|
Title:
|
|
Wind power predictions analysis. Part. 2: economical analysis
|
|
|
|
Author(s):
|
|
|
|
|
|
Published by:
|
Publication date:
|
|
ECN
Wind Energy
|
1-8-2002
|
|
|
|
ECN report number:
|
Document type:
|
|
ECN-I--02-011
|
Other
|
|
|
|
Number of pages:
|
Full text:
|
|
36
|
Download PDF
|
Abstract:
In the first part, the five daily runs of predictions provided by thewind power prediction method developed by ECN have been compared. It
appears that the freshest runs are not always the most accurate: their
level of power is not accurate enough. It is concluded that the run
of predictions including predictions beginning at 12 o'clock (UT) the
previous day turns out to be the most accurate.
Then, we have compared different ways to fill an E-program: submitting
a constant value for the E-program, using one run provided by the power
prediction method (run 12) and updating the E-program each time there
is a new set of predictions available. Here the imbalance costs is not
a good criterion to take a decision in so far as it is not representative
of the amount of imbalance sent to the grid. The best way to proceed
is thus to use the run 12 to predict one?s production for the next day
because it will bring about the smallest amount of imbalance.
In the final part of this study, we have made a method to help the producer
to decide if it is worthwhile making change when E-program change is
not free. This method computes the updating gains thanks to the improvement
of the prediction and to the TenneT imbalance prices model. Then, if
gains are superior to benefits, there is an E-program change. This method
computes the updating gain with a confidence level of 60%.
Back to List