Data mining and wind power prediction: A literature review


Colak I., SAĞIROĞLU Ş., YEŞİLBUDAK M.

RENEWABLE ENERGY, vol.46, pp.241-247, 2012 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Review
  • Volume: 46
  • Publication Date: 2012
  • Doi Number: 10.1016/j.renene.2012.02.015
  • Journal Name: RENEWABLE ENERGY
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Page Numbers: pp.241-247
  • Keywords: Data mining, Data mining techniques, Wind power prediction, Prediction time scales and models, Literature evaluation, TURBINE, MODELS, FUZZY
  • Gazi University Affiliated: Yes

Abstract

Wind power generated by wind turbines has a non-schedulable nature due to the stochastic nature of meteorological conditions. Hence, wind power predictions are required for few seconds to one week ahead in turbine control, load tracking, pre-load sharing, power system management and energy trading. In order to overcome problems in the predictions, many different wind power prediction models have been used to achieve in the literature. Data mining and its applications have more attention in recent years. This paper presents a review study banned on very short-term, short-term, medium-term and long-term wind power predictions. The studies available in the literature have been evaluated and criticized in consideration with their prediction accuracies and deficiencies. It is shown that adaptive neuro-fuzzy inference systems, neural networks and multilayer perceptrons give better results in wind power predictions. (C) 2012 Elsevier Ltd. All rights reserved.