[meepokman]

Hi all,

I'm currently trying to use the User-defined Scalar Transport Equations to solve my simulation problem. Essentially, the four scalar transport equations I'm trying to solve are as per the attached picture, for alpha, Ux, Uy, and Uz.

The equations are very similar to what would be solved for a typical species transport equation, however, there is no diffusion term, and it doesn't follow the velocity of the primary phase, as the species has its own momentum source term for Ux, Uy, and Uz, hence I'm not using the species transport module or the multiphase module natively available in Fluent. I was having divergence issues, however, and couldn't understand why.

I then decided to test my setup with a simple test case, where I only solve the first equation, using the primary phase's flow velocities, to test the solution (as per a more typical species transport equation).

I get two different solutions solving for the convective flux in two ways. The UDF I used is as follows:

DEFINE_UDS_FLUX(eqn1_flux,f,t,i)
{
cell_t c0;
cell_t c1;
Thread *t0;
Thread *t1;
real NV_VEC(psi_vec), NV_VEC(A), flux = 0.0;

c0 = F_C0(f,t);
t0 = F_C0_THREAD(f,t);
c1 = F_C1(f,t);
t1 = F_C1_THREAD(f,t);
F_AREA(A,f,t);

if (BOUNDARY_FACE_THREAD_P(t))
{
NV_DS(psi_vec, = , F_U(f,t),F_V(f,t),F_W(f,t),*,rho_water);
flux = NV_DOT(psi_vec,A);
}
else {
NV_DS(psi_vec, = , C_U(c0,t0),C_V(c0,t0),C_W(c0,t0),*,1.0);
NV_DS(psi_vec, +=, C_U(c1,t1),C_V(c1,t1),C_W(c1,t1),*,1.0);
flux = rho_water*NV_DOT(psi_vec,A)/2.0;
}

/* flux = F_FLUX(f,t); */ /* Second way of calculating */
}

If I solve it by taking the dot product of cell velocities and cell face (my mesh is uniform, so I am just using a simple linear average), I get a different solution than if I use the inherent macro F_FLUX provided by FLUENT. The use of F_FLUX yields a solution that is similar to that solved by a species transport equation as provided by FLUENT.

My issue here is if there's a way to improve the solution obtained by taking the actual dot product of cell_centroid values and cell face normal. I'm asking because I can only calculate the convective flux through the faces for the 4-equation case (my original problem) using the cell centroid values (C_UDSI) for my scalars, since FLUENT doesn't seem to store cell face values for any of the scalars. I'm not sure if that's the reason for the divergence observed earlier, but this disparity seems to be a potential reason.

Thank you for all help and advice provided!