[paweld]

Dear All,

For a couple days I've been trying to simulate a mixing process in a tank using an anchor agitator and get some results with using a scaled (smaller) model.
I had some doubts about the results and when I was searching the Internet I found the following information about SW FS:

"General recommendations for the rotating bodies

(...) Any geometry that deviates from a hub with blades sticking out will not be well suited to be used in a rotating region. In particular, bulk geometries, helixes or geometries with holes. (...) What this means is that the results will diverge from reality. They may look real, but will not match experimental results if they are available. "

The anchor agitator from my simulations surely differs from "a hub with blades sticking out". So, does anyone of you simulate or tried to a similar case? Unfortunately I don't have experimental data to compare my results. Maybe some of you can look at my results and check if they look reasonable.

I will be thankful for any help.


My simulation initial and boundary conditions:
- tank dimensions heigh 120 mm, diameter 100 mm,
- agitator angular speed = 1 rad/s,
- fluid: water, initial temp: 20 deg C,
- tank walls; constant temperature = 100 deg C,
- physical time = 100 s.

Attachments:

http://dc659.4shared.com/download/wW....png?lgfp=3000

http://www.4shared.com/download/QwMH....png?lgfp=3000

http://www.4shared.com/download/XYYD....png?lgfp=3000

[Boris_M]

Hi Pawel,

this is not true. I don't know who wrote this comment but the main limitations in the rotating reference frame or mixing planes or averaging planes approach is that if you have a model like yours and there are stator turbulators reaching into the rotating region then it is not possible to simply define them as stator as this boundary condition only defines a wall boundary laywer of 0 velocity. Also the flow in circumferental direction is averaged and therefore any application where the flow enters from the right into the rotating region and leaves to the left such as in a tangential fan will not work.

The new sliding mesh approach can cope with the second limitation but of course is also limited the first limitation since the mesh is sliding away and would therefore cut the stator turbulator reaching into the rotaing region of and move it with it which of course does not make sense.

Your model however should work with both methods but there are ways to improve the project definitions you have. Since it is a cylinder and basically everything is moving except for the stator walls of the cylinder, you can simple select it as global rotating and define the cylinder walls in contact with the fluid as stator. This way you don't need a rotating region. This eliminates also the interpolations issues you can see on the bottom and top rotating region inteface in your third image.

The sliding mesh appraoch is fully transient and actually shows the model moving. I think this should be more of interest for you if your project is defined as transient to show the mixing over time.
This approach is available in SWFS 2015 or FloEFD V14.

Boris