The Elliptic Blending R-ε model belongs to RANS-RSM type of turbulence models and, hence, does not assume a particular form of

(the kinematic Reynolds stress tensor). It solves the transport equation for
R instead, complemented by the transport equation for the turbulence dissipation rate
ε, and by the elliptic equation for the blending variable, which allows to capture the evolution of turbulence structure properly
Ref. 36. This leads to more accurate prediction of separation, transient phenomena, effects of swirl and curvature. Moreover, correct description of secondary flows is achieved, which eddy-viscosity based models are intrinsically incapable of. The blending procedure additionally improves the model’s ability to take into account the effect of system rotation and characterize behavior in the near-wall region.
The set of Elliptic Blending R-ε model parameters with their default values is
In a rough analogy to The v2-f Turbulence Model, the pressure–strain correlation
Πij and the dissipation tensor
εij are blended between bulk (homogeneous) and near-wall expressions as
where α is the blending variable.
For Elliptic Blending R-ε, the modeling of the homogeneous part of the pressure–strain correlation
Πij involves seven model parameters (which is the same as in
The SSG-LRR Turbulence Model, although

by default)
where is n is the normal vector to the nearest wall. Notice that the relations
Πkk = 0 and
εkk = 2ε still hold.
The modified coefficient in the ε-equation as well as turbulence length scale and turbulence time scale, are given by
In the Elliptic Blending R-ε model, the
Simple diffusion option is not allowed for the
Reynolds stress diffusion model. The calibration of
Hanjalic–Launder,
Mellor–Herring, and
Lumley relative to
Daly–Harlow is made according to
Ref. 35. Adjust
σk if needed.
where lw is the distance to the closest wall.