The double cleavage drilled compression (DCDC) test has been introduced for glass but also extended to test polymer. It is a slender prismatic specimen with a hole in its middle and cracks that are initiated at both poles. It is interesting because the loading is uniaxial compression and the cracks are submitted to opening mode I fracture. It is a simple test to realize but it brings difficulties in terms of analysis. In this work, we have considered DCDC samples of viscoplastic polymer, and applied the loading at a very slow strain rate to neglect the viscosity. The experiments are well reproduced four times, allowing to observe a stress plateau as the cracks grow.
Since the material displays significant plasticity, it was chosen to use a finite element numerical approach to estimate the critical energy release rate. Two approaches were considered the damage phase-field approach that becomes more and more popular since it is easy to implement and does not demand any particular mesh effort, and the cohesive zone elements since it is already implemented in several finite element codes and therefore ready to use. In order to first validate our modeling work, it had been compared to the experimental results obtained on an elastic fragile glass for which analytical solution exist. A first difficulty noticed with the damage phase-field approach is that the possibility to predict damage nucleation that is not realistic. This nourishes the ongoing debate on how damage nucleation should be modeled with phase field (see references in [1]). On this preliminary study, we showed that while the cohesive zone model was able to predict the critical energy release rate (Gc), it was not the case for the phase-field approach. Actually, in the case studied here, the regularization parameter (l) could not be related easily to the material strength, resulting in the existence of several possible couples (l, Gc) to fit the experimental data.
In the case of our elastic-plastic material, it was shown that both approaches were converging toward the same range of values to estimate the strain energy release rate at crack initiation. However, both approaches were not able to correctly reproduce the crack propagation and more work remains to be done here.
The complete study can be found in the paper referenced below, it will help the interested reader to model the DCDC test. It is one of the first study that provides a quantitative comparison between experiments and the phase-field model for elastoplastic material.
This work was part of Arnaud Coq’s PhD.
[1] A. Coq, J. Diani, S. Brach, 2024. Comparison of the phase-field approach and cohesive element modeling to analyze the double cleavage drilled compression fracture test of an elastoplastic material. International Journal of Fracture, 245, 209-222. https://doi.org/10.1007/s10704-023-00755-2.