【摘要】 在SUV产品风阻仿真开发中，雷诺平均方程（Reynolds-Averaged Navier-Stokes,RANS）因其高效性而广泛应用，但现有湍流模型缺乏对复杂分离流动的标定，存在对部分仿真工况分离流动捕捉效果较差、仿真精度不足的问题。本文采用条件流场反演结合符号回归的方法，以shear-stress-transport(SST)湍流模型为基础，以类SUV外流场分离现象的二维经典分离流算例为训练集，构造适用于SUV车型风阻仿真的数据驱动湍流模型，并分别在电动、燃油SUV车型上进行验证，经与试验结果比较，修正模型仿真结果与试验值相对误差<2%。更多还原

【Abstract】 For the SUV product wind resistance simulation development, the Reynolds-Averaged NavierStokes(RANS) equation is widely used due to its high efficiency. However, the existing turbulence models lack calibration for complex separated flows, resulting in poor capture of separated flows and insufficient simulation accuracy for some simulation conditions. In this paper, a data-driven turbulence model suitable for wind resistance simulation of SUV models is constructed based on the shear-stress-transport(SST) turbulence model by using the method of conditional flow field inversion combined with symbolic regression, with the two-dimensional classical separated flow example of the SUV-like external flow field separation phenomenon as the training set. The model is verified on electric and fuel SUV models respectively. Compared with the test results, the relative error of the simulation results of the modified model is less than 2% compared with the test values.更多还原