What y⁺ actually measures
y⁺ is a Reynolds-number-like scaling of the distance from a wall, defined as y⁺ = (u_τ · y) / ν, where y is the distance from the wall, ν is the kinematic viscosity, and the friction velocity u_τ = √(τ_w / ρ) is built from the wall shear stress τ_w. Because it is normalized by the viscous length scale, y⁺ tells you where in the boundary layer your first cell lands — not just how many millimeters off the wall it sits.
Why it drives turbulence modeling
Near a wall the mean velocity profile splits into distinct regions, and your turbulence model has to assume one of them:
| Region | y⁺ range | Meshing approach |
|---|---|---|
| Viscous sublayer | 0–5 | Wall-resolved (target y⁺ ≈ 1) |
| Buffer layer | 5–30 | Avoid — no clean assumption |
| Log-law region | 30–300 | Wall functions |
Wall-resolved models integrate all the way to the wall, so the first cell must sit deep in the sublayer at y⁺ ≈ 1. Wall-function models instead bridge the near-wall region with an analytic log-law and expect the first cell centroid at y⁺ ≈ 30–300. Put the mesh in the wrong band and wall shear stress, heat transfer, and separation prediction all drift.
The catch: you need a solution to know y⁺
y⁺ depends on τ_w, which you do not know until you have run the case. So in practice you estimate it from a flat-plate correlation, build the mesh, solve, then check the realized y⁺ and iterate. See first cell height for that backward calculation.
How AutoMesh-Geo helps
Holding a consistent y⁺ across a complex, curved surface is a meshing problem, not just a solver setting — near-wall spacing has to stay uniform even where the geometry pinches or folds. AutoMesh-Geo builds conforming near-wall cells directly from the surface, so your target resolution holds across the whole model, not only the flat, easy regions.