Dynamic State Estimation for Improving Observation and Resiliency of Interconnected Power Systems

This article investigates a particle filter (PF)-based fully-decentralised dynamic state estimation (DSE) method for interconnected multi-machine power systems. The PF-based observer is developed to dynamically estimate the states of the 7th-order dynamic model of synchronous machines that are either inaccessible and/or highly noisy. It is assumed that the proposed PF-based robust decentralized observer for a particular synchronous generating unit relies on typical output measurements available from phasor measurement units (PMUs) installed at its terminal. The performance of the presented observer is investigated using the benchmark model of the IEEE 68-bus system considering a detailed sub-transient representative model of synchronous machines with different excitation and control systems. The presented estimation framework works successfully and accurately under various transient events, such as load perturbation, faults, and changes in network topology, while accounting for different errors and sampling rates in measurements. The accuracy and robustness of the presented dynamic estimator in the case of Gaussian and non-Gaussian noisy measurements are verified. The paper also develops an approach-based PF to detect bad data and introduces a new metric based on the computation Cramér–Rao Low bound (CRLB) for evaluating the dynamic estimation performance. The introduced PF-based DSE improves the system resiliency by providing the system operator with the monitoring and observation capability of the system in a real-time manner to perform the proper corrective and protective actions in case of any events. The comparative study with other sophisticated dynamic state estimators confirms the brilliance, robustness, and superiority of the presented PF-based dynamic state estimation for multi-machine systems, and its practical and implementation feasibility.