[zacko]

Hello,

i have found an older simulation of mine of a centrifugal pump impeller including a suction chamber and volute.
I transformed the impeller mesh through rotation to get the full impeller. But, instead of using the Glued Adjacent Meshes feature, I created an periodic interface between all blade channels.

My question would be: Will there be a difference between both approaches and how would CFX model this when the full impeller mesh is included but a periodic interface between all channels is used ? Has anyone an idea ?

Thank you for your input and kind regards
Daniel

[ghorrocks]

It would depend on how you set up the periodic boundaries and what you are modelling as to whether it means the simulation is invalid.

But to my thinking it would stuff the simulation up enough that the results are meaningless and it should be thrown out, the problem fixed and the simulation redone.

[zacko]

Originally Posted by ghorrocks

It would depend on how you set up the periodic boundaries and what you are modelling as to whether it means the simulation is invalid.

As far as I see, the periodic interfaces hav been set up using the periodic faces of each passage, e.g.

PHP Code:

     BOUNDARY: R1 Periodic Side 1
       Boundary Type = INTERFACE
       Location = PER1 1 Passage,PER1 1 Passage 2,PER1 1 Passage 3,PER1 1 \
         Passage 4,PER1 1 Passage 5,PER1 1 Passage 6,PER1 1 Passage 7,PER1 2 \
         Passage,PER1 2 Passage 2,PER1 2 Passage 3,PER1 2 Passage 4,PER1 2 \
         Passage 5,PER1 2 Passage 6,PER1 2 Passage 7 


I have attached an image of the impeller, blue are the blades and orange the interfaces.
The result has been actually quite "Okay" in means of the expected hydraulic data such as the head.


I will for sure try to resimulate the setup.

[Gert-Jan]

Around BEP and above, this approach is okay.
If you want to determine the shut off head, I would not give a penny for your results.

[zacko]

Originally Posted by Gert-Jan

Around BEP and above, this approach is okay.
If you want to determine the shut off head, I would not give a penny for your results.

Okay, interesting, so the results could be used.
I'm always very doubtfully about the results outside of the preferred operating region. I usually try to simualte between 70% and 120 % of BEP anyways.

But, nonetheless, I will rerun the simulation and see what comes out of it.