Current Transformer Saturation Detection Using Gaussian Mixture Models 基于高斯混合模型的电流互感器饱和检测.pdfVIP

Current Transformer Saturation Detection Using Gaussian Mixture Models * M. Moghimi Haji, B. Vahidi , S. H. Hosseinian Department of Electrical Engineering Amirkabir University of Technology Tehran, Iran *vahidi@aut.ac.ir ABSTRACT This paper presents a novel current transformer (CT) saturation detection approach based on Gaussian Mixture Models (GMMs). High accuracy is the advantage of this method. GMMs are trained with secondary current of CT. The appropriate performance of the proposed method is tested by simulation of different fault conditions in PSCAD/EMTDC software. The results show that the trained GMMs can successfully detect CT saturation with high accuracy. Keywords: CT saturation, GMM, protective relaying, transient analysis. 1. Introduction Current transformer (CT) saturation distorts the is a perfect sinusoid, the summation of the current secondary current and in consequence leads to and its second-order derivative is zero over time. operating delay or malfunction of protection relays In [5] an impedance-based CT saturation detection (e. g. under reaching of over current relays, algorithm for bus-bar differential protection is overestimation of fault loop impedance in distance suggested. Calculation of power system source relays) [1]. Therefore an appropriate saturation impedance at the relay position is based on a first- detection method is necessary to maintain the order differential equation. This method uses protection system reliability. voltage and current signals of the bus-bar to detect CT saturati