LCOE of single and multi-system tower CSP plants compared
Efficiency or LCOE: which is more important to reduce the cost of Gen-3 CSP systems?
A pair of related papers from ANU in Australia compare the value of efficiency versus Levelized Cost of Energy (LCOE) in Concentrated Solar Power (CSP) systems.
The first paper, published at Solar Energy; titled Techno-economic assessment of a numbering-up approach for a 100 MWe third generation sodium-salt CSP system, is a techno-economic study comparing LCOE and efficiency in multi-unit CSP systems against today’s single 100 MW tower CSP.
The second paper, published at Applied Energy; titled Co-optimisation of the heliostat field and receiver for concentrated solar power plants, focuses on optimizing a heliostat solar field together with the solar receiver for the first time to compare investment in efficiency versus LCOE.
Both studies demonstrate a trade-off between optimizing for the highest efficiency or for the lowest LCOE.
“LCOE is what people are mainly understanding, and although we know that it’s a very poor design criterion for CSP, it’s still what you can report to try to be understood,” said co-author Charles-Alexis Asselineau in a call from Spain, where he is currently at IMDEA Energy, after eight years at ANU in Australia, working on a similar type of solar research for a sustainable future.
“Some components do cost a fair bit for the efficiency they supply to the overall system. And sometimes you’re better off actually scaling down a bit the size of these components and making them slightly less efficient because the cost gains you get in the end are worth it. And also, efficiency is not a simple metric. The subsystems like the field, the receiver, the tower and all the machinery in the tower and then the storage and so on all kind of interact together, and can make it complicated.”
Each of these costs would have a multiplied impact in the first study. Instead of one large CSP tower plant, with one large solar field, this study looked at the economics of CSP design that would instead comprise multiple small systems.
Why smaller modules matter: a small power block is in the CSP future
Size of the s-CO2 turbine compared with a Rankine cycle turbine ©GE
“If supercritical CO2 actually gets to the commercial stage, then it is possible that we will be able to build smaller CSP systems with favorable economics,” said Asselineau.
“And we will not have to compromise on the efficiency of the solar system by always having to go bigger and bigger because of the Rankine cycle. This paper shows that it is worth going through this exercise of looking at smaller sizes and multiple small CSP plants, as with a much smaller supercritical CO2 cycle, you can have an efficient power block with a much smaller turbo and a much smaller balance of plant.”
LCOE of single and multi-system tower CSP plants compared. Image ©Techno-economic assessment of a numbering-up approach for a 100 MWe third generation sodium-salt CSP system
Commercially, the optimal size of tower CSP plants has settled at 100MW for optical reasons and comprises one tall central tower surrounded by one very large solar field of heliostats. But a thermal power block using a Brayton s-CO2 cycle, which is significantly smaller than today’s legacy Rankine cycle system would change that.
“That’s why what we looked at was actually breaking down the whole CSP system into smaller independent systems to see how that performs with supercritical CO2 cycles.” he explained.
The papers were in response to the US DOE call for competition between liquids (molten salts), solids (particles), and gas-based (air, etc) for Gen-3 CSP systems. All three systems would feed into the s-CO2 power block.
The team was investigating the liquid pathway, using liquid metals to transfer the heat.
Asselineau noted that although the DOE chose the particles-based approach to develop further, the CSP industry itself is heavily invested in liquids, with China continuing to develop liquids.
Co-optimizing the solar field with the receiver
In the second study comparing efficiency versus LCOE, the authors introduce a way to simultaneously optimize the heliostat field, the solar receiver on the tower, and the aiming strategy based on annual operational conditions, using an integrated optical, thermal, and mechanical model to accurately simulate the solar subsystem connected to the rest of the plant components, to ensure the solar flux remains neither too much nor too little, year-round.
Co-optimizing the solar field with the receiver in CSP ©Co-optimisation of the heliostat field and receiver for concentrated solar power plants
Co-optimizing the receiver and the solar field as a unit is a first, Asselineau believes:
“I think this is the only formal optimization work that has been published on the topic so far. Prior to this, no-one developed a method that could actually tackle the complexity and the economics together. This was the final work of Dr. Shuang Wang, one of our PhD students at the time.”
That team looked at various combinations of receiver size and solar fields to compare annual solar-to-thermal efficiency, annual field efficiency, annual receiver efficiency, and LCOE for each design parameter variation. Then, they tried out varying each design parameter from 0.8 to 1.2 times the optimized value to double-check that they had found the ideal system.
They had access to a supercomputer able to crunch 700 hours of computation into 72 hours. The studies used a Genetic Algorithm (JEGA in DAKOTA) for the heliostat field and receiver design.
The trade-off between efficiency and LCOE
As an example of the trade-offs, the team found that solar field efficiency increases when the solar receiver is bigger because more heliostats will succeed in reflecting sunlight onto a bigger target. But conversely, a bigger receiver is less efficient, as it tends to lose some heat to ambient air.
But, although a smaller receiver would have more “spillage” (wasted solar reflections that missed the mark), they found that this had relatively little impact on LCOE because the cost reductions on the receiver made up for them.
Overall, the team found that the choice of optimization metric had a significant impact on the design of the components. Considering Gen-3 Liquids hypotheses, combining a larger solar field with a smaller and lower solar receiver than the industry average, just over 173 meters high, would cut the cost of the solar field by 5.2% and of the receiver by 15.5%.
Asselineau added that some important engineering factors, like reliability and project standardization, not covered in the studies, could further improve the case for multiple systems.
Australia’s liquid sodium solar receiver to be on show in October
Sodium-based Vast Solar Combines the Best of Trough & Tower CSP to Win our Innovation Award
