Optimisation of ship form based on seakeeping behaviour using machine learning

Abstract

The analysis of seakeeping behaviour is the study of the movements and forces produced by waves in marinesystems. This is crucial in naval design, as important parameters such as the operability of the ship, passengercomfort, propulsion performance, manoeuvrability, or the operability of equipment and systems depend on it.Traditionally, it has been analysed by means of tests on hydrodynamic experience channels or with numericalmodels. In recent years, with the development of Artificial Intelligence, research works have been appeared inwhich the use of Machine Learning (ML) techniques have been proposed for the study of the seakeepingbehaviour of ships [1,2]. In this work, a pre-trained Artificial Neural Network (ANN) [3] is used to predictseakeeping behaviour. Due to the speed up in predictions offered, a significant number of ships can be analysedwith a reduced computation time, compared to traditional techniques.The main objective is to search for the geometry best adapted to specific sea conditions and operationalprofiles, optimising specific metrics related to their operability. A ship hull form optimisation will be proposed,by the use of techniques such as Genetic Algorithms (GA) or Particle Swarm Optimization (PSO), linked tothe surrogate ANN solver previously developed. Finally, the most relevant conclusions of the work will beshown.