Study of Hull Shape Optimization Using Self-Blending Method

Abstract Aiming at the problems of large number of geometric deformation parameters and large consumption of computational resources in ship hull optimization, this paper proposes a self-blending method, which extracts the characteristic shape of ship profile from the mother ship and is used to blending to form a new profile, thus realizing the three-dimensional surface deformation of the hull with fewer parameters. CFD method and multi-island genetic algorithm are used to optimize the ship hull of a high-speed ship. As to the ship optimized in the paper, 6 blending factors are used to control the blending of the base cross-section extracted from the original ship, and the hull surface is reconstructed from the transverse section generated after blending, and the three-dimensional surface deformation of the hull is realized under the premise of ensuring the smooth of the hull. The objective function (total hull resistance coefficient) is predicted by CFD method, and the uncertainty of the numerical calculation method is verified. For the optimization part, the response surface is introduced as a surrogate model, and the number of calculation conditions is reduced by more than 90%, which greatly reduces the consumption of computing resources. After optimization, the total drag coefficient of the target ship is reduced by 9.76%, which verifies the effectiveness of the method in hull shape optimization.