What is CFDPre?

CFDPre is an open-source collection of object-oriented software tools for calculating boundary layer mesh dimensions for Computational Fluid Dynamics simulations. Among other things, it can be used to:

• Calculate First Layer Thickness, Growth Ratio and Final Layer Thickess

Installation

In your command line, within Python environment:

    pip install cfdpre

• The Python module can also be installed using pip on Windows, macOS, and Linux.

Usage

yhgrcalc supports two flow types:

• flow_type='internal' (default) — fully-developed pipe/duct flow. massflow_kgpersec and hydraulicdia_mm are used to derive velocity and Reynolds number. Skin friction uses the laminar (Re < 2300, 16/Re) or the Haaland (1983) explicit turbulent correlation, valid across the full turbulent range for smooth and rough pipes (roughness_mm) — unlike Blasius, which is only accurate below Re ≈ 1e5. The prism-stack thickness defaults to the pipe radius, but you can set it explicitly with bl_thickness_mm or as a fraction of the radius with bl_thickness_fraction (filling the entire radius with prisms is rarely what you actually mesh).
• flow_type='external' — flow over a flat plate / aerodynamic surface. hydraulicdia_mm is interpreted as a characteristic length (e.g. chord length), and flow_velocity_mpersec is required (the mass-flow/area velocity is meaningless for external flow). Skin friction and boundary layer thickness use the laminar (Blasius) or turbulent (Schlichting) flat-plate correlations; bl_thickness_mm can override the correlation-based thickness.

If the computed growth ratio exceeds ~1.3 (or the boundary layer is too thin to resolve with the requested target_yplus/num_layers), a UserWarning is raised so you can adjust your inputs.

from cfdpre import yhgrcalc
yhgrcalc('Air', 50, 10, 2.5, 125, 1, 8)

Output:

{'fluid': 'Air',
 'temperature [C]': 50,
 'pressure [bar]': 10,
 'massflow [kg/sec]': 2.5,
 'hydraulicdia [mm]': 125,
 'target yplus': 1,
 'number of layers': 8,
 'flow type': 'internal',
 'roughness [mm]': 0.0,
 'dynvisc [N-sec/m^2]': 1.9762497305390764e-05,
 'thermal conductivity [W/m-k]': 0.028357331300649127,
 'specific heat [cp] [J/kg-k]': 1019.3146170790077,
 'density [kg/m^3]': 10.792698589669245,
 'kinematic viscosity [m^2/s]': 1.8310987878701066e-06,
 'flow velocity [m/sec]': 18.875569021507275,
 'reynolds number': 1288541.1444310248,
 'prandtl number': 0.7103701741111368,
 'skin friction coefficient [cf]': 0.002776948990231969,
 'wall shear stress [tau_wall]': 5.339100066909961,
 'boundary layer thickness [delta99] [m]': 0.0625,
 'height of cell centroid from wall [yp] [m]': 2.6034112015544278e-06,
 'first layer height [yh] [m]': 5.2068224031088555e-06,
 'Growth Ratio': 3.6553638704281375,
 'Final Layer Thickness [m]': 0.04540326342524223}

Note: with target_yplus=1 and only 8 layers at this high Reynolds number (~1.29e6), the growth ratio comes out to ~3.66, well above the recommended ~1.3 maximum — this example raises a UserWarning accordingly. In practice you'd either use far more layers (20-40+) or restrict the prism stack to part of the radius. For example, spanning only 30% of the radius:

yhgrcalc('Air', 50, 10, 2.5, 125, 1, 8, bl_thickness_fraction=0.3)  # GR ~3.04

External flow example (flow_velocity_mpersec is required):

yhgrcalc('Air', 25, 1.01325, 2.5, 1000, 1, 15, flow_type='external', flow_velocity_mpersec=30)

Documentation

In progress - not yet made!

• Project Home Page: https://cfdpre.github.io/ [under construction]
• Users Group: https://groups.google.com/g/cfdpre
• Source code: https://github.com/phsheth/cfdpre
• PyPI Page: https://pypi.org/project/cfdpre/

Call for Contributors

The CFDPre project welcomes your expertise and enthusiasm! Better to discuss on the users group before starting to contribute!

Project Log

January 2025:

1. Created Library

June 2026:

1. Added flow_type ('internal'/'external') support. Internal flow uses laminar 16/Re and the Haaland (1983) turbulent correlation (valid to high Re, smooth/rough via roughness_mm); external flow uses Blasius/Schlichting flat-plate correlations and now requires flow_velocity_mpersec. Prism-stack thickness is selectable via bl_thickness_mm or bl_thickness_fraction. Added growth-ratio (>1.3) and boundary-layer sanity warnings.

Project RoadMap:

1. Documentation for existing functionality.
2. Include example data within library.
3. Object-oriented API refactor + input validation.
4. Generate snappyHexMeshDict-ready addLayersControls snippets.
5. Bundled examples, unit tests, CI, and docs site.
6. Parametric/batch sweeps.