Y+ (written as y⁺) is a dimensionless number that tells you how fine your mesh is near a wall in a CFD simulation. It represents the non-dimensional distance from the wall to the center of the first mesh cell, and it determines whether your simulation accurately captures what happens in the boundary layer. Getting y⁺ wrong is one of the most common reasons CFD results near walls turn out inaccurate.
The Y⁺ Formula
Y⁺ is calculated as:
y⁺ = (y × u_τ) / ν
Where y is the physical distance from the wall to the nearest cell center, u_τ is the friction velocity (a measure of how strongly the fluid shears against the wall), and ν is the kinematic viscosity of the fluid. The friction velocity itself comes from the wall shear stress: it’s the square root of the wall shear stress divided by the fluid density.
Because y⁺ combines the mesh spacing, the flow conditions, and the fluid properties into a single number, it gives you a universal way to judge whether your near-wall mesh resolution is appropriate. A y⁺ of 1 in one simulation might require a completely different physical cell height than a y⁺ of 1 in another, depending on the flow speed and fluid involved.
Why Y⁺ Matters: The Three Near-Wall Regions
Fluid behavior near a wall isn’t uniform. The boundary layer breaks into three distinct zones, each defined by y⁺ ranges:
- Viscous sublayer (y⁺ < 5): Right at the wall, the flow is dominated by viscosity. Turbulent eddies are suppressed, and the velocity increases linearly with distance from the wall. This is the thinnest layer and the hardest to resolve.
- Buffer layer (y⁺ between 5 and 30): A transitional zone where both viscous and turbulent effects are significant. The physics here are complex and difficult to model accurately. Most wall function approaches deliberately avoid placing the first cell center in this region because results tend to be unreliable.
- Log-law region (y⁺ from 30 to roughly 200): Turbulence dominates, and the velocity profile follows a predictable logarithmic pattern. Wall functions are designed to work in this zone.
Your target y⁺ value depends entirely on which modeling approach you’re using, and placing your first cell in the wrong zone can quietly ruin your results.
Wall-Resolved vs. Wall Function Approaches
There are two fundamentally different strategies for handling the near-wall region, and each demands a different y⁺ range.
In a wall-resolved approach, you make the mesh fine enough to directly calculate the flow all the way through the viscous sublayer. This means your first cell center needs to sit at y⁺ ≈ 1. It’s more accurate, especially for heat transfer and flow separation predictions, but it requires many thin cell layers stacked near the wall, which increases the total cell count significantly.
In a wall function approach, you skip resolving the viscous sublayer entirely. Instead, you use empirical equations (first proposed by Launder and Spalding in 1972) to bridge the gap between the wall and the outer flow. For this to work, the first cell center must land in the log-law region, typically at y⁺ values above 30. If the cell center accidentally falls in the viscous sublayer while wall functions are active, the results become very inaccurate because the wall function equations assume turbulent, logarithmic behavior that simply doesn’t exist that close to the wall.
Target Y⁺ for Common Turbulence Models
Different turbulence models pair with different near-wall strategies, so your choice of model dictates your y⁺ target.
The k-epsilon model relies on wall functions. You want your first cell center at y⁺ above 35, placing it firmly in the log-law region. Going higher is fine up to a point, though accuracy also depends on overall mesh quality. Dropping below 30 puts you in the buffer layer or viscous sublayer, where the wall functions break down.
The k-omega SST model, by contrast, has no built-in wall functions and is designed to resolve the near-wall region directly. A y⁺ of 2 is typically acceptable, though for boundary layer heat transfer you may need y⁺ closer to 1. Ansys and other software vendors generally recommend at least 10 inflation layers (thin cells stacked near the wall) when using SST to ensure smooth resolution through the boundary layer.
How to Calculate the First Cell Height
Y⁺ is a result of your simulation, not something you set directly. But you can estimate the physical cell height needed to hit your target y⁺ before you build the mesh. The process works backward from the flow conditions:
- Step 1: Calculate the Reynolds number from your flow speed, characteristic length (like pipe diameter or plate length), fluid density, and viscosity.
- Step 2: Estimate the skin friction coefficient. For external flow over a flat plate, a common correlation is C_f = 0.058 × Re^(−0.2). For internal pipe flow, use C_f = 0.079 × Re^(−0.25).
- Step 3: Calculate the wall shear stress: τ_w = 0.5 × C_f × ρ × U²
- Step 4: Get the friction velocity: u_τ = √(τ_w / ρ)
- Step 5: Solve for the physical cell height: y = (y⁺ × μ) / (u_τ × ρ)
This gives you an estimate. After running the simulation, you check the actual y⁺ values on your walls and adjust the mesh if needed. Most CFD software can display y⁺ as a contour plot on wall surfaces, making it easy to spot regions where the mesh is too coarse or too fine.
Common Mistakes With Y⁺
The most frequent error is mismatching the y⁺ range with the wall treatment. Using wall functions with a y⁺ of 1, or running a wall-resolved model with a y⁺ of 50, both produce poor results for different reasons. The second common mistake is assuming a single y⁺ value applies everywhere. In practice, y⁺ varies across the surface because the wall shear stress changes. A mesh that gives y⁺ = 1 in an attached flow region might jump to y⁺ = 10 or higher near a separation point or stagnation zone.
Some modern solvers offer “adaptive” or “blended” wall treatments that attempt to handle a range of y⁺ values gracefully, switching between wall-resolved and wall function behavior depending on the local mesh. These are more forgiving, but understanding the underlying y⁺ requirements still matters for interpreting your results and knowing where to trust them.