Faults in Transformer Winding and Probing their Occurence by S-Transform and RBF Neural  Network Technique

Tripti Sahu; Harshwardhan Bhatia

CALL FOR PAPERS : Dec-2023

Submission Last Date

25-Dec-23

Submit Manuscript Online

FOR AUTHORS

FOR REVIEWERS

ARCHIEVES

DOWNLOADS

Open Access

Faults in Transformer Winding and Probing their Occurence by S-Transform and RBF Neural Network Technique

Author(s):

Tripti Sahu , Indian School of Mines Dhanbad Jharkhand ; Harshwardhan Bhatia , Indian School of Mines Dhanbad Jharkhand

Keywords:

EMTP, S-transformation, TRF Time frequency response, Stock well Coefficient, RBF, transient model of transformer

Abstract:

Now a days significant enhancement of certainty, reliability and economics of the power transformer is an important topic in power system engineering because among all equipment as for expedition used in electrical systems power transformer is one of the most costly piece of equipment. So that accurate diagnosis of fault has been always been a major issue. Simulation studies based on transient model of transformer winding enhance understanding of surge voltage and fault current waveforms. Those are non-stationary in nature. The non -stationary nature of current waveforms inside the transformer winding requires the model parameter to be both time and frequency dependent. it will help to detect type of fault as well as location of occurrence using simulation studies. This paper aims to present a technique for diagnosis of the type of the fault and location of occurrence. . The proposed method is based on a transient model based transformer under impulse test and S-Transform with good time frequency resolution for all frequencies (multi resolution). S-Transform is used to produce a stock well coefficient vector which utilized as inputs testing set to a Radial Basis Function Neural Network (RBF). The fault conditions are simulated by the variation of fault location according to inter turn winding of power transformer or disc. The training is conducted by programming in MATLAB. The robustness of the proposed scheme is investigated by synthetically polluting the simulated voltage signals with White Gaussian Noise. The suggested method has produced fast and accurate results. Estimation of fault location is intended to be conducted in future.