective, a load adapted collector with a high output during spring/fall and a low output dur-
ing summer can be used. Such a collector had an output of 490 MJ/m2 at Top = 75 °C. Finally a concentrating collector
for wall mounting was evaluated with an estimated annual output of 194 MJ/m2 at Top = 75 °C. The concentrator
design concept can also be used for concentrators for PV-modules.
Ó 2005 Elsevier Ltd. All rights reserved.

Keywords: Asymmetric CPC collector; Collector performance/evaluation; Building integration

1. Introduction latter track is followed and an attempt is made to reduce


the investment cost of the solar collector without loosing
The benefit/cost ratio of solar energy can be im- too much in performance. The application of booster
proved in two ways; either by increasing the efficiency reflectors between the collector rows in a flat plate col-
or by decreasing the investment cost. In this work the lector field has previously been used to reduce the energy
cost and the required ground area for a given load (Per-
ers and Karlsson, 1993). A further development of this
concept has lead to a compound parabolic concentrator
*
Corresponding author. Tel.: +46 26 835 29; fax: +46 26 836 (CPC) collector with an internal reflector where a part of
70. the relatively expensive absorber area is replaced by
E-mail address: bjorn.karlsson@vattenfall.com (B. Karls- cheap reflector material. A bi-facial absorber in combi-
son). nation with a lower and/or upper reflector was used to

0038-092X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.solener.2005.04.023
M. Adsten et al. / Solar Energy 79 (2005) 638–647 639

Nomenclature

Ac aperture area [m2] T temperature [K]


b0 incidence angle modifier coefficient Tamb ambient temperature [T]
Ci area concentration Tin temperature at collector inlet [K]
cp heat capacity (water) [J/kg K] Tout temperature at collector outlet [K]
Gb beam radiation [W/m2] Top operating temperature, average of Tin and
Gd diffuse radiation [W/m2] Tout [K]
Kta incidence angle modifier U1 first order heat loss coefficient [W/m2 K]
(mC)e effective collector thermal capacitance [J/ V volume (water) during a 10 min interval [m3]
m2 K] g0b beam efficiency
P power [W/m2] g0d diffuse efficiency
Q energy output during a 10 min interval [MJ] q density (water) [kg/m3]
Rsol solar reflectance h angle of incidence
t time [s] ht transversal projected angle of incidence
Tsol solar transmittance hc acceptance half angle

minimise the absorber area. At high latitudes the annual simulation program to obtain an annual estimate of the
solar radiation is asymmetric over the year, implying thermal performance.
that the collector should be asymmetrically truncated
to maximise the yield. These conclusions have lead to
the development of a design concept with an asymmetri- 2. Measurement system and collector evaluation
cally truncated CPC-collector with a bi-facial absorber method
that does not need any tracking. Since the main objec-
tive was to maximise the reflector to absorber area for 2.1. Measurement system
a given ground area, this collector is called a Maximum
Reflector Collector, MaReCo. In addition to the aim of The measurement facility at the Älvkarleby Labora-
reducing the cost of solar energy the intention is also to tory has the capacity to evaluate up to 15 collectors at
indicate the possibility of special collector designs the same time. It has two main systems, one in which
adapted to various installation- and load conditions. the inlet temperature at each collector can be controlled
Several other studies of asymmetric concentrating in order to get a variety of operating conditions and one
collectors have been reported by, for example, Tripanag- system where all collectors are connected in series with a
nostopoulos et al. (1999, 2000), Norton et al. (1991), common flow, facilitating comparison measurements. In
Welford and Winston (1989), Mills and Giutronich the common flow system, which was used in these mea-
(1978) and Rabl (1976). A study of ultra flat concentra- surements, the water is cooled after each collector to get
tors suitable for building integration has been made by the same inlet temperature for all collectors. With this
Chaves and Collares-Pereira (2000). A collector type setup, where all collectors have the same water flow,
with similar reflector shape but a different absorber set- uncertainties in flow measurements have no affect on
up has been developed by Mills and Morrison (2001). the comparison measurements. The collector inlet and
Three large stand-alone ground mounted MaReCo sys- outlet temperatures were registered together with the
tems have been constructed and are described in Karls- flow of water/glycol. The ambient temperature and solar
son and Wilson (1999). radiation (global and diffuse) measured on a horizontal
In this study prototype MaReCos have been built surface and in the collector plane were recorded. The
and outdoor tested under various conditions: stand- radiation was also measured with a sun tracking pyra-
alone on the ground, wall mounted and finally roof nometer and a tracking pyrheliometer. The diffuse radi-
mounted with different tilt and orientation and also for ation was measured by using a shadow ring. All data
load adaption. The prototypes were designed for Stock- were sampled with a Campbell Scientific data logger
holm climatic conditions and tested at the Vattenfall each 10th second and the sampled values were stored
Laboratory in Älvkarleby (60.6° N, 17.4° E), Sweden. as 10 min mean values.
The results of these outdoor collector tests are described
in this paper. The collector parameters were determined 2.2. Collector evaluation method
with dynamic testing described in Perers (1993, 1995,
1997) and Perers and Walletun (1991). The experimental The collected data were then analysed with statistics-
parameters from the outdoor testing were then fed into a and calculation software. In the analysis the incidence
640 M. Adsten et al. / Solar Energy 79 (2005) 638–647

angle for beam radiation in the collector plane was cal- the measured output. Ideally the dots should form a
culated together with the monitored energy output from straight line y = x.
the prototypes. The collector parameters used to esti- In the next step of the evaluation, the expected an-
mate the annually delivered energy were determined nual energy output was estimated by using the experi-
with the dynamical testing model using Multiple Linear mentally derived collector parameters from Eq. (2)
Regression, MLR, on measured data. The method of obtained from the dynamic model in the simulation pro-
dynamical testing is further described in Perers (1993, gram MINSUN (Chant and Håkansson, 1985). In the
1997) and Perers and Walletun (1991). The dynamic test- MINSUN version used, the collector array model has
ing model has been compared to other test methods and been further developed with additional correction terms
is considered theoretically complete taking almost all and functions using experience from solar collector test-
effects into account (Nayak and Amer, 2000). ing (Perers, 1993). The MINSUN simulation program
Flow, temperatures and irradiation were measured was chosen to perform the calculations presented here
during a number of days with various combinations of because no knowledge about the system outside the col-
irradiation and collector temperatures. Sampling of the lector array is needed. Instead of detailed system infor-
studied parameters were made each 10th second. From mation, the program uses a fixed average operating
these data, the monitored energy output during a (heat carrier fluid) temperature ([Tin + Tout]/2). The
10 min interval, Q, was calculated according to well-defined operating conditions make a comparison
Q ¼ qVcp ðT out  T in Þ=Ac ð1Þ between different collectors more straightforward, since
no system effects are included. If the operating tempera-
The 10 min monitored energy was then converted to an ture is varying within relatively small limits, the system
average collector output power during the 10 min inter- can be satisfactorily simulated using the average operat-
val. A model according to Eq. (2) was then used to de- ing temperature.
scribe the collector output power, P, with irradiation Hourly average climate data for Stockholm, a refer-
and temperature as input, to obtain the collector param- ence year based on SMHI (Swedish Meteorological
eters with MLR. and Hydrological Institute) measurements 1983–1992,
P ¼ g0b K ta Gb þ g0d Gd  U 1 ððT out þ T in Þ=2  T amb Þ was used in the MINSUN simulations. Participants in
the International Energy Agency, Solar Heating and
 ðmCÞe dT =dt ð2Þ Cooling Programme, compiled this reference year. The
where year is characterised by annual global and diffuse irradi-
ation in the horizontal plane of 3.32 and 1.70 GJ/m2
K ta ¼ 1  b0 ð1= cos h  1Þ ð3Þ (922 and 471 kW h/m2), respectively. Stockholm is situ-
Kta is a factor that takes the incidence angle dependence ated at latitude 59° north and longitude 18° east.
into account (Duffie and Beckman, 1991) and b0 is the
incidence angle modifier coefficient. In this work only
the first order heat loss coefficient, U1, was identified. 3. Collector design concept
In order to check the model different diagrams are
drawn. In a modelled/measured diagram shown in To design the solar concentrator, knowledge about
Fig. 1, the modelled output is plotted as a function of the angular distribution of the annual solar radiation
is required. The solar radiation incident on a surface
can be projected into two components; one that is in a
plane defined by the concentrator axis and the normal
of the cover glass, the longitudinal plane, and one com-
ponent that is orthogonal to this plane, the transversal
Modelled power (W/m2)

600
plane, created by the orthogonal of the concentrator axis
500
in the aperture plane and the normal of the aperture
400 plane (Duffie and Beckman, 1991). The longitudinal
component of the solar radiation is parallel to the receiv-
300
ing surface and will therefore not contribute to the en-
200 ergy gain of the collector. The transversal component
100
of the solar radiation will contribute to the energy gain
if the projected transversal angle of incidence is within
100 200 300 400 500 600 700
the acceptance interval of the solar concentrator. The
Measured power (W/m2)
projected angle of incidence is called the transversal pro-
Fig. 1. Modelled power as a function of measured power in W/ jected angle of incidence, ht. The annual radiation is then
m2 for the roof mounted MaReCo. The solid line indicates a sorted according to projected angle of incidence and an
linear fit of the data. annual distribution of irradiation as a function of pro-
M. Adsten et al. / Solar Energy 79 (2005) 638–647 641

jected angle of incidence is obtained. This distribution is demand is high and a low efficiency during summer
then used to design the acceptance angles of the when the heating demand is low. The latter condition
concentrator. prevents over-heating in the system during summer. It
At high latitudes a large part of the projected solar is important that collectors for walls, east/west roofs
irradiation in the transversal plane is found around a and with low summer performance are cheap, since the
peak at the summer solstice. At lower latitudes two output per area unit will be low. The MaReCo collectors
peaks are found; one around the summer solstice and fulfill that requirement due to low materials content.
one at the winter solstice as shown by Rönnelid and The general form of the MaReCo reflector trough
Karlsson (1997). The loss of the winter solstice peak at (Fig. 2) is designed with two parabolas with their optical
high latitudes is due to large solar zenith angles for di- axes defining the lower and upper acceptance angles.
rect radiation during the winter months causing high The reflector consists of three parts. Part A is a lower
absorption of the direct radiation in the atmosphere side parabola extended between points 1 and 4 in
(Rönnelid and Karlsson, 1997). This is an advantage Fig. 2. This parabola has its optical axis along the upper
when designing an asymmetric CPC solar thermal col- acceptance angle and its focal point on point 5 (the top
lector since it can have a narrow acceptance half angle of the absorber). Part B is a circular part between points
covering the summer solstice peak, which leads to a high 1 and 2. This circular part transfers the light onto the
concentration factor. Thus a high fraction of the avail- rear side of the absorber. It replaces an absorber fin be-
able radiation is accepted. tween focus and point 2 (indicated by a dotted absorber
The maximum concentration factor of a two dimen- between points 2 and 5 in Fig. 2) with the rear side of the
sional non-truncated ideal CPC with an acceptance half absorber between point 1 and 5. The lower tip of the ab-
angle hc is given by (Welford and Winston, 1989): sorber can be placed anywhere between points 1 and 2.
1 Part C is a parabolic upper reflector between points 2
Ci ¼ ð4Þ and 3 in Fig. 2. This parabola has its optical axis along
sin hc
the lower acceptance angle and its focus at point 5.
This formula is, however, difficult to apply for an aniso- The position of the cover glass, the truncation, (i.e.
tropic light source, instead the ratio of reflector area to the position of points 3 and 4 along the extended parab-
aperture area is used. To design the collector the annual olas in Fig. 2) is determined by varying the position of
radiation distribution is used together with the bound- the reflector sheet along the extended parabolas to find
ary conditions: the time period during the year that the position where maximum annual irradiation, given
the concentrator is to operate and the limited length of by the annual radiation distribution, onto the aperture
the reflector in the north–south direction. An ideal is obtained. When designing the collector prototypes this
CPC concentrator is designed from the first criterion was achieved by sliding a reflector sheet of a certain
that gives the required acceptance angle. The concentra- width along the parabola/circle form shown in Fig. 2
tion factor is then given by Eq. (4). To accommodate the and measuring the distance between point 3 and 4 in
second limitation the concentrator is truncated to the Fig. 2, i.e. the width of the cover glass and the aperture
given reflector width. The truncation is made asymmet- tilt defined in Fig. 2. The glass width and the aperture
rical to adapt to the asymmetrical annual radiation dis- tilt angle were fed into the MINSUN energy output
tribution. The stand-alone MaReCo is designed in order
to obtain maximum annual output for a given reflector
width.
For some applications the collector depth, the maxi-
mum distance between the plane of the aperture and the
reflector trough, should be minimised to facilitate build-
ing integration. This limitation is found for the roof- and
wall mounted MaReCo. The reflectors are designed with
a circular part behind the absorber that goes over in a
parabola at a point, which is determined by the optical
axis of the reflector. The same design principle can also
be used to create designs for special conditions. In this
paper collectors designed for a south wall and east/west
facing roofs were evaluated since not all existing build- Fig. 2. Sketch of the basic MaReCo design. Part A is the lower
parabolic reflector extended from point 1–4, Part B is the
ings have roofs facing south. A load adapted collector
connecting circular reflector extended from point 1–2, Part C is
was also investigated. In this case a high solar fraction the upper parabolic reflector extended from point. The cover
in the heating system over the whole year is the objec- glass is found between points 3 and 4. The position of the cover
tive. The geometry of the collector is designed to have glass varies along the extended parabola depending on the
a high efficiency during spring and fall when the heating truncation. a is the aperture tilt.
642 M. Adsten et al. / Solar Energy 79 (2005) 638–647

simulation program together with the collector parame- along the reflector profile. The EPS also supported the
ters to calculate the expected annual delivered energy. reflector and determined the shape. The roof- and wall
The configuration with the highest annual output is MaReCos had no thermal insulation. Low iron glass
the optimum position of the reflector sheet in the parab- with normal incidence solar transmittance Tsol = 0.90
ola/circle shape. The non-symmetrical form of the an- (Hellström et al., 2000) was used in the cover glasses.
nual irradiation on a northern latitude means that the The collector parameters obtained from the evaluation
front reflector will be longer than the rear reflector. are found in Table 1 and are further discussed in the fol-
lowing sections. The low concentration and the use of a
bifacial absorber result in relatively low heat-loss factors
4. Description of solar collectors and presentation of even if the collectors are not insulated.
results The stand-alone MaReCo has a rather low zero loss
diffuse efficiency, g0d, as seen in Table 1. This is due to
In this work six different MaReCos have been tested. the high concentration factor, Ci = 2.2. The opposite is
The first two are designed for stand-alone ground found for the roof MaReCo with Ci = 1.5 and a higher
mounting and have an aperture tilt of 30°, facing south. g0d. The diffuse efficiency, g0d, is generally higher than
The concentrator shape is the same for the two, but one expected from the concentration factors and assump-
collector has a Teflon convection suppression film tions of an isotropic sky. This is explained by non-iso-
around the absorber. Three kinds of roof integrated tropic sky, which is brighter inside the acceptance
MaReCos were tested. One standard roof MaReCo de- interval of the collectors. The incidence angle modifier
sign for roofs facing south with a tilt of 30° was tested. coefficient, b0, of the stand-alone MaReCo without Tef-
The second type is for roofs facing east or west. This lon is very high, b0 = 0.51. The angular dependence
type has a tilt of 25°. The last roof model is load adapted model (Eq. (3)) used is not sufficient for taking the biax-
and the active area is reduced to 1/3 during summer. Fi- ial angular effects in the MaReCo into account. A more
nally an integrated vertical MaReCo for south facing sophisticated model is required. There is a difference in
walls was tested. High fin efficiency Sunstrip absorbers zero loss diffuse efficiency between the spring/fall and
(Wäckelgård and Hultmark, 1998) and anodised alu- the summer case for the load adapted collector. This
minium reflectors with solar reflectance Rsol = 0.85 could partially be explained by a high circum-solar radi-
(Hellström et al., 2000) were used in all collectors. The ation at high solar heights outside the acceptance inter-
stand-alone collectors were fitted with expanded poly- val during the summer period. The error in the model is
styrene, EPS, insulation of varying thickness (5–15 cm) described by R2, the coefficient of determination, a value

Table 1
Collector parameters for the evaluated MaReCos and a reference flat plate collector
MaReCo type g0b (–) g0d (–) b0 (–) F 0 U1 mCe R2 (–) Q25°, MJ Q50°, MJ Q75°, MJ
(W/m2 K) (J/m2 K) (kW h/m2) (kW h/m2) (kW h/m2)
Stand-alone 0.59 0.37 0.37 2.4 2980 0.9997 1256 911 670
(349) (253) (186)
Stand-alone 0.64 0.36 0.46 2.2 3380 0.9954 1325 1015 781
Teflon (368) (282) (217)
Roof integr. 0.69 0.56 0.29 2.4 1950 0.9994 1609 1209 925
(447) (336) (257)
East 0.58 0.25 0.13 2.0 6250 0.9996 691 486 349
West 0.60 0.35 0.16 2.0 4890 0.9994 (192) (135) (97)
889 626 436
(247) (174) (121)
Spring/fall 0.56 0.31 0.41 2.6 2230 0.9944 1058 716 490
0.34a 0.23a 0.23a 2.0a 3800a 0.9840a (294) (199) (136)
Wall 0.61 0.27 0.22 2.0 1130 0.9984 781 511 194
(217) (142) (54)
Reference 0.81 0.76 0.15 4.0 6000 – 2228 1156 1012
flat plate (619) (321) (281)
Estimated outputs at 25, 50, and 75 °C average operating temperature calculated with the MINSUN simulation program.
a
Evaluated for summer months.
M. Adsten et al. / Solar Energy 79 (2005) 638–647 643

1.0

0.8
Efficiency

0.6

0.4

0.2

0
0 0.01 0.02 0.03 0.04 0.05 0.06 Fig. 5. Section of the stand-alone MaReCo for Stockholm
[(Tout+Tin)/2-Tamb]/Itot [°C/(W/m2)] conditions. Aperture tilt 30°. Optical axes 20 and 65° defined
from the horizon.
Fig. 3. Experimental efficiency of a spring/fall roof MaReCo.
A least square linear fit is indicated by a solid line.
stand-alone-MaReCo (one with and one without a Tef-
lon convection suppression film around the absorber)
Roof, standard
were tested.
Stand alone Teflon The parameters from the collector evaluation are pre-
Stand alone sented in Table 1 together with estimated outputs for
0.7 Roof, spring/fall
Roof, west average collector operating temperature, Top = 25, 50
Wall and 75 °C. The stand-alone MaReCo without Teflon
0.6 Roof, spring/fall (summer)
has an annual energy output of 670 MJ/m2 at
0.5 Top = 75 °C. The stand-alone MaReCo equipped with
Teflon convection suppression has significantly lower
heat losses compared to the stand-alone without Teflon.
Efficiency

0.4
At Top = 75 °C the annual delivered energy output is
0.3 781 MJ/m2, 17% higher than for the collector without
Teflon. A previous evaluation of a stand-alone system
0.2 with extremely long collectors equipped with Teflon con-
vection suppression (Karlsson and Wilson, 1999) shows
0.1 a beam efficiency of 0.63 and a U-value of 1.7 W/m2 K.

0 4.2. The roof integrated MaReCo


0 0.02 0.04 0.06 0.08 0.1
([Tout+Tin]/2-Tamb)/Itot The standard roof integrated MaReCo is shown in
Fig. 6. This collector is designed by letting the cover
Fig. 4. Fitted efficiency for all six investigated MaReCos. glass start where the circular part of the MaReCo ends;
i.e. the glass is between point 2 and 4 in Fig. 2. No upper
between one and zero found in Table 1. For perfect cor- reflector is used and the inverted absorber is placed just
relation between model output and the simulated or underneath the cover. The whole design is then tilted to
measured values the R2-value is equal to one. the roof angle. All radiation up to the cover glass normal
An example of an efficiency diagram with experimen-
tal data is shown in Fig. 3 for the spring/fall roof MaRe-
Co. The least square fitted efficiency of all six evaluated
collectors during the test period is shown in Fig. 4. It
should be pointed out that the efficiency measurements
cannot be done during normal incidence for this collec-
tors, since this occur outside or close to the border of the
acceptance interval.

4.1. The stand-alone MaReCo

The stand-alone MaReCo for Stockholm conditions


is shown in Fig. 5. The upper acceptance angle is 65° Fig. 6. Section of the roof integrated MaReCo design for a roof
and the lower is 20°. Two similar prototypes of the angle of 30°. Optical axis 90° from the cover glass.
644 M. Adsten et al. / Solar Energy 79 (2005) 638–647

is accepted. The optical axis should be pointed at an


angle of 60° above the horizon. This means that the
geometry is determined by the roof angle. With a 30°-
roof tilt the area concentration Ci is 1.5. The collector
parameters are displayed in Table 1 together with the
simulated annual output. This MaReCo is characterised
by a rather high beam efficiency of 0.69. The roof
MaReCo shows a rather high annual energy output,
925 MJ/m2 at Top = 75 °C, only slightly lower than a flat
plate collector without Teflon (energy output at Top =
75 °C approximately 1012 MJ/m2 from Table 1). Fig. 8. Section of the east/west roof MaReCo designed for a
The specially designed roof MaReCo for east/west roof facing west. Optical axis 70° from the cover glass.
facing roofs has the concentrator axis placed in the
east/west direction, tilted along the roof as shown in
the photo in Fig. 7. The east/west MaReCo accepts radi-
ation in the interval 20 to 90° from the cover glass nor-
mal as seen in Fig. 8. This means that the collector does
not accept the hours of low irradiance before around
7.00 or after 17.00. The area concentration is 2.0.
The results of the evaluation are shown in Table 1.
The evaluation showed that with Stockholm climate
conditions it was more favourable to mount the collec-
tor facing west than east. The solar radiation is basically
symmetric around noon, but the direct fraction is higher
on the west roof. The ambient temperature is higher
after noon, which leads to lower heat losses. According Fig. 9. Section of the spring/fall MaReCo designed for a roof
to the evaluation the expected annual output is rather tilted 30°. Optical axis at 45° from the horizon.
low 349 (east) 436 (west) MJ/m2 at Top = 75 °C, which
means that the collector preferable should operate at pared to the ordinary roof MaReCo, as seen in Fig. 9.
lower temperatures. The beam efficiency is 0.58 (east) The absorber is placed just underneath the cover glass.
and 0.60 (west). Beam radiation hitting the reflector at an angle smaller
The load adapted roof integrated MaReCo, denoted than 15° from the aperture normal will be reflected out
the spring/fall MaReCo, has the optical axis tilted com- of the collector. The absorber is, however, working for
all angles of incidence, which reduces the concentrator
to a flat plate collector with an area equal to the single
sided absorber area for projected incidence angles smal-
ler than 15° to the aperture normal. An area concentra-
tion of Ci = 1.8 is found for the spring/fall MaReCo.
The results from the evaluation are found in Table 1.
The annual output is compiled from two simulations.
From May to August all radiation that hits the reflector
is reflected out of the collector and the only energy that
is utilised is the radiation that hits the absorber directly.
To obtain the annual energy gain per square meter the
spring/fall parameter values from Table 1 are used for
September to April and the summer output is taken
from a simulation using the parameters evaluated during
May–August from Table 1. As seen in Table 1 the beam
efficiency during summer (0.34) is about 60% of that
during spring and fall (0.56). The estimated annually
delivered energy is low, 490 MJ/m2 at Top = 75 °C.
The concept of a seasonally load adapted collector is
also used by another Swedish research group at the
Solar Energy Research Center, Dalarna University.
Fig. 7. Photo of a MaReCo designed for east/west-facing roofs. They use a similar reflector geometry to fulfill the
The white arrow indicates the south direction. requirements (Nordlander and Rönnelid, 2001).
M. Adsten et al. / Solar Energy 79 (2005) 638–647 645

April 10th May 30th July 19th September 7th ingly low cost. The annual load distribution follows
0.6
the delivered energy output distribution. The concentra-
tion factor is Ci = 2.2. The acceptance angle interval is
0.55 from 25° to 90° from the horizon as seen in Fig. 11.
The absorber is placed just underneath the cover glass.
The results of the evaluation are found in Table 1.
Optical efficiency

0.5
The annual energy output is rather low, only 194 MJ/
m2 at Top = 75 °C, and the beam efficiency is 0.61. A
0.45
vertical flat plate collector facing south with the collector
parameters from Table 1 has an annual energy output of
0.4 464 MJ/m2. This is partly explained by the too high opti-
cal axis at 25°, which means that valuable irradiation
0.35 will be reflected out of the collector. The optical axis
should be lowered to 10° and a corresponding concen-
tration ratio of Ci = 1.2.
0.3
100 150 200 250
Day no.
5. Discussion
Fig. 10. Optical efficiency for the spring/fall MaReCo for a
number of days between April 1st and September 20th.
The materiel cost for a stand-alone MaReCo includ-
ing the stand is summarized to 55 Euro and the invest-
ment cost for the large collector field previously
A long term study of the experimental optical effi- installed was approximately 155 Euro/m2. The goal is
ciency of the spring/fall MaReCo at different time peri- to be able to install the collector for 110 Euro in a large
ods is shown in Fig. 10 where the fitted efficiency is field. If it is assumed that the investment cost for a field
shown. The study shows that the efficiency is high for with large flat plate collectors is 180 Euro/m2 and that
September and April and then drops especially for the annual performance of a MaReCo is 75% of a flat
May–July. plate, then the MaReCo has to be installed for a cost
lower than 135 Euro in order to be competitive.
4.3. The wall MaReCo Efforts are made to further reduce the collector cost,
for example by using reflectors consisting of polymer
The wall MaReCo is an alternative for a vertical film with vacuum evaporated on a sheet steel (Brogren
installation due to is low material content and accord- et al., in press). This reflector has a reflectance of alu-
minium and a mechanical stability of sheet steel. This
means that the reflector construction and the stands
can be simplified.
The MaReCo concept can also be used in PV appli-
cations to reduce the cost of electricity produced with
solar cells. A stand-alone MaReCo-system with solar
cells laminated on the thermal absorber has recently
been installed on a flat roof in Stockholm (Karlsson
et al., 2003).
All tested collectors have low U-values. This is ex-
plained by the small absorber surface compared to the
total glazed area. An attempt was made in the evalua-
tion to find also the second order heat loss coefficients
(Duffie and Beckman, 1991) for the collectors, but the
MLR parameter fits were better if only the first order
loss coefficient, U1, was identified and thus a smaller
number of parameters. The effective thermal capacitance
is low in most cases. This is in part explained by the low
collector weight due to the low material content.
A more sophisticated model for the incidence angle
dependency must be used. In this evaluation only the
Fig. 11. Section of the wall MaReCo designed for a south incidence angle modifier, b0, was identified, which is
facing wall. Optical axis at 25° from the horizon. insufficient for very asymmetric collectors, such as the
646 M. Adsten et al. / Solar Energy 79 (2005) 638–647

MaReCo. In this study b0 values of up to 0.51 are found. ergy demand without problems with overheating during
This is because the optical efficiency is influenced by the the summer months. The spring/fall MaReCo is useful
performance in the transverse direction. An attempt to for obtaining high solar fractions on roofs with low tilt
solve this problem using a bi-axial incidence angle mod- angles.
ifier for the incidence angle dependency based on out- The low energy output for the wall MaReCo is partly
door test data is described in Helgesson and Karlsson explained by the unfavourable vertical alignment and
(2002). partly because of the relatively high lower acceptance
Because of the concentration the distribution of radi- angle (25° above the horizon). The heat losses of the wall
ation incident on the absorber is uneven. To get a high MaReCo are supposed to increase due to convection
energy yield from the collector an absorber with high caused by the absorber being positioned in the lowest
collector efficiency is required. A more detailed study part of the collector and hot air transfer of heat from
of the radiation distribution on the absorber for various the absorber to the glazing. The U-value is still surpris-
angles of incidence and absorber angles is required to ingly low for a non-insulated collector.
find a geometry with a more even radiation distribution,
making the collector less dependent on the fin efficiency.
The reflector shapes of the collectors studied here 6. Conclusions
were designed from a diagram showing the energy distri-
bution in various incidence angle intervals on a horizon- The MaReCo design concept is flexible and can be
tal surface from Rönnelid and Karlsson (1997). A adapted to various installation conditions. The aim of
further study of the radiation distribution on the actual the collector is to provide energy at a low cost, which
surfaces is required to investigate if these acceptance is achieved by replacing the expensive absorber, insula-
intervals are optimised. This study could also be ex- tion materials and box with cheap reflectors. A large
panded to investigate if the same concept can be used part of the available solar radiation can be collected
at lower latitudes. without the need for tracking, which would increase
Including a Teflon convection suppression film signif- the investment and maintenance costs. The design con-
icantly reduces the losses, especially at high tempera- cept can be refined through further studies of reflector
tures as seen in Fig. 4. It is, however, important to geometry, collector materials, incidence angle depen-
make sure that the Teflon film does not touch the absor- dence, and absorber alignment. The use of Teflon is
ber surface. recommended since the absorber temperature of a
In this study the standard roof mounted MaReCo MaReCo will be significantly higher than that of a flat
shows a high annual output, 925 MJ/m2 at Top = 75 °C. plate. The standard roof MaReCo and the stand-alone
The high performance of the roof-MaReCo is explained MaReCo are more cost effective or equally cost effec-
by the high optical efficiency (half of the absorber is fac- tive compared to the flat plate collector. The wall
ing the sun) and the low concentration factor. The ther- MaReCo and the special designs for east/west facing
mal losses are low due to the reduced convection with roofs and load adaption do not meet the requirement
the inverted absorber mounting. The same effect is also of cost effectiveness but are interesting niche products
found by for example Tripanagnostopoulos et al. and can compete with standard collectors in the same
(2000). This particular design can be mounted on roofs areas.
with inclination 30° and lower, which is an advantage
compared to flat plate collectors that have an optimum
tilt of 45° for Swedish conditions. Not many roofs have Acknowledgments
a tilt of 45° in Sweden, but rather 30° or lower.
The expected annual output of the east/west facing This work was financed by the Foundation for Stra-
MaReCo is rather low. The evaluation showed a large tegic Research (SSF) within the Energy Systems pro-
difference between an east/west MaReCo facing gram and Vattenfall Utveckling AB in co-operation
east and west. The solar radiation is symmetric around with the Swedish National Energy Administration in
noon, but the losses are in general smaller in the after- the solar heating RD&D program. Gunnar Wilson has
noon since the ambient temperature is higher in the contributed to the development of the collector design.
afternoon.
The spring/fall MaReCo is hard to make cost effec-
tive, since the material in it is the same as in the ordinary References
roof integrated MaReCo but the output is significantly
lower than for the ordinary roof integrated version. This Brogren, M., Helgesson, A., Karlsson, B., Nilsson, J., Roos, A.
collector can, however, be used if the aim rather is a high Optical properties, durability, and system aspects of a new
annual solar fraction than a commercial system. The aluminium-polymer-laminated steel reflector for solar con-
collector area is large to cover the major part of the en- centrators. Sol. Energy Mater. Sol. Cells, in press.
M. Adsten et al. / Solar Energy 79 (2005) 638–647 647

Chant, V.G., Håkansson, R., 1985. The MINSUN simulation Perers, B., 1993. Dynamic method for solar collector array
and optimisation program. Application and Users Guide. testing and evaluation with standard database and simula-
IEA SH&C Task VII, Ottawa. tion programs. J. Sol. Energy 50 (6), 517–526.
Chaves, J., Collares-Pereira, M., 2000. Ultra flat ideal concen- Perers, B., 1995. Optical modelling of solar collectors and
trators of high concentration. J. Sol. Energy 69, 269–281. booster reflectors under nonstationary conditions. Thesis
Duffie, J.A., Beckman, W.A., 1991. Solar Engineering of Uppsala University 1995.
Thermal Processes, second ed. Wiley Interscience, New Perers, B., 1997. An improved dynamic solar collector test
York. method for determination of non-linear optical and thermal
Helgesson, A., Karlsson, B., 2002. Incidence angle dependence characteristics with multiple regression. J. Sol. Energy 59
of an asymmetric collector. Presented at EuroSun in (4–6), 163–178.
Bologna. Perers, B., Karlsson, B., 1993. External reflectors for large solar
Hellström, B., Adsten, M., Nostell, P., Wäckelgård, E., collector arrays, simulation model and experimental results.
Karlsson, B., 2000. The impact of optical and thermal J. Sol. Energy 51 (5), 327–337.
properties on the performance of flat plate solar collectors. Perers, B., Walletun, H., 1991. Dynamic collector models for
In: Proceedings of Eurosun 2000, Copenhagen, Denmark. 1 h time step derived from measured outdoor data. In:
Karlsson, B., Wilson, G., 1999. MaReCo—a large asymmetric Arlen, M.E., Burley, M., Coleman, S.M.A. (Eds.), Proceed-
CPC for high latitudes. Presented at ISES World Solar ings ISES Solar World Congress, 19–23 August, Denver,
Congress 1999 Israel. USA, vol. 2. Pergamon Press, New York, pp. 1221–1226.
Karlsson, B., Helgesson, A., Svensson, L., 2003. Development Rabl, A., 1976. Comparison of solar collectors. J. Sol. Energy
of a MaReCo-hybrid. Contribution to ISES in Gothenburg. 18, 93–111.
Mills, D.R., Giutronich, J.E., 1978. Asymmetrical non-imaging Rönnelid, M., Karlsson, B., 1997. Irradiation distribution
cylindrical solar concentrators. J. Sol. Energy 20, 45–55. diagrams and their use for estimating collectable energy.
Mills, D.R., Morrison, G.L., 2001. Optimisation of minimum J. Sol. Energy 61, 191–201.
backup solar water heating system. In: Proceedings of ISES Tripanagnostopoulos, Y., Yianoulis, P., Papaefthimiou, S.,
2001 Solar World Conference, Adelaide, Australia. Souliotis, M., Nousia, Th., 1999. Cost effective asymmetric
Nayak, J.K., Amer, E.H., 2000. Experimental and theoretical CPC solar collectors. J. Renewable Energy 16, 628–631.
evaluation of dynamic test procedures for solar flat-plate Tripanagnostopoulos, Y., Yianoulis, P., Papaefthimiou, S.,
collectors. J. Sol. Energy 69 (5), 377–401. Zafeiratos, S., 2000. CPC solar collectors with flat bifacial
Nordlander, S., Rönnelid, M., 2001. Load adapted collectors absorbers. J. Sol. Energy 69, 191–203.
for high solar fractions. In: Proceedings of North Sun Wäckelgård, E., Hultmark, G., 1998. Industrially sputtered
Conference 6–8 May, Leiden, The Netherlands. solar absorber surface. Sol. Energy Mater. Sol. Cells 54,
Norton, B., Eames, P.C., Yadav, Y.P., 1991. Symmetric and 165–170.
asymmetric linear compound parabolic concentrators: the Welford, W.T., Winston, R., 1989. High Collection Non-
state-of-the-art in optical and thermophysical analysis. Int. imaging Optics. Academic Press, San Diego, ISBN 0-12-
J. Ambient Energy 12, 171–190. 742885-2.

You might also like