The Spectrum Comparison Technique of DWT for Discriminating between External Fault and Internal Fault in Power Transformer

Abstract

Abstract – This paper proposes a technique for detecting and identifying internal winding fault of three-phase two-winding transformer. A spectrum component obtained from DWT of differential current is analyzed. A ratio between per unit differential current and per unit time is calculated and performed as comparison indicator in order to discriminate between internal fault condition and external fault condition. Various cases based on Thailand electricity transmission and distribution systems are studied to verify the validity of the proposed algorithm. Results show that the proposed technique has good accuracy to detect fault and to identify its position in the considered system. Keywords : Power transformer, External fault, Internal fault, Discrete wavelet transform 1. Introduction To guarantee safety and stability of power grid operating, a precise protection scheme is required. In the literature for fault detection, several decision algorithms have been developed to be employed in the protective relay [1-10]. Most of them have different solutions and techniques. An application of a finite impulse response ANN (FIRANN) as differential protection for a three-phase power transformer is proposed in [1]. In [2], the paper describes a new approach for transformer differential protection that ensures security for external faults, inrush, and over-excitation conditions and provides dependability for internal faults. A new relaying fuzzy logic algorithm to enhance the fault detection sensitivities of conventional techniques is proposed in [3]. The relaying algorithm consists of flux-differential current derivative curve, harmonic restraint, and percentage differential characteristic curve. In [4], a new algorithm based on processing differential current harmonics is proposed for digital differential protection of power transformers. This algorithm has been developed by considering different behavior of second harmonic components of the differential currents under fault and inrush current conditions. In [5], the paper describes a new approach for transformer differential protection that ensures security for external faults, inrush and over-excitation conditions and provides dependability for internal faults. In [6], a new algorithm based on processing differential current harmonics is proposed for digital differential protection of power transformers. This algorithm has been developed by considering different behaviors of second harmonic components of the differential currents under fault and inrush current conditions. In [7], a novel analysis of the currents arising during a turn-to-turn fault in transformer in which a winding is delta-connected is done, so that data acquisition pre- and post-fault conditions may lead to a correct diagnosis. The approach given in this paper is based on the analysis of current sequences which appear in the fault state, and mainly in the nature of a zero sequence current (ZSC) in a delta winding, which is thoroughly discussed. In [8], the paper addresses turn-to-turn faults in power transformer windings. A sensitive detection method for such faults is described. As a result, most research works are interested in only the effects from magnetizing inrush current and the discrimination between magnetizing inrush current and internal faults [1-8], and etc. In addition, wavelet transform has been reported in the literature [9]. The idea of application of wavelet transform to fault diagnosis is not new, and there is a number of research papers related to this idea [9-14]. The advantage of the wavelet transform is that the band of analysis can be fine adjusted so that high frequency components and low frequency components are detected precisely. Results from the wavelet transform are shown both in time domain and in frequency domain. In previous research works [12], an analysis of the spectrum of the transient current signal is performed in order to determine whether the current is a fault or a magnetizing inrush current. The approximated signal of DWT is then employed in the algorithm for a decision unit in the protection scheme.