scholarly journals Deep Learning Enabled Fault Diagnosis Using Time-Frequency Image Analysis of Rolling Element Bearings

2017 ◽  
Vol 2017 ◽  
pp. 1-17 ◽  
Author(s):  
David Verstraete ◽  
Andrés Ferrada ◽  
Enrique López Droguett ◽  
Viviana Meruane ◽  
Mohammad Modarres
Keyword(s):  
Deep Learning ◽  
Fault Diagnosis ◽  
Expert Knowledge ◽  
Rolling Element Bearing ◽  
Hilbert Huang Transform ◽  
Time Frequency ◽  
Public Data ◽  
Feature Extraction And Selection ◽  
Rolling Element ◽  
Image Representations

Traditional feature extraction and selection is a labor-intensive process requiring expert knowledge of the relevant features pertinent to the system. This knowledge is sometimes a luxury and could introduce added uncertainty and bias to the results. To address this problem a deep learning enabled featureless methodology is proposed to automatically learn the features of the data. Time-frequency representations of the raw data are used to generate image representations of the raw signal, which are then fed into a deep convolutional neural network (CNN) architecture for classification and fault diagnosis. This methodology was applied to two public data sets of rolling element bearing vibration signals. Three time-frequency analysis methods (short-time Fourier transform, wavelet transform, and Hilbert-Huang transform) were explored for their representation effectiveness. The proposed CNN architecture achieves better results with less learnable parameters than similar architectures used for fault detection, including cases with experimental noise.

Wigner–Ville distribution based on cyclic spectral density and the application in rolling element bearings diagnosis

2011 ◽  
Vol 225 (12) ◽  
pp. 2831-2847 ◽  
Author(s):  
Y Zhou ◽  
J Chen ◽  
G M Dong ◽  
W B Xiao ◽  
Z Y Wang
Keyword(s):  
Fault Diagnosis ◽  
Spectral Density ◽  
Wigner Distribution ◽  
Rolling Element Bearing ◽  
Rolling Element Bearings ◽  
Accurate Analysis ◽  
Time Frequency ◽  
Bearing Fault ◽  
Rolling Element ◽  
Cross Terms

The vibration signals of rolling element bearings are random cyclostationary when they have faults. Also, statistical properties of the signals change periodically with time. The accurate analysis of time-varying signals is an essential pre-requisite for the fault diagnosis and hence safe operation of rolling element bearings. The Wigner distribution is probably most widely used among the Cohen’s class in order to describe how the spectral content of a signal changes over time. However, the basic nature of such signals causes significant interfering cross-terms, which do not permit a straightforward interpretation of the energy distribution. To overcome this difficulty, the Wigner–Ville distribution (WVD) based on the cyclic spectral density (CSD) is discussed in this article. It is shown that the improved WVD, based on CSD of a long time series, can render the time–frequency distribution less susceptible to noise, and restrain the cross-terms in the time–frequency domain. Simulation and experiment of the rolling element-bearing fault diagnosis are performed, and the results indicate the validity of WVD based on CSD in time–frequency analysis for bearing fault detection.

Wavelet Analysis and Envelope Detection For Rolling Element Bearing Fault Diagnosis—Their Effectiveness and Flexibilities

2001 ◽  
Vol 123 (3) ◽  
pp. 303-310 ◽  
Author(s):  
Peter W. Tse ◽  
Y. H. Peng ◽  
Richard Yam
Keyword(s):  
Fault Diagnosis ◽  
Wavelet Analysis ◽  
Rolling Element Bearing ◽  
Frequency Spectra ◽  
Bearing Faults ◽  
Time Frequency ◽  
Envelope Detection ◽  
Rolling Element ◽  
Element Bearing ◽  
Bearing Characteristic

The components which often fail in a rolling element bearing are the outer-race, the inner-race, the rollers, and the cage. Such failures generate a series of impact vibrations in short time intervals, which occur at Bearing Characteristic Frequencies (BCF). Since BCF contain very little energy, and are usually overwhelmed by noise and higher levels of macro-structural vibrations, they are difficult to find in their frequency spectra when using the common technique of Fast Fourier Transforms (FFT). Therefore, Envelope Detection (ED) is always used with FFT to identify faults occurring at the BCF. However, the computation of ED is complicated, and requires expensive equipment and experienced operators to process. This, coupled with the incapacity of FFT to detect nonstationary signals, makes wavelet analysis a popular alternative for machine fault diagnosis. Wavelet analysis provides multi-resolution in time-frequency distribution for easier detection of abnormal vibration signals. From the results of extensive experiments performed in a series of motor-pump driven systems, the methods of wavelet analysis and FFT with ED are proven to be efficient in detecting some types of bearing faults. Since wavelet analysis can detect both periodic and nonperiodic signals, it allows the machine operator to more easily detect the remaining types of bearing faults which are impossible by the method of FFT with ED. Hence, wavelet analysis is a better fault diagnostic tool for the practice in maintenance.

A Novel Fault Diagnosis Method Based on Time-Frequency Image Recognition

2014 ◽  
Vol 687-691 ◽  
pp. 3569-3573 ◽  
Author(s):  
Wei Gang Wang ◽  
Zhan Sheng Liu
Keyword(s):  
Fault Diagnosis ◽  
Image Recognition ◽  
Rolling Element Bearing ◽  
Support Vector ◽  
Time Frequency ◽  
Rolling Element ◽  
Diagnosis Method ◽  
Element Bearing ◽  
Vibration Time ◽  
Pyramid Matching

A novel intelligent fault diagnosis method based on vibration time-frequency image recognition is proposed in this paper. First, Smooth pseudo Wigner-Ville distribution (SPWVD) is employed to represent the time-frequency distribution characteristics. Then, the features of time-frequency images are extracted by using locality-constrained linear coding (LLC) and spatial pyramid matching. Next, we use the support vector machine to identify these feature vectors for realizing intelligent fault detection. The promise of our algorithm is illustrated by performing above procedures on the vibration signals measured from rolling element bearing with sixteen operating states. Experimental results show that the proposed method can acquire higher diagnosis accuracy compared with the ScSPM method in rolling element bearing diagnosis.

Time-Frequency Analysis System of Rolling Element Bearing Based on Virtual Instrument

2013 ◽  
Vol 278-280 ◽  
pp. 844-847
Author(s):  
Xian Jun Yu ◽  
Yu Guo ◽  
Jun Guo ◽  
Yun Li
Keyword(s):  
Fault Diagnosis ◽  
Virtual Instrument ◽  
Rolling Element Bearing ◽  
Time Frequency ◽  
Fault Features ◽  
Acceleration Sensors ◽  
Rolling Element ◽  
Data Acquisition Card ◽  
Element Bearing ◽  
Analysis System

Rolling element bearing (REB) is one of important components in the condition monitoring and faults diagnosis of machinery. In this paper, a REB fault diagnosis system is presented, which is developed by using LabVIEW. In the system, vibration signals are picked by acceleration sensors and acquired by NI USB data acquisition card at first. Then, the fault diagnosis can be performed in the time-frequency domain by various time-frequency methods. The experiments show that the presented system can be used to extract the bearing fault features and diagnoses the failures effectively.

scholarly journals Method for Fault Diagnosis of Temperature-Related MEMS Inertial Sensors by Combining Hilbert–Huang Transform and Deep Learning

Sensors ◽  
2020 ◽  
Vol 20 (19) ◽  
pp. 5633 ◽  
Author(s):  
Tong Gao ◽  
Wei Sheng ◽  
Mingliang Zhou ◽  
Bin Fang ◽  
Futing Luo ◽  
...  
Keyword(s):  
Deep Learning ◽  
Fault Diagnosis ◽  
Short Term Memory ◽  
Inertial Sensors ◽  
Fault Classification ◽  
Hilbert Huang Transform ◽  
Time Frequency ◽  
Mode Decomposition ◽  
Improved Performance ◽  
Micro Electromechanical System

In this paper, we propose a novel method for fault diagnosis in micro-electromechanical system (MEMS) inertial sensors using a bidirectional long short-term memory (BLSTM)-based Hilbert–Huang transform (HHT) and a convolutional neural network (CNN). First, the method for fault diagnosis of inertial sensors is formulated into an HHT-based deep learning problem. Second, we present a new BLSTM-based empirical mode decomposition (EMD) method for converting one-dimensional inertial data into two-dimensional Hilbert spectra. Finally, a CNN is used to perform fault classification tasks that use time–frequency HHT spectrums as input. According to our experimental results, significantly improved performance can be achieved, on average, for the proposed BLSTM-based EMD algorithm in terms of EMD computational efficiency compared with state-of-the-art algorithms. In addition, the proposed fault diagnosis method achieves high accuracy in fault classification.

Sign in / Sign up

Export Citation Format

Share Document