Enhanced Grid Stability and Demand-Side Optimization through Deep Neural Network-Controlled Vehicle-to-Grid (V2G) Peak Shaving and Load Shifting

Modern power networks face both opportunities and problems from the quick adoption of electric vehicles (EVs) and the growing use of non-conventional resources. To keep the grid stable, controlling peak demands and making sure energy is distributed efficiently provide a significant challenge. By enabling bidirectional power transfer between EVs and the grid, vehicleto-grid (V2G) technology provides a workable option. As a result, EVs become mobile energy storage devices that may optimize energy consumption and lessen grid stress by recharging during off-peak hours (load shifting) and discharging electricity during peak demand (peak shaving).This study suggests a Deep Neural Network (DNN)-based Demand Side Management (DSM) approach for a grid-connected V2G energy storage system. By training the DNN to forecast short-term power use and user behaviour, EV charging and discharging cycles may be controlled in real time. Through advanced V2G operations, the model ensures optimal energy exchange by considering criteria including EV availability, battery State-of-Charge (SOC), grid load patterns, and power price. MATLAB/Simulink simulation results with various residential and business load profiles over a 24-hour period show how successful the suggested approach is. Peak grid power peaked at 166.5 kW without DNN management, however peak shaving based on DNN lowered this to 100 kW. The demand was further spread using load shifting, which produced a smoother load curve. The suggested DNN-based DSM strategy is a viable option for next-generation smart grids as it greatly improves grid stability, lowers operating costs, and makes it easier to integrate renewable energy.