Design and Performance Analysis of Multilayer Neural Network-based Battery Energy Storage System for Enhancing Demand Side Management

Abstract

Fossil fuel-based power plants are harmful to the environment because it releases greenhouse gases. Also, the availability of fossil fuels is going to run out in the near future. Hence, Renewable Energy Sources (RES) play a vital role in power generation to meet the required energy demand. Since RES are ubiquitous, the earth's surface is blessed with a huge amount of energy resources. The electrical energy from RES is harnessed using suitable energy conversion devices such as solar PV, wind turbine, water turbine, etc. However, RES is seasonable and not available in a unique manner in all areas. Therefore, interconnecting to the grid is a crucial task that may raise stability issues. Hence, an accurate Demand Side Management (DSM) system is required to maintain grid stability. Moreover, DSM can reduce peak energy consumption from the grid using an accurate DSM system. The proposed system has Artificial Neural Network (ANN) based battery storage system. ANN controls the battery by using certain critical parameters such as the off-peak period, peak period, and state of charge of the battery. The ANN-based management system has been developed in MATLAB Simulink. From the results obtained, it is understood that the battery management system supplies the electrical energy to the grid during peak periods, reducing the burden on the main grid and storing the electrical energy during an off-peak period. Hence, the required size of generating units is reduced; therefore, ANN-based DSM optimizes the cost.