EEMD Independent Extraction for Mixing Features of Rotating Machinery Reconstructed in Phase Space

Sensors

Zaichao Ma²  Guangrui Wen¹  Cheng Jiang² 

[1] State Key Lab for Strength and Vibration of Mechanical Structures, Xi’an Jiaotong University, Xi’an 710049, China;School of Mechanical Engineering, Xi’an Jiaotong University, No.28 Xianning West Road, Xi’an 710049, China; E-Mails:

关键词: PSEEMD;    EEMD;    mode mixing;    amplitude of added noise;    weak features;   

DOI  :  10.3390/s150408550

来源: mdpi

【 摘 要 】

Empirical Mode Decomposition (EMD), due to its adaptive decomposition property for the non-linear and non-stationary signals, has been widely used in vibration analyses for rotating machinery. However, EMD suffers from mode mixing, which is difficult to extract features independently. Although the improved EMD, well known as the ensemble EMD (EEMD), has been proposed, mode mixing is alleviated only to a certain degree. Moreover, EEMD needs to determine the amplitude of added noise. In this paper, we propose Phase Space Ensemble Empirical Mode Decomposition (PSEEMD) integrating Phase Space Reconstruction (PSR) and Manifold Learning (ML) for modifying EEMD. We also provide the principle and detailed procedure of PSEEMD, and the analyses on a simulation signal and an actual vibration signal derived from a rubbing rotor are performed. The results show that PSEEMD is more efficient and convenient than EEMD in extracting the mixing features from the investigated signal and in optimizing the amplitude of the necessary added noise. Additionally PSEEMD can extract the weak features interfered with a certain amount of noise.

【 授权许可】

CC BY   
© 2015 by the authors; licensee MDPI, Basel, Switzerland.