[lrodrigues]

Let's say I have a 2D problem where two solid domains are separated by a fluid domain. (solid 1|fluid|solid 2). Solid 1 has heat generation, and all heat generated goes through the left wall (zero wall flux at the solid-liquid interface).

How can I make the right wall of solid 2 have the same values of the wall in solid 1 that "collects" the heat? Would an UDF with DEFINE_ADJUST be the best solution? Any thoughts?

Thanks in advance!

[gearboy]

Originally Posted by lrodrigues

Let's say I have a 2D problem where two solid domains are separated by a fluid domain. (solid 1|fluid|solid 2). Solid 1 has heat generation, and all heat generated goes through the left wall (zero wall flux at the solid-liquid interface).

How can I make the right wall of solid 2 have the same values of the wall in solid 1 that "collects" the heat? Would an UDF with DEFINE_ADJUST be the best solution? Any thoughts?

Thanks in advance!

You can use DEFINE_ADJUST to compute the heat value on solid 1 left wall, then use DEFINE_PROFILE to set solid 2 right wall's heat flux same as that. Below is an example to compute the heat flux on each facet of solid 1 left left wall.


Thread *t=Lookup_Thread(domain,left_wall_id);
face_t f;
real A[ND_ND];
begin_f_loop(f,t)
{
F_AREA(A,f,t);
Message("Wall_Heat_Flux=%f W/m2\n",BOUNDARY_HEAT_FLUX(f,t)/NV_MAG(A));
}
end_f_loop(f,t)

[lrodrigues]

Originally Posted by gearboy

You can use DEFINE_ADJUST to compute the heat value on solid 1 left wall, then use DEFINE_PROFILE to set solid 2 right wall's heat flux same as that. Below is an example to compute the heat flux on each facet of solid 1 left left wall.


Thread *t=Lookup_Thread(domain,left_wall_id);
face_t f;
real A[ND_ND];
begin_f_loop(f,t)
{
F_AREA(A,f,t);
Message("Wall_Heat_Flux=%f W/m2\n",BOUNDARY_HEAT_FLUX(f,t)/NV_MAG(A));
}
end_f_loop(f,t)

That makes sense! Thanks for your help, I'll give it a try!

[lrodrigues]

So, after spending some time trying to make it work, it seems something is still wrong.

Here's the code used. Two notes:
First, for convenient reasons, I'm interpreting the UDF instead of compiling it. As soon as I get it running I intend to compile it. From my understandings, in this case there shouldn't be a difference between compiled/interpreted other than the line written as a comment.
Second, as I'm using a mesh in the nm range, the source value needs to be high so that there is an increase in temperature (By running only the DEFINE_SOURCE, the solid increases around 8 K, so the issue isn't here)


#include "udf.h"

real wall_flux_value;

DEFINE_SOURCE(Heat_generation, c, t, dS, eqn)
{
real x[ND_ND];
real y;
real source;

C_CENTROID(x, c, t);
y = x[0];
source = 5e24*y;
dS[eqn] = 0.0;
return source;
}

DEFINE_ADJUST(Heat_Flux_LeftWall, d)
{
Thread *t = Lookup_Thread(d, 13);
face_t f;
real A[ND_ND];
begin_f_loop(f, t)
{
F_AREA(A, f, t);
/* Message("Wall_Heat_Flux=%f W/m2\n", BOUNDARY_HEAT_FLUX(f, t) / NV_MAG(A)); - This line of code doesn't work in interpreted UDF's */
wall_flux_value =+ BOUNDARY_HEAT_FLUX(f, t) / NV_MAG(A);
}
end_f_loop(f, t)
}

DEFINE_PROFILE(Heat_Flux_RightWall, t, i)
{
face_t f;

begin_f_loop(f, t)
{
F_PROFILE(f, t, i) = wall_flux_value;
}
end_f_loop(f,t)
}


When running the case with the UDF, there's no increase in temperature in solid 2. My guess is that the issue must be between storing the value from the DEFINE_ADJUST and using it in the DEFINE_PROFILE. How how I go about doing such? Would an UDM be the right way to go about solving this?

PS: Sorry if it is obvious, but i'm still new to writing UDF's and the whole C language :P