Reinitialization methods. More...

Namespaces
module	mod_lsm_reinitialization_hj_first_order
	Reinitialization method using the classic HJ algorithm.

module	mod_lsm_reinitialization_cp
	Reinitialization method using the CP algorithm.

module	mod_lsm_reinitialization_hcr2
	Reinitialization method using the HCR2 algorithm.

module	mod_lsm_reinitialization_hj
	Reinitialization method using the classic HJ algorithm.

module	mod_lsm_reinitialization_tools
	Reinitialization common tools for algorithms.

module	type_levelset_reinit_cp_parameters
	Level set CP parameters as a structureAlso define default values.

module	type_levelset_reinit_parameters
	Level set reinitialization parameters as a structureAlso define default values.

Classes
type	type_levelset_reinit_cp_parameters::t_levelset_reinit_cp_parameters
	The LS RCP parameters. More...

type	type_levelset_reinit_parameters::t_levelset_reinit_parameters
	The LS reinitialization parameters. More...

Functions
subroutine	mod_lsm_reinitialization_cp::compute_levelset_reinitialization_cp (my_levelset)
	Compute the reinitialization using the closest points method.

subroutine	mod_lsm_reinitialization_hj::integrate_filtered_hj_equation (phi_0, phi_n, band, filter, dtau, my_levelset)
	Integrate with Euler time scheme the HJ equation.

subroutine	mod_lsm_reinitialization_hj::integrate_filtered_hj_equation_rk2 (phi_0, phi_n, band, filter, dtau, my_levelset)
	Integrate with RK2 time scheme the HJ equation.

subroutine	mod_lsm_reinitialization_hj::integrate_filtered_hj_equation_rk2_tvd (phi_0, phi_n, band, filter, dtau, my_levelset)
	Integrate with TVD RK2 time scheme the HJ equation.

Detailed Description

This module provides several methods for resolving the reinitialization problem for level sets.

Two principal methods have been developed:

The Hamilton-Jacobi (HJ) equation ;
The Closest Point algorithm.

Among the first, a supplementary method has been added in order to enhance the precision, consisting in the constrained reinitialization scheme (HCR2).

These methods are accessible through the LS parameters and the UI.

Function Documentation

◆ compute_levelset_reinitialization_cp()

subroutine mod_lsm_reinitialization_cp::compute_levelset_reinitialization_cp ( class(t_levelset_cp), intent(inout) my_levelset )

The level set is reinitialization thanks to the Closest Points method. The reinitialization is by default 4th order (using 4th order derivatives and 4th order interpolation). Outlines:

Initialize partitionning of the domain
Detection of numerical and inherent kink
Smoothing numerical kink
Compute closest point to the interface
Use the computed closest point to the interface to reinitialize the level set & & when, for a point on the interface isn't perturbed by a nearby kink

◆ integrate_filtered_hj_equation()

subroutine mod_lsm_reinitialization_hj::integrate_filtered_hj_equation	(	double precision, dimension(:,:,:), intent(in)	phi_0,
		double precision, dimension(:,:,:), intent(in)	phi_n,
		integer, dimension(:,:,:), intent(in)	band,
		double precision, dimension(:,:,:), intent(in)	filter,
		double precision, intent(in)	dtau,
		class(t_levelset), intent(inout)	my_levelset )

Note: Before we were using WENO5 when the band==1 and WENO3 in the band==2 (nothing elsewhere) but this seems to be very unstable. Now we are only using WENO5 in the whole band.

◆ integrate_filtered_hj_equation_rk2()

subroutine mod_lsm_reinitialization_hj::integrate_filtered_hj_equation_rk2	(	double precision, dimension(:,:,:), intent(in)	phi_0,
		double precision, dimension(:,:,:), intent(in)	phi_n,
		integer, dimension(:,:,:), intent(in)	band,
		double precision, dimension(:,:,:), intent(in)	filter,
		double precision, intent(in)	dtau,
		class(t_levelset), intent(inout)	my_levelset )

Note: Before we were using WENO5 when the band==1 and WENO3 in the band==2 (nothing elsewhere) but this seems to be very unstable. Now we are only using WENO5 in the whole band.

◆ integrate_filtered_hj_equation_rk2_tvd()

subroutine mod_lsm_reinitialization_hj::integrate_filtered_hj_equation_rk2_tvd	(	double precision, dimension(:,:,:), intent(in)	phi_0,
		double precision, dimension(:,:,:), intent(in)	phi_n,
		integer, dimension(:,:,:), intent(in)	band,
		double precision, dimension(:,:,:), intent(in)	filter,
		double precision, intent(in)	dtau,
		class(t_levelset), intent(inout)	my_levelset )

Note: Before we were using WENO5 when the band==1 and WENO3 in the band==2 (nothing elsewhere) but this seems to be very unstable. Now we are only using WENO5 in the whole band.

Namespaces