1.5 Stage compressor Problem

zacko · August 18, 2017, 04:38

Hello there,

I was trying to simulate a 1.5 compressor stage in CFX. The geometry is very easy and the mesh is generated with TurboGrid. I simulated it in steady state (maybe wrong because of the interaction between Rotor and stator, right?)

In the attached screenshots you can see the setup with the interfaces and then the results with streamlines. This looks like the interface from Stator1 to Rotor1 is wrong.
I also changed the interface mode between Stage and frozen rotor. The results are nearly the same.

Can someone give me a hint, what could solve this? Did I choose wrong interfaces ?

Thank you for your help!!!

Have a nice day!

ghorrocks · August 18, 2017, 05:59

It does not look converged to me. You need to run it longer and converge it more.

zacko · August 18, 2017, 06:05

Quote:

Originally Posted by ghorrocks

It does not look converged to me. You need to run it longer and converge it more.

I ran this with 100 iterations. I try it with 500 and report the results.

Thanks for the quick respond!

ghorrocks · August 18, 2017, 06:15

The residuals are the usual guide for how well converged it is. You want it to converge to your residual criteria, which should probably be something like 1e-5.

zacko · August 18, 2017, 06:18

Quote:

Originally Posted by ghorrocks

The residuals are the usual guide for how well converged it is. You want it to converge to your residual criteria, which should probably be something like 1e-5.

Yes, that might be a problem. When 100 iterations were reached the RMS value was only at 1e-4. Now i have set the max value to 1e-6.

zacko · August 18, 2017, 09:01

so 500 iterations are over. I attached the residuals and the streamline view.

The RMS curves don't look good, too in my opinion.
Is it possible that the angular velocity isn't optimal (torque R1= -92Nm) and that these Streamlines occur from this ?

ghorrocks · August 19, 2017, 07:10

You convergence never got to 1e-4, so it is highly likely it is not converged. See the FAQ: https://www.cfd-online.com/Wiki/Ansy...gence_criteria

Note that if you are simulating a point off the designed operating point of the device you will get massive separations and complex flow patterns. It will also be very difficult to converge. So make sure that you are simulating the correct operating point before you fiddle with the convergence parameters.

zacko · August 19, 2017, 07:22

Thank you for your advice !

I was so concentrated on the RMS P Mass that i didn't notice that the other didn't reach 1e-4.

Is it possible to find the operating point through many simulations with parameters? Is there a variable like the zero net torque for turbines to find the operating point ?

Thank you for your time !

ghorrocks · August 19, 2017, 07:25

You can find the operating point by simulation but that is the slow way. The best way is to ask the person who designed it.... (or look at their data sheets). Failing that you can estimate the speed by looking at blade angles and the fluid velocity.

But the best way forward depends on what you are trying to do. What are you trying to do? Why are you doing the simulation?

zacko · August 19, 2017, 07:34

Okay than I will start from zero with designing it with a purpose and with calculations.

There is no particular reason. I did the simulation just for me to keep the "workflow" in mind. So the design of the blades was very random I didn't even choose a specific profile type.
But thats what I'm doing next to achieve a better result.
Hope your not mad that these questions were not for a bigger project or so

Thank you !

ghorrocks · August 19, 2017, 07:58

No, I am glad that you are learning by trying things out. Hopefully you have learn a lot so next time you do it you can do it better.

So make sure the next design is operating at a point where you expect the flow to be fully attached. The operating point will be near there, and the simulation will be much easier to complete properly.

zacko · August 19, 2017, 08:02

Yes, definetly!

I will try my best to implement your tips to achieve good reulst!

Thank you very much

August 18, 2017, 05:59		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	It does not look converged to me. You need to run it longer and converge it more. zacko likes this.

August 19, 2017, 07:10		#7
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	You convergence never got to 1e-4, so it is highly likely it is not converged. See the FAQ: https://www.cfd-online.com/Wiki/Ansy...gence_criteria Note that if you are simulating a point off the designed operating point of the device you will get massive separations and complex flow patterns. It will also be very difficult to converge. So make sure that you are simulating the correct operating point before you fiddle with the convergence parameters. zacko likes this.

August 19, 2017, 07:58		#11
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	No, I am glad that you are learning by trying things out. Hopefully you have learn a lot so next time you do it you can do it better. So make sure the next design is operating at a point where you expect the flow to be fully attached. The operating point will be near there, and the simulation will be much easier to complete properly. zacko likes this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
settlingFoam unstable?	bendel_boy	OpenFOAM Running, Solving & CFD	38	July 8, 2016 05:07
Diverging simulation	czhongrong	OpenFOAM Running, Solving & CFD	45	November 21, 2015 07:17
[blockMesh] Another cylinder question	bendel_boy	OpenFOAM Meshing & Mesh Conversion	5	January 6, 2015 05:09
Stage Interface errors & requirements	Teradil	CFX	0	April 16, 2013 02:36
CFX4.3 -build analysis form	Chie Min	CFX	5	July 12, 2001 23:19

August 18, 2017, 06:15		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	The residuals are the usual guide for how well converged it is. You want it to converge to your residual criteria, which should probably be something like 1e-5.

August 19, 2017, 07:22		#8
zacko Senior Member Daniel Join Date: Feb 2017 Location: Germany Posts: 147 Rep Power: 9	Thank you for your advice ! I was so concentrated on the RMS P Mass that i didn't notice that the other didn't reach 1e-4. Is it possible to find the operating point through many simulations with parameters? Is there a variable like the zero net torque for turbines to find the operating point ? Thank you for your time !

August 19, 2017, 07:25		#9
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	You can find the operating point by simulation but that is the slow way. The best way is to ask the person who designed it.... (or look at their data sheets). Failing that you can estimate the speed by looking at blade angles and the fluid velocity. But the best way forward depends on what you are trying to do. What are you trying to do? Why are you doing the simulation?

August 19, 2017, 07:34		#10
zacko Senior Member Daniel Join Date: Feb 2017 Location: Germany Posts: 147 Rep Power: 9	Okay than I will start from zero with designing it with a purpose and with calculations. There is no particular reason. I did the simulation just for me to keep the "workflow" in mind. So the design of the blades was very random I didn't even choose a specific profile type. But thats what I'm doing next to achieve a better result. Hope your not mad that these questions were not for a bigger project or so Thank you !

August 19, 2017, 08:02		#12
zacko Senior Member Daniel Join Date: Feb 2017 Location: Germany Posts: 147 Rep Power: 9	Yes, definetly! I will try my best to implement your tips to achieve good reulst! Thank you very much