Analysis of Bad Data Detection in Power System State Estimators Considering PMUs

This paper investigates the bad data detection issues in power system state estimation considering SCADA and PMU measurements. The traditional State Estimator (SE) containing only SCADA measurements, two One Phase SEs and one Two Phase SE are evaluated. The SEs are assessed by considering measurements containing gross and gaussian errors and two classical methods for measurement standard deviations specification. The bad data detection is realized with the classical Largest Normalized Residual approach. The main found is that the One Phase SEs are, in general, more accurate and, therefore, the bad data detection approach works better with these SEs. Besides, the performance of bad data detection approach is strongly deteriorated, even for SCADA measurements, when classical standard deviation methods are adopted with the Two Phase SE. In general, considering all evaluated SEs and the classical standard deviation methods the bad data detection presented inadequate performance for PMU measurements. Therefore, it can be concluded that some improvements are required in this topic. Test results on the IEEE-14 bus system are presented.