|
Title:
|
|
Wind power predictions analysis. Part. 1: TenneT imbalance price system. Development of a model for TenneT imbalance price
|
|
|
|
Author(s):
|
|
|
|
|
|
Published by:
|
Publication date:
|
|
ECN
Wind Energy
|
1-8-2002
|
|
|
|
ECN report number:
|
Document type:
|
|
ECN-I--02-010
|
Other
|
|
|
|
Number of pages:
|
Full text:
|
|
44
|
Download PDF
|
Abstract:
To preserve the system balance in the Dutch power system "energy-programmeresponsibility" was set up. The program responsibility is basically
a system in which all market parties involved in the use of the network
send on a daily basis their planned production, transport and use of
electricity to TenneT in the form of so-called E-programmes. Using that
information as a starting point TenneT, the Dutch grid system operator,
can correct any imbalances or internal transmission constrains generated
in the system.
When a producer doesn't produce according to the predictions he has
registered in the E-programme, he has to pay compensations for imbalance
(difference between expected and actual output power) to TenneT. In
wind energy, previsions made in the E-programme are hardly ever equal
to the actual output. For this reason it is assumed that wind energy
producers have to deal with large imbalance costs. Therefore it is important
to have a good understanding of the imbalance pricing system that is
ruling their imbalance.
In this study, we have first made a global statistical description of
imbalance prices for shortage and surplus in the period from February
2001 to January 2002. This first approach has highlighted the fast fluctuations
of imbalance prices as well as the wide range containing these prices.
Then, to have a better understanding of the behaviour we have made a
spectral analysis of these prices. The main harmonics appeared for every
multiple of 24 hour: a daily pattern is thus the first step to describe
these prices. We can distinguish different daily patterns for the weekend
and the rest of the week, as well as for every month of the year. These
daily patterns seem to be linked with consumption in so far that imbalance
prices remain low during the night but increase from 7 o'clock to 20
o'clock.
On the basis of this description, a prediction model was developed.
This model provides for the imbalance price two daily patterns for one
month: one for the weekend and another one for the rest of the week.
A daily pattern includes a mean expected value as well as a minimum
and a maximum expected value. Using the first months of the year 2002,
we have shown that the range defined by the minimum and the maximum
expected price contains the imbalance price with a confidence level
of 85%.
Back to List