Meshing two-phase turbulent flow
 

Hey folks!

I am simulate two-phase turbulent jet flow with VOF-Method. When I use a coarse mesh the simulation converge whereas the solution does not look physically correctly. But when I use a fine mesh, after a while of evaluation, the simulation aborts whereas the interface is captured fine until the time of abort. So is there something specific which I have to consider by meshing the domain? I am using a VOF-Method with and k-e realizable model.

cheers,

Claus

turbulent VOF
 

Hi there,
what kind of VOF scheme are you using? turbulent VOF may provide more smooth surface and the disturbances at the interface may not be simulated that well using RANS modeling. a les type model will provide more accurate results here. btw, hav you tried running a laminar case?

regarding the convergence, i guess you just need to look into the setting / time stepping (variable / fixed ..courant number) etc.

mesh refinement imposes restriction on the Courant number and hence time stepping ..you may want to work on this direction..

/CFDtoy

Are you running steady-state or transiently. Would suggest you need to run transiently to resolve this problem. Also what types of fluids do you have? velocities etc.

Hey peoples!

Currently, i am running it as a transient simulation whereby I am using the interFoam solver form OpenFOAM. The thing is that I am intend to simulate a jet srpaying into a vessel. I know preori which direction the jet takes, therefore, in this direction the mesh is set more dense. However, when I do so after a while of simulation time the simulation aborts. I assume it has something to do with the pressure because when the pressure evaluation step is supposed to start then the simulation stops. When the mesh is coarse the simulation runs with no problems to the end, however, I do not have a physically correct jet, i.e. the jet is distort.

I assume for a two phase flow the mesh has to be set skilful in order to avoid distortions of a phase and pressure problems. Note that the pressure in a two phase model is defined by p*:=p0+rho_i g where p0 is the total pressure, rho_i the density of pahse i and g the garvity.

Cheers,

Claus

Additionally, I want to mention that when the simulation is going the collapse the time steps, adjusted by the Courant No., are getting smaller and smaller and smaller.

Cheers,

Claus

hi everybody,

i' ve a problem simular as idrama, i want to simulate a nozzle problem includes water jet flow in air, 3d problem, i used unstructured mesh, in 2d section model, all the multiphase models are working. but in 3d, same models are not converged, i tried to solve unsteady and used Fluent, after about 30 iteration the continuity goes to 10^12. i take courent number default value, 0.25
i'm waiting for your helps,

thanks

bugra

Hey Burga!

Do you realy mean continuity? I think that is total different to my one because positive continuity means that you have production of mass in your domain. In my case continuity is fine. The problem I have is that the simulations runs into a singularity, i. e. the velocity is going to blow up. The reason for that is: I do not know.

Cheers,

Claus

Hi Claus,

you' re right, that's different, i'm still waiting for replys,

thanks

Bugra

Hi idrama

i am a fresher in OF..I also doing the same job just like you. but with a little difference, the turbulet model i need is LES. and the liquid phase is Non_newtonian liquid in my job. i want to know which solver i should use, and i hope i can get some adivce from you about LES+VOF+Non_newtonian. Thanks a lot ! any help whould be much appreciated!

zhenyan

Hi,

as far as I see, you should use interFoam solver. In the turbulenceProperties dictionary, you just need to select LESModel. In constant/LESModel, e.g. Smagorinsky (You must look which is suitable for your problem), you need to select a LES model and a filter (Try to get Versteeg, Ferziger or Pope). Finally, in the dictionary transportProperties you have to select a model for non-Newtonian fluids: CrossPowerLawCoeffs or BirdCarreauCoeffs. For you problem, you should look for papers which are related to your problem to find best values for mu mu_inf etc.

I hope this helps!

HI idrama

Thanks a lot for your help. I will have a try according to your idea. thank you very much!


zhenyan