Optimal Phasor Measurement Unit Placement Using Machine Learning Technique

The Phasor Measurement Unit (PMU) plays a massive role in dynamic monitoring and wide-area protection of the power system. Although installing PMU on each bus makes sense, it is unrealistic due to the high cost and unnecessary. Given these circumstances, academia and industry have been researching to find the best algorithm for Optimal Phasor Measurement Unit Placement (OPP). The OPP is primarily driven by economic and observability issues, among others. Traditionally, all OPP strategies frame OPP as an optimization problem and use various optimization algorithms to find the best solution. In this paper, a Machine Learning (ML) technique based on a data-driven approach is investigated. The proposed novel algorithm makes use of the Real-time Simulator (OPAL-RT) to extract data and the Random Forest Method to detect the physical anomaly in the power system. The results are further processed to find the best solution for PMU placement. The proposed approach is applied in the IEEE 9-bus system to validate its effectiveness. The regression technique is used to verify the ability of the approach to maximize the observability of the entire system.