Análise do filtro Savitzky-Golay para redução de ruído — uma avaliação crítica em aplicações de estimação fasorial e processamento de sinais

Loading...
Thumbnail Image

Journal Title

Journal ISSN

Volume Title

Publisher

Universidade Federal de Juiz de Fora (UFJF)

DOI

Abstract

This work presents a critical analysis of the Savitzky-Golay (SG) filter’s application as a pre- processing technique for noise reduction in power system signals. The investigation evaluated the filter’s effectiveness in three distinct scenarios: phasor estimation via Discrete Fourier Transform (DFT), resampling of off-nominal frequency signals, and numerical differentiation. Contrary to the initial hypothesis, the results showed that SG pre-filtering was ineffective for phasor estimation under synchronous conditions, as the DFT itself acts as a superior filter for this specific task. In contrast, the SG filter showed excellent performance in time-domain applications, significantly reducing the Mean Squared Error (MSE) in resampling and robustly outperforming other numerical differentiators in the presence of noise, as measured by the Mean Absolute Error (MAE). It is concluded that the effectiveness of the SG filter is critically application-dependent, excelling in time-domain smoothing tasks but proving redundant for frequency-domain analyses where the primary tool already possesses inherent filtering characteristics.

Description

Citation

Endorsement

Review

Supplemented By

Referenced By

Creative Commons license

Except where otherwise noted, this item's license is described as Acesso Aberto