Implementation of Mixing Length Model in FLUENT ( also called zero equation model or Algebraic Model)
I can find that FLUENT-6.1.22 has a mixing length turbulence model which can be activated through the TUI command;
This activates "Turbulent Mixing Length" in the GUI (Define-->Models-->Viscous panel)
How it is implemented in FLUENT? The FLUENT documentation doesn't provide any information about this model. Moreover, what are the general class of problems where it is possible to use the mixing length model in FLUENT? What are its mesh requirements and accuracy?
The mixing-length or the zero-equation turbulence model (also known as the algebraic model) uses the following relation to calculate turbulent viscosity, mu_t:-
mu_t = rho* l^2* S
The mixing length (l) is defined as,
l = min (kd, 0.09*d_max)
where, d is the distance from the wall and the von Karman constant k = 0.419
S is the modulus of the mean rate of strain tensor, defined as,
S = sqrt (2*S_ij *S_ij)
with the mean strain rate S_ij given by,
S_ij = 0.5 *[(du_j/dx_i)+ (du_i/dx_j)]
"What are the general class of problems and when is it possible to use them in FLUENT?
In general, mixing length models are valid only for two-dimensional, unidirectional flows, i.e. no separation, such as flow in a pipe, flow over a flat plate, jets, mixing layers. However, the mixing length must be calibrated for every different type of flow, which very much limits the usefulness of this type of model. The version in Fluent, based on distance from the wall, should be used only for wall bounded, two-dimensional, unidirectional flows. Because there is almost no real industrial CFD problem that satisfies this criteria, in almost every case it is much better to use a model that can be activated without having to use the TUI. These models should provide similar accuracy for the same kind of flows for which the mixing length model is suitable.
"What are its mesh requirement and accuracy?"
The mixing length model, as implemented in Fluent, can only give accurate results in the flows described above. It is a high Reynolds number model and the first grid point should be in the log-layer, but there is almost no basis for using this model in any general industrial CFD problem.
"Where can one use the mixing length model?"
The mixing length model is a very crude model and is not intended for most industrial applications. The reason that it has been made available is to enable users to create turbulence models using user-defined functions. Once enabled, it creates storage for the turbulent viscosity which can be overwritten with user-defined functions. Thus, if one wants to implement a custom turbulence model in Fluent, one can activate the mixing length model and then use the DEFINE_TURBULENT_VISCOSITY function and DEFINE_PRANDTL functions to implement the custom model without having any additional turbulence transport equations activated.