CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Fluids from two Inlets opposing each other (http://www.cfd-online.com/Forums/cfx/73037-fluids-two-inlets-opposing-each-other.html)

Arvind February 25, 2010 04:09

Fluids from two Inlets opposing each other
 
1 Attachment(s)
Hello Every1,

I am trying to run a simulation in which i have a domain as shown in figure. Through Inlet1 i have given velocity( from left to right) and through Inlet2 i want the same fluid to come in the opposite direction(i.e from right to left) but with different velocity. I gave negative values of velocity at Inlet2 thinking it shud work but the solver is terminating with errors.


Pls can any1 tel whether I am heading towards the right direction.:confused::confused:

Timon February 25, 2010 05:58

Hi,

An inlet with negative velocity imposed on it is an outlet. Impose positive values with a direction normal to the boundary.

Tijmus

mach000 February 25, 2010 06:19

Quote:

Originally Posted by Arvind (Post 247315)
Hello Every1,

I am trying to run a simulation in which i have a domain as shown in figure. Through Inlet1 i have given velocity( from left to right) and through Inlet2 i want the same fluid to come in the opposite direction(i.e from right to left) but with different velocity. I gave negative values of velocity at Inlet2 thinking it shud work but the solver is terminating with errors.


Pls can any1 tel whether I am heading towards the right direction.:confused::confused:

Hi,
if you set "normal velocity component" on inlet2 the flux will go automatically from right to left, in your case.
CFX knows that from Inlet BC the flux enters into the domain. A negative component is not raccomanded.

Arvind February 25, 2010 07:10

@ Timon and Mach000


First of all let me thank u for ur reply.......I tried setting normal speed B.C at inlet 2 (using positive value of velocity) but still the solver is terminated with errors after around 10-15 iterations.


Is there any other thing that i am missing out???????

mach000 February 25, 2010 07:23

i suppose yuo have an "overflow".

Is the simulation transient or stationary?

I think you have you have to set a transient one.

You have to try to obtain a small courant number. You can control it modifying the timestep.

Are you initializing the calculation?

Let me know your initial setting

Bye

Arvind February 25, 2010 08:06

@ Mach000



Yes u are right. I am getting a fatal overflow error.

I am working with a steady state simulation only. Should i use transient type?

The problem i am trying to solve deals with Hypersonic velocities so my initial Mach is around 5.5 and i hav used supersonic condition at inlet giving cartesian velocity and pressure. The various B.Cs are as follows

INLET1: Supersonic with Cartesian velocity and Pressure

INLET2 : Subsonic with Normal velocity component(as suggested by u and Timon)

OUTLET: Supersonic

OPENING: subsonic with static pressure(Entrain).

Quote:

Are you initializing the calculation?
I din understand what u mean by initialising the calculation. Can u explain it????


Any help in this regard would be very much appreciated.

Thnx in advance!!!!

sans February 25, 2010 08:34

Hi, whats the timestep your using? try a small timestep in the initial stages and then ramp it up. These issues are discussed in the help files.

mach000 February 25, 2010 08:34

Supersonic?
I've not experience in this kind of simulation.
Could you explain me where are the BC "opening" and "outlet"? From your doc I don't see those BC.
Initializing is a step in which you set the initial fluid dynamics features of the flow. Most of cases, the overflow is due to bad initial conditions. In CFX try to find the icon with "t=0" as symbol.
Let me know

Arvind February 25, 2010 09:01

1 Attachment(s)
@ sans

The time step i am using is 0.1 sec


@mach000

I am working in aerospace domain The model which i have shown is actually a simplified model of a space shuttle .I think now u will be clear why am using Hypersonic velocities.

As u said i have checked the initial conditions ( from that t=o icon) .In that everything is set to automatic with velocity type as cartesian. Should I change that setting from "Automatic" to "Automatic with value" and give ma conditions in der.?


The modified domain showing opening and outlet is attached. Please check it and let me know.

mach000 February 25, 2010 09:08

ok,
know it seems clearer.

AS first step i should define symmetry conditions instead of your opening (top and down).

Then, I should change your outlet with opening, but you can try with outlet, too.

About timestep: In my opinion it is too high.

If you want you can send me your pre file (I've cfx v11). I can check it better

Arvind February 25, 2010 09:13

@ mach000

Thnx for ur prompt replies.

Actually i have defined the front and back surfaces as the symmetry . If I change ma opening to symmetry( as suggested by you) then what shud i assign for the front and back surface?


Wat difference will it make if i change the opening to symmetry and outlet to opening????? Hope I am not loading u with questions:rolleyes::rolleyes:

About timestep: I wil try reducing it.

Quote:

If you want you can send me your pre file (I've cfx v11). I can check it better

Give me ur mail ID so that i can send my pre file.


and one more thing.....Is my simulation a transient one??????

mach000 February 25, 2010 10:35

my mail is cannavacciuolociro@gmail.com

in front and back is correct to assign simmetry boundary condition.

If top and down BC are quite far from the object, they are in a not disturbed zone, where the flux is very regular. So, you don't need to simulate something that enter or go out from your domain. There is another kind of solution that yuo can consider but for now i think we can try with the one i've just try to explain to you.

About the opening condition instead of the opening: when you set outlet, cfx will think that from that boundary the flux can only goes out. If there is something different (for exemple a vortex) some flow will "enter" into the domain from that boundary. This could be a problem for the solver. If you set opening, instead, cfx "knows" that from that boundary the flow could enter or go out.
Sorry for my english, i hope you undestand me.

Arvind February 25, 2010 10:43

Thnx a ton!!!!

I wil change my outlet to opening and my opening to symmetry as u suggest and try solving my problem.


Moreover I wil mail the CFX-PRE file to ur mail. Pls have a look at it and let me know.

mach000 February 25, 2010 10:51

ok!

I'm waiting for your file

Arvind February 25, 2010 10:59

I have sent the file to ur mail......Pls check ur inbox and let me knw whether u r able to access it....Thnx

mach000 February 25, 2010 11:05

i'm checking your files.

I'll answer you from my private address to your.

mach000 February 25, 2010 11:26

i sent you a mail to your private address.


All times are GMT -4. The time now is 02:03.