Optimal control of wind turbine using neural networks

Mahinsasa Narayana, Ghanim Putrus

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

4 Citations (Scopus)

Abstract

Variable-speed, fixed-pitch wind turbines are required to optimize power output performance without the aerodynamic controls. In steady-state, a wind turbine generator system is operated such that the optimum points of wind rotor curve and electrical generator curve coincide. In order to obtain maximum power output of a wind turbine generator system, it is necessary to drive the wind turbine at an optimal rotor speed for a particular wind speed. Therefore, accurate wind speed measurements are required for optimal operation of the wind turbine. In practice, it is difficult to accurately measure wind speed by an anemometer installed closed to the wind turbine, because the wind turbine experience different forces due to wake rotation. Therefore, it is useful to use a wind speed sensor less control strategy. In this study, a Nonlinear Autoregressive Moving Average (NARMA) neural network model is used to identify the combined performance of the wind rotor and generator. Wind speed sensorless optimum control strategy is introduced and comparison study is preformed with a controller that employs a wind speed sensor. According to the obtained results, proposed controller performs as good as to the controller that employed with wind sensor.
Original languageEnglish
Title of host publicationProceedings of the 45th International Universities Power Engineering Conference (UPEC)
Place of PublicationPiscataway, NJ
PublisherIEEE
Pages1-5
ISBN (Print)978-1424476671
Publication statusPublished - Aug 2010
EventUPEC 2010: 45th International Universities' Power Engineering Conference - Cardiff, Wales
Duration: 1 Aug 2010 → …

Conference

ConferenceUPEC 2010: 45th International Universities' Power Engineering Conference
Period1/08/10 → …

Fingerprint

Dive into the research topics of 'Optimal control of wind turbine using neural networks'. Together they form a unique fingerprint.

Cite this