Comparison of Wavelet Based Denoising techniques for Gear Condition Monitoring

Ahmed, Rounaq and Pai, Srinivasa P and Sriram, N S and Vijay, G S (2013) Comparison of Wavelet Based Denoising techniques for Gear Condition Monitoring. In: International Conference on Convergence of Science, Engineering and Management in Education and Research, 26-27 September 2013, Dayananda Sagar Institutions, Bangalore.

[img] PDF
paper14.pdf - Published Version
Restricted to Registered users only

Download (523kB) | Request a copy


Gears are widely used in industrial applications. They are prone to failure due to uninterrupted operations, heavy loads and harsh working conditions. They are the most challenging to identify faults. It is very important that gear faults are identified at an early stage. Vibration signals are widely used in fault diagnosis and condition monitoring of gears. But these vibration signals have lot of noise, because of which useful information is corrupted, that it is impossible to extract useful information from the signals and take proper decisions regarding the condition of the gear. Wavelet Transform (WT) is a latest signal processing tool that can be used to extract features from the noisy vibration signals. This is particularly useful, because the gear vibration signals are non stationary in nature and hence cannot be effectively processed using conventional signal processing techniques like time domain or frequency domain analysis. The wavelet based denoising has proven its ability to denoise signals and images using thresholding. The method consists of decomposing the noisy data into an orthogonal wavelet basis, to suppress the wavelet coefficients smaller than given amplitude and then transforming the data back into the original domain. Accordingly wavelet based denoising techniques have been studied and compared for vibration signals from a gear box. These techniques are effective as they improve the signal-to-noise ratio (SNR) and reduce the root-mean-square error (RMSE). A comparative evaluation of four wavelet based Denoising techniques is discussed. The denoising techniques considered include universal threshold, minimax threshold, heursure threshold and rigorous SURE threshold. The mean squared error (MSE) and Kurtosis have been used to evaluate the performance of these techniques. Different mother wavelet functions have been compared. The results show that with soft thresholding, using universal threshold technique gives the best performance for gear condition monitoring when using Daubechies mother wavelet.

Item Type: Conference or Workshop Item (Paper)
Uncontrolled Keywords: Wavelet Transform; Gear Condition monitoring; Denoising techniques.
Subjects: Engineering > MIT Manipal > Mechanical and Manufacturing
Depositing User: MIT Library
Date Deposited: 02 Jan 2014 07:14
Last Modified: 02 Jan 2014 07:14

Actions (login required)

View Item View Item