Friday, August 28, 2009

PG&E Compressed Air Project - Quick Thoughts

Now that the deadline has passed to file for grid storage projects under the ARRA , we're beginning to see some of the concepts and projects that have applied for funds - see post below - including the 300 MW compressed air project by PG&E I believe PG&E will have to deal with some of the following issues on this project.
  • First, the $25 million requested is only for "initial analysis and design". The anticipated cost for the project will be $368 million. And that's before we see the cost over runs, delays and unexpected problems a huge project like this will invariably incur.
  • Then the utility will have to explain why they want to take "clean" wind energy and make it "dirty". Because, you see, the compressed air will be used for natural gas turbines! The argument is that the compressed air will make the natural gas turbines run more efficiently, requiring less natural gas - which is great if you're trying to make your natural gas turbines more efficient. But we thought the point of wind energy was to produce clean and renewable power - not more fossil fueled power.
  • And how much energy will we lose in this process? We create a certain amount of energy with wind and then burn it away by compressing it and burning natural gas. What is the net delivered energy when all of this is finished? I've seen reports of as little as 54%. So we take wind energy, throw away 40% or more, increase the price volatility through the natural gas market, and add emissions and GHGs. Why is this a good idea?
Utilities like CAES because it increases their power as a utility. They get to spend a bunch of ratepayer money on a huge central power plant which they control. This is why they have problems with distributed advanced batteries like the VRB-ESS. We could install 300 MW of the VRB-ESS in less time then they can get a CAES plant - if they ever get it built at all. And, the batteries would be built where it's needed, close to the load, reducing the need for more transmission wires, reducing the cost of distribution, improving energy security and power quality, and with greater efficiencies - less loss of wind energy - no emissions and no volatility on the cost of power.

5 comments:

Samir Succar said...

Flow batter technology holds promise, but promoting one technology doesn't make it necessary to create false arguments against other technologies.

Its true that CAES burns natural gas, but the fuel efficiency is very high and the total greenhouse gas emission rate of wind+CAES baseload power is much lower than any almost any other baseload power option. Its one fifth the carbon emissions of even the cleanest natural gas plants.

The efficiency of CAES is not 54%; this is simply based on erroneous information and a fundamental misunderstanding of the thermodynamics of CAES. Actual roundtrip efficiencies are actually closer to 75-85%

For more on CAES efficiencies and emission rates see a recent report I authored last year

http://www.princeton.edu/~ssuccar/caesReport.html

for more on wind+storage, see my blog from yesterday

http://switchboard.nrdc.org/blogs/ssuccar/does_wind_need_energy_storage_1.html

Charles R. Toca said...

Samir, I appreciate your comments and I reviewed your report. Your efficiency calculations grappled with the proper academic formula to use in estimating efficiency. Your various assumptions include comparing CAES to natural gas generation without compressed air - CA resulting in much higher efficiencies compared to generation without CA. You also include the compressor losses in other formulations. Your assumed efficiencies run from 66% to 82% and you summarize by saying,” The formulations provided in this section help only to provide a basis for comparison with other storage technologies, but as indicated above, the relevant expression is determined in large part by the application one has in mind." (Pg. 36) You wrote an extensive report and I would recommend it to anyone with interest in the subject.

That being said, I have to wonder if any studies have actually measured the efficiency of the two existing plants. What has been their experience with losses due to air migration, leaks, transformer losses, etc.. Real world efficiency of the VRB-ESS (which is the only technology that I can speak about with some knowledge) is 70-75%.

I also would recommend anyone to the work done by Richard D Moutoux, CU-Boulders Energy Storage Research Group on CAES. He does similar calculations to yours and reports 50% efficiency on an energy to energy basis, and an 88% efficiency assuming natural gas would be used anyway and the purpose of wind energy is to make it more efficient.
http://www.scribd.com/doc/8492873/EESAT-Wind-Integrated-CAES-Paper

So - I would tend to agree that using wind energy to make more efficient natural gas generators is a good idea - if that is the purpose of building a wind farm. However, I believe there are those that would find this counter productive. I would also point out that these reports do not address the energy efficiency losses of producing and transporting natural gas to the generator, or on another note, the increased GHGs caused by leaks of methane in the process, as reported by NREL: http://tinyurl.com/nypaks.

Sam Jaffe said...

Charles,
Your estimation of 54% is much closer to the mark than Samir claims. The problem is that efficiency can mean many different things. Even more precise terminology, like "Round-Trip Efficiency", doesn't work very well with analyzing CAES.

The reason for that is that you have two acts: compressing air and burning natural gas. Those two acts, when combined, have a greater resulting energy output than if they were each done separately. In other words, there's a multiplier effect (to steal a term from economics) in the act of combining compression and combustion.

How to account for the multiplier effect? The shortest answer is that you can't without getting into some complex mathematics, each of which can be played with by someone who has an agenda (i.e. who wants to make a CAES plant look more or less efficient).

The purest way of measuring efficiency in a CAES plant is to look at thermodynamic efficiency. In other words, how much energy is expended from compression (minus the loss of thermal energy during compression), versus the energy value of the compressed air. From that perspective, an ideal CAES plant is physically incapable of achieving better than 50%. In internal models that we have done based on publicly available information, the McIntosh, Alabama plant probably gets less than 40%. We have heard from others in the industry who have modelled that plant as getting less than 30%.Of course, none of that matters when the spread between daytime and night-time electricity is so wide (which it is in Alabama because of all the excess coal-fired generators in that area that overproduce at night). In other words, it doesn't matter how inefficient the process is if you can still make money on it.

I think that the energy industry is shooting itself in the foot by describing CAES as energy storage. There's an energy storage component to it. But what CAES really represents is a night-time load that makes day-time natural gas plants more efficient, cleaner and cheaper to operate. That in itself has tremendous value and the technology should be judged from that perspective.

Anonymous said...

Well to be honest you'd have to mention that CAES is cheaper than flow battery technology. It appears that the most promising concept for CAES is heat storage, and different variants have been proposed, using thermal oil storage, steam compression, solid thermal storage materials like iron and even sodium has been mentioned in blogs.

A thermal storage variant would solve the inefficiency problem, over 75% is possible for a well designed system. And no natural gas use...

Of course the thermal storage does add costs, but it looks like a large scale system could be cheaper than flow batteries.

Charles R. Toca said...

To Anonymous; I guess I'd have to see your calculations to determine if CAES is cheaper than flow batteries - and my frame of reference is the VRB-ESS, which is a one-time capital expense and minimal O&M over 20 years or more. The "cost" for CAES is theoretical, with only two installations, and would have to be greatly dependent on location. And, how much money has to be spent before construction on finding a location, testing, permitting, transmission access, etc. Then add in the costs for O&M of an internal combustion engine or a turbine plus volatile natural gas costs and efficiency losses. Why not just install flow batteries now, where they're needed, instead of focusing on potential large central plant CAES that are as difficult (or more) to site and get built as pumped hydro?