CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   CFX (https://www.cfd-online.com/Forums/cfx/)
-   -   "prevent fluid from flowing into the domain" (https://www.cfd-online.com/Forums/cfx/25444-prevent-fluid-flowing-into-domain.html)

J March 13, 2008 11:09

"prevent fluid from flowing into the domain"
 
Hi guys...

During simulations I got this message:

Notice

A wall has been placed at portion(s) of an OUTLET boundary condition (at 1.3% of the faces, 0.9% of the area) to prevent fluid from flowing into the domain. The boundary condition name is: Outlet2. The fluid name is: Blood1. If this situation persists, consider switching to an Opening type boundary condition instead.

I put as boundary condition the average pressure. In ICEM if I run Build Topology I found a blue triangle in this outlet. Somebody have an idea to correct this problem?

Thanks guys

J

Usman March 13, 2008 16:09

Re: "prevent fluid from flowing into the domain"
 
Wall is being placed on very small area (0.9%) which i think should not affect your solution too much. I would say carry on with the simulation and see how out come is affected.

or

Switch your outlet BC to opening for some iteration and then swich it back to your normal BC. If problem still persists then,

You will have to move your outlet further downstream.

Usman

Neel March 14, 2008 03:46

Re: "prevent fluid from flowing into the domain"
 
Hi Usman,

What if areas are 100% of the faces and 100% of the area?

Usman March 14, 2008 06:57

Re: "prevent fluid from flowing into the domain"
 
In that case my friend you have problem at hand. Switch to opening BC and run it for lets say 40 iterations or time steps. And then switch back to normal BC. In my simulations message under consideration usually disappears when i do that.

Usman

Kushagra March 14, 2008 15:56

Discussion continued.
 
I am relatively new to CFD, hence my comments should be taken as my offence :)..

The impression that I got from many people, that outlet (flow out of domain) is the only physically sensible boundary condition at outflow boundaries. I think it may be right for most of the cases.

Hence the suggestions to

1) extend the outlet pipes

OR

2) to change the 'outlets' to 'openings' for some time and then rechange it to 'outlets'

may work as it helps removing the numerical affect.

BUT for some cases, flow 'in' and 'out' at the same boundary is indeed. Hence using an 'opening' for whole simulation should not be a bad idea. AGAIN remember, 'openings' are not good with LES type models. They make linear solver fail very often (atleast CFX fails for me. While I have read that FLUENT is pretty stable with LES and 'pressure-outlet' BCs which are nothing but 'opening' in CFX).

So I may suggest not to buy the idea of always using outlets. Think about the physics of your problem and then take the decision. Mostly 'outlet' will be the appropriate decision.

And I would love to hear the feedback upon my comments. I just learnt it from my experience so far. There may be need of correction.

Cheers, Kushagra

Kushagra March 14, 2008 15:57

Re: Discussion continued.
 
correction....

I am relatively new to CFD, hence my comments should NOT be taken as my offence :)..


Neel March 17, 2008 05:21

Re: "prevent fluid from flowing into the domain"
 
Hi Guys,

Thank you for advicing me on this. I tried using "opening" initially for 50 iteractions and then switched back to 'outlet' and ran it over the weekend. I still get 100% faces & areas. I also get negative pressures.:(

Usman March 17, 2008 08:04

Re: "prevent fluid from flowing into the domain"
 
Hi Neel

Are you simulating RANS case or is it LES? You can switch to opening BC if your case is RANS as Kushagra mentioned and see how that will affect your results. As a last resort you will have to draw your model again and move your outlet further downstream.

Usman

Neel March 17, 2008 09:31

Re: "prevent fluid from flowing into the domain"
 
Hello guys,

I am not using RANS nor LES, I have chosen k-epsilon model for this.( I am using ansys 10.0)I thought it k-epsilon is a good start.

what do you suggest?

Magnoli March 17, 2008 10:31

Re: "prevent fluid from flowing into the domain"
 
Hi, if you are using k-e, you are using RANS.

Kushagra March 17, 2008 11:38

Re: "prevent fluid from flowing into the domain"
 
You must extend your domain at its outlet up to a good length......this will make sure that the domain does not get affected by the outlet boundary condition...

then if the backflow still occurs, use opening type boundary condition.....and when the simulation converges, see if reverse flow is there...draw vectors at outlets...and see

Now, as you mention that negative pressure (I assume you mean less than pressure at outlet) occurs, it seems to me a physical problem....Though may be a very less amount of reverse flow is expected....Is the area of outlet cross section is very small? then may be the 'imposed' outflow boundary condition is messing up the whole simulation and you are getting 100% backflow...it happened with me once...

As you are using k-e, 'opening' will not give any convergence problem....if you were using LES then you had problem...

try it out and let us know

Neel March 17, 2008 11:47

Re: "prevent fluid from flowing into the domain"
 
Hi,

Thank you for the responses. I understand now perfectly. Thank you Magnoli for confirming. As Kushagra and Usman instructed, I will extend the outlet pipe and use 'outlet'. If i still get backflow then i ll use 'opening'instead. I am wondering if i use 'opening' and 'static pressure of 12000 Pa' will there be any liquid hold up?!? Thanks, Neel


yeeking86 August 1, 2009 00:47

I am having the same problem. How do you change the BC from opening to outlet when it is running? I went to 'Edit Run in Progress' but I am not allowed to change the BC.

smn August 1, 2009 01:45

Quote:

Originally Posted by Neel
;86908
Hi,

Thank you for the responses. I understand now perfectly. Thank you Magnoli for confirming. As Kushagra and Usman instructed, I will extend the outlet pipe and use 'outlet'. If i still get backflow then i ll use 'opening'instead. I am wondering if i use 'opening' and 'static pressure of 12000 Pa' will there be any liquid hold up?!? Thanks, Neel

hi Neel:
are you working on a blood flow?would you please tell me some details about your current work?i am modeling a blood vessel using fsi and i have some problems.

flyingd August 1, 2009 03:20

Hi,
You can change the calculate region, let the outlet is far from your structure.It is Ok!
The second method is change the bounday to opening.It is OK!
But the best method is the first.

yeeking86 August 1, 2009 12:49

Quote:

Originally Posted by flyingd (Post 224929)
Hi,
You can change the calculate region, let the outlet is far from your structure.It is Ok!
The second method is change the bounday to opening.It is OK!
But the best method is the first.

Hi,
Thanks for your advice but how do I change the outlet further from the structure? Do I have to use some kind of expression?
Thank you

flyingd August 2, 2009 01:32

reply
 
You can make wave region of your calculation longer.It is oK!.

bharath August 3, 2009 22:17

Hi guys,
is it really need of changing from opening to outlet after some iterations?how it affects the solution?

flyingd August 3, 2009 23:28

outlet setting
 
Quote:

Originally Posted by bharath (Post 225146)
Hi guys,
is it really need of changing from opening to outlet after some iterations?how it affects the solution?

Hi:
If the calculation of region is enough length, the outlet needn't change.
If there is free surface, the outlet will be set STATIC PRESSURE.
If there is not free surface, the out let will be set ave pressure zero.

C'est Moi January 26, 2014 10:33

Im having the same problem but Im using Ansys CFX 14.0 working to model a van in a flow domain. After 20 iterations out 60, I receive the same error but dont know how to proceed as recommended i.e "how to switch to Opening type BC" see the error notice I received:

Notice ******
A wall has been placed at portion(s) of an OUTLET boundary condition (at 1.0% of the faces, 0.9% of the area) to prevent fluid from flowing into the domain.
The boundary condition name is: outlet.The fluid name is: Fluid 1.

If this situation persists, consider switching to an Opening type boundary condition instead.


All times are GMT -4. The time now is 22:04.