We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Wavelet threshold denoising harmonic detection method based on permutation entropy-CEEMD decomposition.
- Authors
LI Zhi-jun; ZHANG Hong-peng; WANG Ya-nan; LI Xiao
- Abstract
Although the harmonic detection method based on complete ensemble empirical mode decomposition ( CEEMD) can improve the modal aliasing problem of empirical mode decomposition ( EMD) to some extent, it needs to add additional auxiliary noise during the decomposition process, resulting in decomposition. The intrinsic mode function ( IMFS) has false components and noise residuals, which seriously affect the extraction of harmonic feature information. Based on the traditional CEEMD, this paper proposes a PE-CEEMD method based on permutation entropy ( PE) algorithm to improve the deficiency of false components in CEEMD, and solve the problem of noise residue in decomposition. The wavelet threshold denoising ( WTD) method under the threshold function denoised the inherent model function ( IMFS) obtained after decomposition, and extracted the feature information of each harmonic contained in the noise-reduced IMFS. Simulation experiments show that the PE-CEEMD method effectively improves the false component phenomenon in CEMMD. The WTD method under the new threshold function effectively eliminates the influence of residual noise on IMFS feature information extraction and improves the detection accuracy of harmonic signal, and has good anti-noise performance.
- Subjects
PERMUTATIONS; DATA mining; FEATURE extraction; HILBERT-Huang transform
- Publication
Electric Machines & Control / Dianji Yu Kongzhi Xuebao, 2020, Vol 24, Issue 12, p120
- ISSN
1007-449X
- Publication type
Article
- DOI
10.15938/j.emc.2020.12.015