A MEMD-MMFE-based discrimination technique to identify magnetizing current of interconnecting transformer of wind farm

Abstract

This paper studies the time-frequency characteristics of short-circuit currents in a wind farm during the low-voltage ride-through (LVRT) process. The short-circuit current contains low-order harmonics that may affect second harmonic restraint component of transformer differential protection, while the second harmonic restraint is a classical way to block the differential relay during inrush events for the transformer. A maloperation case of the differential protection employed by the interconnecting transformer of wind farm under the LVRT process is investigated in this paper. It indicates that the second harmonic restraint cannot identify the inrush events reliably of the transformer in a wind farm. A visualization algorithm is proposed to identify the magnetizing inrush current of interconnecting transformer in wind farm based on multivariate empirical mode decomposition and multivariate multiscale fuzzy entropy (MMFE). The current is firstly decomposed into several aligned intrinsic mode functions for the signal preprocessing to reduce the effects of low-order harmonics from the short-circuit current of wind farm. Then the MMFE method is used to calculate the entropy area of the preprocessed current signal. By comparing the entropy area of the transformer current with a given restraint area, the inrush current can be discriminated. The simulation results of a practical wind farm verify the accuracy and effectiveness of the proposed algorithm.