Leveraging Full-Field Measurement from 3D Digital Image Correlation for Structural Identification

Within the domain of structural health monitoring (SHM) measurement techniques have primarily relied on discrete sensing strategies using sensors physically attached to the structural system of interest. These sensors have proven effective in describing both global and local phenomena, but are limited to providing discrete response measurements of these systems. With the introduction of novel imaging tools and image analysis techniques, such as digital image correlation (DIC), the ability to measure the full-field response of these systems provides a novel approach to refining structural identification (St-ID) approaches used in SHM. This paper explores this proposed concept through a case study on a series of structural test specimens analyzed using 3D digital image correlation (3D-DIC) for St-ID. Finite element model updating (FEMU) was used as the technique for the structural identification. For the identification process, ABAQUS was interfaced with MATLAB to converge on the optimal unknown/uncertain system parameters of the experimental setup. 3D-DIC results provided a rich full-field dataset for the identification process, which was compared against measurements derived from traditional physical in-place sensors typically used in SHM. In this work a Hybrid Genetic Algorithm (HGA), which combines the genetic algorithm as a global optimization and a gradient-based method as a local optimization, was used for the FEMU based on 3D-DIC results of structural specimen subjected to variable loading. To minimize the error between the full field 3D-DIC measurements and FEA model updating results, an objective function was introduced that included the full-field contributions of strains and deformation response. The evolution of this objective function illustrated satisfactory convergence of the identified parameters and the excellent agreement of the experimental and numerical strain and displacement responses after the model updating process confirmed the success of the proposed approach. The results of this study highlight the advantage of this hybrid approach and provide the foundation for effective deployment of the proposed strategy for large-scale structural systems.