论文部分内容阅读
改进的快速自适应组合网格方法（ＦｌｅｘｉｂｌｅＦａｓｔＡｄａｐｔｉｖｅＣｏｍｐｏｓｉｔｅＧｒｉｄ简称ＦＦＡＣ方法）实质上是利用了多重网格方法原理，用非正常连接法进行处理的区域分解方法。首先ＦＦＡＣ方法由于其网格的加密是在每个网格块内剖分产生的，应用非正常连接的技术来处理相邻网格步长不一致的问题，克服了ＦＡＣ各子域之间不能相互接触的难题。从而也就可以比较容易地对各种驱替前缘用动态加密网格方式进行模拟。其次，ＦＦＡＣ方法则能够根据实际情况而灵活地采用迭代法和直接解法。第三，ＦＡＣ各加密网格层和未加密的网格层是重叠的，这样就增加了网格总数，运行时需要更多的内存。ＦＦＡＣ则用嵌套的方法来处理，可以大大节约内存。第四，ＦＦＡＣ方法中各级网格之间边界值的转换通过在组合网格上求余量来实现而取代了ＦＡＣ方法利用多重网格方法中插值、限制转移的方式，避免了插值、限制转移公式不易确定的难题，大大简化了计算过程。ＦＦＡＣ方法克服了ＦＡＣ方法的不足之处，使得它的实用性更为广泛。这种算法也可用于并行计算。文中阐述了ＦＦＡＣ法的原理和方法，并提供了算例。
The improved fast adaptive combination grid method (FlexibleFastAdaptiveCompositeGrid referred to FFAC method) is essentially the method of regional decomposition using the principle of multi-grid method, which is processed by the abnormal connection method. First of all, FFAC method because of its grid encryption is generated in each grid block, the application of non-normal connection technology to handle the adjacent grid step inconsistency, to overcome the FAC can not interact with each other between sub-domains The problem of contact. Thus, it is relatively easy to simulate various driving fronts with a dynamic encryption grid. Secondly, the FFAC method can flexibly adopt the iterative method and the direct solution according to the actual situation. Third, the FAC’s encrypted mesh layers and the unencrypted mesh layers overlap, thus increasing the total number of grids and requiring more memory at run time. FFAC nested methods to deal with, you can save memory. Fourthly, the boundary value transformation between grids at all levels in the FFAC method is realized by finding the margin on the combination grid instead of using the interpolation method in the multi-grid method to limit the transfer, which avoids interpolation and limitation Difficult to determine the transfer formula difficult problems, greatly simplifying the calculation process. The FFAC method overcomes the deficiencies of the FAC method, making it more practical. This algorithm can also be used for parallel computing. In this paper, the principle and method of FFAC method are expounded and examples are given.