Learning Forum

[nicolo varallo]

Good morning to everyone.

I am simulating the flow field in a cylindrical 3D bubble column using the Eulerian-Eulerian approach.

I need to define a drag coefficient that depends on the distance from the wall.  My idea was to use the C_CENTROID macro to calculate the distance from the column centre and consequently define the drag coefficient, but it did not work.

Can anyone help me?

In the following, how I calculate the spatial coordinate to define the dependency of the drag coefficient on the radial coordinate.

begin_c_loop_all(c,t)
    {
        C_CENTROID(xc,c,t);
        r=pow(xc[0]*xc[0]+xc[2]*xc[2],0.5);  //distance from the column centre. The coordinate y is the column axis
        CD= a*r  // just an example
       }
end_c_loop_all(c,t)

 

 

[Rob]

It's not compiling, giving an error or giving the wrong answer? "Not working" isn't that precise! :) 

[nicolo varallo]

Sorry for that. It gives the wrong answer; the compilation is ok.

I obtain a constant coefficient without any spatial dependency.

 

[Rob]

Looking at https://ansyshelp.ansys.com/account/Secured?returnurl=/Views/Secured/corp/v241/en/flu_udf/flu_udf_MultiphaseDEFINE.html%23flu_udf_sec_define_exchange_property  I don't know if you need the cell loop. 

[nicolo varallo]

Thanks for you reply.

I tried without cell loop, but it is the same. I think the problem is related to the calculation of the distance from the wall. I implemented this simple udf to see if the C_CENTROID works as expected. The udf defines a drag coefficient equal to 3 if the axial coordinate is smaller than 2m, 5 otherwise:

#include "udf.h"

#include "math.h"

DEFINE_EXCHANGE_PROPERTY(WLL,c,t,i,j)

{

Thread *t_a = THREAD_SUB_THREAD(t,j);

Thread *t_w = THREAD_SUB_THREAD(t,i);

real CD;

real xc[ND_ND];

real a=2;

C_CENTROID(xc,c,t);

real y=xc[1];

        if (y < a)

            {

            CD=3;

            }

            else if (y>a) 

            {

            CD=5;

            }

 

begin_c_loop_int(c,t)

{

C_UDMI(c,t,0)=CD;

}

end_c_loop_int(c,t)

return CD;

}

 

The result obtained is wrong as the figure shows:

The withe line corresponds to a distance from the column bottom of 2m.

 

[Rob]

The white line may be 2m from the column bottom but the UDF is looking at the y-coordinate. What range of y does Fluent show (Scale is a good way to check). 

[nicolo varallo]

 

 

y is the axial coordinate, with y comprised between 0 and 2m, x between -0.12 and 0.12m and z between -0.12 and 0.12m.

The contour represents only the region of interest, not the whole geometry.

The problem is that the udf does not correctly compute the distance from the column’s bottom.

 

 

[Rob]

You're not computing a distance, xc[1] is the y-coordinate. So if your domain is 0-2m in the y direction. 

[nicolo varallo]

Why not? If the reference frame is (0,0,0) and coincides with the column bottom, the y-coordinate of each cell centre coincides with the distance of the cell centre from the column bottom. 

[Rob]

Reference frame isn't necessarily inside the domain. What is the minimum y coordinate in your mesh? 

[nicolo varallo]

0m:

[Rob]

Thanks - that's what I wanted to check. You're not working with "wall distance" as that's a different macro, and not useful here. Can you replot the contour and overlay the mesh? 

[nicolo varallo]

Here it is: 

[Rob]

Get rid of the cell loop around the UDM. 

[nicolo varallo]

Now it works. Thank you very much!

[Rob]

I had to ask too!