The Rise of Concentrating Solar Thermal Power

You can say those three hurdles will always remain and not just for CSP, which uses reflectors to concentrate and beam sunlight toward a receiver for producing steam, which then goes to powering turbine-generators for producing electricity. But the three issues have evolved in a marketplace that also has changed in the last few years. Photovoltaic panels have become far cheaper than expected, prompting solar thermal power companies such as Solar Trust of America and SolarReserve to start developing PV projects.

Concentrating photovoltaic (CPV) technology is attracting buyers, and that’s going to spark competition between CPV and CSP for optimal project sites. Both technologies work best in super sunny locals without much clouds – they both use reflectors to concentrate sunlight, and those optics don’t do well with diffused light.

CSP technology developers know competition will only grow more fierce. To win customers, they are improving efficiencies of their equipment to turn sunlight into electricity and adding storage to make a CSP project operate more like a fossil-fuel power plant. The U.S. Department of Energy, in addition to funding power plant projects, also has directed its researchers to explore similar technology issues.

“Our research includes the development of new heat-storage materials that are stable at high temperatures and methods that maximize the thermal energy storage capacity at low costs,” said Cliff Ho, a scientist at the Sandia National Laboratory.

The CSP market so far has a bright future. About 17.54 GW of power projects are under development worldwide, and the United States leads with about 8.67 GW, according to GTM Research. Spain ranks second with 4.46 GW, followed by China with 2.5 GW. 

About 1.17 gigawatts of CSP power plants already are online. Spain is home to 582 megawatts of them, followed by the United States with 507 megawatts. Iran, interesting, takes the third place with 62 megawatts, GTM said.

Front Runners

Major CSP players share similar profiles: they are staffed with experts in power plant engineering, sometimes specifically in CSP plant designs. They also are able to raise the capital to finance research and development and power plant construction. Some of them already have built projects in Spain, where feed-in tariffs provide a sure-source of incomes and an incentive to use energy storage to boost production. These project developers include Solar Millennium, BrightSource Energy, Abengoa Solar, Penglai Electric, Renovalia and NextEra Energy. Solar Millennium, by the way, is part owner of Solar Trust of America, which focuses on U.S. projects.

Although the United States has no feed-in tariffs, which are government-set wholesale electricity pricing designed to guarantee good returns, certain state policies have attracted a cadre of developers. They have flocked mostly to the southwestern region, which offers a combination of sunny climate, state mandates for renewable energy use, and public land that is available for energy development leases.

California, a state that recently passed a law that requires utilities to get 33 percent of their electricity from renewable sources by 2020, has been a magnet. So have Arizona, Nevada and Colorado. The California Energy Commission alone approved nine CPS projects totaling more than 4.1 gigawatts within a four-month period last year. The federal Bureau of Land Management, which also signed off on many of these California projects, approved additional projects in Nevada. One of Abengoa’s key projects, 280-megawatts Solana, will be built on private land in Arizona.

Winning permits only clears one major obstacle for these developers. Raising money is another. Without feed-in tariffs, these CSP projects must compete more on cost in order to win power sales contracts from utilities. As a result, the projects are all more than 100 megawatts in order to reach an economy of scale that keeps the construction and operating costs down. Solar Trust of America is working on the 1,000-megawatt Blythe Solar Project in California, but that project is divided into four power plants of 250 megawatts each.

The need for scale also leads to a high price tag for the overall cost of each project. Many developers applied to a federal loan guarantee program that sprung from the stimulus package in 2009. Nailing that loan guarantees was critical for all these companies who wanted to build their first CSP power plant in the United States. They also had to be able to raise equity for the project because the loan guarantee, which paves the way for the recipients to get loans from the Treasury-run Federal Financing Bank, will at most cover 80 percent of a project’s cost.

The U.S. Department of Energy, which oversees the loan program, has been fond of CSP. It has offer a total of roughly $5.89 billion to four projects; that’s more money and more projects than what the DOE has offered to developers of photovoltaic or concentrating photovoltaic power plants. Solar Trust is set to get about $2.11 billion to build half of Blythe; SolarReserve is finalizing the paperwork for $737 million for the 100-megawatt Crescent Dunes  Solar Energy Project ; BrightSource closed $1.6 billion for the 392-megawatt Ivanpah Solar Electric Generating System; and Abengoa closed about $1.45 billion for Solana.

SolarReserve is waiting for word about its loan guarantee application for another, 150-megawatt project called Rice Solar Energy Project in California, said Tom Georgis, senior vice president of development at SolarReserve.

The loan guarantee program is set to end this Sept. 30. Solar companies would like to see it get more funding. But it’s unclear how likely that will happen, considering that lawmakers and the White House are sparring over what to cut in next year’s budget. Solar industry lobbyists also hope to save a program run by the Treasury Department that covers 30 percent of the cost of a project. CEO of Solar Trust, Uwe T. Schmidt, said his company is mindful that federal funding won’t always been plentiful, and he hinted at the company’s efforts to find other sources of funding.

“We are looking for innovative ways to complement the traditional debt and equity structure,” Schmidt said. “You will see examples of what we mean.”

Key Technologies

Power plant designs that use parabolic trough reflectors and power-tower receivers are the most popular. Solar Millennium and Abengoa Solar are primarily devotees of parabolic trough technology while BrightSource concentrates on the power-tower design. The parabolic trough design features rows of connected reflectors that focus the sunlight onto tubes that run along the length of the reflectors. These tubes contain synthetic oil that flows to a heat exchanger to heat water and produce high-pressure steam. The steam then powers a turbine, which in turn runs a generator to produce electricity.

Parabolic trough power plant developers such as Solar Trust and Abengoa like to call this type of technology “proven” because a series of parabolic trough power plants totaling 354 megawatts materialized in California between 1984 and 1990. But today’s parabolic trough designs are quite different than those from two decades ago. NextEra owns 310 megawatts of this cluster of CSP plants in California. The company also is developing a 250-megawatt, Genesis Solar project in California, for which it has received the construction permit from the California Energy Commission. The largest CSP power plant proposal is the 1,000-megawatt Blythe project by Solar Trust.

Another CSP technology uses a central tower instead of tubes as the receiver. A field of reflectors beams the light to the top of the tower, where a tank of water or molten salt sits. The heated fluid then goes through the similar steps for steam generation and electricity production. BrightSource and SolarReserve both are betting on the success of this technology; BrightSource uses water while SolarReserve uses molten salt. The salt, which keeps heat trapped for hours, can be used for electricity generation after the sun goes down.

Stirling engines make up the third common CSP technology and, unlike parabolic trough and power tower setups, each Stirling engine embodies both the thermal and electric generation mechanisms and uses gas rather than fluid to transfer the sun’s heat. Main components of a Stirling engine include a giant round dish of reflectors that concentrate the sunlight to heat up hydrogen gas or helium inside an engine. The heat gas expands and creates a lot of pressure that is then used to run the piston that then drives the generator to produce electricity. Stirling engine companies include Infinia and Stirling Engine Systems (SES).

Stirling engines from SES seemed close to being deployed commercially by Tessera Solar until Tessera had a hard time raising the necessary financing and sold its two prized projects late last year and earlier this year. One of the buyers, K Road Power, said it will still use Stirling engines but only for a small portion of the Calico project; the rest will use solar panels. Tessera sold Calico as an 850-megawatt project because the project came with an electric grid interconnection agreement for 850 megawatts. But the California Energy Commission cut the size to 663.5 megawatts before issuing the permit, and the commission said the application from K Road to modify the permit in order to use solar panels does not request any change to the size of the project.

The second buyer, AES Solar, told the commission that it wasn't going to use Stirling engines at all for the 709-MW Imperial Valley Solar Project and will use some sort of PV technology instead. But the company then notified the commission last week that it still wanted to hold on to the permit for the solar thermal power plant, so it remains unclear what the company plans to do. AES declined to comment for the story.

Which Works Best

Developers of different technologies will tell you one type is better than the other. There are indeed advantages and room for improvements for all three, Ho said. Temperatures that these technologies can achieve when heating up the heat transfer fluid in the receiver, reflectivity of the reflectors, as well as the sunlight-to-electric conversion efficiencies are some of the metrics.

Parabolic trough plants generally heat the heat transfer fluid to about 390 degrees Celsius, which is lower than the temperatures from power tower plants and Stirling engines, Ho said. Power tower designs can achieve around 550 degrees Celsius and higher. Running a steam turbine at a higher temperature improves its efficiency. The thermal-to-electric efficiency of a parabolic plant is around 38 percent while the efficiency for a power tower plant is up to 42 percent, Ho said.

A power tower plant can end up operating for fewer hours each year than the trough plant, however, because power tower relies on a single receiver. The plant’s output will be compromised if that receiver isn’t working well. A trough plant has many loops of tubes, so one problematic loop won’t cause the whole plant to shut down, Ho said.

If you look at the sunlight-to-electricity efficiency, Stirling engine can do better than parabolic trough or power tower. The paraboloidal dish of the Stirling engine gives it the highest solar concentration ratio, Ho said. As a result, the sunlight-to-electric peak efficiency is about 31 percent for Stirling engine and 22-23 percent for the other two, Ho said. Stirling engines use far less water and needs it for washing the dishes. Parabolic trough and power tower designs, on the other hand, need far more water to condense the steam for re-use after it’s gone through the generator. The latter two can use dry cooling by running fans, but that adds costs and lowers the plant’s efficiency. Fights over water use have prompted BrightSource, Solar Trust and NextEra to incorporate dry cooling in designs for some of their projects.

Stirling engines are modular; each of them is a stand-alone power producer. That can be an advantage for deployment because these engines don’t require a centralized turbine-generator. But it also can be a disadvantage because it can’t achieve a certain economy of scale that is possible with parabolic trough and power tower designs, where one way to increase energy output may be to add more reflectors but not other pieces of equipment. A Stirling engine plant scales more like a PV power plant, Ho noted. Adding storage to Stirling engines will improve the technology’s appeal.

“Currently, without storage, dish-engine systems are similar to PV systems,” Ho said. “Storage will be a differentiating factor between dish-engines and PV modules that can increase the capacity factor of dish-engines and potentially reduce costs.”

Storage is a big selling point for CSP developers these days, particularly since they are having a harder time competing with PV technologies that have become much cheaper in the last two years. Storing thermal energy for use after the sun goes down means a CSP plant is more flexible in adjusting its power output to meet a utility’s demand. PV power plant output can drop significantly in the late afternoon and early evening, when electricity use can spike as people come home from work and turn on TVs and other appliances.

CSP power plants with storage already are running in Spain. Developers such as Solar Trust and SolarReserve have designed storage into their projects in the United States as well. Although including storage means adding costs, the greater ability to provide power on demand makes a CSP plant more valuable than one without storage or with PV, Georgis said.

“Las Vegas’ demand stays warm into the evening, when all the lights are turned on. We can operate like a conventional power plant. We can displace conventional generation,” he added.

Researchers are working on boosting the temperatures of the molten salt so that a smaller amount of it is necessary to produce the same amount of electricity, Ho said. Finding materials that will keep molten salt stable at high temperatures is another goal. Right now, molten salt can start to decompose and cause plugged pipes and valves when it reaches 600-650 degrees Celsius.

Improving the reflectivity of the reflectors is another research focus. Schmidt said the latest reflector design by Solar Millennium and its R&D partners have created larger and more efficient reflectors that come with fewer parts for easy assembly. At Sandia, meanwhile, researchers plan to test a metalized polymer film produced by 3M to see how it compares with glass reflectors, HO said. This project is set to start this summer.

"We want to look at the peak flux, total power, and beam size and shape to see if we can get a tight beam on the receiver with these metalized reflective films,” Ho said.