论文部分内容阅读
高超声速飞行器的系统仿真是一个多物理场高度耦合的过程,仿真过程中,需要完成大规模数据的迭代计算。目前大多采用分布式仿真的方法,但这种方法对于计算效率的提高不明显。为了有效提高计算效率,基于高性能计算机集群系统,分别对网格并行方法和模型并行算法进行了研究,提出了气动流场高效并行仿真解算方法、分区边界数据交换方法以及并行负载均衡技术,并在高超飞行器的系统仿真中进行了验证,验证结果表明,计算精度满足要求,求解方程的速度较快,工程应用效果较好。
Hypersonic vehicle system simulation is a multi-physics highly coupled process, the simulation process, the need to complete large-scale data iterative calculation. At present, most distributed simulation methods are used, but this method is not obvious for the improvement of computational efficiency. In order to effectively improve the computational efficiency, parallel algorithms and model parallel algorithms are studied respectively based on high performance computer cluster system. The method of efficient parallel simulation of aerodynamic flow field, partition boundary data exchange method and parallel load balancing technology are proposed. The simulation results show that the calculation accuracy meets the requirements, the speed of solving equations is faster and the engineering application is better.