Advances and Challenges in Computational Modeling of Dynamic Material Failure: From Single to Multi-Scale Simulations and Their Industrial Applications

Seminar | February 6 | 2-3 p.m. | 3110 Etcheverry Hall

 Dr. C.T. Wu, Livermore Software Technology Corporation (LSTC)

 Department of Mechanical Engineering (ME)

Numerical modeling of material failure remains a formidable challenge to the computational mechanics community. Apparently, the pure continuum-based numerical methods are not able to accurately predict the material failure takes place at the finer scale. In other words, the C1-continuity assumption in most finite element methods is inadequate to describe the kinematic discontinuity of displacement fields at the macro-scale for material separation analysis. In this talk, three numerical methods exhibiting strong displacement discontinuity for single-scale material failure analysis are introduced to bypass the numerical limitations of conventional finite element methods. Smoothed Particle Galerkin (SPG) method, Peridynamics and Extended Finite Element Method (XFEM) are presented to model fracture in solid, composite and shell respectively. On the other hand, while most multi-scale methods are focused on the models of intact materials, their application in material failure analysis is very rare and has become an emergent need for industrial applications. In contrast to the single-scale approach, the multi-scale approach encompasses the material failure modeling with better scale information and relevant physics. In this talk two multi-scale methods are introduced to model the material failure in specific industrial applications such as the joint failure in crashworthiness and the failure in composite materials.

 zohdi@berkeley.edu, 510-642-9172