Identification and classification of fault during power swing using decision tree approach

Abstract

This paper suggests identification and classification of fault at the time of power swing by using decision tree approach for transmission system network. The power swing which occurs during switching in/out of heavy loads, after fault clearance condition etc causes change in active and reactive power. The suggested decision tree method initially retrieves the information about fault/ power swing condition and if it is a fault then it identifies the fault. Different power swing and fault cases are evaluated by simulating 400 kV transmission system. Fundamental values of voltages, currents and zero sequence component of currents of double circuit transmission line are taken as the input features to the decision tree. To validate the suggested method, decision-tree-based scheme is tested and trained for change in fault type, fault resistance, fault distance.