考虑横风工况的货车导流罩多目标优化设计

Multi-objective optimization design of truck fairing considering crosswind condition

环境横风的作用会增加货车的侧向力,因此,设计一款兼顾纵向空气阻力系数(Cd)和侧向空气阻力系数(Cs)的导流罩装置,对提高车辆经济性和行驶稳定性具有重要意义。以简化的货车模型为研究对象,采用拉丁超立方采样(LHS)方法构建300个数据样本,并使用CFD软件计算值作为对应样本的监测值。将数据样本按7:2:1分为训练集、验证集和测试集,采用分布训练模型预测Cd和Cs,训练模型在测试集上对Cd和Cs的拟合优度(R2)分别达到0.87和0.83。再利用构建好的数据样本进行深度学习训练,并将训练好的BP神经网络模型作为NSGA-Ⅱ的目标函数进行多目标优化求解。从NSGA-Ⅱ得到的Pareto前沿解集中,按照Cd最小、Cd和Cs适中和Cs最小的原则,分别选取验证模型A、B和C,通过与CFD仿真值对比,预测值最大误差为1.9317%。其中:模型B和原始模型相比对,Cd降低了12.7167%;模型C和原始模型相比对,Cs降低了11.9957%。

The effect of environmental crosswind will increase the side force of the truck,so designing a guide hood device that takes into account the longitudinal air resistance coefficient(Cd)and side force coefficient(Cs)is of great significance to improve the economy and driving stability of the vehicle.Taking the simplified truck model as the research object,the Latin hypercube sampling method is used to construct 300 data samples,and the CFD software is used to calculate the values as the monitoring values of the corresponding samples.The data samples are divided into training set,verification set and test set according to 7:2:1.The distributed training model is used to predict Cd and Cs,and the goodness of fit(R2)of the training model for Cd and Cs on the test set reached 0.87 and 0.83,respectively.Subsequently,the constructed data samples are used to train a deep learning model,and the trained BP neural network model served as the objective function for multi-objective optimization using NSGA-II.In the Pareto frontier solution set obtained via NSGA-II,three validation models,i.e.,A,B,and C,are selected based on the principles of minimizing Cd,achieving moderate Cd and Cs,and minimizing Cs,respectively.Com-pared with CFD,the maximum error of prediction is 1.9317%.Among them,model B has a 12.7167%reduction in Cd compared to the original model.In model C,Cs is reduced by 11.9957%compared with the original model.