GEARBOX FAULT CLASSIFICATION BY USING FREQUENCY BASED FEATURE EXTRACTION

Year 2020, Volume: 21 , 101 - 107, 27.11.2020

Murat Başaran Mehmet Fidan

https://doi.org/10.18038/estubtda.818794

Cited By: 1

Abstract

Gearboxes are the fundamental elements of rotational systems to provide speed adjustment ratios from a rotating power source to another. In industrial applications, the existence of any kind of faults in rotational systems may be hazardous unless the early detection and maintenance procedures are applied. Incipient types of faults such as a few chipped or worn teeth at the gearbox mechanism may deteriorate and cause the maladjustment of the rotation and even the mechanism may stop to rotate which may cause loss of the production. Preventive maintenance strategies such as monitoring of the vibration signals and comparison of the frequency domain irregularities with normal operation case with healthy gearbox elements is essential to ensure safe and accurate rotational speed transmission in industrial systems. In this work, frequency domain characteristics of three different pinion conditions; healthy, a chipped tooth, and three consequent worn teeth are analyzed, and frequency domain features are proposed for classification of the pinion state. Proposed features obtained from the statistical properties of the coefficients of third level Wavelet packet decomposition. After feature extraction process, classification of the gear condition is made with different Support Vector Machine based classifiers and significant classification success observed with the proposed technique.

Keywords

Fault Classification, Gearbox, Preventive Maintenance, Support Vector Machine, Wavelet Packet Decomposition

References

• [1] Kumar RS, Ray K and Kumar KV. Fault Diagnosis of Industrial Drives using MCSA Techniques, Paper Presented at the 2009 International Conference on Control, Automation, Communication and Energy Conservation, 2009.
• [2] Večeř P, Marcel Kreid, and Šmíd R. Condition Indicators for Gearbox Condition Monitoring Systems, Acta Polytechnica vol. 45, no. 6, 2005.
• [3] Zamanian AH and Ohadi A, Gear Fault Diagnosis Based on Gaussian Correlation of Vibrations Signals and Wavelet Coefficients, Applied Soft Computing, vol. 11, no. 8, pp. 4807–4819, 2011.
• [4] Zamanian AH and Ohadi A.Gearbox Fault Detection Through PSO Exact Wavelet Analysis and SVM Classifier, in 18th Annual International Conference on Mechanical Engineering-ISME2010, 2010.
• [5] Akansu AN, Haddad PA, Haddad RA and Haddad PR. Multiresolution Signal Decomposition: Transforms, Subbands, and Wavelets, Academic press, 2001.
• [6] Vapnik V and Lerner AJ. Generalized Portrait Method for Pattern Recognition, Automation and Remote Control, vol. 24, no. 6, pp. 774-780, 1963.
• [7] Aiserman M, Braverman, EM and Rozonoer L, Theoretical Foundations of the Potential Function Method in Pattern Recognition, Avtomat i Telemeh, vol. 25, no. 6, pp. 917-936, 1964.
• [8] Boser BE, Guyon, IM and Vapnik V. A Training Algorithm for Optimal Margin Classifiers, In Proceedings of the fifth annual workshop on Computational Learning Theory, pp. 14.