 |
Center for Nonlinear Studies |
Computing solid dynamics has issues of stability and accuracy that transcend those in fluids. One of these is numerical and arises from the widely used solution method of radial return to update stress. A second originates in the thermodynamic framework of hypoelasticity and the use of rate dependent constitutive models. I will begin by showing that high strain rate plastic flow is a multiscale phenomenon in the sense that (dimensionless) stress rate can be much larger than the associated strain rate. Then multiple time scale perturbation methods can be used to create alternative stress updates that are stable and accurate while maintaining computational efficiency. An important step in replacing hypoelasticity with a hyperelastic framework is reformulating popular plastic yield criteria in strain space. I will show how the perturbation theory also accomplishes this. This is a talk about resolving current issues in numerics and thermodynamics, which should precede and facilitate the implementation of new material models.
Host: Mark Schraad, T-3: FLUID DYNAMICS AND SOLID MECHANICS, schraad@lanl.gov