PI: Guru Parulkar; Co-PIs: Ram Rajagopal, Arun Majumdar, Charles Kolstad, Nicholas Bambos; Students: Thomas Navidi, Emmanuel Balogun
Future EV charging demands, particularly fast charging, will require coordination between EV chargers, power networks, and vehicles to manage reliability of the grid. Such coordination requires information exchange across multiple devices and stakeholders in the system.
Purely centralized coordination would require data and objectives from all stakeholders in the system to be known by the coordinator. This is not practical as it does not satisfy the autonomy of each stakeholder in the system. Alternatively, traditional decentralized coordination approaches would require each charging facility—or even individual chargers—to be aware of the physical power network topology, transformer, and communication
network constraints preventing the scaling of the solution. Instead, we propose designing a virtual overlay thatdecouples the utility power network and edge devices.
The overlay will enable edge devices to make their own local decisions while incorporating simple local constraints that ensure the voltage and transformer reliability of the network. Furthermore, virtual local storage manages unexpected local fluctuations. The network overlay also enables the information sharing necessary to design a variety of critical applications for the EV ecosystem, including green route planning of vehicles, dynamic pricing for the distribution system, and forecasting power and traffic.