Análise de oscilações de baixa frequência em sinais de sistemas elétricos de potência utilizando VMD e Prony

Abstract

This article addresses the analysis of low-frequency oscillations in power systems using the VMD (Variational Mode Decomposition) and Prony algorithms. Low-frequency oscillations can cause instability in the power system and compromise its operation and reliability when not properly damped. As they are electromechanical oscillations, they can influence the operation of the mechanical parts of turbine-generator sets, causing vibrations and wear. The VMD algorithm is used to decompose the electrical signal into its modal components and attenuate noise, thus separating low-frequency oscillation from other frequency components present in the original signal. The Prony algorithm is used to model the resulting modal components and determine the model parameters, thus characterizing the oscillations in terms of amplitude, frequency, damping factor, and phase. The presented results aim to compare the performance of the method across different scenarios. Signals with random parameters within the commonly observed range of low-frequency oscillations were generated for statistical analysis of the method’s performance across a large volume of tests. Subsequently, the proposed method is applied to synthetic signals used in other articles found in the literature, and its performance is compared to that of other methods. Additionally, data from two electrical systems, one simulated and the other real, are used to assess the method’s performance in situations closer to real-world conditions.The results obtained shows that the combination of these two algorithms allows for efficient and accurate analysis to detect and analyze low-frequency oscillations in power systems.