Application of Parity Mismatches in Detection of Bad Data for Power System State Estimation

Abstract

Measurement residuals have been conventionally used in the detection and identification of gross errors in power system state estimation. Other types of mismatches have been employed in failure detection of complex plants, though their application to power systems have not yet been reported. Parameters in power systems can be used for the purpose of generating such mismatches called parity mismatches in this paper. Parameters calculated from measurements containing erroneous ones differ from their true values resulting in parity mismatches. In this paper, parity mismatches are employed for identification of gross errors in given measurements. A relation between parity mismatches and measurement residuals has been derived. Similar to the method of residuals, normalisation can be effected on parity mismatches. Physical appeal of the parity mismatches enables one to adopt normalisation which improves the detectability of bad data in short lines. A segregated treatment for real, reactive power flows and also the injections can be applied unlike the method of residuals.