[yangxu]

Hello everyone!
I have got this question when I'm simulating a wind turbine. I defined a new Coordinate Frame, and it's origin is set in a rotating domain. And my question is that will the newly defined Coordinate Frame rotate with the rotating domain or just stay fixed?
I have had this question for a long time. I would be very appreciated if someone should get the answer for me!

[Opaque]

For clarification, there is no mechanism to create a coordinate frame in a domain. Coordinate frames are created independently in space. The attributes of the coordinate frame such as type, location and motion are set during creation.

These coordinate frames can be associated (or used) in different locations such as domain, boundaries, subdomain, points, etc. Once they are associated to a location, the input on those locations are referred to the coordinate when needed. For example, coordinates x,y,z are based on the coordinate frame associated to the specific location.

If you associate a coordinate frame to a domain (rotating or stationary), your inputs for the domain refer to such frame. The domain may rotate, but the coordinate frame does not inherit such property.

If you associate a moving coordinate frame (available since ANSYS CFX R14.0) to a boundary in a rotating domain, there may be relative motion between the boundary and the coordinate frame (depending of other setting on the boundary) and you must be aware of its implications.

Hope the above helps,

[yangxu]

Originally Posted by Opaque

For clarification, there is no mechanism to create a coordinate frame in a domain. Coordinate frames are created independently in space. The attributes of the coordinate frame such as type, location and motion are set during creation.

These coordinate frames can be associated (or used) in different locations such as domain, boundaries, subdomain, points, etc. Once they are associated to a location, the input on those locations are referred to the coordinate when needed. For example, coordinates x,y,z are based on the coordinate frame associated to the specific location.

If you associate a coordinate frame to a domain (rotating or stationary), your inputs for the domain refer to such frame. The domain may rotate, but the coordinate frame does not inherit such property.

If you associate a moving coordinate frame (available since ANSYS CFX R14.0) to a boundary in a rotating domain, there may be relative motion between the boundary and the coordinate frame (depending of other setting on the boundary) and you must be aware of its implications.

Hope the above helps,

Thanks very much for your patient reply!
By your reply, I understand that you mean the coordinate frame is independent from domain and that the motion of coordinate frame is set once it's created, so that the motion of domain won't have any effect on the motion of coordinate frame. But it seems that the ANSYS CFX does not support the set of motion of a coordinate frame before the version of R14.0. Does that mean all the coordinate frame doesn't rotate in simulation when I use the ANSYS CFX R13.0? If that's true, the results I have already got in my simulation cannot be explained, for the result seems to indicate that the coordinate frame is under a rotation motion.
Expecting your reply!

[Opaque]

By your reply, I understand that you mean the coordinate frame is independent from domain and that the motion of coordinate frame is set once it's created, so that the motion of domain won't have any effect on the motion of coordinate frame.

Correct!!

But it seems that the ANSYS CFX does not support the set of motion of a coordinate frame before the version of R14.0.

Correct as well.

Does that mean all the coordinate frame doesn't rotate in simulation when I use the ANSYS CFX R13.0?

Correct again..

If that's true, the results I have already got in my simulation cannot be explained, for the result seems to indicate that the coordinate frame is under a rotation motion.

They can be explained if we do not confuse "Coordinate Frame" with "Domain Motion". As I indicated in a previous response, the Coordinate Frame setting indicates input is interpreted in the locator the coordinate frame is associated to. The "Domain Motion" does not change the state of the coordinate frame, but indicates on what observer frame the equation is solved: stationary frame, or rotating frame. When solving the equations in the rotating frame there is no need to move the mesh, and the rotation effects are directly included in the equations. In the post-processor, you visualize the solution in the stationary frame by letting the domain rotate (see CFD-Post options), or in the rotating frame by keeping the domain fixed.

Notice I use "coordinate frame", "stationary frame" , "rotating frame" and was careful not to use "rotating coordinate frame". The latter is where the confusion starts.

Hope the above helps,

[yangxu]

Originally Posted by Opaque

Correct!!



Correct as well.



Correct again..



They can be explained if we do not confuse "Coordinate Frame" with "Domain Motion". As I indicated in a previous response, the Coordinate Frame setting indicates input is interpreted in the locator the coordinate frame is associated to. The "Domain Motion" does not change the state of the coordinate frame, but indicates on what observer frame the equation is solved: stationary frame, or rotating frame. When solving the equations in the rotating frame there is no need to move the mesh, and the rotation effects are directly included in the equations. In the post-processor, you visualize the solution in the stationary frame by letting the domain rotate (see CFD-Post options), or in the rotating frame by keeping the domain fixed.

Notice I use "coordinate frame", "stationary frame" , "rotating frame" and was careful not to use "rotating coordinate frame". The latter is where the confusion starts.

Hope the above helps,

Thanks for your reply! It do help me a lot!
I think you mean that although the coordinate frame may be fixed, the equation can be solved either way, in stationary frame or rotating frame. But I am still a little confused that how can I know in which frame the equation is solved when the domain is set to be rotating or not, or that the equation is actually solved in both way and we can choose in which frame to visualize the solution? By reference from the CFD-Post options, I presume that the equation is solved both way, and the default value of expressions like "force_x()@physical region" is the value in the rotating frame. If that's true, then everything can be explained. And it should be that the motion of the "rotating frame" follows the motion of the rotating domain. Am I right?

[Sobhan]

Originally Posted by Opaque

For clarification, there is no mechanism to create a coordinate frame in a domain. Coordinate frames are created independently in space. The attributes of the coordinate frame such as type, location and motion are set during creation.

These coordinate frames can be associated (or used) in different locations such as domain, boundaries, subdomain, points, etc. Once they are associated to a location, the input on those locations are referred to the coordinate when needed. For example, coordinates x,y,z are based on the coordinate frame associated to the specific location.

If you associate a coordinate frame to a domain (rotating or stationary), your inputs for the domain refer to such frame. The domain may rotate, but the coordinate frame does not inherit such property.

If you associate a moving coordinate frame (available since ANSYS CFX R14.0) to a boundary in a rotating domain, there may be relative motion between the boundary and the coordinate frame (depending of other setting on the boundary) and you must be aware of its implications.

Hope the above helps,

Hello Opaque,
thank you for your contribution. I have the below problem.
I want to use a rotary coordinate system. But under Domain, I can only use the stationary coordinate systems, and the one with rotation does not apear at all.
So I attached some pictures for description, as below:


this coordinate system (as Coord 2) is stationary and doesn't have frame motion.





so the coord 4 (below fig) should rotate about the ( minus gobal y axis). It is the same like coord 2 but with frame motion.





But as you see, it does not appear in menu at all. It shows only the stationaries coordinate systems.



I would appreciate for a reply.
Best Regards
Sobhan

[Opaque]

Domains do not support moving coordinate frames, so they are not listed.

If I recall correctly, moving coordinate frames are only supported at boundaries to create relative motion between the boundary condition, and the domain.

You may want to explain your model in more details, and someone here in the forum may guide you on how to use the existing tools to achieve it.