Start Date

2016 12:00 AM

Description

This paper reports on work to explore model based defect characterization methods for NDE of CFRP composites. The work is examining defect responses obtained with ultrasound and thermography, for the purpose of classifying and characterizing the defect through combined analysis of the multiple-mode data. Analysis is premised on the availability of forward scattering models to predict NDE response to specified defects. The approach to defect characterization identifies a set of parameters describing the defect, then optimizes agreement between NDE measurements and measurement predictions through manipulation of defect descriptors, subject to ancillary measures of defect properties imposed to regularize an otherwise ill-posed inversion. Motivation for the project is the detection and characterization of defects in out-of-autoclave large composite structures. Defects of interest in these structures are delamination, excessive porosity, and fiber misalignment. This presentation will summarize the forward measurement models being adapted for this purpose, and will outline the approach being taken to implement well-conditioned data inversion. Examples of application to actual CFRP laminate damage will be presented.

Language

en

File Format

application/pdf

Share

COinS
 
Jan 1st, 12:00 AM

Multimode Model Based Defect Characterization in Composites

This paper reports on work to explore model based defect characterization methods for NDE of CFRP composites. The work is examining defect responses obtained with ultrasound and thermography, for the purpose of classifying and characterizing the defect through combined analysis of the multiple-mode data. Analysis is premised on the availability of forward scattering models to predict NDE response to specified defects. The approach to defect characterization identifies a set of parameters describing the defect, then optimizes agreement between NDE measurements and measurement predictions through manipulation of defect descriptors, subject to ancillary measures of defect properties imposed to regularize an otherwise ill-posed inversion. Motivation for the project is the detection and characterization of defects in out-of-autoclave large composite structures. Defects of interest in these structures are delamination, excessive porosity, and fiber misalignment. This presentation will summarize the forward measurement models being adapted for this purpose, and will outline the approach being taken to implement well-conditioned data inversion. Examples of application to actual CFRP laminate damage will be presented.