[mkal]

Hello to everyone,
I am trying to do a centrifugal fan analysis with Cfx.Impeller mesh in TurboGrid. As I change the impeller hub diameter as in the first picture, the results are very different. The result is low when I make inlet pipe hub diameter as in the second picture. This hub diameter is determined according to what. I am making mistakes. Please help me.
http://imageshack.com/a/img923/7953/GK4P4D.png
http://imageshack.com/a/img923/7332/pzLsjg.png

[ghorrocks]

You probably have an inaccurate simulation, which makes the results it produces meaningless. Have a look at the FAQ: https://www.cfd-online.com/Wiki/Ansy..._inaccurate.3F

[mkal]

Mr. Gleen, thank you very much.
I checked your grievances, but the same problem persists.
I made mesh independence. All of them are made up of hexa elements. No convergence problem.
I use the Advencion scheme and turbulence numerics high resolution.
I use k epsilon as a model of turbulence.
Interface model frozen rotor results 10% lower. I'm using stage as an interface model.
Does the hub diameter affect the results?
Some fans have a result of 97% while some fans have a 20% error.
Where am I doing wrong. I have not solved this problem for 1 month.
Please help me...

[turbo]

If you want to get some help, provide details on what performance changes made you confused after changing the hub diameter from what to what.

[mkal]

Hi Turbo.
You can find the performance values that come up with the hub diameter change in the next file.
Picture Fan1 : When the diameter of the dhub is changed, the performance values of the Cfx Results vary.
Picture Fan2 : When the diameter of the dhub is changed, the Cfx Results performance values are the same but lower than the experimental results.
Thank you for helping.
I'm sorry for my bad english.

http://imageshack.com/a/img922/3748/ZeY21X.png

[turbo]

A big change between fan1 and fan2 geometry in CFD looks whether the inlet pipe volume is reduced or not. Check how you treated the interface between the pipe and the rotor, especially matching faces. Do not use massflow BC at the exit, but use p-static. In CFD, mass flow BC at the exit is not right in terms of equation nature, while the option was forced to be added only for users' convenience.

[mkal]

Thank you very much for your reply.
Is it normal for Dhub Ø0 to have a lower result?
The same problem happens when I increase the volume of the input tube.
Which should I use FAN1 or FAN2?
Inlet massflow, outlet static pressure, when I apply it also gives wrong result.
Picture for the interface between pipe and rotor. Am I doing it right?
Thank you so much.

http://imageshack.com/a/img924/5622/3OJlo4.png

[turbo]

Above all, your interfaces of S1-S2-R1 look messed up. If more help, provide their definitions with physical hardware shapes. BCs should be p-total-inlet and p-static-exit.

[mkal]

Hi turbo,
In some fans the interfaces worked well this way.
How else can I make an interface?
Thank you.

[turbo]

If you are confident with the interface setting, try it using p-static-exit BC.

[mkal]

Thank you for your time. I used p-static-exit BC but the results were the same. I do not know if there is anything else I can control.

[turbo]

Your post shows the exit-massflow-BC for fan1 and fan2. Try them with pressure BC.

[mkal]

Thank you Turbo. I did what you said, but the problem continues.
Could you send me resources about fan analysis?

[turbo]

Fan CFD is very tricky because of so low compressibility. In other words the map gets quite flat compared to compressors, meaning flow changes at very tiny pressure changes.

[mkal]

How do you do the fan analysis. Impeller mesh Ansys Meshing or Turbogrid? Are you making the impeller geometry complete or periodic? Thank you .

[turbo]

Any of what you asked should work.

[mkal]

Hello turbo.
Thank you so much.

[mkal]

Hello to everyone,
I am doing centrifugal fan analysis with Cfx. It consists of fan, impeller, volute and inlet pipe. I use stage as an interface. I use the turbulence model k-epsilon, advencion scheme high resolution. The propeller mesh Turbogrid, volute and inlet pipe mesh are made in ICEM CFD as a hexa element.
I made a CFD analysis of the 4 fans that were tested. The analysis results of the 3 fans are close to the experimental results. The other Fanda, though I apply the same methods, I get high and absurd values ​​from the test results.

Fan test results:
8 kg / s = 1550 Pa
7.6 kg / s = 2200 Pa
7 kg / s = 3100 Pa
5.8 kg / hr = 4200 Pa
4.3 kg / s = 5100 Pa

Fan CFD results:
8 kg / s = 3600 Pa
7.6 kg / s = 3950 Pa
7 kg / s = 4180 Pa
5.8 kg / s = 4260 Pa
4.3 kg / s = 4600 Pa
I used all boundary conditions such as total pressure input, static pressure output, but the result did not change.
I could not figure out where it was wrong. Please help me . Thank you so much .
http://imageshack.com/a/img923/8560/jfwyrL.png

[mkal]

Hello to everyone, Correction of previous post
I am doing centrifugal fan analysis with Cfx. It consists of fan, impeller, volute and inlet pipe. I use stage as an interface. I use the turbulence model k-epsilon, advencion scheme high resolution. The propeller mesh Turbogrid, volute and inlet pipe mesh are made in ICEM CFD as a hexa element.
I made a CFD analysis of the 4 fans that were tested. The analysis results of the 3 fans are close to the experimental results. The other Fanda, though I apply the same methods, I get high and absurd values ​​from the test results.

Fan test results:
8 kg / s = 1550 Pa
7.6 kg / s = 2200 Pa
7 kg / s = 3100 Pa
6 kg / s = 4200 Pa
4.3 kg / s = 5100 Pa

Fan CFD results:
8 kg / s = 3600 Pa
7.6 kg / s = 3950 Pa
7 kg / s = 4180 Pa
6 kg / s = 4260 Pa
4.3 kg / s = 4600 Pa
I used all boundary conditions such as total pressure input, static pressure output, but the result did not change.
I could not figure out where it was wrong. Please help me . Thank you so much .
http://imagizer.imageshack.us/a/img923/8560/jfwyrL.png