Noise Analysis and Reduction in a Fiber-Grating-Based Acoustic Emission Detection System

Citation data:

Page: 1-121

Publication Year:
2017
Usage 28
Abstract Views 28
Repository URL:
http://digitalcommons.unl.edu/dissertations/AAI10272532
Author(s):
Hu, Lingling
thesis / dissertation description
Detailed and systematic noise analysis and novel noise reduction methods in fiber-grating-based acoustic emission (AE) detection system are presented in this dissertation. The goal of this work is to provide solutions for the detection of small acoustic emission signals, which is of great interest in practical applications.^ A cost-effective way for interrogation in an AE detection system using fiber Bragg gratings (FBGs) as sensors is to convert the AE-introduced wavelength shift to intensity variations of a laser whose wavelength is set on the slope of the grating’s reflection spectrum. Disturbance in such a sensing system, including the wavelength shift introduced by the components inside the system, and the response of the system to external perturbations are discussed herein.^ The small, cost-efficient distributed feedback (DFB) laser diode is an attractive light source for the sensing system. However, it is also a main contributor of noise. The output wavelength of the DFB laser diode results from the interplay between the injected current and the temperature of the semiconductor junction. The low frequency drift of the laser wavelength caused by thermal and mechanical disturbance can be controlled by using a feedback control system. However, the high frequency wavelength fluctuations caused by the spontaneous emission broadens the laser linewidth and can be converted into large frequency noise by using a frequency discriminator, such as πFBGs. ^ To improve the SNR of the system in a cost-effective way, a noise reduction method using two types of reference channels is proposed. Using a matched πFBG as a reference, which works at the same operating point as the πFBG sensor and is isolated from AE signals, the noise from the light source can be completely subtracted from the sensing signals. The reference channel is improved for use in the environment with a large thermal gradient and mechanical fluctuations by interferometers, which have large detection bandwidth. A novel data processing method was developed for each type of reference to efficiently subtract the laser intensity and frequency noise from the sensor signals. It has been demonstrated that the SNR can be improved more than 20 dB by using the proposed noise reduction methods.^