弯曲壁面流动分离及再附的尺度解析模拟

Scale-Resolving Simulations of Flow Separation from Curved Surfaces and Subsequent Reattachment

关键词:  流动分离  RANS/LES混合模拟  应力混合涡模拟  分离涡模拟  周期性山丘流动

Key words:  Flow separation  Hybrid RANS/LES modeling  Stress-blended eddy simulation  Detached eddy simulation  Periodic hill flow

摘要:

为准确预测弯曲壁面发生的流动分离及再附现象，基于模拟应力混合思想构建了新的RANS/LES混合框架，涡黏系数和湍动能耗散项中的长度尺度同步切换、涡解析区动态确定亚过滤尺度模型系数、采用混合过滤尺度匹配动态模型，并据此构造了k-ω基动态应力混合模型。首先利用槽道湍流检验了基准的亚过滤尺度模型和近壁RANS模型，前者解析获得了正确的湍流脉动近壁行为和相干结构，后者以很小代价捕捉到了雷诺应力分量的各向异性和松弛效应。随后利用周期性山算例考核了动态应力混合模型对分离再附湍流的预测精度，并与DDES，IDDES及SAS模型进行了对比。结果表明，三项改进使得动态应力混合模型具有近壁模拟能力且不存在对数律不匹配问题，涡解析能力优于DES类隐含的准代数模型，能够更好地捕捉压力梯度诱导的流动分离和再附，获得更为准确的壁面剪切应力、摩擦系数、压力系数、回流区，分离剪切层相干结构的分辨率更高。

Abstract:

In order to accurately predict the flow separation from curved surfaces and subsequent reattachment， a novel paradigm in hybrid RANS/LES modeling was built on stress-blended eddy simulation. Therefore， a concrete k-ω dynamic stress-blended model was also proposed based on three improvements including synchronous switching of length scales in eddy-viscosity and turbulent kinetic energy destruction term， determining dynamically the model constants in eddy resolved region， and hybrid filter-size matching dynamic model. Firstly， basal subgrid-scale model and RANS near-wall treatment were verified by turbulent channel flow， which reveal that the former could obtain correct near-wall behavior of turbulent fluctuation and coherent structures， and the later could capture anisotropy and relaxation of Reynolds stress at little cost. Secondly， the flow over periodically arranged hills were used as benchmark case to validate the effectiveness of the proposed model compared with DDES， IDDE and SAS models. In a word， dynamic stress-blended model with three improvements proved to be an alternative approach for wall-modeled LES without logarithmic-layer mismatch， as well as yield better LES resolved capacity than quasi-algebraic model in DES class. As a result， the proposed model provides exact prediction of a flow phenomenon that involve separation and reattachment induced by pressure-gradient， also obtains more precise wall shear stress， friction coefficient， pressure coefficient， recirculation zone， as well as separated shear layer with more local coherent structures.