Process Optimization of Composite Flex Beams Using Neural Networks

Abstract

Large scale, intricate shape parts with widely varying cross-sections find difficulty in manufacturing using conventional manufacturing processes. Cavity molding, a process similar to compression molding is unique and often used in rotorcraft industries to manufacture thick flex beam composite parts with high viscosity resins. The curing of thermosetting resins used in these applications is usually accompanied by an exothermic reaction and excessive heat buildup during the polymerization reaction can cause internal stresses to build-up and result in structural defects in thick composite structures. Process optimization of flex beam composite parts manufactured using the cavity molding process will result higher quality parts. In this work, a three-dimensional coupled cure and flow model is developed to investigate the cavity molding process. Comsol Multiphysics, a finite element tool with multiphysics capabilities is used to study the coupled cure and flow phenomenon. Finite element simulations are performed to predict resin flow, resultant temperature distributions, and degrees of cure at various cross-sections. Neural networks based models are developed and integrated with finite element modeling to optimize the cavity molding process, and the optimal process parameters are suggested. Copyright 2013 by Aurora Flight Sciences.

Meeting Name

SAMPE 2013 Long Beach, California May 6-9

Department(s)

Mechanical and Aerospace Engineering

Keywords and Phrases

Composite Parts; Comsol Multiphysics; Conventional Manufacturing; Finite Element Simulations; High Viscosities; Polymerization Reaction; Rotorcraft Industry; Structural Defect

Document Type

Article - Conference proceedings

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 2013 Society for the Advancement of Material and Process Engineering (SAMPE), All rights reserved.

Publication Date

01 Jan 2013

This document is currently not available here.

Share

 
COinS