[kzzang3831]

Hello
I am a senior student at a university.

I wonder why we have to separate the rotating areas.
I think it's to limit the cells and nodes to the rotating areas and divide the more detailed areas to the rotating areas.

I wonder if there is no problem with eliminating the rotating round areas and calculating them.

I'm sorry that I'm not good at English writing because I'm a student in a non-English speaking country.

[Gert-Jan]

The impeller is rotating in reality. Therefore you need to put it in a rotating domain. Why do you want to ignore that in a proper CFD analysis? What is your goal of your CFD analysis?

[kzzang3831]

If I give the condition to rotate only the blades of the propeller, and I don't set the rotating domain, would I get a wrong calculation result?
So far, I have only given moving mesh conditions to propeller blades without setting the rotating domain. And no successful results.

[Gert-Jan]

That is not how it is going to work. Results will be wrong.
Also, to get accurate results, you need to run transient cases. There is no other option than to let the domain rotate and set an interface (Transient-rotor-stator interaction) where the domains meet.

[kzzang3831]

I'll try again based on the feedback you mentioned! Thank you so much for your advice.

[evcelica]

Think about it, and how the solver is working.
If only your blades spin as you say, what would your new mesh look like?
Wouldn't it have to re-mesh every time they rotated to any different angle than original?
So why not instead just rotate that surrounding mesh with the blades and just have a single, rotating mesh region. Then an interface is used between the rotating and non-rotating domain at their circumferential interface.

[kzzang3831]

Hello
I don't think I specified a rotating domain in the tutorial link I attached. So I thought I didn't have to specify a rotating domain either. Is there anything I missed?

https://youtu.be/3Ig1kodreAg?si=A0e3N4pF_KzPowA4

[evcelica]

That tutorial is of a gear pump which is using immersed solid, and those solids are using rotating domains. It is done like that because with two overlapping gears you can't do what you are able to do with your single impeller where you split the mesh into a single rotating domain and the non-rotating one. I guess you COULD perform your model the same way, but with your single impeller, you don't have to. I have no idea how the "immersed solids" actually works, but it is surely not as efficient or accurate as the conventional rotating domain methods you are able to perform with nonoverlapping single impellers.

[kzzang3831]

I've only recently finally understood the difference between MRF and Dynamic mesh techniques, and the pros and cons of each technique. I've had a great realization thanks to the experts here.