[kanak_BUET_ME19]

Hello, I am working on Selective Laser Melting CFD simulation using ansys fluent. And to apply the Marangoni effect (equation photo added below) , I need to calculate Cell temperature gradient.

According to fluent UDF manual , I can access Cell Temperature Gradient using C_T_G(c,t)[0] (for x axis gradient)

But whenever I mention the term C_T_G(c,t) inside udf, after running the simulation Fluent crashes instantly saying "bad termination" ( I will add the error log as a doc link)

My UDF code for the Marangoni part is like this :



DEFINE_SOURCE(x_mom, c, t, dS, eqn)
{
Thread *g, *w ;
g = THREAD_SUB_THREAD(t, 0);
w = THREAD_SUB_THREAD(t, 1);
real source = 0 ;
real recoil_pressure = 0;
real marangoni_flow = 0 ;
real D;
real T = C_T(c,t);
real temp_g_x = C_T_G(c,t)[0] ;
real temp_g_y = C_T_G(c,t)[1] ;
real temp_g_z = C_T_G(c,t)[2] ;
D = (2 * C_R(c,t) / (C_R(c,g) + C_R(c,w))) ;
if (C_VMAG(c,t) != 0)
{
if (T > boiling_temp)
{
recoil_pressure = 0.54 * 101000.0 * exp((latent_vap * molar_mass * (T - boiling_temp)) / (8.314 * T * boiling_temp)) * C_VOF_NX(c, t);
}
if (C_LIQUID_FRAC(c, t) > 0)
{
marangoni_flow = temp_surface_tension * (temp_g_x - (C_VOF_NX(c, t) * ND_DOT(C_VOF_NX(c, t), C_VOF_NY(c, t), C_VOF_NZ(c, t), temp_g_x, temp_g_y, temp_g_z)));
}
}
source = (recoil_pressure + marangoni_flow) * C_VMAG(c, t) * D;
dS[eqn] = 0 ;
return source;
}


Any suggestion what could possibly went wrong ? Thanks!

[San_tu_1]

hope it's not too late to answer

Hey friend, I'm glad we're doing the same thing.
I mean they are all doing LPBF numerical simulations with Fluent.
I have encountered your console error code in my previous work.
I did not encounter this error after reducing the number of CPU cores (i.e. CPU count<maximum CPU) in the startup interface of Fluent.

You can try adding the following command to the console interface regarding your issue with C_T-G (c, t)

/solve/set/expert
Save cell residuals for post-processing? [no]
Keep temporary solver memory from being freed? [no] yes
Allow selection of all applicable discretization schemes? [no]

Friendly reminder:
If your surface tension varies with temperature, there is no need to add additional Marangoni force, as the solver will include additional shear stress (Marangoni convection) caused by changes in the surface tension coefficient

I'm sorry, my English learning is very poor. Most of the content above is machine translated. If there are any expression problems, I hope you can understand