Learning Forum

[Fabricio1]

Dear Mr

I was able to produce a rigid body immersed in fluid and make it move in the -x direction but after the simulation progresses by various time increments it stops by error after 1.2 sec. Can you tell me if I need to improve the mesh around my body or other advice to address this issue?

Best Regards

 

[peteroznewman]

Is it modeled in Fluent or CFX?

I recommend you investigate the Immersed Body method available in CFX.

[Fabricio1]

I am using CFX  the body has to be semi-immersed so it can create a wave by moving like a wedge the way it is used at Kelly Slater's wave pool,. In that case the immersed body can be used?

[peteroznewman]

Kelly Slater's wave pool is awesome!

 

I don't know if the CFX immersed body works with a free surface, you will have to read the documentation.

[Fabricio1]

Dear friend

 

Are you interested in solving this problem with me?. I already can create a typical wave pool  and can create a wave generated by a moving sphere and can simulate a wedge type body moving semi submmersed. But the simulation stops after 1,2 sec and it would be necessary to accelerate this body up to 4m/s so i need 10 sec of simulation. Any sugestions other than the immersed body?

[peteroznewman]

Dear Fabricio,

I have been learning Fluent, and never used CFX.  I have not yet built a VOF model, and will do that one day this year, but I'm busy with an airflow model right now.

Good luck!

[Fabricio1]

I am sending you an image of that wave pool i am working with

[Rob]

Why does it stop? Does it fail, or did it just run out of time steps? 

[Fabricio1]

It says meshing folding problem

[Fabricio1]

the program inform meshing folding problem

[Rob]

If you look at the last data set how fast does the object move, and if you compare it's speed across a cell with the time step how does it look?

[Fabricio1]

I tried both using a rigid body with 1m / s2 acceleration and also a program-defined body (without defining faces previously) and putting a movement defined by 1m / s2 acceleration. In either case the body moves for just over 1 sec and reports error code 2. The movement is noticeable. I was more successful in moving a submerged sphere, which moved longer and could see the wave generated by the movement, but it also stopped by an error  after a while. I have been more fortunate in leaving my body still and accelerating the water around him with one face defined as inlet and other as outlet.

[Fabricio1]

If you want you can give me your email and I will send you the template via gmail

[peteroznewman]

 ANSYS staff do not open models (I am not staff).

If you use a time step that is 10 times smaller, will that let the simulation continue further?

Do you have to keep the body accelerating?  What if you transition from acceleration to constant velocity?  Would that keep the wave going then use the profile of the floor to shape the wave?

[Fabricio1]

I start with a time increment of 0.01sec and then reduce to 0.005 sec and the simulation goes a little further. However when I make the body stand still and accelerate the water towards the body I can use a 0.025 sec time increment without problems. See attached photo

[Rob]

Which suggests that the moving mesh wants a smaller time step: what happens if you drop the time step by a factor of 10 as Peter suggests?

[Fabricio1]

I am trying a mixed approach by making the rigid body moves at 2m / s and making the water moves at 4m / s and using a time increment of 0.0001sec. In addition I activated the surface tension and increased the atmospheric pressure to 3 atm. Let's see how it goes;

[Fabricio1]

It seems that I have achieved a satisfactory solution by combining a 4m / s velocity of water flow with the movement of a solid using the displacement of its mesh with a sinusoidal function only in the -x direction. With that I was able to run several steps of time and generate an incipient wave that I will now refine.

[Fabricio1]

This was the best modeling integrating a water flow and a sinusoidal motion in the -x direction only. Unfortunately, you can't simulate the  Kelly Slater's wave pool perfection. When I get the body to move alone

at a given speed or acceleration the time increment needs to be minimal and weird effects appear as the water is forced out of inertia some vertical water jets are generated and prevent further analysis. Any additional tips?

[Rob]

You do know you can create an isosurface of VOF at (for example 0.5) to show the free surface?   Have a look at the local flow patterns and mesh, there's usually a reason for weird results. 

[Jaredoac]

Hi Fabricio,

I am learning Fluent in order to develop this kind of projects. If you want, we can try to improve the models together.

My email is jaredoac@gmail.com