Range of wall y+ values

rks171 · January 4, 2012, 12:16

So I ran a simulation for 1000 iterations and it looks like my solution has converged. I checked out the range of my wall y+ values and they are from 0.001 to 193, with most values falling below 30. I read that the standard k-epsilon model doesn't resolve the viscous sublayer, so these values would go too small, but I'm using the realizable two-layer k-epsilon model with all y+ wall treatment which, according to my user manual, says it should resolve the viscous sublayer if my mesh is fine enough. However, I'm new to CFD and turbulence modeling, in particular, so I'm not sure if my y+ values that are 1 or 2 orders of magnitude less than 1 are overkill. Will this slow my solution down due to an excessive number of mesh cells near the wall? Will it cause solution inaccuracies? Should I be targeting a specific range of y+ values?

rks171 · January 4, 2012, 17:54

Closer inspection of the surfaces in my model using a scalar plot of the y+ values reveals that the very low <1 y+ values occur in small crevices in the model where I'd expect low velocities or recirculation to occur. It seems reasonable to me that a very low y+ value in these regions would be expected due to the very low velocity.

January 4, 2012, 12:16	Range of wall y+ values	#1
rks171 Member Join Date: Dec 2011 Location: State College, PA Posts: 87 Rep Power: 14	So I ran a simulation for 1000 iterations and it looks like my solution has converged. I checked out the range of my wall y+ values and they are from 0.001 to 193, with most values falling below 30. I read that the standard k-epsilon model doesn't resolve the viscous sublayer, so these values would go too small, but I'm using the realizable two-layer k-epsilon model with all y+ wall treatment which, according to my user manual, says it should resolve the viscous sublayer if my mesh is fine enough. However, I'm new to CFD and turbulence modeling, in particular, so I'm not sure if my y+ values that are 1 or 2 orders of magnitude less than 1 are overkill. Will this slow my solution down due to an excessive number of mesh cells near the wall? Will it cause solution inaccuracies? Should I be targeting a specific range of y+ values?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
UDF to Access Wall Normal Concentration Gradient	Daniel Tanner	Fluent UDF and Scheme Programming	4	February 18, 2015 15:35
[ICEM] Export ICEM mesh to Gambit / Fluent	romekr	ANSYS Meshing & Geometry	1	November 26, 2011 13:11
natural convection	mehrdadeng	CFX	10	February 25, 2011 06:25
Correct values for constants k E in wall functions	jaswi	OpenFOAM Running, Solving & CFD	0	February 21, 2008 12:33
Wall shear stress values for flat plates?	mark hulme	Main CFD Forum	0	February 9, 2004 00:41

January 4, 2012, 17:54		#2
rks171 Member Join Date: Dec 2011 Location: State College, PA Posts: 87 Rep Power: 14	Closer inspection of the surfaces in my model using a scalar plot of the y+ values reveals that the very low <1 y+ values occur in small crevices in the model where I'd expect low velocities or recirculation to occur. It seems reasonable to me that a very low y+ value in these regions would be expected due to the very low velocity.