Application Of Design Space Reduction Technology Based On SOM And Rough Set In Hull Form Optimization

At present,CFD-based optimization technology of ship hull is widely used in the field of energy-saving hull form research and development.In this technique,the CFD numerical simulation is used to calculate the optimization target,and the ship design space is explored by the technique of parametric geometry reconstruction of hull form and the optimization technique.Finally,the optimal hull form meeting the constraint condition is obtained.However,this method often involves a lot of design variables,including intensive computational analysis and simulation process,which leads to too much optimization time and cannot meet the needs of engineering practice.For the above deficiencies,approximate model,efficient optimization algorithm,sensitivity analysis and other technologies have been used in hull form optimization.Although these methods have been successfully applied to ship form optimization,there are still some problems.For example,it is difficult for approximation model technology to establish high precision approximation model when dealing with high dimensional optimization problem,which directly affects the optimization quality.Efficient optimization algorithm technology can reduce the number of optimization iteration,but it is difficult to significantly improve the optimization efficiency.Sensitivity analysis techniques can reduce the dimensionality of optimization problems,but there is a risk of reducing the quality of optimization.In order to improve the quality and efficiency of optimization at the same time,this paper applies the design space reduction technology to hull form optimization,which locates the subspace with better objective performance by data mining.Optimizing in subspace can greatly improve optimization efficiency.Specific research is as follows:1)The basic theory of self-organization mapping is introduced,and the theory is applied to the space reduction technique.The practicability and reliability of the space reduction technique are analyzed by the function example and the KCS container ship wav-making resistance optimization example.2)The basic knowledge of rough set theory is introduced,and some scientific problems of discretization of rough sets are studied.An improved heuristic discretization method based on fuzzy c-means algorithm(FCM)is proposed to make the continuous simulation data have better discretization results.The practicality and reliability of the method are analyzed by some examples.3)Comparing the advantages and disadvantages of two kinds of space reduction techniques based on self-organization mapping and rough set theory,a multi-level design space reduction technique for high dimensional optimization is proposed,which is combined with the advantages of the two methods.Finally,the feasibility of the space reduction technique is verified by the optimization of the total resistance and wake field uniformity of 46000 DWT oil tanker.Through the above research,this paper proposes a multi-level design space reduction method based on SOM and rough set.The method can significantly reduce the design space,it's helpful to improve the efficiency of optimization when focusing on optimization work in the reduced sub-space,which meets the need of the rapid optimization of the ship engineering.