Like other California electric utilities, San Francisco-based Pacific Gas & Electric (PG&E) has been scrambling to meet the state’s renewable portfolio standard (RPS), which requires suppliers to obtain at least 20 percent of their power from renewable energy sources by 2010.
Though the RPS includes a variety of technologies, renewables developers are choosing utility-scale solar power more than any other resource, says Hal La Flash, PG&E’s director of emerging clean technologies.
“We solicit offers for renewable energy every year,” he says. “This year (the response) was 80 percent solar. Part of that is because the best wind and geothermal sites are already under contract. But solar, especially photovoltaic technology, lends itself to a modular approach and you can develop more projects on a smaller scale.”
Without question, solar power is hot these days. PG&E alone has contracted for more than 1,500 MW of photovoltaics (PV), including 250 MW it will own and operate, and nearly 3,800 MW of thermal solar power capacity. Other utilities, including Florida Power & Light and Southern California Edison (SCE), are following suit. And that, in turn, is driving R&D investments designed to improve solar plant output, reduce equipment costs, and challenge the argument that solar power will cause headaches for fine-tuned utility operations.
“Solar power is really booming globally. In Spain they announced 2.5 GW of photovoltaic projects in 2008 alone, which is a staggering number,” says Larry Stoddard, manager of renewable energy at Black and Veatch. “The United States is lagging behind a bit. Right now we have 6,700 MW in solar-thermal PPAs [power purchase agreements] and roughly 2,400 MW in PV PPAs.”
If the trend continues, these numbers might be just the beginning of a huge boom in solar construction. Connection applications for a staggering 31,000 MW of thermal and 14,000 MW of PV have been submitted to the Southwest Power Pool Interconnection, and right-of-way applications for 100,000 MW of solar thermal and 46,000 MW of PV have been submitted to the U.S. Bureau of Land Management.
“Nowhere near that amount of capacity is ever going to be built, but the sheer number of applications demonstrates the feeding frenzy that exists in the United States right now,” Stoddard says.
Indeed, frenzy is the word.
“There are now 34 states with RPS and electric utilities are scrambling to meet their requirements,” says Scott G. Smith, U.S. national clean tech leader with Deloitte & Touche. “Utilities are under a lot of pressure to get renewables contracted, even if they won’t have the actual capacity until later.”
A report released in September by UBS Wealth Management predicts growth in the global market for thermal, or concentrated solar power (CSP), will reach almost 20 GW over the next decade. All the major CSP technologies, from parabolic trough and parabolic dish, to power tower and compact linear Fresnel reflector (CLFR) designs, are about to be demonstrated on a utility scale in the American Southwest, mainly California. All of them are being financed and built by the technology suppliers themselves.
Which design will prove the most cost-effective remains to be seen. Parabolic trough technology, with its curved mirrors that direct heat to absorber tube receivers filled with heat transfer fluid, has been employed in California and other parts of the world for decades. As such, the technology generally is viewed as proven, and massive utility-scale projects are now in development. For example, Israel’s Solel Solar Systems Ltd., acquired by Siemens in October, is building what will be one of the largest parabolic trough plants ever. The 533-MW Mojave Solar Park will cover up to 6,000 acres, or nine miles, of California’s Mojave Desert. The plant, which will include 1.2 million mirrors and 317 miles of vacuum tubing, is scheduled to begin operation in 2011. PG&E has contracted to buy its output.
Similarly, Spain’s Abengoa Solar is building a smaller 280-MW parabolic trough plant 70 miles southwest of Phoenix, near Gila Bend. The Solana Generating Station is unique because it will employ a thermal storage system consisting of large insulated tanks filled with molten salt that will store heat generated during the daytime. The heat will then be used to produce energy during periods of low or no sun, including the evening hours. The plant is scheduled to go into operation in 2011, with Arizona Public Service buying the output.
But as impressive as such projects are, some analysts believe trough technology ultimately will give way to the more advanced CLFR and power tower technologies, both of which are under development.
A 2008 report by San Francisco-based Cleantech Group, which monitors the solar power industry for the investment community, predicts trough-based systems will be most prevalent until 2012 or 2013, but could then be displaced by power towers, CLFR and dish-engine developers.
“Troughs will dominate the first generation of CST,” states Brian Fan, Cleantech’s senior director of research. “It’s project financeable today because we know the costs, we know the technologies, and there are no technology risks. Developers have a roadmap to bring down the cost. But if the power-tower concept is proven in the field in test operations, because of higher thermodynamic efficiency and higher scalability, it will be the next generation of CST plants past 2012.”
Like a trough design, the CLFR technology uses the reflectors to concentrate sunlight on transfer fluid-filled pipes. However, the pipes run directly above the mirrors, which, developers say, increases the amount of heat generated. Power tower designs do away with the transfer fluid and piping altogether. Instead, they use mirrors to direct sunlight directly to a boiler system located within a tower.
Developers of both technologies have 5-MW pilot plants operating commercially. In October 2008, Australian-based developer Ausra started up its Kimberlina CLFR plant in Bakersfield, Calif. And in August, Pasadena-based eSolar brought on-line the 5-MW Sierra SunTower plant in Lancaster, Calif. Like Ausra, eSolar intends to scale-up the technology with a 245-MW solar thermal power plant in the Antelope Valley region of southern California. SCE will purchase the output.
Power tower supplier and developer BrightSource Energy has the boldest development plan to date. The Oakland-based company says it will build and own some 2,600 MW of power tower capacity, with seven plants generating 1,300 MW to be delivered to SCE and seven more generating another 1,300 MW under contract to PG&E. Its first plant, the 440-MW Ivanpah Solar Power Complex, will be located in Ivanpah, approximately 50 miles northwest of Needles, Calif., and about five miles from the Nevada border.
The plant will be constructed in three phases, two 110-MW facilities and one 220-MW facility. The first phase is scheduled to begin construction in early 2010 and be completed by 2012. The second phase will begin construction roughly six months after the start of the first. The plant is located near existing transmission lines and will require relatively minor line upgrades to connect and deliver its power to PG&E.
BrightSource is led by the original management team of Luz International, which developed a group of nine CSPs known as the Solar Energy Generating Systems (SEGS) in the Mojave Desert between 1984 and 1991. Those plants, now owned by FPL Group and SCE, employ parabolic trough technology and generate a cumulative 310 MW. The Luz team is now embracing a power tower design it says can generate more heat, and therefore steam, to drive today’s higher-efficiency steam turbines, says spokesman Keely Wachs.
“The holy grail in this business is high temperatures and high pressure steam. With the trough technology we could get maximum steam temperatures between 300 to 400 degrees [Centigrade]. With the power tower design, we can get 550 to 650 C. You’ve got to reach those temperatures to take full advantage of today’s advanced steam turbine technologies. Otherwise it’s like putting a 486 chip in a new computer,” Wachs says.
Concentrating the solar energy directly on the boiler, he says, results in a 15-percent efficiency gain over a trough design. Further efficiency gains are realized with heliostats that can be raised and lowered to optimally track the sun as its angle changes with the seasons. And though it decreases efficiency and increases the plant’s capital cost, dry cooling will be employed to reduce its water requirements.
“Wet cooling uses 20 times more water,” Wachs says. “Dry cooling supports a closed-loop system that continually feeds the steam cycle. Ivanpah will require 100 acre feet of water, which is the equivalent of 300 homes worth of water a year. Most of the water will be used to wash the mirrors. You do take a hit with dry cooling, but we believe wet technology will have a hard time getting permitted because the West is so water sensitive.”
Demand for PV also is on the rise, as is size of new PV plants. However, unlike with CPS technology, some utilities are taking an ownership position. Both PG&E and SCE, for example, have announced they will build, own and operate 250 MW of PV assets over the next five years.
Growing interest from utilities is driving PV suppliers to lower prices and improve efficiency across the value chain. That means PV cells, panels and tracking systems are becoming cheaper and more efficient and cost less to install, operate and maintain.
While improving cell efficiency basically entails altering chemical recipes, panel costs are further reduced by streamlining manufacturing processes. Still, deciphering which PV technology is best for an installation basically comes down to site location and the amount of available space.
The two primary cell technologies are crystalline poly-silicon and thin film. Crystalline technology is the most expensive to produce but generally offers the best cell efficiency, typically ranging from 16 to 20 percent. Thin film technology is cheaper to produce but also less efficient, at or slightly above 10 percent, depending on the supplier. Therefore, thin film requires at least twice the physical space to produce the same energy as crystalline.
But there are other PV flavors too, including cadmium-telluride (CdTe), a type of thin-film technology that offers higher efficiency, typically around 12 to 16 percent, and copper-indium-gallium diselenide (CIGs), which offers cell efficiency levels comparable to crystalline but involves comparatively unproven manufacturing processes.
With its high efficiency rating, crystalline generally is considered most advantageous where space is limited. But some argue its output can degrade in hot desert climates. Thin film performs well in hot climates and is a good fit where space isn’t an issue.
Regardless of the technology, all cell producers are working to boost cell efficiency and reduce manufacturing costs. For example, earlier this year First Solar, the Tempe, Ariz.-based CdTe-thin film panel supplier, said it has reduced its cost for solar panels to $1 a watt, and claimed the lowest manufacturing cost in the industry. The company reduced its per-panel manufacturing cycle time to roughly 2.5 hours, which it says is one-tenth the time it takes to manufacture a crystalline panel.
San Jose-based crystalline panel supplier SunPower, on the other hand, announced in October it soon would begin producing panels with a 20.4 efficiency rating, which it says is the highest ever efficiency rating for a full-sized solar panel— a claim confirmed by the National Renewable Energy Lab (NREL). It also introduced the T20 tracker, a single axis ground mounted system the company says has fewer moving parts, a simpler mechanical structure, better wind resistance and is pre-assembled to speed installation.
“You’ve got a 1.5 HP motor at the end of a whole string of trackers. It works like a Venetian blind,” says Julie Blunden, vice president with SunPower. “We’ve lowered the manufacturing cost, the O&M costs, and generated more watts per tracker.”
Ironically, the announcements coincided with the commissioning of FPL’s 25-MW DeSoto Next Generation Solar Energy Center in Arcadia, Fla. The plant, now the country’s largest utility-owned PV plant, employs SunPower’s previous solar panel design, rated at 18.7 percent efficiency, and solar tracker equipment.
“In the traditional utility cycle, the time between commissioning a new power plant technology often is measured in years or decades,” says Blunden. “With the solar industry, it’s entirely different because there are multiple opportunities for technological improvements across the value chain. You can measure changes on a quarterly basis. But the customers know that. We tell them, ‘Here’s where the cell efficiency will be a year from now.’”
Over the next five years, SCE expects to blanket some 65 million square feet of unused Southern California commercial rooftops with 250 MW of PV technology. Two facilities have gone commercial so far—a 1-MW system in Chino, and a 2-MW facility in Fontana. First Solar supplied the panels for both. Among the many takeaways of these projects is the ability to measure each system’s impact on the California ISO grid operations.
“In the last year, renewable power has come to represent 16 percent of our total power delivery,” says SCE spokesman Gil Alexander. “As our renewable resources increase, it’s obviously going to be a challenge to keep the grid in balance. These two plants will help us determine how to best integrate a variable, intermittent power supply with conventional, dispatchable resources. We intend to share those findings with the industry.”
The key to such a balancing act is the inverter technology that transforms the direct current (DC) produced by PV panels to alternating current (AC) and sends that current to the distribution grid. Technology advances at this interface are making PV power more grid-friendly. Boston, Mass.-based Satcon, which supplied the inverters to the Fontana installation, announced in September a third-generation product called Satcon Solstice, a utility-scale inverter for large solar farms. With Solstice, both the plant and the interconnection can be monitored and controlled remotely.
Individual circuits continually will monitor the operations of each array, giving the operator the ability to isolate or even shut down individual panel threads that are experiencing technical problems and undercutting the array’s performance. The ability to control each individual thread optimizes the amount of electricity harvested from the PV facility.
At the same time, the AC side of the inverter system is configured to allow the utility to monitor the facility’s output in real time. If a change in the weather is predicted—perhaps thunderstorms are expected in the afternoon—the utility can monitor the output to closely synchronize the time it dispatches other generation to replace lost solar capacity.
“The inverter is the brain and the muscle of the PV system,” says Michael Levi, Satcon’s senior director of marketing. “With a 1-MW system or larger, the inverter gives the operator a window into the PV operations and a control device that gives the utility 100ths of a second reaction capability. Most utilities are reluctant about solar power because they view it as unstable and disruptive. They need reliability and performance, along with an interconnect that lets them on-board the power in a comfortable manner. That’s what this system provides.”
Indeed, with solar plants becoming vastly larger than they once were, utilities require much greater ability to manage and control their output. Not long ago, a 1-MW PV facility was considered a major technological achievement. Now Chicago-based Exelon is building an 10-MW facility on the city’s south side; FP&L owns and operates the 25 MW DeRosa facility; SCE and PGE are each looking to develop and own some 250 MW of PV over the next five years; and CSP developers are looking to add thousands of megawatts in new capacity to the nation’s grid (see Figure 1).
So solar certainly is on a roll, but much work remains to be done.
“I was recently on a plane sitting next to this investment banker type who’s telling me about a solar project his firm had gotten involved in somewhere in the Southwest,” says Ryan Sather, senior manager of generation and energy markets for global business consultant Accenture. “He said the one area they failed to consider was the cost of keeping the mirrors clean. They hadn’t anticipated the dust and the annual cost of having to clean it. It wasn’t part of the financial model and that tripped them up.”
From an electric utility standpoint, Sather’s story is an apt metaphor. While there’s plenty of optimism, there’s still plenty to prove—and the industry is bound to make mistakes.
“Solar makes a lot of sense in certain parts of California and the Southwest,” Sather says. “But utilities are used to dealing with rotating equipment and this is a whole new animal. A lot of questions still need to answered. How big does the plant have to be? What’s a cost-effective capacity rating? How will solar impact grid operations? And what are the O&M costs?”
The current crop of plants might provide some answers, giving utilities the technology experience they need to make the most of solar energy’s expanding future.