|
[Sponsors] |
July 10, 2006, 12:02 |
Fluent Computation
|
#1 |
Guest
Posts: n/a
|
Hi,
I'm facing a sttrange problem- I have a 3d geometry which was meshed partially in Gambit and the remaining in TGrid. I'm trying to run a steady laminar flow case. I input a velocity of 0.175 m/s at the inlet and want to maintain the outlet at atmospheric pressure. These are the only Boundary consitions that I imposed. All the remaining faces are treated as WAll. When I ran the case in Fluent, it just goes on computing and the solution never converged. I also tried reducing the Under- Relaxation factors, changed to second- order upwinding, etc. Nothing worked. Another thing that I noted is that in the plot of residuals, continuity is staying put at 100% error. The X,Y,Z velocities appear to be converging (at least it looks so in the plot), but continuity just remains unchanged. I just cannot understand the reasons for these problems. Any comments/ suggestions? Thanks. |
|
July 10, 2006, 16:46 |
Re: Fluent Computation
|
#2 |
Guest
Posts: n/a
|
A few general tips:
Check your all boundary conditions carefully. Check you donot overspecify things e.g both velocity and pressure at a BC. Check dimensions (units) of your domain as well if they are consistent. Such a problem, you must get a converged solution if code is bug free ( I guess it is) |
|
July 10, 2006, 17:05 |
Re: Fluent Computation
|
#3 |
Guest
Posts: n/a
|
Hi,
I checked all the parameters. I input a steady velocity of 0.175 m/s at the inlet and the outlet is at atmospheric pressure. The grid was created in mm and I scaled it after importing into Fluent. So what could be the problem? It goes on computing for days and finally hangs up because the memory ran out. Thanks. |
|
July 12, 2006, 08:12 |
Re: Fluent Computation
|
#4 |
Guest
Posts: n/a
|
I m quite certain that there is something wrong with your boundary conditions. You may do the following test:
Set reference pressure in your domain 0 Pa (everywhere and also as initial guess before simulation) Now at outlet set the static pressure BC i.e. P_static = 0 Pa. Your problem may be that at inlet you are specifying a velocity and at outlet you are specifying a pressure of 1 atm which does not allow the liquid to come out and therefore continuity is not fullfilled. For an in compresible flow in a CFD solver absolute pressure has no meaning. Only the relative pressure i.e. P - P_ref is what a code cares for. Where P is the pressure in a control volume and P_ref is the reference presure at some point in your domain. A person who has never written a CFD code, unfortunately, do not realize this problem. Hope this helps. VK |
|
July 12, 2006, 08:43 |
Re: Fluent Computation
|
#5 |
Guest
Posts: n/a
|
Hi,
I set up the initial test values that you suggested...... and that's what I did earlier too. But the same problem continues. I don't find any error, but the computation just keeps going on. Thanks, VR |
|
July 12, 2006, 09:37 |
Re: Fluent Computation
|
#6 |
Guest
Posts: n/a
|
In addition to initial test values you have to set reference and boundary pressure equal to 0. I hope you had also done this step.
|
|
July 12, 2006, 09:51 |
Re: Fluent Computation
|
#7 |
Guest
Posts: n/a
|
Yes, I did that too. Nothing happened even then.
|
|
July 12, 2006, 22:09 |
Re: Fluent Computation
|
#8 |
Guest
Posts: n/a
|
Would it help in any way if I specify pressure at both the inlet and outlet? Then I have a very small value (about 1mm Hg at the inlet). So I think velocity at the inlet and pressure at the outlet is the best possible combination. Or will specifying OUTFLOW at the outlet help?
Thanks, VR |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Stopping a Fluent batch job AND saving the data! Possible? | Volker Pawlik | FLUENT | 13 | December 28, 2020 04:16 |
Two questions on Fluent UDF | Steven | Fluent UDF and Scheme Programming | 7 | March 23, 2018 03:22 |
How to open Icem mesh in Ansys Fluent? | emmkell | FLUENT | 27 | February 6, 2018 03:34 |
Fluent 6.3 32bit vs Fluent 12.0 64bit | ibex7 | FLUENT | 7 | April 18, 2011 02:44 |
Advanced Turbulence Modeling in Fluent, Realizable k-epsilon Model | Jonas Larsson | FLUENT | 5 | March 13, 2000 03:27 |