High performance car simulation
 

Good morning everyone,

I am trying to run a simulation of a high-performance car and right now I am trying to make a mesh study. However, I have some doubts that I would really appreciate some help with.

I have calculated the first prism layer height for a y+=1 (0.0103mm) and taken 20% of the total boundary layer height for my simulation (14mm). I have chosen a stretching factor of 1.2 and 12 prism layers. How do these numbers sound?? I am using an airspeed of 40m/s.

Star-ccm+ doesn't allow you to insert the first prism layer height if you want to set up a stretching factor, so I have been playing with the stretching factor and the number of layers to get the first height similar to the one I calculated. Does this approach make sense??

Should I have selected a higher y+? I am expecting to have an important boundary separation, so that is why I decided to use that small value. I have then selected the K-omega SST low y+ treatment, but I am getting reversed flow on the outlet region of my block and a high number of cells with turbulent viscosity limited. what does it mean??
When should I use low y+ treatment and when all y+ treatment??

Thank you very much in advance for the help

1) If you want to capture your first layer height I recommend to use wall thickness option instead of stretch factor and define your near wall thickness and total thickness.

2) Reverse flow at outlet and turbulent viscosity limitation might die out as your iteration increase.

K-omega SST vs K-epsilon
 

Thank you very much for your reply.

After the mesh study, I am doing a turbulent model study. I am comparing the k-epsilon all y+ treatment with K-omega SST all y+ treatment, and I am using the same mesh for both of them of course. However, I am getting very different Cl values. 0.05 with k-epsilon (totally converged) and 0.2 with k-omega, although this last one fluctuates between 0.09 and 0.25 all the time. On the other hand, I am getting mostly the same Cd in both cases without fluctuations. Does it make any sense?? I should be getting similar cl values, right?? and why the CL is not converging as good as CD? I attached a picture of the plots I got with K-omega