A detailed tuning evaluation of sensitivity threshold using LNRT for bad data detection in state estimation

Abstract

The sensitivity threshold, using largest normalized residual test, was often chosen based on past experience without proper technical understanding. Hence, the aim of this paper was to systematically investigate the variation of the sensitivity threshold when bad data of various magnitudes were individually injected into each measurement. Subsequently, the relationship between bad data and sensitivity threshold could be observed. Moreover, the relationship of sensitivity threshold with the redundancy of measurements were also analyzed. Two test cases were built using IEEE 14 bus test system. Simulations results demonstrated that minimum sensitivity threshold should not be fixed as is the case currently.