并列双圆柱绕流的动力学模态分解

Dynamic mode decomposition of flow past parallel twin cylinders

本文对双圆柱绕流的卡门涡街现象进行了非稳态数值仿真,使用动力学模态分解方法(dynamic mode de⁃composition,DMD)对仿真得到的圆柱绕流涡量数据进行了分析和流场预测,研究了不同奇异值截断阶数对预测效果的影响。结果表明:利用DMD方法可以较准确地获得非稳态流动过程中涡量数据的模态结构和对应模态的频率、振幅;与稳定的周期阶段相比,用DMD方法分析处于不稳定周期阶段的圆柱绕流流场特征较困难;增加截断的奇异值数目在完善流场特征的同时也可能造成流场还原效果变差。

The phenomenon of Karman vortex street formed by double cylindrical flow was simulated.The vorticity data were analyzed and the development of the vorticity was predicted using dynamic mode decompo⁃sition method.The effect of number of singular values on the flow field prediction was studied.The results show that dynamic mode decomposition can be used to extract modes,dominant modes and amplitudes at dif⁃ferent frequencies.It is difficult to extract the complete flow field characteristics of unstable periodic stages,and increasing the number of truncated singular values will enable to extract more flow field characteristics but worsen the quality of prediction.