Testing Improved Solar Collector Performance Algorithms. With Solar System Design Software

Ricardo Aguiar*, Ricardo Coelho, Pedro Horta and Maria Joao Carvalho

INETI, Department of Renewable Energies, Lisbon, Portugal
* E-Mail: ricardo. aguiar@ineti. pt, Tel: +351 210 924 602


The influence of the incidence angle 0 on the performance of solar collectors has long been recognised and is accounted for in collector testing, as well as in solar system performance estimation and design, through the incidence angle modifier function, K(0). However, the usual approach taken to measure, describe and use it for computations has been pointed out as oversimplified, especially in the case of 2D collectors, such as CPCs or vacuum tubes with mirrors inside.

For flat plate collectors, standard practice assumes K(0) =[1+ b0 (1/cos 0)c ]+, measures the response of the solar collector at 0= 50°, assumes e. g. K(85°) = 0.85, and uses these data to find b0 and c. This curve is then used for computations, e. g. system performance and sizing.

However, for 2D collectors K(0) can take a non-monotonic shape and be larger than 1, meaning that optical performance is not maximum at perpendicular incidence. In addition, 2D collectors generally present an axis of symmetry, and their response is different in the longitudinal and transverse directions in respect to that axis. While this is generally not a problem when considering beam radiation (B), it complicates the treatment of diffuse radiation – both celestial (D) and ground reflected (R) – and leads to a case-by-case approach, according to collector geometry. Furthermore, this introduces bias when testing flat-plate and 2D collectors: the diffuse radiation absorbed by the different collector types is not taken into account at a levelled basis.

Thus, 2D collector performance comes out affected when the full scope of optical effects over different radiation components can’t be dully decoupled and a calculation based on the irradiation of the aperture plane is used: viz. the instantaneous power per unit area q is estimated as q = no K(0)G – a1 AT – a2 AT 2/G (with n0, au a2 the experimental coefficients, G the global radiation per unit area at the surface of the collector, and AT the temperature jump between hot water exiting the collector and the ambient).

Updated: July 15, 2015 — 12:40 pm