An adaptive H-refinement algorithm for local damage models

Jonathan S. Pitt, Francesco Costanzo

Research output: Contribution to journalArticlepeer-review

1 Scopus citations

Abstract

An adaptive mesh refinement strategy is proposed for local damage models that often arise from internal state variable based continuum damage models. The proposed algorithm employs both the finite element method and the finite difference method to integrate the equations of motion of a linear elastic material with simple isotropic microcracking. The challenges of this problem include the time integration of coupled partial differential equations with time-dependent coefficients, and the proper choice of solution spaces to yield a stable finite element formulation. Discontinuous elements are used for the representation of the damage field, as it is believed that this reduction in regularity is more consistent with the physical nature of evolving microcracking. The adaptive mesh refinement algorithm relies on custom refinement indicators, two of which are presented and compared. The two refinement indicators we explore are based on the time rate of change of the damage field and on the energy release rate, respectively, where the energy release rate measures the energy per unit volume available for damage to evolve. We observe the performance of the proposed algorithm and refinement indicators by comparing the predicted damage morphology on different meshes, hence judging the capability of the proposed technique to address, but not eliminate, the mesh dependency present in the solutions of the damage field.

Original languageEnglish (US)
Pages (from-to)1281-1300
Number of pages20
JournalAlgorithms
Volume2
Issue number4
DOIs
StatePublished - Dec 2009

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Numerical Analysis
  • Computational Theory and Mathematics
  • Computational Mathematics

Fingerprint

Dive into the research topics of 'An adaptive H-refinement algorithm for local damage models'. Together they form a unique fingerprint.

Cite this