Tesi etd-10222024-092311
Link copiato negli appunti
Tipo di tesi
Corso Ordinario Secondo Livello
Autore
MANNINO, FILIPPO
URN
etd-10222024-092311
Titolo
Gradient-based combined size and shuffling algorithms for non-black metal composite optimisation
Struttura
Classe Scienze Sperimentali
Corso di studi
INGEGNERIA - INGEGNERIA
Commissione
relatore Prof. CIARAMELLA, ERNESTO
Relatore Prof. CINI, ANDREA
Presidente Prof.ssa BOGONI, ANTONELLA
Membro Dott.ssa CREA, SIMONA
Membro Prof. ABENI, LUCA
Membro Prof. ANDREUSSI, TOMMASO
Membro Prof. AVIZZANO, CARLO ALBERTO
Membro Prof. CASTOLDI, PIERO
Membro Prof. MICERA, SILVESTRO
Membro Prof. ODDO, CALOGERO MARIA
Membro Prof. RICOTTI, LEONARDO
Relatore Prof. CINI, ANDREA
Presidente Prof.ssa BOGONI, ANTONELLA
Membro Dott.ssa CREA, SIMONA
Membro Prof. ABENI, LUCA
Membro Prof. ANDREUSSI, TOMMASO
Membro Prof. AVIZZANO, CARLO ALBERTO
Membro Prof. CASTOLDI, PIERO
Membro Prof. MICERA, SILVESTRO
Membro Prof. ODDO, CALOGERO MARIA
Membro Prof. RICOTTI, LEONARDO
Parole chiave
- composite
- lamination parameter
- stacking sequence
- structural optimisation
Data inizio appello
09/12/2024;
Disponibilità
parziale
Riassunto analitico
In this research, the applicability and suitability of combined sizing and shuffling optimization for aeronautical composite structures are investigated. Currently, these two optimization phases are treated separately in commercial software such as Altair OptiStruct.
This thesis aims at integrating shuffling optimization phase into a sizing approach by using fibre orientations as design variables, incorporating manufacturing and damage tolerance constraints. Since no commercial software currently supports this, an integrated approach is proposed, using a Python algorithm for the shuffling phase and OptiStruct for the sizing phase, where ply thickness is the design variable. The commercial software remains essential for solving the displacement field and providing constraint sensitivity, such as maximum strain, relative to all ply thicknesses.
Firstly, two main tasks were completed: the first involved developing a method to translate the sensitivity of the constraint, provided by OptiStruct into the sensitivity of the same constraint in relation to fibre orientation of each single ply. To achieve this, two approaches involving the stiffness matrix, and the compliance matrix were considered. In both cases, the Classical Lamination Theory (CLT), using lamination parameters, was selected as the foundation.
The second task was the selection of an appropriate objective function for the Python algorithm, with the chosen approach focused on minimizing the maximum ratio between the strain of elements shortlisted by the commercial software and the corresponding constraint.
Furthermore, based on this formulation, a potential algorithm for integrating the commercial software analysis with the novel Python optimization algorithm is presented. This integration mitigates the risk of being trapped in local minima and allows for transitioning out of infeasible design spaces.
The possibility of using fibre angles as design variables could be groundbreaking. This work serves as a preliminary step towards the future implementation of angle-based design variable optimization in an in-house finite element method (FEM) solver.
This thesis aims at integrating shuffling optimization phase into a sizing approach by using fibre orientations as design variables, incorporating manufacturing and damage tolerance constraints. Since no commercial software currently supports this, an integrated approach is proposed, using a Python algorithm for the shuffling phase and OptiStruct for the sizing phase, where ply thickness is the design variable. The commercial software remains essential for solving the displacement field and providing constraint sensitivity, such as maximum strain, relative to all ply thicknesses.
Firstly, two main tasks were completed: the first involved developing a method to translate the sensitivity of the constraint, provided by OptiStruct into the sensitivity of the same constraint in relation to fibre orientation of each single ply. To achieve this, two approaches involving the stiffness matrix, and the compliance matrix were considered. In both cases, the Classical Lamination Theory (CLT), using lamination parameters, was selected as the foundation.
The second task was the selection of an appropriate objective function for the Python algorithm, with the chosen approach focused on minimizing the maximum ratio between the strain of elements shortlisted by the commercial software and the corresponding constraint.
Furthermore, based on this formulation, a potential algorithm for integrating the commercial software analysis with the novel Python optimization algorithm is presented. This integration mitigates the risk of being trapped in local minima and allows for transitioning out of infeasible design spaces.
The possibility of using fibre angles as design variables could be groundbreaking. This work serves as a preliminary step towards the future implementation of angle-based design variable optimization in an in-house finite element method (FEM) solver.
File
| Nome file | Dimensione |
|---|---|
Ci sono 1 file riservati su richiesta dell'autore. |
|