Stir in a bowl : free surface problem
 

Hello,

I am trying to solve a simple problem with starccm+: stirring in a bowl.

I have only two regions: the bowl, and the spoon.

- I first imported my bowl from solidworks (.x_t file) and then surface and volume meshed it in starccm+.
- I then imported my "spoon" (Solidworks file than I volume meshed before importing as a volume mesh in my simulation)

So I have two different regions: the bowl which is fixed and enclosed (no inlet no outlet, only walls and an open top), and the "spoon" which is rotating around the axis of revolution of the bowl ( like how anybody would stir in a bowl), and which is completely solid, all faces are walls.

My problem lies in the liquid/air mixture, the inital conditions, and the models.

I want to bowl to be partially full of water, but also have a free surface so I can see the little ripples/waves generated by the rotation of my spoon which would ideally be half submerged.

First, I am not able to set the initial conditions so that I can see the water in the bowl, I only end up with the whole bowl filled with an homogeneous mixture of air and water...


Here is what appears in the model section I used:

Eulerian multiphase (liquid and air)
Gravity
Implicit unsteady
K-epsilon turbulence
...
segregated flow
3D
two layer all y+ wall treatment
volume of fluid


and in the initial conditions:

Pressure, constant, 101325.0 Pa.
Volume fraction, constant, [0.3,0.7].


I have read ans studied all the tutorials, especially the rotating fan and the boat one, but still, I can manage to get to water surface when the spoon is moving in my bowl..

Do you have any idea or any advice on how I should do that ?

Thank you very much!!

John

Hey,

does anyone have any idea ? I'm still stuck on that...

Thank you so much!

You probably need to use a field function to set the initial condition for volume of fluids. I can'r remember the syntax right now, but I can imagine you can find it easily in the documentation, but it is going to be something like this:

define a field function called water_vof, and set this expression:
$$Location[2] < <0.07 m> ? 1.0 : 0.0

this means is z-coordinate is smaller than 7cm, then water vof is 1, otherwise, it is 0. Then you can define another field function called air_vof that is equal to:

1.0 - $water_vof

then use these for the initial condition (you need to change it from constant to components, and then each component from constant to field function, then select the appropriate field function for each phase)

I hope this helps.

triple_r

Thank you very much for this !I found the exact syntax and it works !

But now I have another issue: even if I refine the mesh a lot, I still have a very bad looking water surface (created in Derived Parts > Iso surface > 0.5 volume fraction of water), even right after initialization, and it looks like this does not even change when I refine the mesh, it's always as bad.
To be more clean, even if I put the "base size" 0.01m instead of 1m (my geometry is about 100m diameter for the bowl), I have the same bad tetrahedral blocs, not smooth at all.

Do you have any idea on how to make the water surface look very nice and smooth (at least at the initialization stage) ?

Thanks!

I also tried to use different mesh methods, like trimmer + surface remesher + prism layer mesher + extruder, that was used on a boat simulation. No success...

Could you post some pictures of the mesh and the surface? At least when using the trimmer it should work without having a choppy surface.

I'm sorry I can't show the entire thing, but here are some pictures of the surface and the mesh (thinner than the elements we see on the surface picture).
I still cannot make it work...

any idea ? I don't see where I'm wrong..

Another issue is that there is no wave/wake generated by the partially submerged spoon, it moves at a speed such that we should have both waves and wake (Fr=1.4), but nothing happens, water is only elevated in a region very very close to the spoon... I'm in K-epsilon turbulent flow and I've check the viscosity, all is normal.. any thought ?

thank you very much, I'm really stuck on this problem

an easy way to ensure that the initial free surface is 'smooth' is to put an internal interface at the initialized water level, and grow a layer of prisms off the interface. Or use a trimmed mesh rather than polyhedral

Thanks, but it did not change a lot.. Anyway maybe this is a secondary issue at that stage..

I am still having troubles to get something... did one of you guys do something similar ?

I have re-read the tutorials but nothing in there is helping anymore. It seems like such a simple problem, I would be very surprised if nobody got it to work!

Thanks for the help!

I'm sorry, I can't see any issues from your (pretty rough) description. Could you maybe create a generic, non confidential case from your setup (e.g. with a simplified spoon, just an axis with a flat plate at the end, or however your spoon looks like) and post the file here, so I can have a look on it?

Hello,

Here is attached a generic example of what I am trying to solve, with the little cube (half submerged) supposed to run in circles around the y-axis. I have added the x_t parts from Solidworks as well.

The goal is, starting from a state of rest, run the little cube in circle and see the wake, the reflections, and all the waves generated by that motion.
I have created the isosurface, and a new Field function for the initial state of rest.

The mesh is not generated in that file to keep the file as small as possible.

I would appreciate any help on that, I am really stuck now...

Thanks!

Every time I run this example (as well as my other model) I get an Error, floating point exception. And I cannot figure out why, "the specific cause cannot be identified"...

I can have the "spoon" moving around in that "bowl", but it produces no wave, nothing...
Could you help me and tell me if I missed something ?

Thanks!

Do you use VOF model, try VOF to see what will happen.

I do already use VOF and it does not work well... but thanks for the suggestion!

It seems like the water just perfectly slides on the moving object instead of being pushed and moved by it. There is no flow...

Does anybody have any idea ?

I'm sorry to come back to you again guys, but I really cannot make this work... does anybody see why the water wouldn't move at all ?

John,

I looked at your file. I would suggest adding a volume source which spans the expected amplitude of the waves you expect and then refine isotropically to a much finer level. Your cells on the free surface have a high aspect ratio which make them like pancakes and their size of .4m will make is difficult to capture much if any free surface detail. Also your surface size on your box should be decreased.

Andrew