【摘要】 针对可压缩流动数值模拟的加速需求，在作者前期提出的扰动域推进计算框架下，融入多重网格加速技术，提出了多重网格-扰动域推进方法。建立了粗、细网格上动态计算域的更新逻辑与高效算法，并针对多重网格方法的特点，对粗网格初始化、增大对流动态域等步骤进行了改进。设计了新型粗网格生成策略，可消除结构网格粗网格生成对单元数目的限制。典型算例验证表明，多重网格-扰动域推进方法可同时降低粗、细网格上的计算量并减少总迭代步数；通过在粗、细网格上同时采用随数值扰动传播而增大、随求解收敛而缩小的动态计算域，多重网格-扰动域推进方法可在传统多重网格方法加速的基础上再将计算效率提升57.9%。更多还原

【Abstract】 To accelerate the numerical simulation of compressible flows, a new acceleration methodology, named the Disturbance Region Update Method with Multigrid(DRUM-M), is presented, which integrates the multigrid technique into the Disturbance Region Update Method(DRUM) proposed by the authors. The principles and algorithms of updating the Dynamic Computational Domains(DCDs) between the fine and the coarse grids are proposed. Improvements on certain operations, such as the initialization and the extension of the advective DCD, are made based on the characteristics of the multigrid technique. Besides, a new strategy of the coarse grid generation is established for structured grids, capable of eliminating the cell-number restriction of existing methods. Numerical test cases demonstrate that, firstly, DRUM-M can decrease the computational effort per iteration on both the fine and the coarse grids, as well as the total number of iterations; secondly, benefiting from employing DCDs on both the fine and the coarse grids, there are acceleration synergies between the multigrid technique and DRUM; thirdly, compared with the conventional multigrid technique at the same conditions, DRUM-M could achieve a time reduction of 57. 9%.更多还原