ERCOT Event on February 26, 2008: Lessons Learned Page: 9 of 13
This report is part of the collection entitled: Office of Scientific & Technical Information Technical Reports and was provided to UNT Digital Library by the UNT Libraries Government Documents Department.
Extracted Text
The following text was automatically extracted from the image on this page using optical character recognition software:
almost no cost of implementation. This type of assessment works very well for flatter
periods, but decreases in effectiveness for larger ramping periods like the one during this
event.
Most state-of-the-art short-term wind energy forecasts build upon this idea by taking
recent data and using more sophisticated mathematical techniques to produce a forecast
of the wind energy. This is very similar to the short-term load forecast, where a major
predictor is past behavior. This causes a characteristic error of "delay." The forecast tends
to trail reality as it is strongly influenced by a persistence forecast. Alternatively, a Day-
Ahead wind energy forecast may have the correct profile of the wind production, but
shifted in time forward or backward. This would also give larger errors when the wind
production is moving in fast ramps before or after it was expected. In both instances, it is
up to the system operator to prepare for the wind event, and by looking at both the
forecast and actual generation data, act appropriately.
Some of the more advanced short-term wind energy forecasts will use offsite
observations to get signals about the upcoming ramping behavior. With the ideal scenario
of having multiple offsite observations all being a radius of an hour away, an oncoming
ramp could be more accurately forecast an hour ahead of time.
Setting aside the algorithms that come up with a wind energy forecast, the forecast itself
needs to be presented to the system operators in a way that allows them to prepare for
large changes in wind generation. This may include using an appropriate advisory time
frame, as well as an assessment of the variability inside a normal forecasting time frame.
Since day-ahead unit commitment is an evaluation of the entire day, a full-day wind
forecast is usually used in conjunction with a full-day load forecast. During real-time
operation, system operators still need to see a few hours ahead, and be able to prepare for
large increases and decreases of wind generation since it may be possible that slower
start-up units may have to be turned on to accommodate lost generation. It can also give
efficient solutions based on the duration of a wind power increase or decrease. For
example, let's assume the wind forecast predicted that during the February 26th event,
the wind power decreased in hour 18, but by 19:30, the wind generation began to increase
to where it had been previously. The system operator in this case would know that it
might make more sense to have some quick start units with shorter minimum run times
on during this time, rather than to bring up a unit with a long minimum run-time and most
likely, a high startup cost.
Figure 4 shows the hour-ahead resource plan for all wind given by the QSEs, the wind
forecast from the forecast provider to be used in the future ERCOT nodal system
(forecast uses an 80% exceedance level), and the actual wind generation output.6
Upcoming Pages
Here’s what’s next.
Search Inside
This report can be searched. Note: Results may vary based on the legibility of text within the document.
Tools / Downloads
Get a copy of this page or view the extracted text.
Citing and Sharing
Basic information for referencing this web page. We also provide extended guidance on usage rights, references, copying or embedding.
Reference the current page of this Report.
Ela, E. & Kirby, B. ERCOT Event on February 26, 2008: Lessons Learned, report, July 1, 2008; Golden, Colorado. (https://digital.library.unt.edu/ark:/67531/metadc894235/m1/9/: accessed April 16, 2024), University of North Texas Libraries, UNT Digital Library, https://digital.library.unt.edu; crediting UNT Libraries Government Documents Department.