An explicit phase field model for progressive tensile failure of composites

This work is dedicated to a systematic investigation of progressive failure in composite laminate by using a novel phase field model (PFM). In view of the inherent shortcomings of existing implicit solvers for the PFMs that are of less solving efficiency and severe convergence issues, we present in this paper an alternative approach in terms of explicit solvers for the PFMs. We aim to develop a more practical PFM that is effective and able to predict various failure modes in composite laminate. More importantly, the new developed approach is capable of capturing the interaction among failure modes. To achieve this goal, a three-dimensional (3D) complimentary Gibbs energy is introduced to express the potential energy such that distinct failure modes can be characterized separately by the corresponding stress quantities. We reformulate both the crack surface density function and degradation function to characterize different failure mechanisms and to prevent any possible unreasonable or non-physical damage patterns. In addition, we have successfully embedded the widely used Hashin criterion for damage initiation of matrix crack into the PFM, and the mathematical derivations are thus provided. The proposed model performs very well with parallel computing, and the computational efficiency is significantly improved by using multiple threads. The validation shows that different failure mechanisms in the progressive failure of composites such as fiber breakage, matrix cracking, and delamination are well captured and the computed results are in good agreement with the experimental observations.

Automated summary

This work focuses on systematically investigating progressive failure in composite laminate using a novel phase field model. An alternative approach with explicit solvers is presented to address the issues of solving efficiency and convergence in existing implicit solvers for the PFMs. The proposed model effectively captures various failure modes in composites and demonstrates improved computational efficiency with parallel computing.