CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Axisymmetric B.C. (http://www.cfd-online.com/Forums/fluent/29096-axisymmetric-b-c.html)

 cfd guy November 23, 2001 08:26

Axisymmetric B.C.

Hi everybody,
I'm using Fluent 5.5 to model an 2-D axisymmetric case with swirl velocity at inlet. I have the grid, already imported from GAMBIT, and I'm wondering how to set the inlet profile for it. What is the meaning of "flow direction" in the Velocity Panel? Fluent asks me the magnitude (which I have) and three flow directions (axial, radial and tangential) but how to set these flow directions? Which kind of values are necessary, integer, real ?
And another thing, how do I know which angle (theta) Fluent uses to solve the equations in an axisymmetric situation.

Regards,
cfd guy

 Jin-Wook LEE November 26, 2001 05:14

Re: Axisymmetric B.C.

1. It's direction cosine. If you have some difficulty with 'MAGNITUDE and DIRECTION' option, how about to use 'COMPONENTS' option ?

2. Fluent 4.x use one radian for the calculation domain, Fluent 5.x use whole angle(2 * phi radian, that is 360 degree).

Sincerely, Jinwook

 cfd guy November 26, 2001 08:13

Re: Axisymmetric B.C.

Thanks for the suggestion Mr. LEE,
I has tried all options available, but those options gave me erroneous results, so I believe that I'm not setting it correctly.
Imagine a brief situation. Consider a generic 3-D flow (with swirl), the mass flow rate (Q) will be V*w*L (Velocity Vector*width*length). If I use this Q to an axisymmetric situation, I'll decompose it in two relevant components.
Radial Velocity (Vr = Q/area[2-D]), where area will be perimeter(pi*diameter)*length
Swirl Velocity (Vtheta = Q/area[3-D] = V), where area[3-D] is w*L
PS.: I'm imposing how components will be set at BC as mentioned up. And length is associated to an axial property.
If I submit this job to Fluent 5.5.14 (the version I have), it seems that there's no significant swirl in the calculation domain. Setting Inlet BC as COMPONENTS, there's also an extra item, ANGULAR VELOCITY (rad/s), I'm letting null this parameter. Do I need to include it?
Sorry for the long message, but any hints will be appreciated.
Regards,
cfd guy

 Jin-Wook LEE November 26, 2001 21:52

Re: Axisymmetric B.C.

You have some mistake.

1. mass flow rate = RHO * V_ax * Area, where,

V_ax = axial velocity

Area = phi * R * R

2. swirl velocity is not Q/Area. It depends on your system.

3. Radial velocity is entirely independent of mass flow rate. In general zero for engineering calculation.

It might be time consuming to someone who may read this post, but I would like to show you one example.

Let your system as :

Inlet radius = 1m

mass flow rate = 10kg/sec(= 10m3/sec if Rho = 1kg/m3)

You have swirler with 60 degree vane

Then, approximate input is

1. axial velocity = 10/(phi*R*R) = 3.18 m/sec

2. radial velocity = 0.0

3. swirl velocity = 3.18 * tan(60) = 5.51 m/sec

Then swirl number is approximately 1.154 (In this case, Swirl number is ~= (2/3)* tan(60), strong swirl. So you can see the strong swirl effect.

Sincerly, Jinwook

((2/3)

 cfd guy November 27, 2001 05:48

Re: Axisymmetric B.C.

Thanks again Mr. Lee,
You're right, I really did some mistake, I was meaning Volumetric Flow rate, not mass flow, sorry.
It was very clear to me the example you have shown, but as you said, the swirl velocity depends on my system. You have shown a simple case with axial and swirl velocities.And my case is a little bit different, I don't use axial velocity in the inlet, I use radial and swirl velocities, it is an axisymmetric cyclone I want to run. And in two other codes, one user coded and the other one is a commercial code, we use those profiles (of radial and tangential velocity) I mentioned in my previous post for the inlet BC.
So, how can I approximate a simple profile to my specific case using Fluent 5?
Sorry to bother you again, but any help will be really appreciated. Regards,
cfd guy

 Jin-Wook LEE November 29, 2001 21:19

Re: Axisymmetric B.C.

1. Do you think that it is possible to approximate cyclone as axisymmetric ? What kind of cyclone do you want to model ? Cyclone separator or ...... ? O.K. please maintain inlet area(not inlet width) the same as 3D case.

2. Anyway, I can not understand your question, because, under axisymmetric assumption, velocity of your system has only two componets, radial velocity and swirl velocity. So, I can only say that just give the two absolute values.

One example :

Suppose that your cyclone(I assume it as cyclone separator like Stairmand type) has tangential inlet. 90 degree tangential inlet is impossible so the inlet maight be located at the position of 60 to 90 degree. Then 75 degree is mean value. In this case,

V_swirl = V_in * cos(15 degree) V_radial = V_in * sin(15 degree)

But I think that flow field of 2D case is fairly different from that of 3D case because you may change the swirl number of the system without deep consideration of swirl characteristics of the cyclone.

One reference : 'Swirl Flows', A.K. Gupta, D.G. Lilley, N. Syred, ABACUS press.

It's the end of my suggestion. It is up to you, now. Hoping good luck with you.

Sincerely, Jinwook

 All times are GMT -4. The time now is 09:29.