Hi everybody,

Does someone know how to do a loop over the faces of a cell in an UDF ? or what is the macro to name the face of a given cell ?

I have found the instruction "C_FACE(cell,thread,i)" where I suppose that "i" corresponds to face number i of the cell. Therefore, I tried a thing like :

for (i=0;i less C_NFACES(cell,thread);i++)

{

f0=C_FACE(cell,thread,i);

...

}

(I write "less" otherwise errors occurs in the text, but in the C source, I obviously used the sign, thus don't mind about that, the problem is not here!!)

but it doesn't seem to work.

Any help would be really appreciated.

Thanks in advance

Anthony

The integer variables C_NNODES and C_NFACES return the number of nodes or faces, respectively, for a given cell. The integer variable F_NNODES returns the number of nodes associated with a face.

So this is not the appropriate macro for looping over faces in a cell.

Use : begin_f_loop(f,t) end_f_loop(f,t) ---> which loops over faces in a face thread.

Hi Ravi,

Thank you for your contribution.

Unfortunetaly, your answer does not correspond to what I would like to do. I know very well the command : begin_f_loop(f,t) end_f_loop(f,t), which as you mentioned as well, is useful for a loop over faces in a FACE THREAD. But I am concerned with loop over faces of a CELL, not of a thread : this is completely different.

I have been using "begin_f_loop(f,t) end_f_loop(f,t)" efficiently for a long time in simpler situations, when I would like to do a loop over faces of a face thread.

If you have any other suggestion concerning how to write a "LOOP OVER FACES OF A CELL", thank you in advance

Anthony

Ok here's how to do it:

int numbf; cell_t cf; Thread *tf,*f_thread; face_t face;

/* cf,tf is cell & cell thread */

c_face_loop(cf,tf,numbf) {

face = C_FACE(cf,tf,numbf);

f_thread = C_FACE_THREAD(cf,tf,numbf);

/* do what you like with face numbf with

face & f_thread */

/* -- eg temp = F_T(face,thread); */ }

Greg

Thanks Greg !!

Anthony

[edu_aero]

Originally Posted by Greg Perkins
;98638

Ok here's how to do it:

int numbf; cell_t cf; Thread *tf,*f_thread; face_t face;

/* cf,tf is cell & cell thread */

c_face_loop(cf,tf,numbf) {

face = C_FACE(cf,tf,numbf);

f_thread = C_FACE_THREAD(cf,tf,numbf);

/* do what you like with face numbf with

face & f_thread */

/* -- eg temp = F_T(face,thread); */ }

Greg

Hello, to all and thank you for this post because it is exacly what I was looking for:

However, I am struggling to implement the code written here.

I am trying to implement a code able to calculate the mass flow in different planes created by the user, not boundaries.

My idea is integrate through a volumen and then divide for the width of the volumen. The code is created in a way that if the cell is located in the section then a loop through all the faces wants to be done.
Because the mass flow importance for this problem is only the one in the direction Z, I calculate the normal vector of the face an multiply the Z component by the mass flow, for each face.
I am having problems because I need to initialize two times the variables, for the first bucle and for the second bucle.
I would really appreciate to get some help in this problem.
Attached can be found the code.

#include "udf.h"
#include <math.h>
#define Num_div 5

DEFINE_ADJUST(GetMassFlow,d)
{
face_t f;
Thread *t;
cell_t c;

    int n=0;                 
    int max_div=0;
    int numbf;
    real division=0;        // Width of the division
    real kk0=0;                // Minimum variable for each Division
    real kk1=0;                // Maximum variable for each Division
    real x[ND_ND];
    real NormalVector[ND_ND];
    real Area;
    real SUMflow=0.;

    division=0.09/Num_div;                                // Calculating the division    
    max_div=Num_div+1;    

    for (n = 1; n < max_div; ++n)
    { 
/////////////////////////////////////////////////////////////////////////////
    kk0=(n-1)*division+0.01;                            // Minimum variable for each Division
    kk1=n*division+0.01;                                // Maximum variable for each Division
    SUMflow=0;
        
        thread_loop_c(t, d)                                            // Variable that stores the SUM of Mass Flow is initialized to zero
        {
            C_CENTROID(x,c,t);
            if ((x[2]<kk1) && (x[2]>kk0))    // All the parcels are organized in different sections
            {
                //  L O O P - O V E R - A L L - T H E - F A C E S - O F - A - C E L L 
                /////////////////////////////////////////////////////////////////////////////////////////////////////
                cell_t cf;
                Thread *tf,*f_thread;
                face_t face;
                /* cf,tf is cell & cell thread */
                F_AREA(NormalVector,face,tf);
                Area=  sqrt(NormalVector[0]*NormalVector[0]+NormalVector[1]*NormalVector[1]+NormalVector[2]*NormalVector[2]);
                NormalVector[0]=NormalVector[0]/(Area);
                c_face_loop(cf,tf,numbf)
                {
                face = C_FACE(cf,tf,numbf);
                f_thread = C_FACE_THREAD(cf,tf,numbf);
                /* do what you like with face numbf with
                face & f_thread */
                /* -- eg temp = F_T(face,thread); */ 
                SUMflow+=F_FLUX(face,f_thread)*NormalVector[2];
                }
                ///////////////////////////////////////////////////////////////////////////////////////////////////////
                //  L O O P - O V E R - A L L - T H E - F A C E S - O F - A - C E L L 

            }
        }


    Message ("%f\t",SUMflow/division);
    }
    Message ("\n");
}
}

I have already created a code that integrate and get the mass flow of liquid droplets using the DPM model. The idea is using both together as a convengence criteria.

Thank you very much

[diogoloureiromartinho]

Hey,

I think I will need to implement something as the code you wrote. What I dont get is the meaning of the numbf, how does it change and where does it come from to work as an input.

Kind regards,

Diogo