Y+ value for logarithmic region

 Register Blogs Members List Search Today's Posts Mark Forums Read

 January 15, 2020, 16:44 Y+ value for logarithmic region #1 New Member   Join Date: Nov 2018 Posts: 4 Rep Power: 3 Hi all. I am doing simple CFD simulation flow over spherical body to understand the concept of y+ and boundary layer. I have a simple question related to log-law layer. I read that for this region y+ value should be between 30 to 300 range. Now, how I would know that for my current simulation y+ value 35 is better or 50 or something higher value?

 January 15, 2020, 17:04 #2 Senior Member   Lucky Tran Join Date: Apr 2011 Location: Orlando, FL USA Posts: 4,026 Rep Power: 48 First, y+ is a fancy coordinate system. Do not confuse this with the more popular wall y+, which specifically refers to the y+ of the cell centroid of the wall adjacent cell. This wall y+ is the y+ of the cell nearest the wall. The log-layer is indeed the region where the y+ is roughly between 30 and a few hundred. For just about any wall-bounded flow with a boundary layer, you can find this log region no matter if the BL is only a few millimeters thick, a kilometers thick, or millions of kilometers thick. So given any flowfield, you (in theory) should be able to calculate the y+ field. That means you have already solved your problem (i.e. run your simulation first). You can then go to the cells nearest the wall, and find the y+ of just those cells. Then you can inspect the value if it is 30, 50, 100, or whatever. y+ is determined a posteriori, not a priori.

January 25, 2020, 03:00
#3
Member

Vignesh Rajendiran
Join Date: Aug 2016
Location: Chennai, India
Posts: 60
Rep Power: 6
Quote:
 Now, how I would know that for my current simulation y+ value 35 is better or 50 or something higher value?

Technically y+ value of 35 should give a good result compared to 50 as the later mesh is a coarser one. But I would advise you to conduct a grid independency test and decide on the mesh.

You can compare your output with the mesh density and decide which mesh to choose. If the output of 35y+ and 50y+ does not vary much, then I advise you to choose to choose the 50y+ mesh as it would reduce the computational time.

And as mentioned by Lucky Tran, the correct y+ value is calculated after a simulation. But you can get an approximate y+ value using your freestream velocity and make a decision from that (because you need some value to work). Once again this y+ value is not the correct one. You have to calculate the y+ from the simulation results and make changes after that.

Vignesh

 February 2, 2020, 02:37 #4 Senior Member   Hamid Zoka Join Date: Nov 2009 Posts: 264 Rep Power: 14 First cell y+ in fact is an onset that separates the analytical solution of the field (from 0 to let say 35 or something else) from numerical solution when using a high-Re turbulence model. In this context selecting the proper lower bound of y+ is more critical since if y+ gets low values (e.g. less than 12) simulation results will be unrealistic and physically meaningless. In other words, in selection of proper y+ value you should be more concerned with minimum acceptable lower bond. while simulation results is not much sensitive to y+ values above that critical value. Moreover, the highest gradients of flow parameters occurs at near-wall region for which a wall function is used and therefore discussing about "better solution" does not make sense specially when a separated region is seen in the flow field. generally one way to check the first cell y+ value is comparing your results with those calculated by a trusted low-Re turbulence model.

 Tags boundary layer, cfd, log layer, y+ value