On the identification of cohesive zone model for curved crack in mortar

Abstract

This paper proposes an approach to defining the path of a curved crack in a single edge notched specimen with gray level residuals extracted from digital image correlation, followed by the calibration of the parameters of a cohesive zone model. Only the experimental force is used in the cost function minimized via finite element model updating. The displacement and gray level residual fields allow for the validation of the calibrated parameters. Last, the results are confronted with those given by a straight crack to highlight the benefits of accounting for the actual crack path.

Crack growth measurement and J‐integral evaluation of additively manufactured polymer using digital image correlation and FE modeling

“…For the second procedure, the tip of the crack is located along a predetermined path, which was straight in the present case (Figure 5). Otherwise, thanks to correlation analyses (with the use of gray level residuals 58,59 ) the crack path can be tailored to its actual topography. The displacement residuals can also be checked to validate a posteriori the a priori choices.…”

On the calibration of cohesive parameters for refractories from notch opening displacements in wedge splitting tests

“…There is a gradual increase of both residuals with time, which is due to the assumption of a straight crack path in a sample with no lateral grooves to guide crack propagation. Although out of the scope of the present work, this shortcoming could be improved by using the gray level residuals to position cohesive elements along the actual crack path [39]. Even with such hypothesis, the RMS difference is on average 25 times the displacement uncertainty (i.e., ≈ 24 µm), which is deemed small given the present conditions.…”

Fracture energy evaluation of refractories in wedge splitting tests from notch opening displacements

“…This result shows that the selected CZM is able to describe accurately the two studied experiments, as additionally proven by the very close experimental and simulated CMOD vs. force responses shown in Figure 11. The small increase of in the monotonic test is due to the fact that the CZM was probed on a larger part of the experiment (i.e., it was stopped for a post-peak force of 5% the ultimate load) in comparison to the cyclic test (70% the ultimate load) and from deviations to the straight crack hypothesis [46,57]. Once the CZM parameters were calibrated, the NOD and CMOD histories were assessed by following the same way as in the experimental analyses, starting off with the monotonic case (Figure 12).…”