Detection of Voltage Fluctuations in Low-Voltage Power Distribution Networks with Principal Component Analysis

Abstract

This work presents the development and results of an automated event detection strategy based on principal component analysis (PCA) for low voltage distribution grids with the presence of distributed generation (DG) and phasor measurement units (PMUs). The proposed methodology, relying on measurements provided by PMUs installed at different nodes, is capable of correctly identifying and distinguishing abnormal operating conditions (AOC) from normal operating conditions (NOC) without requiring any information about the network topology or electrical parameters of its components. Moreover, it is tested and validated under voltage sags and swells in a real-based power distribution network simulated in MATLAB with PMUs deployed in distinct settings