A review on the application of deep learning in system health management

A review on data-driven fault severity assessment in rolling bearings

A review on signal processing techniques utilized in the fault diagnosis of rolling element bearings

This paper presents a novel signal processing scheme, bandwidth empirical mode decomposition, and adaptive multiscale morphological analysis (BEMD-AMMA) for early fault diagnosis of rolling bearings. In this scheme, we propose a bandwidth based method to select the best envelope interpolation method. First, multiple envelope algorithms are defined and separately subtracted from the original data to obtain the preintrinsic mode functions (PIMFs). Second, an IMF with the smallest frequency bandwidth is selected to be the optimal IMF (OIMF). Third, this OIMF is subtracted from the original signal, and then repeat the sifting process until the residual is a constant or monotonic. Since the OIMF has the smallest frequency bandwidth, the mode mixing phenomenon can be significantly weakened. After that the OIMFs with clear fault information are used to construct the main component of the original signal. Then, the AMMA is introduced to demodulate the constructed main component. Simulation and experimental vibration signals are employed to evaluate the effectiveness of the proposed method. Results show that the proposed method outperforms EMD-AMMA, ensemble empirical mode decomposition-AMMA, and generalized empirical mode decomposition-empirical envelope demodulation in detecting early inner race fault.

Application of Bandwidth EMD and Adaptive Multiscale Morphology Analysis for Incipient Fault Diagnosis of Rolling Bearings

Feature extraction is one of the most important, pivotal, and difficult problems in mechanical fault diagnosis, which directly relates to the accuracy of fault diagnosis and the reliability of early fault prediction. Therefore, a new fault feature extraction method, called the EDOMFE method based on integrating ensemble empirical mode decomposition (EEMD), mode selection, and multi-scale fuzzy entropy is proposed to accurately diagnose fault in this paper. The EEMD method is used to decompose the vibration signal into a series of intrinsic mode functions (IMFs) with a different physical significance. The correlation coefficient analysis method is used to calculate and determine three improved IMFs, which are close to the original signal. The multi-scale fuzzy entropy with the ability of effective distinguishing the complexity of different signals is used to calculate the entropy values of the selected three IMFs in order to form a feature vector with the complexity measure, which is regarded as the inputs of the support vector machine (SVM) model for training and constructing a SVM classifier (EOMSMFD based on EDOMFE and SVM) for fulfilling fault pattern recognition. Finally, the effectiveness of the proposed method is validated by real bearing vibration signals of the motor with different loads and fault severities. The experiment results show that the proposed EDOMFE method can effectively extract fault features from the vibration signal and that the proposed EOMSMFD method can accurately diagnose the fault types and fault severities for the inner race fault, the outer race fault, and rolling element fault of the motor bearing. Therefore, the proposed method provides a new fault diagnosis technology for rotating machinery.

https://www.mdpi.com/1099-4300/19/1/14/pdf

A New Feature Extraction Method Based on EEMD and Multi-Scale Fuzzy Entropy for Motor Bearing

Feature extraction of rolling bearing’s early weak fault based on EEMD and tunable Q-factor wavelet transform

Bearing fault recognition method based on neighbourhood component analysis and coupled hidden Markov model

The rolling bearing’s early stage fault feature is very weak for reasons of the signal attenuation phenomenon between the fault source and the sensor collecting the fault signal and the interferenc...

Fault diagnosis of rolling bearing’s early weak fault based on minimum entropy de-convolution and fast Kurtogram algorithm

The rolling bearing fault signal under strong background noise is very weak because of environmental noise impaction and the attenuation of signal. The feature extraction of rolling bearings’ weak fault is very important in avoiding serious disaster, but it is also very difficult. Sparse decomposition has been used in the fault feature extraction of rolling bearings. But its performance is very poor when the background noise is very strong. This text combines the minimum entropy de-convolution (MED) and sparse decomposition to extract the feature of a rolling bearing’s weak fault. Firstly, the rolling bearing weak fault signal with strong background noise is de-noised using the MED method, and subsequently the de-noised signal is handled by sparse decomposition. Finally, the fault feature extraction method-envelope demodulation is carried on the last given signal and better results are obtained. In conclusion, through simulation and experiment the effectiveness and the feasibility of the proposed method a...

Weak fault feature extraction of rolling bearing based on minimum entropy de-convolution and sparse decomposition

Aiming at the difficulty of identifying weak fault of rolling element bearing (REB) accurately using only one single fault signal evidence domain, a multi-source information deep fusion diagnosis method for REB based on multi-synchrosqueezing transform (MSST) and deep residual convolution neural network (DRCNN) is presented in this paper, which combines the potential application of MSST in fault feature extraction of REB and the advantages. Firstly, the signals of REB under different running states are transformed by short-time Fourier transform (STFT) and MSST, respectively, to obtain the STFT and MSST time-frequency spectrum symptom set. Then, multiple DRCNNs are built to perform feature learning on the obtained time-frequency multi-symptom domain information, thereby a mapping between the local feature space and the fault space is established. Finally, the testing feature vectors obtained same as the processes of training feature vectors are input into the trained DRCNN models for automatic recognition. The validity of the proposed method is verified by experiment, and the overall average recognition success rate of the proposed method reaches above 95%. Besides, its advantage is also compared with the other related methods.

Hongchao Wang

Papers

Feature extraction of rolling bearing’s early weak fault based on EEMD and tunable Q-factor wavelet transform

Bearing fault recognition method based on neighbourhood component analysis and coupled hidden Markov model

Fault diagnosis of rolling bearing’s early weak fault based on minimum entropy de-convolution and fast Kurtogram algorithm

Weak fault feature extraction of rolling bearing based on minimum entropy de-convolution and sparse decomposition

Multi-source information deep fusion for rolling bearing fault diagnosis based on deep residual convolution neural network