Fan MRF modelling with half domain???
 

Hi everyone,

I'm performing the CFD analysis of a demo fan(centrifugal .. the fan blade is simplified to cylindrical disc! ). The fan is axisymmetric with a a period of 90deg angle

I'm modeling the fan rpm using "Multiple reference frame". I
I was hoping the fan quarter model (with the cu faces assigned "Symmetry" BC will work. However, the results did not were incorrect (the contours on the symmetry faces were no same)

So It appeared to me that a 120 deg symmetry model / 180 deg symmetry model should give correct results. I tried the same with a 180deg symmetry model but again the contours on the symm faces are not matching eachother
I tried the mrf on full model , and the results are symmetric (about a cross section about the axis of rotation)

My questions from the above are as follows:
1) Can't symmetry BC used with MRF modelling??
2) Is Expecting the same velocity profile on the symm faces wrong??

Periodic BC
 

Hello,

What you are actually asking, is about modelling a sector of the Turbomachine. This has been done extensively across industries and the methodology is quite well established.

You need to develop a periodic mesh (in which the nodes from one side are periodic to the nodes on the other side of the sector being modelled - when I say periodic, the nodes should be at the same radius. Use TurboGrid to develop this mesh, as it is the easiest to be done, for all turbomachine meshing)

Once you have a nice periodic mesh, you need to define a periodic mesh interface. Check the manuals about how to define this, based on the software that you are using. The results will then match.