Improvement of wind power prediction from meteorological characterization with machine learning models

To mitigate uncertainties in wind resource assessments and to improve the estimation of energy production of a wind project, this work uses a decision tree machine learning model to assess the effectiveness of hub-height wind speed, rotor-equivalent wind speed, and lapse rate as variables in power prediction. Atmospheric data is used to train regression trees and correlate the power outputs to wind profiles and meteorological characteristics to be able to predict power responses according to physical patterns. The decision tree model was trained for four vertical wind profile classifications to showcase the need for multiple calculations of wind speed at various levels of the rotor layer. Results indicate that when compared to traditional power curve methods, the decision tree combining rotor-equivalent wind speed and lapse rate improves prediction accuracy by 22% for the given data-set, while also proving to be the most effective method in power prediction for all classified vertical wind profile types. Models incorporating lapse rate into predictions performed better than those without it, showing the importance of considering atmospheric criteria in wind power prediction analyses. (c) 2021 Elsevier Ltd. All rights reserved.

Automated summary

This study utilizes a decision tree machine learning model to assess the effectiveness of variables such as wind speed, rotor-equivalent wind speed, and lapse rate in energy production prediction for wind projects. The model shows an improvement in prediction accuracy compared to traditional methods, particularly when combining rotor-equivalent wind speed and lapse rate. Including lapse rate in predictions is shown to be important for better performance in wind power prediction analyses.