[CFDParty]

Hi all,

I am try to simulate the filling process of a balloon with water. The data that I have is a collection of frames showing the wall movement across time and the inlet velocity at these frames. However, doing a dynamic wall simulation gives me errors and i'm trying to find a way to seperately simulate the flow state at that time point, but in a static simulation for every frame.

But I can't get my head around how this is correctly done. In this case, I have an inlet, wall and no outlet. The balloon is already filled with water at frame 0 and keeps on filling for the next frames. In a static simulation this means that the pressure should increase as more fluid is flowing in, while the volume doesn't change. This is not the case in the filling balloon, as the balloon will expand because of the flow.

So my question: Is there a way to simulate the flow state for every frame, seperately in a static simulation. Or can I incorporate the elastic properties of the wall inside the fluid, eg. via compressibility?

[FMDenaro]

Originally Posted by CFDParty

Hi all,

I am try to simulate the filling process of a balloon with water. The data that I have is a collection of frames showing the wall movement across time and the inlet velocity at these frames. However, doing a dynamic wall simulation gives me errors and i'm trying to find a way to seperately simulate the flow state at that time point, but in a static simulation for every frame.

But I can't get my head around how this is correctly done. In this case, I have an inlet, wall and no outlet. The balloon is already filled with water at frame 0 and keeps on filling for the next frames. In a static simulation this means that the pressure should increase as more fluid is flowing in, while the volume doesn't change. This is not the case in the filling balloon, as the balloon will expand because of the flow.

So my question: Is there a way to simulate the flow state for every frame, seperately in a static simulation. Or can I incorporate the elastic properties of the wall inside the fluid, eg. via compressibility?

Using water, you have an almost incompressible fluid, you can get a steady (at each frame) solution. The configuration of the elastic walls must be considered.

However, you could mimic a steady configuration if you prescribe a uniform "transpiration" velocity (normal to the fixed walls) that balances exactly the flow-rate in inflow.

[arjun]

Would not a surface tension type approach do it, with strong surface tension coeff?

[CFDParty]

Thank you, so are you suggesting I should do an FSI?

Your second approach also seems really interesting! From a quick search, I dont seem to find how to to this? I'm using fluent.

[arjun]

Originally Posted by CFDParty

Thank you, so are you suggesting I should do an FSI?

Your second approach also seems really interesting! From a quick search, I dont seem to find how to to this? I'm using fluent.

You just use VOF method and the balloon like a big droplet. At certain values of surface tension coefficient the bubble starts to behave like a balloon.

The calculation will require very small dt, because of strong body forces.

[arjun]

I did run a small 2d calculation with surface tension coeff = 1E7 . The idea in principal does work (kind of).


PS: I used wildkatze.

[FMDenaro]

This kind of problem is quite similar to that one studied by Peskin et al. in the immersed boundary method. There are, however, issues in ensuring mass conservation, as well as the need of considering the real mass of the membrane. This fact can be afforded in the immersed interface method.

[andy_]

Originally Posted by CFDParty

So my question: Is there a way to simulate the flow state for every frame, seperately in a static simulation. Or can I incorporate the elastic properties of the wall inside the fluid, eg. via compressibility?

Does the balloon fill rapidly with significant inertia effects or does it fill slowly and essentially pass through a sequence of steady state solutions with no significant inertia effects? If the latter then this is often referred to as quasi-steady state or filling and emptying if you wish to search on how this has typically been tackled in the past.

Another relevant question is the information you wish to extract from the simulation. Some approaches can help or hinder what you wish to be accurately represented in the simulation.

[CFDParty]

Originally Posted by FMDenaro

This kind of problem is quite similar to that one studied by Peskin et al. in the immersed boundary method. There are, however, issues in ensuring mass conservation, as well as the need of considering the real mass of the membrane. This fact can be afforded in the immersed interface method.

Thank you all again for your input. This seems indeed like an accurate approach to simulate the interaction with the wall. However, I am trying to avoid complex FSI simulations, at least for now. The reason for this is that I have all the frames over time and in this case could just go for a dynamic wall simulation with prescribed motion.

I have the inlet flow velocity, and wall position at separate frames over time and the goal is to estimate the flow state in separate simulations for this time points.

The filling of the balloon is a transient process, having changing inlet velocity and changing wall movement. I thought of maybe doing a transient simulation where I keep the changing velocity at inlet (which is the same for every simulation), but with a static wall(which changes for every simulation according to the time point). Then for every simulation extracting the flow state at the time point of that frame.

[FMDenaro]

Originally Posted by CFDParty

Thank you all again for your input. This seems indeed like an accurate approach to simulate the interaction with the wall. However, I am trying to avoid complex FSI simulations, at least for now. The reason for this is that I have all the frames over time and in this case could just go for a dynamic wall simulation with prescribed motion.

I have the inlet flow velocity, and wall position at separate frames over time and the goal is to estimate the flow state in separate simulations for this time points.

The filling of the balloon is a transient process, having changing inlet velocity and changing wall movement. I thought of maybe doing a transient simulation where I keep the changing velocity at inlet (which is the same for every simulation), but with a static wall(which changes for every simulation according to the time point). Then for every simulation extracting the flow state at the time point of that frame.

The best approach would be the FSI formulation. If you want to try a series of steady approach your problem is ensuring mass conservation.

Approaching water as incompressible, you need to have normal velocity of the flow leaving the domain. Otherwise set a sink into the domain that balances the inlet. But I would try the former approach. Spread your flow rate in inlet along all the walls as outflow.