论文部分内容阅读
通过求解RANS方程和Menter’s k-Omega SST两方程湍流模型,以及采用多重网格加速收敛技术、基于多块结构网格的通用数据传输方法和区域分解负载平衡技术,实现CFD软件的并行计算。在国家超算长沙中心的“天河”系统上完成了软件的移植、测试,并实现翼身组合体外形的2048处理器核数、网格规模上亿单元的大规模并行计算,并行效率达到48%,较大幅度地缩短了计算周期,提高了工作效率。通过对DLR-F6的模拟,在气动力系数精确求解、超大规模网格模拟的快速收敛和网格收敛性研究等方面取得了初步结果,为下一步大规模工程实际应用打下了坚实基础。
The parallel computation of CFD software is realized by solving the turbulence model of RANS equation and Menter’s k-Omega SST two-equation turbulence model, as well as the multi-grid acceleration and convergence technology, the common data transmission method based on multi-grid structure and the regional decomposition load balancing technology. Completed software migration and testing on the “Tianhe” system at Changsha National Supercomputer Center in Changsha and achieved large-scale parallel computation of 2048 processor cores with wing body assembly shape and massive cells with millions of cells on grid scale. The parallel efficiency Reached 48%, greatly shorten the calculation cycle, improve work efficiency. Through the simulation of DLR-F6, preliminary results have been obtained in terms of the exact solution of the aerodynamic coefficient, the fast convergence of super-large-scale grid simulation and the study of grid convergence, laying a solid foundation for the next large-scale engineering application.