Research Group:
Speaker:
Lorenzo Fabris
Institution:
SISSA
Schedule:
Tuesday, June 25, 2024 - 14:00
Abstract:
The shipbuilding industry is facing new challenges in hull design, to achieve lower operational and building costs, while adhering to structural stability regulations.Data-driven reduced order modeling has been used to implement a pipeline for the structural optimization of cruise ship hulls, based on proper orthogonal decomposition, Gaussian process regression and Bayesian optimization.We present two improvements: multi-objective optimization based on genetic algorithms, and a hierarchical parameterization refinement procedure based on hull responses.The pipeline has been tested on a simplified hull and a full ship, achieving considerable reduction of the total mass from very coarse initial parameterizations.
