Date of Award

Fall 2021

Project Type

Dissertation

Program or Major

Mechanical Engineering

Degree Name

Doctor of Philosophy

First Advisor

Marko MK Knezevic

Abstract

It is undeniable that metals and metallic alloys are of great significance in automotive and aerospace industry. Metallic materials are aggregates of grains (i.e., crystals) in orders of microns. Understanding the behavior of multi-phase metallic structures plays a vital role in automotive and aerospace design and manufacturing industries. Work presented in this dissertation presents a high-performance crystal plasticity solver, capable of performing simulations of massively large data sets in order of few hours instead of months of run. The developed accelerated solver is capable of performing simulations of massively large data sets representing very high-resolution microstructures with tens of millions of grains. Since the developed high-performance solver is capable of simulations of microstructures with limited boundary conditions and regular geometries, simulation of real-world geometries (e.g., aircraft/car body) under any arbitrary loading condition, necessitates a more evolved computational package that brings our recent developments to perfection. The high-performance micromechanical solver is then incorporated into ABAQUS commercial finite element (FE) simulation package as a user material subroutine allowing us to perform precise and efficient modelling, analysis, and design of mechanical components and assemblies for real-world applications. The implementation is the first parallel meso-scale full-field2 coupling of a hybrid CPU (central processing units)-GPU (graphics processing units) full-field micromechanical elasto-visco-plastic fast Fourier transform (EVPFFT) crystal plasticity (CP) model with a full-field implicit Finite element (FE) in ABAQUS standard to facilitate simulation of large plastic deformation of metallic components with arbitrary geometry and loading boundary conditions. The underlying spectral solver takes the advantage of GPU acceleration utilizing Nvidia high performance computing (HPC) SDK compiler (aka PGI compiler) and compute unified device architecture (CUDA) FFT libraries, facilitating computationally efficient simulations of higher microstructural resolutions. The hybrid full-field2 FE-GPU-EVPCUFFT spectral crystal plasticity package presented herein demonstrates exceptional capabilities in accurate predictions of microstructure-property-linkage, facilitating significant contributions to computational design of materials that could significantly further the objectives of automotive and aerospace industries.

Download

Share

COinS