DocumentCode :
3388507
Title :
Sigma-Sampling Wavelet Denoising for Structural Health Monitoring
Author :
Medda, Alessio ; Chicken, Eric ; DeBrunner, Victor
Author_Institution :
Department of Electrical and Computer Engineering, FAMU-FSU College of Engineering, Florida State University
fYear :
2007
fDate :
26-29 Aug. 2007
Firstpage :
119
Lastpage :
122
Abstract :
Structural Health Monitoring (SHM) techniques are non-destructive evaluation methods that try to detect, locate and assess the structural damage of a structure. The presence of damage in a structure often results in very small changes in the vibration response of the structure. It is very difficult to detect these changes due to their low order of magnitude relative to the vibration signal. A wavelet method is proposed to preprocess the signal prior to damage analysis in order to improve its signal to noise ratio and bring the vibration signal damage signature to a detectable level. Because of the characteristics of the vibration signal, standard thresholding techniques do not yield useful results. Our proposed method estimates the variance from automatically selected, structure-free portions of the vibration response and then uses this value to threshold the wavelet coefficients prior to reconstruction. This novel technique has been called sigma-sampling thresholding.
Keywords :
Background noise; Computerized monitoring; Discrete wavelet transforms; Educational institutions; Noise reduction; Power engineering and energy; Signal processing; Statistics; Wavelet coefficients; Wavelet packets; Array signal processing; Wavelet transforms;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Statistical Signal Processing, 2007. SSP '07. IEEE/SP 14th Workshop on
Conference_Location :
Madison, WI, USA
Print_ISBN :
978-1-4244-1198-6
Electronic_ISBN :
978-1-4244-1198-6
Type :
conf
DOI :
10.1109/SSP.2007.4301230
Filename :
4301230
Link To Document :
بازگشت