X-ray-based Computed Tomography (CT) has been linked with finite-element analysis to provide a capability for structural characterization of as-manufactured parts— especially for the nondestructive evaluation of metal-matrix composite (MMC) material parts. In some cases, this capability might eventually obviate costly experiments, including destructive experiments that are traditionally performed to determine mechanical responses. Though developed primarily for MMCs, this capability could also be applied to other types of composites, metal forgings and castings, and plastic components.

Figure 1. This Cross Section of an Eight-Ply MMC Coupon was made by CT with sufficient resolution to show individual fibers. Images like these can be digitized, then processed (1) by image-analysis software to obtain fiber volume fractions and (2) by applied load to obtain stress distribution as shown in (b) and (c).

The basic idea is to utilize data obtained nondestructively to eliminate the need for mechanical testing of a component. In the present approach, one does this by using a finite-element structural-analysis computer program to predict the behavior of the component under load (including, for example, effects of stress concentrations), in combination with information obtained nondestructively by x-ray CT of the component. In the program, local variations in material properties are approximated by differences among the material-property parameters of the finite elements. The values of the finite-element material parameters are, in turn, calculated from such data as local volume fractions of fibers as determined by analysis of CT imagery.

Figure 2. Two Cross-Sectional CT Images of a Composite Ring show imperfections of the Ti-matrix/SiC-fiber core and the Ti cladding.

To link the CT and finite-element-analysis capabilities, it was necessary to develop software to overlay finite-element meshes on CT images, software to manipulate nodes of finite-element meshes to conform with geometries of tomographic slices, and software to classify and manage input and output data pertaining to each finite element. Two case studies were performed to demonstrate the resulting capability. The first study involved MMC test coupons like that of Figure 1. Image-processing techniques were used to segment the fibers from the matrix, then there was created a finite-element mesh, wherein each finite element had unique stiffness properties based on the volume fraction of fibers in that element as calculated from the segmented CT data. The finite-element analysis showed that concentrations of high-fiber-volume-fraction finite elements produced stress concentrations.

The second case study involved a ring comprising (1) a core made of a composite of SiC fibers in a Ti matrix surrounded by (2) Ti cladding and including (3) a damaged area near the outer part of the core. CT cross-sectional images of the ring (see Figure 2) revealed that the core was not uniformly shaped or positioned within the cladding. Several CT cross-sectional images, including the ones in Figure 2, include a bright line across the core that was later determined to represent overlapping of a titanium foil. Variations in the density of the core as shown in the cross-sectional CT images were later correlated with variations in fiber volume fraction. In this study, a finite-element mesh was created to match the ring geometry, and the stiffness parameters in the finite elements in the damaged area were reduced. The results of the finite-element analysis showed that the damaged area could be expected to give rise to stress concentrations elsewhere in the ring.

This work was done by George Y. Baaklini of Lewis Research Centerand Robert N. Yancey of Advanced Research and Applications Corp. Inquiries concerning rights for the commercial use of this invention should be addressed to

NASA Lewis Research Center, Commercial Technology Office, Attn: Tech Brief Patent Status, Mail Stop 7 –3, 21000 Brookpark Road, Cleveland, Ohio 44135

Refer to LEW-16618.


NASA Tech Briefs Magazine

This article first appeared in the September, 1998 issue of NASA Tech Briefs Magazine.

Read more articles from the archives here.