Output details

High-resolution non-destructive evaluation of defects using artificial neural networks and wavelets

This paper presents artificial neural network as a direct high resolution method for the interpretation of interfering signals from close multiple defects in a component. Close multiple defects in a component can be particularly difficult to characterise and easily misinterpreted. The analysis was underpinned by experimental signal attenuation tests.

The work was partly funded by an EPSRC research studentship and built on the work carried out by a visiting sabbatical fellow from Instituto Politécnico Nacional in Mexico. The method is preferable to existing maximum likelihood parameter estimation methods and yields very good results even in cases with noisy signal.