Ship hull shape optimisation to increase energy efficiency of ships

Abstract

The optimisation of a ship hull shape plays an important role in increasing energy efficiency and reducing fuel consumption in maritime transport. This thesis examines three approaches to hull shape optimisation: design of experiments, the adjoint method and machine learning-based methods. The study begins by outlining the fundamental principles of computational fluid dynamics and hull parametrisation, forming the basis for evaluating hull performance and enabling shape modification. In addition, the advantages and limitations of each optimisation method are briefly discussed.

Design of experiments showed to be a useful tool for exploring different hull concepts and identifying important design parameters, making it a useful tool for early-stage hull design. The adjoint method, in contrast, demonstrated its applicability in optimising existing hull shapes by computing the gradient of an objective function. Machine learning-based methods showed significant potential in accelerating the hull design process by reducing reliance on computationally demanding simulations. Findings show that the complementary use of these methods enables an efficient workflow for optimising a ship’s hull shape.