Prediction of residual stresses and distortion of carbon fiber/epoxy composites due to curing process

Loading...
Thumbnail Image
Issue Date
2011-05
Authors
Tavakol, Behrouz
Advisor
Minaie, Bob
Citation
Abstract

The curing process of carbon fiber/epoxy composite parts has always been a challenging task. One of the major concerns is that the final part geometry deviates from the desired, nominal geometry due to the process-induced residual stresses. The consequent distortion may add an extra process for repair or may lead to part rejection; in either case, the extra cost is an important concern. In addition to conventional remedies, which are usually not efficient or capable of addressing the problem completely, many studies have used finite element analysis to predict the final geometry of composite parts. However, shortcomings of such studies limited the use of finite element modeling to only simple shapes. In this study, a methodology was introduced to take into account several mechanisms of residual stress induction during cure in a fully 3-D coupled thermal-curing-mechanical finite element analysis and accurately estimate the final geometry of laminated composite parts. Different modules were developed in Fortran for different mechanisms and material properties. A ply-property approach was utilized in which experimentally measured ply properties were modeled and used. Due to the limitations of current available techniques to obtain pure thermal strain and cure shrinkage during cure, a method was developed and a series of tests were conducted to decouple such parameters during the cure process. To validate the simulation results, a flat square panel was cured experimentally as well as using simulation, a laser scanner was used to obtain the 3-D distortion pattern of the fabricated panel, and a Python code was written to obtain the estimated distortion from the nodal information. Good agreement was observed between maximum distortions as well as 3-D patterns of distortion estimated by simulation and measured experimentally. Finally, 94% reduction in computational costs was achieved using different hardware and software techniques.

Table of Content
Description
Thesis (M.S.)--Wichita State University, College of Engineering, Dept. of Mechanical Engineering.
publication.page.dc.relation.uri