Digital image correlation and finite element analysis applied to fiber-reinforced composites at the micro-scale

© 2015 International Committee on Composite Materials. All rights reserved. Digital image correlation (DIC) is a widely employed technique to study deformation of fiber-reinforced composites at the macro and meso-levels. It is, however, rarely applied at the micro-level although deformation analysis of the microstructure is of great interest in the field of composites. This is mainly attributed to practical difficulties related to creation of a nano-scale speckle pattern as well as the experimental set-up. The challenge is to combine DIC with SEM, where imaging involves noise, the gray intensity level of microstructure can be uniform over large areas, and it is difficult to maintain the observation window fixed during loading. The current study aims to address these challenges and enhance the methodology for real-time strain mapping of composites at the micro-scale. A large effort of the work is dedicated to deposition of a high-quality nano-scale random speckle pattern, of which the adequacy is assessed through a strain deviation analysis. The DIC parameters are optimized for the study zone. The methodology is applied to a fiber-reinforced composite loaded under transverse three-point bending inside an ESEM. The innovations in this methodology necessitate its validation against more reliable methods. For this, a finite element model of the study zone is created with boundary conditions taken from the experiment. An acceptable level of agreement was observed between experimentally measured and numerically predicted displacement/strain maps. For the studied case, DIC could correctly capture the mean strain, display the effect of microstructure heterogeneity, show the lowest strains inside fibers, and finally detect some of the strain concentrations in the matrix between fibers. Some small-scale strain concentrations e.g. those in between touching fibers, however, could not be detected. Overall, it was concluded that the methodology can be considered as a promising and dependable technique for strain mapping at the micro-scale.status: publishe