Increasing the efficiency of small scale CSP applications

An Interview with Dr Eduardo Zarza

Small scale and modular CSP projects (less than 10 MW) offer a whole new potential market for CSP investment. Modular CSP has a number of significant benefits: materials can be pre-packaged and assembled on site reducing labour requirements; it is more adaptable to different terrains than large scale CSP projects; it doesn’t have the same permitting headaches associated with large scale CSP projects and, relatively speaking, the upfront investment is lower meaning that financing should be a more straight forward matter. And as industrial applications, such as mining and enhanced oil recovery, become more of an area of focus for CSP, modular CSP is likely to increase in traction.
However, there are a number of concerns facing small-scale CSP, primarily in terms of the economies of scale and lower efficiency levels when compared to large scale CSP plants.


Research by Dr Eduardo Zarza, Head of the CSP Research Unit at Plataforma Solar de Almería (PSA) has shown that replacing synthetic oil with Carbon Dioxide (CO2) as a working fluid can significantly improve the efficiency of small scale CSP projects.

CSP Today Speaks to Dr Zarza about PSA research into Carbon Dioxide as a heat transfer fluid in small scale applications of CSP.


CSP Today: Please tell me about PSA research into Carbon Dioxide as a heat transfer fluid in a CSP plant.

EZ: Our research at PSA has focused primarily on small scale applications of CSP whereby the overall efficiency of the project is improved by replacing the thermal oil with CO2.

With CO2 we could implement a Brayton Cycle with high efficiency where a 5 MW plant withCO2 could have similar efficiency to a 50 MW CSP plant with oil as a transfer fluid. This to me is the biggest advantage. Using CO2 we could implement a Brayton cycle with a top temperature of 500˚C giving us a power block gross efficiency of about 40% - which is a similar efficiency as the power block of a 50 MW thermal oil plant. The parameters of the Brayton cycle required for this application have been defined and the pre-feasibility study of such a Brayton cycle have been very encouraging. However, a significant research and development effort would be needed to develop at a commercial level the components of this Brayton cycle.
If we implement a 5 MW CSP plant using thermal oil rather than CO2 as a transfer fluid then the maximum efficiency it could hope to achieve is less than 30%. With CO2 and the analyzed innovative Brayton cycle we could meet market needs of small power plants with efficiency similar to that of larger scale CSP plants.

CSP Today: Has this been considered only for modular CSP plants?

EZ: Yes. The main problem for CO2 is the high pressure loss in the solar field piping. If we have a large field with 50 or 100 km of receiver pipes, the pressure loss with CO2 would be unaffordable. This means CO2 is specifically suited to small scale CSP plants. Concerning the impact on central receivers we have found with a theoretical study that the coupling of the new Brayton cycle to a central receiver using supercritical CO2 could also led to a high efficiency for plants within the power range of 2-5 MWe. However, we have only done experimental research on a parabolic trough plant.

CSP Today: What is the impact of using CO2 on the cost of a CSP plant? 

EZ: For a small plant (less than 10 MW), I think the CAPEX will be slightly higher, but the efficiency would be significantly higher, so there would be a clear benefit – the final cost of electricity would be lower.
The maintenance cost would be similar to that of a plant using oil – the main benefit is the higher efficiency.

CSP Today: Is there an operational CSP plant using CO2 as an HTF?
EZ: Not on a commercial scale – there are only some small test facilities. At the PSA we have a test facility with a nominal thermal output of 400 kW. We have already achieved 510ºC in a parabolic trough system – I think this may be the highest in the world using a compressed gas in the receiver pipes. We have proven that carbon dioxide can be used in a parabolic trough collector to produce thermal energy at 500ºC. With such a high temperature the efficiency of such a high plant would be about 40% when used in the innovative Brayton Cycle we have designed.

CSP Today: Why a Brayton Cycle as opposed to a Rankine Cycle?
EZ: A Rankine cycle needs higher power to achieve good efficiency, because of the limits imposed by small steam turbines. With the new Brayton cycle we can achieve a much higher efficiency at low power. Brayton cycle is also more user-friendly. The engine of an aircraft is a Brayton cycle: they have a turbine which expands the hot air which gives them propulsion – these engines can easily start and shutdown. In a Rankine cycle you need steam turbines, pump, cooling systems, deaerators and so forth. You cannot start it in ten minutes – you need at least half an hour.

CSP Today: Is a Rankine cycle more costly to use than a Brayton cycle, or does the cost depend on the scale of the plant?

EZ: I would say they are quite similar.

CSP Today: Are there any other areas you are researching with regards to CSP plant efficiency and cost reduction?
EZ: There are many possibilities. ESTELA (the European Solar Electricity Association) has just published their strategic research agenda which includes topics such as better receiver pipes, improved thermal storage systems, new working fluids and so on. PSA, DLR and other research institutes have collaborated to produce this agenda. The list of potential areas for improvement is large. However, there is not a single area which will lead to overall improvement. Rather the solution lies in the accumulation of small improvements in many items – no one item could achieve a 30% LCOE reduction. Better receiver pipes with lower maintenance for parabolic troughs could be an improvement. Direct steam generation can be also a significant step forward as replacing thermal oil with water would lower the maintenance cost because of fire risks and the auxiliary systems required by oil systems. There are also companies investigating the technical features of using molten salts as a working fluid.

Using molten salt would result in a lower maintenance cost because the thermal storage medium used in parabolic trough plants these days is molten salt. If molten salt is used in the solar field the working fluid and the storage medium would be the same element – we wouldn’t need heat exchangers in between. This is why many research and development and private companies are researching the technical capabilities of molten salt in trough collectors. The main technical barrier is the high freezing point of 240ºC so at night time the plant operator needs to be very careful to avoid freezing. A salt plug in the solar field piping could be a disaster. There are also companies trying to find new salts with a lower freezing point to avoid these maintenance problems.

There are also companies which are researching increasing the maximum working temperature of thermal oils. At the moment the maximum temperature is 398˚C – above that and the oil starts to degrade. This degradation increases the maintenance costs. This maximum temperature is a significant barrier – the higher the temp in a Rankine cycle the higher the efficiency. Efficiency is therefore limited by the working temperature nowadays.

CSP Today: Are there other areas you are aware of that can reduce the cost of a CSP plant?
EZ: New working fluids need to be found to improve the CSP technology as they allow higher temperatures and as a result efficiencies would be higher. Efficiency and cost move in parallel. Any item improving efficiency will reduce the energy cost.
Another area is the reduction of man power – in civil works and assembly of components. If we develop automatic procedures – we could significantly reduce the CAPEX of a CSP plant.

CSP Today: Is labour reduction affected by the size of the plant?
EZ: There is a clear scale up effect – the bigger the plant the lower the investment cost per KW.
Nowadays we already have parabolic trough collector designs with a significant man power reduction of 25-30% less manpower than a collector design from ten years ago. They are cheaper in assembly and production and there is still room for further man power reduction.

CSP Today: How do you think CSP can achieve a better LCOE?
EZ: There are many ways. But at the end of the day it depends on three major factors that can impact the cost of a CSP plant either reduced CAPEX, increased efficiency and lower O&M costs – these are the most important areas – any improvement applied to one of these areas will reduce LCOE.

For more information on research into CO2 as a working fluid, visit the PSA website