Identification of Three Phase Transformer Abnormal Conditions Using Wavelet Entropy

Traditionally, power system signals have been analyzed by techniques based on Fourier transform and fast Fourier transform for the purposes of identifying abnormal conditions and power quality issues. Distinguishing the inrush currents and fault currents in power transformers is an essential task for protection purposes. Detecting, discriminating and severity ranking of different unbalanced conditions of power transformers may prevent damage of transformers and supplied loads. The paper presents a wavelet-based approach to analyze the inrush currents of a three-phase power transformer in order to detect and rank severity of anticipated unbalanced conditions. As will be shown by the simulated results, the paper reveals that wavelet entropy, which has been adopted in this paper, is a reliable and an efficient tool that facilitates the accurate discrimination of abnormalities in transformer currents and to investigate the unbalanced conditions.