Investigation on enhanced mathematical morphological operators for bearing fault feature extraction

2021 ◽  
Author(s):  
Bingyan Chen ◽  
Yao Cheng ◽  
Weihua Zhang ◽  
Guiming Mei
Keyword(s):  
Feature Extraction ◽  
Morphological Operators ◽  
Bearing Fault ◽  
Fault Feature Extraction

A fractal-dimension-based envelope demodulation for rolling element bearing fault feature extraction from vibration signals

2016 ◽  
Vol 230 (18) ◽  
pp. 3194-3211 ◽  
Author(s):  
Juanjuan Shi ◽  
Ming Liang
Keyword(s):  
Feature Extraction ◽  
Fractal Dimension ◽  
Search Algorithm ◽  
Vibration Signal ◽  
Rolling Element Bearing ◽  
Vibration Signals ◽  
Bearing Fault ◽  
Bearing Condition Monitoring ◽  
Interference Frequency ◽  
Fault Feature Extraction

Vibration analysis has been extensively used as an effective tool for bearing condition monitoring. The vibration signal collected from a defective bearing is, however, a mixture of several signal components including the fault feature (i.e. fault-induced impulses), periodic interferences from other mechanical/electrical components, and background noise. The incipient impulses which excite as well as modulate the resonance frequency of the system are easily masked by compounded effects of periodic interferences and noise, making it challenging to do a reliable fault diagnosis. As such, this paper proposes an envelope demodulation method termed short time fractal dimension (STFD) transform for fault feature extraction from such vibration signal mixture. STFD transform calculation related issues are first addressed. Then, by STFD, the original signal can be quickly transformed into a STFD representation, where the envelope of fault-induced impulses becomes more pronounced whereas interferences are partly weakened due to their morphological appearance differences. It has been found that the lower the interference frequency, the less effect the interference has on STFD representations. When interference frequency keeps increasing, more effects on STFD representations will be resulted. Such effects can be reduced by the proposed kurtosis-based peak search algorithm (KPSA). Therefore, bearing fault signature is kept and interferences are further weakened in the STFD-KPSA representation. The proposed method has been favourably compared with two widely used enveloping methods, i.e. multi-morphological analysis and energy operator, in terms of extracting impulse envelopes from vibration signals obscured by multiple interferences. Its performance has also been examined using both simulated and experimental data.

scholarly journals Rolling bearing fault feature extraction under variable conditions using hybrid order tracking and EEMD

2016 ◽  
Vol 18 (7) ◽  
pp. 4449-4457 ◽  
Author(s):  
Hongkai Jiang ◽  
Qiushi Cai ◽  
Huiwei Zhao ◽  
Zhiyong Meng
Keyword(s):  
Feature Extraction ◽  
Rolling Bearing ◽  
Bearing Fault ◽  
Order Tracking ◽  
Fault Feature Extraction
Sign in / Sign up

Export Citation Format

Share Document