[francesco12]

Hello,

I am performing a simulation of the advective-diffusive mass transfer for the oxygen in a fluid.
I have a bifurcating pipe (an artery and a vein), with a velocity inlet, a velocity outlet, and a pressure outlet as boundary conditions.
I'm using the passive scalar model to simulate the oxygen, and I would like to impose the normal concentration gradient, on the two outlets,
but since I'm using a velocity and a pressure as boundary conditions, I can only impose a specified value on those faces.
For what concerns the other boundaries, I have a concentration sink on the wall of the pipes (with a specified value of 0), and a constant concentration on the inlet.
I have also tried to use the flow-split condition, but sometimes the solution start diverging because of a recirculation on the cells of the boundary layer, even if I try extend the pipe to damp the recirculation.
The Schmidt number I'm using is quite high (2700), and I noticed that if I change the constant value on the outlets from 0 to 1 the solution for the passive scalar remains the same, on the other hand, imposing the gradient would be conceptually more correct.
Does anyone have any suggestion?

Thanks in advance,

Francesco

[siara817]

I hope I understood correct. If you want to have a normal distribution of a phase (oxygen) you can use lagrangian multiphase mode. Then the second phase will be defined as a normal distribution.