Support Vector Machine based bearing defect classification using wavelet denoised vibration signal

Vijay, G S and Kumar, H S and Pai, Srinivasa P and Sriram, N S (2012) Support Vector Machine based bearing defect classification using wavelet denoised vibration signal. NMAMIT Annual Research Journal, 2. pp. 27-32. ISSN 2249-0426

[img] PDF
paper_7_2012__5.pdf - Published Version
Restricted to Registered users only

Download (5MB) | Request a copy

Abstract

Early detection of defects in rotating machinery is of prime importance as it cuts down-time, maintenance cost and avoids possible damage to machinery and human life. Localized defects in bearings like defects on Inner Race (lR), Outer Race (OR) and Ball (B) element cause increase in the energy of the acquired vibration signals. Wavelet based denoising has gained popularity due to its ease of implementation and effectiveness. In this paper, vibration signals acquired from a customized bearing test rig, under one load and two speeds, were subjected to two types of wavelet based denoising. Several time domain and frequency domain features were extracted from the denoised signals. Only a few sensitive features selected by Fisher's Criterion (FC) were used as inputs to train and test a classifier namely, Support Vector Machine (SVM). The SVM performance was found to be comparatively higher when only a few selected sensitive features were used as inputs, than when all extracted features were used. Also, a comparison between the two wavelet based de noising is presented.

Item Type: Article
Uncontrolled Keywords: Rolling Element Bearing, Wavelets, Denoising, Support Vector Machines, Fisher's Criterion.
Subjects: Engineering > MIT Manipal > Mechanical and Manufacturing
Depositing User: MIT Library
Date Deposited: 18 Mar 2013 11:28
Last Modified: 25 Mar 2013 04:57
URI: http://eprints.manipal.edu/id/eprint/79150

Actions (login required)

View Item View Item