Data-Driven Apprehension of Cyber and Physical Anomalies in Distribution System

The modernization of the power system network with the integration of internet of things (IoT) has led to the threat of database attack along with the conventional physical anomalies such as LG-LL-LLG-LLLG faults. If these aberrations are not properly classified, it may lead to incorrect decisions for restoration. In this paper a 5-bus system and reconfigured IEEE 33 bus distribution system is developed in Typhoon HIL real time simulator to create a database for all possible types of aberrations which may exist in distribution system. Heterogeneous synchrophasor datasets obtained from the real time simulator are used for an accurate apprehension of data into faults, database attack and healthy condition of the system network. The tested model is also compared with existing supervised machine learning techniques to validate its efficacy.