mixing plane vs frozen rotor
 

:(Hi, I am simulating a fan system using MRF method. I have attached a picture showing the convergence path. Interesting, the frozen rotor converged nicely but the mixing plane floats on a high level. From your experience, do you think the mixing plane result is usable? How come they behave so differently. Exactly the same mesh, same physics, but the interface type is different.

Thanks.:)

your question is confusing since mrf is a type of motion to apply to a region whereas mixing plane is a type of interface to connect different regions

Thanks Ping, but I am comparing frozen rotor and mixing plane. CCM+ didn't explain well that the default MRF and interface is in fact frozen rotor method.

mrf is much more than frozen rotor since it applies to linear motion too

for the forum to help you we need to understand what you want to do so i will say this again in star-ccm+ terminology

- mrf is a method to apply motion to one or more regions

- mixing plane is a type of interface to connect two regions and does circumferential averaging of the fields and is used normally in combination with the mrf technique on multi-stage turbines with uneven blade numbers

so you cant 'compare' the two

Thanks Ping, I don't want to waste your time. So I have to repeat my first question. I am comparing frozen rotor and mixing plane. You can say default interface vs mixing plane interface otherwise. The physics is very simple, steady state with a single fan and a simple ducting system.

it would help if you explained this correctly in the first place rather than confusing us and remember than frozen rotor is not a term cd-adapco uses in the documentation

a mixing plane interface smears or averages the flow around its circumference, whereas a normal inplace interface transfers the flow from each local cell on one side to the matching cell on the other side of the interface and maybe with some interpolation if the interface is not conformal

if you create a scalar scene and place just the interface in the scene and show axial velocity for example, the above will be very clearly shown

there is no reason to use mixing plane interfaces unless the circular extent of the geometry on either side of the interface is not identical - eg you model one blade of a fan in a sector but connect it to a full cylindrical model of the rest of the domain