Application of k-Nearest Neighbor (kNN) Machine Algorithm for Transformer Fault Classification

Abstract

Power transformer forms a very important link in the power system. A fault in transformer can cause a huge loss to the utility and consumer depending on the duration of the outage. Dissolved Gas Analysis (DGA) acts as a key tool to diagnose transformer fault based on gas ratios. In this paper an effort to predict Power transformer fault more precisely using kNN algorithm has been made.  DGA data of various transformer oil samples were collected and analyzed to select the best kNN algorithm to be used and to observe the prediction accuracy.