CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   inflow into closed container (http://www.cfd-online.com/Forums/main/89351-inflow-into-closed-container.html)

ehooi June 10, 2011 03:36

inflow into closed container
 
Hi,
Is it possible to model, say a cylindrical domain where one one end, there is an inlet with a diameter smaller than the cylinder. At this end, the velocity inlet boundary condition is prescribed. The rest of the wall is assumed to be stationary no slip wall.

In this case, since we have an inlet flow into a closed domain, can the conservation of mass equation still be satisfied? I tried solving this in FLUENT and the solution converges. I am wary of its solution because intuition tells me that conservation of mass is not satisfied since there is no outflow to balance the inflow. Is the solution of FLUENT correct?

Thanks.

Sincerely,
EH

Ford Prefect June 10, 2011 05:40

Incompressible?

ehooi June 10, 2011 06:57

Yes, it is solved with the k-e model.

Ford Prefect June 10, 2011 07:57

Quote:

Originally Posted by ehooi (Post 311446)
Yes, it is solved with the k-e model.

I assume that you mean "Yes and it is solved with the k-e model.", hoping this is not a troll attempt.

Anyways, to your main question: No it should not work in this case, and I assume that your solution has not reached convergence and that you have terminated it before it blew up.

Cheers!

ehooi June 10, 2011 08:01

Quote:

Originally Posted by Ford Prefect (Post 311457)
I assume that you mean "Yes and it is solved with the k-e model.", hoping this is not a troll attempt.

Anyways, to your main question: No it should not work in this case, and I assume that your solution has not reached convergence and that you have terminated it before it blew up.

Cheers!

Well, I am solving a transient problem and my continuity criteria was set at 1e-6. After about 60s, the residual for continuity, velocities, k an e are less than 1e-6. When I plotted the contour for residual for mass imbalance, it is of the order of 10-9.

Ford Prefect June 10, 2011 08:49

If you calculate the mass flux across all boundaries what do you get?

ehooi June 10, 2011 09:11

I assumed you meant mass flow rate? Well, the mass flow rate at all the walls are zero except for the inlet where a value of 0.015kg/s was computed.

I guess it means there is a 0.015kg/s flow into the domain...but how is it that convergent is satisfied? This is puzzling. By the way, the mass flow rate at the interior calculated from fluent is -0.18kg/s...does this imply that mass vanishes somehow at the interior?

Ford Prefect June 10, 2011 11:18

Ok, I did a quick test with Fluent and I get the same results as you. However when I try the same in COMSOL and PHOENICS the solution diverges as expected. I have no good explanation as to why this works in Fluent. I will think more on the matter during the weekend. A cold beer is a good catalyst for "problems" like this ;)

Cheers!

ehooi June 13, 2011 01:25

Quote:

Originally Posted by Ford Prefect (Post 311502)
Ok, I did a quick test with Fluent and I get the same results as you. However when I try the same in COMSOL and PHOENICS the solution diverges as expected. I have no good explanation as to why this works in Fluent. I will think more on the matter during the weekend. A cold beer is a good catalyst for "problems" like this ;)

Cheers!

Hi,
If you happen to check the mass flow rate on the boundary, there is a positive flow rate into the container. However, if you calculate the mass flow rate at the interior as well, then there is a sink, which is somehow greater than the net mass flow rate at the boundary (at least in my model). Really puzzling...

Ford Prefect June 13, 2011 05:03

Quote:

Originally Posted by ehooi (Post 311714)
Hi,
If you happen to check the mass flow rate on the boundary, there is a positive flow rate into the container. However, if you calculate the mass flow rate at the interior as well, then there is a sink, which is somehow greater than the net mass flow rate at the boundary (at least in my model). Really puzzling...

Yes it is, however as I see it the net flow rate is unaffected by the value in the interior. Strange.


All times are GMT -4. The time now is 11:32.