You set up a perfect mesh, chose k-ω SST, hit run — and your turbulence quietly dies before the flow even reaches the object. The cause is almost always the same: bad turbulence inlet boundary conditions. Two-equation models like k-ε and k-ω SST need starting values for turbulence at the inlet, but those values (k, ε, ω) are hard to guess directly. The trick every CFD engineer uses is to specify two intuitive quantities instead — turbulence intensity (I) and length scale (ℓ) — and convert them. This guide gives you a free, validated calculator that does exactly that, plus the theory behind every number.
Table of Contents
The Turbulence Inlet BC Calculator
Choose your method, enter the flow details, and get k, ε, ω and the eddy viscosity ratio instantly — ready to paste into ANSYS Fluent, OpenFOAM, STAR-CCM+, COMSOL or SimScale. The pipe mode even estimates intensity and length scale for you from the Reynolds number.
🌀 Turbulence Intensity & Length Scale Calculator
Why Turbulence Inlet Conditions Matter
Two-equation RANS models — k-ε, k-ω SST and their relatives — solve transport equations for turbulence quantities. Like any differential equation, they need boundary conditions. At a velocity inlet, you must tell the solver how much turbulence is entering.
If you get this wrong, one of two failures happens:
- Turbulence decays — you specified too little, and the incoming turbulence dies out before reaching your object, under-predicting mixing, heat transfer and separation.
- Turbulence explodes — you specified too much, artificially thickening boundary layers and smearing out the very flow features you care about.
Turbulence Intensity (I)
Turbulence intensity is the ratio of the velocity fluctuations to the mean velocity, as a percentage:
Guideline values (from ANSYS Fluent and Altair documentation):
| Flow type | Typical intensity I |
|---|---|
| External / free-stream (clean) | ~0.05 – 1% (low) |
| Internal flow (pipes, ducts) | 1 – 5% (medium) |
| Complex / turbomachinery | 5 – 20% (high) |
For fully developed pipe flow, there's a handy empirical correlation that needs only the Reynolds number:
At Re ≈ 100,000 this gives I ≈ 3.8%. The calculator's Auto-Pipe mode applies this for you.
Turbulence Length Scale (ℓ)
The turbulence length scale represents the size of the largest, energy-containing eddies. For fully developed internal flow, the widely used estimate (mixing-length convention, as in Fluent and OpenFOAM) is:
where Dh is the hydraulic diameter. For other geometries: at a turbine-stage inlet, ℓ is often taken as ~5% of the channel height; in grid-generated turbulence, roughly the grid bar size; and for a wall boundary layer, based on the boundary-layer thickness.
The Conversion Formulas
With I and ℓ known, the turbulence quantities follow directly (Cμ = 0.09):
Here k is the turbulent kinetic energy, ε the dissipation rate (used by k-ε models), and ω the specific dissipation rate (used by k-ω SST). Enter k and ε for a k-ε model, or k and ω for k-ω SST.
Two Methods: Length Scale vs Viscosity Ratio
There are two standard ways to specify inlet turbulence, and choosing the right one matters:
| Intensity + Length Scale | Intensity + Viscosity Ratio | |
|---|---|---|
| Best for | Internal flow (pipes, ducts, channels) | External flow (airfoils, vehicles, buildings) |
| Extra input | Length scale ℓ (from Dh) | Eddy viscosity ratio μt/μ |
| Typical ratio | Produces μt/μ ≈ 100+ | μt/μ ≈ 1–10 |
| Why | Clear geometric length available | No meaningful length scale |
For external free-stream boundaries, the NASA Turbulence Modeling Resource recommends setting the eddy viscosity ratio directly between about 1 and 10 — enough eddy viscosity to keep the solver stable without overpowering boundary-layer development.
Worked Example
Water flows through a 50 mm pipe at 2 m/s (ν = 1.004×10⁻⁶ m²/s). Using Auto-Pipe mode:
- Re = U·D/ν = 2 × 0.05 / 1.004×10⁻⁶ ≈ 99,600
- I = 0.16 · Re−1/8 ≈ 3.8%
- ℓ = 0.07 × 0.05 = 0.0035 m
- k = 1.5 · (2 × 0.038)² ≈ 0.0087 m²/s²
- ε = 0.090.75 · 0.00871.5 / 0.0035 ≈ 0.038 m²/s³
- ω = √0.0087 / (0.090.25 · 0.0035) ≈ 48.5 s⁻¹
So at your inlet you'd enter k = 0.0087 and ε = 0.038 for a k-ε model, or k = 0.0087 and ω = 48.5 for k-ω SST. Try it in the calculator — it matches to the decimal.
Common Mistakes
- Entering I as a fraction instead of a percent (or vice versa) — 5 vs 0.05 changes k by 10,000×. The calculator expects percent.
- Setting length scale to zero or the full pipe diameter — use ~7% of Dh, not 100%.
- Using length scale for external flows where no sensible length exists — switch to the viscosity-ratio method.
- Ignoring turbulence decay — even correct inlet values decay over distance; place your inlet close enough, or account for the decay.
- Mismatched model inputs — feeding ε to a k-ω model or ω to a k-ε model. Match the quantity to the model.
- Guessing wildly high intensity — 50% intensity is almost never physical; internal flows are usually 1–5%.
- Forgetting units — velocity in m/s, length in m, ν in m²/s. Mixed units silently corrupt k, ε and ω.
Frequently Asked Questions
What is turbulence intensity in CFD?
The ratio of the RMS turbulent velocity fluctuations to the mean velocity, as a percentage. Below 1% is low, above 10% is high. It's one of the two quantities most used to set turbulence at a CFD inlet because it's easy to estimate.
How do you calculate turbulence intensity for pipe flow?
Use I = 0.16 · Re−1/8, with Re based on the hydraulic diameter. At Re ≈ 100,000 this gives about 3.8%. When no data exists, 1–5% is commonly assumed for internal flows.
What is the turbulence length scale?
The size of the largest energy-containing eddies, ℓ. For fully developed internal flow it's about 7% of the hydraulic diameter (mixing-length convention). It must never be zero, which would imply infinite dissipation.
How do you convert turbulence intensity to k, ε and ω?
k = 1.5·(U·I)²; ε = Cμ0.75·k1.5/ℓ; ω = √k/(Cμ0.25·ℓ), with Cμ = 0.09. These give physically consistent inlet values.
When should I use eddy viscosity ratio instead of length scale?
For external flows (airfoils, vehicles, buildings) where defining a length scale is hard. The ratio μt/μ is typically 1–10 externally and 100+ internally. Use length scale for internal flows with a clear hydraulic diameter.
Why do turbulence inlet boundary conditions matter?
k-ε and k-ω SST models need starting turbulence values at the inlet. Unrealistic values make turbulence decay or grow incorrectly, changing separation, mixing, drag and heat transfer. Deriving them from intensity and length scale gives a correct, stable start.
Conclusion
Turbulence inlet boundary conditions are one of the quietest sources of error in CFD — easy to get wrong, and rarely flagged by the solver. The professional workflow is simple: estimate the two intuitive quantities, turbulence intensity and length scale, then convert them into the k, ε and ω your model actually needs. Use the calculator above to do it in seconds, match the method to your flow (length scale for internal, viscosity ratio for external), and your turbulence model will start from correct, stable physics every time.
Set the inlet right, and everything downstream has a chance to be right too.
For more CFD theory, meshing guides and free engineering calculators, explore Free CFD Tutorial. If this tool helped you, please share it with your classmates and colleagues.


Comments
Post a Comment