A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Exploring Machine Learning Techniques for Short-Term Wind Power Forecasting of Multiple Wind Parks
Wind power is a renewable energy source that is growing rapidly globally. A disadvantage of wind power is its uncertainty as the energy output can vary greatly, a consequence of the stochastic nature of the wind. Accurate forecasts of wind power production can mitigate the problem of uncertainty. Methods within Machine Learning can be used to perform such forecasting. The aim of this study was to investigate the performance of several Machine Learning algorithms for short-term wind power prediction using an open data set. The study also investigated which input feature had the biggest impact on the output power. The explored algorithms were Autoregression, Linear Regression (LR) with the Least Absolute Shrinkage and Selection Operator (LASSO) method, Random Forrest Regression and Gradient Boosting. The study concluded that the Gradient Boosting model demonstrated the best performance among the investigated algorithms and that wind speed, wind direction and hour of the day were the most important input features.
Exploring Machine Learning Techniques for Short-Term Wind Power Forecasting of Multiple Wind Parks
Wind power is a renewable energy source that is growing rapidly globally. A disadvantage of wind power is its uncertainty as the energy output can vary greatly, a consequence of the stochastic nature of the wind. Accurate forecasts of wind power production can mitigate the problem of uncertainty. Methods within Machine Learning can be used to perform such forecasting. The aim of this study was to investigate the performance of several Machine Learning algorithms for short-term wind power prediction using an open data set. The study also investigated which input feature had the biggest impact on the output power. The explored algorithms were Autoregression, Linear Regression (LR) with the Least Absolute Shrinkage and Selection Operator (LASSO) method, Random Forrest Regression and Gradient Boosting. The study concluded that the Gradient Boosting model demonstrated the best performance among the investigated algorithms and that wind speed, wind direction and hour of the day were the most important input features.
Exploring Machine Learning Techniques for Short-Term Wind Power Forecasting of Multiple Wind Parks
Tubulekas, Alexis (author)
2022-01-01
Theses
Electronic Resource
English
DDC:
690
Study on Medium-Term and Short-Term Wind Power Forecasting Methods
| British Library Conference Proceedings | 2013
Wind Power Short-Term Forecasting Method Based on LSTM and Multiple Error Correction
| DOAJ | 2023
| DOAJ | 2021
Adaptabilities of three mainstream short-term wind power forecasting methods
| American Institute of Physics | 2015