Intelligent Systems for Renewable Energy Forecasting and Management

Abstract

In the current study, a critical and analytical review of existing models of artificial intelligence for renewable energy utilization is provided, with specific reference to solar and wind energy forecasting. Unlike other studies that are directed toward developing new models of renewable energy forecasting and experimentally proving them, the current study is directed toward critically evaluating existing models of machine learning and deep learning that are applied for renewable energy forecasting. A range of models, including traditional statistical models, supervised learning models, and advanced models of deep learning like LSTM, CNN, hybrid models of CNN-LSTM, and Transformer models, are considered for review. By critically synthesizing existing literature, the study identifies the advantages and disadvantages of existing models of renewable energy forecasting with respect to forecasting accuracy, generalization, data dependency, computational complexity, and reproducibility. Additionally, it is shown how differences in data, forecasting horizons, and evaluation metrics impact the comparison of models of renewable energy forecasting. By summarizing recent developments in methodological approaches, the current review is directed toward providing structured insights that are helpful for future developments in intelligent renewable energy forecasting models.