Koopman Kalman Particle Filter for Dynamic State Estimation of Distribution System

Abstract

Dynamic state estimation (DSE) plays an important role in the real-time control and monitoring of distribution systems, which are high-dimensional space–time systems. The degree of nonlinearity of distribution system increases drastically with the availability of sustainable energy and the diversification of load types, thereby reducing the accuracy of the standard linear model. Under the existing schemes, the measurement noise may not follow a Gaussian distribution. Therefore, a Koopman operator-based Kalman particle filter (KKPF) is proposed herein for estimating the dynamic states of a distribution system. The KKPF performs data-driven dynamic state estimation based on the Koopman operator theory, which does not rely on the distribution system model and can be applied to high-dimensional systems. Furthermore, this method can be applied to dynamic systems and noise with both Gaussian and non-Gaussian distributions. The KKPF method provides accurate estimation results when only a small number of measurements are available. IEEE 141 system and an improved 141 system were used to test the performance of the proposed KKPF compared to the EKF, CKF, and PF. The test revealed that the proposed KKPF can obtain accurate results under high-dimensional and non-Gaussian noise environments.