What y+ Means in CFD

Velocity changes rapidly near a no-slip wall. The fluid velocity is zero at the wall and rises through a thin region before joining the outer flow. The physical thickness of that region changes with viscosity, density, velocity, pressure gradient, and wall shear. A distance of 0.1 mm may be extremely fine for one case and far too coarse for another. y+ scales the cell-center distance by a local viscous length so meshes from different flows can be compared on a physically meaningful basis.

The quantity uses wall-normal distance y, friction velocity u_tau, and kinematic viscosity nu. Friction velocity is not the external flow speed; it is a velocity scale derived from wall shear stress. Because wall shear is usually unknown before solving, first-cell-height estimates rely on correlations or an assumed skin-friction coefficient. The achieved y+ must therefore be checked after the solution and the mesh revised when necessary.

Very close to a smooth wall, viscous stress dominates and mean velocity expressed in wall units is approximately proportional to y+. This is the viscous sublayer. Farther out, turbulence production becomes more important and a logarithmic velocity relation can describe part of the boundary layer. Between them lies a buffer region where neither simple limit is fully dominant. The numerical bounds are approximate and depend on flow, roughness, pressure gradient, and modelling convention.

A mesh intended to resolve the wall should place the first cell center within the viscous sublayer and include enough layers to capture the steep velocity gradient. A wall-function mesh instead places the first cell in a range where the selected wall relation is valid. Landing in the buffer region unintentionally can be problematic because a basic wall function may not represent that zone accurately, while the mesh is not fine enough to resolve the sublayer.

Modern solvers may use scalable, blended, or automatic wall treatments that reduce sensitivity to a strict band, but they do not eliminate the need to understand y+. The implementation, model constants, roughness treatment, thermal wall function, and compressibility corrections are solver-specific. Use the solver's own theory and verification material, not a target copied from an unrelated tutorial.

For a wall-resolved low-Re approach, a target near 1 is common because the first cell center should lie in the viscous sublayer. Values substantially above that can under-resolve the gradient and compromise the model's intended near-wall integration. Reaching y+ near 1 also requires adequate layer count, gradual growth, sufficient total prism or inflation thickness, and acceptable cell quality. One tiny first cell followed by abrupt growth is not a complete boundary-layer mesh.

For conventional wall functions, targets in the logarithmic region are common, often expressed as greater than about 30 with an upper range constrained by the model and flow. A value such as 30 is not a universal switch. Low local Reynolds numbers, separation, strong pressure gradients, heat transfer, roughness, and complex geometry can narrow the valid region. Some enhanced treatments accept intermediate values, but their behavior must be verified for the selected solver.

The appropriate target is therefore a modelling decision. Record turbulence model, near-wall option, expected flow regime, target band, first-cell-height estimate, layer count, growth ratio, and total layer thickness together. If different walls serve different physical roles, they may need different mesh strategies. A rotating blade, duct wall, symmetry plane, and slip wall should not be treated as one generic surface set.

A preliminary estimate begins with velocity, characteristic length, density, and viscosity. Calculate Reynolds number, select an appropriate skin-friction correlation, estimate wall shear, derive friction velocity, and rearrange the y+ definition to y = y+ nu/u_tau. The result is the distance to the first cell center. If the mesher asks for first-layer thickness rather than center distance, the relationship depends on cell construction; for a wall-adjacent prism of uniform thickness, the center is roughly halfway through the layer.

Correlations are approximations. An external flat-plate estimate does not necessarily represent an internal duct, separated wake, rotating machine, or impinging jet. Use a conservative local velocity where high shear is expected and calculate a range rather than one exact height. Property variation can matter in heated walls, compressible flow, and multiphase cases. Geometry scaling and mesh-unit mistakes are especially dangerous because a millimetre-to-metre error changes first-cell height by three orders of magnitude.

Inflation or prism layers should cover a meaningful part of the boundary layer, not only the first cell. Estimate boundary-layer thickness where possible, choose enough layers, and keep growth gradual. Inspect aspect ratio, skewness, orthogonality, layer collapse, and abrupt transitions. A nominal first height is not achieved everywhere if the mesher removes layers around tight corners or small gaps.

Consider air at density 1.2 kg/m3 and dynamic viscosity 1.8e-5 Pa s flowing at 20 m/s over a 1 m characteristic length. The Reynolds number is about 1.33e6. Suppose a preliminary correlation gives skin-friction coefficient Cf = 0.004. Estimate wall shear as tau_w = 0.5 rho V squared Cf = 0.5 x 1.2 x 400 x 0.004 = 0.96 Pa. Friction velocity is sqrt(0.96/1.2), about 0.894 m/s.

Kinematic viscosity is 1.8e-5 / 1.2 = 1.5e-5 m2/s. For target y+ = 1, the cell-center distance is y = 1 x 1.5e-5 / 0.894, or about 1.68e-5 m, equal to 0.0168 mm. For target y+ = 50, the estimate is about 0.839 mm. These values differ by a factor of fifty because they support different near-wall strategies.

Neither value is final. Build the mesh, solve to a meaningful state, plot area-weighted and local y+ over every relevant wall, and inspect the distribution rather than one average. If high-shear regions exceed the resolved target or wall-function regions fall into the buffer zone, revise the local mesh and repeat. Report both target and achieved range with the turbulence and wall-treatment settings.

Post-processing should include contour plots and distributions, not only a global maximum or mean. A small high-y+ patch at a leading edge may be more important than a large low-shear area that dominates an average. Check minimum, maximum, percentiles, area fractions within the target band, and the locations of excursions. Exclude boundaries where y+ is not meaningful, such as symmetry or slip surfaces, from summaries.

Convergence also matters. y+ depends on wall shear, so an unconverged or transiently fluctuating solution can change the reported field. For steady simulations, inspect residuals and monitored engineering quantities before accepting y+. For unsteady simulations, examine time histories or statistically converged averages. A mesh-independent output study should include the near-wall resolution strategy and not merely reduce cells uniformly.

When heat transfer is important, thermal near-wall treatment and thermal y+ concepts may require additional checks. Rough walls alter velocity laws and effective origins. Separated and reattaching regions can produce low or changing wall shear. Treat these as physical features to analyze rather than trying to force every cell to one identical y+.

The first mistake is choosing a target before choosing the turbulence model and near-wall treatment. The second is using free-stream velocity directly in place of friction velocity in the y+ formula. Free-stream speed can support a correlation for wall shear, but it is not u_tau. The third is accepting the estimated first-cell height without checking achieved y+ after solving.

Another mistake is focusing only on the first layer. Too few layers, excessive growth, insufficient total thickness, and collapsed prisms can undermine the boundary-layer representation even when the first center lands at the target. Users also confuse first-layer height with center distance or enter geometry in millimetres while the solver interprets metres. Unit auditing must happen before meshing.

Finally, one target does not fit every CFD case. Wall functions, low-Re integration, roughness, heat transfer, separation, transition models, and compressibility impose different requirements. Avoid reporting y+ as a quality badge without the model, mesh, achieved distribution, convergence state, and output sensitivity.

Use the Reynolds Number Calculator to establish the flow scale and verify the viscosity basis. Use the Y-Plus Calculator to connect target y+, wall shear or friction velocity, viscosity, and cell-center distance. The First Cell Height Calculator supports the practical mesh dimension. Keep every assumption visible, especially the skin-friction correlation and whether the requested mesh value is a layer thickness or center distance.

The Turbulence Intensity Calculator can support an inlet estimate, but inlet turbulence and near-wall resolution are separate choices. For OpenFOAM setup, the Y-Plus Wall Treatment Helper can organize wall-function selections and checks without executing a solver. After simulation, use the solver's y+ field and wall-shear results as the authoritative achieved values. Iterate the mesh and compare engineering outputs such as force, pressure loss, and heat transfer.

Related in this workflow: Y-Plus Calculator, First Cell Height Calculator, Reynolds Number Calculator, Turbulence Intensity Calculator, OpenFOAM Y-Plus Wall Treatment Helper.

First-cell-height calculations are estimates based on simplified wall-shear relations. They cannot predict the final local y+ field in complex geometry, separation, strong acceleration, rotation, buoyancy, transition, roughness, or conjugate heat transfer. Solver definitions and wall-function ranges differ. Always consult the documentation for the exact model and version used.

A satisfactory y+ distribution does not prove that a CFD result is accurate. Domain extent, boundary conditions, turbulence modelling, discretization, time step, convergence, mesh independence, material properties, and validation remain necessary. Use the guide for planning and diagnosis, then verify consequential simulation decisions with established CFD practice and qualified professional judgement.