CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   BFS...again (http://www.cfd-online.com/Forums/main/1179-bfs-again.html)

Jan Ramboer August 19, 1999 04:16

BFS...again
 
Hello,

Here I am again with my backward facing step. I tried a few boundary conditions on the outlet. Now when I used the boundary condition on the Rieman invariants described by Hirsch, Internal and External flows. for the outflow. on the inflow I used an imposed velocity and temperature. Now it converge very wel but when convergence is almost reached it suddenly blows up. Is it because the outflow and inflow boundary conditions are incompatible?

Greetings,

Jan

John C. Chien August 19, 1999 04:27

Re: BFS...again
 
(1). Don't expect readers to dig up or guess at what you did before, if you expect interested readers to answer your question. (2). It is essential to state your problem in terms of: steady vs transient, laminar vs turbulent, compressible vs incompressible, and this Rieman invariant method. (3). Readers answer your question not just for you, it is for the rest of the readers. It is really expensive on Internet.

Jan Ramboer August 19, 1999 05:40

Re: BFS...again
 
OK let me state my problem again, I am solving a 2D compressible turbulent backward facing step (baldwin lomax model). At inlet I imposed a uniform velocity and temperature. 20 stepheights from the corner to develop the boundary layer. Reynolds is 37000 based on stepheight. At the outlet I used the rieman invariant boundary condition, (Rn=-abs(Vn)+2c/gamma-1). However when convergence is nearly reached, it explodes. Could it be that the two boundary conditions are not compatible?

Greetings,

Jan

John C. Chien August 19, 1999 12:15

Re: BFS...again
 
(1). I don't have the time to worry about whether the condition is properly coded. I am assuming that you are using the transient approach. (2). The suggestion is: find out the location where the calculation diverges. Plot the flow field ( or xy plot) before it diverges. Study the flow field evolution before it diverges. (3) In this way ,you can narrow down the location and the cause of the problem. You can also try a uniform mesh (uniform in x, uniform in y, separately) to elminate the mesh related causes. It is also important to mention whether it is transient formulation, or pressure based, or density based. The method of solution is also important along with the initial guess, the number of iteration at the time the solution diverges.


All times are GMT -4. The time now is 10:13.