Offshore Wind Power Forecasting—A New Hyperparameter Optimisation Algorithm for Deep Learning Models

The main obstacle against the penetration of wind power into grid is its high variability in terms speed fluctuations. Accurate forecasting, while making maintenance more efficient, leads to profit maximisation traders, whether for a turbine or farm. Machine learning (ML) models are recognised as an accurate and fast method prediction, but their accuracy depends on selection correct hyperparameters. incorrect choice hyperparameters will make it impossible extract maximum performance ML models, which attributed weakness forecasting models. This paper uses novel optimisation algorithm tune long short-term memory (LSTM) model forecasting. proposed improves prediction accelerates process. Historical data offshore Scotland utilised validate compare outcome with regular tuned by search. results revealed significant effect models’ performance, improvements RMSE 7.89, 5.9, 2.65 percent, compared persistence conventional search-tuned Auto-Regressive Integrated Moving Average (ARIMA) LSTM