Dynamic State Estimation for Power Networks by Distributed Unscented Information Filter

摘要

This paper presents a distributed unscented information filtering (UIF) method for state estimation of interconnected nonlinear dynamic systems. The UIF method is an information filter based on unscented transformation (UT), which is a nonlinear estimation method. Therefore, the estimation accuracy of UT-based distributed UIF method should be higher than that of linearization-based distributed maximum a posteriori (MAP) method. When implementing the distributed UIF algorithm, we first calculate the local estimate by UIF method based on the local observations, and then gradually integrate the neighboring information by iterative method to obtain a more accurate distributed estimate. The IEEE 118-bus system is adopted to conduct a series of simulations to evaluate the performance of the proposed distributed UIF method, the centralized UIF estimator, local UIF estimator and the distributed MAP estimator. Simulation results show that the proposed distributed UIF approach achieves a worse estimation accuracy than the centralized UIF, but is better than both the distributed MAP estimator and the local UIF estimator.