Multidisciplinary optimization of an offshore aquaculture vessel hull form based on the support vector regression surrogate model

The shape of a ship's hull has a significant impact on the resistance and propulsive performance of the ship, especially for wide ships. In this study, experimental and numerical analyses of an offshore aquaculture vessel were carried out, while multidisciplinary design optimization aimed to improve the resistance performance and the wake field quality of the vessel. Numerical analysis was performed using a Reynolds-averaged-Navier_Stokes (RANS) solver, and the numerical method was validated by comparing the numerical results from the original shape to the experimental results. Three support vector regression (SVR)-based surrogate models with different numbers of sample points were built, and the influence of the sample set size on the prediction precision was determined. An SVR surrogate model with 60 sample points was applied and optimized to replace direct numerical simulations. An optimization framework using Latin hypercube sampling (LHS), the free-form deformation (FFD) method, the SVR surrogate model, and the non-dominated sorting genetic algorithm II (NSGAII) was developed. Nine design variables were employed to modify the shape of the vessel. Because of the trade-off between the minimum resistance and the minimum circumferential nonuniformity of the wake flow, the optimal solution of the hull shape was selected from the Pareto-optimal solutions to balance the two objectives. Model tests for the optimized shape were then performed to validate the design results. The results showed that the resistance and the circumferential nonuniformity of the wake flow of the optimized shape were reduced by 1.59% and 17.80%, respectively, when compared to the results of the original shape.