CSP stands for “concentrated solar power”. It describes the process of concentrating sunlight on to a working fluid, using either mirrors or lenses. The hot working fluid then heats water in a boiler, flashing it to steam. In turn, the steam drives a turbine to generate power.
Some CSP systems, typically using molten nitrate salts, can store the heat in the working fluid for hours of days. Molten salt heat storage enables the generation of 12 or more hours of electricity at night, after the sun has gone down.
Compressed solar irradiance map from the Global Solar Atlas 2.0 (click here for full attribution)
Concentrated solar power is primarily a technology for the subtropics and nearby regions. Only direct sunlight can be concentrated with mirrors or lenses. Diffuse sunlight will not work.
Although the tropics are hotter than the subtropics, the additional heat causes significantly higher evaporation and humidity, which leads to more clouds, thus reducing the number of hours each year with clear skies, and the potential power available from concentrated solar technology. Compare the reduced power available at the equator vs the subtropics on the map above.
The geographic restriction to regions in, or close to, the subtropics seems to severely curtail the use of CSP. But it turns out that these regions are where a considerable proportion of the world’s population lives, perhaps 40% or more.
There are two main types of types of CSP systems, described below.
Parabolic trough systems, such as in the picture above, use long parabolic reflectors to concentrate sunlight on to a long thin tube containing the working fluid. Trough systems were more common until a few years ago, but generally do not include storage.
The most popular type of CSP system recently has been solar tower thermal, as shown in the image of the Hami, China system at the top of this web page.
Solar tower thermal systems use a field of reflecting heliostats, perhaps of up to 1 sq km, to reflect sunlight on to a tower, through which molten salt (usually nitrates) passes. The final temperature of the salts will be in the range of 500 to 1000°C (of which the upper limit relates to proposed rather than actual working fluids). The final temperature of the salts must be kept high enough for them to remain molten after generation is complete. Initially, weeks or months may be taken to melt the salts before the plant can become operational.
For a further description of these and of other types of CSP systems see Wikipedia : Concentrated solar power – Current technology.
The highest efficiency obtained is around 35%, which is higher than the 33.2% Shockly-Queisser efficiency limit for any single junction solar cell. Much work is ongoing to increase maximum working fluid temperatures, which, in turn, increases the thermodynamic efficiency and thus reduces the LCOE (levelised cost of electricity) of generation.
According to the chart in the IRENA document Renewable Power Generation Costs in 2021 Figure B5.1 p136), between 2010 and 2019, weighted average CSP working fluid temperature have increased from 396°C to 485°C, with more to come.
Higher working temperature fluids and other improvements have resulted in a turbine efficiency (not overall system efficiency) increase from 38.1% to 44.0%.
With sufficient hot salt storage and suitably chosen heliostat and generation capacity, the number of clear days in the location limits the capacity factor of a CSP or hybrid CSP / solar PV system. The Atacama Desert in Chile has over 300 clear days per year, so the annual capacity factor there can be as high as 30/365 = 82%, similar to the highest point on the charts above.
Even though hybrid CSP plus solar PV systems can provide 24 hour power, some form of backup is still required to provide reliable grid power all year round.
Both peaks in air conditioning demand and maximum output from CSP (and solar PV) systems are on clear, sunny days. Thus, they correlate well with each other.
The implication is that, in places where CSP is effective, the maximum load will be lower during days where the sky is not clear, and where CSP output is low.
Thus, the backup capacity required for CSP is lower than might be supposed. For a region mainly powered by CSP, backup capacity does not need to cover the peak hours of annual demand, just the maximum demand during (cloudy) days and subsequent night where CSP is not effective. How different these are depends on the fraction of homes which have air conditioning.
According to an IEA 2018 report on air conditioning, in the USA and Japan, 80% of homes have air conditioning. But in hot countries (such as the subtropics suitable for CSP), only 8% of home have air conditioning.
Currently 1.5 billion homes have air conditioning, expected to rise to 5.6 billion by 2050.
In India, air conditioning is currently 10% of electricity demand but is expected to rise to 45% of demand by 2050. Thus variable solar generation must rise considerably, but backup dispatchable generation needs to rise by somewhat less.
These IEA estimates for India give some indication of the likely situation in other subtropical countries.
The region near the equator will be different, because it is much more often cloudy. Thus, CSP is less suitable for equatorial regions, as CSP does not work in diffuse sunlight.
In the geographies where CSP is viable, solar PV power with no storage is invariably considerably cheaper than the likely future cost of CSP.
The US DoE target is to develop CSP technologies which can bring the cost of CSP down to $50/MWh. However, a better solution is to install CSP in conjunction with solar PV.
Solar PV can provide cheap daytime power. But for nighttime power it is expensive to add enough grid battery storage to provide complete overnight power coverage. Future CSP at $50/MWh should provide competitive night time power which would make average power prices competitive with fossil fuels.
The combination should provide 24 hour power, except on the occasions when the daytime weather is overcast. Although weather isn’t often overcast in the subtropics, it does happen.
The LCOE of CSP in the last decade or so has reduced considerably as indicated in the IRENA chart (Figure 5.7 from Renewable Power Generation Costs in 2021 p133), despite the provision of increasing durations of storage in these systems. Further, it has been achieved with a relatively limited cumulative installed capacity increase, certainly relative to the huge capacity increases in solar and wind installations over the same time. Most likely, a significant part of the reductions are due to technology improvements.
There are two additional hybrid solar PV plus CSP project bids with significantly lower costs.
The first of these is the fourth phase of the DEWA (Dubai and Electricity Water Authority) Mohammed bin Rashid Al Maktoum Solar Park, which combines CSP from a 600 MW parabolic base with 11 hours of storage and a 100 MW solar tower thermal with a record 15 hours of storage, with 250 MW from solar PV.
The LCOE of the 700 MW of CSP is $73 US/MWh, and of the 250 MW of solar PV is $24 US/MWh.
There was also a recent bid in the Atacama Desert, Chile, by EIG of a combined CSP and solar PV system, split between Likana (CSP) and Pampa Union (solar PV), with an average power price less than $34/MWh. The original proposal was for a system delivering 200 MW of solar tower CSP with 12 hours of storage, and 200 MW of solar PV, providing 24 hour power for around 80% of days.
The CNE (Chilean National Energy) Commission rejected the hybrid bid, not because there was anything wrong with the bid, nor because it was not good value for money, but because the auction design did not take account of the times when renewable energy was generated, and the source of any necessary overnight power. So solar PV bids, incapable of providing power from renewables at night, easily undercut the EIG hybrid solar CSP + PV bid. Hopefully, the benefits of such a hybrid solar CSP + PV bid will be recognised in the design of future CNE auctions.
The Atacama Desert is one of the sunniest place on the planet, with over 300 clear days per year, one reason why so many major astronomical telescopes are built in Chile. Thus it is highly suitable for CSP. Other subtropical locations are also suitable, but the Atacama Desert is the best.
Given that the same auction produced solar PV bids of around $13/MWh, one implication of the bid price is that the CSP power component of the bid could cost around $34 + (34 – 13) = $55/MWh, though this figure should be treated as only indicative, rather than definitive.
China now has a fast track pipeline of 30 hybrid CSP projects with the CSP system scheduled to complete before mid 2024. However, many Chinese renewable energy projects, scheduled for earlier in the year, get pushed into Q4.
The total capacity across all hybrid CSP projects includes 3 GW of CSP with storage, 20 GW of solar PV, and 1.5 GW of wind power. Hybrid projects are located in Qinghai, Gansu, Jilin, Xinjiang, Tibet and elsewhere, as part of gigawatt-scale renewable energy complexes in these provinces.
Full Global Solar Atlas condensed map attribution
The solar irradiance map is obtained from the Global Solar Atlas 2.0, a free, web-based application, developed and operated by Solargis s.r.o. on behalf of the World Bank Group, utilizing Solargis data, with funding provided by the Energy Sector Management Assistance Program (ESMAP). For additional information see the Global Solar Atlas.