Fraunhofer-Gesellschaft

Publica

Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Estimation of available active power in wind farms for operating reserve provision

Presentation held at the Conference and Exhibition "Brazil Windpower 2013", 3-5 Septermber 2013, Rio de Janeiro, Brazil
 
: Kaminski Küster, Kristie; Schneider, Dominik; Siefert, Malte; Speckmann, Markus

:
Fulltext urn:nbn:de:0011-n-3600567 (739 KByte PDF)
MD5 Fingerprint: 0338b2f0c0394418b4c17c02c62a2c5c
Created on: 17.9.2015


2013, 8 pp.
Brazil Windpower Conference and Exhibition <2013, Rio de Janeiro>
English
Presentation, Electronic Publication
Fraunhofer IWES ()
wind integration; operating reserve; wind farms; available active power; nacelle wind speed correction

Abstract
There are several reasons why wind farms should provide operating reserve, even though this is generally associated with a limitation of the wind energy yield in the wind plant. Among others, the provision of operating reserve by wind can: avoid turning off renewable generation in favor of conventional provision of ancillary services; be a cheaper alternative to attend the demand for operating reserve as opposed to extra storage or conventional plants running at poor efficiency; enhance competition in the balancing power market; and increase economic feasibility to renewable energy investment by generation of extra revenue.The available active power (AAP) of a wind farm is defined as the power a wind farm could produce if it had not been curtailed. The exact estimation of the AAP is crucial for the provision of operating reserve by wind farms as a method to prove the offer and delivery of operating reserve.Until now only few techniques for the estimation of the AAP are available. In this paper five different techniques are investigated and compared: Last Measured Value, Adjusted Power Curve, Site-Specific Power Curve, Reference Turbine and Physical Model. For the tests, ten-minute-averaged data from a wind farm in Brandenburg (Eastern Germany) measured over 21 months were used. Four criteria of comparison are defined: accuracy under normal operation, accuracy under curtailment, consideration of changes in the wind farm’s composition and costs. The method Physical Model showed a normalized Root Mean Square Error (nRMSE) smaller than 1.7% for all curtailment cases and durations.This innovative research features the comparison and analysis of the different AAP estimation techniques using historical data from a real wind farm, and is the first work to introduce the Reference Turbine method implemented with Artificial Neural Networks.

: http://publica.fraunhofer.de/documents/N-360056.html