Scholarlink

<p>The effectiveness of monitoring and control systems relies on the signal processing techniques used to estimate oscillatory modes and signal components. Time–frequency representation methods such as short-time Fourier transform (STFT) and S-transform are well suited for analyzing power system signals, but there is a trade-off between time and frequency resolutions. Optimizing the window length using advanced methods can help address this challenge. However, fully adaptive techniques in both time and frequency domains are highly complex. This paper focuses on addressing time–frequency resolution, delay, and complexity in processing power system signals. The use of adaptive Gaussian kernels improves the resolution without causing any phase shift, allowing for the estimation of frequency components using the phase angle. Then, the amplitude is compensated for each dominant mode. As a result, the proposed method eliminates repetitive calculations for estimating the magnitude of all frequency points across the spectrum to ensure an accurate report of modes with a desired number precision/scale. This also reduces the window length, leading to reduced complexity and delay.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Fast Fourier-based analysis of power system signals with high time–frequency resolution and minimal delay

  • 期刊论文

  • 发表日期：

    2025年2月14日

查看全文
  • Mohammad Zarei,
  • Mohammad Karimadini,
  • Mohsen Nadjafi,
  • Abolfazl Salami

摘要

The effectiveness of monitoring and control systems relies on the signal processing techniques used to estimate oscillatory modes and signal components. Time–frequency representation methods such as short-time Fourier transform (STFT) and S-transform are well suited for analyzing power system signals, but there is a trade-off between time and frequency resolutions. Optimizing the window length using advanced methods can help address this challenge. However, fully adaptive techniques in both time and frequency domains are highly complex. This paper focuses on addressing time–frequency resolution, delay, and complexity in processing power system signals. The use of adaptive Gaussian kernels improves the resolution without causing any phase shift, allowing for the estimation of frequency components using the phase angle. Then, the amplitude is compensated for each dominant mode. As a result, the proposed method eliminates repetitive calculations for estimating the magnitude of all frequency points across the spectrum to ensure an accurate report of modes with a desired number precision/scale. This also reduces the window length, leading to reduced complexity and delay.