Hybrid of Persistence and Neural Network Algorithm Approach for Load and Renewable Energy Forecasting

Hybrid of Persistence and Neural Network Algorithm Approach for Load and Renewable Energy Forecasting

Abstract: Microgrids, which combine distributed renewable energy sources (RESs), energy storage systems, & load management methodologies, have lately risen to prominence as a building element for smart grids. Because renewable energy sources are inconsistent, smart microgrids face various issues, including reliability, power quality, and supply-demand balance. As a result, anticipating energy production from renewable energy sources (such as wind turbines and solar panels) is becoming increasingly important for the power grid's efficient and continuous operation, as well as for maximizing RES use. Smart microgrids include energy demand forecasting, which aids in energy production planning & energy exchange with the commercial grid. In this article, an ANN with a distance technique is used to estimate load power while accounting for the penetration of renewable energy sources, and the average variation produced is analyzed. Eventually, a comprehensive Microgrid design in MATLAB/Simulink is shown, including the power sources, their power electronics, and a load and mains simulation. The presented method provides correct estimates, is more economical, and has a good generalization capability, as evidenced by the experimental analysis.

Keywords: Renewable energy sources, Microgrid, persistence, Artificial Neural Network