[MrZZQi]

Dear all,

I am using UDF to update the temperature in a solid zone from the temperature in another solid zone (attachement1). The UDF can be successfully compiled. A strange thing happens that although the temperature could be updated when the UDF is compiled as "Execute on demand" or "Execute at end" (attachement2). However, when I simulated 1 more time step, the temperature recovered to the origin value (before update, (attachement3).

Can anyone help me with this problem? Thanks in advance.

Following is my UDF code.

#include "udf.h"
/* Mapping temperature profiles between solid zones of slabs */
DEFINE_ON_DEMAND(map_T_between_slabs)
{
    Domain* d;
    cell_t c1;
    cell_t c2;
    Thread* t1;
    Thread* t2;
    real local_cor[ND_ND],target_cor[ND_ND];
    real local_t,tar_t;
    real half_range;
    real z1, z2, nt, z_tar;
    real delta_x;
    delta_x = 0.02; 
    half_range = 0.0002; 
    d = Get_Domain(1);
    t1 = Lookup_Thread(d, 10);  // Solid Zone 1
    t2 = Lookup_Thread(d, 9);  // Solid Zone 2

    nt = 0.0;
    begin_c_loop(c1, t1)
    {
        tar_t = 0.0;
        nt = 0.0;
        C_CENTROID(local_cor, c1, t1);
        z1 = local_cor[2];

        z_tar = z1 - delta_x;
        begin_c_loop(c2, t2)
        {
            C_CENTROID(target_cor, c2, t2);
            z2 = target_cor[2];

            if ((z2 >= z_tar - half_range) && (z2 <= z_tar + half_range))
            {
                tar_t = tar_t + C_T(c2, t2); 
                nt = nt + 1.0;       /* count number */
            }
        }
        end_c_loop(c2, t2)
        local_t = tar_t / nt;
        C_T(c1, t1) = local_t;
    }
    end_c_loop(c1, t1)
}

Best regards,
Zuo

[AlexanderZ]

could be problem in the logic of model

you may freeze temperature using fixed values at cell zone conditions and apply profile (define_profile macro) to specify temperature values

[MrZZQi]

Hi Alexander,

Thanks for your reply. The simulation is a solid at position 1 moving to position 2 periodically. The temperature of solids changes so I cannot use fixed values.

Do you suggest that the setting of my case is wrong or that Fluent cannot handle this problem? I also tried a case that just set the fluid temperature in a region, but the same problem occurred. I wonder if there is a very tiny mistake in my code because I think UDF surely can do these things.

DEFINE_ON_DEMAND(map_T_between_slabs)
{
Domain* d;
cell_t c1;
cell_t c2;
Thread* t1;
Thread* t2;
real local_cor[ND_ND],target_cor[ND_ND];
real local_t,tar_t;
real half_cell_size;
real z1, z2, nt, z_tar;
real delta_x;
delta_x = 0.02;
half_cell_size = 0.0002;
d = Get_Domain(1);
nt = 0.0;

/* loop over all cell threads in the domain */
thread_loop_c(t1, d)
{
begin_c_loop(c1, t1)
{
C_CENTROID(local_cor, c1, t1);
z1 = local_cor[2];
if ((z1 >= 0.0105) && (z1 <= 0.0195))
{
C_T(c1, t1) = 310;
}
}
end_c_loop(c1, t1)
}
}

Thank you!

Update: This code works in prue fluid simulations, so the problem may related to the solid-fluid coupling, e.g. the wall and shadow-wall. Will post the solution if solved.

[MrZZQi]

The problem arises because the assigning of temperature will quickly diverge after 5 iterations (in my case). The solution for me is quite hacky, which is executing the DEFINE_ADJUST macro every 2 iterations by adding a sentence like "if (N_ITER % 2 ==0){code}".

Hopes this can help.

Zuo