Automated network topology extraction based on graph theory for distributed microgrid protection in dynamic power systems

Abstract

Unlike conventional grids, microgrids may utilize different connections and the overall topology can be variable. Considering this, it is required to develop a new protection concept/scheme for safe operation. Maintaining proper selective operation of the relays in these dynamic microgrid structures is a challenge itself. This requires monitoring the connections and updating time delays of the relays which will ensure the desired hierarchy in the system. In this paper, a novel approach has been taken where electrical networks are modeled according to graph theory. Smart algorithms, such as network graph discovery, local manager selection, and protection coordination strategy, are run to automatically detect topology changes and ensure proper protection operation. Furthermore, distributed nature of this method mitigates the risks associated with central controller-based schemes. The developed method is applicable to all power system operations, and it poses a unique implementation in postdisaster recovery. After a disaster or terror attack, this self-diagnosis, self-healing system can identify healthy sections and run them as a standalone system until the relief arrives. The ability of the protection system to be run as a distributed control makes sure that any healthy part of the system can be restructured and utilized, without the dependency, on any central controller or connection.