Refinement of Wind Turbine Power Prognostication via Ensemble Learning: Synergizing Long Short-Term Memory and Random Forest Models

Wind energy is now a pivotal part of renewable energy systems, where precise forecasting of wind turbine power output is crucial for effective wind farm management. This paper proposes a new method that integrates Support Vector Regression (SVR) and Multi-Layer Perceptron (MLP), along with an ensemble model merging Long Short-Term Memory (LSTM) and Random Forest. Meteorological data, such as wind speed, direction, temperature, and failure time, collected from wind farms, form the dataset. These parameters are used as input features for machine learning models, while the output represents the wind turbine’s power output. Our findings demonstrate that the ensemble approach, combining LSTM and Random Forest, surpasses individual SVR and MLP models. This technique advances predictive maintenance strategies and optimizes renewable energy systems, promoting the transition to a sustainable energy future.