Towards mechanism-based simulation of impact damage using Exascale computing

ORAL

Abstract

Over the past 60 years, the finite element method has been very successful in modelling deformation in engineering structures. However the method requires the definition of constitutive models that represent the response of the material to applied loads. There are two issues. Firstly, the models are often difficult to define. Secondly, there is often no physical connection between the models and the mechanisms that accommodate deformation. In this paper, we present a potentially disruptive two-level strategy which couples the finite element method in the macroscale with cellular automata in the mesoscale. The cellular automata are used to simulate mechanisms, such as crack propagation. The stress-strain relationship emerges as a continuum mechanics scale interpretation of changes at the micro- and meso-scales. Iterative two-way updating between the cellular automata and finite elements drives the simulation forward as the material undergoes progressive damage at high strain rates. The strategy is particularly attractive on large-scale computing platforms as both methods scale well on tens of thousands of CPUs.

Authors

  • Anton Shterenlikht

    University of Bristol

  • Lee Margetts

    University of Manchester

  • Samuel McDonald

    University of Manchester

  • Neil Bourne

    University of Manchester, CMEC, The University of Manchester, Research Complex at Harwell Rutherford Appleton Laboratory, Didcot, Oxfordshire, OX11 0FA, United Kingdom., UMARI, School of Materials, University of Manchester, Oxford Road, Manchester M13 9PL,, School of Materials, University of Manchester, Rutherford Appleton Laboratory, Didcot, Oxfordshire, OX11 0FA, United Kingdom., Manchester X-ray Imaging Facility, School of Materials, University of Manchester, Oxford Road, Manchester M13 9PL, United Kingdom., CMEC, School of Materials, University of Manchester, Rutherford Appleton Laboratory, Didcot, Oxfordshire, OX11 0FA, United Kingdom.