翼型失速特性的RANS方法与DES方法数值模拟对比分析

Is DES Approach Better than RANS Approach for Airfoil Stall Simulation?

RAN S(R eyno lds-averaged N av ier-Stokes)加湍流模型是当前计算飞机粘性流场的最常用方法,数值实践说明要计算大分离流动,需要更高级的方法例如LES(Large Eddy S im u lation)或DN S(D irect N S S im u lation)。然而实际雷诺数下,LES和DN S对网格的要求太高,以至目前还难以应用。DES(D etached-Eddy S im u lation)方法结合了RAN S和LES的优点,通过对Spalart-A llm aras湍流模型中长度尺度的修正,在近壁面它体现为RAN S模型的特点,而在远离物面处又保持LES的亚格子模型的特性。论文对比了采用RAN S和DES方法数值模拟翼型失速特性的能力,并与实验结果进行了对比。结果表明,对大分离流动的数值模拟,DES方法体现出更强的能力。

Purpose. DES （Detached-Eddy Simulation） approach is theoretically a good way of taking advantage of the merits of both RANS （Reynolds Averaged Navier-Stokes） and LES （Large Eddy Simulation） approaches. For engineers, it is important to know for a certainty that this is really so. We take the problem of stall and show that DES approach is really good for two of three different types of NACA airfoils. In the full paper, we explain in detail what we did for the problem of stall; here we give only a briefing. Even though LES method is expected for massively separated flow, it is not satisfactory because of the enormous requirement of grid points for practical aviation flow with large Reynolds number. In DES approach, the Spalart-Allmaras turbulence model reduces to a RANS formulation near the wall, and to a subgrid model of I.ES away from the wall. It is found that： （1） before the stall, DES gets almost the same results as RANS for three airfoils; （2） near the stall, for NACA64A-006 and NACA631-012, DES results are much closer to experimental data than RANS; and for NACA633-018, they are almost the same because it is trailing-edge stall type with small separation region when the stall occurs.