Date of Award
Doctor of Philosophy
Engineering Science and Mechanics
Donald O. Thompson
Advances in component life prediction techniques have prompted increased interest in quantitative nondestructive characterization of flaws in engineering materials. Flaw characterization techniques utilize a signature from the fLaw; In ultrasonics, the signature is estimated from noise-corrupted experimental measurements of the scattered acoustic wave field resulting from insonification of the fLaw; Estimating the flaw's signature involves removing the effects of the measurement system in the presence of noise. In the frequency domain, the flaw's signature is called a scattering amplitude. The purpose of this work is to evaluate an optimal Wiener filtering approach to scattering amplitude estimation;The scattering amplitude estimation problem is described with stochastic models in which noise and scattering amplitude are assumed to be random variables. The optimal Wiener filter is derived for the general case where scattering amplitude and noise are assumed to be uncorrelated, Gaussian random variables with known mean and variance. This derivation yields the important result that the filter determines an optimal estimate as the weighted average of the information derived from measurement of the scattered acoustic field and prior information about the flaw distribution. Experimental procedures for measuring flaw signals and for measuring acoustic noise are stated. Noise and scattering amplitude are analyzed as random variables with emphasis on evaluation of the assumptions made in deriving the filter. It is shown that acoustic noise has zero mean and is reasonably uncorrelated and Gaussian. Scattering amplitude is shown to be correlated and non-Gaussian for a lognormal distribution of volumetric scatterers. A novel approach is used to create noise-corrupted flaw signals which utilizes either measured or simulated flaw signals along with measured acoustic noise. The performance of the optimal Wiener filter in determining scattering amplitude estimates from noise-corrupted flaw signals is evaluated and compared to the presently employed on-optimal form of the Wiener filter. The optimal Wiener filter is shown to provide improved scattering amplitude estimates by filtering out acoustic noise and by utilizing prior flaw information;ftn;[superscript]1DOE Report IS-T 1257. This work was performed under contract No. W-7405-Eng-82 with the U.S. Department of Energy.
Digital Repository @ Iowa State University, http://lib.dr.iastate.edu/
Steven Phelps Neal
Neal, Steven Phelps, "A prior knowledge based optimal Wiener filtering approach to ultrasonic scattering amplitude estimation " (1988). Retrospective Theses and Dissertations. 8877.