Detection and Classification of Internal Faults in Power Transformers using Tree-based Classifiers

TL;DR

This paper presents a decision tree-based method for detecting and classifying internal faults in power transformers, achieving high accuracy through feature extraction from differential currents and comparison of multiple classifiers.

Contribution

It introduces a novel approach combining feature extraction from differential currents with decision tree classifiers for accurate fault detection and classification in power transformers.

Findings

Decision Tree detects faults with 100% accuracy.

Gradient Boost classifier performs best among tested classifiers.

Features from time and frequency domains effectively distinguish fault types.

Abstract

This paper proposes a Decision Tree (DT) based detection and classification of internal faults in a power transformer. The faults are simulated in Power System Computer Aided Design (PSCAD)/ Electromagnetic Transients including DC (EMTDC) by varying the fault resistance, fault inception angle, and percentage of winding under fault. A series of features are extracted from the differential currents in phases a, b, and c belonging to the time, and frequency domains. Out of these, three features are selected to distinguish the internal faults from the magnetizing inrush and another three to classify faults in the primary and secondary of the transformer. DT, Random Forest (RF), and Gradient Boost (GB) classifiers are used to determine the fault types. The results show that DT detects faults with 100\% accuracy and the GB classifier performed the best among the three classifiers while classifying the internal faults.