Design a Hybrid Model of Mel-Spectrogram and VGG16 Based Convolutional Neural Network for Diagnosing Motor Bearing Faults
DOI:
https://doi.org/10.5281/zenodo.8071505Keywords:
Rolling bearing, Mel-spectrogram, VGG16, Deep Learning, Predictive maintenanceAbstract
Owing to rolling bearings' significance as one of the most frequently utilized components of industrial machinery. Therefore, it is crucial to establish a system to check the bearing's condition. In this research, a hybrid method of spectral feature extraction combined with Convolutional Neural Network (CNN) designed to classify faults. Firstly, Mel-spectrogram applied as a method for pre-processing by transforming the raw vibration data. Secondly, a Mel-VGG16 employed as a classifier to detect the bearing faults. The proposed technique tested on the CWRU benchmark dataset to bearings in different rotating speeds. The case study results demonstrated that the proposed model could obtain a higher testing accuracy.
References
A. Althubaiti, F. Elasha, and J. A. Teixeira, “Fault diagnosis and health management of bearings in rotating equipment based on vibration analysis–a review” Journal of Vibroengineering, 46-74, 2022.
A. Khadersab, S. Shivakumar, “Vibration analysis techniques for rotating machinery and its effect on bearing faults” Procedia Manufacturing, 247-252, 2018.
M. Sohaib, J. M. Kim, “Reliable fault diagnosis of rotary machine bearings using a stacked sparse autoencoder-based deep neural network” Shock and Vibration, 2018.
R.Yan, R. X. Gao, “Multi-scale enveloping spectrogram for vibration analysis in bearing defect diagnosis” Tribology International, 293-302, 2009.
H. Liu, D. Yao, J. Yang, and X. Li, “Lightweight convolutional neural network and its application in rolling bearing fault diagnosis under variable working conditions” Sensors, 19(22), 4827, 2019.
G. Hong, D. Suh, “Supervised-Learning-Based Intelligent Fault Diagnosis for Mechanical Equipment” IEEE Access, 116147-116162, 2021.
“Bearing Data Center | Case School of Engineering | Case Western Reserve University.” [Online]. Available: https://engineering.case.edu/bearingdatacenter. [Accessed: 04-Apr-2022].
G. Hong, D. Suh, “Supervised-Learning-Based Intelligent Fault Diagnosis for Mechanical Equipment” IEEE Access, 116147-116162, 2021.
S. Lee, H. Yu, H. Yang, J. Yang, G. Lim, K. Kim, and J. Kwon, “Apply VGGNet-based deep learning model of vibration data for prediction model of gravity acceleration equipment”. ArXiv preprint arXiv, 2005.10985, 2020.
Published
Issue
Section
License
Copyright (c) 2022 AINTELIA Science Notes Journal
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
COPYRIGHT NOTICE
Authors submitting a manuscript do so on the understanding that if accepted for publication, copyright of the article shall be assigned to Aintelia® Science Notes Journal (ASNJ).
By submitting their work, authors agree to the following terms:
-
Copyright Transfer: Copyright of the published article is transferred to Aintelia® Science Notes Journal. The journal reserves the right to publish, reproduce, distribute, and archive the work.
-
Licensing: While the journal retains the copyright, the article is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0). This allows third parties to share and adapt the work for non-commercial purposes, provided the original work and the journal are properly cited.
-
Author Rights: Authors retain the right to use their article for their own scholarly needs, such as including it in a thesis or dissertation, presenting it at conferences, or distributing it to students for educational purposes, provided that the journal is cited as the original publisher.
