Anomaly Zone Determination for Topology Error Processing in Power System State Estimation

During the last decade, the development of topology error identification methods in power system real-time modeling has been driven by the detailed representation of suspect regions of the network and also by the adoption of a two-stage approach to perform state estimation. Most research efforts deal with specific methodologies to be used in the first or in the second stage. This paper proposes a unified approach capable of determining the region of impact of topology errors in the first stage of state estimation, as well as to perform topology error identification carried out in the second stage. Both stages are formulated as optimization problems, whose solutions are obtained by Hachtel’s method, so that Lagrange multipliers are made available in both cases. Correlation indices previously proposed in the literature are expressed as a function of those multipliers and employed to determine an anomaly zone, which contains all modeling errors. In addition, graph-theoretical algorithms are used to extend the anomaly zone, so as to ensure connectivity and other topological properties required at the second stage. The proposed methodology is tested through several simulations conducted with the IEEE 30 bus test-system.