A Smart Grid-oriented Energy Optimization System Using Renewable Energy Sources and Machine Learning

Abstract

This study establishes a framework for implementing active demand-side control in houses inside smart grids that
incorporate distributed Renewable Energy Sources (RES) production and energy storage systems. This methodology
results in a decision-making mechanism that regulates the battery to minimize consumer energy expenses. It delays
the development of the electrical grid when the peak loading period corresponds with the elevated daily electricity
rate. The decision-making platform is a certified Neural Network (NN) trained with optimum data applicable in
households fulfilling particular parameters, including location, power tariff, and consumption profiles aligned with
standards certified by the nearby electricity company. Three consumption characteristics and energy production
profiles were constructed and integrated to verify this process. The findings indicate that the Artificial NN (ANN)
based decision-making mechanism manages the battery effectively to minimize the cost of electricity.