Hermite transform-based algorithm to discriminate magnetizing currents in transformers

Abstract

This work proposes a new method based on the Hermite transform with the aim of identifying transient events in power transformers caused by energizations and current transformer (CT) saturation. The Hermite transform is a signal processing technique that allows the representation of signals in the time–frequency domain as well as at multiple resolution levels. Gaussian derivatives analyze the signals in the proposed scheme according to the Hermite transform theory. The newly introduced approach is based on a differential scheme considering three functions, the first one to analyze the positive half-cycle and the second one for the negative half-cycle. The latter is a new blocking function for the detection of transient conditions, that is, the operation command is disabled; otherwise, the third function will confirm if the condition is an internal fault. To validate the performance of the proposed method, several fault conditions are addressed, such as load switching, internal and external faults, current transformer saturation, and single-phase-to-ground faults. Results show the effectiveness of the proposed method, which improves the differential protection performance to discriminate transient conditions from internal faults.