species flux calculation for species balance

[wengzf20]

Hello, I am confused about how the species flux is calculated in the do_species_balance feature, and I would be grateful if you could help to provide some explanations.

Take the IsothermalSoretChannel case as an example, at each time step:

1. The Pre-SDC stage

• initTemporals() set m_domainRhoYFlux to zero ($F_{tot} = 0$).

• computeDifferentialDiffusionTerms(AmrOldTime, diffData) include the species diffusion, i.e. $F_{tot} += 0.5*F_{diff}^{old}$.

2. At the last step of SDC (m_sdcIter == m_nSDCmax),

• computeDifferentialDiffusionTerms(AmrNewTime, diffData) include the species diffusion, i.e. $F_{tot} += -0.5*F^{new}_{diff}$.

• computeScalarAdvTerms(advData) includes the species convection flux, i.e. $F_{tot} += \rho U_{mac} Y_i$

• differentialDiffusionUpdate(advData, diffData); includes the species diffusion flux, i.e. $F_{tot} += F^{final}_{diff}$

3. The post SDC stage

• calcDivU(is_initialization, computeDiffusionTerm, do_avgDown, AmrNewTime, diffData) will correct the species diffusion again, $F_{tot} += -0.5*F^{post}_{diff}$.

Finally, the species flux is
$F_{tot} = \rho U_{mac} Y_i + F_{diff}^{final} + 0.5 F_{diff}^{old} - 0.5 F_{diff}^{new} - 0.5 F_{diff}^{post}$

I assumed that the 1st and 2nd terms would be major contributions, and the 3rd and 4th terms would make a correction, but the last term seems somehow overlap with the 4th term.

I plan to consider the diffusion in the inlet boundary condition, and found that the species flux from do_temperols does not match the prescribed species mass flow rate.

[baperry2]

I believe that you are correct and it is a bug that the final term is included in the summation of fluxes. I'll create a PR with a fix.

One thing to note is that with inlet boundaries, unless the inlet covers the entire boundary, you must define a zero_visc function in order to prevent unwanted diffusion through the portions of the boundary that are intended to be walls, for example:

PeleLMeX/Exec/Production/DiffBunsen2D/pelelmex_prob.H

Line 338 in 019098b

static void zero_visc(

[wengzf20]

Great! Thank you, Bruce. Let's move to #609 to fix this issue.