Machine Learning Implementation on Wind Speed Prediction

İlhan A.

1st International Conference on Frontiers in Academic Research, Konya, Turkey, 18 - 21 February 2023, pp.205-210, (Full Text)

  • Publication Type: Conference Paper / Full Text
  • City: Konya
  • Country: Turkey
  • Page Numbers: pp.205-210
  • Ankara Yıldırım Beyazıt University Affiliated: Yes

Abstract

In this study, the measured hub-height wind speed of a wind turbine is forecasted using machine learning. Accordingly, the wind speed data has been obtained from an installed wind turbine of a wind power plant that is located in the Republic of Kosovo. A cumulative of 2,000 wind speed data has been used in historical time-series predictions performed by long-short term memory, adaptive neuro fuzzy inference system with fuzzy c-means (FCM), subtractive clustering (SC), and grid partitioning (GP). The results of 102 computed models have indicated that the best wind speed predictions have been obtained during the utilization of the ANFIS-SC algorithm. The accuracy of the predictions has been evaluated considering the mean absolute error (MAE), root mean square error (RMSE), as well as the correlation coefficient (R). In this context, it was determined that the SC tool of the ANFIS resulted the wind speed predictions with the superior statistical error outcomes corresponding to 0.2562 MAE, 0.3047 RMSE, and 0.9990 R.