高级检索

一种求解汽车外流场问题的可扩展数值算法

A Scalable Numerical Method for Simulating the External Flows around Cars

  • 摘要: 受外型复杂、雷诺数高等因素影响, 汽车外流场流动的数值计算规模巨大且难以精确求解。发展高效并行算法以利用超级计算平台资源来数值求解外流问题成为该领域的研究热点。文章提出一种全隐格式的可扩展并行 Newton-Krylov-Schwarz 算法对某真实汽车的外流场流动问题进行计算。通过与风洞试验以及主流计算流体力学软件的计算结果对比验证了算法的正确性。并行数值计算结果显示, 文章的算法在数千处理器规模下仍具有很好的并行可扩展性。

     

    Abstract: The high-fidelity numerical simulation of unsteady flows around cars are a very challenging computational problem because of the huge computation caused by high Reynolds number and complex geometry. In this paper, we presented a scalable parallel Newton-Krylov-Schwarz based fully implicit algorithm for the full unsteady incompressible flows around cars and compared our results with results obtained from commercial CFD software. Our algorithm shows very good parallel scalability on a supercomputer with over two thousand processors.

     

/

返回文章
返回