[syble]

Hello,

I'm modelling a boundary where one gas species enters at a certain rate while another species exits the boundary at a different rate.

My plan was to write a udf that loops through all the elements on the boundary and for every second face specify the flow into the domain and for every other face specify the boundary out of the domain (at the moment I just have one gas species to simplify the problem and I'm just altering the flow direction).
To access each second face I keep track of the number of faces I have looped over and if this count is odd then the flow is in one direction and if the count is even then the flow is in the other direction.

This is my udf:

DEFINE_PROFILE(flowRate,t,nv)
{
face_t f;
cell_t c0;
int count = 0;
int mod;
int fluxValue = 2.0;

begin_f_loop (f,t)
{
count += 1;
mod = count%2;
printf("count: %d\n", count);
printf("modulus: %d\n", mod);

if(mod == 0) // even
{
F_PROFILE(f,t,nv) = fluxValue ;
printf("EVEN cell \n");
}
else // odd
{
F_PROFILE(f,t,nv) = -1.0*fluxValue ;
printf("ODD cell \n");
}
}
end_f_loop (f,t)
}


I used this udf to define the inlet boundary of a simple pipe where the outlet of the pipe was set to zero relative pressure. I expected that the net flow over the inlet would be zero since the inward flux specified by half the faces would be balanced by the other half of the faces. However when I look at the results the mass flow rate over the inlet is equal to -fluxValue which makes me think that only the negative condition is being set on all the faces in the boundary.

Am I using the begin_f_loop and F_PROFILE correctly? Does this specify the boundary condition on each face?

The simulation converges in about 300 iterations to RMS residuals for all solution variables to 10^-4.

Thanks in advance for any help.