[koad]

I have a a Fluent simulation of CVD on the inner wall of a pipe. It seems to work, but when I try to calculate the total mass flow to the pipe wall, I get strange answers.

The simulation is fairly simple. 2D Laminar flow with multi-component diffusion in a long, straight pipe. A carrier gas and a reactive species (call it AB) come in the inlet. I'm using a reaction boundary condition at the wall. The only chemical reaction is on the wall of the pipe:

AB(g) -> A(s) + B(g)

Species A deposits on the wall, and B diffuses into the gas. Only a few percent of AB is consumed, so at the outlet there is a mix of the carrier gas, AB, and B.

The problem comes in post-processing. If I use Results->Reports->Surface Integrals->Mass Flow Rate to calculate the total mass flow at each boundary, I get something like

Inlet: 1
Outlet: -0.98
Wall: -0.02

That makes sense. But if I use Results->Reports->Surface Integrals->Flow Rate->Species->Mass Fraction of Species B to calculate the mass flow of B to the boundaries I get something like

Inlet: 0
Outlet: -0.01
Wall: -0.00001

So Fluent is saying that species B is flowing out of the system but not into it, which would violate mass conservation.

Summing the mass flows for the individual species gives the same answer as calculating the total mass flow, so that is consistent. At the wall Fluent achieves this by calculating a net flow of carrier gas to the wall, even though it is not involved in the reaction.

I've tried doing the same calculations in CFD-Post, using the massFlow and massFlowInt functions in a table, and I get the same answers.

Does anyone have any ideas on what I might be doing wrong? Or is this a defect in Fluent? I have tried running the simulation for more iterations to reduce the residuals, and decreasing the mesh size, but neither changes the answer.

On a related note, I tried changing the surface washcoat factor on the wall from 1 to 2. This doubles the mass flow of species B to each boundary, as it should, but mass is still not conserved. Also, the total deposition rate of species A, as computed with Results->Reports->Surface Integrals->Integral->Species->Surface Deposition Rate of A, is independent of the washcoat factor. So apparently Fluent doesn't take the washcoat factor into account when doing surface integrals.

[koad]

OK, so I think I figured out the problem for myself. The problem is that

Results->Reports->Surface Integrals->Flow Rate->Species->Mass Fraction of Species B

does not integrate the mass flow of species B. It calculates the integral of the total mass flow time the mass fraction of species B. Which is NOT the same thing if there is diffusion. With diffusion, the mass flux of species B is

rho_B*v_B = Y_B(rho*v) - rho*D_B*grad(Y_B)

where rho and v are the total mass density and velocity, rho_B and v_B are the mass density and velocity of species B, and Y_B and D_B are the mass fraction of species B. So if D_B = 0 you can get the mass flux of species B by multiplying the total mass flow by Y_B, but otherwise no.

So this begs the question: Is there any way to calculate the mass flux of an individual species in Fluent or CFD-Post when diffusion is important? Do I have to write a UDF?

[seanread]

Did you have any luck with this? I am also trying to compute species flux at the wall in fluent

[koad]

Sorry, no. I contacted ANSYS and they confirmed that it is a bug that they will fix at some point.

[LuckyTran]

I don't know what the status of the bug is but a workaround is to use a custom field function for the integrand and integrate the custom field function.