Why?    Applications    Other dishes

Despite many years of investigation and some promising initiatives, dish systems are still the smallest and least mature segment of the CSP market at present. However there are some compelling arguments as to why they have good future potential and some understandable reasons why they have not yet achieved it.

Compared to the other concentrator approaches, dishes offer advantages including:

• The highest optical efficiency
• The highest optical concentration and hence the highest thermal efficiency for a given operating temperature
• The ability to reach high temperatures for advanced chemical processes
• A modular product allowing systems of any size to be assembled from multiples of standard units
• The ability to carry out development work on single units at relatively modest expense

As long as construction methods can be found to produce dishes at similar costs per unit area as their competitors, then they offer great potential. The reason that widespread deployment has not occurred so far can be that  CSP as an industry is still at a stage that all options have not been fully explored. To date the bulk of the effort on dish systems has been on coupled dish plus receiver mounted Stirling engines. This particular configuration has struggled in the absense of an energy storage solution and in the face of aggressive cost competition from photovoltaics.

However with many other promising energy conversion applications possible,  the dish sector offers great potential for investors.

Examining efficiencies

Dish optical efficiency is considerably higher than the trough, Fresnel or tower systems because the mirror is always pointed directly at the sun, whereas the trough, Fresnel and tower suffer from a reduction in projected area due to a frequent low angle of incidence of the solar radiation (so called cosine losses).

Being a point focus solution means that dishes have inherently higher levels of concentration that the linear trough or Fresnel systems. In comparison to tower systems, they have an advantage in concentration via the potential for approximating more closely the ideal optics of a paraboloid. Higher concentration means that receivers have smaller aperture areas and this then means that for any given operating temperature, the thermal losses from re-radiation and convection are reduced.

To make the performance argument in a quantitative manner, results from established validated models of the various system types can be compared. A well respected and freely available approach to modelling performance is the US National Renewable Energy Laboratory’s “System Advisor Model” (SAM) (https://sam.nrel.gov/). Using some of the available CSP system models in SAM to examine operation over a typical year at using solar data for Barstow California gives the following results:

The tower, trough, Fresnel and dish Stirling systems modelled all have no storage and a Solar Mutiple of 1 to provide a most direct comparison of the relative solar field performance. The results are annual average efficiencies, design point efficiencies are higher in each case. On the basis of this comparison, a dish system has a clear performance advantage that arises from its higher optical efficiency.

Relative costs

The performance advantages are widely acknowledged, the question that is not settled however is if the added complexity of a dish means that its cost per unit area will be too high to justify the performance advantage. Every CSP technology developer will do their best to produce a system at lowest cost. It is hard to draw general conclusions when comparing specific offerings from specific companies. A more fundamental approach to the argument is to look at the various sub-components that make up a concentrator and estimate the fundamental differences between the concentrator types on the assumption that they were all cost optimised on a common basis. The following qualitative assessment by the author is offered in this regard:

A dish with its added complexity is estimated to  cost 46% more to construct per unit area of aperture than a trough. However, dividing by their relative system efficiencies and re-normalising suggests that a Dish field may still be able to produce power for a cost of energy 15% less than a trough field and arguably the lowest of all the CSP options.

The estimated cost of energy differences between the CSP types on this basis are not huge. Thus other specific design factors can be the major determinant. At present the most significant driver in this regard is the cost of energy storage. Tower systems with molten salt out perform trough systems economically because their higher operating temperature allows them to use an equivalent salt inventory to store nearly three times as much energy. For dishes to become a major force in CSP  a storage solution equally effective as salt on a tower needs to be proven commercially.

Optimum size

The development of large dishes at ANU was based on the premise that large dishes are more cost effective per unit aperture area than small ones. At the begining of the project to design and build a generation II Big Dish, the ANU team revisited the justification for this. Calculating a true optimum size requires consideration of how all the individual cost elements scale with size.

The cost of a dish will be made up of contributions from the various parts of it. Each major component will itself have a fixed cost component and a variable cost component that will have some functional dependence on dish size. Typically cost contributions will have components proportional to R, R²  and R³. An obvious R² contribution comes from the cost of mirror panels which are essentially and area dependant cost and so have no impact on the optimum size. The major R³ cost contribution comes from the dish structure. This arises because structure cost is largely about resisting wind loads. The force is proportional to the area (R²) but it also acts to create a moment that must be resisted and this introduces an extra factor of R. So the optimisation is essentially a trade off between fixed and R proportional costs (control systems, actuators, motors, foundations etc), which encourage a larger dish  and the R³ structure costs which favour a smaller dish.

The bold curve corresponds to the best estimate of R³ dependance for the basic ANU dish design.

Most dishes built are however smaller than the ANU Big Dish. The reason is partly explained by the extra challenge that design and prototype construction represents. In addition this determination does not take into account the specific needs of a pre-existing receiver or engine (eg Stirling) unit which has been a driver for many dish systems.  There is also a second order effect that for small systems or smaller early production runs, a smaller dish design means that larger numbers of smaller components can be fabricated or purchased and so benefit sooner from cost efficiencies in manufacture.

In conclusion

Dish systems have a real performance advantage that appears to outweigh potential complexity based construction cost increases. There is a strong case for continued further investigation of dish systems. Proving cost effective energy storage systems must be a high priority. There is also a strong case that larger dishes will be more cost effective than small ones for utility scale (eg more than 10 – 20MW) CSP applications.