基于本征正交分解的DrivAer快背车非定常尾迹分析

Unsteady Wake Analysis of Fastback DrivAer Automotive Model based on Proper Orthogonal Decomposition

摘要:

在非定常数值模拟的基础上，采用本征正交分解(proper orthogonal decomposition, POD)对DrivAer快背车非定常尾迹进行了研究。改进的延迟分离涡模拟(improved delayed detached-eddy simulation, IDDES)仿真结果与实验结果的对比表明该数值方法有效。对汽车尾迹回流区进行POD分析发现，前2阶模态能量占比分别为6.78%和5.61%，合计占总能量的12.39%，模态系数的频谱分析说明前2阶模态对应同一种低频的拟序流动结构，主频为0.216 (30Hz)；相位分析表明两者相位差为0.455π。对前2阶模态重构的涡量场进行相位平均，发现其对应的拟序流动结构由车体尾迹下剪切层周期性运动主导；前273阶POD模态(能量占比93%)可以很好地重构原流场，极大地降低了自由度。

Abstract:

Based on the unsteady simulation results, proper orthogonal decomposition (POD) was used to analyze the unsteady wake of fastback DrivAer automotive model. The comparison between improved delayed detached-eddy simulation (IDDES) results and the experimental data verifies the effectiveness of simulation method. The POD results of wake recirculation region show that the energy contribution of the first two modes is 12.39% , 6.78% and 5.61% respectively.The spectrum analysis indicates that the first two modes correspond to the same low-frequency quasi-sequential flow structure with a dominant frequency of 0.216(30Hz). Phase analysis show that they have a temporal phase shift of 0.455π. The phase average of the vorticity reconstruction field based on the first two modes reveals that quasi-sequential flow structure is found to be dominated by the periodic motion of the shear layer on the underside of the wake. The first 273 POD modes (93% of the energy) can reconstruct the original flow field well and greatly reduce the degree of freedom.