Transmission Line Fault Pattern Recognition using Decision Tree based Smart Fault Classifier in a Large Power Network

With the advent of smart grid, the power system operation, monitoring and controls are becoming more intelligent and computer assisted. In-line with the basic objective and viewpoint of the smart power grids, transmission line fault pattern recognition and fault clearance must be done out more intelligently, judiciously and automatically, with less operator intrusion. The present paper proposes an organized and smart fault classifier with a combination of reliable preprocessing technique for key attribute selection from the power system recorded waveforms and employs a dependable decision tree based classification algorithm to acquire fault detection precision, even when the power network under consideration is large. PSCAD/EMTDC has been used for conducting a case study on simulated IEEE 30-bus system. Maximum observability of the IEEE 30-bus network was achieved using the concept of depth-of-maximum observability of the buses, where fault recorders were placed at strategic locations for continuous monitoring of the whole power network. Discrete wavelet transform (DWT) was performed for preprocessing of the recorded signals and for reliable attribute selection. The extracted key features were fed to decision tree algorithm for fast and precision fault recognition. The results confirm that the method suggested is highly proficient in classifying major power transmission network faults with precision and rapidity.