Hybrid Solar Energy Forecasting with Supervised Deep Learning in IoT Environment

Microgrids, which integrate decentralized “renewable energy sources (RESs)”, energy storage devices, and load management methodologies, have arisen lately as a building ingredient for smart grids. The intermittent nature of renewable energy sources offers many barriers to smart microgrids, including "reliability, voltage quality, and supply-demand balance". Thus, predicting power output from renewable energy sources "(such as wind turbines and solar panels)" has become more vital for the efficient and continuous functioning of the electrical grid, as well as for attaining optimum RES utilization. Smart microgrids also incorporate energy demand forecasting, which assists in the planning of power production and energy trading with the commercial grid. Models based on “MLand DL” are promising ways of forecasting consumer needs and energy production from “renewable energy sources”. In this context, this research article presents a full review of the available DL-based algorithms for “wind turbine and solar panel power forecasting, as well as electric power demand forecasting”. It also goes through the datasets that were used to train and test the different DL-based prediction models, enabling future researchers to pick the correct datasets for their projects. A secondary way of data collecting is considered for this work to obtain information from various papers and journals.