CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   SIMPLER boundary condition (http://www.cfd-online.com/Forums/main/5379-simpler-boundary-condition.html)

Ramesh November 6, 2002 19:58

SIMPLER boundary condition
 
Hi

I've written a SIPLER code for a 2D flow in a pipe. The transient part is very important for my problem. I have notice the radial velocity for the incopresible flow is nealy zero far from inlet however, its value increases as it approaches the outlet(Say from 10^-18 to 10^-14). Also I can not reach steady state solution, I get close but the term dv/dt does not become zero. Your help on this matter is greatly appreciated.

mukhopadhyay November 6, 2002 23:20

Re: SIMPLER boundary condition
 
The header of your posting appears to have the answer - check outlet BC.

Ramesh November 7, 2002 00:43

Re: SIMPLER boundary condition
 
Well, according to Patankar it is not necessary to impose and velocity boundary condition at the outlet. Knowing the outlet pressure should be enough. However, I observed the strange behavoir of radial velocity and am not able to reach the steady state!

STN November 7, 2002 04:39

Re: SIMPLER boundary condition
 
Could you post the Reynolds number?

Is it in the range of Turbulent flow.

If yes which type of turbulent model, do you use?

These information may indicate your problem much more clear to all CFD man in this forum.

Ramesh November 7, 2002 14:23

Re: SIMPLER boundary condition
 
The flow is laminar, Re=400. Thanks for the help.

seaspace November 8, 2002 01:24

Re: SIMPLER boundary condition
 
Could you mail me your code for reference? Thx!

Anthony Wachs November 8, 2002 03:43

Re: SIMPLER boundary condition
 
Hi Ramesh,

If I read well, you are talking in your mail about radial velocity whose magnitude increases from 10^-18 to 10^-14 ?

My opinion is that you do not have to care about such velocity magnitude of 10^-14 or 10^-18 (or anyway any value less than 10^-8 !!), this is definitively zero for your computer since this is less than the usual machine precision.

What I am trying to say is that 10^-14 for a computer is zero for you. Usually, if you are performing calculations in single precision, than velocity value less than 10^-8 have nearly no meaning (for double precision, the limite is 10^-16) and could be considered to be zero.

I think your code is working very well, there is probably nothing that need to be modified.

You are likely to be a bit too optimistic or to expect too much from your code (I'm just kidding, obviously)

I hope it could help,

Anthony

Ambrose November 8, 2002 13:06

Re: SIMPLER boundary condition
 
Well, actually I have very similar problems as Ramesh does. Mine is steady, incompressible, Poiseulle-Couette flow with Re=100. Uniform flow in and fully developed flow out with no information on Pressure; pressure at exit is not given. The problem is that the axial velocity profile at exit is decreasing as the length of the pipe is getting longer; 10% error. One of the things that I suspect is the mass is decreasing from the inlet to exit. Also, have to keep relaxation factor for velocities very small like less than 0.1. Surprisingly, results are really good with 0.01. However, I have no idea why. I looked at my codes very carefully again and again, but everything looks fine. I appreciate any comments.

Ambrose

Ramesh November 8, 2002 14:56

Re: SIMPLER boundary condition
 
Hi Anthony,

Thanks for your comments, If that is the case I still have problem to reach steady state solution. As I mentioned the term dvz/dt (z is the axial direction) never becomes zero (again it ossilats from 10^-13 to 1-^-15). Thanks for your help.


All times are GMT -4. The time now is 06:54.