Page 85 - RSE - Results of the Apollon Project

Results of the APolloN PRoject ANd coNceNtRAtiNg PhotovoltAic PeRsPective

Shading losses. Only self-shading has been considered: fat horizon and no obstacles. For c-Si modules “shading by
strings with 50% electrical effect” has been assumed. For HCPV modules “linear shading” (loss proportional to
shaded area) has been assumed because they typically have a bypass diode for each cell. Note the difference in
losses between the dense and the sparse layout.
For Giza the shading losses are generally lower than for Ragusa because the average sun height above the horizon
is higher. For Tucson shading losses are similar to those for Ragusa.
2
temperature losses. For c-Si modules a NOCT of 45°C at Voc with a Uc=28.8 W/m K were defned in PVSYST. The
module has a Pmpp temperature coeffcient of -0.47%/°C. The tracked modules receive more light and thus heat
up more than the fxed modules.
2
For HCPV modules a NOCT of 75°C at Voc with a thermal exchange coeffcient Uc=13.1 W/m K were defned in
PVSYST. The module has a Pmpp temperature coeffcient of -0.04%/°C. The temperature losses for the HCPV
modules are around 3 times lower than for c-Si.
Infuence of wind-speeds on cell temperature was not used.
Module Mismatch losses. 2.1% was assumed for c-Si (PVSYST default value). For HCPV 0.5% was used, assuming
application of the technology developed in the APOLLON project to reduce such losses (use of CPV MPPT devices).
Soiling losses. 2.5% has been assumed for c-Si; 3.5% for HCPV.
Spectrum losses. These exist only for HCPV and are mainly due to the variation of the effciency with Air Mass (AM),
DNI-level and Ambient Temperature (Tamb). These variations are due to changes in the spectrum and to some
structural deformations in the optical parts. They are diffcult to assess. In the simulations the infuence of AM
was defned to be 1 below AM1.7, decreasing linearly to 0.58 at AM8. The DNI-level infuence was defned as 1
2
above 600 W/m , decreasing linearly to 0.9 at DNI=0. Infuence of Tamb was not used.
In the sparse layout the modules receive more light (less shading) but all of this light is gained when the sun
is close to the horizon (high AM), thus increasing average spectrum losses. In Giza the losses are slightly lower
than in Ragusa because the average higher sun gives an average lower Air Mass. In Tucson losses are even lower,
because the average DNI-level is higher.
Comparing the AC-energy production between Ragusa and giza (Table 16 ) the energy increase is lower for HCPV
(3,363/2,884=16.6%) than for (fxed) c-Si (2,371/2,007=18.1%), notwithstanding the higher temperature losses for
c-Si. This can be explained by comparing the amounts of useful incident energy (see Table 14): +18.7% for Global
Inclined, +18.0% for Global Normal and +15.4% for Direct Normal. In Giza the percentage of diffuse is higher than
in Ragusa and as a consequence DNI increases less than GNI.
Comparing the AC-energy production between Ragusa and tucson (Table 16) the energy increase is much higher
for HCPV (3,697/2,884=28.2%) than for (fxed) c-Si (2,285/2,007=13.8%). Again, the explanation can be found in
the amounts of useful incident energy (see Table 14): +12.8% for Global Inclined, +15.9% for Global Normal and
+26.9% for Direct Normal. In Tucson the percentage of diffuse is far less than in Ragusa (and Giza) and as a
consequence DNI increases more than GNI.
Hence, when comparing locations suitable for HCPV, in addition to the DNI-level, the percentage of diffuse light
should be checked as well.

Fixed Pv, tracked Pv and hCPv, installation Costs Comparison

This chapter deals with the installation Costs and the LCOE-calculation.
From insider information, the costs in Romania for the turn-key installation of 1 MWp fxed c-Si systems are
currently around 1.0-1.1 euros/Wp. These systems are selling there from 1.4-1.7 euros/Wp. According to a report of
49
L.B.N.L. of July 2013 most prices for ground-mounted fxed c-Si systems >2 MWp (utility-scale), sold in the U.S. in 2012,
ranged from 2.5 dollars/Wp to 4.0 dollars/Wp. There are indications that prices will continue to go down in coming
years. Costs for such systems may also be slightly lower in Europe than in the US. For LCOE-cost comparison in this
document, 2 installed system cost levels for fxed c-Si have been considered: 1.40 euros/Wp and 1.60 euros/Wp.
Tracked c-Si needs a (low-precision) tracker. According to note 46, the costs for a tracked c-Si system exceed
those for a fxed system by 0.3 dollars/Wp to 0.5 dollars/Wp. For LCOE-cost comparison in this document, 2 installed
system cost levels for tracked c-Si have been considered: 1.70 euros/Wp and 2.00 euros/Wp.
For the installation costs of HCPV three different principles have been considered: a system based on fresnel
lenses, a beam splitter system with dichroic mirrors, and a system with off-axis mirrors.

49 http://emp.lbl.gov/sites/all/fles/lbnl-6350e.pdf.

84