摘要
烟雾在高精度网格中与大量障碍物交互十分耗时.为改善障碍物周围的局部视觉细节并进一步提高模拟的实时性,引入多重网格技术与结点分组方法相结合.使用结点分组方法处理障碍物边界条件;在每个障碍物周围,包围一张局部网格,在每个时间步,为每张网格分别求解N-S方程组并在局部和全局网格间传递信息.实验表明,采用文中方法,在障碍物局部模拟精度和单独使用结点分组方法相近的前提下,模拟速度提高了近3倍,达到了更高的实时性.
It is time-consuming to simulate the smoke that interacts with large numbers of obstacles in the high resolution grid.To further improve the real-time of the simulation and retain the local visual details around obstacles,we combine the overlapping grids method with the node-grouping method.The node-grouping method is used to handle the boundary condition of obstacles,and each obstacle is surrounded by a local grid.In each time-step,we solve the N-S equations for each grid and transfer information between the local grid and the global grid.The experimental result shows that our method could obtain competitive local visual details around obstacles with the node-grouping method,and the simulation speed is almost three times faster than that using the node-grouping method only.
出处
《小型微型计算机系统》
CSCD
北大核心
2011年第6期1118-1121,共4页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(60970073)资助