[JMM_UC3M]

Hi folks,

I am creating my first UDF and I have a big problem, I do not understand how to read the face wall temperature . Can anyone help me?. The code is almost ready except how to input this temperature. Many thanks in advance.
PS: If i set up temperature at 500 K all calculus are ok, but I would need to read the real one!

UDF below!
#include "udf.h"
#define QMAX 407400 /* W/m2 */
#define EEPSILON 0.1
#define TAU 5.67e-8
#define T_AIRE 300
#define H_AIRE 40 /* convective coef.W/m2K */

/* profile for heat flux*/
DEFINE_PROFILE(heatflux_profile, t, i)

{
/* variable declaration */
real centroide[ND_ND]; /* vector position */
real x;
real y;
face_t f;
cell_t cell;
real seno; /* heat distribution */
real T_p; /* wall temperature */


begin_f_loop(f, t)
{
F_CENTROID(centroide,f,t);
x=centroide[0]; /* valor del cntroide en la componente x */
y=centroide[1]; /* valor del cntroide en la componente y */
seno=y/sqrt((x*x)+(y*y));


T_p = 500; /* I NEED TO FIND THE FUNCTION IN ORDER TO READ THE WALL TEMPERATURE */

if (0 < y)
{
F_PROFILE(f,t,i)=(QMAX*seno)+TAU*((T_p*T_p*T_p*T_p )-(EEPSILON*T_AIRE*T_AIRE*T_AIRE*T_AIRE))+H_AIRE*(T_ p-T_AIRE);

} else {
F_PROFILE(f,t,i)=0;
}
}
end_f_loop(f, t)
}

[JMM_UC3M]

Small correction:

#include "udf.h"
#define QMAX 407400 /* W/m2 */
#define EEPSILON 0.1
#define TAU 5.67e-8
#define T_AIRE 300
#define H_AIRE 40 /* convective coef.W/m2K */

/* profile for heat flux*/
DEFINE_PROFILE(heatflux_profile, t, i)

{
/* variable declaration */
real centroide[ND_ND]; /* vector position */
real x;
real y;
face_t f;
cell_t cell;
real seno; /* heat distribution */
real T_p; /* wall temperature */


begin_f_loop(f, t)
{
F_CENTROID(centroide,f,t);
x=centroide[0]; /* valor del cntroide en la componente x */
y=centroide[1]; /* valor del cntroide en la componente y */
seno=y/sqrt((x*x)+(y*y));


T_p = 500; /* I NEED TO FIND THE FUNCTION IN ORDER TO READ THE WALL TEMPERATURE, DO NOT WHY F_T(f,t) DOES NOT WORK */

if (0 < y)
{
F_PROFILE(f,t,i)=(QMAX*seno)-TAU*((T_p*T_p*T_p*T_p )-(EEPSILON*T_AIRE*T_AIRE*T_AIRE*T_AIRE))-H_AIRE*(T_ p-T_AIRE);

} else {
F_PROFILE(f,t,i)=0;
}
}
end_f_loop(f, t)
}

[pakk]

You say F_T does not work, but can you be more specific? What happens if you do that?

[JMM_UC3M]

Sure!! (btw many thanks!!! I am really stacked). And just one comment, the mesh I have is not good because I was trying to run the UDF and then I was gonna work on the mesh and so on, hope it is not the problem because when I set up T_p = 500 K it works...

Please find below the error , it seems it´s about the UDF in line 21

"Divergence detected in AMG solver: temperatureWarning: C:\\Users\\usuario\\Desktop\\udf\\heatflux_profile - copia (21).c: line 21: Warning: C:\\Users\\usuario\\Desktop\\udf\\heatflux_profile - copia (21).c: line 21: Warning: C:\\Users\\usuario\\Desktop\\udf\\heatflux_profile - copia (21).c: line 21:
Error at host: floating point exception
Warning: C:/Users/usuario/AppData/Local/Temp/heatflux_profile - copia (21).c.19192.0.c: line 21: Warning: C:/Users/usuario/AppData/Local/Temp/heatflux_profile - copia (21).c.19192.0.c: line 21: Warning: C:/Users/usuario/AppData/Local/Temp/heatflux_profile - copia (21).c.19192.0.c: line 21:
Error at Node 0: floating point exception
Error: floating point exception
Error Object: #f
Registering Udf, ("C:\Users\usuario\Desktop\udf\heatflux_profile - copia (21).c")

Calculation complete."

That´s the UDF I have used with the right function to read the wall temperature

#include "udf.h"
#define R 0.01
#define QMAX 40740 /* W/m2 */
#define EEPSILON 0.1
#define TAU 5.67e-8
#define T_AIRE 300
#define H_AIRE 40

/* profile for heat flux*/
DEFINE_PROFILE(BC, t, i)

{
/* declaración de variables */
real centroide[ND_ND]; /* vector de posición */
real x;
real y;
face_t f;
cell_t cell;
real seno; /* para meter la distribucion de calor */
real T_p; /* temperatura de la pared */


begin_f_loop(f, t)
{
F_CENTROID(centroide,f,t);
x=centroide[0]; /* valor del cntroide en la componente x */
y=centroide[1]; /* valor del cntroide en la componente y */
seno=y/sqrt((x*x)+(y*y));


T_p = F_T(f,t);


if (0 < y)
{
F_PROFILE(f,t,i)=(QMAX*seno)-TAU*((T_p*T_p*T_p*T_p)-(EEPSILON*T_AIRE*T_AIRE*T_AIRE*T_AIRE))-H_AIRE*(T_p-T_AIRE);

} else {
F_PROFILE(f,t,i)=0;
}
}
end_f_loop(f, t)
}

[pakk]

Is that in the first iteration?

And try this: initialize your simulation, don't run, and plot the heat Flux on your relevant boundary. Is that what you expected?

[JMM_UC3M]

Hi Pakk, I think it was the mesh.... I cannot beleive it... I am running the simulation, but first checks seems good.... I will replay in a few hours heheheh.... I created a good mesh, but now I dont have the floating point (it was in the second iteration BTW)

[JMM_UC3M]

Just to close out this issue, I guess there were two different problems:
1) The mesh
2) just in case I created alll "z" coordinates in the positive side, just in case
Additionally, a pair of formulas were not ok, but it was not the problem... Below the UDF for a tube of a solar receiver, were you can see the heat flux, and the two losses (convective and radiation)

#include "udf.h"
#define QMAX 407400 /* W/m2 */
#define EEPSILON 0.9
#define TAU 5.67e-8
#define T_AIRE 300
#define H_AIRE 40

/* profile for heat flux*/
DEFINE_PROFILE(receiver, t, i)

{
/* declaración de variables */
real centroide[ND_ND]; /* vector de posición */
real x;
real y;
face_t f;
cell_t cell;
real seno; /* para meter la distribucion de calor */
real T_p; /* temperatura de la pared */


begin_f_loop(f, t)
{
F_CENTROID(centroide,f,t);
x=centroide[0]; /* valor del cntroide en la componente x */
y=centroide[1]; /* valor del cntroide en la componente y */
seno=y/sqrt((x*x)+(y*y));


T_p=F_T(f,t); /*500*/


if (0 < y)
{
F_PROFILE(f,t,i)=(QMAX*seno)-TAU*((T_p*T_p*T_p*T_p)-(EEPSILON*T_p*T_p*T_p*T_p))-H_AIRE*(T_p-T_AIRE);

} else {
F_PROFILE(f,t,i)=0;
}
}
end_f_loop(f, t)
}



PS: Thanks pakk