Moore’s Law is a seductive but slippery concept. Strictly speaking, it refers to the 1965 observation by Intel founder Gordon Moore that the number of transistors possible in an integrated circuit doubles about every two years. But much to Moore’s chagrin, the term often is misused. “If [Al] Gore invented the Internet,” he once joked, “I invented the exponential.”
Moore’s law is casually, and often erroneously, applied throughout the digital technology sector to any exponential increase (speed, density) or reduction (cost, size) over time. Recently, the notion has ventured even farther afield and is increasingly mentioned in relation to the cost and performance of photovoltaics (PV).
“The PV buzz isn’t really analogous to Moore’s Law, which is just an engineering statement,” says Bill Sweet, editor of Spectrum, the flagship magazine of IEEE—the Institute of Electrical and Electronics Engineers. “This PV law is about costs, and of course costs depend on a lot of things. However, when you say PV Moore’s Law, you do get a lot of attention.”
On the surface the argument is simple. Global PV production grows by about 50 percent a year, so capacity doubles about every 18 months. According to the PV version of Moore’s Law, every time capacity doubles, the cost per watt comes down 20 percent. Project those gains over time, and somewhere around 2015, PV will achieve cost parity with traditional central generation.
Of course it’s not that clear-cut. PV isn’t a single technology, but an increasingly broad range of products, making any sweeping generalizations dubious. Further, government subsidies have a substantial effect on cost and proliferation.
Still, the caveats and qualifications don’t negate the fundamental point behind the Moore’s Law fixation: Every day PV technology is getting more effective and less expensive. Researchers and manufacturers may cite different numbers, but the numbers all point in the same direction.
“When we went public we made a commitment to our investors that compared to 2006 we’d reduce our installed system cost by 50 percent by 2012, and that we’d be two thirds of the way there by 2010,” says Julie Blunden, vice president of public policy and corporate communications for the solar manufacturer, SunPower Corp. “We’re on track to meet that.”
That kind of cost reduction, coupled with PV’s site flexibility and speed to market, is transforming the way the utility sector uses solar. Utilities that once reluctantly dabbled in retail solar and sun farms only to satisfy regulatory requirements are beginning to actively pursue these markets.
Moreover, even investor-owned companies are beginning to explore the potential of utility-owned PV for distributed generation. Once shunned, distributed PV generation actually may prove to be the fastest and cheapest way to expand capacity, meet renewable standards and manage peak load.
In the United States, distributed generation traditionally is considered, even in the public policy domain, as residential and commercial solar, and small-scale industrial combined heat and power systems. The concept of small, utility-owned generation plants spread throughout the grid and particularly near load centers rarely enters the discussion. When it does, usually it gets the stiff arm.
Reticence is understandable. Central generation affords economies of scale, proven and familiar technology and a business model that utilities understand. Distributed generation is perceived as a third-party-owned asset that cuts into the rate base and adds a complicated and unprofitable layer to the system. But as emissions restrictions tighten, renewables standards increase and the cost and efficiency of PV improve, then the calculus shifts.
“In other parts of the world, ‘distributed’ is actually the inherent nature of energy production and distribution in general, especially anywhere that incurs any type of social unrest or has difficulty maintaining centralized infrastructure,” says Anne-Marie Borbely-Bartis, co-author of the book, Distributed Generation: Power Paradigm for the New Millennium. “In places like Afghanistan, parts of Central and South America and Africa, if you build a centralized system, it’s simply going to be torn down again during the next social upheaval.”
The same attributes that make distributed generation a third-world solution—speed, flexibility and security—have real appeal in a first-world preoccupied with load growth, carbon constraints and cyber-attacks. But realizing those benefits in a mature grid isn’t a simple matter of plug-and-play.
The difficulty with a robust national distributed generation system is the need for a grid that can accept two-way control of the electrons themselves. That requires fully automated sub-stations at the transmission and at the distribution level. The cost of automating all of the sub-stations across the country is in the hundreds of billions of dollars.
That process of automation already is in progress—albeit in a piecemeal and sometimes halting fashion—as North America gradually builds the smart grid. The perhaps unintended result of the capital and technological investment utilities now are making to enable smart meters, demand response and greater efficiency in general is a framework that also makes viable distributed generation.
“Some enlightened utilities are looking at this because other options are limited. The traditional approach of, ‘lets go build another big coal plant and string some wires up,’ is less of an option right now,” says David Sweet, executive director of the World Alliance for Decentralized Energy [no relation to the IEEE’s Bill Sweet]. “There are a lot reasons —climate change, renewable portfolio requirements, the need to solve some of these intermittency issues. The smart grid isn’t just a bunch of switches and meters, it’s also how you make power available where and when people need it.”
For example, an oft-cited benefit of distributed solar is its relative proximity to the end user, and the efficiencies that can result from that close proximity. A lot of power bleeds out of the system between a central power plant and its customers, but distributed power is used locally, dramatically reducing line loss. Factor in solar’s afternoon peaking profile, and its value in load management is amplified.
Such complementary relationships are real considerations, but they’re not the kinds of things that tend to drive investor-owned utilities to change their generation profile. Lawmakers, regulators and even public power utilities and cooperatives might be more apt to follow big-picture environmental or efficiency principles, but IOUs look primarily to the bottom line. For them, the most significant impetus toward distributed solar stations is a lot simpler, and more compelling. It’s the stark logic inherent in Moore’s Law.
“Eventually, producing your own electricity will literally cost less than having it produced by somebody else off-site, and having it shipped to you—at which point, why would you opt for centralized power?” says Borbely-Bartis.
“There’s actually an enormous business opportunity built into this, if utilities were interested in seeing it as something other than a threat,” she says. “Why shouldn’t utilities be the owner-operators of these rooftops? They would be responsible for replacing them, for any maintenance that might be required. It actually becomes a capital asset built into the rate base, as though it were a central power plant.”
A handful of utilities are doing just that.
To comply with North Carolina’s ramping renewable energy regulations, Duke Energy formulated a two-pronged strategy. One side is a third-party-owned solar farm—traditional central generation. The other is close to the model Borbely-Bartis describes.“We think of these installations truly as mini-power plants,” says Duke spokesman Dave Scanzoni. “By that I mean, these will be Duke Energy-owned, Duke Energy-maintained, and Duke Energy-installed. It’s just like a regular coal or nuclear plant in terms of the ownership and maintenance.”
The difference is location: Duke’s solar plants will be sited on the property of utility customers. Duke will pay customers a monthly rental fee based on the size of the installation and the amount of energy generated.
The $50 million, 8-MW project was approved by the North Carolina Utilities Commission in May, and the utility currently is selecting sites. Duke will begin installing between 100 and 400 distributed solar plants on the roofs and grounds of schools, homes, shopping centers and factories this fall. The total number of plants ultimately depends on the size and location of the sites chosen.
It’s a relatively small first step, but it might be just the beginning of a long-term program: Duke expects the solar infrastructure to last 25 years.
“We think distributed generation will be a growing sector of the renewable world,” Scanzoni says. “It won’t replace base-load power plants, but it gives us another option for power production. During peak periods—when it’s in the upper 90s on summer days—that’s when the sun is most available. It complements our generation mix to have solar panels generating electricity during those hours.”
A key aspect of the program is that it doesn’t involve any net-metering relationship with site owners, which make the installations particularly attractive for Duke from a business perspective. “The power goes back into the grid just like the regular base-load model,” Scanzoni says.
Additionally, with Duke and other utilities facing the prospect of federal renewable electricity standards, distributed solar generation offers a fast way to meet regulatory requirements. However, Scanzoni says Duke sees value in the business model beyond compliance, and plans to seek ways to implement it throughout the company’s service territory.
“Building these facilities is a lot simpler than building a large solar farm, in terms of the siting issues and transmission lines,” he says. “We’ll study this project to see how these sites integrate into the grid. It will provide a laboratory to analyze ways to expand as we see fit. There’s no question, it will be a growing part of the mix going forward.”
Duke’s project is a start, but it’s tiny compared to what’s slated in California.
The California Solar Initiative (CSI), announced by Governor Schwarzenegger in 2004, mandates a million solar roofs in the state by 2018 with a total capacity of 3 GW. The CSI aims to make the state a world leader in solar generation, and it’s ahead of schedule. Some 50,000 customers already are engaged in solar net metering, generating 500 MW of power. The program is directed toward consumers, but has sparked surprising interest in other quarters.
“We didn’t expect to have all three utilities propose to own their own solar distributed generation,” says SunPower’s Blunden.
In addition to the 3 GW of net metered solar, the California Public Utilities Commission (CPUC) has approved 1 GW of central-station solar farms and 2 GW of utility-owned distributed PV.
“A couple of years ago we had just a couple of megawatts installed,” Blunden says. “California now has an incremental 6-GW commitment by about 2015. The great thing about solar is that it’s really fast to market. We as an industry put 3 GW into Spain in 18 months. So if you want lots of solar fast, you can do it with PV.”
One of the major utilities targeting distributed PV is Southern California Edison. In March 2008, the company proposed installing 2 square miles of advanced PV panels on about 150 large commercial roofs. In June, the CPUC not only approved the project, it doubled it.
“In five years we’ll have four square miles of panels on about 300 roofs, generating 500 MW,” says SCE spokesman Gil Alexander. The first already is operating in Fontana, east of Los Angeles. “We went ahead and started even before we got regulatory approval. It has 33,700 panels, 2 MW of generating capacity. The day we plugged it into the grid last fall, it became the largest single solar PV installation in California.”
Additional installations will prioritize load-growth areas, such as Fontana and Chino in California’s burgeoning Inland Empire region. The other main criterion for site selection is the similarity of roofs. Like Duke Energy, SCE will lease from customers, but the utility is relying on uniformity to keep engineering and installation costs down.
Ideal sites won’t have a lot of power demand of their own. “Otherwise, the owner might be thinking, ‘Boy if I just had my own panels on my roof, I could cut down this big electric bill,’” Alexander says. “We offer a monthly check to lease their roof. It’s added revenue for them, and it’s valuable, unused southern California real estate that we can use for solar generation. Then, literally, we’ll run a wire off the roof, under the parking lot and to the nearest power pole, and the juice goes into the neighborhood.”
SCE’s distributed-solar business model treats the facility just like any other power plant, and the company plans to invest $875 million in new installations. Alexander is careful to point out, however, that it’s not the beginning of other major distributed-generation projects for SCE.
Under California’s decoupled regulatory scheme, SCE is allowed to generate only 30 percent of the power it sells; it must purchase the rest on the market. Therefore, its generation projects are carefully selected to test and cultivate markets. In this case, the seed nature of the project even was explicit in the CPUC’s approval.
“When we move into generation, there’s always some unique one-off rationale, and this project is motivated primarily by a desire to stimulate the market,” Alexander says. “When the public utility commission approved our project, they also said to do a competitive solicitation among independent power producers, invite them to match us and sign long-term contracts for the power they produce on their roofs.”
As a utility administrator of the CSI, SCE witnessed retail level activity on the small end of the market, and it observed strong interest and rapid development in central-station solar. But the company saw a gap in the market.
“We noticed there was almost no activity in the middle, and we thought we could prime the pump,” Alexander says. “We think the [commercial solar roof] project, because of its scope, will help bring down the price of solar rooftops for everybody. It will influence the market and maybe even the transform the market. Who knows?”
The speed of SCE’s actions seems to demonstrate the company’s commitment to the concept. The distributed solar project went from initial proposal to switching on the first roof in just six months—a pace that surprises even an unabashed solar booster like Blunden.
“You couldn’t have convinced me two years ago that we would see what we see right now in California with PV. It’s absolutely stunning,” she says. “Utilities often aren’t really happy about transformations, whether it’s wholesale energy markets, divestiture of assets or letting people use your billing systems to bill retail customers. And yet here we are with utilities embracing the idea that they’re going to own solar.”