Solid and immersed solid

ghorrocks · February 10, 2020, 16:28

Here's a quote from the CFX documentation:

Quote:

For transient cases, immersed solids do not interact properly with fluid domains that involve compressibility or multiphase flow.

In other words, you cannot do what you are proposing to do.

Gert-Jan · February 10, 2020, 16:42

I am not really interested in the exact geometry (and prefer not to open a zip file from this forum from someone I don't know at all).

I'd rather see a sketch/drawing of the geometry and the physics that you want to include. Movements, fluids, solid, forces, etc. And show what the fuid is that is not allowed to move. Then we can think of how to model such a thing.

Ali Di · February 10, 2020, 18:25

Dear Glenn,

Actually, I choose the immersed solid because I could rotate it around three axes (x, y and z) while it moves along these axes (x, y and z).

Therefore, if I cant do such the simulation with immersed solid, can I do that with the solid domain (since, solid domain can interact fluid domain properly), Or you recommend me to use another application such as Fluent or COMSOL?

Ali Di · February 10, 2020, 18:32

Quote:

Originally Posted by Gert-Jan

I am not really interested in the exact geometry (and prefer not to open a zip file from this forum from someone I don't know at all).

I'd rather see a sketch/drawing of the geometry and the physics that you want to include. Movements, fluids, solid, forces, etc. And show what the fuid is that is not allowed to move. Then we can think of how to model such a thing.

Dear Gert,

I attached a sketch of the geometry. And about physics of the problem, I should say that there is no complicated conditions. I just want to simulate the movement of water and air inside the immersed solid, while immersed solid moves with sinusoidal movement.

Gert-Jan · February 11, 2020, 01:43

In one of the previous threads you said: "the two-phase fluid doesn't move with movement of immersed solid". If you think it won't move, than ignore it and include it in the solid. Only make a correction for the density.

But now you say you want to model "the movement of water and air inside the immersed solid, while immersed solid moves with sinusoidal movement".

Nevertheless, If you want to model fluid inside and outside simultaneously, I would not do that using immersed solids because of Glenn's comment. Probably through deforming meshing.
Moreover, if you know the sinusoidal movement, then ignore the water on the outside, and only model the two- phase flow inside.

urosgrivc · February 11, 2020, 02:58

hello,

Do it like this:
You just need ONE domain for this problem
Get rid of the outermost domain and the solid one and set the air and water in the same domain, just a simple cylinder with proper initial conditions (both are fluids) there is no solid in this example if I understand corectly.

Then just make a cylinder half filed with water half air and set outer walls to a sinusoidal velocity (you don't need a solid for this) include gravity and correct models maybe even surface tension when you see it is working.

begin with a really small timestep or even a steady-state solution because two-phase simulations like this tend to explode at the beginning (first few timesteps are problematic).

You will need to play around with the models to get good results as air will be mixed with water a lot

Consider using symmetry, this will help the computational speed, if the movement is just sin up and down then this is an axisymmetric problem and a 2.5D can be used this would be very fast to compute, doing this in full 3D will be a massive job.

Regarding the mesh, I don't think you will see a lot of improvement in CFX using hex mesh. Hex mesh or even pure 2D axisim in Fluent would make the simulation a lot better but CFX isn't affected much. Do put some inflation on the walls though.
Have fun and experiment...

Ali Di · February 11, 2020, 05:59

Thank you Gert,

Yeah, I want to simulate movement of water and air, but the fluid domain (water and air) doesn't move correctly with immersed solid. Therefore I cant simulate it correctly.

About considering the fluid domain as a solid domain or immersed solid. In my case, I suppose it doesn't work. Because, I need a 3D render of water.Volume Fraction, which is not available for solid domains.

And about ignoring outside space and immersed solid. It doesn't work too, because I know the exact movement of the solid (immersed solid), not movement of the fluid (water and air). Actually, this is the movement of the solid which acts on fluid.

urosgrivc · February 11, 2020, 06:09

Stop talking about immersed solid, what immersed solid do you have?
The container? or what do you think an immersed solid is in your case?

The container is not part of the fluid simulation, just the velocity of the walls of the inside of the container is the part of CFD simulation.

Again this is a ONE domain multiphase simulation

"And about ignoring outside space and immersed solid. It doesn't work too, because I know the exact movement of the solid (immersed solid), not movement of the fluid (water and air). Actually, this is the movement of the solid which acts on fluid."

again this is not an immersed solid
Only Walls act on the fluid not the solid as a volume so there is no solid domain needed if you know the movement of the walls

and why did you have additional "space" in Attached Thumbnails, What does that do?
I think that you need to rethink this or I dont understand what you are doing

Is your plan to simply mix the air and water together by moving the container up and down?

Ali Di · February 11, 2020, 06:11

Thank you urosgrivc for your complete and helpful explanation.

Ali Di · February 11, 2020, 06:17

So, I need to simulate movement along x, y and z axes and rotating around these axes (x, y and z), simultaneously. Do you think that I can do this with just one domain (fluid domain)?

Ali Di · February 11, 2020, 06:30

Dear urosgrivc,

The outer space is introduced, because the container need to immerse inside a fluid domain. And if I don't introduce it, CFX considers the fluid inside the container as the fluid domain.

Moverover, mixing water and air is not my plan. I need to study on sloshing of water.

urosgrivc · February 11, 2020, 06:33

isn't this exactly what you are looking for?

https://www.youtube.com/watch?v=oIFfPEWHkxw

google -> vof sloshing tank cfx

You need to set the domain to move so goolge-> domain motion cfx, mesh motion cfx

You need to "fake" the container as it is not needed

Ali Di · February 11, 2020, 06:40

Yes, this is kind of what I want to do. But as you can see in the video the fluid domain just moves along x, y and z axis. But, I want to rotate the fluid around x, y and z axes simultaneously, too.

urosgrivc · February 11, 2020, 06:46

I am sure it is possible to move it in any way you like.

I mean if 1+1 works why wouldn't 2+3? if you know what I mean
it is the same thing just in other directions

Ali Di · February 11, 2020, 06:51

Thank you dear urosgrivc,

There should be a way to do what I want. I just need to find it.

February 11, 2020, 02:58		#26
urosgrivc Senior Member urosgrivc Join Date: Dec 2015 Location: Slovenija Posts: 365 Rep Power: 11	hello, Do it like this: You just need ONE domain for this problem Get rid of the outermost domain and the solid one and set the air and water in the same domain, just a simple cylinder with proper initial conditions (both are fluids) there is no solid in this example if I understand corectly. Then just make a cylinder half filed with water half air and set outer walls to a sinusoidal velocity (you don't need a solid for this) include gravity and correct models maybe even surface tension when you see it is working. begin with a really small timestep or even a steady-state solution because two-phase simulations like this tend to explode at the beginning (first few timesteps are problematic). You will need to play around with the models to get good results as air will be mixed with water a lot Consider using symmetry, this will help the computational speed, if the movement is just sin up and down then this is an axisymmetric problem and a 2.5D can be used this would be very fast to compute, doing this in full 3D will be a massive job. Regarding the mesh, I don't think you will see a lot of improvement in CFX using hex mesh. Hex mesh or even pure 2D axisim in Fluent would make the simulation a lot better but CFX isn't affected much. Do put some inflation on the walls though. Have fun and experiment... Last edited by urosgrivc; February 11, 2020 at 04:35.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Immersed Solid and FSI	Pierrm	CFX	0	March 26, 2018 05:36
Immersed Solid Valve Expressions	shzinwoyk	CFX	2	January 5, 2018 00:15
Flow over a flat plate as an immersed solid	hamed.majeed	CFX	4	September 8, 2016 14:40
Immersed solid, vs solid-fluid interaction to determine forces on a submerged solid	amrbekhit	CFX	0	January 8, 2015 16:39
Immersed Solid Momentum Sources inside function	Martinw	CFX	1	October 10, 2013 17:14

February 10, 2020, 16:42		#22
Gert-Jan Senior Member Gert-Jan Join Date: Oct 2012 Location: Europe Posts: 1,885 Rep Power: 27	I am not really interested in the exact geometry (and prefer not to open a zip file from this forum from someone I don't know at all). I'd rather see a sketch/drawing of the geometry and the physics that you want to include. Movements, fluids, solid, forces, etc. And show what the fuid is that is not allowed to move. Then we can think of how to model such a thing.

February 10, 2020, 18:25		#23
Ali Di New Member Ali Join Date: Dec 2019 Posts: 18 Rep Power: 6	Dear Glenn, Actually, I choose the immersed solid because I could rotate it around three axes (x, y and z) while it moves along these axes (x, y and z). Therefore, if I cant do such the simulation with immersed solid, can I do that with the solid domain (since, solid domain can interact fluid domain properly), Or you recommend me to use another application such as Fluent or COMSOL?

February 11, 2020, 01:43		#25
Gert-Jan Senior Member Gert-Jan Join Date: Oct 2012 Location: Europe Posts: 1,885 Rep Power: 27	In one of the previous threads you said: "the two-phase fluid doesn't move with movement of immersed solid". If you think it won't move, than ignore it and include it in the solid. Only make a correction for the density. But now you say you want to model "the movement of water and air inside the immersed solid, while immersed solid moves with sinusoidal movement". Nevertheless, If you want to model fluid inside and outside simultaneously, I would not do that using immersed solids because of Glenn's comment. Probably through deforming meshing. Moreover, if you know the sinusoidal movement, then ignore the water on the outside, and only model the two- phase flow inside.

February 11, 2020, 05:59		#27
Ali Di New Member Ali Join Date: Dec 2019 Posts: 18 Rep Power: 6	Thank you Gert, Yeah, I want to simulate movement of water and air, but the fluid domain (water and air) doesn't move correctly with immersed solid. Therefore I cant simulate it correctly. About considering the fluid domain as a solid domain or immersed solid. In my case, I suppose it doesn't work. Because, I need a 3D render of water.Volume Fraction, which is not available for solid domains. And about ignoring outside space and immersed solid. It doesn't work too, because I know the exact movement of the solid (immersed solid), not movement of the fluid (water and air). Actually, this is the movement of the solid which acts on fluid.

February 11, 2020, 06:09		#28
urosgrivc Senior Member urosgrivc Join Date: Dec 2015 Location: Slovenija Posts: 365 Rep Power: 11	Stop talking about immersed solid, what immersed solid do you have? The container? or what do you think an immersed solid is in your case? The container is not part of the fluid simulation, just the velocity of the walls of the inside of the container is the part of CFD simulation. Again this is a ONE domain multiphase simulation "And about ignoring outside space and immersed solid. It doesn't work too, because I know the exact movement of the solid (immersed solid), not movement of the fluid (water and air). Actually, this is the movement of the solid which acts on fluid." again this is not an immersed solid Only Walls act on the fluid not the solid as a volume so there is no solid domain needed if you know the movement of the walls and why did you have additional "space" in Attached Thumbnails, What does that do? I think that you need to rethink this or I dont understand what you are doing Is your plan to simply mix the air and water together by moving the container up and down?

February 11, 2020, 06:11		#29
Ali Di New Member Ali Join Date: Dec 2019 Posts: 18 Rep Power: 6	Thank you urosgrivc for your complete and helpful explanation.

February 11, 2020, 06:17		#30
Ali Di New Member Ali Join Date: Dec 2019 Posts: 18 Rep Power: 6	So, I need to simulate movement along x, y and z axes and rotating around these axes (x, y and z), simultaneously. Do you think that I can do this with just one domain (fluid domain)?

February 11, 2020, 06:30		#31
Ali Di New Member Ali Join Date: Dec 2019 Posts: 18 Rep Power: 6	Dear urosgrivc, The outer space is introduced, because the container need to immerse inside a fluid domain. And if I don't introduce it, CFX considers the fluid inside the container as the fluid domain. Moverover, mixing water and air is not my plan. I need to study on sloshing of water.

February 11, 2020, 06:33		#32
urosgrivc Senior Member urosgrivc Join Date: Dec 2015 Location: Slovenija Posts: 365 Rep Power: 11	isn't this exactly what you are looking for? https://www.youtube.com/watch?v=oIFfPEWHkxw google -> vof sloshing tank cfx You need to set the domain to move so goolge-> domain motion cfx, mesh motion cfx You need to "fake" the container as it is not needed

February 11, 2020, 06:40		#33
Ali Di New Member Ali Join Date: Dec 2019 Posts: 18 Rep Power: 6	Yes, this is kind of what I want to do. But as you can see in the video the fluid domain just moves along x, y and z axis. But, I want to rotate the fluid around x, y and z axes simultaneously, too.

February 11, 2020, 06:46		#34
urosgrivc Senior Member urosgrivc Join Date: Dec 2015 Location: Slovenija Posts: 365 Rep Power: 11	I am sure it is possible to move it in any way you like. I mean if 1+1 works why wouldn't 2+3? if you know what I mean it is the same thing just in other directions

February 11, 2020, 06:51		#35
Ali Di New Member Ali Join Date: Dec 2019 Posts: 18 Rep Power: 6	Thank you dear urosgrivc, There should be a way to do what I want. I just need to find it.