Improved hull design with potential-flow-based parametric computer experiments

Abstract

There is a significant market pull and technology push to incorporate computational fluid dynamics (CFD) in the early vessel design phase. In this phase, one allows for large-impact changes in the hull geometry. The computational cost of CFD deployment in this early phase has two important drivers. Firstly, for an individual CFD simulation, accuracy is correlated with cost—both in man and CPU hours. Secondly, when we increase the number of shape parameters, we face the ‘curse of dimensionality’, an exponential increase of the number of individual simulations that are required to obtain a target accuracy. By applying design and analysis of computer experiments (DACE) methodology, as well as by employing potential flow simulations, we can gather valuable insight in the performance of the vessel within a short amount of time. In this paper, we present the multi-objective automatic shape optimisation of a combatant with eight design parameters, as well as the inverse design of an offshore patrol vessel with seven design parameters.