A Comparative Study of Various Methods of Bearing Faults Diagnosis Using the Case Western Reserve University Data

Boudiaf, Adel; Moussaoui, Abdelkrim; Dahane, Amine; Atoui, Issam

doi:10.1007/s11668-016-0080-7

Cited by 110 publications

(64 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Any feature has its own limitations [9], e.g., the time domain feature could not detect the faulty component, the frequency domain feature is unable to identify the location of damage, the envelop analysis requires prior knowledge and professional experience and the wavelet tree feature requires pre-selection of the suitable mother wavelet and appropriate level of decomposition.…”

Section: Introductionmentioning

confidence: 99%

An Improved Bearing Fault Diagnosis Method using One-Dimensional CNN and LSTM

2018

SV-JME

View full text Add to dashboard Cite

Section: Introductionmentioning

confidence: 99%

An Improved Bearing Fault Diagnosis Method using One-Dimensional CNN and LSTM

2018

SV-JME

View full text Add to dashboard Cite

“…In case of passive-mode, the expert knowledge-based features are manually analyzed for fault identification [5]. On the contrary, the active-mode supplies the extracted features to a machine-learning based classification/fault identification model [6].…”

Section: Introductionmentioning

confidence: 99%

“…Moreover, performance of machine learning -based classifiers relies heavily on the represent ability and quality of the features extracted from raw data. In literature, several signal processing and statistical based vibration features have been proposed, examples include wavelet packet transform (WPT), Fast Fourier Transform (FFT), cepstrum information, Short Time Fourier Transform (STFT), empirical mode decomposition (EMD), time-domain statistical features (TDSF) [5]. All these feature representations have their respective strengths and limitations and are extensively reviewed by [5].…”

Section: Introductionmentioning

confidence: 99%

“…In literature, several signal processing and statistical based vibration features have been proposed, examples include wavelet packet transform (WPT), Fast Fourier Transform (FFT), cepstrum information, Short Time Fourier Transform (STFT), empirical mode decomposition (EMD), time-domain statistical features (TDSF) [5]. All these feature representations have their respective strengths and limitations and are extensively reviewed by [5]. Several fault-diagnostic models have been proposed by combining aforementioned vibration-features with suitable classifiers such as SVM, ANN, BPNN, PNN, Fuzzy Inference, ANFIS, multinomial logistic regression.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Robust Feature Extraction on Vibration Data under Deep-Learning Framework: An Application for Fault Identification in Rotary Machines

Shaheryar¹,

Yin²,

Ramay³

2017

IJCA

View full text Add to dashboard Cite

Mechanical failures in rotating machinery (e.g. wind turbines, generators, motor-derives etc.) may result in catastrophic failures. Different mechanical faults induce characteristic vibrations in the equipment structure. Online vibration monitoring helps mitigate catastrophic failures through early detection and identification of underlying mechanical faults. However, extracting characteristic vibration features that improve fault classification performance and are robust to various noises in the vibration signals is a challenging task. Various statistical and signal processing-based vibrationfeatures have been proposed in the literature. These vibrationfeatures were devised on the basis of prior knowledge about characteristics of vibration signals from different fault types. Recently, automatic feature extraction through unsupervised learning in deep neural architectures has resulted in state of art performance on image and speech recognition tasks. So, Instead of feature-engineering, we, here, hypothesized that feature learning on raw vibration signal possibly will extract vibration-features that can improve fault identification performance of subsequent classifier. To the purpose, we explored Convolutional Neural Network for unsupervised feature learning on vibration signals and Denoising AutoEncoder for extracting vibration features that are robust and invariant to the noises in vibration signals. We proposed a Hybrid deep-model consisting of a Multi-channel Convolutional Neural Network followed by a stack of Denoising Auto-Encoders (MCNN-SDAE) with a single classification layer at the top. We compared the fault identification and classification performance of the proposed model with other models employing tradition statistical and signal processing based vibration-features. We validated the performance of all models on a benchmark vibration data collected from an experimental test-rig specifically designed to study vibration characteristics of bearing related faults.

show abstract

“…Meanwhile, various nonlinear multi-body dynamic models have been proposed, due to Hertzian contact, unbalanced rotor effects, radial internal clearance, raceway spalls and the size of the rolling elements [4][5][6][7]. Diagnosis analyses cover the exploration and development of approaches for the feature extraction and identification for rolling bearing faults from the viewpoint of signal processing or information theory, such as statistical processing, fractal dimension, linear discriminant analysis, cepstrum analysis, time-frequency analysis, supervised learned processing and so on [8][9][10][11][12][13][14][15][16][17][18][19][20]. Few studies however have investigated and extracted the characteristics of the vibration signals on rolling bearings in terms of dynamics theory.…”

Section: Introductionmentioning

confidence: 99%

Fault Detection for Vibration Signals on Rolling Bearings Based on the Symplectic Entropy Method

Lei

Meng

Dong

2017

Entropy

View full text Add to dashboard Cite

Abstract:Bearing vibration response studies are crucial for the condition monitoring of bearings and the quality inspection of rotating machinery systems. However, it is still very difficult to diagnose bearing faults, especially rolling element faults, due to the complex, high-dimensional and nonlinear characteristics of vibration signals as well as the strong background noise. A novel nonlinear analysis method-the symplectic entropy (SymEn) measure-is proposed to analyze the measured signals for fault monitoring of rolling bearings. The core technique of the SymEn approach is the entropy analysis based on the symplectic principal components. The dynamical characteristics of the rolling bearing data are analyzed using the SymEn method. Unlike other techniques consisting of high-dimensional features in the time-domain, frequency-domain and the empirical mode decomposition (EMD)/wavelet-domain, the SymEn approach constructs low-dimensional (i.e., two-dimensional) features based on the SymEn estimate. The vibration signals from our experiments and the Case Western Reserve University Bearing Data Center are applied to verify the effectiveness of the proposed method. Meanwhile, it is found that faulty bearings have a great influence on the other normal bearings. To sum up, the results indicate that the proposed method can be used to detect rolling bearing faults.

show abstract

A Comparative Study of Various Methods of Bearing Faults Diagnosis Using the Case Western Reserve University Data

Cited by 110 publications

References 29 publications

An Improved Bearing Fault Diagnosis Method using One-Dimensional CNN and LSTM

An Improved Bearing Fault Diagnosis Method using One-Dimensional CNN and LSTM

Robust Feature Extraction on Vibration Data under Deep-Learning Framework: An Application for Fault Identification in Rotary Machines

Fault Detection for Vibration Signals on Rolling Bearings Based on the Symplectic Entropy Method

Contact Info

Product

Resources

About