Speaker
Description
Phase-field models for brittle fracture have become a widely adopted tool for simulating crack initiation and propagation without explicit tracking algorithms. As their popularity has grown, so has the variety of available implementation strategies, encompassing different solvers, coupling schemes, irreversibility treatments, and software platforms. While each approach has been individually validated in the literature, direct and systematic comparisons remain scarce, leaving practitioners without clear guidance on which strategy best suits their specific problem in terms of accuracy, robustness, and computational cost.
This work addresses this gap through a comprehensive cross-framework benchmarking campaign for the AT2 phase-field fracture model, implemented in the open-source library FEniCSx and the commercial solver Abaqus. In the work, a progressive suite of benchmark problems is proposed to expose the practical strengths and limitations of each approach under controlled conditions. In FEniCSx, the history-variable method is contrasted against bound-constrained variational inequality solvers. In Abaqus, monolithic and staggered coupling schemes are evaluated through a heat-transfer analogy for the implicit solver.
The results reveal that while all strategies converge to framework-independent solutions when properly configured, they differ substantially in iteration count, time-step sensitivity, and robustness during crack propagation. These findings could help researchers and engineers selecting the most reliable numerical tool for their phase-field fracture analyses.