Previsão de geração fotovoltaica com redes long short-term memory: uma revisão da literatura

Abstract

The forecasting of photovoltaic generation is essential for optimizing the operation of electrical systems and ensuring efficient energy management. In recent years, machine learning and neural network-based methods have been widely used to improve the accuracy

of such predictions, which is crucial for systems with renewable sources, such as micro- grids. This paper provides a literature review on the techniques applied to photovoltaic

generation forecasting, with an emphasis on approaches using Long Short-Term Memory

(LSTM) networks. Additionally, the main challenges and current trends in the field are dis- cussed, aiming to provide a solid theoretical foundation for future research and applications.