I am studying external aerodynamics on Ground vehicles. And I am using Star CCM+ for my simulation. To achieve better drag and lift results, I got to know that Y+ plays an important role. So how can I specify the first wall distance in Star CCM+?
I tried to change the number of prism layers, prism layer stretching, Prism layer total thickness. But I didn't find any improvement.
I do not use Star CCM+ but in general you need to keep y+ ~ 1 in case you wish to resolve the boundary layer. That is probably your case since to determine the drag and lift forces correctly, you should do that. Besides, the choice of the turbulence model is important. Not every turbulence model is fit to be applied to turbulence modeling in the boundary layer.
As you probably know, y+ is a function of the distance of the first grid point from the wall. So if you increase the number of cells within the boundary mesh layer while keeping the thickness of this layer constant, you must see a change in the value of y+.
Teorically you should just keep the total thickness of the layer and increase the number of elements in the direction normal to the boundary. Otherwise, if you are not able to reduce the first layer thickness and, consequently the y+, use a wall function insteas of a wall treatment. With the wall functions, y+ can lie between 50 and 500.
Franco
I have given a similar "advice" in other forum of Researcj Gate, but I think that recover it can be helpful for you... anyway...
You can compute preliminary values for u* by means of approximate formulations for the skin friction coefficient. You can see full details in "Boundary-Layer Theory" from Schlichting, and brief details in my PhD project. I have uploaded my PhD thesis in Research Gate, and then, I think that you can download it. The name of the work is Theoretical - numerical study of flows with strong streamlines curvature, PhD Thesis ETSEIB, UPC Barcelona Spain.
By the way, the u* value estimations are useful for defining suitable values for y+, when you use wall functions or some kind of near-wall treatment. Some preliminary values for:
Standard wall function: 40
Good Morning Mr. Nani:
All previous advice is good. I would only add the following.
1. Preliminary runs with coarse grids will help you in figuring out how many cells you will need to resolve the boundary layer.
2. Follow the y+ guidelines of any near wall turbulence model you decide to use. If using standard wall functions Launder & Spalding developed the model to recover the law of the wall or y+ > 100. Nevertheless I have found that SWF perform well down up to y+ > 10-20 depending on the type of flow.
3. If you want to resolve the boundary layer around the object on external flow I strongly recommend getting as low as y+ > 0.1. In this case you should use the Spalart Allmaras turbulence model. In case you are not familiar with it. This model is essentially a two layer model. It uses damping functions near the wall (I do not remember the original proposers of the damping functions) and it turns to the traditional k-e away from the wall.
Good Luck!
Larry
Mr Nani,
I hope you Calculated first Layer Height,If not
Y plus=(Density*Friction velocity*Initial Height)/Dynamic Viscosity.
With All of the above parameters known,we can find out Initial Height.
With the obtained initial height, input the value in STAR CCM+
Continua>>>Mesh>>>In prism layer mesher properties change streching mode to wall thickness.
Then In reference values under Thickness of Near wall Prism layer input this value.
hi Rusheendra Nani
in star ccm you needn't Calculate first Layer Height. because in star in scenes nod in scalars you can plot y+ . if you wanna y+ on some line you can define line in derived parts and then create report y+ on this line.
if you wanna change first layer height gonna in meshes nod in prism layer nod change distribution mod to wall thickness.
other guys give good advice's for y+ range.
best regards
Hey,
It depends on your choice of turbulence closure. If you select k-epsilon turbulence model with wall function near wall, y+ should be around 30 to 50. However, if yoy use Spallart-Allmaras turbulence closure or k-omega, u should use y+ close to 1. As these two models are supposed to work for the viscous sub layer.
Thanks, for more details, [email protected]
Abhishek Kumar Pandey , for K -epsilon y+ should be in range between 30 to 150, i guess its typo from your end.
- Badges
- Science topic