PV Grid Parity Not as Close as Hoped?

IEEE Spectrum blogger Bill Sweet writes that despite encouraging statistics, PV grid parity is not as close as it seems.

Alexander E. Braun, Senior Editor -- PV Society, 11/16/2009

Writing about grid parity in IEEE Spectrum's Energy Wise, blogger Bill Sweet mentioned a recent Lawrence Berkeley National Laboratory report that concluded that the installed cost of photovoltaic systems declined by >30% from 1998 to 2008, from $10.80/W to $7.50/W. Sweet indicated that although this appears encouraging, it is not.

"According to one eye-catching allegation, there's a kind of Moore's Law in photovoltaics, which holds that costs come down by 20% with every doubling of installed capacity," Sweet wrote. "Rates of installation have varied in the last decade, of course, both year-to-year and world region to world region. What is more, the very idea of a photovoltaic Moore's Law is a bit slippery." The writer added that it is conservatively postulated that if the installed PV base has doubled roughly every two and a half years in the last decade, average photovoltaic costs should have come down about 60% since 1998, not 30%.

Horizontal inline sputter coater for architectural glass, with an annual output of up to 8 million square meters. (Source: Applied Materials)

Sweet observed that grid parity is put at $1/W. "That may be a somewhat too demanding standard, considering that photovoltaics work best on rooftops or integrated into construction material, so that electricity is consumed at the point of production, eliminating transmission and distribution costs. But even if grid parity were put at $2/W and installation costs declined at a rate of 30% per decade from the currently estimated $7.50/W, it would take PV electricity until roughly mid-century to become economically competitive."

European Union energy planners expect PV grid parity to be reached around 2015 in southernmost countries such as Spain and Portugal — at least when PV materials are used in solar concentrating systems. However, according to Sweet, there is no basis to expect conventional PV to become competitive that soon. "Not only are photovoltaics manufacturing costs not in the ballpark right now, they're so far from the ballpark, there's no way of knowing whether they'll ever be in the ballpark."

Sweet qualifies all this by pointing out that installation costs are not the best way of evaluating PV — or any other energy-generating system's — cost-effectiveness. "The generally accepted and standard measure of generating costs is the levelized busbar cost — that is, the cost of electricity at the point where it is fed into the grid, taking all cost factors into account (investment, financing, operation, maintenance, and so on)," wrote Sweet, adding that ideally, to evaluate claims made about grid parity or a PV Moore's Law, it would be necessary to measure performance in terms of levelized costs and in the same units seen on monthly utility bills — cents per kilowatt-hours.

"There are at least two insurmountable obstacles to our actually doing that," Sweet said. "One is that photovoltaic electricity is too new and its cost elements are changing too fast for anybody to make reliable estimates of its average levelized costs over time. Just as importantly, operators of PV plants are often very cagey about how much it is costing them to run the installations. That leaves us for better or worse with estimated installed costs. The good thing about using them as a metric is that the size and cost of any given PV plant is generally a matter of public record. So there are a lot of trade groups and energy monitoring organizations that add up the wattage and costs every year, on a global basis. The bad thing is that it would be a full-time job to figure out whose estimates are most reliable."

As an example, Sweet pointed to Daniel Yergin's Cambridge Energy Research Associates estimated aggregate that world PV installation costs were at ~$7/W in 2004, wind slightly under $1/W. According to Marketbuzz/Solarbuzz, average world PV installation costs came to ~$6.2/W in 2008, four years later. "That's an improvement of 11.5% in four years, much too slow a rate of improvement to give us grid parity any time in the foreseeable future," Sweet said.

Not all readers agreed, however. Mark Pinto, senior vice president, corporate CTO and general manager of the Energy and Environmental Solutions (EES) and Display businesses at Applied Materials Inc. (Santa Clara, Calif.), thought Sweet dramatically oversimplified the analysis, mixing retail and utility costs, and drawing a much too conservative conclusion with respect to solar economic viability. "The price (not cost) of $7/W is a retail price for a consumer rooftop — the historically predominant type of installation until recently," Pinto argued. He then pointed that the grid parity definition used by the blogger relates to cost to generated base load utility power. "You must compare retail panel prices to retail electricity costs," he said. "And retail electricity costs are a function of time of year and time of day (solar being most productive when retail electricity costs most)." According to Pinto, in California, the cost of electricity for the average consumer in mid-summer makes solar viable.

Pinto also indicated that, for utilities, large-scale solar installations cost less than half of residential (currently ~$3.50/W) and that this should be compared not against base load generation costs (e.g., from coal) but peak generation costs such as for gas turbines or diesel generation. "Further, since you are buying electricity for the next 25+ years with a solar panel, you need to compare against expected prices for the next 25+ years." He finished by saying that in some significant parts of the world we are near or even below peaking parity with utility-scale solar today.

Richard Reis, principal engineer at Conservation Engineering (Silver Springs, Md.), said that Sweet's argument of installed-cost grid parity overlooks two important facts: Fossil fuel power plants must pay continuous fuel costs, whereas the sun sends no bills. Thus energy cost, not installed cost, is a better metric. "Fossil fuel, nuclear, and large hydro plants impose high external costs on everyone and on nature," Reis wrote. "For coal, these externalities are $0.03/kWh for non-greenhouse air pollution, $0.06/kWh for land pollution from coal mining, and $0.02/kWh for greenhouse gas emissions." Reis added that these $0.11/kWh do not include the environmental costs of disposing of bottom ash and byproducts of smokestack scrubbers. When this is considered, solar energy begins to look much more competitive.