# [ICEM] y-plus calculation & transmission into icem cfd

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

 May 7, 2016, 06:22 y-plus calculation & transmission into icem cfd #1 New Member   Join Date: Mar 2016 Posts: 4 Rep Power: 10 Hello Everybody, actually I'm trying to model a fluid flow over a pipe in icem and convert it into cfx. It's my first try to do some simulation in ansys software. I decide to do a tetra mesh in the volume and a prism layer on the walls of the pipe. I wondered how big the cells in the prism layer has to be and found out that there is a size called y plus. I understand the theory of boundaries. I searched a long time for the calculation and I think my calculating method is right. The simulation is an exercise for university so it would be very nice if someone could confirm my method. At first I calculate the Reynolds number of the system: - Re (= ~1e4 => turbulent) , and the velocity v [m/s] - lambda [1] - tau [N/m^2] - y [m] for y+=5 - delta= (12.64*D)/(Re)^(3/4) [m] - y=h0 * (q^n-1)/(q-1) = h0*7.4416 with q=1.2 and n=5 - height of the wall nearest cell: h0=y/7.4416 1.: Now I'm not sure if I can explain (especially with my english) where my problem is. So the boundary layer theory says there are 3 layers (sublayer, bufferlayer, loglayer) and for these 3 layers there are defined y+ values (0-5, 5-30, >30). Do I have to calculate different y values with the different y+ values? But in icem I can only define the 'height' (height of the first cell, so h0) and the 'height ratio' (= q, so 1.2). Am I right? 2.: I defined here n=5, so there are 5 layers of the prism layer. How can I find out how many layers I need? 3.: What does the size delta mean? Theoretically it's the height of the viscous layer. It's independent of y+, but depends on the Reynolds number. So does it mean, that I have to do such a big prism layer? And then in some way to calculate how many layers I need and how big is the height h0? I'm still not sure which turbulence model I have to choose, but I think k-omega would be good for good resolution of near-wall regions. How does the choosed turbulence model effects my calculation of y+? It would be so great to get an answer! I really tried to find solution for my own, but now I don't think I can find it by myself. Thanks in advance! greets -smaru

 May 9, 2016, 02:56 #2 New Member   Join Date: Mar 2016 Posts: 4 Rep Power: 10 Can someone help me?

 Tags boundary layer, calculate, wall, yplus