Fault diagnosis of rotating machinery using Gaussian process and EEMD‐treelet

Wu, Edmond Q.; Wang, Jin; Peng, Xian‐Yong; Zhang, Peidong; Law, Rob; Chen, Xi; Lin, J. H.

doi:10.1002/acs.2952

Cited by 14 publications

(5 citation statements)

References 41 publications

(64 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Furthermore, the time-varying features of non-stationary signals could be characterized. TFA methods are wildly applied in radar, sonar and astronomical, biomedicine, and mechanical engineering areas [1][2][3][4][5][6] , etc. The conventional TFA methods are roughly divided into linear and quadratic transforms, and all of them have respective drawbacks.…”

Section: Refining the Time-frequency Characteristic Of Non-stationary...mentioning

confidence: 99%

Refining the time–frequency characteristic of non-stationary signal for improving time–frequency representation under variable speeds

Liu

Xiang

Jiang

et al. 2023

Sci Rep

View full text Add to dashboard Cite

Time–frequency ridge not only exhibits the variable process of non-stationary signal with time changing but also provides the information of signal synchronous or non-synchronous components for subsequent detection research. Consequently, the key is to decrease the error between real and estimated ridge in the time–frequency domain for accurate detection. In this article, an adaptive weighted smooth model is presented as a post-processing tool to refine the time–frequency ridge which is based on the coarse estimated time–frequency ridge using newly emerging time–frequency methods. Firstly, the coarse ridge is estimated by using multi-synchrosqueezing transform for vibration signal under variable speed conditions. Secondly, an adaptive weighted method is applied to enhance the large time–frequency energy value location of the estimated ridge. Then, the reasonable smooth regularization parameter associated with the vibration signal is constructed. Thirdly, the majorization–minimization method is developed for solving the adaptive weighted smooth model. Finally, the refined time–frequency characteristic is obtained by utilizing the stop criterion of the optimization model. Simulation and experimental signals are given to validate the performance of the proposed method by average absolute errors. Compared with other methods, the proposed method has the highest performance in refinement accuracy.

show abstract

Section: Refining the Time-frequency Characteristic Of Non-stationary...mentioning

confidence: 99%

Refining the time–frequency characteristic of non-stationary signal for improving time–frequency representation under variable speeds

Liu

Xiang

Jiang

et al. 2023

Sci Rep

View full text Add to dashboard Cite

show abstract

“…However, the average number of sets is limited, and the residual white noise leads to poor completeness of algorithm decomposition, which affects the subsequent diagnosis results. In addition, when the fault signal noise is relatively large, the EEMD method requires a higher number of iterations to reduce the residual noise, causing an increase in computing load [15]. In response to the shortcomings of EEMD, Zhang et al [8] adopted complementary EEMD (CEEMD) to improve the measurement accuracy of reference velocity for the laser Doppler velocimeter.…”

Section: Introductionmentioning

confidence: 99%

Design of a progressive fault diagnosis system for hydropower units considering unknown faults

Chen,

Zheng,

Deng

et al. 2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

To address the misidentification problem of signals containing unknown faults for hydropower units, a progressive fault diagnosis system is designed. Firstly, in view of the non-stationary and nonlinear vibration signals of hydropower units, the method of complementary ensemble empirical mode decomposition is used to process the normal and fault vibration signal samples, and the intrinsic mode function (IMF) and residual components with different frequencies are obtained. Then the IMF energy moment is calculated and used as the feature vector. Furthermore, a classifier (IMF-K1) is constructed based on the feature vector samples of the normal vibration signals of hydropower units, fault symptom indicators, and K-means algorithm to determine whether the hydropower unit is faulty; a classifier (IMF-K2) is constructed based on the feature vector samples of the fault vibration signals of hydropower units, fault symptom indicators, and K-means algorithm to determine whether the hydropower unit has the known fault; a classifier (IMF-bidirectional long short-term memory neural network (BiLSTMNN)) is constructed to distinguish the fault type of hydropower units by combining the eigenvector samples of known fault vibration signals, fault symptom indicators, and BiLSTMNN. Finally, a progressive fault diagnosis system for hydropower units is constructed using IMF-K1, IMF-K2, and IMF-BiLSTMNN, and comparative experiments are designed using the sample data from the rotor test bench and actual hydropower unit. The results show that the designed progressive fault diagnosis system has greater effectiveness in mining signal features and high fault diagnosis accuracy.

show abstract

“…This is realized by adding Gaussian white noise to change the characteristics of the extrema, and variants of this algorithm have been developed. Wu et al [ 11 ] proposed a feature learning detection method based on EEMD and a Gaussian process classifier. Trelet was used for data dimension reduction as the Gaussian process input, and it is optimized by bacterial foraging optimization.…”

Section: Introductionmentioning

confidence: 99%

Subway Gearbox Fault Diagnosis Algorithm Based on Adaptive Spline Impact Suppression

Yang

Yao

et al. 2021

Entropy

View full text Add to dashboard Cite

In the signal processing of real subway vehicles, impacts between wheelsets and rail joint gaps have significant negative effects on the spectrum. This introduces great difficulties for the fault diagnosis of gearboxes. To solve this problem, this paper proposes an adaptive time-domain signal segmentation method that envelopes the original signal using a cubic spline interpolation. The peak values of the rail joint gap impacts are extracted to realize the adaptive segmentation of gearbox fault signals when the vehicle was moving at a uniform speed. A long-time and unsteady signal affected by wheel–rail impacts is segmented into multiple short-term, steady-state signals, which can suppress the high amplitude of the shock response signal. Finally, on this basis, multiple short-term sample signals are analyzed by time- and frequency-domain analyses and compared with the nonfaulty results. The results showed that the method can efficiently suppress the high-amplitude components of subway gearbox vibration signals and effectively extract the characteristics of weak faults due to uniform wear of the gearbox in the time and frequency domains. This provides reference value for the gearbox fault diagnosis in engineering practice.

show abstract

Fault diagnosis of rotating machinery using Gaussian process and EEMD‐treelet

Cited by 14 publications

References 41 publications

Refining the time–frequency characteristic of non-stationary signal for improving time–frequency representation under variable speeds

Refining the time–frequency characteristic of non-stationary signal for improving time–frequency representation under variable speeds

Design of a progressive fault diagnosis system for hydropower units considering unknown faults

Subway Gearbox Fault Diagnosis Algorithm Based on Adaptive Spline Impact Suppression

Contact Info

Product

Resources

About