[donny1991]

hey guys,

im new in openfoam. i have to simulate two gases in a melting furnance. i know i have to use the buoyantBoussinesqSimpleFoam, but i dont know how to integrate two different gases in the hotRoom tutorial. the gases are not reacting. The temperature difference accure natural convection. The Temperature of the floor is 1023 K an the ceiling 623 K.

Please i need your help.

regards
Donny

[Tobi]

Dear donny,

first hint, buossinesq approximation is in my opinion the wrong choice. You should take the buoyantSimpleFoam but to clear my mind just one question, why you think boussinesq is the right approach?

[donny1991]

Hey Tobi,

thank you for the answer. i read in another forum that this solver is the right choice for natural convection. Therefore i tried it just for one gas and it worked, but i do not know how to do it with two gases.

By the way, you have a privat message from me.

best regards
Donny

[Tobi]

Hey guys,

I was writing via messages with donny and now I will summarize.

• buoyantBoussinesq...Foam are using the Boussinesq-Approximation , you only should use them if the density change in your fluid is not high and the temperature range (dT) is low. Check out some references... e.g. Ferziger & Perić
• Donny want to implement a second gas, therefore the very simple approach would be to implement a passive scalar "S" that is limited between 0 and 1. (like alpha in VOF). This scalar represent the two gases (S = 0 -> gas1; S = 1 -> gas2; 0 < S < 1 -> mixture of gas1 and gas2)
• After that, you have to implement all thermodynamic data for gas1 and gas2 into the solver. Therefore checkout the createField.H and create new fields for rho1, rho2, nu1, nu2 (if incompressible assumption) or mu1, mu2 etc.
• The next step is to create the actual thermodynamic variables based on gas1 and gas2 and the passive scalar S:

rho = rho1 * S + rho2 * (1-S);

• You have to choose if you use openFOAM thermodynamics (then you normally should have two) or you implement your own polynoms (I think you also could use icoPoly.... for that). That means you have either to create a second thermodynamic object or you decouple all thermodynamic calculation with your polynoms.

This model would be nice to handle and very easy. Another idea at the moment would be, to use some combustion solver and unset the reaction (kinetics). Once I did something like that with some old collegue who was exactly doing the same. The result between the combustion solver and the solver mentioned above was very close and a lot of faster.

The easiest approach would be: implement only one passive scalar and use one fluid inside. Doing this leads to a first guess but hence we only use thermodynamics out of one gas, you will be not as accurate as in the method I mentioned above.

If the thermodynamic data are very similar (cp, rho, nu, mu ...) then the passive scalar should work fine.


This would be my suggestions.

[donny1991]

Hey Tobi,

thank you for this helpful answer. Im now using the reactionFoam solver without reaction and only with O2 and N2.
I changed all BC's and the properties for my case. I tried to run the simulation but i got this.


Reading g
Creating reaction model

Selecting combustion model PaSR<psiChemistryCombustion>
Selecting chemistry type
{
chemistrySolver ode;
chemistryThermo psi;
}

Selecting thermodynamics package
{
type hePsiThermo;
mixture reactingMixture;
transport sutherland;
thermo janaf;
energy sensibleEnthalpy;
equationOfState perfectGas;
specie specie;
}

Selecting chemistryReader foamChemistryReader
chemistryModel: Number of species = 2 and reactions = 1
Selecting ODE solver KRR4
Reading field U

Reading/calculating face flux field phi

Creating turbulence model.

Selecting turbulence model type RASModel
Selecting RAS turbulence model kEpsilon
Floating point exception: 8

i dont know how to fix this problem.

Under the following link you can download my case. thank you!!

meltingfurnance.zip

[Tobi]

Hi,

maybe I will check it but as I told above, the bad thing using a reaction solver is, that it is soooo slow. Finally I am not sure if you have buoyancy included. So you have to check it yourself. My suggestion to you, build your own solver. Its so much faster, clean (you know whats going on) and really not a big deal to implement sth. like that into OpenFOAM.