[Wong0912]

Hi all,

I'm simulating centrifugal compressor in steady state using SST with the high speed (compressible) wall heat transfer model. Frozen rotor interface is used. The mesh has about 1.3m elements generated in Turbogrid and Y+ was set to 1. 10% of relative tip clearance is also applied in Turbogrid.

At my design speed of 205krpm, the compressor choke at 0.0547 kg/s. As you can see from the figure below, at 0.0546 kg/s, both inlet and outlet mass flow are equal, efficiency and pressure ratio are converged.

[IMG][/IMG]

However, I have convergence issues at my highest flow rate running at 230krpm. I used total pressure and total temperature at the inlet, mass flow rate at the outlet as my boundary conditions. This time, mass flow of 0.0555 kg/s is used. However the simulation keeps diverging as shown below. The inlet mass flow doens't match the outlet.



I tried to run at 0.045 kg/s, no convergence as well:



Similar error at 0.040 kg/s:



Things I've tried so far:
1) Reducing the time scale factor to 0.5
2) Using ramping function for the 1st 100 time step for rotational speed and mass flow
3) Initialize from my design point convergence
4) Finer mesh to 2.5m elements
5) Increasre the RAM factor

However, I still couldn't get any convergence and the solver always end with a fatal overflow error. Can someone please guide my how to solve this issues ?

[ghorrocks]

Please post images directly on the forum. Do not use third party download sites. Here is how to do it: https://www.cfd-online.com/Wiki/Ansy...n_the_forum.3F

Have you read the FAQ on convergence?
1) https://www.cfd-online.com/Wiki/Ansy...gence_criteria
2) https://www.cfd-online.com/Wiki/Ansy...do_about_it.3F

In addition to the FAQ comments, in your case you have to consider compressible flow issues as well. If you try to define a mass flow rate higher than the choked flow of a duct it will always diverge. Your mass flow rates are very close to the choked flow so you will always be at risk of this.