Dear all, I am attempting to model the movement of some aerofoil blades in a cycloidal mechanism. This consists of a number of blades, each located in their own domain (blade domain), circular in shape so the domain/mesh can be rotated.. these domains are located in a larger cicular domain(rotating domain) which i also want to rotate. This larger domain is located in a fixed domain where the inlet and outlets are located. So far i have been able to get a model where the rotating domain and blade domains spin around a central axis. However.. the setup only allows domian rotation around one point to be specified. For the blade domains, i need to tell them to rotate around the fixed axis.. as well as there own axis which is being rotated about the fixed axis. I hope this is possible using CEL, i have done the tutorials using CEL but they do not include this scenario and i dont know how to go about programming it to rotate the meshes in such a way. Can anyone tell me if this is possible? My lecturer, (im a university student), has currently got me looking at fortran files from a junctionbox routine with a view to modify them to achieve this. Any help or advice to achieve this in the simplist way would be great!

Thankyou!

R.Stringer

Dear Robert,

Rotating domains do not rotate any meshes.. The meshes are fixed in the frame, the equations are modified to account for the non-inertial terms, and the domain interface overlap is recomputed every time step (assuming you using transient rotor-stator).

Are you using single rotating frame of reference, or MFR (multiple frame of reference)?

Opaque.

You can do this by specifying the larger domain as rotating and using a transient rotor-stator type interface between this and the stationary domain. For the smaller domains rotating within this rotating frame, just specify the appropriate equations of mesh motion to rotate them. It is actually simplified by the fact that the solver never really rotates the larger domain, the coordinate system within that domain is essentially moving with it.

-Robin

Hi Opaque and Robin, the domains with the blades in are not on the same axis as the larger rotating domain, as i already explained they need to be rotated about a fixed axis while at the same time rotate about their own axis, which is being moved for each time step. The setup only allows the user to specify the domain to rotate about one axis at any one time, and i want to tell the domian to rotate about two axis.. a fixed axis and a moving axis(that is being rotated about the fixed axis). I have already contacted ansys and they confirm that the software cannot simulate a moving axis in the solver.. however, they say it is possible using a junctionbox. This is far from ideal however as i want to change many variables and would need to generate an enormous amount of meshes. I am hoping.. using fortran or CEL it is possible to get cfx to perform the domian rotation, or move the mesh points so that the blade walls are relocated at the position they should be after each timestep, and return to the solve. If there is a method of specifiying axis points and domain rotation in a CEL expression then perhaps what ure saying is possible robin, please let me know if it is! I hope that is a more clear desciption. hope u can help!

Thankyou!

R.Stringer