The Generation Professionals Group is for utility professionals who work in biomass, coal, gas/oil, hydro, natural gas, or nuclear power generation fields. 

Post

Why the Recent Texas Reliability Event Raises No Concerns About Much Higher Wind Penetration

Rob Gramlich's picture
President Grid Strategies LLC

Rob Gramlich has been working to drive grid modernization and expansion efforts since the mid-1990s. Through senior positions in government and the private sector, he has developed and led...

  • Member since 2005
  • 8 items added with 4,305 views
  • Apr 2, 2008
  • 914 views
On February 26, 2008, a drop in frequency on Texas's transmission grid caused the Electric Reliability Council of Texas (ERCOT) to put in place an Emergency Electric Curtailment Plan. The event was reported in some media outlets as having been caused by a sudden drop in output from wind projects. In fact, as the information below makes clear, other factors had a greater impact in this particular incident.

Over the 40-minute period preceding the start of load curtailment, wind generation declined by 80 MW relative to its schedule, non-wind generation decreased by 350 MW relative to its schedule, and load rapidly increased to a level that was 1,185 MW more than forecast. More generally, disturbances of this type routinely involve conventional power plants.

The incident in reality shows that the key to successful electric system management lies in a diverse power supply, of which wind can reliably be a large part. The American Wind Energy Association has pulled together the following background information and highlights from the report issued by ERCOT on the event. This provides both perspective on the event and insights into the issues going forward.

Complex electric systems are comprised of many dynamic parts, including power plants, fuel supplies, power delivery infrastructure, and aggregate levels of consumption. Maintaining the reliability of an electric system requires robust operating procedures to handle contingencies, including multiple simultaneous events. As this report establishes, ERCOT has such procedures in place and they work. Yet, particularly as regional fuel sources and power options are diversified, there are improvements to grid operating procedures that can and should be made.

As an action item in response to this event, ERCOT has determined to incorporate wind forecasting into short term planning to better inform its capacity procurement needs. The wind industry supports and applauds that conclusion, as it will enable continued diversification of power supplies while maintaining the robustness and reliability of the grid. As a GE study commissioned by ERCOT recently reported, with the use of wind forecasting wind energy deployment in Texas can be more than tripled to at least 15,000 MW without negative consequences for system reliability.

Some observations regarding the February 26, 2008 ERCOT event:

  • The drop in frequency was successfully managed and no customers involuntarily lost power. The event was the result of several factors including sudden output shortfalls from conventional generators and a much slower decline in wind output. To the extent that changes in wind were a contributing factor during the passage of the February 26 cold front that can be resolved by integrating the use of wind forecasting into electric system management, as ERCOT is planning to do.
  • Wind is not dispatchable, but it is generally predictable, and therefore can be successfully and reliably integrated into electric system management. Wind forecasting accurately predicted ERCOT's February 26 decrease in wind generation. This event illustrates that wind forecasting is an important tool for the successful management of large amounts of wind generation on a grid. ERCOT’s wind forecasting system is under development and they plan to accelerate its implementation. When it is in place, such changes in wind output would not be expected to contribute to any system problems.
  • Unlike other sources of generation that can go offline in 1/60th of a second, wind's declining output tends to be gradual over a matter of hours, giving system operators more time to respond to changes. Fossil and nuclear power plants can and frequently do trip offline instantaneously, as the power outage on February 26 in Florida demonstrated. In ERCOT, 13 conventional generating plants instantaneously tripped offline during the week following February 26th. In the largest of these incidents, 420 MW, 500 MW, 540 MW, 582 MW, and 650 MW were instantaneously lost because a conventional generating unit tripped offline. The February 26th incident was preceded by the loss of a 150 MW non-wind generating unit that tripped offline at 5:44 PM. In comparison, wind generation decreased steadily on February 26 by about 8 MW/minute, during a 3-hour interval.
  • The event was caused by a combination of factors, one of which was non-wind generators falling short of their scheduled output. The activation of responsive reserves and interruptible load reserves at 6:49 PM on February 26 immediately followed a 40-minute period during which conventional generation dropped by 350 MW relative to its schedule. Although wind was consistently below its scheduled output during this time period, from 6:27 to 6:50 wind output gradually increased by 40 MW relative to its schedule.
  • Another factor that contributed to the event was the rapid rate of load increase that evening. The load increase on the evening of February 26th was 3,150 MW larger than the previous evening’s load increase. As a result, ERCOT’s short-term load forecast underestimated the actual load at 6:45 PM by 1,185 MW.
  • To sum up, over the 40-minute period preceding the start of load curtailment, wind generation declined by 80 MW relative to its schedule, non-wind generation decreased by 350 MW relative to its schedule, and load rapidly increased to a level that was 1,185 MW more than forecast.

The use of “demand response” is not unusual and is a proven tool in cost effectively maintaining grid reliability. ERCOT is a leader in the use of paid voluntary interruptible loads (demand response) to accommodate significant variations in load, the sudden loss of conventional generation plants, and to avoid subjecting customers to involuntary power outages. For example, on December 12, 2007, three conventional generation plants totaling 1,022 MW tripped offline. Frequency on the ERCOT grid dropped from 60 to 59.787 Hz, significantly worse than the drop from 60 to 59.85 Hz that was registered on February 26. According to ERCOT, this qualified as a potential NERC Disturbance Control Standard event, while the February 26th event did not. In the December 12th incident, ERCOT was able to prevent an involuntary power outage by deploying all of the interruptible load reserves that were available.

ERCOT's full report

GE’s study

Rob Gramlich's picture
Thank Rob for the Post!
Energy Central contributors share their experience and insights for the benefit of other Members (like you). Please show them your appreciation by leaving a comment, 'liking' this post, or following this Member.
More posts from this member
Discussions
Spell checking: Press the CTRL or COMMAND key then click on the underlined misspelled word.
Ferdinand E. Banks's picture
Ferdinand E. Banks on Apr 3, 2008
Rob, what is your reaction to the following observation: Denmark has the highest electricity price in the industrial world, and apparently three times as high as the average electricity price in the U.S.

Fred

Rob Gramlich's picture
Rob Gramlich on Apr 4, 2008
Fred, thanks for spurring the discussion on this topic. To make sure I could respond to your observation accurately, I looked up historical data on the price of electricity in Denmark and compared it to data showing the electricity generation mix in those years. This data clearly shows that, adjusting for inflation, the price of electricity increased drastically in the early 1980's and has fairly consistently remained at that level ever since. Because wind energy did not become a significant part of Denmark's generation mix until the late 1990's, it would be difficult to blame wind energy for Denmark's high electricity prices. Instead, the major cause of this price increase in the early 1980's appears to have been a decision by the government to impose heavy taxes on electricity use.

It's interesting that you bring up Denmark, as the country serves as an excellent example of how wind forecasting makes it possible for a region to obtain a very large percentage of its electricity from wind energy without any negative impacts on the reliability of the electric grid. Denmark obtains more than 20% of its electricity from the wind, and at certain moments almost all of the country's electricity is provided by wind. Given that the U.S. is a vastly larger country with a much more diverse wind resource, through the use of wind forecasting we should be able to increase our use of wind energy many times over without any negative impact on the electric grid's reliability.

Len Gould's picture
Len Gould on Apr 7, 2008
I don't suppose Denmarks access to much larger reliable sources in neighbouring countries, including France's huge nuclear baseload generation and Sweden and Norway's quick-response storage hydro might be factors? Pretty hard to design comparable systems for the entire US.

I think a smart grid and integrated market system is the only real alternative if N America is really determined to do a significant percentage wind generation.

Rob Gramlich's picture
Rob Gramlich on Apr 8, 2008
Len, thanks for adding to the discussion. The key to integrating large amounts of wind energy is taking advantage of the flexibility that is already built into the electric grid. This flexibility has been built into the grid to accommodate variations in electric demand, which changes on time scales ranging from seconds to years. Often these variations are unexpected, as are the sudden shortfulls in electricity supply that can be caused by the loss of a generating plant. Flexible generators -- power plants that can vary their output on demand -- are used to make sure that changes in electricity demand are accommodated with changes in supply. These same flexible generators can be used to accommodate unexpected changes in wind output, and this is how wind energy has been successfully integrated in the U.S. and around the world.

A number have studies have found that there is enough flexibility in the US grid to support a much larger amount of wind generation. An excellent compilation of these studies can be found on the Utility Wind Integration Group's website, at the following link: http://www.uwig.org/opimpactsdocs.html . Of course, reforms that will make the U.S. electric grid smarter and more integrated are welcome changes that will improve the grid's reliability for all users while at the same time making it easier to integrate even larger amounts of wind.

Jim McDowall's picture
Jim McDowall on Apr 8, 2008
Rob, thanks for an interesting article. What you describe as "wind generation declined by 80 MW relative to its schedule" is described by ERCOT as "wind production dropped from over 1700 megawatts (MW) three hours before the event, down to 300 MW at the point the emergency procedures were activated." One view seems to say that the glass is almost full, while the other says it's 80% empty.

You're absolutely correct in saying that ERCOT needs to build wind forecasting into its system planning, but the fact remains that integrating high levels of wind generation into the system does impose strains on other resources. Wouldn't it be more productive for AWEA to be involved in a frank discussion about coping with variability, rather than simply stating that there are "no concerns about much higher wind penetration"?

Rob Gramlich's picture
Rob Gramlich on Apr 9, 2008
Jim, Thanks for your comment. As far as how we characterize such events, we were reacting to some statements in the press implying that wind development might be limited as a result. We think that view needs to be challenged. We try to be very up front about the challenges system operators face and we agree with you that frank discussions are needed about the challenges of dealing with increased variability on the system. Operators need better tools and market structures to help them do their job and we've been working with NERC, IEEE, FERC, WECC, UWIG and others to help on this agenda. We're open to doing more or clarifying anything wind can and cannot do. We hope to see everyone interested in this issue at next week's UWIG meeting--that is the best forum for such discussions (see www.uwig.org).
Jim Beyer's picture
Jim Beyer on Apr 9, 2008
I will admit not fully understanding all the details, but according to the report, the total load at the time of the event was about 31,000 MW. And the loss from the idle wind turbines was 80 MW. This is less than 0.2% of the load. It's not clear what the total wind input to this grid is, but I would assume it's much less than 1%. I don't see this event as being very significant one way or another with respect to wind generation and grid stability.

Given what transpired, a very high wind generation component to the grid, say 10% or 3100 MW, could be troubling. If you had a sudden loss of 3000 MW, instead of 430 MW, the problem would be much more severe.

Maybe I'm missing something here.

Jim McDowall's picture
Jim McDowall on Apr 9, 2008
Jim, ERCOT just filed a study on transmission plans for future wind generation ranging from 12,000-24,000 MW, so if that scenario transpires the challenges will become greater. Balancing wind output with flexible generation, as Rob mentions above, is OK as long as it's reasonably clean generation - otherwise the whole thing is kinda pointless. Hydro, pumped or otherwise, is probably the best bet but it's not going to be available in all locations. A lot of people in the storage community are looking at compressed air energy storage (CAES) as a potential savior.

I'm a fan of wind generation but believe it has to be implemented not just on its own, but in conjunction with whatever clean balancing solution makes the most sense.

Jack Ellis's picture
Jack Ellis on Apr 9, 2008
My colleague Ellen Wolfe and I recently reviewed a California ISO's report (http://www.caiso.com/1ca5/1ca5a7a026270.pdf) that talks about the challenges of integrating renewable resources. The CAISO concluded that while it would have to procure additional amounts of ancillary services, the incremental purchases were well within the capability of existing resources for renewable penetration rates of up to 20%.

We agree - the additional amounts of flexibly dispatchable resources that are required to accommodate a lot more wind generation are surprisingly modest. In fact, the debate over whether existing resources are adequate is somewhat misplaced - if the price is right, owners of conventional steam and combined cycle plants will gladly take actions that shorten the service lives and increase maintenance costs for their plants, and they could almost certainly do so at a lower overall cost than some of the storage technologies that are thought to be necessary. The carrying cost of a new combustion turbine, exclusive of fuel and O&M, is about $150/kW/year in California and storage costs at least twice as much, whereas the average cost of ancillary services in California has remained under $100/kW/year since the end of the energy crisis (http://www.caiso.com/1f9c/1f9c8d5f22300.pdf, page 3). With the higher demand for ancillary services that would accompany higher renewable energy penetration levels we might expect ancillary services prices to increase, but we expect them to remain well under the cost of new resources, especially if California can every figure out how to use demand management for some of its balancing needs.

I'm not suggesting that wind can be integrated for free. Particularly in California, substantial amounts of conventional generation will be required to provide a backstop for intermittent generation on hot summer days when demand is high because the wind stops. It appears, though, that the cost of balancing the grid around wind output won't be quite as onerous and the technical problems won't be quite as severe as conventional wisdom suggests.

Donald Harker's picture
Donald Harker on Apr 10, 2008
The third bullet strikes me as not being an accurate comparison. The nuclear and coal plants are tripping based on safety systems or grid disturbances not availability of fuel supply. Wind would trip instantaneously as well based on a signal to isolate generation for grid integrity. A more proper comparison would be a nuclear plant limited in output by an LCO or some type of equipment degradation.

Don Harker

Rob Gramlich's picture
Rob Gramlich on Apr 10, 2008
Don, thanks for your feedback. The third bullet point addresses the impact on the grid from the system operator's perspective, which is what matters from a grid reliability standpoint. If a nuclear or coal plant trips offline for internal safety reasons or to protect its equipment from a perceived grid disturbance (as they often do), the grid operator likely won't know or care why it has tripped offline. All he/she will know and care about is that a large amount of generation was suddenly lost (1,000+ MW for the biggest plants). You are correct that a wind turbine can suddenly trip offline for the same reason, although in that instance one would only lose 1.5 or 2 MW, not 1,000 MW, which is a world of difference for a system operator.
Ferdinand E. Banks's picture
Ferdinand E. Banks on Apr 21, 2008
Everything is relative in this old world of ours, Rob, but it so happens that I don't agree with you about the development of Denmark's electricity price. The key things for me are (1) in a mainstream economics textbook sort of world the Danes would have petitioned for the nuclear reactors at Barsebäck to stay open instead of insisting that they close, which would have helped to reduce the price of Danish electricity, and (2) despite constant talk about how wonderful wind is, the Swedes are dragging their feet when it comes to building new installations. They are dragging their feet because they know that increasing the size of the wind sector would be a mistake.

Having said let let me note that I like the way you immediately reply to most comments on your article. Having that option helps to keep the know-nothings in their place.

Fred

Terry Embury's picture
Terry Embury on Apr 22, 2008
The main problem is that all of the wind is in the west zone of ercot, with very little transmission capacity to get it to the major load zones in Houston and North zones. Looking at total ercot load at the time is interesting but not at all the problem. The problem is the small amount of load in the west zone as compared to the very large, and getting larger amount of wind generation in the west zone.

This problem is not at all like wind in Calif which is on a main 500kv line from Phoenix to LA, with a major load pocket nearby.

Transmission systems need to get upgraded on a massive scale from west zone to Houston and/or North to be able to move all of that wind to the load pockets, otherwise they quickly overwhelm the diapatchable generation as a percent of total generation to load in the west zone.

The true costs of wind should include the required transmission and additional ancillaries upgrades required to make it work: dispatchable units waiting on stand-by or LSL to pick up load at much less efficient heat rates, intra-day gas scheduling around wind swing, additonal gas storage for the dispatchable units to deal with wind swing, etc.

While I'm a fan of renewables, it doesnt appear that the full costs to the system or the end user or ratepayer is being fully disclosed.

Bradley Collard's picture
Bradley Collard on Jul 8, 2008
Yes, ERCOT will need to allow utilities to build transmission lines to the load centers from the wind regions, as well as promote wind closer to load centers. While nukes and coal burners tripping causes some instability, the key here is that nukes and coal burners are figured into the multiple contingencies that Rob talks about.

Wind is protected in that it does not have to secure spinning reserves. Wind is a fuel source--it is time we started treating it as such. Wind does not stop blowing all at once across the state, but does stop blowing in regions as it did in February. While 80 MW does not seem like a large amount, it can make the difference when you are running the system tighter. ERCOT was already experiencing difficulties and the wind contributed to the problems. Part of the contingency plans that Rob refers to includes ERCOT issuing Emergency Electric Curtailment Plan (Emergency Operations) when the order to drop firm load can come into play as a last resort--that means nieghborhoods.

I am a big believer that wind will play an important role in America's energy future--it does have a place, but lets be responsible that includes looking at alternatives to back up wind, nuclear, coal, gas, and combustion turbines. Wind does have its rightful place in the mix of energy resources to supply power in many states.

Get Published - Build a Following

The Energy Central Power Industry Network is based on one core idea - power industry professionals helping each other and advancing the industry by sharing and learning from each other.

If you have an experience or insight to share or have learned something from a conference or seminar, your peers and colleagues on Energy Central want to hear about it. It's also easy to share a link to an article you've liked or an industry resource that you think would be helpful.

                 Learn more about posting on Energy Central »