Derivative-based Dynamic State Estimation of Synchronous Generator using Extended Kalman Filter

Information about the states of a synchronous generator plays a crucial role in monitoring, controlling, and fault detection and changes in the power system. However, due to technical difficulties, this information is not always readily available from direct measurements. This paper proposes a method for dynamic state estimation (DSE) using an Extended Kalman Filter (EKF) approach based on linearization. The method aims to estimate generator states using data measured at the generator terminal. A comprehensive 7
^th
-order sub-transient model is employed to thoroughly depict the behavior of synchronous generators in different system conditions, providing detailed insights into how they react to faults or changes within the system. The accuracy of estimated states is measured by comparing them with actual synchronous generator states. The resilience of EKF is comprehensively evaluated within a Single Machine Infinite Bus (SMIB) system by considering various scenarios of changes including short-circuit faults, and different levels of process and measurement noises. The achieved results demonstrate the proposed EKF approach’s ability to deliver precise state estimations for the sub-transient synchronous generator model, relying solely on terminal measurements. The achieved Mean Squared Error (MSE) values range from a very low minimum of $3.80\times 10^{-6}$ to a moderately low maximum of $3.06\times 10^{-2}$, further confirming the EKF’s estimation accuracy.