Prediction of short term power output of wind farms based on least squares method

S. Dutta, T. J. Overbye

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Ability to reasonably predict the power outputs of wind farms enables a better management of reserves and control mechanisms of turbines. Wind power forecasting is also important in assessing the wind energy potential of a region at the planning stages. This paper focuses on the prediction of power generation from wind farms in the short term. It proposes a least squares based method of predicting the total power output of a group of wind farms distributed over a region. A test study has been conducted by predicting the wind power output of four hypothetical wind farms located at four counties in the state of Illinois, the Bureau County, Henry County, Mercer County and Knox County. The wind speeds at the latter three counties predicted from the measured wind speeds at the former are compared to the actual measured wind speeds. Results validate the effectiveness and accuracy of the proposed method for predicting the power output of wind farms in the short term. Comparison with the Persistence Model shows that the proposed model yields superior short term wind speed predictions.

Original languageEnglish (US)
Title of host publicationIEEE PES General Meeting, PES 2010
DOIs
StatePublished - Dec 6 2010
EventIEEE PES General Meeting, PES 2010 - Minneapolis, MN, United States
Duration: Jul 25 2010Jul 29 2010

Publication series

NameIEEE PES General Meeting, PES 2010

Other

OtherIEEE PES General Meeting, PES 2010
CountryUnited States
CityMinneapolis, MN
Period7/25/107/29/10

Keywords

  • Cross correlation
  • Prediction of power output
  • Varying time lag cross correlation
  • Wind power forecasting

ASJC Scopus subject areas

  • Energy Engineering and Power Technology

Fingerprint Dive into the research topics of 'Prediction of short term power output of wind farms based on least squares method'. Together they form a unique fingerprint.

Cite this