基于建筑风场数值模拟的网格层块加密法

在建筑风场的数值模拟中,当前普遍采用的离散网格多是计算前一次性布置的固定网格,通常很难适应实际流场变量的变化要求.为提高数值模拟的精度,基于结构化同位网格系统及控制容积离散微分方程的方法,将适应性网格局部加密(AMR)的思想引入到采用压力校正迭代算法的建筑风场模拟中,提出了一种半自适应的层块网格加密方法.该方法可结合误差分析对误差较大的区域网格实行自动判别并实施逐层块状加密.算例分析表明,该方法能在较高的效益下提高数值解的精度.

A layer-patch grid refinement method for numerical simulation of wind flow around buildings

The grid division format now commonly used in the numerical simulation of wind flow around buildings is a fixed format that was determined and arranged before computation.Such a grid format is often difficult to fit in with the change of wind-velocity and wind-pressure gradients in different airflow regions.In order to improve the accuracy and efficiency of the numerical prediction,a semiself-adaptive layer-patch grid refinement method is proposed based on the nonstaggered grid system and the structured orthogonal grid scheme,with the finite-volume method achieving the mumerical discretization of the differential equations and the pressure correction iterative algorithm fulfilling the solutiom of the discretized equations.Using the method all control volumes that are required to be refined for a certain layer of grid can be automatically distinguished by the error analysis.These volumes are built up manually to form several square blocks/patches and the refinement is then carried out to these blocks.This procedure is repeated to the above refined layer of grid until the final convergence criterion is reached.Numerical examples are presented.The obtained results show that with the proposed refinement method the accuracy and efficiency of the numerical solutions are considerably improved.