Session: 03-08-01: Computational Modeling and Simulation for Advanced Manufacturing

Paper Number: 119328

119328 - Design for Additive Manufacturing (Dfam) Paradigm in Robotic Manufacturing of Composite Laminates: An Exemplar Problem Using Steered Fiber Paths 

Robotic Automated Fiber Placement (RAFP) technology has revolutionized composites
manufacturing and is quickly replacing traditional hand-layups. RAFP technology manufactures
CFRP laminates by using prepreg slit tapes that are fed into a roller at a constant rate, which
can then be deposited onto a tool surface or on to previously laid down material. RAFP is
suitable for manufacturing large parts and contours, and the benefits of RAFP includes minimal
scrap, reduced cost and increased time efficiency of the manufacturing process. In addition to
these, AFP opens up the space for structural design using optimal steered fibers that can be
derived for maximizing specific structural performance indices. The first constraint in a steered
fiber optimization process is to ensure that the parametrization of the courses shall produce a
continuous path for steering. Additional parameters depend on the machine used for
manufacturing and include- course width, tow width, minimum radius of curvature, tolerance
on gaps and/or overlaps among others. Collectively, all these factors can be termed as
Manufacturing Signatures (MS).
In an automated, additive manufacturing process like RAFP, it is important to consider
the MS in the design and analysis of the parts manufactured. The current work identifies these
MS and incorporate them in the numerical models for the optimization problem of deriving
steered fiber paths for a rectangular plate with an elliptical cutout under remote uniaxial tensile
loading. Though pristine optimal solutions for such a problem have been reported in literature,
most studies neglect the manufacturing aspect of fiber steering using RAFP machines. As an
extension to the previous works by the authors, Bézier curves are used for representing the
centerlines of the fiber paths, which explicitly ensure continuity of the final fiber paths, as well
as provides a robust way to describe the design variables using the Bézier control points. Since
the expressions for a Bézier curve and its local tangent and local radius of curvature is available
in closed form equations, it also enables easy accommodation for modeling course/ path width
as well as machine minimum radius constraints in the design. Further, a concept called global
manufacturing mesh is introduced to help reduce the number of optimization variables further,
but also ensuring that in the numerical design space, only two adjacent courses are allowed to
overlap. Once the optimization problem and the numerical models are established, a surrogate
model is implemented using Radial Basis Functions (RBF). This surrogate model is used in
conjunction with a global optimization algorithm like Genetic Algorithm (GA) to obtain the final
distribution of fiber courses that minimizes the stress concentration around the cutout.

Presenting Author: Avinkrishnan Ambika Vijayachandran University of Michigan

Presenting Author Biography: Dr. Avin Vijay is currently an Assistant Research Scientist at the Department of Aerospace Engineering at the University of Michigan. Dr. Vijay obtained his Ph.D. in Aerospace Engineering from the University of Michigan, focusing on composite structures. Dr. Vijay's current research interests include In-space Additive Manufacturing (ISM), Design Optimization and Design for Additive Manufacturing (DfAM), structural stability, experimental mechanics, and machine learning applications in design and design optimization.

Authors:

Avinkrishnan Ambika Vijayachandran University of Michigan
Anthony Waas University of Michigan