Event Classification in Microgrids using Harmonic Synchrophasors

Abstract

The rapid advancements in monitoring technologies have opened up new possibilities for deploying Phasor Measurement Units (PMUs) in distribution networks. This development holds particular significance in active distribution networks, as it greatly enhances the availability of information to support network operations. However, the integration of renewable energy sources into existing distribution networks introduces a significant challenge in the form of increased harmonic content. These two factors, the abundance of complex data and the presence of harmonics, pose challenges for effectively managing modern systems. Hence, this research aims to propose a novel approach that utilizes PMU data, specifically harmonic synchrophasors, to improve event classification in distribution networks with distributed generation penetration. Simulations of typical scenarios in a 20kV and 11-bus benchmark model are conducted, resulting in a dataset comprising 2430 examples across four different event classes. The classification task is performed using optimized machine learning models, considering various feature scenarios. The robustness of the proposed method is assessed by evaluating the impact of factors such as measurement point distance, PMU location, and noise in the measured data on the overall accuracy. Preliminary results demonstrate the suitability of harmonic synchrophasors for enhancing state-of-the-art event classification methodologies in active distribution systems.