Boundary conditions for waterAirflatInterface2d

[CWhalley_1]

Hi all,

I have been trying to set up inlets and an outlet in a model based on the waterAirflatInterface2d example however I have not had any success so far.
I initially tried using,
setInterpolatedVelocityBoundary<T,DESCRIPTOR,BulkDynamics>
at the inlets. This boundary could be initialised, however after the first step it returned ‘-nan’ for the ‘avEnergy’ and ‘avRho’ for each lattice and also gives uMax=0 for each lattice.
As this did not work I have moved to trying to use,
setAdvectionDiffusionTemperatureBoundary<T,DESCRIPTOR>
at the inlets. This boundary condition is not being recognised and an error is showing when I attempt to ‘make’ the example.

Could someone please help me understand where I am going wrong and how I can introduce inlets into the system? At the moment I am only focussing on the inlets however once I have made these boundaries work I will then set up the outlet.

Kind Regards,

Callum

[TimBingert]

Hi Callum,

I want to emphasize again, that this model for thermodynamic multicomponent multiphase mixtures has been introduced to OpenLB without boundary conditions as you can see from the example cases. Even for the inlet, it might need extra work to have a functioning velocity boundary condition. E.g. how does the pseudopotential force work on your bc as there will be some neighbouring nodes that are missing? Or as this model has density ratios, how do you determine the density at the inlet, it needs to be set as well as the velocity as in these multiphase models the density is not uniform everywhere.

If you would just like to have a simulation with water and air without the exact thermodynamics, you could also have a look into the free surface models. They are easier, but have the drawback that you only solve the momentum equation for the liquid. In the future, there will be high density ratio phase field models in OpenLB, too, with which it will be possible to simulate water-air systems too (without the mixture thermodynamics though).

Kind regards
Tim

[CWhalley_1]

Hi Tim,

Thank you for your response, I did come to the same conclusion and have shifted to using the free surface models.

For these free surface models however, I have been unable to implement a cross flow of air to interact with the water jet being injected. I have been trying to use multiple lattices and coupling them with the air flow on one lattice and water flow on the other however I have not had any success.
Is coupling 2 lattices possible for the free surface models?

Thanks for your assistance,
Callum

[TimBingert]

Hi Callum,

the idea of the free surface models is that you are only solving the momentum equations for one phase, i.e. the liquid. If you want to do a crossflow of water and air you need to solve momentum equations for both phases. So a phase field model would be more appropriate for you. There is a so-called “free energy” model in OLB but it only has equal densities for the two phases yet. In the next release, there will be high density ratio phase field models that can be used for your problem.

Best
Tim

[Author]

Posts