[Pourya.heidari73]

DEAR all
I want to simulate the laser welding with the FLUENT. I just activate the solidification and melting (without VOF). I face a problem to import the Marangoni effect into my simulation. In the BOUNDARY CONDITION (MARANGONI STRESS), I found the formula for surface tension gradient variation base on the temperature. Since I could not import the surface tension gradient by UDF in the MARANGONI STRESS, I try to import this effect by UDF in the SPECIFIED SHEAR. I attach the formula for this part and shot screen of specified stress. To be honest, I do not know which Macro I should use to write this formula. I try DEFINE_PROPERTY and DEFINE_PROFILE. In my UDF, I consider the surface tension graidient as a constant. Everytime I run this code, Fluent would close. It is very painful. Please help me to find the error of this UDF.
Here you can find the UDF that I wrote for this part.
DEFINE_PROFILE(marangonix, t, i)
{
face_t f;
cell_t c;
Thread *tc;
real grad_temp, gama_temp;
begin_f_loop(f,t)
{
c=F_C0(f,t);
tc = THREAD_T0(t);
//tc=F_C0_THREAD(f,t);
grad_temp= C_T_G(c,tc)[0];
gama_temp=-0.00043;
F_PROFILE(f,t,i)=gama_temp*grad_temp;
}
end_f_loop(f,t)
}

DEFINE_PROFILE(marangoniy, t, i)
{
face_t f;
cell_t c;
Thread *tc;
real grad_temp, gama_temp;
begin_f_loop(f,t)
{
c=F_C0(f,t);
tc = THREAD_T0(t);
//tc=F_C0_THREAD(f,t);
grad_temp= C_T_G(c,tc)[1];
gama_temp=-0.00043;
F_PROFILE(f,t,i)=gama_temp*grad_temp;
}
end_f_loop(f,t)
}

DEFINE_PROFILE(marangoniz, t, i)
{
face_t f;
cell_t c;
Thread *tc;
real grad_temp, gama_temp;
begin_f_loop(f,t)
{
c=F_C0(f,t);
tc = THREAD_T0(t);
//tc=F_C0_THREAD(f,t);
grad_temp= C_T_G(c,tc)[2];
gama_temp=-0.00043;
F_PROFILE(f,t,i)=gama_temp*grad_temp;
}
end_f_loop(f,t)
}


I really appreciate, if you help me to solve this issue.

[redlemon44]

Probably best using FLOW-3D WELD for this (https://www.flow3d.com/products/flow3d-weld/) as all the physics are already there and no udfs are needed

[ankit_tiwari]

I have written the following code for a similar problem (following this thread). These gradients are though cell center temperature gradients and I am working on calculations of face center gradients. There is a trick to get this code work (see end of the post)


#include "udf.h"
#include "mem.h"
DEFINE_PROFILE(marangonix, t, i)
{
face_t f;
cell_t c;
Thread *tc;
real grad_temp[ND_ND],gama_tempx;
gama_tempx = -0.000199; /*surface tension gradient with temperature for Nb*/
begin_f_loop(f,t)
{
c = F_C0(f,t);
tc = THREAD_T0(t);
NV_V(grad_temp, =, C_T_G(c,tc));
if (F_T(f,t) > 2741)
{
F_PROFILE(f,t,i)=gama_tempx*grad_temp[0];
}
else
{
F_PROFILE(f,t,i)= 0;
}
}
end_f_loop(f,t)
}
DEFINE_PROFILE(marangoniy, t, i)
{
face_t f;
cell_t c;
Thread *tc;
real grad_temp[ND_ND], gama_tempy;
gama_tempy = -0.000199; /*surface tension gradient with temperature for Nb*/
begin_f_loop(f,t)
{
c=F_C0(f,t);
tc = THREAD_T0(t);
NV_V(grad_temp, =, C_T_G(c,tc));
if (F_T(f,t) > 2741)
{
F_PROFILE(f,t,i)=gama_tempy*grad_temp[1];
}
else
{
F_PROFILE(f,t,i)= 0;
}
}
end_f_loop(f,t)
}


To run this code:

• Run Command "solve/set/expert/" in fluent console. And say "yes" for "Keep Solver Memory from being freed".
• This will keep gradient values stored which will be called from UDF.
• If we hook this UDF directly after initialization. Fluent will give error and exit as variable values are not there for calculation.
• Run calculation for one tiny time step say 1E-6 seconds without hooking UDF. This will store values of gradients.
• Stop the solution after one time step and now hook your UDF. This shall now work