CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   impeller motion in an open raceway pond (http://www.cfd-online.com/Forums/fluent/99329-impeller-motion-open-raceway-pond.html)

hossein65 March 31, 2012 17:02

impeller motion in an open raceway pond
 
Dear all
I am using Fluent 6.3 to simulate a 3D impeller in an open raceway pond for microalgae culturing (you can see the picutres by simple search in google); I have done these steps in creating the model in Gambit:
1- whole channel creation
2- a cylinder indicating as the moving mesh area
3- subracting the cylinder from the whole model (retaining the cylinder)
4- creating the impeller
5- subracting the impeller (without retaining the impeller)
I devided the whole model to two areas: down=water and up=air
zones:
up (air) & down(water) & the cylinder including the impeller===>fluid
B.Cs:
the two sides and the surface of the inner cylinder: interface
the two sides and the sufrace of the outer cylinder (the cylinder wich is the result of the volume subtraction): interface
in Fluent, I connected the two interface areas with "Grid interface"
water and air zones are stationary and the cylinder is moving mesh

the problem is when the cylinder rotates, the water inside the cylinder doesn't mix with the water in the channel.
I know that all other things are ok; like the VOF, patching, ....

Is there any suggestions?
my email address: hossein.amini.che@gmail.com

-mAx- April 10, 2012 14:22

can you post a sketch of your model, and color the differents parts for better understanding.
But I can say that if you want a moving cylinder (with rotation) in your domain, you may disconnect your cylinder from the other domain (else the sliding mesh won't work)

hossein65 April 10, 2012 17:08

1 Attachment(s)
This is it;
I should notice that the 3 volumes are seperate, but st. zone1 and zone2 are connected. The boundary conditions are:
1- for the moving mesh, I have chosen the whole cylinder as an interface
2- for the cylinder just inside the channel also interface that fluent seperates the interfaces automatically as it belongs to two different zones.

all zones are set to fluid

hossein65 April 10, 2012 17:10

I forgot to mention, I want the impeller inside the cylinder operate as an agitator, that's what I desire. I don't care what happens to the cylinder. (I mean the motion of the impeller is important. that's why I have set the boundary surfaces as interface)

hossein65 April 11, 2012 03:55

1 Attachment(s)
This is the figure of interface boundary condition I have set. this boundary condition also has been done for the cylinder inside the channel. Then, in FLUENT, I connect the two interfaces with "grid interface".

-mAx- April 11, 2012 04:07

you have to increase complexity of your model step by step.
I don't understand where comes fluid in.
Any way
First concentrate you on the sliding mesh.
What is the surface (white) which is in the middle of your domain, but not splitting the whole domain?
Then if you only have your red mixer which is rotating, forget stationnary etc...
you only have to work with interfaces.
Define the red generated surface as interface 1 and the white generated surface as interface 2.
Define the red rotating volume as separated fluid domain, for being able later to define it as rigid body in Fluent
When moving back your rotating (red) volume, then interface 1 and interface 2 should be superposed.
Only connect them back in Fluent by defining grid interfaces.
Once your mesh is well rotating, and fluid isn't stopped at interfaces, then introduce vof

hossein65 April 11, 2012 07:10

let me be more anal about the model:
1- there is no fluid inlet. So, imagine a pool (elliptic shape)
which has a height of 0.5m and filled with water up
to 0.2m. There is an impeller which circulates the
water inside the pool (this is the definition of the
geometry of an open raceway pond for microalgae culturing)
2- the white surface inside the pool is just a wall
for dividing the pool to have a circular characteristic (you can see it in the real model below).
after these all, I should tell you that I do know that I have
to use VOF, I have seperated the mixer domain and the channel domain
as fluid zones (in which mixer is as a moving mesh not stationary).
I know that I have to use grid interface in FLUENT to connect the two
zones (stationaries(air and water) and moving(mixer))
but after setting these all , I got a problem with the phase transformation
between two zones. you know, the mixer zone acts like there is no
connection between the mixer and the channel and all of the cylinder
around the mixer has a role like WALL boundary conditions. So the mixer can't mix the whole water in the channel

hossein65 April 11, 2012 07:19

1 Attachment(s)
here is a real model of what I am going to simulate


All times are GMT -4. The time now is 06:54.