Wednesday, March 11, 2009

Storing Sunlight

An article in MIT's Technology Review on a new liquid battery technology got me rethinking an assumption I've been making for some time concerning the synergy between renewable energy and better batteries. The article's author makes a similar assumption, suggesting that with bigger, cheaper batteries, electricity from solar power might be supplied around the clock. But while this new battery, consisting of a combination of molten metals and molten salts, looks clever, I'm not sure it would make sense to use it to store solar electricity in the way the author envisions. The main impediment to the large-scale application of solar power today is not so much its cyclical nature--which storage can address--but its high cost per generated kilowatt-hour, compared to other technologies. The power likeliest to be stored for later delivery won't be the most expensive, but the cheapest.

My focus here is not on the rooftop solar panels being installed on homes. Storage isn't an issue in most such cases, unless you're in a remote location or insist on grid independence. Net metering--the ability to sell excess electricity back to the grid and buy power from it when the sun isn't shining--typically offers a much better deal for homeowners than batteries, by effectively using the grid as free storage. Since rooftop solar has the inherent advantage of competing with retail, rather than wholesale electricity prices, I'm more interested in the utility-scale solar installations springing up all over. These compete directly with the output of gas-fired simple-cycle turbines, the standard "peaking" power plant technology. Utility solar projects currently cost around $6,000 per installed kilowatt (kW) based on several recent project announcements turned up by a quick web search. Even with the 30% solar investment tax credit and a site in a sunny location, such as Florida, that results in an amortized cost of generation of roughly $0.25 per kilowatt-hour (kWh), based on a 20-year life and 6% interest rate. That might be acceptable for peak demand periods, such as hot, sunny afternoons, but it doesn't compare very well to off-peak wholesale power costs from other technologies, including wind and gas turbines, let alone coal or nuclear power.

Nor is the cost per kWh the only barrier solar power must overcome, in order to be competitive around the clock, even if the cost of storing it were negligible--which is certainly not the case today. The capital involved in amassing enough capacity to serve a given market 24/7 is much higher for utility-scale solar power than for other technologies because solar's capacity factors, reflecting the fraction of time when these facilities are available and generating peak power, often average below 20%. In the Florida example above, a solar array would receive an average amount of sunlight equivalent to 4-4.5 hours of peak sun per day. That equates to a capacity factor between 17-19%. Replacing the baseload power from a 500 MW coal-fired power plant operating at an average capacity factor of 80% would require 2,200 MW of solar power plus a commensurate amount of storage. So at $6,000/kW, a solar power plant capable of generating as many kWhs as a $1.5 B coal-fired plant would cost $13.2 B, excluding the cost of delivering power when needed, instead of when the sun happens to be shining. (It also implies a very high cost per ton for the avoided CO2 emissions.)

With current solar technology, the entire proposition of storing lots of solar power looks impractical and unnecessary. Using large-scale, cheap storage--of whatever technology, whether batteries, compressed air, or pumped water--to time-shift renewable power makes much more sense when applied to lower-cost generation from wind power, the normal output of which also has a much poorer overlap with typical daily and seasonal power demand curves than solar power. In most markets, solar power should be going after the premium associated with the afternoon demand peak. Solar needs little or no storage for that, other than to buffer the effects of cloudiness or extend its output by an hour or two on either side of its natural output peaks. That looks easiest with solar thermal technology, which stores energy as heat, rather than electricity. As a result, developers of new batteries should not pin their hopes on the growth of a market for storing solar power.

No comments: