May 24, 2009
On The Dependability Of Wind Energy During Peak Demand

Naturally on the Sunday before Memorial Day holiday I'm sitting at home reading the North American Electric Reliability Corporation (NERC) 2009 Summer Reliability Assessment for the North American electric power grid. I'm sure I'm not the only one doing this. Well, on page 52 of the PDF file I espy mention how in the US Midwest in one one grid region at peak demand time (hot summer afternoon) the worst case they've seen with wind generation was only 2% of nameplate capacity. That's pretty bad.

The variable resources for the MRO-U.S. (wind generation) expected to be available at peak times is 1,130 MW, based on 20 percent of nameplate capacity of 5,924 MW. For wind generation, nameplate capability is assumed as maximum capability, although simultaneous output of geographically disperse wind farms at 100 percent nameplate capability is highly unlikely. 20 percent of nameplate capacity is used by the Midwest ISO when determining capacity of variable generation. 20 percent is also assumed available at peak load by the MRO Model Building Subcommittee when building peak models. Historically, the Midwest ISO has recorded a maximum output of about 65 percent of wind nameplate capacity operating simultaneously throughout the Region during peak demand. The Midwest ISO has also recorded approximately 2 percent of wind nameplate capacity operating simultaneously throughout the Region during peak demand. Saskatchewan, which has about 172 MW of nameplate wind, and Manitoba Hydro, which has about 100 MW of nameplate wind, do not count wind resources for reliability/capacity purposes.

I wonder what year that was. The more recent it was the worse it looks for wind power reliability since in more recent years the number of wind farms and their geographical dispersion has increased.

I would like to know whether new wind farms get better and worse deals from electric power buyers if the new wind farms are respectively further away from and closer to other wind farms. The idea here is that the more geographically dispersed the wind farms the lower the chance that the wind will not be blowing at all of them at the same time.

I would also like to know what sort of biomass is burned to generate 331 MW of biomass electricity. Wood?

The biomass portion of resources for the MRO expected to be available at peak times is 331 MW.

Update: In a comment post Greg F points to Ontario wind farm output data and he created a graphical view of Ontario wind power output from May 1 to May 15. You can see that the wind power output is highly variable swinging from 0 to over 750 MW over the course of a few days with a very fast ramp-up where wind soared. Any back-up for wind has to swing up and down very quickly. I fail to see from this data how wind can get counted as baseline power for Ontario.

What we need: graphs like Greg's that are for different wind markets aggregated at different levels. When Ontario has no wind how far do you have to go to get to an area with heavy wind blowing? Is Alberta far enough? Kansas?

Update II: A key study in the debate over the reliability of win for baseload power was published by Stanford researchers Cristina Archer and Mark Jacobson in December 2007. They argued that with long distance transmission lines connecting very geographically dispersed wind farms that wind would become highly reliable.

Wind power, long considered to be as fickle as wind itself, can be groomed to become a steady, dependable source of electricity and delivered at a lower cost than at present, according to scientists at Stanford University.

The key is connecting wind farms throughout a given geographic area with transmission lines, thus combining the electric outputs of the farms into one powerful energy source. The findings are published in the November issue of the American Meteorological Society's Journal of Applied Meteorology and Climatology.

Wind is the world's fastest growing electric energy source, according to the study's authors, Cristina Archer and Mark Jacobson, who will present their findings Dec. 13 at the annual meeting of the American Geophysical Union in San Francisco. Their talk is titled "Supplying Reliable Electricity and Reducing Transmission Requirements by Interconnecting Wind Farms."

Then 33% percent of average wind farm output could be used as baseload power. Note that is not 33% of max output (aka nameplate output). That is 33% of actual average output.

The researchers used hourly wind data, collected and quality-controlled by the National Weather Service, for the entire year of 2000 from the 19 sites. They found that an average of 33 percent and a maximum of 47 percent of yearly-averaged wind power from interconnected farms can be used as reliable baseload electric power. These percentages would hold true for any array of 10 or more wind farms, provided it met the minimum wind speed and turbine height criteria used in the study.

My understanding of this (and someone correct me if I'm wrong) is that then a group of wind farms that average 30% of nameplate output could then have 10% of their nameplate output go into providing baseload. That's not a lot. How much does a 2.5 MW turbine cost fully installed? That'd provide 250 KW of baseload. 400 of them would provide 1 GW. With a price for that 400 plus the cost of long distance transmission lines we could start doing comparisons with nuclear power plants.

The Stanford research about the need for geographical dispersion for dependable wind power strikes me as less optimistic than it seems at first blush. One big problem: High quality wind is not equally dispersed geographically. Not all high quality wind farm sites in the high plains of the US will have sufficiently distant matching sites to back up each other. Balancing new wind sites in core wind areas with other more distant sites becomes a problem if use of wind for baseload power is the goal.

Share |      Randall Parker, 2009 May 24 05:54 PM  Energy Wind


Comments
James Bowery said at May 25, 2009 6:43 AM:

Ammonia fuel network has some good proposals -- not the least of which is Solid-State Ammonia Synthesis (SSAS). SSAS is basically a reversible ammonia-air fuel cell. They tell me they don't require any precious metals.

Michael B. said at May 25, 2009 1:46 PM:

Here's an idea. What if they just built them higher off the ground so that they would have more exposure to the wind because the higher up you go, the greater the wind. That being said...Why not build wind towers on top of skyscrapers!? Shouldn't that be a great way to take advantage of economies of scale?!

K said at May 25, 2009 2:25 PM:

Michael B:

I think it would be more expensive to put wind towers on skyscrapers than to build the same height as a simple steel structure.

And despite knowing little of architecture I will hazard a guess that the upper stories of tall building are not designed to have much weight above them. The lower stories must support the higher but the top stories have to support very little.

Wind towers would also put some horizontal load on the building, especially in high winds. Whether that would amount to anything I can't say.

Nick G said at May 25, 2009 5:20 PM:

I wonder what year that was.

Me too. We really need quite a bit more data to judge: if such an event happens very rarely, it's much easier to deal with (with Demand Response, as well as with resources that would be expensive if used often, such as hospital diesel generators).

We really need raw, hourly wind data: has anyone seen it?

I would like to know whether new wind farms get better and worse deals from electric power buyers

It looks like it varies depending on the ISO (see footnotes on page 8/12), but mostly it appears to depend on actual performance over several years.

Bone Pickens said at May 25, 2009 6:30 PM:

We have always known that wind is unreliable, and is a blasted migraine headache for grid managers. Almost everyone who promotes wind is an airhead who knows nothing about the electric power business. The rare exceptions have an economic and political power stake, like Boone Dickens.

Michael B. said at May 25, 2009 9:17 PM:

The problem of roofs not being strong enough to support wind towers, perhaps new skyscrapers, when built could be made with the towers being built into the building, and be part of the building's frame. It'll be more expensive I'm sure. I don't know if it's practical though. What if towers were built on high grounds like on mountain tops? I realize that the cold/permafrost could be an issue as well as stringing electrical lines to run down to where people actually live.

I think it should be noted that the Midwest was never really believed to have that much potential as an area that could use much wind power. Other parts of the country make better candidates.

Also, this problem could be partly solved by dynamic pricing and consumption.

K said at May 25, 2009 11:05 PM:

Michael B:

Can't argue with what you say. Certainly almost anything can be built on top of a new or strenghtened building. You have it exactly right, the question is not possibility but practicality. Some things are not worth the price.

Some wind installations have been built along mountain ridges. And others in mountain passes. Anywhere that land contour boosts wind speed may offer enough advantage to offset the (usually) higher costs of building in rougher terrain.

I favor a great expansion of nuclear myself and cutting consumption by insulation and more efficient applicances. Solar concentration for utilities, solar PV for homes, and, rather reluctantly, more biofuels from biomass. We won't get perfection. IMO ending energy imports is the best single move we can make.

Engineer-Poet said at May 26, 2009 5:31 AM:

If you think the mechanical vibrations and stroboscopic effect of a massive wind turbine would be acceptable to tenants of a skyscraper, you're dreaming.  The irregular airflow around a prism-shaped building would be no friendlier to the wind turbine, I suspect.  And who would write a liability policy to pay damages if the unit e.g. sheds a blade in the middle of a city?

As for wind power during demand peaks, really... this is comparing Indy cars and dump trucks.  The purpose of wind is to cut fuel use and cost (and various emissions including carbon) in the rest of the system.  Today's backup is the same peaking capacity used for those hot, windless summer days which idle the wind farms anyway (if you haven't ever looked at a weather map during a big Midwest high-pressure system, do so).  Tomorrow's backup may be CAES, or it may be slashed by DSM such as ice-storage A/C.

Nick G said at May 26, 2009 10:48 AM:

The purpose of wind is to cut fuel use and cost

I think we can defend the usefulness of wind more strongly than that.

That 2% was apparently a one-time minimum that may no longer be relevant; it was for only one of 9 US regions; DSM is very cheap and effective; time-of-day metering is extremely likely to reduce the size of peak load; peak capacity already exists; and backup that is used rarely can be very cheap.

An important point: the US ISO's generally accept wind as proving peak capacity. On average, they give it a capacity credit of about 16%, which is more than half of it's average output. That's not bad.

Rob said at May 26, 2009 1:52 PM:

In 2007, Jim Detmers, VP of the California ISO gave a talk at Stanford in which he said this:

"Wind is not produced on peak. This last summer, when we went across the summer peak, I had 3,000 megawatts of capacity of wind. How much did I have on the summer peak, back in August? No, no, no, I didn't have zero. I had a total of 63 out of 3,000. And we're investing all of this money in wind..."

My source for this is the audio of the talk at iTunes U, here's the URL for those who wish to listen: http://deimos3.apple.com/WebObjects/Core.woa/Browse/itunes.stanford.edu.1291176865.01291176868.1427408599?i=1204795563

This speaks only of California, of course, but I know that Texas almost had an outage last year when the wind in west Texas unexpectedly quit blowing. Wind is a scam, text excerpts from the speech can be found here: http://roborant.info/main.do?entry=1386

TexasJew said at May 26, 2009 2:08 PM:

Wind IS a scam.
I live in West Texas and have deriven throught the multitudinous wind farms on many occasions when the blades were at a dead stop.

The oil pumpjacks around the wind towers, however, were all going up and down, of course. That's a reliable energy source for you!

Nick G said at May 26, 2009 2:12 PM:

Rob,

Many ISO's don't like wind. This isn't surprising: it makes their lives harder. Utilities don't like the best solutions to wind variability (inter-grid balancing, DSM, time-of-day pricing, etc), because utilities are paid for their investments, not for fine-tuning of demand.

BTW, 3GW isn't very much - it's not surprising that it sees high variance.

David Gillies said at May 26, 2009 2:30 PM:

Three figures of merit are:
1) cost per installed watt of nominal generation capability ($/W)
2) peak to mean ratio i.e. inverse of availability of nominal generation capability
3) power concentration (W/m^2)

In none of these does wind come even close to prime power sources like coal or nuclear. Fiddling around the edges does not alter this. To suppose that wind can ever be anything other than a supplement is an appeal to magic.

Bi-Coloured-Python-Rock-Snake said at May 26, 2009 3:07 PM:

What sort of biomass is burned to generate 331 MW of biomass electricity. Wood?

I'm not sure if it falls under the category of "biomass" or not, but there are still a few waste-to-energy incinerators running. Also, the big pulp and paper mills run their operations by burning wood waste on-site, so perhaps that's also figured in.

Curly Smith said at May 26, 2009 3:15 PM:

Peak electrical demand is during the heat of the day in summer and the dark of night in winter but peak wind periods are in spring and fall. In fact, where I live there is virtually no wind during the heat of a summer day so all of the wind turbines in the world would be useless. Consequently, to supply 1000 MW an "environmentally responsible" utility company would need to install 1000 MW of coal, gas or nuclear generated power PLUS (1000 MW * 1/ efficiency factor) of wind power. The utility company would thus install total capacity of 2000 MW (efficiency factor = 100%) up to 6000 MW (efficiency factor = 20%).

geowalsh said at May 26, 2009 3:20 PM:

I am not an expert and I don't play one on tv but I do know that the extra structural reinforcement needed for mechanical vibrations and stroboscopic effects and other stressor, and the effort to balance legal liabilities can be ground down to a dollar amount. Would anyone be surprised if that dollar amount greatly exceed the dollar amount of energy produced?

All the hysteria about carbon emissions non with standing, If the system can't pay its own way then it will eventually be dropped in favor of those that can. If large scale wind power is to be truly effective then it must find a way to store the energy created for later use. A farmer, rancher, or other land owner could make it work by using the windmill to drive a compressor that pumps air into a tank. The compressed air could then run a generator. If the tank is large enough and the run up to its use is long enough you wind power become a workable solution. But this kind of small scale application is very different from wind commercial wind farms.

If enough small scale users were to appear the total amount of power of power could become, over time, statistically significant. But hey, who am i to think such thoughts?


Greg F said at May 26, 2009 3:22 PM:

If you have any doubt that wind farms are a real poor way to generate electricity then just wait for a heat wave in the summer or a cold snap in the winter. Then go check the national wind maps. These are the two times when the demand is the highest.

http://tinylink.com/?wnv1A7EzrZ
Will British weather provide reliable electricity?

Abstract:
There has been much academic debate on the ability of wind to provide a reliable electricity supply. The model presented here calculates the hourly power delivery of 25 GW of wind turbines distributed across Britain's grid, and assesses power delivery volatility and the implications for individual generators on the system. Met Office hourly wind speed data are used to determine power output and are calibrated using Ofgem's published wind output records. There are two main results. First, the model suggests that power swings of 70% within 12 h are to be expected in winter, and will require individual generators to go on or off line frequently, thereby reducing the utilisation and reliability of large centralised plants. These reductions will lead to increases in the cost of electricity and reductions in potential carbon savings. Secondly, it is shown that electricity demand in Britain can reach its annual peak with a simultaneous demise of wind power in Britain and neighbouring countries to very low levels. This significantly undermines the case for connecting the UK transmission grid to neighbouring grids. Recommendations are made for improving ‘cost of wind’ calculations. The authors are grateful for the sponsorship provided by The Renewable Energy Foundation."

See also "E.ON Netz Wind Report 2005" (They are the largest producer of wind power in the world).
http://www.windaction.org/documents/461

Page 7:
"The weather situation determines the wind level. Both cold wintry periods and periods of summer heat are attributable to stable high-pressure weather systems. Low wind levels are meteorologically symptomatic of such high pressure weather systems. This means that in these periods, the contribution made by wind energy to meeting electricity consumption demand is correspondingly low."

Also notice chart 3 also on page 7:

"Wind power feed-in in the E.ON control area".

Not exactly a picture of reliability.

Michael B. said at May 26, 2009 3:46 PM:

I think wind only has viability right now in areas that are already high off the ground. Why not put giant wind farms on mountain tops, where the wind almost never stops, and transfer the electricity to populated areas? It won't be a very complete solution, but even if it consistently provides us with 1-2% of our energy needs then it would still be worthwhile to have them.

I agree with you geowalsh. The only problem is that it's not really worth it for everyday people to build wind towers. It's expensive and they don't get much return.

Nick G said at May 26, 2009 4:07 PM:

David,

Those measures are modestly useful, but far from the most important.

The important thing is cost/KWH. Wind does just fine on that measure when compared to new coal or nuclear.

1) aka nameplate capacity: Natural Gas is much lower than wind. Wind is much lower than coal or nuclear - what does this tell us?
2) Natural Gas is higher on this than wind, nuclear is lower. What does this tell us?
3) This is meaningless. What's the power concentration of oil-bearing rock?

AST said at May 26, 2009 4:29 PM:

I'm waiting for all the green groups to spring on us that wind turbines are sucking the energy out of our weather patterns and contributing to catastrophic climate change, either melting icecaps or a new ice age.

Why are we listening to people whose only evidence is their computer models of something that's too complex and delicately balanced to make any valid modeling possible?

Michael B. asks: "Why not put giant wind farms on mountain tops, where the wind almost never stops, and transfer the electricity to populated areas?" Because elites with nice views are already complaining about wind farms destroying their views. In a lot of places putting structures in such places is illegal because it impinges on someone's skyline. Wind turbines also require connections to the power grid, meaning that they must be served by high tension power lines and service roads for maintenance. Roads and power lines are anathema to pro-wilderness groups, but they're not mentioning that now because what they really want is to shut down all power plants burning fossil fuels.

If we are dumb enough to rely on vague promises like those made by wind energy proponents without asking obvious questions about cost and reliability, and then pay subsidies to the T. Boone Pickenses of the world to build these whirlygigs, we deserve to be driven back to the pre-industrial era.

Nick G said at May 26, 2009 4:52 PM:

Curly Smith said: peak wind periods are in spring and fall.

Do you have data on that? I'd love to see it. From what I've seen this varies by region.

to supply 1000 MW an "environmentally responsible" utility company would need to install 1000 MW of coal, gas or nuclear generated power PLUS (1000 MW * 1/ efficiency factor) of wind power.

Not really. The peak capacity already exists - why would you build it again? If you're building wind for environmental reasons, then all you need is the KWHs - the peak capacity factor of roughly 55% of average output (based on the latest NERC figures) is just a bonus.

If large scale wind power is to be truly effective then it must find a way to store the energy created for later use

Not really. See my earlier comments.

th said at May 26, 2009 5:17 PM:

quoth nick g:
"if such an event happens very rarely, it's much easier to deal with (with Demand Response, as well as with resources that would be expensive if used often, such as hospital diesel generators)."

That, of course would be algae-based biodiesel.

"Many ISO's don't like wind."

What an understatement.... out of 9022 windmill turbines installed as of this report, 1292 on grid, a whopping 7800 off, a better mousetrap it ain't.
http://www.windpoweringamerica.gov/pdfs/2007_annual_wind_market_report.pdf

Good news here for robert redford fans, it can be as high as 10%, page 10, lines 13 thru 34, Texas and NY weigh in.
http://www.nhsec.nh.gov/2008-04/documents/090105testimony_linowes.pdf


Lockestep said at May 26, 2009 5:46 PM:

"the peak capacity already exists - why would you build it again?"

Peak 2009 exists, but peak 2012 does not. Building wind and selling it as available for peak...or even worse, building in 20% of nameplate wind capacity to your n-1 calculations...is begging for blackouts.

Randall Parker said at May 26, 2009 6:42 PM:

th,

I'm looking at that Annual Report on U.S. Wind Power Installation, Cost, and Performance Trends: 2007 and what I see on page 4 is that those 1292 on-grid turbines the 7800 off-grid turbines are meant to be turbines that mostly generate local power. They are together a small fraction of all wind turbines installed in 2007. Also, the on-grid subset is much larger per turbine. Go and read page 4 again.

Engineer-Poet said at May 26, 2009 7:06 PM:

Geo. Walsh, you need to do a little searching and reading on "CAES" (Compressed Air Energy Storage).  Other folks are way ahead of you, though the personal-scale things you're thinking about aren't going to be economical.

Michael B., mountain tops don't have very much area, and the wind running over sharp ridges is turbulent and causes troublesome mechanical stresses; this has been a known problem since the early 20th century.  The best areas for wind appear to be flat or gently rolling, such as the US's Great Plains and areas over or near large stretches of open water.  This isn't to say that W. Virginia may not have some good mountain sites, because it's close to major markets and the savings on transmission lines may pay for beefed-up mechanicals.

Greg F, if you've got a limited amount of gas to burn, what's your best option:  using it in simple-cycle peaking turbines at a best-case efficiency of 46% for the latest units (and likely much less), or at 80% gas-to-electric efficiency in a CAES system pumped up with wind power?  Which option will give you the most predictable cost of energy, and the best energy security?

Randall Parker said at May 26, 2009 7:24 PM:

E-P,

Seems to me that wind power increases the need for peaking capacity. Wind blows less on hot summer afternoons. So to the extent that wind displaces coal and natural gas baseload wind increases the need for peaking power. That seems inescapable.

Storage mechanisms for wind power help some. But they can at best shift wind power around during a day or two. If the wind isn't blowing all night and day during a really hot period that doesn't help. It is the wind that blows in the winter that doesn't blow in the summer that is problematic.

Engineer-Poet said at May 26, 2009 7:48 PM:

Quoth RP:

Seems to me that wind power increases the need for peaking capacity. Wind blows less on hot summer afternoons.
No, you can keep some existing FF plants around for the lulls.  You just run them a lot less and burn a lot less fuel.  If your forecasts are good, you have plenty of time to get them started before they need to be on-line.
Storage mechanisms for wind power help some. But they can at best shift wind power around during a day or two.
You're thinking of pumped hydro, which typically has less than 1 day of full-power storage.  CAES systems have much broader limits and could easily be sized for a week or more.  The major question is how big a reservoir you want to create, and how much energy you're willing to invest to pump it up the first time.

BTW, the 80% gas-to-electric efficiency of CAES assumes that none of the heat of compression can be recovered.  The energy inputs to the latest CAES designs appear to be about 3/8 electric, 5/8 fuel; the overall efficiency is about 50%.  If you can recover 1/3 of the heat of compression, the gas-input-to-electric-output ratio goes to 1:1; it may be feasible to do quite a bit better than that.

Last, the wind lulls occur when highs dominate the weather.  Highs mean clear skies.  Guess what works great when you've got clear days?

Randall Parker said at May 26, 2009 8:12 PM:

E-P,

It is my (admittedly limited) understanding that the base load coal and natural gas plants can't spin up quickly. Now, if you can forecast that you have a hot afternoon with little wind coming up you can spin them up. But how many hours does it take and how much energy do you expend during the spin-up and spin-down phases?

To put it another way: can we really count on converting base load plants into load following plants to serve as back-ups for wind power?

And another thing: to the extent that we idle base load plants much of the time in order to get power from wind we also require that the capex for those baseload plants get spread over a smaller total number of kwh produced. So capex expenses per kwh go up.

Also, it isn't enough to just keep existing FF plants around to back up wind power. As demand rises the incremental wind power will need incremental back-up from other types of power.

Clear skies: Sure, solar is a pretty good fit for afternoon peaks. Though it ends up generating too much in the morning and not enough in the evening. Also, solar has to become cheap and that still hasn't happened yet. Once solar becomes cheap it can cut into the need for peak following dispatchable fossil fuel burning plants. But when that happens still lies at least a few years into the future.

Mike Kelley said at May 26, 2009 8:14 PM:

There is no way in hell we can afford to keep conventional power plants sitting around to use when the wind dies. The price of electricity would be astronomical. Who do you use for manpower to run, say a coal plant, when it is shut down some of the time? No decent operator would do it part-time. All this liberal, pie-in-the sky stuff is going to murder our economy.

Greg F said at May 26, 2009 8:50 PM:

EP wrote:
"Greg F, if you've got a limited amount of gas to burn, what's your best option: using it in simple-cycle peaking turbines at a best-case efficiency of 46% for the latest units (and likely much less), or at 80% gas-to-electric efficiency in a CAES system pumped up with wind power?"

If you haven't noticed were not building CAES systems. Red herring.

James Bowery said at May 27, 2009 8:12 AM:

Greg F writes:

...wind farms are a real poor way to generate electricity then just wait for a heat wave in the summer or a cold snap in the winter. Then go check the national wind maps. These are the two times when the demand is the highest.

An excellent observation that supports solar in general. I'm particularly interested in solar updraft technologies because:

1) their low efficiency per area can be compensated for by their low cost per area
2) their main driver of efficiency is how high you can build the updraft structures
3) their collector areas can serve as massive greenhouses
4) their effluent can increase local rainfall
5) their inherent thermal control characteristics raise real estate value around them (the greenhouses actually contain thousands of acres of thermally controlled, continual gentle breeze).
6) their carbon footprint is negative within a few years
7) all of the above are clearly subject to industrial learning curves which can justify immediate large scale investment with the only serious risk being the cost of construction of the first 1km tall tower -- which isn't really that great given the fact that reinforced concrete chimney's are well within the the known art and the compressive strength with some of the more advanced reinforced concretes is several times greater than necessary.

Nick G said at May 27, 2009 2:47 PM:

Lockestep said Peak 2009 exists, but peak 2012 does not.

Peak demand in 2009 is likely to be substantially lower than 2008. Peak 2012 isn't likely to be signficantly higher than 2008. If we wanted to permanently freeze (or lower) peak demand, we just need to give everyone time-of-day metering.

Engineer-Poet said at May 27, 2009 3:34 PM:

Quoth RP:

To put it another way: can we really count on converting base load plants into load following plants to serve as back-ups for wind power?
They do a bit of load-following now, though IIUC the newest and most efficient plants have the slowest ramp rates.  We can't expect them to do what the simple-cycle gas turbines do (cold to full power in 15 minutes).  Whether we can expect them to take frequent thermal cycles at all is a good question; they may not be designed for it.

The obvious thing to do with these plants is to decommission them at end of life and replace them with something like wind+CAES.  The ISEP CAES system will have about 50 hours of storage, which allows plenty of time to get remaining plants fired up even if it takes half a day.

Quoth Mike Kelley:

There is no way in hell we can afford to keep conventional power plants sitting around to use when the wind dies. The price of electricity would be astronomical.
You think so?
  1. What do you think those simple-cycle gas turbines do most of the time?
  2. A fair amount of the nation's coal capacity is small, older plants which are fully amoritized.  These are often used as intermediate capacity because they can be fired up relatively quickly.  They are already "sitting around" a fair fraction of the time.
  3. Larger, more modern coal-fired plants take longer to start, but if coal gets expensive it will be worth it to use them only when prices are higher.  More and more of these will be fully amortized as time goes on, and the only costs will be fuel and O&M.
So no, the cost will not be "astronomical", it will be closer to the rates you're already paying for peak power today.

Quoth Greg F:

If you haven't noticed were not building CAES systems.
Check out the Iowa Stored Energy Park, due to go into service in 2011 (I found nothing about schedule delays).  There are already plants operating in Alabama and Germany.

Greg F said at May 27, 2009 5:57 PM:

Still a red herring EP. The vast majority of wind power installations have no stored energy aspect. I would be willing to bet none of the promotional sites will reveal the conversion efficiency of compressed air.

Engineer-Poet said at May 27, 2009 7:20 PM:

The vast majority of wind farms need no stored-energy aspect (and legal nits like the PTC probably stand in the way anyway).  It's only when wind starts exceeding 20% of local generation, or the local grid is weak, that you need storage.  It's the Iowa Stored Energy Park for a reason.

The latest paper I saw for a study gave figures indicating roughly 50% overall efficiency (37.5% electric input, 62.5% gas input); the overall electric output was 1/3 greater than the input.  Given a gas-to-electric efficiency of 80% and throttling capability as good as a simple-cycle gas turbine, the CAES plant has both top-notch flexibility and efficiency; not even CCGT can beat the efficiency, and CCGT throttles poorly.

B. Grandy (AKA B Dubya) said at May 28, 2009 8:24 AM:

Something else to think about.
Experience with wind turbine installations in the Tug Hill region of NY is demonstrating that these designs are not very reliable, with catastrophic blade and bearing failures after a very short run time. Way too much loading on the mechanical structures and, when they go, you get the added benefit of electrical faulting of the generator as a one in four chance scenario. If these are connected en mass into the grid, how many shots on goal (cascading grid failure) will it take before this green power turns to black out?

I thought Denmark had already provided the test case for wind turbine generation and base load reliability. My understanding is that, even though their installation is extensive, it will not support peaking, nor will it support base loading. They sell off hour power from wind turbines, when its actually available, to Norway, as it is of little use at o dark thirty in Denmark.

The wind turbine generation scheme in progress here is a heavily subsidized, green feel-good scam. In the actual power market, it is a loser.
Sort of like biofuels from food grains.

Engineer-Poet said at May 28, 2009 9:09 AM:

I did a search for "tug hill" wind farm failures and came up with nothing to back up your claims.  Further, they ring false; even if a generator shorted pursuant to a mechanical failure, the fault currents would only trip its own breakers, not the entire farm.  Grid reserve requirements include losing the biggest generator in operation at the time; losing 2-3 MW is down in the noise of grid demand variations.

Greg F said at May 28, 2009 2:53 PM:

EP wrote:
"The vast majority of wind farms need no stored-energy aspect..."
They just need conventional power station backup. Go read the EON report I posted earlier.

EP wrote:
"Further, they ring false; even if a generator shorted pursuant to a mechanical failure, the fault currents would only trip its own breakers, not the entire farm."

Your full of it EP. From the same EON report:
"Even simple grid problems can lead to significant failures in wind power production. Large thermal power stations do not disconnect from the grid even following serious grid failures, instead they generally trip into auxilliary services
supply and until then, "support" the grid. Wind farms, however, have so far disconnected themselves from the grid even in the event of minor, brief voltage dips. Experience shows that this can lead to serious power failures:

• On 29 January 2004, a two-phase line fault occurred in the 220kV grid in the Oldenburg region and resulted in split second-long voltage dips in the region concerned. This produced a sudden loss of around 1,100MW of wind power feed-in.

• On 15 September 2004, a crane caused an earth fault on an extra-HV line in Hamburg. The resulting brief voltage dip of a few tenths of a second meant that approx. 600MW of regenerative power disconnected from the grid in the Hamburg
region."

Nick G said at May 28, 2009 3:38 PM:

E.on Netz....sigh.

That annual report was remarkably misleading - they greatly exaggerated their problems. For instance, they said that wind requires 90% backup ("traditional power stations with capacities equal to 90% of the installed wind power capacity must be permanently online in order to guarantee power supply at all times.")

That's misleading to the point of being dishonest. What they're really saying is that their wind farms provide a capacity credit of 10% of nameplate. Given that their average capacity factor is only 18% (yes, I agree that's low), that's pretty good: it's 55% of average production.

That report will provide fodder for anti-wind activists for a long time...

Nick G said at May 28, 2009 3:42 PM:

To add to that comment: even a single wind farm doesn't go above 85% production as a practical matter; just a little regional geographic dispersion reduces that to around 60% (see the 2007 NERC report, by regions); and modest weather forecasting reduces the uncertainty factor much further. Finally, Demand Side Management (as was used in that famous Texas incident recently) can handle most of the rest of the variance.

So, to say that wind farms need 100% backup is really being dishonest.

Randall Parker said at May 28, 2009 5:58 PM:

B. Grandy,

Since several companies provide blades, turbines, and other equipment for wind farms one can't generalize from specific problems at one farm to all wind farms. The makers of these parts each have their own specific failure modes based on materials and designs and fabrication methods.

I happen to hear about specific problems at one wind equipment maker (and can't say which) and they've got lots of engineers solving their problems. One can't really use a report from 1, 2, 3 years ago to know what the state of play is for reliability of new stuff or of fixed stuff.

Greg F,

Average achieved output and patterns of variation of output differ between wind farms. A single farm of a single company tells you nothing about the industry as a whole. Average wind capacity in, say, up state New York is different than average wind capacity on the high plains. Ditto distribution by time of day.

I'm curious to know a lot more about the particulars and am unwilling to take a few data points as definitive. I took a single data point to use to spark discussion with this post. Sure nuff, here's the discussion.

I'm hoping others will chime in with empirical data about big an area one needs to avoid total output dips below some X percent at various times of the data. I've yet to see the empirical data on this.

Greg F said at May 28, 2009 6:06 PM:

Nick G,
Reserve (usually called spin reserve) is not the same as online (meaning not down for maintenance). See chart #8, page 10.

Greg F said at May 28, 2009 6:11 PM:

Randall,

Averages are misleading. Your lights won't stay on because there is on average enough power. Which happens to be the point the advocates ignore.

Greg F said at May 28, 2009 6:29 PM:

Here is the grid:
http://www.npr.org/news/graphics/2009/apr/electric-grid/

Turn off the proposed lines since they don't exist. On the next heat wave of the summer compare to the wind maps.

Randall Parker said at May 28, 2009 8:10 PM:

Greg F,

We agree. Averages are misleading. So where's a good source of data for minimum outputs of various wind farms and how synchronized the minimums were?

I'm really open to persuasion on this one. I want to know how (un)reliable wind power is. I want to know how much that (un)reliability varies by region. I want to know whether the proposed offshore wind farm for Delaware is more or less reliable than the wind farms in Texas and the Dakotas.

Nick G said at May 29, 2009 8:36 AM:

Greg,

Reserve (usually called spin reserve) is not the same as online (meaning not down for maintenance).

Yes, I agree. Please elaborate on how that affects what we're saying.

See chart #8, page 10....Averages are misleading.

Yes, we're all agreed that wind power has variance. You, I, Randall, E-P, we all know that, really we do.

The ISO's quoted in the NERC report know it too. After pretty exhaustive analysis, they've concluded that regional wind power deserves a capacity credit of roughly 55% of average output. that's not bad.

Randall, Greg,

The NERC report has relevant info here. For instance, the 2% minimum and 65% maximum tells us something: at the level of this region, (and in this period of the day and year) wind output has never gone below 2% and over 65%. So, the range of variation isn't 100%, it's 63%. This is good news. That band would be narrower if you used the proper statistical values, of 3 or 6 sigma, or whatever ISO's use for capacity credits. It would be narrower at the national level.

Yes,, it would be nice to get more data. Ideally, it would be regional or wind-farm hourly data for several years. Greg, seen any?

Greg F said at May 29, 2009 3:43 PM:

There is hourly data for Ontario Canada which is a rather large area. I estimate more than 500 miles across. I downloaded "Hourly Wind Generator Output" which goes back to March of 2006.

http://www.theimo.com/imoweb/marketData/marketData.asp

Plotting all of it looses the detail so I plotted May 1st to the 19th of this year (500 hours), which is the most recent data.

http://i116.photobucket.com/albums/o25/techtipmail/OntarioMay2009Hourly.jpg

The horizontal axis is hours, vertical is in mega-watt-hours. Ontario is part of the north eastern grid.

Nick,
Chart #8, page 10 is actual data for 15 days. Obviously it is at least hourly. I imported the chart into a CAD program so I could scale it. The first day (Mon 5.1) dips to less than 3% of installed capacity. Even taking into account the 85% you mentioned earlier you would still need more than 90% in conventional power plant capacity to cover for the bad days.

Now go to the grid map. Mouse over the question mark on the 700 - 799 and you will see that anything below is generally limited to under 100 miles. What is obvious is, even if the wind was blowing in Ohio, there would be no way to get a significant amount of power to Ontario. Again, when the first summer heat wave hits go look at the wind maps. It will be obvious that the 'wind blowing someplace else' is to far away to cover for where the wind isn't blowing.

There are other problems not obvious to those with no knowledge of electrical theory. Power factor, which is the phase difference between the voltage and current, changes over distance. Unless you have DC transmission lines this will significantly limit the distance you will be able to transmit power. Wind generators, due to weight considerations, has less control over the power factor.

Nick G said at May 30, 2009 10:20 AM:

There is hourly data for Ontario Canada

Thanks for finding that.

which is a rather large area. I estimate more than 500 miles across.

Canada is a funny place: it's provinces are enormous, but it's population is small, and concentrated in the south. In this case, the wind farms are concentrated around Windsor and Toronto in a relatively much smaller area. Further, as of 2008 they only had about 1.2GW wind (nameplate), which really isn't much.

So, this is a good start, but hardly enough to make generalizations.

Even taking into account the 85% you mentioned earlier you would still need more than 90% in conventional power plant capacity to cover for the bad days.

First, you need to do a much more careful analysis than just eyeballing a few days: you need to look at the parts of the year and day where capacity is actually needed, and you have to use the statistical approach preferred by the local ISO to forecast capacity credit. This may not feel intuitive, but the absolute minimum isn't the correct approach. Keep in mind that all forms of generation have variance. Nuclear, for instance, can "trip" in seconds - that can remove a full GW from the grid for days. That's why Ireland, for instance, chooses not to use nuclear: that kind of variance is too much for a small grid.

2nd, if generation can't get above 85%, then you don't need to "back up" any more than that. Actually, you don't need to back it up at all, because you never plan for that kind of generation as part of peak capacity - you just have to manage the change in generation, as the winds die down and wind generation comes down from it's peak. The best way to do that is to use Demand Management: ramp up demand during that peak (say, by signaling to commercial refrigeration to set it's temp down a couple of degrees, and for electric vehicles to charge at maximum), and then ramp demand down as the wind dies. Very cheap, very effective, well tested, widely used. It's what Texas used recently to deal with such a problem.

More later...

Greg F said at May 30, 2009 12:19 PM:

Nick G Wrote,
"In this case, the wind farms are concentrated around Windsor and Toronto in a relatively much smaller area."

And what is a "relatively much smaller area"? No "generalizations" Nick, I want a number. What radius constitutes "concentrated"?

Nick G Wrote:
"First, you need to do a much more careful analysis than just eyeballing a few days:"

I have yet to see any analysis by you Nick.

Nick G said at May 30, 2009 2:49 PM:

what is a "relatively much smaller area"?

Well, take a look at page 4 of http://www.ieso.ca/imowebpub/200811/CanWEA2008-KKhan.pdf . We see that there are couple of roughly 200MW wind farms 300 and 400 miles away from London, and then there are a cluster of about 800MW in the center, in a rough ring of about 75 miles radius.

Given that the US has about 1,050GW in generation, and is about 6,000 miles wide, in statistical terms that's small.

I have yet to see any analysis by you

Well, 1st, some it the ISO's have done it for us. You were questioning their analysis, and I'm suggesting to you that if you want to question it, you'll have to do a similarly very detailed analysis. 2nd, I've offered some analysis above. 3rd...you're right, it would be nice to see more analysis of the effects of geographic dispersion. It's out there - I'll see if I can find it again.

Randall Parker said at May 30, 2009 4:37 PM:

Greg F, Nick G,

This debate reminds me of what sounds like the best study on wind reliability that I've come across. See my 2nd update to this post. That Stanford study doesn't strike me as sufficient by itself to prove the case it is trying to make. We really need data over several years rather than just the year 2000 data that study uses.

Nick G said at June 1, 2009 3:39 PM:

Randall,

The Stanford study doesn't seem to be very ambitious. They look at a small system of wind farms with nameplate capacity only 1.5GW, or the equivalent of a medium sized coal plant, distributed over an area with a radius of about 250 miles.

They demonstrate rigorously what we already know intuitively: a number of sources will have a lower ratio of variance to mean production than a single source. That's all they're trying to do, and it doesn't really tell us much about the limits of the value of interconnection of truly widely separated wind sources.

Randall Parker said at June 1, 2009 5:50 PM:

Nick G,

So far the Stanford study is the best I can find in the public domain. Obviously we need something bigger. Also as obviously: We need cost data on long range DC transmission lines. If distant wind farms can together produce useful baseload capacity at what percentage of nameplate and at what cost of power lines?

Maybe the ISOs have done better modeling internally. But so far I can't tell just how reliable wind can be. But I do know that most wind promoters who have opinions on the subject know less than I do.

Nick G said at June 2, 2009 10:10 AM:

I think I saw some better studies - I'll try to find them. A few thoughts:

So far we've been discussing the value of adding multiple windfarms, presumably with output either non-correlated or only partly correlated. It's worth noting that many windfarms are negatively correlated.

Only a small % of a region's wind power would need to be transferred between regions in order to provide balancing, and possibly not as far as one might think. Sometimes it's just a matter of a number of sub-regions getting their power, on average, 100 miles from their west, rather than 100 miles from their east, and in effect you've transferred power from the western edge of the overall region to the eastern.

We really don't need much more peak capacity - perhaps none at all for many years, with good time-of-day pricing and DSM. That renders most of this argument moot, at least as a boundary: we can use existing generation if we have to as a backup.

Wind farm peak capacity credits are a little like getting a dog to talk: the interesting thing isn't how well the dog talks, but that it talks at all. The fact that even a small cluster of wind farms can have a 1/3 of average capacity credit, or wind at a regional level have an average 55% credit, is important.

I see local wind capacity credit as solving roughly half the intermittency problem; long-distance transmission solving about 1/4, and DSM solving the rest. DSM alone could make an enormous contribution: think 220M EV's (with 2.2TW peak demand or output) doing a dance of load balancing with the grid. This doesn't even touch the legacy peak capacity which could provide backup - this we'd want to minimize to minimize CO2 emissions, but that wouldn't be hard with DSM as a short-term factor: we'd only need it for very unusual, long-term lulls.

Nick G said at June 2, 2009 10:25 AM:

Maybe the ISOs have done better modeling internally.

I think we can rely on that. The ISO's are pretty conservative - if they say something positive about wind, I think we can rely on it. Now, if they say something negative, I think we can treat them as a "hostile witness", and critique it. For instance, the E.on Netz 2005 report, which was obviously biased.

OTOH, it would be interesting to look at the ISO methodology. The Stanford study might be relevant there - I'll have to look at it again.

Nick G said at June 2, 2009 10:33 AM:

Other thoughts:

Don't forget that solar is substantially negatively correlated with wind. That also proportionately reduces the problem.

Remember, in general what we're talking about is reducing statistical variance. Existing generation also has it - that's part of the point of the Stanford study: a medium-sized coal plant of the size of the wind farm cluster in question will be out due to unpredictable and unpreventable failure during 7% of the year. That says that our wind farm cluster can provide zero output for 7% of the year, and do no worse than the coal plant.

Wind is non-dispatchable (from a certain point of view), but it's not unpredictable - that makes integration into the grid much easier than otherwise.

Randall Parker said at June 2, 2009 9:50 PM:

Nick,

A few points:

- Solar is more predictable than wind.

- Solar positively correlates with peak demand while wind negatively correlates with peak demand.

- Wind's predictability varies from site to site.

- Wind's consistency (roughly speaking, what percentage of the time it blows) varies from site to site.

- Wind's correlation with peak demand varies from site to site.

- Not all wind sites can be matched up with other sites to raise baseload availability.

Nick G said at June 3, 2009 9:28 AM:

No question, some wind sites are better than others, in all of those terms: output variance; mean; correlation with demand; and auto-correlation with other sites. Fortunately, we have a lot of sites to choose from, especially in the US.

Yes, wind is negatively correlated with demand. OTOH, PHEV/EV demand will be primarily at night. If we have new demand of about 70GW on average, and place all of that in 8 hours, that's 210GW of new demand at night: that will even out the daily cycle quite nicely. Of course, we won't move that demand quite that neatly, but we'll probably get pretty close, if we want to.

Yes, diurnal and seasonal solar is more predictable than wind. OTOH, hourly output is less predictable, especially for small geographical areas: cloud formations can have sharp edges and move fast, so that individual sites and small regions can see very fast, hard to predict changes in output. Wind can generally be treated as negative demand (the rate of change is similar to the currently demanded load), but solar has a higher rate of change.

This is, of course, an argument for emphasis of large solar plants in areas with very low rates of cloud formation; and emphasis on widely scattered smaller solar.

Nick G said at June 3, 2009 10:05 AM:

Another thought: wind is negatively correlated with demand - that's an argument for combining wind and solar in a renewable grid. Wind is stronger at night, on average (though only slightly), so one would consider wind as baseload and solar as peaking.

It's important to remember that the current demand curve is artificial: it's the result of flat residential pricing, and very crude time-of-day I/C pricing. A lot of demand could move to the night to match wind, or to noon to match solar, if necessary.

DSM, especially from PHEV/EVs, could easily smooth out almost any diurnal variance. Unusual, multi-day lulls in output could be handled primarily with FF/biomass plants that would still only provide well below 10% of KWHs.

Engineer-Poet said at June 3, 2009 7:12 PM:

I cannot help but note that Greg F's list of problems have nothing to do with Tug Hill, or even the class of problems claimed by B. Grandy.  Further, they are probably due to over-sensitive fault protection systems (which could be re-calibrated).  Finally, a 120 MW wind farm requires less spinning reserve than having a 1.6 GW nuclear plant on the grid.  As long as failures are independent, you only need reserve equal to the biggest producer on-line at the moment.

And you're right, average power won't keep the lights on.  However, simple tweaks like adding ice storage to domestic refrigerators allows those loads to follow supply.  Electric vehicles with smart chargers would radically increase that flexibility.

Greg also said:

What is obvious is, even if the wind was blowing in Ohio, there would be no way to get a significant amount of power to Ontario.
This is not at all obvious, given that there is a very large loop of transmission lines which cross from Ontario to Michigan.  Any serious development of wind resources in Lake Huron would develop an even bigger intertie as a byproduct.

Engineer-Poet said at June 7, 2009 7:42 PM:

Randall, correcting your 2nd update:  creating 1 GW of baseload at 10% of capacity would require 4000 turbines at 2.5 GW each, not 400.

The big problem with large-scale wind, and particularly the wind-reliant grid pushed by some of the green types, is the massive scale of the transmission infrastructure required to move the power.  There is a little talk about this, particularly when talking about the advantageous position of offshore wind close to coastal population centers, but not nearly enough.

The answer to many of these problems is nuclear power, particularly molten-salt reactors like LFTR.

Nick G said at June 9, 2009 9:47 AM:

the massive scale of the transmission infrastructure required to move the power.

This seems to be about $.25 per nameplate watt, or 10-15% of wind farm costs. That doesn't seem bad.

Estimates for transmission costs in Texas ($.26/nameplate W):
http://www.cnbc.com/id/25708335

Engineer-Poet said at June 9, 2009 8:07 PM:

$0.25 per nameplate watt would be about $2.50 per baseload watt.  That's roughly the cost of nuclear these days, IIRC.

Randall Parker said at June 9, 2009 8:38 PM:

Nick G,

There's the cost of getting wind electric from a remote area to a city. But that's not the cost of getting it from where it is to many different cities so that it can back up other wind turbines in other areas.

What's the cost of transmission lines that are long enough and efficient enough that they can enable distant wind farms to serve as back-ups for each other?

Wind has scaling problems. We need nuclear. We need concentrating solar at a lower price point.

Nick G said at June 10, 2009 12:32 PM:

$0.25 per nameplate watt would be about $2.50 per baseload watt.

1) That's just looking at the value of capacity factor. The value of the KWHs is much more important.

2) That's assumes a CF of only 10%. The 2007 NERC report says that ISO's are giving about 16%.

That's roughly the cost of nuclear these days, IIRC.

1) I thought the recent costs of nuclear were around $6/average watt. Sure, the Chinese are much lower, but what about western, and esp US costs?

2) that doesn't include operating costs, which are around 1.9 cents per KWH for existing plants.

Randall,

There's the cost of getting wind electric from a remote area to a city.

The $.25/KWp is the cost of getting the wind to a major nearby grid. I think that's all that's needed for the moment, and most of what's needed for the long-term.

What's the cost of transmission lines that are long enough and efficient enough that they can enable distant wind farms to serve as back-ups for each other?

Well, as we've seen, we can get 30% capacity credit (of average output) just by linking some small windfarms within a small area, and 55% capacity credit just by linking some windfarms within a region. That's taking us 55% of the way to where we need to be. It looks to me like a reasonable expansion of grid to grid links; aggressive (but very cost-effective) DSM; and retention of existing capacity (combined with time-of-day pricing to cap growth of peak demand) will get us the rest of the way.

It's really not that hard. Mainly it's a new way of doing business, which involves doing things smarter, rather than the traditional brute-force method of just building infrastructure (mainly generation).

Randall Parker said at June 10, 2009 7:03 PM:

Nick G,

I think you are optimistically guessing just how far the wind electricity needs to be transportable to contribute to baseload.

I also suspect that the correlations on wind intensity between wind farms varies by region. That Stanford study from year 2000 data seems far too small a basis for making conclusions about wind baseload potential for a variety of regions.

My position is that we do not know and we do not even know if people in the ISOs know.

Engineer-Poet said at June 10, 2009 7:31 PM:
The $.25/KWp is the cost of getting the wind to a major nearby grid. I think that's all that's needed for the moment...
I'd agree with that.
... , and most of what's needed for the long-term.
And violently disagree with that.  The bulk of the available wind energy in the USA is many hundreds of miles from major areas of consumption, so the per-watt baseload cost is going to be higher (maybe much higher) as we exploit significant amounts of the resource.  Also, as Randall has noted, you need more and more distance between wind farms if you expect them to back each other up to supply some minimum production.  This means lines which can carry the full immediate surplus from the area the full distance.  We would be far better off building smaller lines and using storage.

PWRs may be $6/watt, but MSRs would probably be half that or less.  Supercritical CO2 turbines are intrinsically cheaper than steam turbines because they are so much smaller, as well as being more efficient.  The more I look at this, the more I think it's our future and we let the ideologues stonewall it at our peril.

Interested said at June 11, 2009 12:59 PM:

Stumbled on this discussion while I was browsing around for some wind-related information. It seems to be a very open and honest discussion. I hope you don't mind if I throw out some questions. I don't want to ruin your discussion, so feel free to ignore me if you like, just looking for some more info.

I'm relatively new to the whole wind energy thing, and am not sure what my opinion is yet, but I've been trying to find out as much as I can. I have a few questions, and was hoping I could get some help with some answers here, as I've been having trouble getting answers anywhere else. These questions are probably a little basic for the types of discussions that are going on here, but I'm going to throw them out there. Any good source information you can provide me with to help me out with my questions would be much appreciated as well.

1: Who usually owns the wind turbines on wind farms? I've been able to determine that the land owners usually just "lease" the land, but is the owner of the turbines typically a power company or a 3rd party that then sells the electricity to power companies? If it is a 3rd party owner, how is the price determined that is paid for the energy (in $/kWh) and what are typical values of that price (I'm assuming it varies daily like other commodities).

2: It looks like wind turbines only get close to generating nameplate capacity in high wind conditions, and generate significantly less under lower wind conditions. Due to my understanding, wind speeds typically follow a Weibull/Rayleigh distribution, so most turbines spend the majority of their time operating in the "upward sloping" portion of their power curves. What would the effect be if it were possible to optimize a wind turbine to generate, say 20% more (just to throw out a number) energy in that region between cut-in wind speed and where it reaches nameplate capacity. I guess what I'm getting at here in a round-a-bout way, is why are wind turbines designed to operate most efficiently in the least likely wind scenarios? Seems kind of counter intuitive to me. Could a system be designed that extracts more energy out of the lower wind speeds (probably need bigger diameter) but shut down (to avoid damage) during high wind speeds?

3: What kind of technologies are out there to actively manage wind power. I know we are kind of at the mercy of mother nature here, but it seems that over a large enough number of wind turbines, the variations in each wouldn't have as much of an effect. Is the frustrations on the grid operators mostly because they have no control over anything? How much would it help them to be able to control (if that were possible) and what kind of control would they need. (Yeah, this question is kind of theoretical. I'm just trying to get a better understanding of the issues by examining the theoretical limits and the improvements if those limits were removed).


That's all I will bother you with for now. As I said before, I don't want to disrupt your conversation, so I'll understand if no one responds. Thanks for having this discussion available for anyone to read. I'm approaching my investigation from more of an engineering perspective to start with before I move to practicality/implementation issues, so most of the discussions above are based on things I haven't reseached yet. Anyways, if anyone can help me out here, I'd really appreciate it.

Thank you.

Nick G said at June 11, 2009 11:55 PM:

Randall,

I think you are optimistically guessing just how far the wind electricity needs to be transportable to contribute to baseload.

Well, we have some evidence. We have the Stanford study, we have the ISOs' analyses. There's quite a bit more out there, including a lot in Europe - they've been working on this quite hard, and for a while. Here's a study from Ontario: http://www.ieso.ca/imoweb/pubs/consult/windpower/wpsc-20090128-Wind-Variability.pdf
I've seen several other studies as well - I'll try to find more.

we do not even know if people in the ISOs know.

Why do you feel that way? The ISO's aren't theorists: they have to make this stuff work, on a daily basis. They're very conservative - if they say something positive about wind, I can't see any reason not to believe them.

E-P,

The bulk of the available wind energy in the USA is many hundreds of miles from major areas of consumption

West Texas wind is in the process of being brought to the eastern parts of ERCOT, as well as CA. Midwestern wind isn't far from Chicago & the whole Midwest. Chicago is connected to the East coast grid. The East is the farthest from the Midwest wind corridor, but it does have quite a lot of indigenous wind resources, albeit not quite as high quality.

This means lines which can carry the full immediate surplus from the area the full distance.

I think that overstates it. You just need balancing amounts, and they can move from grid edge to grid edge. Further, power import/export should only be one part of a whole menu of solutions. Eventually DSM should be very large: think of 220M EV's, each of which can soak up or put out 10KW of power for an hour, or .5KW for 20 hours: 2.2TW of power for an hour, or 110GW for 20 hours. And, the DSM part is free: the EV owners have already paid for it. The V2G part is slightly more theoretical and costly, but it would certainly be usable for unusual, high value occurences.

What's the risk if these measures don't work? We keep the 1,050GW of existing capacity - we have all the base-load, medium load and peak capacity we need - we really only need to displace FF.

We would be far better off building smaller lines and using storage.

Some buffering would certainly make sense. Again, a large menu of solutions.

MSRs...I think it's our future

Sounds good. What would be the first step for policy makers?

Interested,

why are wind turbines designed to operate most efficiently in the least likely wind scenarios?

I'm not sure what you mean. I think you're just looking at the power curve (power being cube of wind speed).

Could a system be designed that extracts more energy out of the lower wind speeds (probably need bigger diameter)

That's the natural evolution of wind turbines: every generation is about 40% larger.

What kind of technologies are out there to actively manage wind power.

You can feather the turbines, and turn down the power output. On the one hand, the marginal cost is zero, so you hate to do so, but you can if the alternative is an unacceptable risk of destabilizing the grid due to an unusual drop in wind speed. This may be a perfectly good strategy: you can toss away .5%-1% of wind output in exchange for reducing peak output disproportionately.

Interested said at June 12, 2009 5:46 AM:

Nick,
Thanks for addressing some of my questions. I have a couple of follow ups.

"why are wind turbines designed to operate most efficiently in the least likely wind scenarios?
Could a system be designed that extracts more energy out of the lower wind speeds"

You're right, these questions are based on the power curves I've seen. I'm sort of reaching back to some of the aerodynamics courses I took in undergrad, but it appears to me, based on the power curves, that blade airfoil shapes are designed to extract max power at nameplate capacity. I'm making an assumption that the power output is directly related to rpm, and thus directly related to wind speed, since it is the lift created by flow over the blade that causes rotation. Would it not be possible to design a blade that produces a higher lift-to-drag ratio at lower wind speeds, causing a higher rate of rotation at lower wind speeds, therefore creating a higher energy output. Of course, there must be some good reason that manufacturers aren't doing this. I'm not questioning their design, just trying to find out why they are designed the way they are. My "somewhat educated" guess is that an airfoil with a high L/D ratio at low wind speeds would stall out at high speeds, meaning that they would have to shut down and might not ever reach nameplate output. What I am interested in seeing is why the increase in power output at lower wind speeds is less (on an annual MWh basis) in this situation, then in the current design.

Thanks!


Nick G said at June 12, 2009 8:52 AM:

Would it not be possible to design a blade that produces a higher lift-to-drag ratio at lower wind speeds

Good question. I suspect that would optimize that range of the curve, but reduce KWHs more in the upper range, due to the cube-power relationship.

There's a lot of work currently on this kind of optimization. One interesting approach is to imitate whale flippers.

Nick G said at June 13, 2009 11:31 AM:

Good thought - I like it too. Also, the report from the NERC "Integration of Variable Generation Task Force" is pretty good: http://www.nerc.com/files/IVGTF_Report_041609.pdf

Good items include:

"Additional flexible resources, such as demand response, plug-in hybrid electric vehicles, and storage capacity, e.g. compressed air energy storage (CAES), may help to balance the steep ramps associated with variable generation. These resources allow grid operators to quickly respond to changes in variable generation output without placing undue strain on the power system."

and

"It is important to distinguish between variability and uncertainty when discussing planning and operations of the bulk power system. The effects of variability are different than the effects of uncertainty and the mitigation measures that can be used to address each of these are different."

Randall Parker said at June 13, 2009 8:36 PM:

Nick,

I deleted that post that had "solar panels" as the name of the commenter. I think it was spam.

Your "Integration of Variable Generation Task Force": Very interesting. I'm reading it.

My biggest concern about wind electric: I worry that political machinations will cause the costs of the variable generation to not be born by the variable generation providers. Cost allocation is a really big problem. I'm not at all confident it will be done fairly and efficiently.

Nick G said at June 14, 2009 1:01 PM:

Cost allocation is a really big problem.

What makes you think so? Have you seen anything quantitative?

I've seen such arguments from anti-wind and pro-nuclear partisans, but it didn't seem very sensible.

1st, power providers are paid separately for KWH and capacity credits (and other grid services). If wind doesn't provide reliable peak power, then it doesn't get paid for it. If nat gas or other generation sits around, doing nothing but providing the occasional backup or peak power, it gets paid for the privilege. That largely addresses concerns about providers of "good" generation (i.e., FF or nuclear) providing unreimbursed services for "bad" generation (that wily, unreliable wind power).

2nd, the external costs and subsidies received by other sources are so large that an occasional grid provided service (i.e., transmission and balancing) isn't very large in comparison. Think of coal's CO2 and other emissions, occupational hazards, and mountaintop removal; oil's military services (roughly $2T for Iraq alone!); nuclear's military and DOE R&D (the single largest thing the DOE does!), Price-Anderson liability cap and last but not least, proliferation dangers (think Iran, N Korea and the Iraq war, the primary public justification for which was...WMD).

I'd dearly love to see a thorough accounting of external costs, subsidies and improperly allocated costs. I saw one done by the EU, but it was badly incomplete. Have you seen something good?

Randall Parker said at June 15, 2009 12:18 AM:

Nick,

On page 16 of that variation generatio task force report they show how wind increases the need for dispatchable ramping capability. So wind requires bigger transmission systems, tax credits, and more dispatchable ramping capability - probably from burning natural gas.

Nick G said at June 15, 2009 11:26 PM:

Randall,

There's no new information here. It's well known that wind blows as hard or harder at night. This is arguably a bigger problem for wind than variance, and it's a problem it shares with nuclear.

Look at the footnote - it says: "If we assume that conventional generation resources provide all the ramping capability for the system..."

Picture 220M EV's providing night time demand and soaking up all that variance. They're not here yet, but they'll ramp up pretty much in parallel with wind.

Nick G said at June 16, 2009 8:31 AM:

Also, while the addition of wind capacity may not increase the amount of peaking capacity by as much as other kinds of plant, it doesn't decrease the amount of peaking capacity, or create a need for more of it.

There may be instances in which taking advantage of wind production, especially during night time base-load conditions, creates more production variance and makes System Operators prefer a switch from less flexible plant (like coal) to more flexible plant (like Nat Gas). On the one hand, 1) as NG is much cleaner, this is a good thing, and 2) this doesn't require more plant, just changes the mix and utilization of existing plant. In the case of NG plant, it's typically less than 25%, and often as low as 7%.

Randall Parker said at June 16, 2009 7:39 PM:

Nick,

Wind cuts the demand for non-wind power. A base load plant's ROI is lowered when it competes at night with wind farms that are selling at below 0 cents per kwh.

Old baseload plants get retired eventually. They'll get retired sooner when their ROI goes down due to wind (and due to other subsidized power sources). Plus, fewer new baseload plants will get built as demand rises.

Reduced supply of baseload will raise electric power rates during times when the wind does not blow even as more wind power lowers electric power rates when the wind does blow. Reduced supply of baseload will increase demand for peaking plants. Yes, that means more natural gas for electric power. Though those natural gas plants won't be as efficient as baseload natural gas plants.

Wind is ramping up already. EVs are rare. Yes, EVs will boost night time demand eventually. When will electric power rates start becoming variable for home owners?

Engineer-Poet said at June 16, 2009 10:17 PM:

Nuclear isn't affected by the diurnal cycle, so I don't see how it shares any "problems" with wind.  What nuclear shares with wind is a negligible fuel cost, so it makes no sense to reduce generation unless the grid is oversupplied.  Wind's problem is that it cannot be scheduled and its costs are mostly up front; nuclear's problem is that it has high capital costs and a long interval between breaking ground and going on line.

Both nuclear and wind argue for market-based electric rates and opportunistic usage.  "Ice Bear" air conditioners, electric vehicles selling regulation services, water heaters which switch from gas to electric when the rates are right... all of these things make sense when the price of juice changes according to the immediate situation of supply and demand.

Nick G said at June 17, 2009 10:34 AM:

Old baseload plants get retired eventually. They'll get retired sooner when their ROI goes down due to wind

Not if they're needed. Old coal and nuclear plants are much, much cheaper than new plants - they'll stay for quite a while, unless we take them out and shoot them.

(and due to other subsidized power sources).

How is wind more subsidized than other sources, esp coal?

Plus, fewer new baseload plants will get built as demand rises.

That's not a bug, that's a feature. As E-P notes, we need variable pricing, not more plant.

more natural gas for electric power.

Probably we'll see a bit of a transfer from coal to nat gas. That's a good thing. We'll see less FF for electrical generation, and that's very good.

Wind is ramping up already. EVs are rare.

Yes, and wind isn't large enough to need them yet.

When will electric power rates start becoming variable for home owners?

It's starting right now. PG&E is installing smart meters for all of their customers. The 2005 energy act mandated the availability of smart meters for allutilities.

Nuclear isn't affected by the diurnal cycle, so I don't see how it shares any "problems" with wind.

I think we see this fundamentally the same way. What I meant was: both wind and nuclear have a problem with unwanted night time production. In the US this isn't that important because nuclear isn't as big as, say, France. OTOH, I think it's clear that some nuclear advocates have a prejudice against wind because the two sources are going to compete for night time customers.

Bill Begosh said at October 5, 2009 7:14 AM:

It is my understanding that when the wind velocity increases beyond design parameters, the blade pitch is changed (feathered?) to eliminate damage to the wind generator. If this is true, why not employ a mechanism such as centrifugal clutch,torque converter to connect the generator to a greater load such as a water pump. This extra load would control the blade RPM and the pump would fill a water tower located in the wind farm. When there is no wind, the water could be released to power the generator with out the blades rotating. I would think this method would increase the efficiency of wind energy in general by reducing "down time"
I am NOT in favor of wind power because of the infrastructure requirements and the idea of "centralized power." Also I think wind generators are a form of visual pollution. Existing high tension towers already alter our landscape. Can you imagine all of those towers in motion?
All that said, if we are going to have wind power, I believe we should make it more efficient.

Nick G said at October 5, 2009 4:30 PM:

Bill,

I think that the feathering protects the blades and tower as well: they all have to be sized to a certain maximum load. Keep in mind that wind turbines are already over-sized to capture those high-power periods - that's why their average utilization is a relatively low 15-45%.

Dan Brown said at November 12, 2011 7:08 PM:

There are many issues with large wind farms that need space and wind and are a long way from power users. It needs to become socially acceptable for small residential windmills to be common place along side solar panels. Wind and solar work well together, the wind is often blowing when it is raining or cloudy.

Post a comment
Comments:
Name (not anon or anonymous):
Email Address:
URL:
Remember info?

                       
Go Read More Posts On FuturePundit
Site Traffic Info
The contents of this site are copyright ©