Adaptive Crack-Tip Localization of Phase-Field Fracture Computations
Abstract
Phase-field models of fracture provide a unified framework for simulating crack nucleation and propagation, but their computational cost remains high because the governing equilibrium and phase-field equations are typically iteratively solved over the entire domain at every staggered iteration. We present an extension of the recently developed adaptive mesh refinement and coarsening framework that localizes not only spatial resolution but also repeated solution effort to a small active patch around the crack tip for linear elastic brittle materials. The inactive exterior is eliminated through a Schur complement and condensed onto the patch interface, and its factorization is reused as long as the patch remains fixed. The phase field is frozen in the exterior, introducing only a small approximation in the crack wake. The method is implemented in FEniCSx and evaluated for quasi-static crack propagation, dynamic crack branching, three-dimensional mixed-mode crack growth, and coalescence of multiple cracks in a unit cell. Across these problems, the condensed formulation reproduces the reference solutions with differences in global quantities below 0.5%, while substantially reducing the number of active degrees of freedom. In the quasi-static surfing problem, wall time speedups increase from 6.9 to 22.8 as the domain size is doubled, demonstrating increasing computational advantage for large-scale phase-field fracture simulations.
Related
- Coupled Phase-Field Fracture and Discrete Dislocation Dynamics for Short-Crack Propagation
- Adaptive Mesh Coarsening for Efficient Phase-Field Fracture Simulations
- A Phase-Field Study of Desiccation Crack Pattern Maturation under Drying-Wetting Cycles
- Spatiotemporal Adaptive Eigenspectrum Remapping in Transcranial Ultrasound Localization Microscopy
- Matrix-free phase-field modeling of fracture in micromechanical testing simulations of inelastic materials
BibTeX
@misc{dahal2026adaptive,
title = {{Adaptive Crack-Tip Localization of Phase-Field Fracture Computations}},
author = {Aarosh Dahal and Aditya Kumar},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.22888},
eprint = {2610.22888},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.22888},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - UNPB
TI - Adaptive Crack-Tip Localization of Phase-Field Fracture Computations
AU - Aarosh Dahal
AU - Aditya Kumar
PY - 2026
DA - 2026/10/16
PB - arXiv
DO - 10.48550/arxiv.2610.22888
UR - https://arxiv.org/abs/2610.22888
AB - Phase-field models of fracture provide a unified framework for simulating crack nucleation and propagation, but their computational cost remains high because the governing equilibrium and phase-field equations are typically iteratively solved over the entire domain at every staggered iteration. We present an extension of the recently developed adaptive mesh refinement and coarsening framework that localizes not only spatial resolution but also repeated solution effort to a small active patch around the crack tip for linear elastic brittle materials. The inactive exterior is eliminated through a Schur complement and condensed onto the patch interface, and its factorization is reused as long as the patch remains fixed. The phase field is frozen in the exterior, introducing only a small approximation in the crack wake. The method is implemented in FEniCSx and evaluated for quasi-static crack propagation, dynamic crack branching, three-dimensional mixed-mode crack growth, and coalescence of multiple cracks in a unit cell. Across these problems, the condensed formulation reproduces the reference solutions with differences in global quantities below 0.5%, while substantially reducing the number of active degrees of freedom. In the quasi-static surfing problem, wall time speedups increase from 6.9 to 22.8 as the domain size is doubled, demonstrating increasing computational advantage for large-scale phase-field fracture simulations.
ER -