[Kenna]

Hello everyone,
I am looking quite long for a fitting solver for the simulation of a Plasma-Welding-Simulation and I am running out of time for my thesis. I want to simulate the thermodynamic part together with two mixing gases - no electromagnetic effexcts needed. I know buoyantPimpleFoam would be good for one phase and I started with it. But the multiphase solver seem al to be insufficient. I am simulating with argon as the inert welding gas and air as the second gas.

One flaw of the Euler-Solver is, they only allow one phase to be continuous and the other has to be dispersed, meaning also no diffusion-effects, furthermore the can't use the turbulence libraries.

The InterFoam-Solver looked good first too, but the seem mostly only able for liquids have no heat transfer, allow only constant properties for the species (Diffusivity, sutherland-viscosity isn't enough either )

I was wondering if the chemistry solver works somehow with more than one phase. I dont understand them well. But are the chemicals all homogen mixed? is there only a distinction between burnt and not burnt?

I guess I will need to construct my own solver. If yes, could you give me some tips what existing solvers I should use? Maybe InterFoam with buoyant Foam?


I would really appreciate your help.

Kind Regards
Andreas Daun

[wyldckat]

Greetings Andreas,

Pretty complicated thesis, if you only have 6 months or less to perform.
Anyway, there isn't much I can do to help, except to suggest the following:

1. http://openfoamwiki.net/index.php/Handy_links - there you'll find links to workshops and courses, most of which supply presentations, reports and source code.
2. A quick search in Google:
   

   openfoam welding

   lead me to this presentation: http://www.openfoamworkshop.org/2012...SlidesOFW7.pdf - Anton Kidess is this forum member: akidess

Best regards,
Bruno

[akidess]

You want to simulate two mixing gasses, i.e. just a single phase? Then you need an additional scalar transport equation for the concentration. This is fairly simple to add to buoyantPimpleFoam.

[r08n]

There are a few theses/presentations about welding simulations:

http://www.tfd.chalmers.se/~hani/pdf...aLicThesis.pdf
http://publications.lib.chalmers.se/...cal_147262.pdf

but, I guess, you already found them?

[Kenna]

Thank you for you ideas!

wyldckat, I think a further look at those workshops and courses might help me. I already forgot about them again.

Originally Posted by akidess

You want to simulate two mixing gases, i.e. just a single phase? Then you need an additional scalar transport equation for the concentration. This is fairly simple to add to buoyantPimpleFoam.

akidess, thats an interesting Idea. So If I understand correctly, I implement equations very close to the Temperatur/Energy-Equation. I could look at this tutorial about making your own solver with adding Temperatur. How_to_add_temperature_to_icoFoam

From thinking about it, I think those concentrations would move identically to the temperature. Diffusion like thermal conductivities, and both dragged with the velocity field. One would be like high temperature and Zero like low temperature.

r08n, the thesis from MARGARITA SASS-TISOVSKAYA was a great inspiration for me so far, even if I don't need the electromagnetic effects. I guess the people from Chalmers University wouldn't have a problem to solve this.

I got also the advice from my professor to fix the turbulence model from compressibleTwoPhaseEulerFoam. Hard to decide which idea to start on. The Idea from akidees got my big interest. I would only need to add an other thermophysicalProperties file and interpolate the fluid properties regarding to the concentration. But it didn't work. You can read it here http://www.cfd-online.com/Forums/ope...nce-model.html

[akidess]

All you have to do is replace thermal diffusivity with mass diffusivity, and you're there. The equations are identical. Once you've set up the code, you might have to start worrying about boundedness, high Peclet number (if your diffusion coefficient is very small) and perhaps strongly varying material properties (though argon and air shouldn't be *too* different), but that's stuff for a later topic

[Kenna]

I looked now quite a while at the code, implemented a alpha-field with equation in BuoyantPimpleFoam. And I gave up at the sheer complexity of implementation a the second thermoType. (even if I could look it up in compressibleTwoPhaseEulerFoam)

An old idea came back into my mind using one of the combustion solver. They look like they will work and I could kick myself for not realising it earlier.

They calculate multiple gases and use for every gas a thermoType - or "thermoReactionType"

So I will need to learn which solver to choose. The names of the solver from OF2.2.0 are different from OF1.7.0 but I assume, they can still do the same jobs?

So which is the fitting one for my welding application?
(I will also need later to insert a heat source like here: adding-internal-heat-source)

chemFoam (this one not)
engineFoam (is this like coldEngineFoam?)
fireFoam (looks promising)
PDRFoam (does not write down the fraction of the gases)
reactingFoam (is this like rhoReactingFoam?)
XiFoam (only not-burnt and burnt phase?)

So, maybe someone can advice me a fitting one? I will just try with luck some solver until them.

Greetings,
Andreas