基于支持向量机的高速滑行艇航态优化  

Sailing State Optimization of High-speed Planning Hull Based on SVM

作　　者：林焰[1,2] 何靖仪 LIN Yan;HE Jing-yi(School of Naval Architecture & Ocean Engineer,Dalian University of Technology, Dalian Liaoning 116024, China;State Key Laboratory of Structural Analysis for Industrial Equipment,Dalian University of Technology, Dalian Liaoning 116024, China)

机构地区：[1]大连理工大学船舶工程学院,辽宁大连116024 [2]大连理工大学工业装备结构分析国家重点实验室,辽宁大连116024

出　　处：《船海工程》2019年第4期1-7,共7页Ship & Ocean Engineering

基　　金：国家重点实验室专项经费(S18315)

摘　　要：引入支持向量机(SVM)分类算法研究滑行艇纵向稳定滑行与纵向运动失稳之间的分界线,验证重叠网格法与Realizable k-ε湍流模型在计算滑行艇阻力及自由升沉纵倾时的准确性,使用该方法对一组以体积弗劳德数Fr▽及重心纵向位置l cg为设计变量的样本点进行数值计算,并使用SVM法对计算结果进行分类,提出基于“偏离度约束”的SVM,使用该改进算法拟合分界线,使用修正后的分界线对重心进行优化,减阻效果可达5.15%,同时可求得在原设计方案的重心下,航速提升至Fr▽=7.16。The support vector machine (SVM) classification method was introduced to study the demarcation line between longitudinal steady sailing state and longitudinal motion instability of planning hull. The accuracy of overset grid method and realizable k-ε turbulence model was verified for calculating ship resistance, free heave and pitch. Numerical calculations of a set of sample points with volume Froude number Fr▽ and longitudinal position of center of gravity l cg as design variables were carried out, and the SVM method was used to classify the calculated results. Meanwhile, an SVM method based on “departure constraint” was proposed, the demarcation line was fitted by using this improved algorithm. The verification results showed 5.15% reduction in total resistance by optimizing the position of gravity center by applying the modified demarcation line, and the design speed can be increased to Fr▽ =7.16 under the gravity center of the original design.

关 键 词：高速滑行艇 重叠网格 CFD SVM 海豚运动 航态优化 

分 类 号：U661.31[交通运输工程—船舶及航道工程] U674.942[交通运输工程—船舶与海洋工程]