An abnormal data identification algorithm based on mutual checking between PMU and SCADA

Abstract

With the development of the smart grid, the data center generates a large amount of abnormal data while obtaining massive multi-source data. Based on the analysis of PMU and SCADA measurement data, this paper proposes an abnormal data identification algorithm based on measurement data checked reciprocally. According to the difference between PMU and SCADA, the probability density function based on the t-distribution is obtained, and the confidence interval of the mean of the two sources data is obtained with 95% confidence. Nodes outside this range are regarded as abnormal nodes. In order to improve the robustness of the algorithm in forecasting samples containing abnormal data, the p-MCC-ELM algorithm is proposed. Forecasting the PMU time series through p-MCC ELM, determine whether the source of the abnormal data nodes is PMU based on the error comparison, and finally determine the abnormal data location. Based on actual PMU and SCADA measurement data, the method proposed in this paper is verified, and the calculation results show that the proposed method can accurately identify the location and source of abnormal data nodes.