Thursday, May 23, 2013

Can Energy Storage Make Wind and Solar Power As Reliable As Coal?

Wind and solar power generated 3.5% and about 0.1%, respectively, of US electricity last year.  These figures represent large increases from much smaller levels in the last decade as the cost of these technologies declined significantly, particularly for solar photovoltaic (PV) modules. However, other barriers to wider deployment remain, including their intermittent output.  Energy storage is often portrayed as the killer app for overcoming the intermittency of renewables, and a number of interesting developments have occurred on this front, including a new "hybrid" wind turbine with integrated storage from GE. To what extent could more and cheaper storage enable wind and solar to function as the equivalent of high-utilization, baseload generation? 

Assessing that potential requires, among other things, recognizing that energy storage is neither new nor monolithic. Nor is the intermittency of renewable energy a single challenge.  For example, the output of a wind turbine and the wind farm in which it operates varies on time scales of minutes, hours and days, as well as months and years.  The output of a PV installation varies somewhat more predictably, but no less dramatically. 

Generating companies and project developers have an array of new storage options, involving various battery technologies, flywheels, and compressed air. Pumped storage, in which water is pumped uphill and generates power later when it flows back downhill, is an old, though hardly obsolete option and already operates on a large scale. According to the National Hydropower Association the US has 22,000 MW of installed pumped storage. This, too, is expanding and remains one of the cheapest forms of power storage in terms of cost per megawatt-hour (MWh) delivered.   Enough new projects have received preliminary permits to more than triple that figure, in 23 states.

All of these storage alternatives have limitations or drawbacks.  Batteries and flywheels, while very responsive, are still expensive.  Compressed air storage often relies on unique local geological features, and some versions essentially function as a supercharger for a gas-fired turbine, resulting in some emissions. Pumped storage works well at a variety of scales but is less responsive than batteries, has a larger physical footprint, and requires suitable terrain. 

What makes GE's "brilliant turbine" with battery storage look clever is that, with the help of predictive models, it requires a very small amount of battery storage--perhaps as little as that in an electric car--to smooth the output of the turbine for 15 minutes to an hour. That provides significant benefits, including financial ones, in terms of integrating it predictably into the power grid. However, it doesn't transform the turbine into a fully dispatchable generator capable of sending power to the grid whenever demanded.  That would require storing much more energy per turbine and delivering it at rates sufficient to replace the entire output of the installation for at least several hours, along the lines of concentrated solar power installations with thermal storage.

Even these techniques don't get us to the point at which a dedicated wind farm or solar installation could replace a baseload coal-fired power plant of similar capacity running 80% of the time.  For starters, energy storage doesn't alter the total amount of energy collected from the wind or sun.  In an area with good onshore wind resources, generating the same energy as 100 MW of coal capacity would take around 267 MW of wind turbines, because the wind doesn't blow at optimum speed all the time, and other times it doesn't blow at all. The wind farm would also need enough storage to absorb any output over 100 MW, and then make up any shortfalls below 100 MW for the longest duration that would be expected.  The figures for a solar installation would be similar. It just doesn't sound very practical, unless storage became dirt cheap.

Fortunately for renewable energy developers, that isn't what grid operators expect of wind or solar.  In most situations the local grid takes their output whenever it's available, though not necessarily at the price that a generator capable of committing its capacity in advance or responding on demand would receive.  So there's a financial incentive for renewables to add a bit of storage to "firm up" some capacity, while bulk storage appears to be more desirable as a separate asset available to the grid, just like a "peaking" gas turbine, to support multiple renewable sources. Of course in that case there's no guarantee that the power stored would come from renewables.  It's likelier to come from whatever is the cheapest off-peak generation in that market.

So while it's easy to see how improved energy storage can enhance the economics of renewable energy and enable it to be integrated into the grid to a greater extent than otherwise, it's less obvious that even cheap, large-scale energy storage is a panacea for intermittent renewables like wind and solar.  It might even have greater benefits for low-emission but more reliable forms of generation, such as nuclear and geothermal, by allowing them routinely to shift a set portion of their output into more valuable segments of the regional power market. 

Disclosure: My portfolio includes investment in GE, which makes products mentioned above.

2 comments:

Ed Reid said...

Geoff,

Assuming that a wind generator would deliver its rated capacity for 8 hours per day, on average, it would require storage capacity equal to the total daily output of the generator to make that capacity dispatchable, ignoring losses into and out of storage and deviations of daily output from the average. While that storage would make the output of the wind generator more valuable, it would also make it far more expensive.

The issue becomes far greater if the wind generation is to displace baseload power, as some in the wind industry and the environmental community suggest. Baseload replacement would require 3 wind generators to provide the output of 1 generator 24/7, plus storage capacity equal to approximately 2/3 of the average daily output of the 3 generators, again ignoring in/out losses and deviations from average daily output.

Achieving four nines reliability for the supported grid would require both a greater number of turbines and greater storage capacity. This, again, would increase the value of the output, but also the cost.

jtf said...

Worth noting that CAES isn't just difficult to implement due to the need for local geology to cooperate - I used to live in upstate NY, where more than a few CAES schemes were scuppered because the alternative use for old solution-mined salt caverns is for natural gas storage. Needless to say, the CAES folks were priced out.