a question on FVM
 

hello,

iam a student and recently completed my project at NAL,bangalore,INDIA. i worked on a code called RANS2D developed by a scientist there.it employed FVM and i have some general questions pertaining to it ; 1. Why is there large numerical diffusion error using UPWIND discretisation scheme? 2. what does the term "blow off" in a code mean? 3. How are boundary conditions in the grid adjacent to a wall employed? for eg in a circular cylinder with a body fitted radial polar grid how is the wall condition of "no slip" incorporated in the discretized equation?

Re: a question on FVM
 

(1). It seems to me that you need some basic knowledge about numerical methods and fluid dynamics. Why not pick some books from the forum/books section first and do some reading. (2). If you are using first-order upwind, then it is less accurate than the second-order central difference. This accuracy issue can be interpreted as the added diffusion effect. It has been discussed in the previous messages here. (3).I don't know anything about the term "blow-off" in the code. Could you explain the situation when the term is used? Perhaps, it is related to the "divergence" of the iteration process. In other words, the iteration process fails to converge. (divide by zero also can cause the problem.) (4). "no slip" condition means that the velocity at the wall is zero, which is the consequence of the viscous fluid flow. Since the grid is body-aligned, you can set the velocity along the wall mesh line to zero. (5). I think, you are still outside the CFD world. You will have to study the CFD technology piece-by-piece in order to get some feeling about CFD issues.