欢迎访问《兵工学报》官方网站,今天是

兵工学报 ›› 2025, Vol. 46 ›› Issue (10): 250411-.doi: 10.12382/bgxb.2025.0411

• • 上一篇    下一篇

一种改进的高效并行爆轰冲击波动力学算法及验证与确认

熊俊*, 代仪军**(), 宫翔飞, 刘鲁峰   

  1. 北京应用物理与计算数学研究所, 北京 100094

An Improved Efficient Parallel DSD Algorithm and Its Verification & Validation

XIONG Jun*, DAI Yijun**(), GONG Xiangfei, LIU Lufeng   

  1. Beijing Institute of Applied Physics and Computational Mathematics, Beijing 100094, China
  • Received:2025-05-27 Online:2025-11-05

摘要:

在均匀二维/三维笛卡尔网格上,数值离散和求解基于爆轰冲击波动力学(Detonation Shock Dynamics, DSD)理论的level set方程,得到的炸药稳定传播爆轰波阵面到达时刻,可用于流体动力学数值模拟的程序燃烧算法。简化和改进了DSD边界条件算法,包括边界节点模板选取方法,及边界节点排序方法。研制了消息传递接口与共享存储多线程混合并行程序DSDLS,可高效求解大规模爆轰问题。利用一系列解析解、半解析解,以及爆轰实验,对算法及程序进行了验证与确认。数值模拟得到的爆轰波走时结果与精确解、实验数据相符。超10亿节点网格大规模并行测试表明,DSDLS程序具有良好的并行效率及可扩展性。

关键词: 爆轰冲击波动力学, 均匀笛卡尔网格, 混合并行, 数值模拟

Abstract:

The equation of level set function based on DSD (Detonation Shock Dynamics) theory is numerically discretized and solved on uniform 2D/3D Cartesian grid. The obtained TOA (Time of Arrival) for the stable detonation front in the high explosive can be used by program burn algorithm in the hydrodynamic simulations. The boundary condition algorithm, including the selection method of boundary node stencils, and the method of boundary node sort, is simplified and improved based on the previous works. The hybrid MPI (Message-Passing Interface) and OpenMP (Open Multi-Processing) parallel code DSDLS is developed for the efficient solutions of large-scale explosive detonation problems. A series of analytic solutions, semi-analytic solutions, and explosive detonation experiments are used for verification and validation of the proposed improved DSD algorithm, which indicate that the detonation front TOA of numerical simulations conform to the exact solutions and the measured data of experiments. The results of large-scale parallel test on the grid of over 109 nodes indicate that DSDLS is of good parallel efficiency and parallel scalability.

Key words: DSD, uniform Cartesian grid, hybrid parallel method, numerical simulation