Traditionally, the electric power system is operated in a centralized way, i.e. decisions on how to dispatch generation are made at central control center. In this project, a distributed approach based on the consensus algorithm to make generation and load dispatch decisions is investigated.
The goal of the project is to implement a completely distributed decision making process in which each participant (generator or load) only shares very limited information with its neighbors. This can be thought of as having flexible consumers for electric power. The project will also extend the existing distributed approach which is currently focused only on generator dispatch to include load dispatch and non-dispatchable generation resources. Further investigations will be made into the study of convergent behavior for larger scale networks and for cases in which participants maliciously share incorrect or inaccurate data with neighbors. The aim is to generate a robust implementation of a distributed approach to dispatch generation that can be feasible with imperfect knowledge in the real world.