Modeling and Characterization of Transformers Internal Faults Using Finite Element and Discrete Wavelet Transforms

This paper investigates the behavior of three phase transformers with internal faults under sinusoidal and nonsinusoidal operating conditions. The terminal behavior of the transformer was investigated by coupling the finite element transformer's physical model with external electric circuit equations. Two types of internal faults were modeled turn-to-ground faults, and turn-to-turn faults with and without arcing. A discrete wavelet transforms (DWT) procedure was then used to extract the feature of the transformer fault currents. The key advantages of the DWT are its ability to provide a local representation of the current signal for normal and faulty modes, as well as its applicability to handle nonstationary fault signals. This characterization gives us an efficient way to design and evaluate detection algorithms for transformers internal faults and the system they operate in

I. Introduction

Early detection and diagnosis of transformer internal faults are desirable for condition assessment, maintenance schedule, and improved operational efficiency of the transformer. Transformer fault diagnosis requires accurate modeling under fault conditions, particularly internal faults. Nearly 70% of transformer failures are caused by undetected turn-to-turn faults. These faults usually begin as undetected turn-to-turn faults in a coil, which progress to phase-to-phase or phase-to-ground short circuit faults. Because of costly repair, down time, and safety consideration, early detection of transformer internal faults is highly desirable.