CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Main CFD Forum (https://www.cfd-online.com/Forums/main/)
-   -   Multiblock boundary conditions (https://www.cfd-online.com/Forums/main/3771-multiblock-boundary-conditions.html)

John August 7, 2001 01:19

Multiblock boundary conditions
 
Hi,

I'm thinking of converting a code running on a single-block structured mesh, to a multiblock version using overlapping block boundaries, and know very little about multiblock.

Since the overlap region uses values from the surrounding blocks, does this mean that these boundaries must be treated explicitly rather than implicitly?

John C. Chien August 7, 2001 02:42

Re: Multiblock boundary conditions
 
(1). Assuming that you have a 1-D second-order equation, so, you need two boundary conditions . (one on each end) (2). For a single block formulation, you put in two interior points, and you can solve the equation and obtain the solution for these two interior points. (3). For a two block formulation, you make the left b.c. point, the 1st and the 2nd interior points as the first block, and you also make the 1st, 2nd and the right b.c. point the second block.(4). If you treat the 1st bolck's new b.c. point implicit(unknown), then the 1st block will have two unknowns and one known left B.c., which is not possible to solve the second -order equation in the 1st block. (5). The same argument applies to the second block. (6). So, if you make the new b.c. point of the block-A (which is actually the interior point of block-B) implicit (unknown), then you will not have enough boundary conditions to solve the equation in the block-A. (7). You will have to assume that it is known from the previous iteration solution of block-B.


All times are GMT -4. The time now is 14:53.