Electronic instrument transformer is an important process-level device in the smart substation, which takes charge of collecting voltage and current information in the power system and transmitting it to the process layer network for measurement, protection, and further usage by control devices. As the digital interface between electronic instrument transformer and process layer network, merging unit (MU) plays an important role in electronic instrument transformer error adjustment. This paper proposed a method combined with self-calibration and Multivariate Exponentially Weighted Moving Average (MEWMA) control charts for evaluating the state of merging unit, and the detection ability of MEWMA control chart is analyzed. In order to predict the states of merging unit accurately during the long-term operation, an optimized states prediction method considering multiple features is proposed. In all experiments, the results agree with theoretical predictions. The analysis of data in different environments using MEWMA control charts optimized with different parameter weights shows that the optimized control chart is more sensitive to identifying abnormal states than traditional methods and can detect MU state anomalies earlier than traditional judgment methods.
Qi et al. (Mon,) studied this question.