Using an adaptive LMS filter to remove background noise in acoustic monitoring of machines
Date
1992
Authors
Carney, Matthew Scott
Major Professor
Advisor
Mann, J. Adin
Committee Member
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
In order to improve the amount of bearing signal that is measured, the background noise must be muffled or the bearing signal must be increased. This would then improve the ratio of the bearing sound level to the background noise level, which will be defined as the signal to noise ratio (SNR). In a real measurement environment, trying to subdue the background noise with some form of shielding is at best feeble. Even the best reflectors would allow noise in and the measurement would become very intrusive on the measurement site. The bearing signal itself cannot be increased since it is the signal we are trying to measure. Therefore, the noise has to be reduced or removed from the acquired data. This process is known as filtering. If there was an a priori knowledge of the noise sources and how they behaved, then a finite, predetermined filter could be used. However, every different measurement situation and environment has different noise sources and noise characteristics, thus a predetermined filter is not practical. A filter that can adjust to any changes in the system or environment is needed. This type of filter is known as an adaptive filter. More specifically, what is needed is an adaptive filter that will cancel or reduce any noise for any given environment.
The purpose of this research was to implement an adaptive noise cancelling (ANC) routine that could improve the bearing signal to noise ratio in acoustically measured data in order to analyze the bearing signal with greater accuracy. The ANC routine must be able to function properly in reverberant environments, such as mine shafts.
Series Number
Journal Issue
Is Version Of
Versions
Series
Academic or Administrative Unit
Type
thesis