CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Fin inside supersonic flow (http://www.cfd-online.com/Forums/fluent/73780-fin-inside-supersonic-flow.html)

Liol Calvert March 16, 2010 17:52

Fin inside supersonic flow
 
Hi,
As a project to end our studies in engineering in Poitiers (France), we had to modelize the behavior of a fin inside a supersonic flow (Mach 5).
http://yfrog.com/jbtestwioj
First, we wanted to know the temperature,pressure and velocity conditions near the fin. Here is a picture of the modelisation :
http://img695.imageshack.us/img695/96/testwio.jpg
As you can see, the fin is on the right, not meshed, surrounded by the chock wave.

We solved this situation for the 3 seconds-long flight, with a time step of 0.001 s, in order to know the temperatures into the boundary layer.
Next step was to mesh the fin (so we will have 2 faces, one solid and one fluid), to chararacterize the heating of the fin, knowing flux correlations on the fin boundaries.

Here his our modelization :
http://img689.imageshack.us/img689/865/modr.jpg
We've got these named selections :
solid
fluid
inlet
wall-fluid (automatically created by fluent)
wall1 (where correlations must be applied)
wall2

We tried to write an UDF to use boundrary layer temperatures into correlations, but we never made it to work. in Wall1, we choose the UDF in thermal conditions, but the fin doesn't see its temperature increase, and convergence criteria is very very high (quite impossible to go under 1E-2).

Here is the code we wrote, with a very simple correlation to try to make it work.

Code:

#include "udf.h"

DEFINE_PROFILE(temperature,thread,index)
{
real x[ND_ND];
real y;
face_t f;

begin_f_loop (f,thread)
    {
    real T= C_T(f,thread);
    F_CENTROID(x,f,thread);
        F_PROFILE(f,thread,index)= 100*(300-C_T(f,thread));
    }
end_f_loop (f,thread)
}

We had to use Fluent (told by our project director), yet we hadn't any formation on this software (Ansys Workbench 12). We're used to work with Matlab, so, even with the Ansys User Guide and Tutorial, it's pretty hard to do anything with valuable content... :confused:

krishna086 March 18, 2010 14:17

Free Stream velocity
 
Hi,

Just curious. I am modeling a similar kind of geometry. But this is incompressible flow. Do you know any method to specify the free stream conditions.

Also let me what options for Density, Viscosity, Thermal Conductivity, and Cp you chose while simulating your model.

Thanks.

Liol Calvert March 18, 2010 14:53

Hi,

We chose to specify "Pressure farfield", on Inlet, to specify the free stream to 5 mach and 1013 Pa.

For the others parameters, we let the default values chosen by Fluent, excepted the fluid nature, which was "ideal gas"

krishna086 March 18, 2010 20:56

Thank u Liol. But the pressure far field BC is only for compressible flows. That is where the problem is. I am dealing with incompressible flows.

If u come to know any other BC for Mach < 0.3 , let me know.

denizen March 19, 2010 03:10

i think conjugate heat transfer will be better.

krishna086 March 19, 2010 07:23

Thanks denizen. So when you say conjugate heat transfer, you mean to sandwich the wall between two adjacent fluid zones.

Let me know.

denizen March 19, 2010 09:16

it means energy equation will solved everywhere in domain (solid - fin and fluid - gas) and heat transfer on the wall will be automaticaly accounted

krishna086 March 19, 2010 13:01

ok. I got u. But I thought you are giving a solution to my problem of BC. My problem is about the outer boundary declaration in case of subsonic flows.

Because if I restrict the flow field inside a defined domain, the no slip on walls causes flow retardation to the bulk of the flow and thus produces different results when compared to a wedge situated in a free stream flow. (flow at infinity).

Let me know if you have an answer to this problem.

Thanks.

denizen March 22, 2010 10:31

could you explaine more detailed?


All times are GMT -4. The time now is 00:24.