Indian Journal of Science and Technology
Year: 2015, Volume: 8, Issue: 26, Pages: 1-7
V. Shanmukha Priya1*, P. Mahalakshmi2 and V. P. S. Naidu3
1 Department of Mechanical Engineering, NITK, Surathkal - 575025, Karnataka, India; [email protected]
2 School of Electrical Engineering, VIT University, Vellore - 632014, Tamil Nadu, India; [email protected]
3 Multi-Sensor Data Fusion (MSDF) Lab, CSIR-NAL (National Aerospace Laboratories), Bangalore - 560037, Karnataka, India; [email protected]
Background/Objectives: Condition monitoring is one of the important functions to be carried out in the maintenance of any machine. In condition monitoring, there are several techniques among which the most commonly used technique for rotating machines is the vibration analysis. Methods/Statistical analysis: Discrete Wavelet Transform is used to decompose the vibration signal into 9 levels. For each level, mean ±std (standard deviation) are computed for both approximated and detailed coefficients. Findings: Bearing data obtained from the bearing test rig of Case Western Reserve University are used to test the algorithm. The standard of coefficients in level to 3 shows distant classification of faults. The levels which show clear classification among the bearings are those frequency bands in which the characteristic defect frequencies of faults occur. It is inferred that, the wavelet decomposition classifies the ball defect clearly than the frequency domain methods. Application/Improvements: Wavelet based bearing health condition monitoring technique can be used for bearing fault diagnosis and it can be extended for prognosis.
Keywords: Bearing Health Diagnosis, Condition Monitoring, Failure Analysis, Vibration Analysis, Wavelet Decomposition
Subscribe now for latest articles and news.