CSP in Spain: A place in the sun

Parabolic trough reflectors Source: Andrew Duke

Recent years have seen spectacular growth in concentrated solar power deployments in southern Spain, where projects benefit not just from high insolation but also a generous tariff regime. The challenge now is to drive down costs so that CSP can one day stand subsidy-free, writes Geoff Nairn.

Geoff Nairn

Concentrated solar power (CSP) uses mirrors to concentrate sunlight and generate heat. This energy is captured in a liquid or gas and is typically used to generate electricity via a conventional steam cycle.

Unlike photovoltaic farms or wind energy, which has grown to be Spain’s third largest power source, CSP plants can cost-effectively store energy that cannot immediately be used.

That is important in Spain, which has a second demand peak in the evening, and most new CSP projects incorporate storage so that they can keep generating electricity several hours after the sun has gone down, or even right through the night.

But, while CSP is more dispatchable than other renewable energy sources, it is also currently more costly. So Spain is the focus for efforts to drive down costs, both through economies of scale and improvements in technologies.

“All the CSP technologies are expensive so a lot of research seeks to reduce component costs and optimise production and installation,” says Eduardo Zarza, head of R&D for solar concentrating systems at the Plataforma Solar de Almeràƒ­a (PSA), Spain’s leading solar energy research centre, which conducts research into all four CSP technologies.

Parabolic troughs more established

The most mature of these is the parabolic trough design, which accounts for 93 per cent of the 2500 MW of new CSP capacity that Spain has authorised up to 2013. While the other three technologies ” solar tower, Fresnel collector and Stirling dish ” all have commercial potential, the financial backers of Spain’s CSP projects have opted to reduce their risks through parabolic trough’s longer track record. In the USA, parabolic trough plants date back to the 1980s.

“With a tower system, for example, it is difficult to get project finance because no one knows how long the receiver will last,” says Frank Dinter, head of solar at the German utility RWE, which is an investor in several Spanish renewable projects.

To benefit from Spain’s generous feed-in CSP tariff, currently €0.28 ($0.41) per KWh for 25 years, CSP plants cannot be bigger than 50 MW. This size limit is seen as less than optimal given the current maturity of parabolic trough technology and a restriction on potential economies of scale. Several costs in a CSP project are not proportional to its size. For example, a 200 MW turbine cost less than four times as much as a 50 MW version. Dinter estimates that a 200 MW plant would be around 25 per cent cheaper per megawatt than a 50 MW plant.

Andasol 3 solar field

RWE has a 25 per cent stake in Andasol 3, a 50 MW parabolic trough CSP plant near the Andalucian town of Guadix. The solar field at Andasol 3 consists of 7296 solar collectors arranged in eight banks of 304 parallel rows aligned north to south. Each solar collector comprises a parabolic mirror with a Dewar receiver tube running horizontally along its focal line. A hydraulic drive moves the collector rows in an arc to track the sun from east to west during the day.

Synthetic oil is pumped through the receiver tubes to absorb the sun’s heat, reaching a maximum of 393 à‚°C when it exits to the heat exchanger. The heat creates steam that drives a turbine and generates electricity using a conventional steam cycle.

Unlike earlier generations of parabolic trough plant, Andasol 3 incorporates molten-salt heat storage. When the plant is generating more heat than is needed to produce electricity, some of the hot oil is siphoned off to a storage circuit, in which a second heat exchanger heats up a nitrate salt mixture as it is pumped from a cold tank to a hot tank. To produce electricity once the sun goes down, the flows are reversed and energy is transferred back from the hot salt to the oil.

The heat stored in the 28 500 tonnes of salt can provide an additional seven hours of power at full load in summer evenings and an extra three hours in the winter. “If we reduced the capacity we can run at 24 hours but we would only be producing 30 MW during the night,” says Dinter.

The price of storage has traditionally been high and it complicates the plant design, but Dinter says adding storage to Andasol 3 allows it to operate 4000 hours a year instead of just 1000 without storage. One of parabolic trough plants that use synthetic fluid as transfer liquid is the relatively low working temperature of 393 à‚°C; the fluid degrades above 400 à‚°C. This low temperature limits the overall steam cycle efficiency to around 38 per cent.

“You can boost the power block efficiency by four percentage points simply by working at higher temperatures,” says Peter Màƒ¼rau, Siemens project manager for molten salt technology.

Siemens gets behind molten salt

The German engineering giant is promoting molten salt as a working fluid as enables plants to work at higher temperatures. Siemens is involved in a research project at the University of Evora in Portugal that will build a test facility using molten salt as the transfer medium.

The 300 metres-long loop will be able to operate at temperatures above 500 à‚°C and will test different types of salt as the transfer liquid. A similar 5 MW demonstration plant is already operating in Sicily.

“Molten salt has significant potential to bring down the levelized cost of electricity (LCOE),” says Màƒ¼rau. Another advantage of using molten salt both a working fluid and for storage is that plant design is simplified as the oil-to-salt heat exchanger is not needed.

Eliminating the exchanger allows the salts in the hot storage tank to reach higher temperatures than in an oil-based plant. The size of the tank can thus be reduced as less salt is needed to store a given amount of energy, leading to a cost saving of around 30 per cent on the tank component.

The big drawback with molten salt is that it freezes at around 220 à‚°C. Care has to be taken to ensure the salt does not solidify in the solar field piping during the night. “That is quite a challenge over a big solar field”, admits Màƒ¼rau.

Researchers are looking to develop new salts with lower freezing points but the attraction of the current mixture ” 60 per cent sodium nitrate and 40 per cent potassium nitrate ” is that the ingredients are cheap. “There is a lot of research into new storage materials but molten salt is currently the favourite,” says Màƒ¼rau. The existing salt mixture is also environmentally benign and — unlike synthetic oil ” does not catch fire. Andasol has already suffered a fire in the solar field due to escaping oil.

Researchers at the PSA are investigating other types of working fluids for parabolic trough plants. Direct steam generation is one option that would allow parabolic plants to operate at higher temperatures.

Efficiency gains down the line

Using steam would also greatly simplify plant design through the elimination of the main heat exchanger. More expensive receiver tubes are needed to withstand high-pressure steam but the switch to steam could reduce total plant costs by 5 per cent and increase efficiency by up to 7 per cent, according to a recent report on CSP from consultancy AT Kearney.

However, the big drawback with using steam is that an efficient and high-capacity storage solution needs to be developed for steam. Zarza says researchers at the PSA are also investigating using a compressed gas such as carbon dioxide or nitrogen as the working fluid. But plants would have to be radically redesigned to work with gas instead of liquid.

The mirrors and receiver tubes are critical components in a parabolic trough plant and so the subject of much innovation. The current thick-glass parabolic mirrors offer 93.5 per cent reflectivity. By 2015, AT Kearney expects that new mirror technology could boost that figure to 95 per cent, which translates into an increase in overall plant efficiency of 3.5 per cent. More precise bending of the mirrors could also deliver a further 2 per cent gain in efficiency.

However, there are a limited number of manufacturers for the mirrors and receiver tubes, just three in the latter case, so competition in the supply of key components is currently limited.

Fresnel CSP technology cheaper but less efficient

This is where the linear Fresnel CSP technology has a key advantage. Unlike the better-known parabolic-trough design, linear Fresnel plants have a much simpler and therefore cheaper solar field design.

Long strips of flat mirrors focus the reflected sunlight on the solar receiver tube, through which saturated steam circulates at up to 285 à‚°C and 70 bar. As the sun moves, the mirrors rotate but the receiver tube remains fixed.

The drawback of a Fresnel plant is that it only captures 65 per cent of the sunlight that a parabolic trough plant captures. The business case for Fresnel technology therefore hinges on lower costs.

“If a Fresnel plant costs 20 per cent less but produces 35 per cent less electricity than a parabolic trough plant, it is still not competitive,” says Zarza.

Novatec Solar, a German company, has been operating a small-scale 1.4 MW Fresnel plant, Puerto Errado 1 (PE1), in the Murcia region of Spain since 2009. A larger 30 MW plant, called Puerto Errado 2 and majority-owned by two Swiss utilities, is being built and will go live in March 2012.

Martin Selig, founder of Novatec Solar, argues that while Fresnel technology is less mature, it has significant potential for cost reductions once the components are mass manufactured. As it operates at lower temperatures, the receiver tube is much simpler than the Dewar tube technology used in parabolic trough receivers. Similarly, the manufacture and installation of flat mirrors can more easily be automated.

While linear Fresnel was originally conceived as low-cost, low-temperature technology, experts see its relatively low thermal efficiency of around 26 per cent as a potential commercial stumbling block.

Novatec therefore plans an evolution of its technology that uses superheated steam to boost turbine cycle efficiency. This will be done by adding an additional high-temperature loop to its PE1 demonstration plant, with redesigned piping and new collectors that can handle superheated steam at 450 à‚°C.

The solar power tower

The third significant commercial CSP technology in Spain is the solar tower, which uses a circular arrangement of ground-based heliostats to focus sunlight onto a tower-mounted receiver.

The PSA research centre has had a small-scale tower in operation for over 25 years. Spain’s first commercial tower plant, Abengoa’s PS10 near Seville, started operating in 2007. The 11 MW plant has very limited storage — sufficient to ride out 30 minutes of cloud cover ” and uses saturated steam as the transfer medium.

The Andasol 3 solar plant in Spain Source: Andrew Duke

The new generation of tower technology is represented by Gemasolar, a 17 MW tower built by Torresol Energy, a joint venture between Spanish engineering firm Sener (60 per cent) and Abu Dhabi’s Masdar (40 per cent). Gemasolar was due to go live this spring.

Gemasol uses molten salts both as the heat transfer medium and for storage of up to 15 hours. That means that while the rated power of Gemasolar is only 17 MW, it can produce as much energy as a 50 MW parabolic design, due to its longer hours of operation.

By using molten salts, Gemasolar works at higher temperatures than previous generations of tower plant such as the PS10. At 560 à‚°C, the efficiency of a molten salt tower plant improves by around 24 per cent compared to a steam-powered predecessor.

Juan Ignacio Burgaleta, head of technology at Torresol Energy, argues that one of the big advantages of a central tower design over other designs is its simpler operation and maintenance. In a parabolic trough or Fresnel plant, the transfer fluid must travel through 80 km of collector pipes before reaching the power block. In a tower plant, the transfer fluid is confined to a much smaller circuit comprising the central tower and the nearby storage system.

Researchers are already looking beyond today’s tower plants to new designs that can work at up to 800 à‚°C using ambient air as a transfer fluid.

That would boost the efficiency of the plant by as much as 13 per cent. Temperatures could be pushed even higher using new materials but the more innovative the receiver, the more difficult it is to obtain project finance.

Stirling dish: the most efficient of all?

The final competing approach in CSP is Stirling dish technology, which has yet to be deployed commercially in Spain. This uses a parabolic dish to focus sunlight onto a Stirling engine and, theoretically, offers the highest efficiencies of all CSP technology.

Stirling dish technology is inherently small-scale and commercial systems typically generate around 2.5 kW. That makes it more suitable for off-grid applications, although Zarza of the PSA says many dishes could be located together to create a larger grid-connected system.

Stirling technology offers perhaps the most potential for cost reduction, by shifting manufacture to low-cost countries and using more off-the-shelf components, for example.

But the big commercial stumbling block is the poor long-term reliability of Stirling engines, which have to withstand temperatures of up to 700 à‚°C and pressures of 150 bar, says Zarza. “At the moment the engine components suffer a lot so we are going to need new materials,” he says.

One big advantage of the Stirling dish is that it uses less water than other solar technologies. Water use can be a controversial issue in southern Spain. The best locations for CSP plants are arid regions with little cloud. However, such locations often experience severe water problems and a plant such as Andasol 3 uses 500 000 m3 of water a year, mostly for cooling the steam cycle but also for cleaning the mirrors.

The investors in Novatec Solar’s PE2 plant insisted on air cooling to avoid any controversy over water use, even though the decision reduces the economic return.

“Air cooling costs much more and it reduces the output by 5-6 per cent,” says Selig.

Opinions are nevertheless divided on this issue. RWE’s Dinter says water cooling is essential to boost the thermodynamic efficiency of the steam cycle of parabolic trough plants like Andasol 3 with a relatively low inlet temperature.

“With dry cooling, you cannot reduce the outlet temperature as much as with water,” he says.

Torresol Energy’s Burgaleta says that even though the Gemasolar central tower plant works at higher temperatures, water access was not a problem, and so the designers opted for water cooling. However, he says one of Torresol’s central tower projects planned for the future will have air cooling instead.

As Spain is now discovering, CSP far from being a single technology but embraces a range of designs and key technologies, each with different operating characteristics, risk profiles and trade-offs. “There is no clear winner,” says Siemens’ Màƒ¼rau.

Even without any radical technological breakthrough, improvements in technologies and greater economies of scale are expected to drive a 30 per cent reduction in the cost of CSP-generated electricity in Spain by 2015. And by 2025, costs may have fallen by as much as 50 per cent, at which point CSP plants will finally be in a position to substitute conventional sources in Spain’s energy mix.

More Power Engineering International Issue Articles
Power Engineering International Archives
View Power Generation Articles on PennEnergy.com

No posts to display