[yida16]

Hi guys,

I am now trying to solve a windage problem (air friction loss problem) using Fluent in 2D. I am trying to use sliding mesh now as MRF doesn't seem to work, because I think there's suppose to be a vortex in the flow between the teeth of the rotor when it rotates and MRF won't give me that.

I am now modeling using sliding mesh, attached is my setup, I set a moving mesh around my teeth, called it rotating fluid, the rest of the fluid is absolute stationary. I am not sure if this set up is ok at all as I haven't seen anyone doing it this way. Can someone please help me?

Thanks

[Bionico]

Hi yida,
your setup looks ok, you have to define the interface zones, like MRF anyway.
I suggest reading Fluent Tutorial about Sliding mesh

Regards

[ghost82]

Sorry,
but I think it will not work as your rotor zone will intersect the stator during rotation; different zones must not intersect during rotation.

Daniele

[CFD-fellow]

Hi
Daniele is right
In both MRF and sliding mesh methods, the separating surface(in 3D problems) and edge (in 2D problems) should be respectively created from a revolution of edge and point about axis of rotation.So your domains are completely wrong.
Use single rotating reference frame and rotate the whole domain.
Regards

[yida16]

Originally Posted by ghost82

Sorry,
but I think it will not work as your rotor zone will intersect the stator during rotation; different zones must not intersect during rotation.

Daniele

Hi Daniele,

Thanks. So what should be a correct way to set up the problem? If I use SRF/MRF, there will just be steady flow, if no inlet/outlet is defined. And the torque on the teeth are just close to zero

Should I set a inlet/outlet? If so, how should I set it up?

Thanks

[yida16]

Originally Posted by CFD-fellow

Hi
Daniele is right
In both MRF and sliding mesh methods, the separating surface(in 3D problems) and edge (in 2D problems) should be respectively created from a revolution of edge and point about axis of rotation.So your domains are completely wrong.
Use single rotating reference frame and rotate the whole domain.
Regards

Hi CFD-fellow,

Thanks for the reply, same as the question to Daniele,
SRF/MRF will give me steady flow and I will see torque close to zero on rotor surface. The teeth of the rotor was supposed to generate vortex in the flow. In addition, I cannot use wall motion for the rotor surface since the teeth of the rotor are normal to the direction of flow

I really appreciate your help

[yida16]

Originally Posted by CFD-fellow

Hi
Daniele is right
In both MRF and sliding mesh methods, the separating surface(in 3D problems) and edge (in 2D problems) should be respectively created from a revolution of edge and point about axis of rotation.So your domains are completely wrong.
Use single rotating reference frame and rotate the whole domain.
Regards

I mean, this is like a immersed solid problem in CFX, a rotor rotating in air bounded by walls. Is it possible to do it in Fluent?.

Thanks

[CFD-fellow]

Yes its a common task to be handled by Fluent.
>Mesh your domain so that you have just one domain.(because of small gap between teeth tip and wall, it is not recommended to create two domains, but as i said your domains are incorrect and you cant judge on Fluent results by this)
>In settings of your domain choose moving mesh for its motion(this is sliding mesh and the rotor really rotates and solver is transient)
>Also you can set moving reference frame to this domain and use its results as an initial condition for moving mesh.
>Be sure that in this way the difference between MRF and sliding mesh torque would not be so much.

[yida16]

Originally Posted by CFD-fellow

Yes its a common task to be handled by Fluent.
>Mesh your domain so that you have just one domain.(because of small gap between teeth tip and wall, it is not recommended to create two domains, but as i said your domains are incorrect and you cant judge on Fluent results by this)
>In settings of your domain choose moving mesh for its motion(this is sliding mesh and the rotor really rotates and solver is transient)
>Also you can set moving reference frame to this domain and use its results as an initial condition for moving mesh.
>Be sure that in this way the difference between MRF and sliding mesh torque would not be so much.

Hi CFD-fellow

Thanks for your advice. I tried your way but it seems that the flow is steady under your method. Attached are the pressure and velocity contours. I also tried sliding mesh, but it reaches steady flow also. From my understanding, the teeth on the rotor are suppose to generate vortex in the flow, I am not seeing that effect at all. I don't have much knowledge in fluid dynamics but from one paper about gear windage loss, I see that effect, and this problem is just like that.

Any good suggestions?
Thanks for your patience

[yida16]

Originally Posted by CFD-fellow

Yes its a common task to be handled by Fluent.
>Mesh your domain so that you have just one domain.(because of small gap between teeth tip and wall, it is not recommended to create two domains, but as i said your domains are incorrect and you cant judge on Fluent results by this)
>In settings of your domain choose moving mesh for its motion(this is sliding mesh and the rotor really rotates and solver is transient)
>Also you can set moving reference frame to this domain and use its results as an initial condition for moving mesh.
>Be sure that in this way the difference between MRF and sliding mesh torque would not be so much.

Here's the paper
http://www.wseas.org/multimedia/jour...012/54-646.pdf

He also used Fluent, and defined teeth as rotating wall but fluid seems to be stationary... I don't know how he did that because there are walls normal to the direction of rotation and the rotating wall in fluent will ignore those parts.

[CFD-fellow]

Please attach a picture of display>vectors>relative velocity or velocity magnitude.
I havent read the paper(the paper and their authors both are not expert) completely, but if their domain is stationary and the teeth are rotating wall, it is completely wrong.

[yida16]

Originally Posted by CFD-fellow

Please attach a picture of display>vectors>relative velocity or velocity magnitude.
I havent read the paper(the paper and their authors both are not expert) completely, but if their domain is stationary and the teeth are rotating wall, it is completely wrong.

Hi CFD-fellow,

Have you seen the contours above?
Attached is the vectors for velocity. The flow is steady, which I think is not correct for my case. The tooth is supposed to generating interrupts in flow according to my limited knowledge.

Thanks

[CFD-fellow]

Hi
I suppose that you know the pressure and suction sides definition in turbomachinery. In your pressure contour the pressure on both suction and pressure sides are equal in the same radial position and this is not true. There is something wrong in your setup that I cant guess until seeing your complete setup and mesh.
Fluent tutorials also will be so helpful.

[yida16]

Originally Posted by CFD-fellow

Hi
I suppose that you know the pressure and suction sides definition in turbomachinery. In your pressure contour the pressure on both suction and pressure sides are equal in the same radial position and this is not true. There is something wrong in your setup that I cant guess until seeing your complete setup and mesh.
Fluent tutorials also will be so helpful.

Hi CFD-fellow,
Thanks
I looked upon the pressure and suction.I was wondering why the two sides are same too. But I defined a moving wall with 0 velocity relative to the fluid, so I guess that will just make the wall stationary with respect to the fluid, which will make the pressure same?? For my rotor, the two sides of the teeth are exactly the same shape, so it is probably different from common turbo problem.

Thanks

[yida16]

Originally Posted by CFD-fellow

Hi
I suppose that you know the pressure and suction sides definition in turbomachinery. In your pressure contour the pressure on both suction and pressure sides are equal in the same radial position and this is not true. There is something wrong in your setup that I cant guess until seeing your complete setup and mesh.
Fluent tutorials also will be so helpful.

I forgot to mention that I don't have inlet/outlet... Is that the issue. I honestly don't know how set them, since I am not pushing air in any direction?

[ghost82]

What is the rotational speed?

[yida16]

Originally Posted by ghost82

What is the rotational speed?

Hi Daniele,

It is 15000RPM.