A Hybrid Observability Analysis Method for Power System State Estimation

Abstract

For observability analysis algorithm of large-scale power system state estimation, the traditional topological algorithm is complicated and may experience combinatorial explosion problems. Meanwhile, the computation speed and numerical stability of the numerical algorithm are greatly affected by the scale of power systems. In an effort to solve these problems, this study proposed a topological-numerical hybrid observability analysis algorithm using SCADA measurements. In this paper, a basic conclusion that there is at most one unknown complex voltage of an observable island is presented; the theoretical basis and the concept of recursive combination of observable islands are described. Then, three combinability rules are proposed and recursively applied to directly judge whether the local observable islands can be combined or not. Moreover, a minimum number of equivalent branches with power flow pseudo-measurements are added inside the remaining islands to form a final simplified connected network. Finally, the numerical observability analysis method based on the decoupled DC model is carried out to obtain the final observability analysis results. Tests show that the proposed hybrid algorithm has obvious advantages in regard to the execution speed for large-scale systems and also avoids the numerical instability in large-scale power grids.