[cuikahlil]

Hello!

I would like to ask a question regarding the appropriate boundary conditions for my simulation. I am simulating a steady granular flow while being fully immersed in ambient fluid. An infinite and continuous flow of particles while completely surrounded by fluid whose velocity and pressure simply reacts to the motion of the particles inside. A diagram of the system is shown in the first image (the encased segment).

Now we find that the velocity profile of the particles are extremely sensitive to the fluid boundary conditions. Once again, the fluid only need to ambient and will only react ti the particle motion. We have two choices for the fluid domain BCs: free atmosphere which is implemented via pressureIntetOutlet on both left and right walls, and periodic boudaries.

The second image shows the Fluid velocity profiles using free atmosphere, while the fourth image is the one using cyclic boundaries. The free atmosphere was considered since tit satisfied the condition of ambience - all velocity and pressure field simply reacted to particle motion since no explicit conditions were set at the walls. Fluid veloity and pressures freely came in and out of the walls as a reaction of the solid moving within (as far as I understand that is what happens). The periodic boundaries were chosen since we the flow is steady state for both solid and fluid phase ans since the solid moves in periodic boundaries, why not the fluid too? Both of them result to different particle velocity profiles and we re not certain which one reflects reality. Note that the moving particles are at most 10mm large in diameter.

I would simply like to know your opinion on the matter and which of the BCs would seem more logical when compared to a real or natural scenario. I would gladly provide more details if necessary.

Thank you!

[FMDenaro]

Periodicity does not depend on a steady condition but it assumes you are simulating a portion of the total volume that repeats itself infinitely in each direction of periodicity. The chosen lenght of periodicity can have impact on the solution.

[cuikahlil]

Yes but I only need the system to reach a steady state - both solid and fluid phase achieving time invariant velocities as a mixture. As long as it allows the mixture to achieve this state, that should be fine right? Would you say that the first condition is more appropriate for the system we want to simulate?

[FMDenaro]

Originally Posted by cuikahlil

Yes but I only need the system to reach a steady state - both solid and fluid phase achieving time invariant velocities as a mixture. As long as it allows the mixture to achieve this state, that should be fine right? Would you say that the first condition is more appropriate for the system we want to simulate?

If you decide to adopt the periodicity, you must be aware of the fact that this condition implies a time-evolving volume of fluid not a spatially evolving.
From your first figure seems you are interested in a spatially evolving case.

[cuikahlil]

Actually I would say my case is simply time-evolving since it is only like I am simulating a particular section of the flow that is in steady-state.