Evaluation of residual stress gradients by diffraction methods with wavelets: a neural network approach

Harald Wern; Marc Ringeisen

doi:10.1117/12.339863

8 February 1999 Evaluation of residual stress gradients by diffraction methods with wavelets: a neural network approach

Harald Wern, Marc Ringeisen

Proceedings Volume 3585, Nondestructive Evaluation of Aging Materials and Composites III; (1999) https://doi.org/10.1117/12.339863
Event: Nondestructive Evaluation Techniques for Aging Infrastructures and Manufacturing, 1999, Newport Beach, CA, United States

Abstract

The presence of residual stress gradients is often revealed by x-ray diffraction analysis. Because x-rays always detect an averaged information due to absorption, in the past some approaches have been employed to retrieve the true-z- profiles from the measured (tau) -profiles where (tau) is in general the 1/e information depth of the diffracted intensity. However, all problems which can be described as so called inverse problems like x-ray diffraction analysis are often extremely ill-conditioned. What makes the wavelet basis interesting is that individual wavelet functions are quite localized in space and simultaneously they also are quite localized in frequency. This particular kind of dual localization achieved by wavelets renders operators to be sparse to some high accuracy, when transformed into the wavelet domain. However, the number of required wavelet coefficients for the representation of a residual stress gradients in general still exceeds the number of measurements. therefore, a multilayer feed forward neural network approach has been investigated. With the implementation of a fast backtracking algorithm, suitable learning rates can be achieved. The advantages of this neural network approach, which to the authors knowledge is first introduce in the field of residual stress analysis, will be discussed.

Citation Download Citation

Harald Wern and Marc Ringeisen "Evaluation of residual stress gradients by diffraction methods with wavelets: a neural network approach", Proc. SPIE 3585, Nondestructive Evaluation of Aging Materials and Composites III, (8 February 1999); https://doi.org/10.1117/12.339863

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available