Date of Award

Spring 2020

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Mechanical & Aerospace Engineering

Program/Concentration

Mechanical Engineering

Committee Director

Oleksandr Kravchenko

Committee Member

Krishnanand N. Kaipa

Committee Member

Mileta M. Tomovic

Committee Member

Gene Hou

Abstract

Automated fiber placement is a state-of-the-art manufacturing process that allows for complex layup patterns and can quickly place, cut, and restart composite tows. However, with this type of manufacturing process layup defects are inevitable, and manufacturing defects propagate after curing. Process modeling is the considered approach for exploring the defect prediction. Two different but related works were conducted, which are the thermochemical and hyperelastic model and the residual deformation model. For the model before cooling, a hyperelastic model and a thermo-chemical were made to simulate the compaction and heat transfer. Temperature dependent properties that are a function of degree of cure and glass transition temperature were made. An exothermic reaction was simulated with the thermo-chemical model which shows the composite temperature exceeds the applied temperature. Currently, the hyperelastic model shows some behavior in manufacturing induced defect deformation. The residual stress analysis involves two tow-steered composite shells, one with overlapping plies and one without. The coefficient of thermal expansion was measured using digital image correlation and homogenized temperature dependent properties was derived from the results. The stacking sequence, orientation, and thickness was used to create a piece wise stiffness matrix on a simple shell element and varying modeling conditions were applied for a more robust model. The results of the residual deformation show that non-linearity and temperature dependence did not affect the resultant geometry. Out-ofroundness data shows that the shells with overlaps can be simulated through shell element while shells without overlaps did not show good qualitative results.

DOI

10.25777/j2pk-ke03

ISBN

9798641453392

Share

COinS