Which is cheaper -- Rooftop solar or utility-scale solar?

The Brattle Group,with support from the Edison Electric Institute, just released a study concluding that utility-scale solar photovoltaic (PV) systems in the United States are more cost effective than residential-scale (rooftop) PV systems in achieving the economic and policy benefits we all expect to come from the widespread use of solar energy.

But it may not be that simple and depends on what goal you have as a user. The study was commissioned by First Solar, one of the largest solar companies in America. Titled Comparative Generation Costs of Utility-Scale and Residential-Scale PV in Xcel Energy Colorado’s Service Area, this study is the first to focus on a “solar to solar comparison of equal amounts of residential-scale and utility-scale PV solar deployed on an operating utility system.”

The study found that the cost of generating energy from 300 MW of utility-scale PV solar is roughly one-half the cost per kWh of electricity produced from an equivalent 300 MW of 5kW residential-scale systems when deployed on the Xcel Energy XEL +1.34% Colorado grid. Utility-scale solar remained more cost effective in all scenarios considered in the study, scenarios having different tax credits, monetizations, and inflation rates.

A new study concluded that utility-scale solar photovoltaic (PV) systems in the United States are more cost effective than rooftop PV systems in achieving the economic and policy benefits we all expect to come from the widespread use of solar energy. Photo courtesy of Energy Northwest The study also concluded that 300 MW of PV solar deployed in a utility-scale configuration avoids approximately 50% more carbon emissions than an equivalent amount of residential-scale PV solar.

The large difference in costs between utility- and residential-scale systems was attributed to economies of scale and greater solar electric output resulting from optimized panel orientation and tracking assumed for utility-scale systems. The improved orientation and tracking of utility-scale solar resulted in a higher capacity factor than for rooftop solar.

Using actual historic data from Xcel Energy Colorado, the study compared the per-MWh customer supply costs of adding 300 MW of PV panels in the form of either 60,000 distributed 5kW rooftop systems owned or leased by retail customers, or 300 MW of utility-scale solar power plants that sell their entire output to Xcel Energy Colorado under long-term power purchase agreements. The study found that projected utility-scale PV power costs will range from 6.6¢/kWh to 11.7¢/kWh across all scenarios, while projected power costs for a typical, customer-owned rooftop PV system will range from 12.3¢/kWh to 19.3¢/kWh.

“Over the last decade, solar energy costs for both rooftop and bulk-power applications have come down dramatically,” said Dr. Peter Fox -Penner, Brattle principal and co-author of the study. “But utility-scale solar will remain substantially less expensive per kWh generated than rooftop PV. In addition, utility-scale PV allows everyone access to solar power. From the standpoint of cost, equity, and environmental benefits, large-scale solar is a crucial resource.”

But there is another aspect to rooftop solar that does not seem to have been captured in this analysis – the fact that rooftop solar does not take any additional space, does not take any additional hook-up or transmission lines, and does not take any additional buffering, or load-following, beyond what the local grid already has. Rooftop solar is truly distributed while utility-scale solar is not. And individual rooftop solar users become more self-sufficient and reduce the need to build additional utility-scale generation.

The study also used existing power-purchase agreements with utility-scale PV which vary from market to market, and include more than just actual production costs. The numbers for the rooftop systems did not include many of the subsidies other states, like Washington, provide to rooftop customers, such as buy-back at 54¢/kWh.

The study does point out this lack on page 11, “It is important to understand that all of our cost results include only the customer-paid costs for the generation from equal amounts of PV capacity deployed in two configurations in one particular utility service area. A complete tally of the differences between equal amounts of the two types of PV capacity would require that these two resource options be alternatively embedded in a complete, subsequently optimized integrated resource plan (IRP) for Xcel Energy Colorado or other systems of interest. When optimized, such an IRP would reflect the effects of each PV option on system costs and potential benefits such as savings (or incremental reinforcement costs) on transmission and distribution outlays, and differences in ancillary service costs.”

Warren Buffet recently purchased theworld’s largest photovoltaic solar array in Bakersfield, California. It is a 5-square-mile 579 MW PV array that cost a little over $2.2 billion. Assuming a capacity factor of 25% over the expected 25-year lifespan, this utility-scale PV will generate 32 billion kWhs in total:

579 MW x 1000 kW/MW x 8766 hours/year x 0.25 x 25 years = 32 billion kWhs 

There are no obvious fuel costs, but PV solar has O&M costs of about 1.3¢/kWh, which comes to about $400 million over the life of this array. So to produce 32 billion kWhs at about $2.3 billion means a life-cycle cost of 7¢/kWh. This is getting close to the range of normal baseload providers like coal, nuclear and hydro, which have life-cycle costs of 5.1¢/kWh, 4.1¢/kWh and 2.7¢/kWh, respectively.

Note that these are actual costs to produce a kWh, not levelized costs, and do not include finance issues, subsidies or special markets, taxes and other costs that are not construction, production or fuel. Levelized costs for utility-scale solar, coal, nuclear and hydro are 13.0¢/kWh, 9.8¢/kWh, 9.3¢/kWh and 11.7¢/kWh, respectively, very different from actual costs.

In contrast, the citizens of Washington State have recently spent $50 million to install 14 MW of rooftop solar. In addition to the 30% installation tax credit, the State of Washington is giving 54¢ for every kWh given to the grid. Since that is about a third of what is produced, there is no cost for electricity for these people, and this return means the payoff-period is now just several years.

As one of these citizens, I just put a 4 kW rooftop solar array on my house. The system I put on my roof in February is the latest and greatest, and over its lifespan will produce about 150,000 kWhs of electricity for about 13¢/kWh. However, I have been selling a third of that back to the grid at my present net metering rate of 54¢/kWh, which more than covers the electricity I buy from the grid at only 7¢/kWh. Over the next several years, I’ll make about $6,000 in profit after subtracting the original capital costs of the array. Assuming this net metering windfall ends in 2021, and I have little O&M and decommissioning costs, I’ll get about a third of my electricity free for the subsequent 20 years.

Since the electricity I do buy from the grid is hydro and nuclear, I have a ridiculously low carbon foot-print to boot. In the end, I will have gotten this 150,000 kWhs for 0¢/kWh. The power I do buy, will still be at about 7¢/kWh.

So maybe the utility-scale solar is best for America, but my rooftop system suits me just fine.

Post-Post addition: A professor at MIT just pointed out that “grid costs are larger for distributed solar and to look at the MIT Future of Solar Energy study. With distributed solar, the local distribution has to be sized to peak solar rather than peak demand. With lots of solar, peak solar is bigger than peak demand implying much more expensive distribution grid, and distribution grids are more expensive than the main grid. For very small amounts of solar, no real impact one way or the other.Also personally, it is a good financial move to put in rooftop solar in Washington State or Massachusetts that have huge subsidies. However, those subsidies are basically a regressive tax. Massachusetts is the extreme case where the big rooftop systems are only on big very-expensive houses in the surburbs,” as pointed out in my previous post on net metering. 

http://www.forbes.com/sites/jamesconca/2015/07/30/which-is-cheaper-rooftop-solar-or-utility-scale-solar/?ss=energy