Numerical and Parametric Optimization of Ship Hull Form

Abstract

Optimization has become a crucial part of modern design practices, with advancements in numerical methods enabling designers to analyze multiple designs and identify the best options. However, the iterative nature of optimization can be time-consuming and computationally demanding, especially for complex designs with numerous criteria and constraints. This challenge is particularly evident in hull form optimization, where design possibilities are limitless. To address this issue, a surrogate-based optimization strategy is proposed. This approach can replace expensive numerical simulations with an equivalent surrogate (or meta) model, trained using data from a limited set of initial simulations. The model can be further refined using suitable optimization algorithms. Implementation of this strategy however, requires seamless interaction among various tools integrated within a common framework. This thesis presents the development of such a framework for hull form optimization. It integrates parametric designs, numerical methods, surrogate models, and optimization algorithms under a common Python-based framework to streamline the optimization process.