Reasearch On Hull Form Optimization Based On Support Vector Regression Algorithm

Hull form optimization has always been one of the most important problems concerned by ship industry.Traditional optimization method is high cost,long cycle,and can only obtain “preferred” results,which cannot meet the increasing demand for ship performances.With the deepening of interdisciplinary integration,simulation-based design(SBD)has gradually shown its advantages.In this paper,the key technologies of SBD are summarized in detail.Forcused on the approximation theory,the construction method of surrogate model based on Support Vector Regression is studied in detail.On this basis,a complete hull optimization framework is established.The main researches of this paper are shown as follows:A hull form optimization framework with high-precision numerical prediction method is established: in this paper,the key technologies such as geometric parametric deformation,numerical calculation,optimal algorithm and approximation theory are integrated to build a complete framework for hull form optimization;support vector regression algorithm,which has good generalization and is suitable for small sample set regression prediction,is used to build a proxy model.The trade-off relationship between calculation accuracy and the time cost of optimization is well balanced;In order to improve the calculation accuracy and expand the scope of application of the optimization system,unsteady viscous flow CFD method is used to evaluate the hydrodynamic performance of the ship;The grid division strategy is determined by sensitive study to ensure the reliability of numerical calculation.Moreoverm the calculation accuracy is verified by model test.The thought of optimization is used to construct the surrogate model: The parameters have a great impact on the prediction accuracy of support vector regression algorithm.In order to improve the prediction accuracy of the surrogate model and balance the contradiction between the accuracy of numerical calculation and the efficiency of optimization,the parameters of SVR algorithm are selected more reasonably by optimization calculation in this paper.After establishing the complete optimization framework,the performance and advantages of the optimization method in hull form optimization are verified by the following examples:Taking the classical hull form Series 60 ship as an example,the resistance performance optimization at different sailing speeds was carried out in this paper.By adding the bulbous bow and changing the geometry form of the mid-ship section,the wave propagation in the middle and high-speed section can be restrained and the viscous resistance at both speed section can be reduced.The results show that the optimal design can improve the resistance performance of series 60 ship at all sailing speed.In addition,the influence of sample set capacity on the accuracy of SVR surrogate model is also studied,which verifies the prediction accuracy of SVR algorithm in the case of small sample set,and shows the advantages of SVR algorithm in hull form optimization.Taking the bow thruster as the research target,the thrust efficiency,the fluctuating pressure of the tunnel wall and the differential pressure force of the ship are optimized by changing the chamfer angle and radius of the tunnel inlet and adjusting the position of the pressure compensation tube.After optimization,the fluctuating pressure amplitude is obviously reduced,and the differential pressure of the hull is also balanced.For this kind of time-consuming unsteady problem,although the accuracy loss of the SVR surrogate model is observed due to the extremely low number of samples,it can also predict the influence trend of each design variable on the objective functions accurately.Through two different optimization procedures,the rapidity of the large-scale ship with low sailing speed was optimized.First,a double-objective optimization was carried out focusing on the resistance performance and the wake field quality,and the modified ship shows a slightly improve in resistance performance and significantly improve in wake field quality.Second,a single objective optimization was carried out by taking the received power of the real ship propeller as the optimization objective.Both the coupling of the propeller and the hull and the wake correction due to scale effect are considered.The single objective optimization is performed based on SVR surrogate model After optimized,the receiving power of the optimal ship is reduced by 8.43% and the rapidity performance is greatly improved.The successful practice of the above three optimization examples verifies the significant advantages of the optimization method based on support vector regression algorithm in hull form optimization design,which greatly reduces the optimization time under the premise of ensuring good optimization results;In addition,using the optimization method,optimization examples involving difficult and time-consuming problems such as complex flow or hullpropeller interference are also well solved,the application scope of SBD based optimization method is widened significantly.