simulation of free surface of stirred tank using vof
 

Hi alls,
I want to simulate free surface of stirred tank.
I've solved my problem but it seems that results are wrong.
System should be stable after a period of time, but did not stable.
I think thak I don’t do a part of problem.
What should I do that I have not done?
Step by step of my problem:
Read msh file
1) Make grid interfaces (rotor-stator interface)
2) Set multiphase model to VOF with default parameters
3) Set K-E with default parameters
4) Copy water-liquid from materials database
5) Define Air as primary phase
6) Define water as secondary phase
7) Operating condition: active gravity and set -9.81 to Z, (0 0 26.5) as reference pressure location
8) Boundary condition:
A) Rotor fluid: whether moving mesh or mrf , 200rpm
B) Rest fluid: stationary
C) Impeller: stationary
D) Shaft: absolute rotational, 200rpm
E) Top of tank: symmetry
F) Tank walls: no slip
10) initialize, init with default parameters
11) Adapt/ region/ (Xmin=-12, Xmax=12) (Ymin=-12, Ymac=12) (Zmin=0, Zmax=21.5)/ mark
12) initialize/ patch/ volume fraction of water in marked zone is set to 1
13) all parameters of solution is default of fluent
14) time step=0.001 or 0.0001
15) iterate

pictures of your weird results?

they are here: http://www.cfd-online.com/Forums/flu...tml#post393765

i would say interfaces issue -- are those spots exactly where interfaces-intersection are ?

which spots?

the four red you see on each picture...

yes, spots and interfaces are overlap.

then here is the issue.
how many interfaces did you define?

two interface, 1)rotor perimeter 2) part of rest tank that contact with rotor (You said to me earlier )
http://www.cfd-online.com/Forums/ans...nk-gambit.html

I would say 3:
*rotor perimeter
*top cap of rotor
*bottom cap of rotor
No?
Anyway are you dealing with moving mesh? if yes is it moving properly?

Exactly, I've done all these things and moving is proper

So 2 or 3 interfaces defined in Fluent? :D

I defined rotor perimeter, top cap of rotor, bottom cap of rotor as one interface in GAMBIT and every three are defined rotor interface(as one interface) in fluent
Stator interface is defined similar to rotor:confused:

It is not correct in regards of topology, because for one interface all surfaces should slide on each other

defined in gambit:
top cap rotor as interface 1
top cap stator as interface 2
bottom cap rotor as interface 3
bottom cap stator as interface 4
perimeter rotor as interface 5
perimeter stator as interface 6

Then in fluent:
grid interface A using interfaces 1 & 2
grid interface B using interfaces 3 & 4
grid interface C using interfaces 5 & 6

Thank you so much
I'll do the things you said and report the results to you

Jamal,
also, if you use mrf you don't need any interface!Create interfaces, as explained by max, only if you use sliding mesh.
If you use mrf you have only to define 2 fluid zones (rotor and stator), without interfaces.

EDIT: "if you use mrf you don't need any interface!" --> if you have a conformal mesh

What if the mesh is not conformal? Whether it is mrf or sliding mesh, interfaces should be defined as mrf can be thought as a special case of sliding mesh.

Yes Far, if you have a non conformal mesh, then interfaces are necessary also for mrf; however, in my cases, if I used mrf, I preferred to not create interfaces, as in post processing some discontinuities may appear on non conformal interfaces, between fluid zones.

Looking at the mesh model, I think it is quite simple to not create interfaces.

Daniele

I agree..... Thats perfect

Thanks all to your responses