Element-deletion-enhanced digital image correlation for automated crack detection and tracking in lattice materials

Architected materials can exhibit remarkable combinations of stiffness, strength, and toughness, yet their application is currently limited by an incomplete understanding of how cracks initiate and propagate through their discrete architecture. Elucidating the mechanisms that underpin these processes is challenging because lattice failure is governed by highly localized strains within slender beams, which fall outside the resolution and assumptions of optical methods developed for continuum solids, such as digital image correlation (DIC). Thus, characterizing crack propagation within lattices requires measurement strategies capable of resolving local deformations while accounting for both the intrinsic topological discreteness and the progressive formation of material discontinuities during failure. This work introduces a global DIC framework in which the correlation problem is solved directly on the lattice mesh. Damaged elements are systematically removed based on prominent DIC residual variations resulting from the initiation of dark discontinuities in the speckle pattern. Validations on 3D-printed regular and imperfect triangular lattices under mode-I loading demonstrate that the approach accurately captures damage initiation and tracks propagation alongside macroscopic load variations. Introducing the damage criterion to track damage within regular triangular lattices lowers global residuals by 40% at the end of the test, confirming the method validity. Notably, the final number of automatically detected broken struts overestimates the manual visual estimate by only three struts, a minor discrepancy within the uncertainty margin of visual observation. Ultimately, we propose an alternative element-removal approach based on a critical failure strain, which can be either pre-defined or estimated directly by the residual-based method.