An Integrated Experimental and Computational Investigation of Fragmentation in Transparent Polymers

Abstract

Major Goals: The major goals of this project are to: 1) advance the capabilities of gradient-damage models to provide robust simulations of fragmentation, with a particular emphasis on representing localized failure surfaces; 2) develop new methodologies to address the stochastic nature of fracture and fragmentation for quasi-uniform loading conditions; 3) construct a relatively simple, quasi two-dimensional fragmentation experiment, allowing for state-of-the-art measurements of stress fields over a range of strain rates, that can be widely employed for model validation. Accomplishments: We made significant progress in this project advancing goals 1) through 3). On goal 1, we have established the ability of cohesive-based gradient-damage models to represent pervasive failure phenomena under quasi-uniform states of stress. These problems are challenging to deal with for continuum-mechanics theories because in reality the bifurcations are driven by defects at length scales that are simply below the level of resolution in the models. This is where goal 2 was essential, because what is needed is to introduce variations in material or fracture properties at the macro-scale. Methods for doing this in a manner that is robust and allows for spatial convergence in quantities of interest are not trivial, however. We developed new approaches in this work that connect variations in material properties to the resulting fracture patterns. The new approach relies on polynomial chaos expansion and a global sensitivity analysis to measure which factors in the variation of the output (i.e. damage fields) are connected to the variation of each input (Young's modulus, fracture toughness) or any combination thereof. For cohesive-based gradient damage models, the studies we performed in this project demonstrated how both variations in critical fracture energies and thresholds for fracture influence fracture patterns.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 30, 2023
Accession Number
AD1224097

Entities

People

  • John Dolbow

Organizations

  • Duke University

Tags

Readers

  • Computational Modeling and Simulation
  • Materials Science (Mechanical Engineering).