Power System Dynamic State Estimation by Unscented Kalman Filter with Guaranteed Positive Semidefinite State Covariance

TL;DR

This paper introduces a novel unscented Kalman filter method that ensures the positive semidefiniteness of the state covariance matrix, improving power system dynamic state estimation accuracy and reliability.

Contribution

It proposes a new approach to maintain positive semidefinite covariance matrices in UKF for power systems, enhancing estimation stability and performance.

Findings

Validated on NPCC 48-machine 140-bus system

Demonstrated improved estimation accuracy

Confirmed robustness of the method

Abstract

In this paper an unscented Kalman filter with guaranteed positive semidefinite state covariance is proposed by calculating the nearest symmetric positive definite matrix in Frobenius norm and is applied to power system dynamic state estimation. The proposed method is tested on NPCC 48-machine 140-bus system and the results validate its effectiveness.