CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLOW-3D (http://www.cfd-online.com/Forums/flow-3d/)
-   -   Spinning cylinder in viscous liquid - a serious doubt (http://www.cfd-online.com/Forums/flow-3d/71519-spinning-cylinder-viscous-liquid-serious-doubt.html)

doctsh January 6, 2010 03:21

Spinning cylinder in viscous liquid - a serious doubt
 
I have made a cylinder rotating about its axis in a highly viscous liquid (to mimic low reynolds no). according to theory,the cylinder has to move axially forward but in my CFD simulation it goes front and then come back... what could be the reason? I made a mesh as in the GMO sphere example and have given 6 DOF and prescribed rotational motion along cylindrical axis and coupled translational motion along cylndrical axis. other motions i made as prescribed and given 0. I have enabled GMO model and viscosity model. Is there anything wrong with this set up?

Plz some one help me...

MuxaB January 6, 2010 15:52

Quote:

Originally Posted by doctsh (Post 241601)
I have made a cylinder rotating about its axis in a highly viscous liquid (to mimic low reynolds no). according to theory,the cylinder has to move axially forward but in my CFD simulation it goes front and then come back... what could be the reason? I made a mesh as in the GMO sphere example and have given 6 DOF and prescribed rotational motion along cylindrical axis and coupled translational motion along cylndrical axis. other motions i made as prescribed and given 0. I have enabled GMO model and viscosity model. Is there anything wrong with this set up?

Plz some one help me...

Looks like you did it all right. What makes it theoretically go along the axis?

doctsh January 11, 2010 02:20

1 Attachment(s)
Thanks a lot. I checked the theory, due to high viscous force there is accumulated azimuthal velocity driving rod in forward direction given by (pls see the attached picture)

Any idea why it is not keep going? Should I give any specific meshing? As I am not welversed in meshing, can you recommend what structure of mesh will be appropriate for my problem (cartesian/cylindrical or fit to geometry/automesh)?

thanks in advance


 

jhiggs January 12, 2010 16:20

The last time I tried to use cylindrical coordinates, the 0 degree boundary and the 360 degree boundary did not allow flow to pass through from one to the other. I ended up using Cartesian.

MuxaB January 20, 2010 12:33

The opposite boundaries must be of type periodic for the communication between them to occur.

Regarding meshing the GMO, there is nothing special required for the mesh. Just make sure it resolves the geometry well and leaves room around it for the flow to develop. If the mesh is uniform, so much the better.

Still not sure what the problem looks like. Can you send an image or a drawing?

doctsh January 21, 2010 13:56

1 Attachment(s)
see the attached figure:

Cylinder of radius 20nm rotating with its axis around 600 rad/s...

otd January 22, 2010 10:20

I don't understand the physical setup.

Why should the cylinder move? Is it mounted on springs or some sort of arrangement that lets it slide in the xy plane? The diagram shows nothing about the suspension of the rotating cylinder.

The data appear to show an oscillation about a steady state solution. Is it possible that vortices are forming in the corners of the square experimental tank? I'm suggesting circulating in the corners about a vertical axis near each corner.

It could be helpful if you described the physical setup fully. As it stands, it appears that you're comparing two-dimension analysis with three-dimensional data.

doctsh January 22, 2010 12:43

Hi

It is actually a nanoscale motor. which drives the cylinder through spinning. when it spins, under low reynolds number environment, due to viscous force it goes forward like a screw. so no spring or any other set up. it is just a cylinder spinning independently.

MuxaB January 28, 2010 12:52

Could you please send me your prepin file?

doctsh January 30, 2010 14:49

thanks for your insights
 
1 Attachment(s)
Please see the attached prepin file.

Thanks for your support.

Please remember that i am a beginner and tried to implement what i learnt through gmo sphere tutorial.

MuxaB February 1, 2010 00:37

Quote:

Originally Posted by doctsh (Post 244311)
Please see the attached prepin file.

Thanks for your support.

Please remember that i am a beginner and tried to implement what i learnt through gmo sphere tutorial.

Thanks for the file. Is the equation for the angular velocity from their book on Hydrodynamics? What section of it and page no?

doctsh February 1, 2010 03:51

1 Attachment(s)
Thanks again for supporting me.

The equation is given in page 56 of FLUID MECHANICS
Second Edition
by
L. D. LANDAU and E. M. LIFSHITZ

I have also attached those pages as pictures.

Please have a look on a research paper which explains this in detail:
Mesoscopic Modeling of Bacterial Flagellar Microhydrodynamics (available at
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1630491/)

I hope this two information will be helpful to understand the problem

thanks

MuxaB February 1, 2010 10:17

thanks Kumar,

this will help me to understand the problem. give me a few days to look at it.

MuxaB February 1, 2010 23:42

Hi,

The solution in the Landau book is for the flow of fluid between two rotating cylinders. This kind of motion does not create any propulsion for the cylinders themselves. The paper, on the other hand talks about a rotating helix, in which case there may be propulsion in the axial direction since the helix acts like a screw.

What you need to start with is create a helix in a CAD tool and import it into FLOW-3D as an STl file.

What school do you study at?

doctsh February 2, 2010 00:29

Thanks for your information.

I have done CFD simulation according to an experiment explained in the paper (please have a look on equation 4, and graph: figure-1) :http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1630491/

Here they have showed that there exist a net azimuthal velocity using experimental set-up. I just tried the same in the Flow3d, though I could already observe the same in helix, i could not do for cylinder as explained by them.

By the way I am an Indian physics teacher, teaching in Oman, but still a student in CFD and bacterial motion.

Thanks for your continuous support.

MuxaB February 2, 2010 10:17

Quote:

Originally Posted by doctsh (Post 244514)
Thanks for your information.

I have done CFD simulation according to an experiment explained in the paper (please have a look on equation 4, and graph: figure-1) :http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1630491/

Here they have showed that there exist a net azimuthal velocity using experimental set-up. I just tried the same in the Flow3d, though I could already observe the same in helix, i could not do for cylinder as explained by them.

By the way I am an Indian physics teacher, teaching in Oman, but still a student in CFD and bacterial motion.

Thanks for your continuous support.

Net axial or azimuthal velocity? The paper explicitly says that for a smooth cylinder there is no axial motion.

doctsh February 2, 2010 21:44

Ok, sorry I understand. I assumed that the graph (figure -1) shows the velocity, of the cylinder's front surface, so cylinder could move forward in a highly viscous environment (low reynolds number).
How will I get data from my simulation for V versus r as in figure -1?

Thanks again

MuxaB February 10, 2010 10:40

Quote:

Originally Posted by doctsh (Post 244688)
Ok, sorry I understand. I assumed that the graph (figure -1) shows the velocity, of the cylinder's front surface, so cylinder could move forward in a highly viscous environment (low reynolds number).
How will I get data from my simulation for V versus r as in figure -1?

Thanks again

You can just plot say y-velocity component as a function of x using the 1-D plotting tool in the GUI.

doctsh February 19, 2010 09:41

Is there anyway to measure the force produced by a GMO object in flow3D?

Thanks


All times are GMT -4. The time now is 12:13.