Centrifugal compressor analysis at Choke and Surge condition

YOUNS · July 12, 2018, 05:26

Hello.

I'm working on centrifugal compressor of the turbocharger for car.
Geometry has 4 domains :
1. Inblock(Inlet domain) 2. Passage(Blade) 3. Outblock(Diffuser) 4. Volute

Condition for analysis is Inlet total pressure and Outlet mass flow rate.
0. steady state
1. Inlet total pressure: 1atm(reference pressure: 0atm)
2. Outlet mass flow rate: 26.7 g/s(@Surge), 69.9 g/s(@Choke)
3. Rotating Speed: 230,000 rpm
4. Frozen Rotor condition at all interfaces.
5. Passage domain and Outblock(Wall Configuration: Yes) domain are
rotating condition.

Design point of this compressor is 46.0 g/s and 230,000 rpm.
In this case, I have a very good converged result and analysis result at surge condition is so so.
But, Problem is choke condition.
I tried to input for initial condition of converged result at design point.
But.... In this case isn't convergence...
So, I tried to condition that has 57.95 g/s condition then It is convergence.
Now, I tried to 63.925 g/s...

How to set analysis condition if i should get converged result?

Thanks for read.

ghorrocks · July 14, 2018, 07:01

Choked flow is often tricky to converge. The standard tips for improving numerical stability apply:
* Improve mesh quality
* Smaller time steps
* Double precision numerics
* Better initial condition

YOUNS · September 20, 2018, 22:07

Quote:

Originally Posted by ghorrocks

Choked flow is often tricky to converge. The standard tips for improving numerical stability apply:
* Improve mesh quality
* Smaller time steps
* Double precision numerics
* Better initial condition

Thanks your reply

I have a one more question.
To my knowledge, it recommend that setting the time step is 1/RPM(rad/s) from cfx guide.
But, I used to very large time step(about 20) than recommend value.
This time step give good convergence.

How about your opinion this situation?
I am confuse always

Thanks for read
Have a nice day!!

ghorrocks · September 21, 2018, 06:34

If it is a steady state simulation and it converges with a large time step then that is OK. It means your simulation is quite stable and should converge quickly.

July 12, 2018, 05:26	Centrifugal compressor analysis at Choke and Surge condition	#1
YOUNS New Member seunghyeon, yun Join Date: Apr 2018 Posts: 18 Rep Power: 8	Hello. I'm working on centrifugal compressor of the turbocharger for car. Geometry has 4 domains : 1. Inblock(Inlet domain) 2. Passage(Blade) 3. Outblock(Diffuser) 4. Volute Condition for analysis is Inlet total pressure and Outlet mass flow rate. 0. steady state 1. Inlet total pressure: 1atm(reference pressure: 0atm) 2. Outlet mass flow rate: 26.7 g/s(@Surge), 69.9 g/s(@Choke) 3. Rotating Speed: 230,000 rpm 4. Frozen Rotor condition at all interfaces. 5. Passage domain and Outblock(Wall Configuration: Yes) domain are rotating condition. Design point of this compressor is 46.0 g/s and 230,000 rpm. In this case, I have a very good converged result and analysis result at surge condition is so so. But, Problem is choke condition. I tried to input for initial condition of converged result at design point. But.... In this case isn't convergence... So, I tried to condition that has 57.95 g/s condition then It is convergence. Now, I tried to 63.925 g/s... How to set analysis condition if i should get converged result? Thanks for read.

July 14, 2018, 07:01		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	Choked flow is often tricky to converge. The standard tips for improving numerical stability apply: * Improve mesh quality * Smaller time steps * Double precision numerics * Better initial condition __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

September 21, 2018, 06:34		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	If it is a steady state simulation and it converges with a large time step then that is OK. It means your simulation is quite stable and should converge quickly. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.