Autodesk CFD equivalent of wall lift-off in viscous units?

[Anonymous]

Hello,

Does Autodesk CFD have an equivalent to the "wall lift-off in viscous units," "wall lift-off (in length units)", and "dimensionless distance to cell center" that COMSOL provides?  If so, how do I check them?

I use Autodesk CFD and have been reading up on turbulence models and mesh sizing in general.  I came across an article by COMSOL where they mention checking the "wall lift-off in viscous units" during their analysis.  If they are below a value (11.06), this indicates that the mesh along the walls of a model is likely fine enough.  They also recommend checking the "wall lift-off (in length units)" and the "dimensionless distance to cell center."

Having these values to check seems really helpful for validating model setup.  It's more empirical and less "by feel" than most recommendations I've seen so far.

Here is the original article: http://www.comsol.com/blogs/which-turbulence-model-should-choose-cfd-application/

Thank you,

Eric

[Jon.Wilde]

Hi,

I would tend to use the Nodal Aspect Ratio in conjunction with Y+ values to check the mesh. The latter depends on the turbulence model that has been applied.

Hope that helps - happy to chat more if you need anything else.

Jon

[Anonymous · Accepted solution]

I was not able to find an explicit definition of "wall lift-off in viscous units", but it appears to be equivalent to the y+ value which is commonly referenced in CFD-related discussions and literature. To plot the y+ value in Simulation CFD, enable it through the "Results quanitities" button in the Solve dialog box. The appropriate y+ values for your model depend on the turbulence model you are using (particularly the wall functions). 11.06 is not a universal target value for all cases.

For more reading, see the following articles:

https://en.wikipedia.org/wiki/Law_of_the_wall

http://www.computationalfluiddynamics.com.au/turbulence-part-3-selection-of-wall-functions-and-y-to-...

The "wall lift-off (in length units)" also lacks an explicit definition but appears to be referring to general mesh resolution near the wall. The nodal aspect ratio may assist in this case by ensuring that there are no skewed elements near the wall. The nodal aspect ratio can be plotted by enabling the stream function in the results quantities section of the solve dialog box. 

I am not able to see the difference between "dimensionless distance to cell center" and "wall lift-off in viscous units" since neither are defined. According to the article you provided, "dimensionless distance to cell center" should be unity when directly resolving the boundary layer (no wall functions). This seems consistent with the definition of y+. 

[Anonymous]

Thank you both for such quick and thorough replies!  I was just typing a response to Jon when I saw dschwalbach's entry.

I believe your posts answer my main question:  it sounds like I should be checking y+ and the nodal aspect ratio to help ensure a good mesh.

I am still a little unclear on how to check these values once I enable them in the Solve dialog, though.  dschwalbach mentions "plotting" them, do you mean by using an xy plot on a plane?  If so, where do you recommend I sample from, nearby the wall?  It seems like it would be pretty tedious to check the whole model this way.  I was hoping they would be viewable as a global variable, but I don't see them listed on the test case I'm currently running.

Thanks again for the help!

Eric

[Anonymous]

You cannot plot y+ values while the simulation is running. After your simulation has completed, set y+ as the global variable and view the global results. Adjust the scale accordingly to see which regions on your model need mesh refinement. 

[Anonymous]

Ah, gotcha. Once I stopped the simulation the values showed up immediately. Thank you both for your help!