Design of a Machine Learning Model to Enhance the Arming of the System Integrity Protection Scheme of the Brazilian North-Southeast HVDC Bipoles

This paper presents the modeling and implementation of a customized Machine Learning (ML) model designed to take advantage of synchrophasor data to enhance the arming procedure of a critical System Integrity Protection Scheme (SIPS) of the Brazilian Interconnected Power System (BIPS). This model allows risk-averse decision-making, mitigating loss of selectivity conditions. Implementation has been achieved using applications developed in the Open and Extensible Control and Analytics (openECA) software environment. Results are obtained using simulations on a Real-Time Digital Simulator (RTDS) set-up, which has been provided with control and protection replicas of the high voltage direct current (HVDC) systems of the BIPS.