Image-Based Tomography of Structures to Detect Internal Abnormalities Using Inverse Approach

Image-based techniques have been extensively deployed in the fields of condition assessment and structural mechanics to measure surface effects such as displacements or strains under loading. 3D Digital Image Correlation (3D-DIC) is a technique frequently used to quantify full-field strain measurements. This research uses 3D-DIC to detect interior anomalies of structural components, inferred from the discrepancy in constitutive properties such as elasticity modulus distribution of a three-dimensional heterogeneous/homogeneous sample using limited full-field boundary measurements. The proposed technique is an image-based tomography approach for structural identification (St-Id) to recover unseen volumetric defect distributions within the interior of a 3D heterogeneous space of a structural component based on iterative updating of unknown or uncertain model parameters. The approach leverages full-field surface deformation measurements as ground truth coupled with a finite element model updating process that leverages a novel hybridized optimization algorithm for convergence. This paper presents a case study on a series of structural test specimens with artificial damage. A computer program was created to provide an automated iterative interface between the finite element model and an optimization package. Results of the study illustrated the successful convergence of the selected objective function and the identified elasticity modulus distributions. The resulting updated model at later stages of loading was also shown to correlate well with the ground truth experimental response. The results illustrate the potential to detect subsurface defects from surface observations and to characterize internal properties of materials from their observed mechanical surface response.

Automated summary

This research uses 3D-DIC to detect internal anomalies of structural components and infer material properties based on boundary measurements. The study demonstrates successful detection of subsurface defects and correlation with experimental responses during loading.