Enhanced Bad Data Processing for Hybrid Block-Orthogonal State Estimators

This paper introduces a new strategy for bad data processing in connection with hybrid block-orthogonal state estimators in order to cope with the significant performance degradation exhibited by the conventional $J\left( {\hat x} \right)$-test when applied to large networks. The proposed method advocates the partitioning of the phasor measurement set and the underlying network into disjoint clusters. This allows that independent $J\left( {\hat x} \right)$-tests be performed on each network cluster, what improves the sensitivity of the detection test to bad data occurring in the phasor measurement set. A number of case studies conducted on the IEEE 300-Bus benchmark system are employed to evaluate the effectiveness of the enhanced strategy.