부산대학교 도서관

The phasor measurement unit (PMU) is widely employed in modern power systems, which enables real-time power system status monitor and risk identification. However, the bad data caused by PMU failure may mislead the risk identification. This paper proposes an efficient power system risk identification method considering PMU bad data. First, this paper describes the risk identification problem based on the analysis of the characteristics of PMU bad data and the power system contingency events. Then, the Teager-Kaiser energy operator with the 3Sigma criterion is proposed for the initial PMU anomaly data screening. Next, based on the difference between the bad data and the contingency events, the Pearson correlation coefficient is adopted for the calculation of the spatial similarity. Finally, a local outlier probability algorithm is presented to analyze the metrics. Case studies are carried out on IEEE 39 and IEEE 118 systems. Results demonstrate that the proposed method has an efficient risk identification accuracy with PMU abnormal data.