Learning Forum

[Paco Sheeran]

Hi,

I'm a newbie to fluent. I am trying to model the melting of a warm object through a block of ice and see how fast it will melt said ice and get to the bottom based on the temperature at the surface of the object. When I simulate this in fluent using solidification/ melting, the liquid fraction shows that ice is melting below the object, but it is not moving, it's basically floating. I checked that gravity is on but it still doesn't move. Is this simulation possible with fluent and if so what is the procedure for it? Thank you

Paco 

[Rob]

Read how the solidification model works: it won't melt to give you floating ice. 

[Paco Sheeran]

 

I’m sorry I might not have explained it very well, or I am not quite understanding your explanation sadly. I’m not exactly looking to simulate ice floating. More so trying to simulate a warm object melting through the stationary ice. The warm object I simulated in the ice stays in position even when the ice around it has melted. It should be falling down with gravity if there is no ice holding it up. At least that was my assumption.

I believe it is possible for Ansys fluent to do this because I have seen the concept published in journal articles such as the one linked below

https://link.springer.com/article/10.1007/s12206-022-1249-5

Does this make sense, thank you

 

[Rob]

Ah, so a falling solid?  OK. Still not easy, but have a look at the 6DOF (possibly 1DOF) solver and moving mesh. 

[Paco Sheeran]

Hi again,

I tried that 6 dof solver and moving mesh and I think I made some progress, thank you. I just keep running into what appears to be a common error which is negative cell volume detected. I have tried what appears to be the typical problem which is too big of a mesh and timesteps too fast but neither of those really seem to work. I attached a picture of the setup. Do you have any thoughts on why this may be happening when trying to do melting/solidification simulations? The other solutions I have found on other forums don't really seem to work and most of those solutions are not for melting/solidification. Thank you

 

[Rob]

I can't remember if full remeshing is availble for hex/quad elements. Check the documentation, you may want to retry with tet/tri. 

[Paco Sheeran]

Ok I tried with both Triangles and Tri and does not seem to be changing anything. Could there be any other source of the error? Thanks

[Rob]

The above are hex cells, so I assume you also tried with tri as asked. If you work out the motion speed how many time steps does it take to cross a cell? 

[Paco Sheeran]

Yeah so I have tried it with all the possible mesh types I see in mesh, I showed the tri on this one I think.

So the motion speed is sort of what I am trying to find with the simulation in the first place so I am not entirely sure how to calculate it. If I do the same simulation without 6dof on though I find that after 5000 time steps of 0.001s the liquid/solid boundary layers appears to have moved about 1.5 cells. So knowing that I am assuming it would take about 3333 times steps to get across the cells currently. Is that a reasonable way to assume that?

[Rob]

More-or-less. However, if you move more than one cell length the risk is you turn the cells inside out: the nodes can't reconnect (remesh) before they're corrupted.  

[Paco Sheeran]

Hi again, sorry for the late response. I have been trying to slowly piece my way through all the possible issues and learn more about fluent.

I think I have figured out why I am getting the error, but I am not sure how to solve it. I took what you said about moving across one cell length and I shrunk the cells near the box small enough that when I start the simulation many of them are already liquid when I leave dynamic meshing off, so it runs just fine then.

When I go to turn on dynamic meshing though, the box for some reason jumps high or low after just a couple of time steps as seen in the preview mesh picture. This causes certain cells to fail, which makes sense. I am just not sure what part of the dynamic mesh settings is causing the box to jump. I want the box to slowly move down as the liquid fraction of the cells below it changes but it seems to just ignore that. I left pictures of my dynamic mesh setup that I think should work. The “box_inside” is the wall of the center box while the “box_itself” is the ice I expect to melt. Is there something I am missing here? If you need another picture of settings that I am not showing please let me know and I’ll happily show it.

[Rob]

If it's suddenly jumping it may be down to the pressure in the system. Careful selection of the operating density may help, but if you're causing a phase change & density alteration with the melting then you could be triggering the jump because the fluid has to go somewhere. 

[Paco Sheeran]

I see, when you mean the operating density, are you referring to the density of the water/ice? I have set that to both a constant value for water and a polynomial that changes with temperature and both have yielded the same bad result with both. Is that density to be found somewhere else or am I thinking about the density of this phase-changing material wrong? 

Also, if I am understanding you correctly, I need an outflow on the top of the ice mesh so that fluid may exit? 

 

Thank you

 

[Rob]

If the density changes you'll need somewhere for the extra material to go, or come from. A small pressure outlet should be sufficient. However, that may not be the solution. 

Operating density is a value the solver uses to anchor some of the equations and calculate terms for hydrostatic pressure: it's covered in the manual. It shouldn't do much here.