When most people talk about solar power in the U.S., they tend to think narrowly about photovoltaics. But it’s a big world out there. Even in the U.S., researchers are working on interesting alternatives, particularly for big commercial bulk power facilities intended as alternatives to nuclear or fossil-fuel power plants.
One of the issues with photovoltaics is just how far they can scale. The largest photovoltaic plant in the U.S., Florida Power and Light’s 25-MW Next Generation Solar Energy Center in DeSoto County, comprises more than 90,000 SunPower solar panels on 180 acres. The world’s largest photovoltaic plant, the 60-MW Olmedilla Photovoltaic Park (Fig. 1) in Olmedilla de Alarcón, Spain, uses more than 160,000 conventional solar photovoltaic panels.
Olmedilla may represent a limit on the maximum capacity of photovoltaic solar. Or it may not. These technologies have a way of leapfrogging each other. Recently, First Solar in Tempe, Ariz., announced that it would build a 2000-MW solar power plant in China. First Solar makes thin-film solar cells in a continuous process, rather than on wafers. However, its cells use tellurium, the relative scarcity of which could limit product volumes.
Efficiency is another issue related to scaling. In terms of wafer-based cells, SunPower offers a conversion ratio of 23.4%, while the average for the rest of the market ranges from 12% to18%. SunPower’s efficiency comes from metalizing the collection grid on the back of the wafer, rather than the front, as it is on other manufacturers’ wafers. Back-collection increases the photon-capture area of the wafer. What makes this possible is SunPower’s process technology, which allows the charge carriers created by the solar-cell junctions when they are hit by photons to migrate all the way through the thickness of the wafer.
However, Sharp Solar has recently announced cells with 35.8% conversion efficiencies using a triple-junction compound solar cell. (Multiple band-gaps allow the junctions to respond to a broader range of photon energy levels—that is, a broader light spectrum.) This approach has been used for decades in powering expensive military satellites, but the manufacturing cost put it out of reach for commercial use.
Beyond that, experimenters at DuPont and the University of Delaware have achieved 42% efficiencies by using multiple band-gap collectors and spectral-splitting optics, according to the university.
Until FirstSolar’s Chinese photovoltaic plant is built, photovoltaics seem limited to a scale of tens of megawatts and hundreds of acres. At the same time, as production economies of scale kick in, photovoltaic panels appear to be the ideal fit for what the Smart Grid visionaries call “distributed resources,” domestic rooftop systems like the 2.5 kW on my own roof, or the distributed 2.5 MW that Intel announced it would install in the first half of 2010 on factory roofs in Arizona, California, New Mexico, and Oregon.
LINEAR CONCENTRATED SOLAR
Non-photovoltaic methods can be lumped into the concentrated solar thermal (CST) category. The common approaches include linear, power tower, and dish-engine systems.
Linear systems generally comprise multiple parallel rows of solar collectors. They may be stationary, or they may be motor-driven to pivot around a north-south axis, following the sun from east to west through the day. Parabolic trough systems can focus sunlight by 30 to 100 times, heating a fluid in a pipe that runs down the focal line of each trough. A heat exchanger then boils water to produce superheated steam to run turbine generators. It’s possible to run the same turbines at night or during cloudy days using natural gas as a fuel, providing continuous capacity.
Linear systems are sometimes used with thermal energy storage (TES) systems. This typically involves a salt mixture of 60% sodium nitrate and 40% potassium nitrate, which melts at 221°C. The TES system keeps the mixture as a liquid and cycles it between “cold” storage at 288°C and “hot” storage at 566°C.
In the solar plant, the hot salt is used at night to take over the job of superheating steam for the turbine generator. A 100-MW turbine can be run for four hours by pumping molten salt between hot and cold tanks roughly 30 feet tall and 80 feet in diameter.
Linear Fresnel-reflector concentrating systems are an alternative to troughs. They employ flat or slightly curved mirrors and Fresnel lenses to concentrate sunlight onto a pipe. They aren’t as optically efficient as parabolic troughs, but they can be cheaper because they can heat water for the turbines directly. They also require less land because design engineers don’t have to worry about one trough shading another.
The largest non-photovoltaic installation (Fig. 2) in the world is a collection of nine trough systems in California. The Solar Energy Generating Systems (SEGS) 364-MW collection of nine solar-trough power plants, operated and partially owned by NextEra Energy Resources, comprises SEGS I-II (44 MW), located at Daggett, SEGS III-VII (150 MW), located at Kramer Junction, and SEGS VIII-IX (160 MW), located at Harper Lake, all near Barstow, Calif., in the Mojave Desert. Taken together, the nine plants cover almost eight square miles.
This is somewhat mature technology. The plants were built between 1984 and 1990. SEGS VIII and IX, at Harper Lake, are the newest. They each have a peak capacity of 80 MW, which is more than double the capacity of most of the older plants. (All except SEGS I have peak outputs of 30 MW. SEGS I has a maximum 14-MW output.)
The concentrator mirrors are half-pipes. The working fluid in the Rankin-cycle steam turbine is water that is heated by a synthetic oil, which is heated to more than 400°C as it is pumped through the pipes at the foci of the troughs. The two-stage heating avoids over-pressurization in the heat-collection piping.
Power tower systems are awesome sights. They use hundreds or even thousands of large, sun-tracking flat heliostat mirrors, covering acres to concentrate sunlight onto a receiver at the tops of tall towers. The concentrated sunlight heats pressurized water or molten salt to run steam generators.
Tower systems are massive, and they cannot be separated into geographically dispersed elements like the SEGS trough farms. This can lead to conflict with environmentalists who favor non-fossil-fueled power on the one hand, but are concerned about wildlife on the other. That’s what is happening with a planned 400-MW power tower that aims to use some 6.25 square miles of desert near Las Vegas—desert that is habitat to the desert tortoise.
In February, The New York Times reported that the developer of the first new solar power plant in California in 20 years had proposed revamping the project to defuse concern over its effect on the animal. BrightSource Energy now will submit a new design plan that shrinks the size of the 4000-acre Ivanpah Solar Energy Generating Station by 12%, the Times also said.
This change should reduce the number of desert tortoises that must be relocated. It also will avoid an area populated by rare plants. The portion of the project that would have the greatest effect on wildlife will be cut by 23%. And, the plant’s electricity generation will fall from 440 MW to 392 MW, the Times reported. The Sierra Club said that it supports the project, but would still like to see it moved.
To get a sense of the physical scale of this project, Ivanpah’s 4000 acres or 6.25 square miles is enormously bigger than FPL’s 180 acres of photovoltaics in DeSoto County. Another way to look at it is that 400 MW is roughly one-fifth the capacity of nearby Hoover Dam. Lake Mead, behind Hoover Dam, covers about 250 square miles when it’s full—about 40 times more area than Ivanpah. (The lake has more functions than generating electricity, though.)
With respect to Ivanpah and the environmentalists, there’s a ticking clock. The Times also noted that BrightSource Energy has to begin construction of the Ivanpah power plant by the end of the year before it can qualify for a loan guarantee from the United States Department of Energy.
Environmentalists in other countries have raised less of an uproar about wildlife, but they’ve also been dealing with smaller-scale tower installations. The largest solar power-tower installations (Fig. 3) in the world are in Spain. The 11-MW Planta Solar 10 (PS10) and 20-MW PS20, which are located in Sanlucar la Mayor near Seville, represent the first phase of a very large construction project. The whole facility will be completed by 2013 and will produce around 300 MW.
The existing facilities are co-located. PS10 has 624 heliostats, each with a 120-m² surface-area parabolic mirror. The mirrors are focused on a 115-m tower, heating water pipes that provide 200 m² of water-cooled energy-exchange surface area. The tower uses water rather than molten salt. Even so, the steam retains enough enthalpy to allow generation at half capacity for an hour or longer after dark.
With 1255 two-axis sun-tracking heliostats and mirrors, PS20 has double the capacity of PS10. The PS20 tower is 165 m high and also uses steam, but it employs second-generation technology in the receiver, in its control system, and in energy storage.
Dish/engine systems (Fig. 4) are another photovoltaic alternative. Their optimum size, however, seems to be more on the scale of a distributed power generator (1 to 40 kW) than a bulk-power plant, though they can be scaled up in arrays. They are expected to find applications in emerging global markets.
In a dish/engine system, a single concentrator (which may actually be a mirror array) continuously tracks the sun, reflecting its solar energy onto a receiver where it is absorbed, converted to heat, and applied to an engine/generator. This is often a Stirling-cycle engine.
Recall that in the Stirling cycle, the working gas is alternately heated and cooled by constant-temperature and constant-volume processes. Stirling engines usually incorporate an efficiency-enhancing regenerator that captures heat during constant-volume cooling and replaces it when the gas is heated at constant volume. The best of the Stirling engines achieve thermal-to-electric conversion efficiencies of about 40%. Stirling engines are a leading candidate for dish/engine systems because their external heating makes them adaptable to concentrated solar flux and because of their high efficiency.
A great deal of insightful information about dish/engine design is available from SolarPACES (Solar Power and Chemical Energy Systems), a program of the International Energy Agency. According to SolarPACES, the thermal receiver cooling fluid, usually hydrogen or helium, may be both the heat transfer medium and the working fluid for the engine.
Directly illuminated Stirling receivers adapt the heater tubes of the Stirling engine itself. Because of the high heat transfer capability of high-velocity, high-pressure helium or hydrogen, direct-illumination receivers can absorb approximately 75 W/cm². One design challenge is balancing the temperatures and heat addition between the cylinders of a multiple-cylinder Stirling engine. Liquid-metal, heat-pipe solar receivers are used to achieve that balance.
In a heat-pipe receiver, liquid sodium metal is vaporized on the absorber surface of the receiver and condensed on the Stirling engine’s heater tubes, evening out the temperature and permitting a higher engine working temperature. It is also possible to make a heat-pipe receiver that isothermally transfers heat by the evaporation of sodium on the receiver/absorber and condensing it on the engine heater tubes.
Stirling cycle engines used in solar dish systems use a hydrogen or helium working gas at temperatures of over 700°C and pressures as high as 20 MPa. Several mechanical configurations implement these constant-temperature and constant-volume processes. Most involve the use of pistons and cylinders. Some use a displacer (a piston that displaces the working gas without changing its volume) to shuttle the working gas back and forth from the hot region to the cold region of the engine.
For most engine designs, power is extracted through a rotating crankshaft. An exception is the free-piston configuration, where the pistons are not constrained by crankshafts or other mechanisms. They bounce back and forth on springs. A linear alternator or pump extracts the power from the power piston. Several excellent available references describe the principles of Stirling machines.
Some future dish/engine designs may use gas turbines, in which efficiency is over 80%, instead of reciprocating Stirling engines. In the most successful of these designs to date, the receivers have used “volumetric absorption,” in which the concentrated solar radiation passes through a fused silica “quartz” window and is absorbed by a porous matrix. Still, the Stirling engine is today’s norm.