Scholarlink

<p>The analysis of power quality disturbance signals by complementary ensemble empirical mode decomposition (CEEMD) leads to uneven distribution of extreme points and residual white noise that cannot be eliminated due to noise interference and lack of adaptivity, which in turn triggers the problem of modal aliasing. A power quality disturbance detection algorithm (PVR-EPD-CEEMD-WTD) is proposed, which uses peak-to-valley ratio (PVR) and energy peak distribution (EPD) to improve the CEEMD (PVR-EPD-CEEMD), and integrates with wavelet threshold denoising (WTD). Firstly, WTD is used to preprocess the noise signal for noise reduction to reduce the interference of noise on the selection of extreme points. Secondly, the PVR-EPD-CEEMD algorithm is constructed, which is embedded with PVR and EPD based on CEEMD, and adaptively determines the optimal noise amplitude and the set of total averaging times to ensure a more uniform distribution of extreme points of the signal, and eliminates the residual white noise. Finally, under the noise environment, simulation experiments are carried out on the time domain perturbation, frequency domain perturbation and composite perturbation signals, and the results show that, compared with the EEMD, VMD, CEEMD and WTD algorithms, the deviation of the starting and ending moments of perturbation in this paper’s algorithm is 1% and 0.5%, and high-frequency harmonic perturbation (the fifth harmonic) can also be accurately detected, the time domain and frequency domain distributions of the intrinsic modal function (IMF) are clear, effectively suppressing the modal aliasing problem. At the same time, compared with other algorithms, the signal-to-noise ratio (SNR) increased by 26.5%, and the root mean square error (RMSE) decreased by 43.8%, which further verifies the accuracy of the algorithm in this paper.</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

WTD and PVR-EPD to improve CEEMD for power quality disturbance detection

  • 期刊论文

  • 发表日期：

    2025年4月1日

查看全文
  • Fuyan Guo,
  • Wensen Yang,
  • Yue Wang,
  • Fei Dong,
  • Jiao Chen,
  • Zheng Qin

摘要

The analysis of power quality disturbance signals by complementary ensemble empirical mode decomposition (CEEMD) leads to uneven distribution of extreme points and residual white noise that cannot be eliminated due to noise interference and lack of adaptivity, which in turn triggers the problem of modal aliasing. A power quality disturbance detection algorithm (PVR-EPD-CEEMD-WTD) is proposed, which uses peak-to-valley ratio (PVR) and energy peak distribution (EPD) to improve the CEEMD (PVR-EPD-CEEMD), and integrates with wavelet threshold denoising (WTD). Firstly, WTD is used to preprocess the noise signal for noise reduction to reduce the interference of noise on the selection of extreme points. Secondly, the PVR-EPD-CEEMD algorithm is constructed, which is embedded with PVR and EPD based on CEEMD, and adaptively determines the optimal noise amplitude and the set of total averaging times to ensure a more uniform distribution of extreme points of the signal, and eliminates the residual white noise. Finally, under the noise environment, simulation experiments are carried out on the time domain perturbation, frequency domain perturbation and composite perturbation signals, and the results show that, compared with the EEMD, VMD, CEEMD and WTD algorithms, the deviation of the starting and ending moments of perturbation in this paper’s algorithm is 1% and 0.5%, and high-frequency harmonic perturbation (the fifth harmonic) can also be accurately detected, the time domain and frequency domain distributions of the intrinsic modal function (IMF) are clear, effectively suppressing the modal aliasing problem. At the same time, compared with other algorithms, the signal-to-noise ratio (SNR) increased by 26.5%, and the root mean square error (RMSE) decreased by 43.8%, which further verifies the accuracy of the algorithm in this paper.