Aerodynamic optimization of a luxury cruise ship based on a many-objective optimization system
Penghui Wang,Fei Wang,Zuogang Chen,Yi Dai
DOI: https://doi.org/10.1016/j.oceaneng.2021.109438
IF: 5
2021-09-01
Ocean Engineering
Abstract:In the multi-objective optimization discipline (MOD), simultaneous optimization with four or more objectives is referred to as many-objective optimization. Compared with the two-objective and three-objective optimization, many-objective optimization brings a series of new challenges, such as the deterioration of the global search ability of an optimization algorithm, the difficulty of the visualization of Pareto solutions, and the increase of the computational burden. To address these challenges, in this study, an efficient many-objective optimization system was proposed, and this system was utilized to improve the aerodynamics of a Vista-class cruise ship at four crucial wind angles. In the process of design optimization, a parametric model with eight design variables was selected as the initial ship form. The uniform design (UD) sampling technique was employed to design a group of transformed ship forms. The Reynolds-averaged Navier-Stokes (RANS) solver was used to evaluate the aerodynamics of transformed ship forms, and the nearest neighbor mesh (NNM) interpolation method was utilized for the initialization process of each numerical calculation to reduce the computational cost of a single simulation. With the numerical results of all transformed ship forms, four aerodynamic surrogate models were established, using a combined method based on a particle swarm optimization and a radial basis neural network (PSO-RBFNN), to replace large-scale numerical simulation. In addition, the Sobol' method was introduced to conduct the sensitivity analysis of the design variables. A series of mature genetic algorithms (GAs) were applied for the single-point, two-point, and four-point optimization of the aerodynamics of a cruise ship in sequence. Additionally, the optimal ship form was selected from the Pareto solutions of the four-point optimization based on the quantitative results of the analytical hierarchy process (AHP). Eventually, the dedicated experimental results of the optimized ship form showed that the aerodynamics at the four wind angles were improved together, confirming the effectiveness of the many-objective optimization system.
engineering, civil, ocean, marine,oceanography