Comparative Study of Machine Learning Models for Power System Fault Identification and Localization

Monitoring and maintenance of faults in the power system are very complex because its configuration keeps on changing due to the inclusion of new loads and distributed generation (DGs)/microgrids in the existing power system. As a result, in the event of a fault, manual fault diagnosis is time- consuming and tedious. Machine Learning-based fault diagnoses are being proposed by the researcher to replace previously used model-based fault diagnoses. ML enables timely identification and localization of faults which can prevent unattended long- duration power failure. Various researches have been done on different ML models for power system fault diagnosis. In this paper, a brief overview of recently popular ML-based fault diagnosis techniques has been discussed and a comparative study of these models on the IEEE 9bus system has been reported with a stepwise procedure of how they were implemented.