Classification of Faults in Power Transformer using Fuzzy Logic and Wavelet Transforms

Abstract

With the day to day increase of the power and the increasing rate of industrialization. The amount of power to be developed and the safety of the power transformers have increased manifolds. For optimum results it is required to have nearly a no – fault operation of power transformer. The objective of this thesis is to design a controller or method which could predict the unwanted outages in a small interval of time and with accuracy. The controllers like PI, PID are in use in industries over a decade and found to be very useful also. But in this case these controllers are not found to give reliable results due to the oscillating and non – periodic nature of the power transients. Hence, fuzzy controller being an intelligent controller could be used for this purpose of detection of inrush and fault current. The wavelet transform with its ability to determine information from transient signals both in frequency and time domain is also used for detailed analysis of various power transformer transients. Further it is found that wavelet transform is a very effective tool for detailed analysis of these transients and also with fuzzy logic we have obtained accurate and very useful results.