Bearing fault detection using relative entropy of wavelet components and artificial neural networks

Schmitt, Helder Luiz; Silva, Lyvia R. B.; Scalassara, Paulo Rogério; Goedtel, Alessandro

doi:10.1109/demped.2013.6645767

Cited by 13 publications

(7 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Several works have followed this suggestion and developed interesting current-based techniques for the FD of different types of bearing damage. Reference [48] proposes a method that combines the entropy analysis of wavelet signals and neural networks for bearing fault detection and classification.…”

Section: Fault Diagnosis Of Mechanical Failures 1) Fault Diagnosismentioning

confidence: 99%

Advances in Electrical Machine, Power Electronic, and Drive Condition Monitoring and Fault Detection: State of the Art

Riera-Guasp

Antonino-Daviu

Capolino

2015

IEEE Trans. Ind. Electron.

486

234

View full text Add to dashboard Cite

Section: Fault Diagnosis Of Mechanical Failures 1) Fault Diagnosismentioning

confidence: 99%

Advances in Electrical Machine, Power Electronic, and Drive Condition Monitoring and Fault Detection: State of the Art

Riera-Guasp

Antonino-Daviu

Capolino

2015

IEEE Trans. Ind. Electron.

486

234

View full text Add to dashboard Cite

“…These years, some researchers have a strong interest in relative entropy in engineering field. Schmitt et al [19] put forward a predictability analysis method based on relative entropy measures for bearing fault detection and the result was satisfying. The relative entropy is an index which measures two probability distribution deviation.…”

Section: Introductionmentioning

confidence: 96%

Rolling Bearing Degradation State Identification Based on LCD Relative Spectral Entropy

Bao

2016

J Fail. Anal. and Preven.

View full text Add to dashboard Cite

In the interest of obtaining an effective bearing degradation feature from complex, nonlinear, and nonstationary vibration signals, a new analytical methodology based on local characteristic-scale decomposition (LCD) and relative entropy theory is proposed. On the one hand, LCD is a new and relatively excellent time-frequency analysis method to analyze practical vibration signals polluted by noise. On the other hand, relative entropy theory is a good way to characterize different degradation states by calculating the probability distribution difference between the degradation signals and the normal signal. Combining the above two theories, two new degradation features named LRNE and LRQE are extracted to indicate the bearing degradation trend from normal state to even failure state. The noise resistance ability and extensive applicability of both the features are verified by simulation signal. For further analysis of experimental vibration signals, the two features have a satisfying performance to characterize different bearing degradation states. With the help of gray relational analysis and fuzzy C-means clustering, the proposed two characteristics can identify different bearing degradation states of inner ring fault mode with high accuracy. In the end, the two features are applied to doing bearing failure analysis with the full-life bearing data. The results show that the LRNE and LRQE are sensitive to bearing degradation trend in the whole life of bearing.

show abstract

“…This can be expensive and also result in many false judgments. In the normal state and fault state of a wavelet neural network [16,17] analysis, the fault states are classified according to the characteristics of the neural network acquired through Fourier transforms. However, the use of a neural network requires a huge database of learned examples.…”

Section: Introductionmentioning

confidence: 99%

Intelligent Ball Bearing Fault Diagnosis by using Fractional Lorenz Chaos Extension Detection Method with Acceleration Sensor

Tian¹,

Fu²,

Li³

et al. 2018

Preprint

View full text Add to dashboard Cite

In this study we used a non-autonomous Chua’s Circuit, and the fractional Lorenz chaos system together with a detection method from Extension theory to analyze the voltage signals. The measured bearing signals by acceleration sensor were introduced into the master and slave systems through a Chua’s Circuit. In a chaotic system minor differences can cause significant changes that generate dynamic errors, and extension matter-element models can be used to judge the bearing conditions. Extension theory can be used to establish classical and sectional domains using the dynamic errors of the fault conditions. The results obtained were compared with those from Discrete Fourier Transform analysis, Wavelet analysis and an integer order chaos system. The diagnostic ratio showed the fractional order master and slave chaos system calculations. The results show that the method presented in this paper is very suitable for monitoring the operational state of ball bearing system to be superior to the other methods. The diagnosis ratio was better and there were other significant advantages such as low cost and few.

show abstract

Bearing fault detection using relative entropy of wavelet components and artificial neural networks

Cited by 13 publications

References 17 publications

Advances in Electrical Machine, Power Electronic, and Drive Condition Monitoring and Fault Detection: State of the Art

Advances in Electrical Machine, Power Electronic, and Drive Condition Monitoring and Fault Detection: State of the Art

Rolling Bearing Degradation State Identification Based on LCD Relative Spectral Entropy

Intelligent Ball Bearing Fault Diagnosis by using Fractional Lorenz Chaos Extension Detection Method with Acceleration Sensor

Contact Info

Product

Resources

About