S. Ramana K. Joga, Pampa Sinha, Manoj K. Maharana, Dwarkadas P. Kothari, and Chitralekha Jena
Fault identification, DTCWT, machine learning, power system protection, zone selection
In this paper, a novel scheme is suggested for identifying the location of short-circuit faults that occur in distribution system. The proposed genetic algorithm and graph theory–based method is designed such a way that it splits electrical distribution system into protection zones containing buses, protection relays with measuring devices. Proposed methodology also decreases the calculation burden in dealing with large number of data sets. Genetic algorithm– based heuristic search method is used to place measuring devices at optimal location, and it is carried out in MATLAB. A new signal processing technique named dual tree complex wavelets transform is used for feature extraction, and support vector machine–based machine learning classifier is used for pattern recognition. IEEE33 bus radial distribution system and IEEE13 bus feeder test systems are tested for validating the proposed methodology and all the simulation work carried out in MATLAB Simulink.
Important Links:
Go Back
