Date of Award

Fall 2022

Document Type

Open Access Dissertation

Department

Mechanical Engineering

First Advisor

Michael A. Sutton

Abstract

In this work, application of DIC in combination with finite element modeling is demonstrated for efficient material characterization for complex heterogeneous materials and tow wrinkle formation during automated fiber placement.

The first part of the work describes the integration of DIC measurements with finite element models for direct property identification in heterogeneous material systems. The material property identification is based on solving partial differential equations (PDE) of equilibrium for unknown elastic material properties with known boundary conditions and full field strain data (Eg. DIC). The PDEs are solved numerically using Petrov-Galerkin finite element procedure. The classical Bubnov-Galerkin method is shown to exhibit spurious oscillations in the result for noisy strain input and for problem involving discontinuous elastic modulus distribution. The formulation is developed for uniaxial applications using basic theoretical constructs, resulting in a computational framework that has a matrix form [A]{E} = {R}, where the [A] matrix components are known functions of measured axial strains and axial positions, components of the vector {R} are known functions of axial body forces, applied loads and reactions and components of the vector {E} are the unknown elastic material properties at discrete locations along the length of the specimen. For a series of 1D material property identification procedure with known axial strains at discrete locations and various levels of random noise, results are presented to demonstrate the accuracy and noise sensitivity of the methodology. Finally, experimental measurements for a heterogeneous bone specimen are compared to our one-dimensional model predictions, demonstrating that the predictions at each load level of interest are in very good agreement with independent estimates at each location along the specimen length.

In the second part, an experimental investigation using StereoDIC to measure out-of-plane wrinkle formation and in-plane deformations in a prepreg slit tape during automated fiber placement (AFP) is described. Based on observations made during the experiments, additional experiments are performed using a rigid double cantilever beam specimen and StereoDIC to measure the Mode I and Mode II traction-separation relationships for tow-to-tow bonding. These were followed by two additional sets of experiments to quantify the viscoelastic response of the tow matrix material.

Results from simulations and experiments for automated fiber placement (AFP) of prepreg slit tape (tow) on a flat surface with different tow path radii of curvature are presented with emphasis on characterization of wrinkle formation. AFP simulations modeled bonding of the slit tape using the "sticky contact" definition in Abaqus and the measured mixed mode cohesive traction-separation relationships. Comparison of the measured and predicted wrinkle shapes, amplitudes and wavelengths to experimental measurements shows excellent agreement for a 6.35 mm wide IM7/8552-1 prepreg tow placed on a flat surface using four different radii of curvature. In addition, the simulations are shown to be capable of capturing the mechanism of wrinkle formation using a generally accepted damage model. Careful inspection of the stress and deformation conditions at the tow-substrate interface under the compaction roller reveals a combination of Mode II and Mode III tractions, with significant damage predicted under the roller due to the mixed mode conditions.

COinS