Concurrent Optimization for Laminates with filtered manufacturing constraints
N.B. Fontenai (TU Delft - Aerospace Engineering)
S.R. Turteltaub – Mentor (TU Delft - Aerospace Engineering)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
The composite additive manufacturing industry keeps growing every year, alongside the need for more lightweight, problem-specific structural components in EV, wind, space, sports, and other industries. This need has led to a great deal of research on the optimisation of composites, as they offer far more lightweight design possibilities and are more problem-direction-specific, which makes them more efficient. This, however, comes with some difficulties that are addressed in this thesis, namely the manufacturability of such designs.
The thesis combines the SIMP method for topology optimisation with an ANDFO approach to fibre orientation optimisation methods. The SIMP method provides a well-defined variable parameter that can be used to retrieve a manufacturable design without grey areas when implemented with a power scale, as in this thesis. The ANDFO, on the other hand, uses normal distribution weight functions to guide the fibre orientation towards a discrete value without requiring multiple design parameters per layer.
The model is initially compared for its manufacturability efficiency against other optimisation methods, such as level-set or Heaviside methods, as well as against lamination parameters and other methods for laminate design.
The model, which works with a mix of heuristic and concurrent combination methods, is shown to have verified results when tested against existing literature and demonstrates some improvement in both manufacturability and compliance levels. The randomness test, however, still shows that even with the use of a GCMMA solver, the non-convexity of the problem causes the randomised initialisation point to yield different results. This effect is mainly observed in the fibre orientations, as they are more susceptible to such variations.