13.2.1 Classical concentrators
The progress in solar cells development since its invention has been huge. The investment needed has also been important. Present cells have become much more efficient and the technology much more advanced. The next-generation approaches discussed in this book try to further cell development, increasing cell sophistication if necessary.
However, PV concentrator development has not been accompanied by such progress. This can be seen in the fact that most PV concentrator systems have been based either in the parabola or in the Fresnel lens (see figure 13.11): the parabola has been known since Menaechmus (380-320 BC) and its optical properties since Apolonius (262-190 BC), and Galileo (1564-1642) used lenses for making the first refractive telescope. The thinning of such lenses to avoid
 |
 |
Figure 13.11. Fresnel lenses and parabolic reflectors are the classical concentrators in PVs.
absorption was proposed by Georges de Buffon in 1748 (and later optimized by Fresnel in 1822).
The parabola and Fresnel lens perform far from the theoretical concentration-acceptance angle limits. The question is: Are they good enough or do we need to come closer to the limit? As discussed in section 13.1.4, we guess that they are not good enough, especially due to the present high concentration trend.
Non-imaging optics came into the PV field as an option for low (static) concentration [31] or as a CPC-type secondary optical element for higher concentration PVs [32]. The first full non-imaging device for non-static concentration was a non-imaging linear Fresnel lens with a curved aperture [33].
Among all these non-imaging systems, only the addition of the CPC-type secondary to a classical primary could work for very high concentration systems (>1000x) with a sufficient acceptance angle (a90 > 1°) [26]. However, apart from the lack of illumination homogeneity and the practical problems related to flow-line mirrors, this solution is necessarily non-compact (depth H to entry aperture diameter D ratio (H/D) & 1.5 for the Fresnel lens and H/D & 1 for the parabolic mirror). This is due to the design approach, in which the secondary and primary are designed separately and, thus, the secondary does not consider the actual rays exiting the primary but considers the primary as a Lambertian source [34]. The smaller the H/D ratio is, the lower accuracy of this approximation will be and, thus, the concentration-acceptance angle product will be reduced.
