CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   FLUENT (https://www.cfd-online.com/Forums/fluent/)
-   -   Residuals - why are some transport equations not associated with residuals? (https://www.cfd-online.com/Forums/fluent/200916-residuals-why-some-transport-equations-not-associated-residuals.html)

juliahartig April 16, 2018 20:00

Residuals - why are some transport equations not associated with residuals?
 
Hey everyone!

I have a conceptual question. From my understanding, Fluent solves every transport equation by linearizing it using the formula here (equation 25.18-1), which would imply that every linearized equation is associated with a residual. In practice, though, this doesn't seem to be the case. For example, I know that there is a species mass fraction (Y) transport equation for the species model, or a species volume fraction (alpha) transport equation for multiphase flow problems, but I have never seen any residuals for these equations when I enable the models. Am I missing something? Why would these equations not have residuals?

LuckyTran April 16, 2018 22:13

Go to your residual monitors box and see if maybe you unchecked it.

juliahartig April 16, 2018 22:19

LuckyTran, thanks for the reply! Unfortunately I've already checked the residual monitors window and that doesn't seem to be it. All of the residuals are currently being displayed - there is no option to toggle on residuals for the transport equations that I gave examples of above.

LuckyTran April 17, 2018 16:24

There should be residuals.

Go to solution controls => equations =>
And make sure you have not deactivated any solvers.

Otherwise, you must be running some very strange combination of settings. Further debugging would require you to basically post all your solver settings.

juliahartig April 27, 2018 12:30

2 Attachment(s)
I looked at the solution controls window and it appears that all solvers are activated so that shouldn't be an issue (I've attached some screenshots here). After going through the Fluent documentation, I think that perhaps some of the numerical equations I identified are either lumped in with other residuals or their residuals are suppressed. For example, there are no residuals associated with the discrete phase variables (particle mass, particle velocity, particle temperature and particle position) so as far as I can tell, because they are computed using a separate solver, Fluent simply suppresses those outputs.

There are still some that I'm unsure about (for example, the Fluent documentation lists separate PDEs for species concentration transport [the Fickian-diffusion equation] and volume fraction, so why is there only one residual associated with each species?) so I think it's doing something to relate some of these equations which they don't explain in the documentation.

Anyway, it's not terribly important for me to understand this. The program is giving me reasonable results so I don't think there are any problems there. It's just curious to me that I can't follow the solver process using what they've provided!

LuckyTran April 27, 2018 13:35

Discrete phase model is a lagrangian particle tracking, it doesn't use a transport equation.

How many residuals do you expect for species transport? The only variable is the species. So there is one for N2 and one for CH4. I'm guessing these are the species you are transporting? What else are you transporting that needs species?

You don't even have a volume fraction equation for there to be a residual.

Temperature is in the energy equation.

So you see, all your residuals are there.

juliahartig April 28, 2018 18:31

That's true. I guess I was thinking that since some of the discrete particle equations are still differential equations, they would be associated with some kind of "error" that would be printed to the console, but since the equations are all ODEs with time as the only independent variable, I guess the only error would be associated with the time step chosen.

So according to the Fluent user manual, both the mixture and Eulerian multiphase models solve a volume fraction transport equation as shown here, but I've never seen a residual for this, so either the documentation is too outdated and it doesn't do this anymore, or something else is going on.

I think the thing that throws me off about the energy transport equation as shown here is that there are two unknowns (energy E and temperature T) and one transport equation. But now that I'm looking at it, I think they just use tabulated enthalpy values and set E=h for incompressible flow so that temperature becomes the only unknown.

LuckyTran April 30, 2018 01:50

The eulerian multiphase does have a residual for volume fraction and some other things. But that's not what you're simulating... If you want the Eulerian multiphase models and options, then turn them on. Stop playing with this discrete phase stuff.

If you derive it yourself you'd find that the energy equation as 3 unknowns (internal energy in time-derivative, enthalpy in the divergence term, and laplacian of temperature in the conduction term). For the most part, Fluent uses the temperature-based formulation, which requires the user to provide the specific heat as a function of temperature. When using the pressure based solver, the pressure work (the p/rho) and kinetic energy term (v^2/2) are not enabled by default. You could enable them though. For the density based solver they're always enabled.

But regardless, that doesn't change that the transported property is energy so the residual should be calculated in terms of energy imbalance (energy is conserved, not temperature).

juliahartig May 2, 2018 17:11

Yeah after reading your responses, my questions sound kind of silly to me haha. Not sure why I got so many of the Fluent models confused with each other.

I realized the reason I was expecting a volume fraction residual is because I was confusing the species mixture and multiphase mixture models, but they are fundamentally different things.


All times are GMT -4. The time now is 21:49.