[emmakateh]

Hi all, I am working on a code that calculates the gradient of a UDSI in fluent using udf. When I use this code and export the gradients on nodes and cell centers, I cannot figure out how fluent is determining the gradient so I cannot validate and verify the gradient. I have looked in the manual at the three ways it calculates the gradients, but when I try to mimic those, I get different results.

I am using a simple 3x3x3 cube to try to validate the results. Can anyone explain if my code is wrong or how fluent is extrapolating points outside the mesh?

#include "udf.h"
#include "sg.h"
#include <math.h>

enum
{
HMAG,
N_REQUIRED_UDS
};

enum
{
GHX,
GHY,
GHZ,
N_REQUIRED_UDM
};

DEFINE_INIT(HFIELD, domain)
{
Thread *t;
cell_t c;
face_t f;

//Calculating T for Cell Centers
thread_loop_c(t,domain)
{
begin_c_loop(c,t)
{
C_UDSI(c,t,HMAG) = 9.5;
}
end_c_loop(c,t);
}

thread_loop_f(t,domain)
{
if (THREAD_STORAGE(t,SV_UDS_I(HMAG))!=NULL)
begin_f_loop(f,t)
{
F_UDSI(f,t,HMAG) = 20;
}
end_f_loop(f,t);
}

//Calculating T Derivatives
int p;
for (p=0; p<n_uds; p++)
{
uds_derivatives(domain,p);
}


thread_loop_c (t,domain)
{
begin_c_loop (c,t)
{
C_UDMI(c,t,0) = C_UDSI_G(c,t,HMAG)[0];
C_UDMI(c,t,1) = C_UDSI_G(c,t,HMAG)[1];
C_UDMI(c,t,2) = C_UDSI_G(c,t,HMAG)[2];
}
end_c_loop (c,t)
}

thread_loop_f (t,domain)
{
if (THREAD_STORAGE(t,SV_UDS_I(HMAG))!=NULL &&
T_STORAGE_R_NV(t->t0,SV_UDSI_G(HMAG))!=NULL)
{
begin_f_loop (f,t)
{
F_UDMI(f,t,0) = C_UDMI(F_C0(f,t),t->t0,0);
F_UDMI(f,t,1) = C_UDMI(F_C0(f,t),t->t0,1);
F_UDMI(f,t,2) = C_UDMI(F_C0(f,t),t->t0,2);
}
end_f_loop (f,t)
}
}
}



void uds_derivatives(Domain *domain, int n)
{
MD_Alloc_Storage_Vars(domain, SV_UDSI_RG(n), SV_UDSI_G(n), SV_NULL);
Scalar_Reconstruction(domain, SV_UDS_I(n), -1, SV_UDSI_RG(n), NULL);
Scalar_Derivatives(domain, SV_UDS_I(n), -1, SV_UDSI_G(n), SV_UDSI_RG(n), NULL);
return;
}

[AlexanderZ]

shouldn't you get data after computation?
DEFINE_INIT(HFIELD, domain) is a macro to initialize case, gradients are not calculated at that moment, however, I could be wrong

[emmakateh]

once I validate the gradient it will be used in a DEFINE_SOURCE loop as I need the magnetic field established to run my simulation to see how the fluid reacts with the magnetic field. This is how I have seen other udfs calculate the gradient of a scalar