Feature extraction and recognition for rolling element bearing fault utilizing short-time Fourier transform and non-negative matrix factorization

2014 ◽  
Vol 28 (1) ◽  
pp. 96-105 ◽  
Author(s):  
Huizhong Gao ◽  
Lin Liang ◽  
Xiaoguang Chen ◽  
Guanghua Xu
Keyword(s):  
Feature Extraction ◽  
Fourier Transform ◽  
Matrix Factorization ◽  
Rolling Element Bearing ◽  
Short Time Fourier Transform ◽  
Bearing Fault ◽  
Rolling Element ◽  
Element Bearing ◽  
Short Time ◽  
Non Negative Matrix Factorization

scholarly journals A Combined Short Time Fourier Transform and Image Classification Transformer Model for Rolling Element Bearings Fault Diagnosis in Electric Motors

2021 ◽  
Vol 3 (1) ◽  
pp. 228-242
Author(s):  
Christos T. Alexakos ◽  
Yannis L. Karnavas ◽  
Maria Drakaki ◽  
Ioannis A. Tziafettas
Keyword(s):  
Fourier Transform ◽  
Fault Detection ◽  
Image Classification ◽  
Rolling Element Bearings ◽  
Short Time Fourier Transform ◽  
Bearing Fault ◽  
Bearing Fault Detection ◽  
Rolling Element ◽  
Computational Resources ◽  
Short Time

The most frequent faults in rotating electrical machines occur in their rolling element bearings. Thus, an effective health diagnosis mechanism of rolling element bearings is necessary from operational and economical points of view. Recently, convolutional neural networks (CNNs) have been proposed for bearing fault detection and identification. However, two major drawbacks of these models are (a) their lack of ability to capture global information about the input vector and to derive knowledge about the statistical properties of the latter and (b) the high demand for computational resources. In this paper, short time Fourier transform (STFT) is proposed as a pre-processing step to acquire time-frequency representation vibration images from raw data in variable healthy or faulty conditions. To diagnose and classify the vibration images, the image classification transformer (ICT), inspired from the transformers used for natural language processing, has been suitably adapted to work as an image classifier trained in a supervised manner and is also proposed as an alternative method to CNNs. Simulation results on a famous and well-established rolling element bearing fault detection benchmark show the effectiveness of the proposed method, which achieved 98.3% accuracy (on the test dataset) while requiring substantially fewer computational resources to be trained compared to the CNN approach.

scholarly journals Sampled and discretized of short-time Fourier transform and non-negative matrix factorization: the single-channel source separation case

2020 ◽  
Vol 9 (1) ◽  
pp. 41-48
Author(s):  
Jans Hendry ◽  
Isnan Nur Rifai ◽  
Yoga Mileniandi
Keyword(s):  
Fourier Transform ◽  
Matrix Factorization ◽  
Single Channel ◽  
Source Separation ◽  
Short Time Fourier Transform ◽  
Time Frequency ◽  
Frequency Representation ◽  
Short Time ◽  
Better Than ◽  
Non Negative Matrix Factorization

The Short-time Fourier transform (STFT) is a popular time-frequency representation in many source separation problems. In this work, the sampled and discretized version of Discrete Gabor Transform (DGT) is proposed to replace STFT within the single-channel source separation problem of the Non-negative Matrix Factorization (NMF) framework. The result shows that NMF-DGT is better than NMF-STFT according to Signal-to-Interference Ratio (SIR), Signal-to-Artifact Ratio (SAR), and Signal-to-Distortion Ratio (SDR). In the supervised scheme, NMF-DGT has a SIR of 18.60 dB compared to 16.24 dB in NMF-STFT, SAR of 13.77 dB to 13.69 dB, and SDR of 12.45 dB to 11.16 dB. In the unsupervised scheme, NMF-DGT has a SIR of 0.40 dB compared to 0.27 dB by NMF-STFT, SAR of -10.21 dB to -10.36 dB, and SDR of -15.01 dB to -15.23 dB.

Roller Bearing Fault Feature Extraction Based on Compressive Sensing

2018 ◽  
Author(s):  
Huibin Lin ◽  
Jianmeng Tang ◽  
Chris Mechefske
Keyword(s):  
Feature Extraction ◽  
Compressive Sensing ◽  
Learning Algorithm ◽  
Sampling Rate ◽  
Roller Bearing ◽  
Rolling Element Bearing ◽  
Bearing Fault ◽  
Rolling Element ◽  
Element Bearing ◽  
Fault Feature Extraction

Compressive sensing (CS) theory allows measurement of sparse signals with a sampling rate far lower than the Nyquist sampling frequency. This could reduce the burden of local storage and remote transmitting. The periodic impacts generated in rolling element bearing local faults are obviously sparse in the time domain. According to this sparse feature, a rolling element bearing fault feature extraction method based on CS theory is proposed in the paper. Utilizing the shift invariant dictionary learning algorithm and the periodic presentation characteristic of local faults of roller bearings, a shift-invariant dictionary of which each atom contains only one impact pattern is constructed to represent the fault impact as sparsely as possible. The limited degree of sparsity is utilized to reconstruct the feature components based on compressive sampling matching pursuit (CoSaMP) method, realizing the diagnosis of the roller bearing impact fault. A simulation was used to analyze the effects of parameters such as sparsity, SNR and compressive rate on the proposed method and prove the effectiveness of the proposed method.

Feature Extraction of Faulty Rolling Element Bearing Based on Time Synchronous Average and Cepstrum Edit

2014 ◽  
Vol 889-890 ◽  
pp. 666-670
Author(s):  
Zong Tao Li ◽  
Yan Gao ◽  
Xiang Zhou ◽  
Yu Guo
Keyword(s):  
Feature Extraction ◽  
Phase Spectrum ◽  
Rolling Element Bearing ◽  
The Real ◽  
Discrete Frequency ◽  
Rolling Element ◽  
Frequency Components ◽  
Element Bearing ◽  
Inverse Procedure

The cepstrum edit scheme for the vibration feature extraction of the faulty rolling element bearing (REB) is studied in this paper. By combined the time synchronous average (TSA) and the real cepstrum to localize and edit the cepstral lines of the original vibration, the unwanted discrete frequency components can be removed. Then, a corresponding inverse procedure is designed, in which the edited cepstrum and the original phase spectrum are employed to reconstruct the edited vibration for the REB feature extraction. Simulation verified the scheme positively.

Sign in / Sign up

Export Citation Format

Share Document