[robinfisichella]

I have a geometry of a car with open wheels and exposed suspension.

I have meshed it using a trimmer mesh with prism layers.

Reynolds number is 8.7x10^6 reference length is 4m

Most of the body is streamlined though the suspension components and wheels are more bluff.

Im not really sure what to do with y+ values, after one 1000 iteration run i have visulised the y+ and it ranges from mostly very small almost zero values to 1025. The high values are seen around the suspension components, and their connections to the main body.

I want to capture the physics of any turbulence created by the model but i also want the detail of the boundary layer as skin friction could be very important. What y+ value range would people recommend, and which physics model is appropriate for the range.

I want to use the spalart allmaras model if possible.

[robinfisichella]

Anyone please.

[DoHander]

Try to keep your y+ from 100 to 150. If you have y+ in the range 5-30 you should refine or coarsen your mesh in this region.

Do

[robinfisichella]

Originally Posted by DoHander

Try to keep your y+ from 100 to 150. If you have y+ in the range 5-30 you should refine or coarsen your mesh in this region.

Do

Thanks, as you can see on my jpeg the y+ is well beyond that range and much smaller in some cases as well!

I dont understand why it is so wayward, the prism layers are something like 15 layers in a thickess of 0.04m = 40mm that should be right for the boundary layer on this.

Also how did you come up with the range? I really need to know, is it just experience?

[galtmann]

Those y+ ranges come from the wall treatment your turbulence model uses. I don't know the terminology that STAR uses, but in Fluent they say "enhanced wall treatment" or "standard wall functions". If standard wall functions are used, then you want to aim for a y+ of 120-150, with ranges of 60-250 acceptable. This is because in order for the wall function to work, it has to be in the logarithmic area of the boundary layer. If enhanced wall treatment is used, then that means that the turbulence model will resolve the entire boundary layer. Here you want most of your cells in around y+ of 1, with ranges up to 5 acceptable, because this is the laminar sublayer area of the boundary layer. Any cells in the 5-60 range is dangerous because this is in the transition area of the boundary layer.

http://www.iafr.eu/TESI/2_file/image063.jpg

Hope this helps.