mp-reinit: Parameter description

Contents

• base

• time stepping

• adaptive meshing

• reinitialization

• normal vector

• curvature

• output

• profiling

• application specific parameters

---

🔷 base

Parameter	Type	Default	Description
case name	string	not_initialized	Sets the base name for the application that will be fed to the problem type.
dimension	integer	2	Defines the dimension of the problem
number	string	double	Floating point number format. Currently, only ‘double’ is explicitely instantiated. Allowed values: - double
global refinements	integer	1	Defines the number of initial global refinements
do print parameters	boolean	True	Set this parameter to true to list parameters in output
verbosity level	integer	1	Sets the verbosity level of the console output: 0: silent: for non-robust tests and benchmark runs; 1: minimal: for robust tests; 2: detailed; 3: full
fe	object		See table

base: fe

Parameter	Type	Default	Description
type	string	FE_Q	Finite Element.FE_Q: hexahedral continuous finite element with polynomial degree p; FE_SimplexP: tetrahedral continuous finite element with polynomial degree p; FE_Q_iso_Q1: hexahedral continuous finite element with p subdivisions containing linear elements; FE_DGQ: hexahedral discontinuous finite element with polynomial degree p Allowed values: - not_initialized - FE_Q - FE_SimplexP - FE_Q_iso_Q1 - FE_DGQ
degree	integer	1	Defines the degree p of the finite element type. If “type” is “FE_Q_iso_Q1” this parameter defines the number of subdivisions.

---

🔷 time stepping

Parameter	Type	Default	Description
start time	number	0.0	Defines the start time for the solution of the levelset problem
end time	number	1.0	Sets the end time for the solution of the levelset problem
time step size	number	0.01	Sets the step size for time stepping. For non-uniform time stepping, this parameter determines the size of the first time step.
max n steps	integer	10000000	Sets the maximum number of melt_pool steps
time step size function	string	0.0*t	Set an analytical function to determine the time step size. For the prediction of the new time increment, the old time is used.

---

🔷 adaptive meshing

Parameter	Type	Default	Description
do amr	boolean	False	Set this parameter to true to activate adaptive meshing
do not modify boundary cells	boolean	False	Set this parameter to true to not refine/coarsen along boundaries.
upper perc to refine	number	0.0	Defines the (upper) percentage of elements that should be refined
lower perc to coarsen	number	0.0	Defines the (lower) percentage of elements that should be coarsened
max grid refinement level	integer	12	Defines the number of maximum refinement steps one grid cell will be undergone.
min grid refinement level	integer	-1	Defines the number of minimum refinement steps one grid cell will be undergone.
n initial refinement cycles	integer	0	Defines the number of initial refinements.
every n step	integer	1	Defines at every nth step the amr should be performed.
min cells marked to refine	integer	1	Minimum number of cells that must be marked for refinement/coarsening before the mesh is updated.
min indicator threshold to refine cell	number	0.0	Minimum indicator value required for a cell to be considered for refinement.
solution transfer average values	boolean	False	Set this parameter to true to average the contribututions to the same DoF coming from different cells during solution transfer.

---

🔷 reinitialization

Parameter	Type	Default	Description
fe	object		See table
predictor	object		See table
linear solver	object		See table
enable	boolean	True	Set to true to activate reinitialization.
type	string	olsson2007	Sets the type of reinitialization model that should be used. Allowed values: - olsson2007 - elliptic - geometric
interface thickness parameter	object		See table
elliptic	object		See table
geometric	object		See table
hyperbolic	object		See table

reinitialization: fe

Parameter	Type	Default	Description
type	string	not_initialized	Finite Element.FE_Q: hexahedral continuous finite element with polynomial degree p; FE_SimplexP: tetrahedral continuous finite element with polynomial degree p; FE_Q_iso_Q1: hexahedral continuous finite element with p subdivisions containing linear elements; FE_DGQ: hexahedral discontinuous finite element with polynomial degree p Allowed values: - not_initialized - FE_Q - FE_SimplexP - FE_Q_iso_Q1 - FE_DGQ
degree	integer	-1	Defines the degree p of the finite element type. If “type” is “FE_Q_iso_Q1” this parameter defines the number of subdivisions.

reinitialization: predictor

Parameter	Type	Default	Description
type	string	none	Choose a predictor type: none: use old value as initial guess; zero: se zeros as initial guess; linear_extrapolation: calculate the predictor by a linear combination from the two old solution vectors; least_squares_projection: least squares projection (WIP) Allowed values: - none - zero - linear_extrapolation - least_squares_projection
n old solutions	integer	2	Choose the number of old solution vectors considered.This parameter is only relevant for least squares projection.For all other predictors, this parameter will be set appropriately.

reinitialization: linear solver

Parameter	Type	Default	Description
solver type	string	CG	Set this parameter for choosing an iterative linear solver type. Allowed values: - CG - GMRES
preconditioner type	string	Diagonal	Set this parameter for choosing a preconditioner type. Allowed values: - Identity - AMG - ILU - Diagonal
max iterations	integer	10000	Set the maximum number of iterations for solving the linear system of equations.
rel tolerance	number	1e-12	Set the relative tolerance for a successful solution of the linear system of equations.
abs tolerance	number	1e-20	Set the absolute tolerance for a successful solution of the linear system of equations.
do matrix free	boolean	True	Set this parameter if a matrix free solution procedure should be performed.
monitor type	string	none	Set the monitor type of the linear solver. Allowed values: - none - reduced - all

reinitialization: interface thickness parameter

Parameter	Type	Default	Description
type	string	proportional_to_cell_size	Choose the value type of the interface thickness parameter. Allowed values: - proportional_to_cell_size - absolute_value - number_of_cells_across_interface
val	number	0.5	Defines the value of the chosen interface thickness parameter type.

reinitialization: elliptic

Parameter	Type	Default	Description
penalty parameter	number	0.0	Penalty parameter for the enforcement of the initial position of the zero level-set iso-surface during the elliptic reinitialization.
fixed point iteration	object		See table

reinitialization: elliptic: fixed point iteration

Parameter	Type	Default	Description
max n steps	integer	5	Sets the maximum number of fixed point iterations.
tolerance	number	2.22507e-308	Set the tolerance for reinitialization. If the maximum change of the level set field exceeds the tolerance, reinitialization steps will be performed.

reinitialization: geometric

Parameter	Type	Default	Description
verbosity	integer	0	Choose the verbosity level. 0 means silent, 1 means verbose.
max distance	number	1.61792e+308	Maximum distance from the zero-level-set where the signeddistance function is reconstructed.

reinitialization: hyperbolic

Parameter	Type	Default	Description
pseudo time stepping	object		See table
Continuous Galerkin	object		See table
Discontinuous Galerkin	object		See table

reinitialization: hyperbolic: pseudo time stepping

Parameter	Type	Default	Description
n initial steps	integer	-1	Defines the number of initial reinitialization steps of the level set function. In the default case, the number is set equal to the number of max n steps.
pseudo time step size	number	-1.0	Sets the reinitialization time step size. By default, it is computed from the cell size.
pseudo time step factor	number	1.0	Factor on the reinitialization time step size that is computed from the cell size.
max n steps	integer	5	Sets the maximum number of reinitialization steps.
tolerance	number	2.22507e-308	Set the tolerance for reinitialization. If the maximum change of the level set field exceeds the tolerance, reinitialization steps will be performed.

reinitialization: hyperbolic: Continuous Galerkin

Parameter	Type	Default	Description
implementation	string	meltpooldg	Choose the corresponding implementation of the reinitialization operation. Allowed values: - meltpooldg - adaflo
tangential diffusion factor	number	0.0	Factor that multiplies the normal diffusion factor, i.e., the diffusion length, to obtain the diffusion factor in the tangential direction.

reinitialization: hyperbolic: Discontinuous Galerkin

Parameter	Type	Default	Description
factor diffusivity	number	0.25	Set the factor for diffusivity.
IP diffusion	number	100.0	Set the internal penalty for diffusivity.
use const gradient in RI	boolean	False	Set if the Godunov gradient should be updated every reinitialization step.
do CFL based time stepping	boolean	False	Sets a flag if the time stepping should be based on the CFL condition.
time integration scheme	string	LSRK_stage_5_order_4	Determines the general time integration scheme for the pseudo-time integration of the reinitialization equation. Allowed values: - not_initialized - LSRK_stage_1_order_1 - LSRK_stage_3_order_3 - LSRK_stage_5_order_4 - LSRK_stage_7_order_4 - LSRK_stage_9_order_5 - implicit_euler - explicit_euler - crank_nicolson - bdf_1 - bdf_2 - bdf_3 - bdf_4 - bdf_5 - bdf_6 - imex
IMEX integration scheme	string	not_initialized	If an IMEX integration scheme is specified, the integration in pseudo time of the reinitialization is done with an implicit-explicit scheme. This means that the diffusion part is treated with the IMEX integration scheme and the Hamiltonian is treated with the general time integration scheme. When choosing an implicit scheme with A-stability, larger time steps can be chosen, only limited by the stability of the Hamiltonian part. This is done since the diffusion part is the most restrictive part for explicit time integration schemes. If a scheme is set, the time step calculation based on a CFL number assumes an A-stable scheme and only calculates the time step based on the Hamiltonian. Allowed values: - not_initialized - LSRK_stage_1_order_1 - LSRK_stage_3_order_3 - LSRK_stage_5_order_4 - LSRK_stage_7_order_4 - LSRK_stage_9_order_5 - implicit_euler - explicit_euler - crank_nicolson - bdf_1 - bdf_2 - bdf_3 - bdf_4 - bdf_5 - bdf_6 - imex
CFL	number	1.0	Set a CFL number for the pseudo-time stepping in reinitialization.
avoid zero division smoothed signum	number	1e-16	Sets a constant to avoid zero division in the computation of the smoothed signum.
signum smoothness paramater	number	2.0	Sets the smoothness parameter for the smoothed signum.
use directed diffusion stabilization	boolean	False	Sets a flag if directed diffusion stabilization should be used for reinitialization.
hyperbolic weighting function_type	string	smoothed_signum	Sets the type of weighting function for the hyperbolic part of the reinitialization equation. Allowed values: - smoothed_signum - initial_levelset
use spatially constant diffusion	boolean	True	Sets a flag if a spatially constant diffusion should be used for reinitialization.
use interface movement penalization	boolean	False	Sets a flag if a penalization of the interface movement should be used.
gradient error time derivative threshold	number	1e-16	Sets the threshold in the time derivative when a reinitialization procedure reaches a stationary point.

---

🔷 normal vector

Parameter	Type	Default	Description
filter parameter	number	2.0	normal vector computation: damping = (cell size)² * filter parameter
implementation	string	meltpooldg	Choose the corresponding implementation of the normal vector operation. Allowed values: - meltpooldg - adaflo
verbosity level	integer	-1	Sets the maximum verbosity level of the console output. The maximum level with respect to the base value is decisive.
compute normalized vector	boolean	False	If set to true, the normal vector resulting from the filtering equation will be a unit vector.
narrow band	object		See table
predictor	object		See table
linear solver	object		See table
Discontinous Galerkin	object		See table

normal vector: narrow band

Parameter	Type	Default	Description
enable	boolean	False	Set this parameter to true to compute the normal vector only in the interfacial region.
level set threshold	number	1.0	If narrow band is enabled to true this parameter determines the level set treshold for the narrow band.

normal vector: predictor

Parameter	Type	Default	Description
type	string	none	Choose a predictor type: none: use old value as initial guess; zero: se zeros as initial guess; linear_extrapolation: calculate the predictor by a linear combination from the two old solution vectors; least_squares_projection: least squares projection (WIP) Allowed values: - none - zero - linear_extrapolation - least_squares_projection
n old solutions	integer	2	Choose the number of old solution vectors considered.This parameter is only relevant for least squares projection.For all other predictors, this parameter will be set appropriately.

normal vector: linear solver

Parameter	Type	Default	Description
solver type	string	CG	Set this parameter for choosing an iterative linear solver type. Allowed values: - CG - GMRES
preconditioner type	string	Diagonal	Set this parameter for choosing a preconditioner type. Allowed values: - Identity - AMG - ILU - Diagonal
max iterations	integer	10000	Set the maximum number of iterations for solving the linear system of equations.
rel tolerance	number	1e-12	Set the relative tolerance for a successful solution of the linear system of equations.
abs tolerance	number	1e-20	Set the absolute tolerance for a successful solution of the linear system of equations.
do matrix free	boolean	True	Set this parameter if a matrix free solution procedure should be performed.
monitor type	string	none	Set the monitor type of the linear solver. Allowed values: - none - reduced - all

normal vector: Discontinous Galerkin

Parameter	Type	Default	Description
penalty factor	number	100.0	Set the jump penalty factor of the diffusion term

---

🔷 curvature

Parameter	Type	Default	Description
enable	boolean	True	Set this parameter to true if curvature should be computed. This is required in case of surface tension forces.
do curvature correction	boolean	False	Set this parameter to true if the curvature value at the discrete interface i.e. where the level set is 0, should be extended to the interface region.
filter parameter	number	2.0	curvature computation: damping = (cell size)² * filter parameter
implementation	string	meltpooldg	Choose the corresponding implementation of the curvature operation. Allowed values: - meltpooldg - adaflo
verbosity level	integer	-1	Sets the maximum verbosity level of the console output. The maximum level with respect to the base value is decisive.
narrow band	object		See table
Discontinous Galerkin	object		See table
predictor	object		See table
linear solver	object		See table

curvature: narrow band

Parameter	Type	Default	Description
enable	boolean	False	Set this parameter to true to compute the normal vector only in the interfacial region.
level set threshold	number	1.0	If narrow band is enabled to true this parameter determines the level set treshold for the narrow band.

curvature: Discontinous Galerkin

Parameter	Type	Default	Description
penalty factor	number	100.0	Set the jump penalty factor of the diffusion term

curvature: predictor

Parameter	Type	Default	Description
type	string	none	Choose a predictor type: none: use old value as initial guess; zero: se zeros as initial guess; linear_extrapolation: calculate the predictor by a linear combination from the two old solution vectors; least_squares_projection: least squares projection (WIP) Allowed values: - none - zero - linear_extrapolation - least_squares_projection
n old solutions	integer	2	Choose the number of old solution vectors considered.This parameter is only relevant for least squares projection.For all other predictors, this parameter will be set appropriately.

curvature: linear solver

Parameter	Type	Default	Description
solver type	string	CG	Set this parameter for choosing an iterative linear solver type. Allowed values: - CG - GMRES
preconditioner type	string	Diagonal	Set this parameter for choosing a preconditioner type. Allowed values: - Identity - AMG - ILU - Diagonal
max iterations	integer	10000	Set the maximum number of iterations for solving the linear system of equations.
rel tolerance	number	1e-12	Set the relative tolerance for a successful solution of the linear system of equations.
abs tolerance	number	1e-20	Set the absolute tolerance for a successful solution of the linear system of equations.
do matrix free	boolean	True	Set this parameter if a matrix free solution procedure should be performed.
monitor type	string	none	Set the monitor type of the linear solver. Allowed values: - none - reduced - all

---

🔷 output

Parameter	Type	Default	Description
directory	string	./	Sets the base directory for all output.
write frequency	integer	1	Every n timestep that should be written
write time step size	number	1.79769e+308	Write output output every given time step. If this parameter is set, the output write frequency is deactivated.
output variables	string	all	Specify variables that you request to output.
do user defined postprocessing	boolean	False	Set this parameter to true to enable user defined postprocessing.
paraview	object		See table
particles	object		See table

output: paraview

Parameter	Type	Default	Description
enable	boolean	False	Set this parameter to true to activate paraview output.
filename	string	solution	Sets the base name for paraview output files.
n digits timestep	integer	4	Number of digits for the frame number of the vtu-file.
print boundary id	boolean	False	Set this parameter to true to output a vtu-file with the boundary id.
output subdomains	boolean	False	Set this parameter to true to output the subdomain ranks.
output material id	boolean	False	Set to true to output the material id.
write higher order cells	boolean	True	Set this parameter to false to write bi- or trilinear data only. Set this parameter to true to write higher order cell data. Note: higher order cell data can only be written for hexahedron meshes and 2 or 3 dimensions.
n groups	integer	1	Number of parallel written vtu-files.
n patches	integer	0	Control number of patches to enable high-order.

output: particles

Parameter	Type	Default	Description
enable	boolean	False	Set this parameter to true to activate particle paraview output.
filename	string	particle	Sets the base name for particle output files.

---

🔷 profiling

Parameter	Type	Default	Description
enable	boolean	False	Set this parameter to true if profiling should be enabled. It will be automaticallyenabled for verbosity level >=1.
write time step size	number	10.0	Write profiling output every given time step size. If this parameter is set, the specified parameter for write frequency is overwritten.
time type	string	real	Choose the type of time measure to write profiling information. Allowed values: - real - simulation

---

🔷 application specific parameters

Parameter	Type	Default	Description
do update normal vector	boolean	False	Set this parameter to true to update the normal vector each pseudo-time step.

---