[jins9158]

The tendency between the interpretation and the experiment is so different. While reviewing the results, I came up with a question, and I am posting this in frustration.

I am currently working on an interpretation of the sirocco fan.
The analysis conditions were applied as follows.
Since the sirocco fan is interpreted as a steady state condition, the residual is at the level of 10e^-3. And, for the monitoring value, the result value used the average value of 50 iterations.



1. steady analysis
2. Frozen rotor
3. inlet condition: Total Pressure 0Pa
4. Oulet condition: outlet mass




After checking the analysis results, the mass balance at the inlet and outlet was satisfactory.
However, the mass at the rotational interface differed from the inlet and outlet mass values by about 5 to 10%.
(ex, inlet mass = 0.01kg/s, rotating interface = 0.011kg/s)

Note that the mass values entering and exiting the face of the rotating interface are the same.

It's so frustrating because I can't go to the next step in this state.

Thank you very much for your comments.

[ghorrocks]

Your simulation has not converged very tightly. Have you read the FAQ on this? https://www.cfd-online.com/Wiki/Ansy...gence_criteria

Also, please post your output file as an attachment to this forum. Also please attach an image of what you are modelling.

[Gert-Jan]

Please show the surfaces from interfaces. Have you put all in one interface? Maybe you have severe recirculations over your interface, making integrating a hassle.

Also, my experience is that steady state can be a waste of time here. Better perform a transient analysis. This might take a bit more time, but waiting for a steady state setup to provide acceptable results is more time consuming, frustrating and eating your motivation. Take 2 revolutions in 144 steps (5° per step) as a start and see where you get. This might take 10 loops per timestep, making a total of 1440 iterations. That is better than the 1000 iterations that you have spent now on your steady state.
You can also take 2° per time step. Then if CFX converges every time within 5 iterations (or even less), you still spend only 1440x1.25=1800 iterations. That is better value for money than useless steady state.

I have seen steady state setup on fans with massflows going everywhere, except to a converged solution. But these transform to a stable solution, using a transient setup.