Use of Artificial Neural Networks for Short Term Wind Power Predictions
By Alla Sapranova
Uni Research is developing models for nowcasting based on artificial intelligence (AI) and artificial neural network (ANN), in particular. The main objective of this work is to enrich the management of wind turbines and wind farms by providing a reliable and fast running tool that is capable to reduce uncertainty in short term forecast by filling a gap between predicted and real data (like wind speed or power output).
ANN is a non-physical based approach for dealing with nonlinear tasks, and especially with forecasting and pattern recognitions. ANN-based models are fault tolerant in the sense that they are able to handle noisy, incomplete and non-linear data sets. Once trained, ANN-based model will perform reliable predictions and generalizations instantly.
On March 2012, as a result of NORCOWE - Fraunhofer cooperation field data from Nordland Windfarm became available. Received is the data for wind speed, direction, power, electrical energy export/import, load, operational time for 3 identical wind turbines (out of total 11 in the park) type Tacke TW-500 with hub-height 40 m and wind speed data from a 40m high metmast. All variables are available as 1) 5 min average, minimum, and maximum values for total of 14 hours observation period for a single day in 2007, and 2) 10 Hz frequency data for 3, 2 and 3 hours of observation periods for 2 days in 2007. Read full article
24. 09. 2018
Lys fremtid for flytende offshore vindparker spår industrien....
13. 07. 2018
Open positions related to offshore wind at University of Bergen. ...
17. oktober 2018
Workshop in collaboration with GCE NODE...