A Parallel Framework for Robust State Estimation Using Node-Breaker Substation Models

This paper presents a parallel implementation of a least absolute value (LAV) based linear state estimation (SE) algorithm using node-breaker (NB) models. The proposed method automatically corrects topology errors during the SE process, therefore, does not need any pre- or post-SE processing. Although leverage measurements can be problematic for LAV estimators, NB modeling is shown to be less susceptible to leverage measurements than bus-branch (BB) modeling. To deal with the increase in computational complexity of NB models, a zone-partitioning algorithm combined with parallel processing is utilized. The IEEE 300-bus system is used to demonstrate both the topology correction capability and the computational efficiency of the proposed method.