I am modelling a rotor stator cavity.I have modelled the fluid domain as rotating in this I have mentioned the rotor wall as stationary no slip boundary and for stator wall as Counter rotating wall boundary.Is it correct or I have to change the boundary condition.can any body clarify the doubt.

Hi Kalyanasundaram,

Your stator must be in a stationary frame of reference, the rotor in a rotating frame and a frame change interface specified between the two meshes.

A counter-rotating wall is stationary in the stationary frame of reference, thus counter-rotating relative to the rotating frame of reference. Wall motion, however, can only be specified tangential to a wall.

Regards, Robin

I just want to know, how swirl decays in an annular long cylinders, which are counter rotating. The inlet is having pre-swirl. Is there any empirical relation for this from boundary layer theory? I need some quick help on this.

Hi Sarma,

There will be an analytical answer for the flow you describe provided you can use simple inlet conditions and a simple geometry. Get out your fluid machanics textbooks and give it a shot. There is no need to go to empirical relations.

Regards, Glenn Horrocks

[cfdesf1990]

Hi everybody

I am simulating a fan and I have some questions.Since the casing of fan is not symmetric so I have to model all the geometry.

1- For boundary conditions of inlet and outlet I have read some sources but each of them has a different condition. For the inlet it is usually recommend using total pressure but for outlet some suggest mass flow and others static pressure. which one is better? Actually, I can test all of them but because of the second question it became more complicated.

2- Since the fluid domain is rotating and the impeller is stationary. I think I have to specify the counter rotating wall option. but I do not know for which wall ( impeller and casing) it should be applied?

I really appreciate any help.

Regards

[ghorrocks]

1. The CFX documentation has a section of choice of boundary condition in recommendations for flow modelling. I recommend you read it.

2. The counter rotating wall option is only available on rotating frames of reference, and means that the wall has a tangential velocity which counter rotates against the frame motion. This means the wall velocity is stationary in the absolute frame.

[cfdesf1990]

Thanks a million Glenn for taking time to help others.
I actually get confused because of the rotating frame. As you know, in a centrifugal fan, there is a volute, a casing, an outlet duct, and an inlet in the geometry. Based on some thread and my experience, I understood that it is not correct to use a rotating frame for the impeller (frozen rotor) so I have to use a stationary frame for the impeller and rotating frame (with opposite direction) for fluid.
Therefore, as you said for the casing I should use the counter rotating wall. How about the impeller, volute wall, and outlet duct wall? Since the flow in the volute and duct are not rotating but I use one fluid domain in my simulation so All the domain is rotating.

The geometry in like(with one inlet) the following :

http://www.cfd-online.com/Forums/att...tor-model2.jpg

Regards

[ghorrocks]

Why are you keeping the rotor still and rotating the casing? This sounds really weird to me. If the rotor is the bit which rotates then it goes in the rotating fame of reference. If you put the stationary part in the rotating frame of reference then the Coriolis and centripetal terms will be applied to the wrong domain - so this is definitely wrong.

If I assume the rotor rotates and the casing is stationary then the casing walls are just normal walls so no special treatment is needed. On the rotor you have walls which rotate with the domain (leave these as default) and walls which are actually stationary (so set these as counter rotating).

[cfdesf1990]

Thanks Glenn for your help.

Actually, I agree with you and I think it does not make sense to use those setting I mentioned. but I set the rotor as a rotating solid domain and the fluid as the stationary one. but based on the result it can be understood that the solid domain does not rotate at all. I do not know what is wrong with that but I found the following thread in which two users said that for such a case we have to use fluid domain rotation with a stationary solid domain. So I used this type of settings.

http://www.cfd-online.com/Forums/cfx...id-domain.html

Over all, I think I should correct my mistake in setting the rotating solid domain. So just one question remains that I think it probably solves the problem. What do you mean from "walls" in you sentence those should be set as counter rotating walls?

Originally Posted by ghorrocks

and walls which are actually stationary (so set these as counter rotating).

Best Regards

[ghorrocks]

This is sounding like an XY problem to me..... (http://mywiki.wooledge.org/XyProblem)

Can you please explain what you are trying to model, why you are modelling it and what you want to learn from the model. Then we might be able to help you select some appropriate options to model it.

[cfdesf1990]

I am going to simulate a centrifugal fan which rotates with 1500 rpm and I know the static pressure. I want to calculate and compare delta p (pressure) between inlet and outlet that based on the experimental result it is 1000 pa and mass flow is 3.6 kg/sec with simulation results.
I know this case is not complicated or something new but the only problem I have is the rotating impeller does not rotate although I set it as a rotating solid domain (frozen rotor).
I also set the fluid domain as stationary. The inlet and outlet BCs are clear (actually inlet and outlet BCs are not important in this question since I will try two different BCs.)

So I have two questions:
1-Now I want to know the BC at walls(on impeller, casing, volute, outlet duct). Which I think one of them should be set as the counter rotating wall?

2- Do you have any idea about the solid domain which does not rotate?

I hope I can clarify my question well.

Thanks

[ghorrocks]

You do not need to model the rotor as a solid (unless you want to know the temperature in the rotor). You just model the fluid domain around the rotor. See the CFX tutorials on rotating machinery for what this looks like.

For your questions:
Q1 - I cannot answer this without seeing your geometry.
Q2 - You do not model a solid domain so this question is irrelevant.

[cfdesf1990]

Thanks, dear Glenn for you help.

I attached my geometry in previous thread as :
http://www.cfd-online.com/Forums/att...tor-model2.jpg

I do not want to have a thermal simulation so I think, as you said, I must define one stationary domain only in which all walls are no- slip except the impeller wall that should be set as no-slip wall with angular velocity so there is no need for any interface?
Am I right?

[ghorrocks]

No, that does not sound correct. You have to put the rotor in a rotating frame of reference. Have a look at the CFX tutorials for how to do this. They have several examples of rotating machinery.

[cfdesf1990]

Thanks, dear Glenn.

I found two films from youtube for the simulation of the centrifugal pump and I think they are similar to my case.

I will do those settings I hope I can simulate my case.

Best Regards