Date of Award

Fall 11-2022

Access Type

Thesis - Open Access

Degree Name

Master of Science in Aerospace Engineering

Department

Aerospace Engineering

Committee Chair

Ali Y. Tamijani

First Committee Member

Alberto W. Mello

Second Committee Member

Jeff R. Brown

College Dean

James W. Gregory

Abstract

Researchers are unlocking the potential of Continuous Fiber Reinforced Composites for producing components with greater strength-to-weight ratios than state of the art metal alloys and unidirectional composites. The key is the emerging technology of topology optimization and advances in additive manufacturing. Topology optimization can fine tune component geometry and fiber placement all while satisfying stress constraints. However, the technology cannot yet robustly guarantee manufacturability. For this reason, substantial post-processing of an optimized design consisting of manual fiber replacement and subsequent Finite Element Analysis (FEA) is still required.

To automate this post-processing in two dimensions, two (2) algorithms were developed. The first one is aimed at filling the space of a topologically optimized component with fibers of prescribed thickness. The objective is to produce flawless fiber paths, meaning no self-intersections, no tight turns, and no overlapping between fibers. It does so by leveraging concepts from elementary geometry and the Signed Distance Function of a topologically optimized domain. The manufacturable fiber paths are represented using Non-Uniform Rational Basis Splines, which can be readily conveyed to a 3D-printer.

The second algorithm then calls a meshing routine to spatially discretize the topologically optimized domain. It takes input from the first algorithm to automatically create and append, orientations and material flags to the spatial elements produced by the meshing routine. Finally, it generates output that is then input to FEA software. The software is written in the C-programming language using the PETSc library. A load case is validated against MSC NASTRAN.

Share

COinS