改进高斯过程的客车侧风工况下气动造型优化研究  

Aerodynamic modeling optimization of bus under crosswind condition based on improved Gaussian process

作　　者：王婷婷[1] 姚铭宽 邵旭 秦东晨[1] 陈江义[1] WANG Tingting;YAO Mingkuan;SHAO Xu;QIN Dongchen;CHEN Jiangyi(School of Mechanical and Power Engineering,Zhengzhou University,Zhengzhou 450000,China)

机构地区：[1]郑州大学机械与动力工程学院,郑州450000

出　　处：《重庆理工大学学报（自然科学）》2024年第4期15-23,共9页Journal of Chongqing University of Technology：Natural Science

基　　金：国家自然科学基金项目(51705468)。

摘　　要：高速客车的行驶安全性受侧风影响较大,为改善高速客车的气动特性和行驶稳定性,提出了一种改进的高斯过程回归模型对客车进行气动造型优化。该模型中,使用一种自动核构造算法,根据数据特征自动构造核函数,采用基于万有引力加速机理的自适应遗传算法对超参数进行优化,克服了传统近似模型精度低的问题。结果表明:改进高斯过程回归模型的预测值均位于95%置信区间,精度较高,实际应用中能够减少仿真和实验的次数,具有较好的工程实用价值。基于协同优化的思想对客车气动造型进行优化,客车的气动升力系数降低了22.56%,侧向力系数降低了18.53%,气动阻力系数降低了4.51%。The driving safety of high-speed buses is greatly affected by crosswind.To improve the aerodynamic characteristics and driving stability of high-speed buses,this paper proposes an improved Gaussian process regression model to optimize the buses’aerodynamic modeling.In this model,an automatic kernel construction algorithm is employed to automatically build the kernel function according to the data characteristics,and an adaptive genetic algorithm based on the gravitational acceleration mechanism is adopted to optimize the hyperparameters.It addresses the problem of low accuracy of the traditional approximate model.Our results show the predicted values of the improved Gaussian process regression model are all high in accuracy,with all in the 95%confidence interval.In its applications,it reduces both the number of simulations and experiments with huge potential in engineering field.Finally,based on collaborative optimization,the aerodynamic shape of the bus is optimized with its aerodynamic lift coefficient down by 22.56%,its lateral force coefficient down by 18.53%,and its aerodynamic drag coefficient down by 4.51%.

关 键 词：客车 稳态侧风 气动造型 高斯过程回归 遗传算法 

分 类 号：U469.1[机械工程—车辆工程]