Digital volume correlation technique for characterizing subsurface deformation behavior of a laminated composite

Abstract

In the present study, the digital volume correlation (DVC) technique is used to study the deformation behavior occurring inside a carbon fibre-reinforced epoxy composite. While a tensile test on a symmetric angle-ply laminated composite specimen is being performed, X-ray images are recorded using synchrotron radiation. The radiographs are processed to reconstruct 3D mu CT images. The reconstruction process utilizes a filtered back-projection (FBP) algorithm combined with a phase retrieval method to enhance the contrast of the CT images. Phase contrast is crucial since the natural microstructural pattern is analogous to a speckle pattern for digital image correlation (DIC) analysis. A slight misalignment of the specimen, typically negligible at the macroscopic scale, is corrected in the CT images prior to the DVC analysis. The accuracy of the DVC results is improved by implementing a sub-voxel registration approach to achieve a resolution less than one pixel. The DVC results are compared with finite element analysis (FEA) results. Although the overall results agree well with the FEA results, few subvolumes are prone to fail to find their correct pairs within DVC volumes. Raw 2D tomograms are individually examined and compared to show that a high correlation coefficient value does not always guarantee a correct match. Detailed discussions about the misleading correlations are provided and thereby the validations and limitations of the DVC (and DIC) technique are addressed.