[Yadu krishnan]

Trying to give Neumann boundary condition for heat flux using below UDF and getting error Error: received a fatal signal (Segmentation fault).


#include "udf.h"
#include "mem.h"
DEFINE_PROFILE(Heat_flux,thread,position)
{

cell_t c;
face_t f;
real x[ND_ND];
real cell[ND_ND];
int k = 1;
real wf;
real nc;
real temp;
real tempgrad ;
real Twall=300;
real delta;

begin_f_loop(f, thread)

{
F_CENTROID(x,f, thread);
wf = x[1];
c = F_C0(f,thread);
C_CENTROID(cell,c, thread);
nc = cell[1];
delta = wf-nc;
temp = C_T(c, thread);
tempgrad= (temp - Twall)/delta ;
F_PROFILE(f, thread, position) = k * tempgrad;
}
end_f_loop(f, thread)


}




Tried initializing and run a few iterations before loading UDF and not working.

[pakk]

temp = C_T(c, thread);

Here is the problem: "thread" is a face thread, not a cell thread. You still need to find the cell thread for cell c.

I forgot how that works.

[aved]

Hi!
I am doing a heat transfer to air from the hot wall which is almost a Rayleigh flow case kinda simulation. I wanted to do a validation but couldn't find a good experimental paper. Can anyone suggest me a paper?
Thanks!

[AlexanderZ]

as Pakk said you need to define cell and thread of the finite volume adjusted to face you are looping
you need to use macros:
F_C0(f,t);
THREAD_T0(t);
final code

#include "udf.h"

DEFINE_PROFILE(my_heat_flux,thread,position)
{

cell_t c0;
Thread *t0;
face_t f;
real x[ND_ND];
real cell[ND_ND];
int k = 1;
real wf;
real nc;
real temp;
real tempgrad ;
real Twall=300;
real delta;

begin_f_loop(f, thread)

{
F_CENTROID(x,f, thread);
wf = x[1];
c0 = F_C0(f,thread);
t0 = THREAD_T0(thread);
C_CENTROID(cell,c0, t0);
nc = cell[1];
delta = wf-nc;
temp = C_T(c0, t0);
tempgrad= (temp - Twall)/delta ;
F_PROFILE(f, thread, position) = k * tempgrad;
}
end_f_loop(f, thread)


}

[Yadu krishnan]

Thanks. There is no error showing now. But when I load the udf for heat flux boundary condition and running the simulation, solution is not converged for cold flow without reaction. Below are the model parameters and BCs.



Energy equation - ON
Viscous - Laminar
Species transport - ON , H2&air mixture,No volumetric reactions


inlet BC : velocity = 0.3 m/s , Mass fraction of h2 =0.6,

Mass fraction of o2=0.23


wall BC : above UDF
Hybrid initialization.


The temperature contour is same from the beginning and not changing even after 5000 iterations and attached herewith.

[AlexanderZ]

check the units of heat flux you are applying

[Yadu krishnan]

Heat flux units are W/m2 which is in mismatch with UDF output. So if I access thermal conductivity of wall(W/mK) and store it in variable 'k' , will it work?


Or do i need to make temperature gradient non dimensional.