Agent-Based Distributed Computing for Power System State Estimation

Centralized grid operations are now facing data storage and computation complexities more than ever. These are further compounded by communication saturation and related issues. These have instigated the transition towards distributed computing paradigm. This paper proposes a novel agent-based distributed (node-node) state estimation that overcomes the complexities of centralized operations. The proposed methodology sketches distributed constraint satisfaction algorithm that enables an agent to estimate its local representative bus's state only by local computation and asynchronous message exchange with connected neighboring agents. The case studies demonstrate the performance, efficacy and solution accuracy of the proposed methodology by comparing it with centralized and distributed algorithms.