[cecilia_xiao559]

Hello all,

I am trying to model the flow in a rotor, from what i've read so far, using MRF is the correct way to do it. Can someone explain to me why set the walls to be rotating walls is not correct?

Thank you~

[Achin]

Hello Cecilia,

I don't think one method is strictly not correct or correct. When solving with moving walls, you are still in the inertial frame of reference but when solving with MRF, you transform your frame of reference to that of the moving part.

One benefit of using MRF is that while with "moving wall approach" your simulation is unsteady as it is with respect to inertial frame but with MRF approach, you can solve it as a steady simulation (if the rotor speed is not changing) as now you are in rotor's frame.

Of course, the equations in MRF get changed and they take into account the relative velocities.

[cecilia_xiao559]

Originally Posted by Achin

Hello Cecilia,

I don't think one method is strictly not correct or correct. When solving with moving walls, you are still in the inertial frame of reference but when solving with MRF, you transform your frame of reference to that of the moving part.

One benefit of using MRF is that while with "moving wall approach" your simulation is unsteady as it is with respect to inertial frame but with MRF approach, you can solve it as a steady simulation (if the rotor speed is not changing) as now you are in rotor's frame.

Of course, the equations in MRF get changed and they take into account the relative velocities.

Hello Achin,

Thank you for your reply. It does makes more sense after you explained it. i still have a few questions maybe you can help me with:
(1)i've tried both method for the same B.C.s and ended up with similar flow pattern with different velocity magnitude etc. It looks like the velocity from rotating wall is w.r.t. the wall (it doesnt have the rotating omega*r component) and the rotating cell is with respect to the ground?
(2)with rotating cell, i set up the frame motion under "cell zone conditions" tab for that part, left with all the walls with zero rotating speed set up in the boundary conditions tab. is that correct?
(3) for the rotating cell, if i want to plot the relative velocity after simulation, say with respect to the rotating wall boundaries (i.e i want to see the circulation pattern in the end-turn cavities), how do i display it? it looks like something wrong with my setup because it plotted the exact same thing regardless plotting velocity or relative velocity.

Thank you in advance for your help.

Cecilia

[Achin]

Hello Cecilia,

Sorry I didn't see your post earlier. Here are my few inputs on your post.

Whether the wall should be stationary or moving depends on your case. Suppose you have a fluid zone which has a Frame Motion(FM), and say there is a wall (may be a turbine blade) in the zone. If you keep it as stationary, then it won't move w.r.t the enclosing fluid zone(which is having FM). But it would move w.r.t the inertial frame.

But as an example of turbine blades, you definitely want this wall to move with your fluid zone, then you have to set it to moving wall, then retain the "Relative to fluid motion" and give the rotational velocity as 0. This means that the wall moves at the same speed as your fluid zone.

To check the circulation, you might want to view the velocity vectors w.r.t to the fluid zone with frame motion. For that click on "Reference Values" in the tree structure and it opens a task page where you can select the rotational zone in the reference zone drop-down. Then go to vectors, and select velocity and relative velocity magnitude.

[cecilia_xiao559]

Originally Posted by Achin

Hello Cecilia,

Sorry I didn't see your post earlier. Here are my few inputs on your post.

Whether the wall should be stationary or moving depends on your case. Suppose you have a fluid zone which has a Frame Motion(FM), and say there is a wall (may be a turbine blade) in the zone. If you keep it as stationary, then it won't move w.r.t the enclosing fluid zone(which is having FM). But it would move w.r.t the inertial frame.

But as an example of turbine blades, you definitely want this wall to move with your fluid zone, then you have to set it to moving wall, then retain the "Relative to fluid motion" and give the rotational velocity as 0. This means that the wall moves at the same speed as your fluid zone.

To check the circulation, you might want to view the velocity vectors w.r.t to the fluid zone with frame motion. For that click on "Reference Values" in the tree structure and it opens a task page where you can select the rotational zone in the reference zone drop-down. Then go to vectors, and select velocity and relative velocity magnitude.

Hey Achin,

Thank you very much for your explanation. this makes more sense now and the relative velocity is what I am looking for.

Another question for my model, i am trying to simulate the flow pattern inside a rotor for a large scale generator, which includes endturn cavities, radial vents etc, so i am not modeling any blades. I am having a issue with the interface in the airgap between the rotating frame and the inertial frame (i made half of the airgap belongs to the rotating frame and half belongs to the inertial frame). i dont thing there will be significant back flow so i am thinking about using mixing plane to deal with it. is it sounds like a good idea?

Thank you,

Cecilia