Hi, guys! I am modeling an rotaing fan in a large round duct. Multiple Reference Frame is used, but how large is the rotaional region is still a big problem? Does anybody have some idea about it? Is there any strict rules to define the rotaional region? What is the length of the rotational region in the axis direction, and what is the diameter of it? Could you give me some of your ideas? Thank you!

Bono

I think a lot of people are doing this, do you define the rotational region arbitrarily? If not, how do you define it? Any body has some ideas to share? THX!

What do you mean by "rotational region"? Isn't the geometry of your model given? You don't have a choice.

Thanks for responding, Mani. I am using multiple reference frame, so I have to define the rotational fluid zone, but I am not sure how large it should be. The fluid is rotating aroud and with the ratating fan, I mean. So I have to seperate such a region from the geometry as rotational region. Have you ever simulated such case that a fan in a lareg duct? I am doing it recently.

When using MRF, i do not think the size should matter, the size matters when we are using sliding mesh, because in sliding mesh approach, the mesh moves and the sizes are important for enforcing continuity. About the formulation, you can eaither chose to solve your equations in each frame individually and then change to global reference frame. Or you can change values to global reference frame and then solve all of them together. It all depends on you. of course these two values are related by: v = v_r + (vec_omega x vec_r )+ v_t where v is absolute value, and v_r is the velocity in the relative non-inertial reference frame, and v_t is the translational velocity of the noninertial reference frame

Thank you very much, ZXAAR! I really appriciate your response!

actually if you look into documents of any good software like fluent etc. you will find the formulation, That gives better idea how exactly it is solved.

I have found that the spinning domain size has a significant impact on results. My situation involves a reasonably large engine cooling fan, within a shroud, that transitions from axial to radial flow on the left hand side of the fan curve. With a small domain, the fan curve is reasonably represented up to the knee. However, the dP results get progressively worse â€" under predicted - as you move to the left. By increasing the size of the domain, the prediction on the LHS gets better.

Make your domain BIG.

I see. If the transformation between the reference frames is done correctly, the size really shouldn't matter. If you did find differences, depending on the size of the various regions, I would suspect there is something wrong the formulation. In that case, your simulation would be somewhat arbitrary and not very trustworthy.

I found the same result like Paulh, when the shroud of the roationla region expands the result varies. So according to ZXAAR and Mani, the size of the ratiaonal region does not matter, why there are single reference frame and multiple reference frame? If the size does not matter, can we use only single reference frame in the case that a fan in a large room or round duct. I mean can we expand the rotational region near to the walls of the room or the duct, though most of the region is not rotating? I am just starting to do these kinds of cases recently and really want good advice from you! If there is something wrong, please correct me. Thanks!