Fluid Mechanics Simulation And Hull Shape Opimization Of Yacht Based On CFD

As an important part of the overall design,the hull shape optimization is significance to travel performance and economic practicality of the yacht.The total resistance is an important indicator of the hydrodynamic performance of the yacht.Therefore,the market competitiveness can be greatly improved if the yacht with the least resistance can be designed.Aiming at the yacht body,the hydromechanical simulation analysis and optimization study of the yacht hull shape are carried out to reduce the total resistance in this paper.Main research contents include following parts.Firstly,the 3D model of the yacht is builded,and the boundary conditions is specified after cleaning up the yacht model.The CFD software is applied for numerical simulations of different working conditions,and then the analysis of the flow field characteristics is carried out.Secondly,the optimization of a yacht hull shape is carried out based on CFD considering the total resistance.The design variables are sampled by Latin hypercube sampling method and free-form deformation(FFD)method is applied to change the yacht hull shape.After the geometric reconstruction,the finite element simulation model of the yacht is established and the corresponding total resistance is sampled by CFD simulation.Pearson Correlation Coefficient is applied to analyze the relationship between design variables and optimization objective.The surrogate model is established based on the quadratic response surface model and stacking ensemble learning model,and the grid search strategy is used to optimize the stacking ensemble learning model.The mean absolute error and relative mean absolute error are used as the model evaluation criteria.The prediction abilities of the quadratic response surface model and stacking ensemble learning model are compared under the 5 folds cross validation,respectively.Finally,the established surrogate model is optimized by genetic algorithm and then the optimal results is verified by CFD simulation.Experimental results have verified that the proposed stacking ensemble learning model has better prediction ability than the quadratic response surface model.The optimization results show that the stacking ensemble learning based surrogate model can significantly reduce the total resistance of the yacht by 5.81%,with a high reliability that the relative error is only 0.65%.And while reducing resistance,the yacht’s load capacity is improved by increasing the yacht’s drainage volume by 1.73%.Moreover,compared with the optimization results of the surrogate models established by the first layer of stacking ensemble learning model,the stacking ensemble learning model has a higher prediction accuracy and generalization ability.In this paper,the proposed optimization method based on FFD can be applied to any shape of ships,and more importantly,the accuracy of optimization results is significantly improved by using the stacking ensemble model.The optimization of ship shape is the cross of multi-disciplinary,and various methods of machine learning introduced in this paper will provide more ideas in the future.