Large Eddy Simulation (LES) models

Large Eddy Simulation models. More...

Namespaces
module	enum_les
	Enumeration for Large Eddy Simulation models.
module	variables_turbulence_les
	Define variables associated to the Large Eddy Simulation models.

Functions
subroutine, public	mod_compute_filtered_fields::compute_filtered_scalar_field (filtered_field, field, filter_ratio)
	Apply an explicit filter to a scalar field using 1D mask.
subroutine	mod_compute_les_subgrid_viscosity::compute_les_subgrid_viscosity (velocity)
	Compute (dynamic) turbulent viscosity thanks to Large-Eddy Simulations models.
double precision function	mod_wale_functions::compute_g_square (velocity_gradient, ii, jj)
	Compute the variables necessary to calculate the turbulent viscosity in the WALE model.

Detailed Description

Large Eddy Simulation models.

Function Documentation

compute_filtered_scalar_field()

subroutine, public mod_compute_filtered_fields::compute_filtered_scalar_field	(	double precision, dimension(:,:,:), intent(out)	filtered_field,
		double precision, dimension(:,:,:), intent(in)	field,
		integer, intent(in)	filter_ratio )

Apply an explicit filter to a scalar field using 1D mask.

The multidimensional filter is constructed as the tensor product of a 1D discrete mask mask_filter given in module private variable. Available filter-to-grid ratio are Δ/Δx=1,2,4,8,16. Composite Simpson is using as quadrature rule with the following number of intervall used for each filter-to-grid ratioi (Δ/Δx), respectivelly (1,1,2,4,8). The filtering is performed successively along each spatial direction, resulting in an efficient implementation compared to full stencil naive implementation (avoiding cache misses).

For a 1D mask \( w \) of size \( N \), the 2D kernel is:

\[ W(i,j) = \frac{w_i \, w_j}{n} \,. \]

with \( n \) the norm of the kernel

\[ {\left(\sum_i w_i\right)^2} \,. \]

For a single 2Δ interval, the 1D Simpson weights are: \( w = \{1,4,1\} / 6. \) In 2D or 3D, the weights become the tensor product of the 1D Simpson weights. For instance, in 2D:

This produces the classical 3×3 kernel:

  !!   /  1   4   1  \
  !!   |  4  16   4  |  / 36
  !!   \  1   4   1  /
  !! 

Warning

This filter assumes a uniform Cartesian grid; using it on non-uniform meshes may lead to incorrect results.

Parameters

[in,out]	filtered_field	Filtered field.
[in]	field	Input vector field to be filtered.
[in]	filter_ratio	Filter-to-grid ratio.

compute_g_square()

double precision function mod_wale_functions::compute_g_square	(	double precision, dimension(:,:), intent(in)	velocity_gradient,
		integer, intent(in)	ii,
		integer, intent(in)	jj )

Compute the variables necessary to calculate the turbulent viscosity in the WALE model.

compute_les_subgrid_viscosity()

subroutine mod_compute_les_subgrid_viscosity::compute_les_subgrid_viscosity

(

type(t_face_field), intent(in)

velocity

)

Compute (dynamic) turbulent viscosity thanks to Large-Eddy Simulations models.