湍流DSRFG与LES风场沿计算域流向变化的影响因素

为生成满足桥梁风工程大涡模拟（LES）要求的入口湍流风场，以丹麦大带桥桥址风特性为例，采用离散再合成的随机流动生成（DSRFG）方法合成了满足目标湍流度、积分尺度、脉动风速谱及空间相关性等参数的各向异性湍流；讨论了DSRFG方法在生成湍流风场上关键参数的合理取值；基于Fluent平台，通过自主开发的用户自定义函数（UDF）程序将生成的湍流风场赋给大涡模拟的入口边界，基于LES研究了不同网格尺寸和时间步长取值，入口湍流风场在计算域流向的变化规律。研究结果表明：DSRFG方法能生成满足桥梁LES模拟要求的指定湍流特性风场，产生的风场风谱和速度分量统计值与目标值吻合较好；入口湍流风场特性在计算域流向有较好的维持，脉动风谱在低频段与目标谱吻合较好，高频段出现一定衰减，而衰减起始频率随网格尺寸和时间步长的减小而增大。最后拟合了网格尺寸与脉动风谱衰减起始频率的关系曲线，建议了LES合适的网格尺寸和时间步大小，相关研究结果可为湍流风场模拟和桥梁风工程大涡模拟提供重要参考。

Investigation of Turbulence Generation by DSRFG and Factors Affecting Wind Field Along Computational Domain in Application of LES

The discretizing and synthesizing random flow generation (DSRFG) method to generate a turbulent wind field that satisfies the large eddy simulation (LES) requirement for bridge wind engineering, particularly to study the wind characteristics at the Great Belt East Bridge site, was presented. The DSRFG method to generate anisotropic turbulent flow with turbulent intensity, integral scale, wind spectrum, and spatial coherence in accordance with the specified wind spectrum at the Great Belt East Bridge site, was studied and a suitable selection of the key parameters in the DSRFG method to generate wind field were discussed. The generated turbulent wind was imposed on the inflow boundary of the LES computational domain based on User Defined Functions, in the commercial software Fluent. and the variation pattern of the turbulent wind along the computational domain was investigated under consideration of multiples grid sizes and time step sizes. Results show that a turbulent wind field satisfying the prescribed turbulent characteristics for the LES of flow around a bridge girder can be synthesized by DSRFG, and the statistical characteristics of both the spectra and velocity components agree well with the target values. Furthermore, the characteristics of inflow turbulence are accurately maintained along the computational domain. The simulated spectra of velocity components are in accordance with the target spectrum at lower frequencies. However, the turbulence energy decays at higher frequencies, and it was found that the frequency at which wind spectrum begins to fall increases with decreases in mesh size and time step size. An equation giving the function of the mesh size and the wave length of wind spectrum was provided, and the suitable grid size and time step size for LES were suggested. The results of this investigation provide important information to generate the desired turbulent wind strength when performing LES for bridge wind engineering.