A Complex Domain Gaussian Belief Propagation Method for Fully Distributed State Estimation

Abstract

To alleviate the communication, storage, and computation burden on the control center and make full use of edge computing resources, fully distributed state estimation has received increasing interest recently. This paper intends to improve the efficiency and robustness of the fully distributed state estimation by introducing a meter-level method based on the Gaussian belief propagation theory. Specifically, we propose a complex domain factor graph, which extends the state variable vector from voltage phasors to multiple electrical quantities, including voltage phasors, current phasors, voltage magnitudes, and active/reactive power, enabling the direct processing of nonlinear measurement models and significantly reducing the number of iterations. Furthermore, based on the M-estimation theory, we innovatively incorporate multiple robust functions to the Gaussian belief propagation method to enhance the robustness of the proposed fully distributed estimator. The effectiveness of the proposed method is demonstrated under various operation conditions.