A multi-agent based approach to power system dynamic state estimation by considering algebraic and dynamic state variables

In this paper an agent-based modeling for the power system dynamic state estimation is proposed that is able to take advantages of hybrid measurement data. Multiple execution tasks are distributed among interacting agents which each agent is supposed to carry out a specific computation or functionality. The algebraic state variables of power system and the dynamic state variables of synchronous generators are considered in the proposed method. Artificial neural network is applied for deriving a parameterized process model of the algebraic state variables. The process model of the dynamic state variables is based on the fourth-order dynamic model of the synchronous generator. The dynamic state estimation problem is solved by using unscented Kalman filters. The effectiveness of the proposed method is confirmed through simulations while different scenarios are considered. The results are compared with some widely used approaches to power system dynamic state estimation. Further, since the proposed approach is benefited from agent based modeling, it is less time-consuming and can be implemented through modular configuration which is more desirable from software and hardware engineering points of view.