Deep learning based techniques to enhance the performance of microgrids: A review

Abstract

In the last few years, carbon emissions and energy demand have increased dramatically around the globe due to a surge in population and energy-consuming devices. The integration of renewable energy resources (RERs) in a power supply system provides an efficient solution in terms of low energy cost with lower carbon emissions. However, renewable sources like solar panels have irregular nature of power generation because of their dependence on weather conditions, such as solar radiation, humidity, and temperature. Therefore, to tackle this intermittent nature of solar energy, power prediction is necessary for efficient energy management. Deep learning and machine learning-based methods have frequently been implemented for energy forecasting in the literature. The current work summarizes the state-of-theart deep learning-based methods that are proposed to forecast the solar power for proper energy management. We also explain the methodologies of solar energy forecasting along with their outcomes. At the end, future challenges and opportunities are uncovered in the application of deep and machine learning in this area.