You are here

On Manufacturing Constraints for Tow-steered Composite Design Optimization

TitleOn Manufacturing Constraints for Tow-steered Composite Design Optimization
Publication TypeJournal Articles
Year of Publication2018
AuthorsBrooks, TR, Martins, JRRA
JournalComposite Structures
Volume204
Pagination548 - 559
ISSN0263-8223
KeywordsAutomated fiber placement, Manufacturing constraints, optimization, Tow steering, Variable angle tow
Abstract

Automatic fiber placement machines have made it viable to manufacture composites where fiber angles vary continuously—tow-steered composites. The additional freedom provided by tow-steered composites has the potential to increase structural performance; however, this approach comes with added design complexity. Numerical optimization can address this complexity, though it is critical that constraints are enforced to ensure that the resulting optimal tow-steered designs are manufacturable using current automatic fiber placement machines. In this work, we consider two manufacturing constraints: tow path curvature and gaps/overlaps. To develop these constraints, we consider a general tow-steered layer pattern as a 2D unit vector field, where the field streamlines represent the tow paths laid down by the automated fiber placing machine. This mathematical formulation provides a relationship between tow path curvature, gap/overlap propagation rate, vector curl, and divergence. These relationships also lead to a constraint on the minimum cut/add length of a tow for a given tow-steered pattern. We demonstrate the developed constraint formulations on two analytical examples, as well as on a structural optimization. We also explore the relationship between curl and divergence of rotated tow patterns. This analysis leads to the conclusion that layups featuring such patterns require strict constraints on bounds to ensure satisfaction of manufacturability requirements. Finally, we use these relationships to motivate a family of gap/overlap-free and curvature-free tow-steered patterns.

URLhttp://www.sciencedirect.com/science/article/pii/S0263822318308158
DOI10.1016/j.compstruct.2018.07.100
Citation KeyBROOKS2018548
Full Text