CFD Turbulence Models — k-ε, k-ω SST, LES, DES
Compare k-ε, k-ω SST, Spalart–Allmaras, DES and Smagorinsky LES turbulence models. See which are validated, what each costs, and run them in your browser.
A turbulent flow carries energy across scales that no mesh can resolve. Turbulence modelling is the practice of deciding what to compute explicitly and what to represent by a closure. The choice is not a preference — k-ε and k-ω SST disagree substantially on separated flows, and picking the wrong one is a common source of confidently wrong answers.
k-ε is the general-purpose RANS workhorse: two transport equations for turbulent kinetic energy k and its dissipation rate ε, robust and well-behaved away from walls, but historically weak on the near-wall treatment that separated flows depend on. k-ω SST fixes exactly that by blending a near-wall-favourable k-ω model with the free-stream k-ε model through a blending function, and it is the right default for boundary layers and adverse pressure gradients.
Spalart–Allmaras is a single-equation model tuned for external aerodynamics on a single path to the wall — economical and widely used for aircraft. DES and DDES extend RANS to separated regions by switching to a subgrid model where the grid cannot sustain a RANS assumption, which buys accuracy at the cost of grid-induced numerical noise in the switching region.
Smagorinsky sub-grid modelling is the classical LES closure, and CFD Lab also offers multiple-relaxation-time collision operators on the LBM path for sub-grid scale modelling in lattice formulations. Every model in CFD Lab carries a maturity badge backed by a canonical case; where a model runs but has no validation case yet, the badge says MVP rather than claiming coverage it does not have.
Frequently asked questions
Which turbulence model should I use?
k-ω SST for boundary layers, separation and adverse pressure gradients. k-ε for free shear layers and simple internal flows. Spalart–Allmaras for streamlined external aerodynamics on a single path to the wall. LES or DES only when you can afford the mesh resolution they require.
What is y+ and why does it matter?
y+ is the dimensionless wall distance in viscous units. Below about 30 the wall is resolved and you use a wall function; above about 30 the first cell centre sits inside the viscous sublayer and the model equations are integrated to the wall instead. A k-ω SST case can fail purely because y+ is badly placed, with no modelling error at all.
Is LES accurate enough for a browser?
LES needs far more resolution than RANS to resolve most of the spectrum, and the Smagorinsky closure is available on the 2D LBM path with limited benchmark coverage. For general engineering work k-ω SST at a converged grid is usually the better accuracy-per-hour choice.
Try it now
The page you are reading is a static, crawler-readable summary. The interactive tool runs in your browser — nothing is uploaded.
Open CFD Lab Back to home