Smart Grid Integration of Electric Vehicle Charging Using Reinforcement Learning and Power Flow Analysis

The increased demand for electricity and the fluctuating EV charging patterns make it challenging to assimilate high electric demands in the power grid. To mitigate this problem, researchers are highly active in devising desirable chargers that charge EVs optimally without compromising system stability incentives. In this project, RL and Power Flow analysis present a new path for reinforcing the integration of electric vehicles into the power grid. RL is a machine learning trick that allows agents to learn optimal actions from interacting with their environment. In this case, the power grid and EV charging loads is an example environment. An RL agent operates under these constraints and demonstrates how to decide when and in what manner the EVs will be charged by considering grid limitations, the personal preferences of the user for whom he is submitting a confirmation, and electricity costs. The RL agent can then use this data to determine how much and when the EVs should minimize disturbance on the grid while making it cheap for vehicle owners from a look-ahead perspective by analyzing load flow, voltage levels & power losses. Furthermore, this way, EVs can communicate with the RL agent, which opens up demand response functionalities. In the end, this collaborative effort of RL and power flow analysis performs well enough to address several obstacles faced by EV charging on the grid, promising a better-suited integration of EVs into our energy system that is reliable and cost-effective.