Wall and Boundary Scale Divergence - All other terms converging nicely

AeroheadG8 · November 16, 2018, 12:35

Hello,

I am a long time user of CFX (~10 years), but have never encountered this issue before, so am in need of some help.

My Wall and Boundary Scale equations are diverging in my run, while all other residuals and user points are trending perfectly. Since this isn't affecting any flow variables, I'm unable to properly diagnose the location in CFX-Post. I have tried this with two different inflation layer schemes, after finding that the first one had transitional y+ values (between 2 and 40). The next mesh put 20 layers in the same thickness where 12 were previously.

I am modelling a centrifugal compressor with a slightly non-traditional layout due so some flow regulating IGV's on the axial inlet. Those IGV's are isolated in their own annular passage (radius ratio only around 1.3), and therefore the stationary domain is applied around them with well-posed axial sector stage interfaces on the upstream and downstream sides . The rotor domain encapsulates the entirety of the rest of the problem with an axial inlet and radial outlet. Both stationary and rotating walls are present in the rotor side mesh, with counter-rotating walls applied appropriately. There is also a free-slip "straw" extending axially from the tip of the compressor nose to the inlet. Inflation layers wrap from this straw onto the viscous surfaces.

K-epsilon solves the problem quite well. SST diverges within 45 iterations as Wall and Boundary scale go to 1e+25. This is independent of time step (since I don't believe that equation adheres to the timescale value).

Thank you very much in advance!

ghorrocks · November 16, 2018, 17:21

The wall scale equation rarely causes a problem so the fact you are having issues with it suggests something is unique in your case. I recall the wall scale equation is just a simple diffusion equation so should be very stable.

The only thing I can suggest causing this would be poor mesh quality. It probably is just one or two elements which are really bad and causing local divergence. I would do some mesh diagnostics to try to find these bad elements and fix them up.

There are also a few parameters you can use to control the wall scale equation but this is not really addressing the real problem. I suspect you need to fix those bad elements and then everything will work fine.

Gert-Jan · November 16, 2018, 18:00

About 10 years ago I have seen such a problem before. If I remember it correctly, it had to do with the free slip condition on certain walls. I would set them back to no-slip to see if the problem disappears.

November 16, 2018, 12:35	Wall and Boundary Scale Divergence - All other terms converging nicely	#1
AeroheadG8 New Member Adam Weaver Join Date: Nov 2018 Posts: 1 Rep Power: 0	Hello, I am a long time user of CFX (~10 years), but have never encountered this issue before, so am in need of some help. My Wall and Boundary Scale equations are diverging in my run, while all other residuals and user points are trending perfectly. Since this isn't affecting any flow variables, I'm unable to properly diagnose the location in CFX-Post. I have tried this with two different inflation layer schemes, after finding that the first one had transitional y+ values (between 2 and 40). The next mesh put 20 layers in the same thickness where 12 were previously. I am modelling a centrifugal compressor with a slightly non-traditional layout due so some flow regulating IGV's on the axial inlet. Those IGV's are isolated in their own annular passage (radius ratio only around 1.3), and therefore the stationary domain is applied around them with well-posed axial sector stage interfaces on the upstream and downstream sides . The rotor domain encapsulates the entirety of the rest of the problem with an axial inlet and radial outlet. Both stationary and rotating walls are present in the rotor side mesh, with counter-rotating walls applied appropriately. There is also a free-slip "straw" extending axially from the tip of the compressor nose to the inlet. Inflation layers wrap from this straw onto the viscous surfaces. K-epsilon solves the problem quite well. SST diverges within 45 iterations as Wall and Boundary scale go to 1e+25. This is independent of time step (since I don't believe that equation adheres to the timescale value). Thank you very much in advance!

November 16, 2018, 17:21		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	The wall scale equation rarely causes a problem so the fact you are having issues with it suggests something is unique in your case. I recall the wall scale equation is just a simple diffusion equation so should be very stable. The only thing I can suggest causing this would be poor mesh quality. It probably is just one or two elements which are really bad and causing local divergence. I would do some mesh diagnostics to try to find these bad elements and fix them up. There are also a few parameters you can use to control the wall scale equation but this is not really addressing the real problem. I suspect you need to fix those bad elements and then everything will work fine. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Wrong flow in ratating domain problem	Sanyo	CFX	17	August 15, 2015 06:20
Radiation interface	hinca	CFX	15	January 26, 2014 17:11
Question about heat transfer coefficient setting for CFX	Anna Tian	CFX	1	June 16, 2013 06:28
Error finding variable "THERMX"	sunilpatil	CFX	8	April 26, 2013 07:00
Multicomponent fluid	Andrea	CFX	2	October 11, 2004 05:12

November 16, 2018, 18:00		#3
Gert-Jan Senior Member Gert-Jan Join Date: Oct 2012 Location: Europe Posts: 1,827 Rep Power: 27	About 10 years ago I have seen such a problem before. If I remember it correctly, it had to do with the free slip condition on certain walls. I would set them back to no-slip to see if the problem disappears.