CUSATIS COMPUTATIONAL SERVICES INC. — Department of Defense STTR Phase I: A21C-T020

Description

This project focuses on the development of computational tools for the multiscale simulation of engineered wood products under impulsive loading conditions resulting from blasts and impacts. The proposed computational tools are based on a multiscale model for wood that simulates failure mechanisms at the length scale of the wood cellular structure. The adopted fine scale model, entitled Connectors & Beams Lattice (CBL) model, simulates wood fibers with nonlinear Timoshenko beams transversely connected with elasto-fracturing connectors. The constitutive equations governing the behavior of both beams and connectors account for strain rate effects. The rate dependence of wood behavior at the meso-scale is assumed to be caused by two different physical mechanisms. The first is a dependence of the fracture process on the rate of crack opening, and the second is the viscoelastic deformation of the unfractured material. In this study, the first mechanism is described by the activation energy theory applied to the ruptures occurring along the crack surfaces, and the second mechanism is modeled by classical viscoelastic theories. The developed model will be calibrated and validated against experimental data gathered from the literature. These include compression and fracture tests performed at various strain rates.  In parallel with the model development, the project will pursue computational tool development, which will aim at both supporting government agencies as well as entering the market with a novel computational product. In Phase I, the research team will implement the strain-rate dependent CBL model in a plug-in for Abaqus. Subsequently, a feasibility study will be conducted to assess opportunities and challenges associated with the implementation of the developed computational framework in plug-ins for other commercial software packages, such as Ansys LS-DYNA and Ansys Mechanical, and government supported software packages, such as EPIC, DYNA3D, and MOOSE. In addition, the research team will explore the deployment of the plug-ins in governmental High Performance Computing (HPC) clusters and commercial cloud platforms. The cloud deployment, which will be pursued fully  in Phase II, will allow the seamless integration of the newly developed model with pre-processing and post-processing tools and will facilitate the solution of practical problems through the automation of time-consuming input-output tasks.