Reflector construction
Another critical point for the outcome of the project was the reflector manufacturing. Despite the low geometrical concentration ratio of the device (about 11), due to the large focus distance, the optical precision required it is similar to parabolic troughs that work with concentration ratios several times higher. On the other hand, in order to integrate the collector in building roofs, weight must be kept as low as possible.
For the reflecting surface, thick glass mirrors with permanent curvature were discarded almost from the beginning due to their high weight, high production costs for small manufacturing series and the low curvatures required. Therefore the two main alternatives considered were:
1. Glass on Metal Laminates (GOML)
2. Solar grade specular aluminium sheets
 |
Eventually the second option was chosen mainly because of weight considerations. Aluminium sheets can be used in sandwich structures filled with polyurethane foam. In this process polyurethane is injected between two metal sheets inside a mould and the pressure due to expansion of the foam forces the metal sheets to take the shape of the mould (Fig. 6).
 |
With this method small reflector pieces (1 x 1.5 m approx.) can be easily produced (Fig 7.a) and then this pieces are assembled to the reflector frame (see Fig. 5. and 7.b) to form the complete reflector of 4.5 x 6 m.
7.a)
An accurate control of the reaction conditions of the foam is of critical importance for the optical quality of the obtained surface because incomplete reactions lead to surface distortions. Nevertheless, under controlled conditions the optical quality obtained is excellent. The figure 8 shows a comparison between a theoretical radiation distribution on the focus, estimated with ray tracing, for a prefect geometry and assuming a material dispersion of о = 7.5 mrad, and the experimental curve obtained from a photograph of the focus for normal incidence.
 |
Despite the importance of this component for the performance of the collector, at the beginning of the project it was decided to try to use market available solutions whenever possible, instead of developing a new component.
Nevertheless, as can be seen in figure 2, at different times of the day the solar radiation reaches the receiver from a different direction. Therefore either the absorbent surface should be cylindrical or it should be oriented to the centre of the reflector for each incident angle.
The orientation of the receiver surface could be an interesting option, particularly if the concentrator were to be used in combined thermal/PV applications, but it was decided not to implement it in the first prototypes in order to reduce the amount of mechanisms.
Therefore a receiver with a cylindrical absorber with a diameter between 45 and 55 mm was required. The only type of absorber available with this geometry were the Sydney evacuated collectors. In this kind of tubes, the heat is primarily absorbed by the inner glass layer and then transferred to a U shaped copper tube or to a “heatpipe” through an aluminium fin. In the figure 9.a a section of the tube with a U-pipe is shown.
The main advantages of this kind of collector are its low cost and a very good sealing of the evacuated volume. On the other hand, the low conductivity of the glass absorber requires high temperature increments in order to evacuate the incident radiation. This problem is increased by the fact that it is very difficult to ensure a perfect contact between the glass and the aluminium fin, and between the aluminium fin and the copper tube. In the figure 9.b the simulated temperature distribution is shown for the case of the concentrated radiation reaching the tube laterally. In this model, only conduction between the different elements has been taken into account, and an air filled gap of 0.1 mm is assumed at both the glass-aluminium and the aluminium-copper interfaces. Although a more complex model would be required in order to accurately predict the absolute temperature values, the results obtained are representative of the relative temperature gradients between the glass zone and the metal parts.
Those relatively large temperature gradients can produce a low efficiency of the collector and breaking of the glass tube due to thermal stresses. Nevertheless it is difficult to theoretically predict those problems because they are highly sensitive to factors, such as the perfect contact between the different parts, that are very difficult to control. Therefore it was decided to experimentally evaluate the collectors for normal incidence (Fig. 10).
For an input fluid at a ambient temperature (21°) and a normal direct radiation of 750 W/m2, the efficiency obtained for the reflector-receiver system was about 70 %, and the stagnation
 |
 |
temperature of the tubes was 297 °С. Due to time issues no other efficiency measurements were carried out, and the obtained values were considered to be sufficient for the first prototype.
Fig. 10. Experimental setting
Regarding to the breaking of the tubes, during the stagnation tests no breaking of the tubes was observed. During the efficiency tests, only two failures were observed, and both of them were produced during manipulations or sudden changes in the fluid regime that could be easily avoided during normal operation. Therefore it was decided to use standard Sydney evacuated tubes for the receiver of the collector.
 |
Nevertheless, during the first tests of the first collector prototype several tube failures have been observed. Most of them have been originated at either one of the tube tips (Fig. 11). Although a complete study of the causes of the breaking is still not completed, it is likely that most of the failures were related to the thermal stresses induced at the tube tips. Whether it is possible to reduce those thermal stresses with an improved fin design is still under analysis.
