文献类型：期刊文献

英文题名：Research on MCSA-based feature enhancement and diagnosis method for AC motor shaft misalignment using third-order transient energy and principal component reconstruction

作者：Liao, Zhiqiang[1,2]; Cai, Renchao[1]; Yan, Zhijia[1]; Huang, Zhende[1]; Chen, Peng[3]; Song, Xuewei[1,2]

机构：[1] Guangdong Ocean University, Naval Architecture and Shipping College, Zhanjiang, 524088, China; [2] Guangdong Provincial Key Laboratory of Intelligent Equipment for South China Sea Marine Ranching, Zhanjiang, 524088, China; [3] Mie University, Graduate School of Environmental Science and Technology, Tsu-shi, 514-8507, Japan

年份：2025

卷号：25

期号：20

起止页码：37761

外文期刊名：IEEE Sensors Journal

收录：EI(收录号：20253819175002)、Scopus(收录号：2-s2.0-105015875979)

语种：英文

外文关键词：Alignment - Electric fault currents - Electromagnetic transients - Extraction - Failure analysis - Fault detection - Induction motors - Power quality - Principal component analysis - Signal processing - Transients

外文摘要：To address the issue of weak shaft misalignment fault features caused by strong harmonic interference in the motor stator current signal, a fault feature enhancement method based on third-order transient energy-enhanced principal component reconstruction is proposed. First, a Third-order Transient Energy Operator (TTEO) is constructed using higher-order difference terms and combined terms to capture the higher-order variations and instantaneous energy of the signal. Then, the Intrinsic Time-scale Principal Component Reconstruction (ITPCR) is employed to capture local time-scale characteristics and extract fault features. Finally, integrating fault characteristic frequency calculation methods, a motor shaft misalignment diagnosis model is proposed that comprehensively considers higher-order signal characteristics and localized temporal properties. The effectiveness of the proposed method was validated through experiments conducted on a motor shaft misalignment test bench. The results demonstrate that the proposed method can not only effectively enhance fault features but also accurately identify fault features of different severity levels, achieving reliable diagnosis of shaft eccentricity faults of different severity levels. Furthermore, comparative experiments involving five filtering methods and five energy enhancement methods were performed. The experimental results indicate that the proposed method exhibits superior feature extraction capability and demonstrates strong feasibility. ? 2001-2012 IEEE.