A time‐domain statistical approach for harmonics separation and analysis. (20th July 2016)

• Record Type: Journal Article
• Authors: He, Chuan; Shu, Qin; Liu, Tianqi; Han, Xiaoyan
• Is Part Of: International transactions on electrical energy systems. Volume 27:Number 1(2017:Jan.)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2016-07-20
• Subjects: power quality -- harmonics -- principal component analysis -- single channel signal analysis -- Euclidean embedding dimension; Electric power -- Periodicals; Electric power systems -- Periodicals; Electrical engineering -- Periodicals; 621.3
• DOI: 10.1002/etep.2239
• Languages: English
• ISSNs: 2050-7038

Abstract

Summary: The paper proposes a novel time‐domain statistical method to separate and analyze harmonics based on single channel analysis and principal component analysis. The proposed method separates the fundamental wave and harmonics to provide specific time‐domain information. Two identification methods are proposed to estimate the number of harmonics within the power system signal, even with high noise. This paper validates the effectiveness of the proposed method through mathematical deduction and studies the selection of Euclidean embedding dimension. The robustness of the proposed method is tested on several simulated synthetic signals and a real‐world signal. Simulation results show that the proposed method can detect harmonics with small amplitude in the presence of noise and has better separated waveform and less execution time than wavelet‐packet transform and empirical mode decomposition. It is also shown that the proposed method can detect sudden voltage changes and separate disturbances.