基于网格缩小的非结构网格梯度重构及制造解精度测试与验证

Investigation on accuracy of gradient reconstruction and flow simulation on general unstructured grids based on downscaling method

  • 摘要: 网格收敛性研究是验证与确认中的重要手段,本文针对一般非结构网格难以自相似加密的问题,实现了基于网格缩小的精度测试方法。在典型的各向同性和各向异性拉伸网格上,利用网格缩小精度测试方法分别考察了梯度重构精度以及制造解流动模拟精度,将网格缩小精度测试结果与传统的网格加密精度测试结果以及理论分析结果进行对比,验证了网格缩小精度测试方法与网格加密精度测试方法及理论分析的一致性。最后将网格缩小精度测试方法应用到各向异性拉伸网格粘性制造解精度测试中,得到了预期中的结果,并初步考察了梯度重构方法、网格类型对数值模拟精度的影响,显示出网格缩小精度测试方法在一般非结构网格精度测试上的优势,具有一定应用前景。

     

    Abstract: With the great improvement in computer technology and numerical methods, computational fluid dynamics has progressed significantly. Even though it is easy and fast to obtain discretized results via numerical simulations, the validity and accuracy of the results need to be carefully verified and validated. Grid convergence study is one of the important approaches in verification process; however, traditional methods cannot realize consistent refinement on general unstructured grids, causing difficulties in grid convergence study on general unstructured grids. In order to solve this problem, the downscaling accuracy testing method (DS method) was adopted in this paper. The DS method was successfully verified by comparing with traditional mesh refinement method and analytical stencil analysis for three cases:a) gradient reconstruction accuracy on isotropic grids; b) gradient reconstructure accuracy on anisotropic grids; c) flow simulation accuracy of a manufactured solution on isotropic grids. The DS method was implemented for accuracy tests on anisotropic stretched grids commonly used in the simulation of viscous flows, and the results were consistent with predictions, which demonstrate the advantage of the DS method on accuracy verification with general unstructured grids.

     

/

返回文章
返回