• global-local
• global local
• fusoliera
• prandtlplane

In this thesis, an optimization strategy has been developed to design a portion of fuselage structure, for which its geometric parameters are the optimization variables. Finite Elements (FE) models are used to evaluate the behaviour of the structure through a global-local approach, allowing to account for local phenomena on specific limited zones of the model.
The aim of the design procedure is the mass minimization of the fuselage by meeting, simultaneously, both global and local design requirements. The global ones focus on the maximum displacement as well as on technological restrictions imposed on the design variables. The local ones take into account the first linear/eigenvalue buckling load of the most solicited stiffened panel, that is automatically extracted, by means of a specificc criterion, from the global model.
The optimization process, based on Genetic Algorithms, uses then two FE models: a global model where the fuselage geometry is modelled through shells and beams elements and a local model, which is a stiffeened panel whose geometry is suitably extracted from the global model. This features stringers and frames modelled through shell elements in order to highlight local buckling phenomena, and a suitably refined mesh.
To test the validity of design procedure, it was applied to a portion of fuselage structure of a traditional aircraft, whose data are retrieved from literature.
First results provide a feasible configuration that allows a weight reduction of 5.8%.