摘要
为提高横风下高速列车的气动特性和运行安全性,对列车头型进行多目标自动优化设计。以横风下列车的气动阻力、气动升力和气动横向力为优化目标,提取了6组优化设计变量。运用最优拉丁超立方采样法,并利用网格驱动变形技术快速得到计算样本。通过构造设计变量建立关于优化目标的径向基函数神经网络近似模型,采用遗传算法NSGA-Ⅱ进行多目标优化设计,获得Pareto前沿解集。与初始模型相比,优化后高速列车的气动阻力降低2.12%,气动升力降低7.29%,气动横向力降低3.61%。结果表明,多目标优化设计可显著改善高速列车在横风条件下的气动性能,同时提升其运行的安全性。
A multi-objective automatic optimization design of high-speed train head shape was proposed to improve the aerodynamic performance and the safety of train operation under crosswind.Six groups of design variables were extracted to optimize the aerodynamic resistance,aerodynamic lift,and aerodynamic transverse force of trains under crosswind.The Optimal Latin hypercube design was adopted to sample,and the grid driven deformation technique was adopted to get the sample quickly.By constructing the radial basis function neural network approximation model of design variables about optimization objectives,Pareto frontier solution set was obtained by using genetic algorithm NSGA-Ⅱ for multi-objective optimization design.Compared with the initial model,the aerodynamic drag of the optimized train is reduced by 2.12%,the lift force is reduced by 7.29%,and the transverse force is reduced by 3.61%.The results show that the multi-objective optimization design can improve the aerodynamic performance and safety of high-speed trains under crosswind.
作者
季玲
李瑜
乔巍
JI Ling;LI Yu;QIAO Wei(CRRC Dalian Institute Co.,Ltd.,Dalian 116023,China)
出处
《大连交通大学学报》
2025年第1期57-64,共8页
Journal of Dalian Jiaotong University
关键词
横风效应
高速列车空气动力学
神经网络近似模型
多目标优化
crosswind aerodynamic effect
high-speed train aerodynamic
neural network approximation model
multi-objective optimization
