A Study of Logistic Regression-Based Discrimination Method of False Overcurrent Alarm of 500kV High-voltage Shunt Reactor

Abstract

Based on the analysis of influence of transformer measuring error on accuracy of overcurrent alarm of 500 kV high-voltage shunt reactor, a logistic regression-based discrimination method of false overcurrent alarm of high-voltage shunt reactor is proposed. This method conducts Logistic regression by analyzing the voltage of the busbar where high-voltage shunt reactor is located as well as the alarm signal data. For low accuracy of instrument reactor, two false alarm criteria are defined according to regression coefficients. High-voltage shunt reactor on one actual 500kV transmission line is taken as the case study, measured data and simulation data are used for verification. Results of the case study show that without any need of blackout experiment or any additional measuring equipment, this method could rapidly recognize accuracy of overcurrent alarm of high-voltage shunt reactor with strong industrial application value.