PHASTA Wiki - User contributions [en]

TCNEQ Version

2022-03-18T01:10:37Z

Jopo9795: /* Post-Processing */

== Background ==
The following information relates to the use of the thermochemical nonequilibrium (TCNEQ) version of PHASTA written in terms of entropy variables. The reader is referred to the following for additional information.

* F. Chalot, T.J.R. Hughes, and F. Shakib, '''"Symmetrization of Conservation Laws with Entropy for High-Temperature Hypersonic Computations,"''' Computing Systems in Engineering, 1(2-4):495–521, 1990.

* J. Pointer, '''"Influence of Interpolation Variables and Discontinuity Capturing Operators on Inviscid Hypersonic Flow Simulations Using a Stabilized Continuous Galerkin Solver,"''' Ph.D. dissertation, University of Colorado, Boulder, CO, 2022.

== Pre-Processing ==
In this section, details of the meshing and model attributes are provided. For simulation cases where the gas is modeled as a single species, the ''scalar'' quantities for boundary and initial conditions are not required. Currently, capability exists to simulate a gas with number of species (''nsp'') ≤ 5. When 2 ≤ ''nsp'' ≤ 5, the ''scalar'' quantities are used to govern the composition of the gas.

=== Meshing ===
Within the Simmodeler utility, the mesh can either be created or loaded from an existing .cas file. Below are steps for loading a mesh from a .cas file:
# Launch Simmodeler (for this example, SimModeler7.0-190604 is used)
# File > Import Discrete Data > (select .cas file to import) > (keep defaults and click OK) > (select YES to keep volume mesh)
# Save .sms and .smd files
# Attributes can now be assigned to the model as normal

=== Boundary Conditions ===
Below are the recognized boundary conditions that can be applied for the current version:
* comp1/comp2/comp3 - Specification of one/two/three components of velocity, [m/s]
* temperature - Specification of translational-rotational temperature, [K]. By default, vibrational temperature is held in equilibrium with this value and nonequilibrium is controlled through simulation inputs.
* surfID - When value is set to 702, the boundary is treated as a slip wall. If using this option, include a boundary layer mesh along the surface to ensure the wall normal direction is accurately computed.
* scalar_1 - Mole fraction of species 2 of the gas
* scalar_2 - Mole fraction of species 3 of the gas
* scalar_3 - Mole fraction of species 4 of the gas
* scalar_4 - Mole fraction of species 5 of the gas
* pressure - Specification of static pressure over a surface, [Pa]
** Used to compute mole fraction of species 1 of the gas with Dalton's Law of partial pressures and subtracting the summation of the other mole fractions from a value of 1
* heat flux - set to zero for adiabatic wall boundary condition

=== Initial Conditions ===
Below are the required initial conditions for the current version:
* initial velocity - Components and magnitude of flow velocity, [m/s]
** If a supersonic outlet condition is used, set such that flow is initialized Mach > 1
* initial temperature - Value used to set translational-rotational temperature, [K]
* initial scalar_1 - Initial value of species 2 mole fraction
* initial scalar_2 - Initial value of species 3 mole fraction
* initial scalar_3 - Initial value of species 4 mole fraction
* initial scalar_4 - Initial value of species 5 mole fraction
* initial pressure - Static pressure of the gas, [Pa]
** For multi-species flows, this value is used in combination with the initial scalar values to compute the mole fraction of species 1

== Simulation Inputs ==

Below is an example of the input script for the current version of the code. Capability is included for handling multi-species flows up to number of species ''nsp'' equal to 5. This feature has been tested and shown to work for five species mass conservation equations at one time. However, when chemical species production governed by the finite-rate chemistry module is active, the solution becomes unstable. Additional work on this feature is needed.

=== Example of solver.inp file: ===

<nowiki># PHASTA_HYP Version 1 Input File

#SOLUTION CONTROL
#{
Equation of State: Compressible
Number of Timesteps: 100
Time Step Size: 3.2339656949004e-08

Limit Density: 0 0.01 0.1 # solution limiting on variables [switch, min, max]
Limit u1: 0 0. 2.8e3
Limit u2: 0 0 0
Limit u3: 0 0 0
Limit Temperature: 0 230 3500 # also limits vibrational temperature
#}

#OUTPUT CONTROL
#{
Number of Timesteps between Restarts: 250
Print Error Indicators: True # shock error stored in column 6, DC factor \nu stored in column 10
Error Indicator Threshold: 0.01 # err > thresh*err_max is flagged as 1 (i.e. identified for refinement)
# --> smaller values = narrower flagged region along shock
Number of Error Smoothing Iterations: 0 # ierrsmooth
Load and set 3D IC: False # load the flowfield from a file as the initial condition
Position Tolerance on IC Load: 1e-7 # sets the tolerance for matching node locations while loading the initial condition
#}

#MATERIAL CONTROL
#{
Viscosity: 1.95508431704028e-5 # dynamic viscosity (only used if nsp.eq.99 for air mixture)
Thermal Conductivity: 26.6843390135759e-3 # only used if nsp.eq.99 for air mixture
Viscous Control: None #Viscous #None

#}

#REACTING FLOW
#{
Number of species: 1 # nsp
#.......currently only allowing 1<=nsp<=5.
#.........and make sure scalars passed from simmodeler are in correct slots
#
Species IDs: 1 # specIDs, length of array must equal
# nsp (IDs numbered in order:
# N2,O2,NO,N,O
# ID:99 for Air molecule
# e.g. Species IDs: 1 3 <<< would give N2 and NO
Ref Entropy Conditions: 1e3 230 230 0 #[P0,T0,T0vib,S0]
Ref Entropy Mole Frac: 1 0 0 0 0 # composition of gas used as reference entropy condition
# >> must sum to zero
Allow reactions: False # chemical reactions, ichem = 1 if True
Chemical heat release: False # chemical heat release, iqtot = 1 if True
Limit on reaction step: 0.00001 # rlim (limits change in species cs per step)
Tolerance to global time: 0.01 # ttol, chem solver is advanced in time until time diff < ttol*dt_global
Temperature threshold: 500 # Tth (below which, reactions ignored)
Reaction solver MIN steps: 5 # nstepmin, minimum number of time steps
Reaction solver MAX steps: 100 # nstepmax, maximum number of time steps
Two Temperature coefficient: 0.5 # qta (Tvib**qta*T**(1-qta))
Exclude vib energy: True # ivib0 = 1 if True
Exclude vib source: True # ivibS0 = 1 if True
Tvib BC Ratio: 1.0 # at any BC with T set, Tvib = tvibBC * T
Vibrational Temperature IC: -1 # set negative value to force Tvib = T, otherwise positive value set as IC value
#}

#LINEAR SOLVER
#
Solver Type: GMRES sparse
Number of GMRES Sweeps per Solve: 1 # replaces nGMRES
Minimum Number of Iterations per Nonlinear Iteration: 10 # minIters
Number of Krylov Vectors per GMRES Sweep: 100 # replaces Kspace
Tolerance on Momentum Equations: 0.01 # epstol(1), affects etol for Hessenberg problem

#}

#DISCRETIZATION CONTROL
#{
Weak Form: SUPG # alternate is Galerkin only for compressible
Time Integration Rule: First Order # 1st Order sets rinf(1) -1
Tau Matrix: Matrix-Ent-Adv
Include Viscous Correction in Stabilization: False # if p=1 idiff=1
# if p=2 idiff=2
Tau Time Constant: 1.0
Tau C Scale Factor: 1.0 # taucfct best value depends
Number of Elements Per Block: 64 #ibksiz

#}

#DISCONTINUITY CAPTURING
#{
Discontinuity Capturing: DC-quadratic # Current Options: DC-mallet, DC-minimum, DC-quadratic, DC-yzbeta
Multiplier for DC factor: 1 # scales DC variable in e3DC
Discontinuity Capturing Scheme: 1 # 0: discontinuous, 1: continuous (L2 projection)
Include Source Term in DC: 0 # 1: sets idcSRC to 1
Write DCqpt: 0 # if > 0, writes out data at a quadrature point iDCqpt

#----Parameters for YZBeta DC operator ----
Beta Value: 1 # 1: smoother , 2: sharper, 12: compromise between 1 and 2
YZB Farfield Conditions: 1e5 2119 10 10 300 # [Pressure, X-Vel, Y-Vel, Z-Vel, Temperature]
YZB Farfield Mole Frac: 1 0 0 0 0 # mole fractions at reference condition
# [xN2,xO2,xNO,xN,xO]
# must sum to 1, must be length 5
Include Umod Term: 1 # 0: no, 1: yes
Mach Adjustment Bm Value: 1 # 0: off,1: smoother shock, 2: sharper shock
Mach Adjustment Bj Value: 6 # 0: off,1: smoother shock, 2: sharper shock
Include Time Term in Z: 1 # 0: no, 1: yes
#------------------------------------------
#}

#STEP SEQUENCE
#{
Step Construction : 0 1 0 1
#}</nowiki>

== Post-Processing ==

An example of the ''flow.pht'' file is provided to demonstrate the ordering of the variables that can be viewed in Paraview. Note that both the ''errors'' and ''DCqpt'' fields would need to be output to the restart file for the example below to work. Saving these two fields is controlled through the solver.inp options.

=== Example of flow.pht file ===

<nowiki><?xml version="1.0" ?>
<PhastaMetaFile number_of_pieces="24">
<GeometryFileNamePattern pattern="24-procs_case/geombc.dat.%d"
has_piece_entry="1"
has_time_entry="0"/>
<FieldFileNamePattern pattern="24-procs_case/restart.%d.%d"
has_piece_entry="1"
has_time_entry="1"/>
<TimeSteps number_of_steps="1"
auto_generate_indices="1"
start_index="2010"
increment_index_by="50"
start_value="0"
increment_value_by="1">
</TimeSteps>
<Fields number_of_fields="9">
<Field phasta_field_tag="solution"
paraview_field_tag="rho_N2"
start_index_in_phasta_array="0"
number_of_components="1"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="solution"
paraview_field_tag="rho_O2"
start_index_in_phasta_array="1"
number_of_components="1"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="solution"
paraview_field_tag="velocity"
start_index_in_phasta_array="2"
number_of_components="3"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="solution"
paraview_field_tag="temp_vib"
start_index_in_phasta_array="5"
number_of_components="1"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="solution"
paraview_field_tag="temperature"
start_index_in_phasta_array="6"
number_of_components="1"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="errors"
paraview_field_tag="nu"
start_index_in_phasta_array="9"
number_of_components="1"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="DCqpt"
paraview_field_tag="DCqpt1"
start_index_in_phasta_array="0"
number_of_components="1"
data_dependency="1"
data_type="double"/>
</Fields>
</PhastaMetaFile></nowiki>

TCNEQ Version

2022-03-18T01:04:29Z

Jopo9795: /* Simulation Inputs */

== Background ==
The following information relates to the use of the thermochemical nonequilibrium (TCNEQ) version of PHASTA written in terms of entropy variables. The reader is referred to the following for additional information.

* F. Chalot, T.J.R. Hughes, and F. Shakib, '''"Symmetrization of Conservation Laws with Entropy for High-Temperature Hypersonic Computations,"''' Computing Systems in Engineering, 1(2-4):495–521, 1990.

* J. Pointer, '''"Influence of Interpolation Variables and Discontinuity Capturing Operators on Inviscid Hypersonic Flow Simulations Using a Stabilized Continuous Galerkin Solver,"''' Ph.D. dissertation, University of Colorado, Boulder, CO, 2022.

== Pre-Processing ==
In this section, details of the meshing and model attributes are provided. For simulation cases where the gas is modeled as a single species, the ''scalar'' quantities for boundary and initial conditions are not required. Currently, capability exists to simulate a gas with number of species (''nsp'') ≤ 5. When 2 ≤ ''nsp'' ≤ 5, the ''scalar'' quantities are used to govern the composition of the gas.

=== Meshing ===
Within the Simmodeler utility, the mesh can either be created or loaded from an existing .cas file. Below are steps for loading a mesh from a .cas file:
# Launch Simmodeler (for this example, SimModeler7.0-190604 is used)
# File > Import Discrete Data > (select .cas file to import) > (keep defaults and click OK) > (select YES to keep volume mesh)
# Save .sms and .smd files
# Attributes can now be assigned to the model as normal

=== Boundary Conditions ===
Below are the recognized boundary conditions that can be applied for the current version:
* comp1/comp2/comp3 - Specification of one/two/three components of velocity, [m/s]
* temperature - Specification of translational-rotational temperature, [K]. By default, vibrational temperature is held in equilibrium with this value and nonequilibrium is controlled through simulation inputs.
* surfID - When value is set to 702, the boundary is treated as a slip wall. If using this option, include a boundary layer mesh along the surface to ensure the wall normal direction is accurately computed.
* scalar_1 - Mole fraction of species 2 of the gas
* scalar_2 - Mole fraction of species 3 of the gas
* scalar_3 - Mole fraction of species 4 of the gas
* scalar_4 - Mole fraction of species 5 of the gas
* pressure - Specification of static pressure over a surface, [Pa]
** Used to compute mole fraction of species 1 of the gas with Dalton's Law of partial pressures and subtracting the summation of the other mole fractions from a value of 1
* heat flux - set to zero for adiabatic wall boundary condition

=== Initial Conditions ===
Below are the required initial conditions for the current version:
* initial velocity - Components and magnitude of flow velocity, [m/s]
** If a supersonic outlet condition is used, set such that flow is initialized Mach > 1
* initial temperature - Value used to set translational-rotational temperature, [K]
* initial scalar_1 - Initial value of species 2 mole fraction
* initial scalar_2 - Initial value of species 3 mole fraction
* initial scalar_3 - Initial value of species 4 mole fraction
* initial scalar_4 - Initial value of species 5 mole fraction
* initial pressure - Static pressure of the gas, [Pa]
** For multi-species flows, this value is used in combination with the initial scalar values to compute the mole fraction of species 1

== Simulation Inputs ==

Below is an example of the input script for the current version of the code. Capability is included for handling multi-species flows up to number of species ''nsp'' equal to 5. This feature has been tested and shown to work for five species mass conservation equations at one time. However, when chemical species production governed by the finite-rate chemistry module is active, the solution becomes unstable. Additional work on this feature is needed.

=== Example of solver.inp file: ===

<nowiki># PHASTA_HYP Version 1 Input File

#SOLUTION CONTROL
#{
Equation of State: Compressible
Number of Timesteps: 100
Time Step Size: 3.2339656949004e-08

Limit Density: 0 0.01 0.1 # solution limiting on variables [switch, min, max]
Limit u1: 0 0. 2.8e3
Limit u2: 0 0 0
Limit u3: 0 0 0
Limit Temperature: 0 230 3500 # also limits vibrational temperature
#}

#OUTPUT CONTROL
#{
Number of Timesteps between Restarts: 250
Print Error Indicators: True # shock error stored in column 6, DC factor \nu stored in column 10
Error Indicator Threshold: 0.01 # err > thresh*err_max is flagged as 1 (i.e. identified for refinement)
# --> smaller values = narrower flagged region along shock
Number of Error Smoothing Iterations: 0 # ierrsmooth
Load and set 3D IC: False # load the flowfield from a file as the initial condition
Position Tolerance on IC Load: 1e-7 # sets the tolerance for matching node locations while loading the initial condition
#}

#MATERIAL CONTROL
#{
Viscosity: 1.95508431704028e-5 # dynamic viscosity (only used if nsp.eq.99 for air mixture)
Thermal Conductivity: 26.6843390135759e-3 # only used if nsp.eq.99 for air mixture
Viscous Control: None #Viscous #None

#}

#REACTING FLOW
#{
Number of species: 1 # nsp
#.......currently only allowing 1<=nsp<=5.
#.........and make sure scalars passed from simmodeler are in correct slots
#
Species IDs: 1 # specIDs, length of array must equal
# nsp (IDs numbered in order:
# N2,O2,NO,N,O
# ID:99 for Air molecule
# e.g. Species IDs: 1 3 <<< would give N2 and NO
Ref Entropy Conditions: 1e3 230 230 0 #[P0,T0,T0vib,S0]
Ref Entropy Mole Frac: 1 0 0 0 0 # composition of gas used as reference entropy condition
# >> must sum to zero
Allow reactions: False # chemical reactions, ichem = 1 if True
Chemical heat release: False # chemical heat release, iqtot = 1 if True
Limit on reaction step: 0.00001 # rlim (limits change in species cs per step)
Tolerance to global time: 0.01 # ttol, chem solver is advanced in time until time diff < ttol*dt_global
Temperature threshold: 500 # Tth (below which, reactions ignored)
Reaction solver MIN steps: 5 # nstepmin, minimum number of time steps
Reaction solver MAX steps: 100 # nstepmax, maximum number of time steps
Two Temperature coefficient: 0.5 # qta (Tvib**qta*T**(1-qta))
Exclude vib energy: True # ivib0 = 1 if True
Exclude vib source: True # ivibS0 = 1 if True
Tvib BC Ratio: 1.0 # at any BC with T set, Tvib = tvibBC * T
Vibrational Temperature IC: -1 # set negative value to force Tvib = T, otherwise positive value set as IC value
#}

#LINEAR SOLVER
#
Solver Type: GMRES sparse
Number of GMRES Sweeps per Solve: 1 # replaces nGMRES
Minimum Number of Iterations per Nonlinear Iteration: 10 # minIters
Number of Krylov Vectors per GMRES Sweep: 100 # replaces Kspace
Tolerance on Momentum Equations: 0.01 # epstol(1), affects etol for Hessenberg problem

#}

#DISCRETIZATION CONTROL
#{
Weak Form: SUPG # alternate is Galerkin only for compressible
Time Integration Rule: First Order # 1st Order sets rinf(1) -1
Tau Matrix: Matrix-Ent-Adv
Include Viscous Correction in Stabilization: False # if p=1 idiff=1
# if p=2 idiff=2
Tau Time Constant: 1.0
Tau C Scale Factor: 1.0 # taucfct best value depends
Number of Elements Per Block: 64 #ibksiz

#}

#DISCONTINUITY CAPTURING
#{
Discontinuity Capturing: DC-quadratic # Current Options: DC-mallet, DC-minimum, DC-quadratic, DC-yzbeta
Multiplier for DC factor: 1 # scales DC variable in e3DC
Discontinuity Capturing Scheme: 1 # 0: discontinuous, 1: continuous (L2 projection)
Include Source Term in DC: 0 # 1: sets idcSRC to 1
Write DCqpt: 0 # if > 0, writes out data at a quadrature point iDCqpt

#----Parameters for YZBeta DC operator ----
Beta Value: 1 # 1: smoother , 2: sharper, 12: compromise between 1 and 2
YZB Farfield Conditions: 1e5 2119 10 10 300 # [Pressure, X-Vel, Y-Vel, Z-Vel, Temperature]
YZB Farfield Mole Frac: 1 0 0 0 0 # mole fractions at reference condition
# [xN2,xO2,xNO,xN,xO]
# must sum to 1, must be length 5
Include Umod Term: 1 # 0: no, 1: yes
Mach Adjustment Bm Value: 1 # 0: off,1: smoother shock, 2: sharper shock
Mach Adjustment Bj Value: 6 # 0: off,1: smoother shock, 2: sharper shock
Include Time Term in Z: 1 # 0: no, 1: yes
#------------------------------------------
#}

#STEP SEQUENCE
#{
Step Construction : 0 1 0 1
#}</nowiki>

== Post-Processing ==

<nowiki><?xml version="1.0" ?>
<PhastaMetaFile number_of_pieces="24">
<GeometryFileNamePattern pattern="24-procs_case/geombc.dat.%d"
has_piece_entry="1"
has_time_entry="0"/>
<FieldFileNamePattern pattern="24-procs_case/restart.%d.%d"
has_piece_entry="1"
has_time_entry="1"/>
<TimeSteps number_of_steps="1"
auto_generate_indices="1"
start_index="2010"
increment_index_by="50"
start_value="0"
increment_value_by="1">
</TimeSteps>
<Fields number_of_fields="7">
<Field phasta_field_tag="solution"
paraview_field_tag="rho"
start_index_in_phasta_array="0"
number_of_components="1"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="solution"
paraview_field_tag="velocity"
start_index_in_phasta_array="1"
number_of_components="3"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="solution"
paraview_field_tag="temp_vib"
start_index_in_phasta_array="4"
number_of_components="1"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="solution"
paraview_field_tag="temperature"
start_index_in_phasta_array="5"
number_of_components="1"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="errors"
paraview_field_tag="nu"
start_index_in_phasta_array="9"
number_of_components="1"
data_dependency="0"
data_type="double"/>
</Fields>
</PhastaMetaFile></nowiki>

TCNEQ Version

2022-03-18T00:54:57Z

Jopo9795: /* Simulation Inputs */

== Background ==
The following information relates to the use of the thermochemical nonequilibrium (TCNEQ) version of PHASTA written in terms of entropy variables. The reader is referred to the following for additional information.

* F. Chalot, T.J.R. Hughes, and F. Shakib, '''"Symmetrization of Conservation Laws with Entropy for High-Temperature Hypersonic Computations,"''' Computing Systems in Engineering, 1(2-4):495–521, 1990.

* J. Pointer, '''"Influence of Interpolation Variables and Discontinuity Capturing Operators on Inviscid Hypersonic Flow Simulations Using a Stabilized Continuous Galerkin Solver,"''' Ph.D. dissertation, University of Colorado, Boulder, CO, 2022.

== Pre-Processing ==
In this section, details of the meshing and model attributes are provided. For simulation cases where the gas is modeled as a single species, the ''scalar'' quantities for boundary and initial conditions are not required. Currently, capability exists to simulate a gas with number of species (''nsp'') ≤ 5. When 2 ≤ ''nsp'' ≤ 5, the ''scalar'' quantities are used to govern the composition of the gas.

=== Meshing ===
Within the Simmodeler utility, the mesh can either be created or loaded from an existing .cas file. Below are steps for loading a mesh from a .cas file:
# Launch Simmodeler (for this example, SimModeler7.0-190604 is used)
# File > Import Discrete Data > (select .cas file to import) > (keep defaults and click OK) > (select YES to keep volume mesh)
# Save .sms and .smd files
# Attributes can now be assigned to the model as normal

=== Boundary Conditions ===
Below are the recognized boundary conditions that can be applied for the current version:
* comp1/comp2/comp3 - Specification of one/two/three components of velocity, [m/s]
* temperature - Specification of translational-rotational temperature, [K]. By default, vibrational temperature is held in equilibrium with this value and nonequilibrium is controlled through simulation inputs.
* surfID - When value is set to 702, the boundary is treated as a slip wall. If using this option, include a boundary layer mesh along the surface to ensure the wall normal direction is accurately computed.
* scalar_1 - Mole fraction of species 2 of the gas
* scalar_2 - Mole fraction of species 3 of the gas
* scalar_3 - Mole fraction of species 4 of the gas
* scalar_4 - Mole fraction of species 5 of the gas
* pressure - Specification of static pressure over a surface, [Pa]
** Used to compute mole fraction of species 1 of the gas with Dalton's Law of partial pressures and subtracting the summation of the other mole fractions from a value of 1
* heat flux - set to zero for adiabatic wall boundary condition

=== Initial Conditions ===
Below are the required initial conditions for the current version:
* initial velocity - Components and magnitude of flow velocity, [m/s]
** If a supersonic outlet condition is used, set such that flow is initialized Mach > 1
* initial temperature - Value used to set translational-rotational temperature, [K]
* initial scalar_1 - Initial value of species 2 mole fraction
* initial scalar_2 - Initial value of species 3 mole fraction
* initial scalar_3 - Initial value of species 4 mole fraction
* initial scalar_4 - Initial value of species 5 mole fraction
* initial pressure - Static pressure of the gas, [Pa]
** For multi-species flows, this value is used in combination with the initial scalar values to compute the mole fraction of species 1

== Simulation Inputs ==

Example of solver.inp file:

<nowiki># PHASTA_HYP Version 1 Input File

#SOLUTION CONTROL
#{
Equation of State: Compressible
Number of Timesteps: 100
Time Step Size: 3.2339656949004e-08

Limit Density: 0 0.01 0.1 # solution limiting on variables [switch, min, max]
Limit u1: 0 0. 2.8e3
Limit u2: 0 0 0
Limit u3: 0 0 0
Limit Temperature: 0 230 3500 # also limits vibrational temperature
#}

#OUTPUT CONTROL
#{
Number of Timesteps between Restarts: 250
Print Error Indicators: True # shock error stored in column 6, DC factor \nu stored in column 10
Error Indicator Threshold: 0.01 # err > thresh*err_max is flagged as 1 (i.e. identified for refinement)
# --> smaller values = narrower flagged region along shock
Number of Error Smoothing Iterations: 0 # ierrsmooth
Load and set 3D IC: False # load the flowfield from a file as the initial condition
Position Tolerance on IC Load: 1e-7 # sets the tolerance for matching node locations while loading the initial condition
#}

#MATERIAL CONTROL
#{
Viscosity: 1.95508431704028e-5 # dynamic viscosity (only used if nsp.eq.99 for air mixture)
Thermal Conductivity: 26.6843390135759e-3 # only used if nsp.eq.99 for air mixture
Viscous Control: None #Viscous #None

#}

#REACTING FLOW
#{
Number of species: 1 # nsp
#.......currently only allowing 1<=nsp<=5.
#.........and make sure scalars passed from simmodeler are in correct slots
#
Species IDs: 1 # specIDs, length of array must equal
# nsp (IDs numbered in order:
# N2,O2,NO,N,O
# ID:99 for Air molecule
# e.g. Species IDs: 1 3 <<< would give N2 and NO
Ref Entropy Conditions: 1e3 230 230 0 #[P0,T0,T0vib,S0]
Ref Entropy Mole Frac: 1 0 0 0 0 # composition of gas used as reference entropy condition
# >> must sum to zero
Allow reactions: False # chemical reactions, ichem = 1 if True
Chemical heat release: False # chemical heat release, iqtot = 1 if True
Limit on reaction step: 0.00001 # rlim (limits change in species cs per step)
Tolerance to global time: 0.01 # ttol, chem solver is advanced in time until time diff < ttol*dt_global
Temperature threshold: 500 # Tth (below which, reactions ignored)
Reaction solver MIN steps: 5 # nstepmin, minimum number of time steps
Reaction solver MAX steps: 100 # nstepmax, maximum number of time steps
Two Temperature coefficient: 0.5 # qta (Tvib**qta*T**(1-qta))
Exclude vib energy: True # ivib0 = 1 if True
Exclude vib source: True # ivibS0 = 1 if True
Tvib BC Ratio: 1.0 # at any BC with T set, Tvib = tvibBC * T
Vibrational Temperature IC: -1 # set negative value to force Tvib = T, otherwise positive value set as IC value
#}

#LINEAR SOLVER
#
Solver Type: GMRES sparse
Number of GMRES Sweeps per Solve: 1 # replaces nGMRES
Minimum Number of Iterations per Nonlinear Iteration: 10 # minIters
Number of Krylov Vectors per GMRES Sweep: 100 # replaces Kspace
Tolerance on Momentum Equations: 0.01 # epstol(1), affects etol for Hessenberg problem

#}

#DISCRETIZATION CONTROL
#{
Weak Form: SUPG # alternate is Galerkin only for compressible
Time Integration Rule: First Order # 1st Order sets rinf(1) -1
Tau Matrix: Matrix-Ent-Adv
Include Viscous Correction in Stabilization: False # if p=1 idiff=1
# if p=2 idiff=2
Tau Time Constant: 1.0
Tau C Scale Factor: 1.0 # taucfct best value depends
Number of Elements Per Block: 64 #ibksiz

#}

#DISCONTINUITY CAPTURING
#{
Discontinuity Capturing: DC-quadratic # Current Options: DC-mallet, DC-minimum, DC-quadratic, DC-yzbeta
Multiplier for DC factor: 1 # scales DC variable in e3DC
Discontinuity Capturing Scheme: 1 # 0: discontinuous, 1: continuous (L2 projection)
Include Source Term in DC: 0 # 1: sets idcSRC to 1
Write DCqpt: 0 # if > 0, writes out data at a quadrature point iDCqpt

#----Parameters for YZBeta DC operator ----
Beta Value: 1 # 1: smoother , 2: sharper, 12: compromise between 1 and 2
YZB Farfield Conditions: 1e5 2119 10 10 300 # [Pressure, X-Vel, Y-Vel, Z-Vel, Temperature]
YZB Farfield Mole Frac: 1 0 0 0 0 # mole fractions at reference condition
# [xN2,xO2,xNO,xN,xO]
# must sum to 1, must be length 5
Include Umod Term: 1 # 0: no, 1: yes
Mach Adjustment Bm Value: 1 # 0: off,1: smoother shock, 2: sharper shock
Mach Adjustment Bj Value: 6 # 0: off,1: smoother shock, 2: sharper shock
Include Time Term in Z: 1 # 0: no, 1: yes
#------------------------------------------
#}

#STEP SEQUENCE
#{
Step Construction : 0 1 0 1
#}</nowiki>

== Post-Processing ==

<nowiki><?xml version="1.0" ?>
<PhastaMetaFile number_of_pieces="24">
<GeometryFileNamePattern pattern="24-procs_case/geombc.dat.%d"
has_piece_entry="1"
has_time_entry="0"/>
<FieldFileNamePattern pattern="24-procs_case/restart.%d.%d"
has_piece_entry="1"
has_time_entry="1"/>
<TimeSteps number_of_steps="1"
auto_generate_indices="1"
start_index="2010"
increment_index_by="50"
start_value="0"
increment_value_by="1">
</TimeSteps>
<Fields number_of_fields="7">
<Field phasta_field_tag="solution"
paraview_field_tag="rho"
start_index_in_phasta_array="0"
number_of_components="1"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="solution"
paraview_field_tag="velocity"
start_index_in_phasta_array="1"
number_of_components="3"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="solution"
paraview_field_tag="temp_vib"
start_index_in_phasta_array="4"
number_of_components="1"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="solution"
paraview_field_tag="temperature"
start_index_in_phasta_array="5"
number_of_components="1"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="errors"
paraview_field_tag="nu"
start_index_in_phasta_array="9"
number_of_components="1"
data_dependency="0"
data_type="double"/>
</Fields>
</PhastaMetaFile></nowiki>

TCNEQ Version

2022-03-18T00:50:49Z

Jopo9795: /* Simulation Inputs */

== Background ==
The following information relates to the use of the thermochemical nonequilibrium (TCNEQ) version of PHASTA written in terms of entropy variables. The reader is referred to the following for additional information.

* F. Chalot, T.J.R. Hughes, and F. Shakib, '''"Symmetrization of Conservation Laws with Entropy for High-Temperature Hypersonic Computations,"''' Computing Systems in Engineering, 1(2-4):495–521, 1990.

* J. Pointer, '''"Influence of Interpolation Variables and Discontinuity Capturing Operators on Inviscid Hypersonic Flow Simulations Using a Stabilized Continuous Galerkin Solver,"''' Ph.D. dissertation, University of Colorado, Boulder, CO, 2022.

== Pre-Processing ==
In this section, details of the meshing and model attributes are provided. For simulation cases where the gas is modeled as a single species, the ''scalar'' quantities for boundary and initial conditions are not required. Currently, capability exists to simulate a gas with number of species (''nsp'') ≤ 5. When 2 ≤ ''nsp'' ≤ 5, the ''scalar'' quantities are used to govern the composition of the gas.

=== Meshing ===
Within the Simmodeler utility, the mesh can either be created or loaded from an existing .cas file. Below are steps for loading a mesh from a .cas file:
# Launch Simmodeler (for this example, SimModeler7.0-190604 is used)
# File > Import Discrete Data > (select .cas file to import) > (keep defaults and click OK) > (select YES to keep volume mesh)
# Save .sms and .smd files
# Attributes can now be assigned to the model as normal

=== Boundary Conditions ===
Below are the recognized boundary conditions that can be applied for the current version:
* comp1/comp2/comp3 - Specification of one/two/three components of velocity, [m/s]
* temperature - Specification of translational-rotational temperature, [K]. By default, vibrational temperature is held in equilibrium with this value and nonequilibrium is controlled through simulation inputs.
* surfID - When value is set to 702, the boundary is treated as a slip wall. If using this option, include a boundary layer mesh along the surface to ensure the wall normal direction is accurately computed.
* scalar_1 - Mole fraction of species 2 of the gas
* scalar_2 - Mole fraction of species 3 of the gas
* scalar_3 - Mole fraction of species 4 of the gas
* scalar_4 - Mole fraction of species 5 of the gas
* pressure - Specification of static pressure over a surface, [Pa]
** Used to compute mole fraction of species 1 of the gas with Dalton's Law of partial pressures and subtracting the summation of the other mole fractions from a value of 1
* heat flux - set to zero for adiabatic wall boundary condition

=== Initial Conditions ===
Below are the required initial conditions for the current version:
* initial velocity - Components and magnitude of flow velocity, [m/s]
** If a supersonic outlet condition is used, set such that flow is initialized Mach > 1
* initial temperature - Value used to set translational-rotational temperature, [K]
* initial scalar_1 - Initial value of species 2 mole fraction
* initial scalar_2 - Initial value of species 3 mole fraction
* initial scalar_3 - Initial value of species 4 mole fraction
* initial scalar_4 - Initial value of species 5 mole fraction
* initial pressure - Static pressure of the gas, [Pa]
** For multi-species flows, this value is used in combination with the initial scalar values to compute the mole fraction of species 1

== Simulation Inputs ==

Example of solver.inp file:

<nowiki># PHASTA_HYP Version 1 Input File

#SOLUTION CONTROL
#{
Equation of State: Compressible
Number of Timesteps: 100
Time Step Size: 3.2339656949004e-08

Limit Density: 0 0.01 0.1 # solution limiting on variables [switch, min, max]
Limit u1: 0 0. 2.8e3
Limit u2: 0 0 0
Limit u3: 0 0 0
Limit Temperature: 0 230 3500 # also limits vibrational temperature
#}

#OUTPUT CONTROL
#{
Number of Timesteps between Restarts: 250
Print Error Indicators: True # shock error stored in column 6, DC factor \nu stored in column 10
Error Indicator Threshold: 0.01 # err > thresh*err_max is flagged as 1 (i.e. identified for refinement)
# --> smaller values = narrower flagged region along shock
Number of Error Smoothing Iterations: 0 # ierrsmooth
Load and set 3D IC: False # load the flowfield from a file as the initial condition
Position Tolerance on IC Load: 1e-7 # sets the tolerance for matching node locations while loading the initial condition
#}

#MATERIAL CONTROL
#{
Viscosity: 1.95508431704028e-5 # dynamic viscosity (only used if nsp.eq.99 for air mixture)
Thermal Conductivity: 26.6843390135759e-3 # only used if nsp.eq.99 for air mixture
Viscous Control: None #Viscous #None

#}

#REACTING FLOW
#{
Number of species: 1 # nsp
#.......currently only allowing 1<=nsp<=5.
#.........and make sure scalars passed from simmodeler are in correct slots
#
Species IDs: 1 # specIDs, length of array must equal
# nsp (IDs numbered in order:
# N2,O2,NO,N,O
# ID:99 for Air molecule
# e.g. Species IDs: 1 3 <<< would give N2 and NO
Ref Entropy Conditions: 1e3 230 230 0 #[P0,T0,T0vib,S0]
Ref Entropy Mole Frac: 1 0 0 0 0
Allow reactions: False # chemical reactions, ichem = 1 if True
Chemical heat release: False # chemical heat release, iqtot = 1 if True
Limit on reaction step: 0.00001 # rlim (limits change in species cs per step)
Tolerance to global time: 0.01 # ttol, chem solver is advanced in time until time diff < ttol*dt_global
Temperature threshold: 500 # Tth (below which, reactions ignored)
Reaction solver MIN steps: 5 # nstepmin, minimum number of time steps
Reaction solver MAX steps: 100 # nstepmax, maximum number of time steps
Two Temperature coefficient: 0.5 # qta (Tvib**qta*T**(1-qta))
Exclude vib energy: True # ivib0 = 1 if True
Exclude vib source: True # ivibS0 = 1 if True
Tvib BC Ratio: 1.0 # at any BC with T set, Tvib = tvibBC * T
Vibrational Temperature IC: -1 # set negative value to force Tvib = T, otherwise positive value set as IC value
#}

#LINEAR SOLVER
#
Solver Type: GMRES sparse
Number of GMRES Sweeps per Solve: 1 # replaces nGMRES
Minimum Number of Iterations per Nonlinear Iteration: 10 # minIters
Number of Krylov Vectors per GMRES Sweep: 100 # replaces Kspace
Tolerance on Momentum Equations: 0.01 # epstol(1), affects etol for Hessenberg problem

#}

#DISCRETIZATION CONTROL
#{
Weak Form: SUPG # alternate is Galerkin only for compressible
Time Integration Rule: First Order # 1st Order sets rinf(1) -1
Tau Matrix: Matrix-Ent-Adv
Include Viscous Correction in Stabilization: False # if p=1 idiff=1
# if p=2 idiff=2
Tau Time Constant: 1.0
Tau C Scale Factor: 1.0 # taucfct best value depends
Number of Elements Per Block: 64 #ibksiz

#}

#DISCONTINUITY CAPTURING
#{
Discontinuity Capturing: DC-quadratic # Current Options: DC-mallet, DC-minimum, DC-quadratic, DC-yzbeta
Multiplier for DC factor: 1 # scales DC variable in e3DC
Discontinuity Capturing Scheme: 1 # 0: discontinuous, 1: continuous (L2 projection)
Include Source Term in DC: 0 # 1: sets idcSRC to 1
Write DCqpt: 0 # if > 0, writes out data at a quadrature point iDCqpt

#----Parameters for YZBeta DC operator ----
Beta Value: 1 # 1: smoother , 2: sharper, 12: compromise between 1 and 2
YZB Farfield Conditions: 1e5 2119 10 10 300 # [Pressure, X-Vel, Y-Vel, Z-Vel, Temperature]
YZB Farfield Mole Frac: 1 0 0 0 0 # mole fractions at reference condition
# [xN2,xO2,xNO,xN,xO]
# must sum to 1, must be length 5
Include Umod Term: 1 # 0: no, 1: yes
Mach Adjustment Bm Value: 1 # 0: off,1: smoother shock, 2: sharper shock
Mach Adjustment Bj Value: 6 # 0: off,1: smoother shock, 2: sharper shock
Include Time Term in Z: 1 # 0: no, 1: yes
#------------------------------------------
#}

#STEP SEQUENCE
#{
Step Construction : 0 1 0 1
#}</nowiki>

== Post-Processing ==

<nowiki><?xml version="1.0" ?>
<PhastaMetaFile number_of_pieces="24">
<GeometryFileNamePattern pattern="24-procs_case/geombc.dat.%d"
has_piece_entry="1"
has_time_entry="0"/>
<FieldFileNamePattern pattern="24-procs_case/restart.%d.%d"
has_piece_entry="1"
has_time_entry="1"/>
<TimeSteps number_of_steps="1"
auto_generate_indices="1"
start_index="2010"
increment_index_by="50"
start_value="0"
increment_value_by="1">
</TimeSteps>
<Fields number_of_fields="7">
<Field phasta_field_tag="solution"
paraview_field_tag="rho"
start_index_in_phasta_array="0"
number_of_components="1"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="solution"
paraview_field_tag="velocity"
start_index_in_phasta_array="1"
number_of_components="3"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="solution"
paraview_field_tag="temp_vib"
start_index_in_phasta_array="4"
number_of_components="1"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="solution"
paraview_field_tag="temperature"
start_index_in_phasta_array="5"
number_of_components="1"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="errors"
paraview_field_tag="nu"
start_index_in_phasta_array="9"
number_of_components="1"
data_dependency="0"
data_type="double"/>
</Fields>
</PhastaMetaFile></nowiki>

TCNEQ Version

2022-03-17T23:00:51Z

Jopo9795: /* Simulation Inputs */

== Background ==
The following information relates to the use of the thermochemical nonequilibrium (TCNEQ) version of PHASTA written in terms of entropy variables. The reader is referred to the following for additional information.

* F. Chalot, T.J.R. Hughes, and F. Shakib, '''"Symmetrization of Conservation Laws with Entropy for High-Temperature Hypersonic Computations,"''' Computing Systems in Engineering, 1(2-4):495–521, 1990.

* J. Pointer, '''"Influence of Interpolation Variables and Discontinuity Capturing Operators on Inviscid Hypersonic Flow Simulations Using a Stabilized Continuous Galerkin Solver,"''' Ph.D. dissertation, University of Colorado, Boulder, CO, 2022.

== Pre-Processing ==
In this section, details of the meshing and model attributes are provided. For simulation cases where the gas is modeled as a single species, the ''scalar'' quantities for boundary and initial conditions are not required. Currently, capability exists to simulate a gas with number of species (''nsp'') ≤ 5. When 2 ≤ ''nsp'' ≤ 5, the ''scalar'' quantities are used to govern the composition of the gas.

=== Meshing ===
Within the Simmodeler utility, the mesh can either be created or loaded from an existing .cas file. Below are steps for loading a mesh from a .cas file:
# Launch Simmodeler (for this example, SimModeler7.0-190604 is used)
# File > Import Discrete Data > (select .cas file to import) > (keep defaults and click OK) > (select YES to keep volume mesh)
# Save .sms and .smd files
# Attributes can now be assigned to the model as normal

=== Boundary Conditions ===
Below are the recognized boundary conditions that can be applied for the current version:
* comp1/comp2/comp3 - Specification of one/two/three components of velocity, [m/s]
* temperature - Specification of translational-rotational temperature, [K]. By default, vibrational temperature is held in equilibrium with this value and nonequilibrium is controlled through simulation inputs.
* surfID - When value is set to 702, the boundary is treated as a slip wall. If using this option, include a boundary layer mesh along the surface to ensure the wall normal direction is accurately computed.
* scalar_1 - Mole fraction of species 2 of the gas
* scalar_2 - Mole fraction of species 3 of the gas
* scalar_3 - Mole fraction of species 4 of the gas
* scalar_4 - Mole fraction of species 5 of the gas
* pressure - Specification of static pressure over a surface, [Pa]
** Used to compute mole fraction of species 1 of the gas with Dalton's Law of partial pressures and subtracting the summation of the other mole fractions from a value of 1
* heat flux - set to zero for adiabatic wall boundary condition

=== Initial Conditions ===
Below are the required initial conditions for the current version:
* initial velocity - Components and magnitude of flow velocity, [m/s]
** If a supersonic outlet condition is used, set such that flow is initialized Mach > 1
* initial temperature - Value used to set translational-rotational temperature, [K]
* initial scalar_1 - Initial value of species 2 mole fraction
* initial scalar_2 - Initial value of species 3 mole fraction
* initial scalar_3 - Initial value of species 4 mole fraction
* initial scalar_4 - Initial value of species 5 mole fraction
* initial pressure - Static pressure of the gas, [Pa]
** For multi-species flows, this value is used in combination with the initial scalar values to compute the mole fraction of species 1

== Simulation Inputs ==

Example of solver.inp file:

<nowiki># PHASTA_HYP Version 1 Input File

#SOLUTION CONTROL
#{
Equation of State: Compressible
Number of Timesteps: 100
Time Step Size: 3.2339656949004e-08

Limit Density: 0 0.01 0.1 # solution limiting on variables [switch min max]
Limit u1: 0 0. 2.8e3
Limit u2: 0 0 0
Limit u3: 0 0 0
Limit Temperature: 0 230 3500 # also limits vibrational temperature
#}

#OUTPUT CONTROL
#{
Number of Timesteps between Restarts: 250
Print Error Indicators: True # shock error stored in column 6, DC factor \nu stored in column 10
Error Indicator Threshold: 0.01 # err > thresh*err_max is flagged as 1 (i.e. identified for refinement)
# --> smaller values = narrower flagged region along shock
Number of Error Smoothing Iterations: 0 # ierrsmooth
Load and set 3D IC: False # load the flowfield from a file as the initial condition
Position Tolerance on IC Load: 1e-7 # sets the tolerance for matching points for loading the initial condition
#}

#MATERIAL CONTROL
#{
Viscosity: 1.95508431704028e-5 #<<<<< this is required to run (value used if nsp.eq.99)
Thermal Conductivity: 26.6843390135759e-3
Viscous Control: None #Viscous #None

#}

#REACTING FLOW
#{
Number of species: 1 # nsp
#.......currently only allowing 1<=nsp<=5.
#.........and make sure scalars passed from simmodeler are in correct slots
#
Species IDs: 1 # specIDs, length of array must equal
# nsp (IDs numbered in order:
# N2,O2,NO,N,O
# ID:99 for Air molecule
Ref Entropy Conditions: 1e3 230 230 0 #[P0,T0,T0vib,S0]
Ref Entropy Species: 1 2 # uses IDs above (NOT ACTIVE YET)
Ref Entropy Mole Frac: 1 0 0 0 0
Allow reactions: False # ichem = 1 if True
Chemical heat release: False # iqtot = 1 if True
Limit on reaction step: 0.00001 # rlim (limits change in species cs per step)
Tolerance to global time: 1.0 #0.01 # ttol
Temperature threshold: 500 # Tth (below which, reactions ignored)
Reaction solver MIN steps: 5 # nstepmin
Reaction solver MAX steps: 100 # nstepmax
Two Temperature coefficient: 0.5 # qta (Tvib**qta*T**(1-qta))
Exclude vib energy: True # ivib0 = 1 if True
Exclude vib source: True # ivibS0 = 1 if True
Tvib BC Ratio: 1.0 # at any BC with T set, Tvib = tvibBC * T
Vibrational Temperature IC: -1 # set negative value to force Tvib = T
#}

#LINEAR SOLVER
#
Solver Type: GMRES sparse <<< default, but why does it have to be set?
Number of GMRES Sweeps per Solve: 1 # replaces nGMRES
Minimum Number of Iterations per Nonlinear Iteration: 10 # minIters
Number of Krylov Vectors per GMRES Sweep: 100 # replaces Kspace
Tolerance on Momentum Equations: 0.01 # epstol(1), affects etol for Hessenberg problem

#}

#DISCRETIZATION CONTROL
#{
Weak Form: SUPG # alternate is Galerkin only for compressible
Time Integration Rule: First Order # 1st Order sets rinf(1) -1
# Time Integration Rule: Second Order # Second Order sets rinf next
# Time Integration Rho Infinity: 0.5 # rinf(1) Only used for 2nd order
# Tau Matrix: Diagonal-Franca-Entropy #itau=4
Tau Matrix: Matrix-Ent-Adv
# Tau Matrix: Diagonal-Shakib-Entropy
Include Viscous Correction in Stabilization: False # if p=1 idiff=1
# if p=2 idiff=2
Tau Time Constant: 1.0
Tau C Scale Factor: 1.0 # taucfct best value depends
Number of Elements Per Block: 64 #ibksiz

#}

#DISCONTINUITY CAPTURING
#{
Discontinuity Capturing: DC-quadratic # Current Options: DC-mallet, DC-minimum, DC-quadratic, DC-yzbeta
Multiplier for DC factor: 1 # scales DC variable in e3DC
Discontinuity Capturing Scheme: 1 #0: discontinuous, 1: continuous (L2 projection)
Include Source Term in DC: 0 # 1: sets idcSRC to 1
Write DCqpt: 0
#----Parameters for YZBeta DC operator ----
Beta Value: 1 # 1: smoother , 2: sharper, 12: compromise between 1 and 2
YZB Farfield Conditions: 1e5 2119 10 10 300 # [Pressure, X-Vel, Y-Vel, Z-Vel, Temperature]
YZB Farfield Mole Frac: 1 0 0 0 0 # mole fractions at reference condition
# [xN2,xO2,xNO,xN,xO]
# must sum to 1, must be length 5
Include Umod Term: 1 # 0: no, 1: yes
Mach Adjustment Bm Value: 1 # 0: off,1: smoother shock, 2: sharper shock
Mach Adjustment Bj Value: 6 # 0: off,1: smoother shock, 2: sharper shock
Include Time Term in Z: 1 # 0: no, 1: yes
#------------------------------------------

#}

#STEP SEQUENCE
#{
Step Construction : 0 1 0 1
#}</nowiki>

== Post-Processing ==

<nowiki><?xml version="1.0" ?>
<PhastaMetaFile number_of_pieces="24">
<GeometryFileNamePattern pattern="24-procs_case/geombc.dat.%d"
has_piece_entry="1"
has_time_entry="0"/>
<FieldFileNamePattern pattern="24-procs_case/restart.%d.%d"
has_piece_entry="1"
has_time_entry="1"/>
<TimeSteps number_of_steps="1"
auto_generate_indices="1"
start_index="2010"
increment_index_by="50"
start_value="0"
increment_value_by="1">
</TimeSteps>
<Fields number_of_fields="7">
<Field phasta_field_tag="solution"
paraview_field_tag="rho"
start_index_in_phasta_array="0"
number_of_components="1"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="solution"
paraview_field_tag="velocity"
start_index_in_phasta_array="1"
number_of_components="3"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="solution"
paraview_field_tag="temp_vib"
start_index_in_phasta_array="4"
number_of_components="1"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="solution"
paraview_field_tag="temperature"
start_index_in_phasta_array="5"
number_of_components="1"
data_dependency="0"
data_type="double"/>
<Field phasta_field_tag="errors"
paraview_field_tag="nu"
start_index_in_phasta_array="9"
number_of_components="1"
data_dependency="0"
data_type="double"/>
</Fields>
</PhastaMetaFile></nowiki>

TCNEQ Version

2022-03-17T22:41:04Z

Jopo9795: /* Simulation Inputs */

== Background ==
The following information relates to the use of the thermochemical nonequilibrium (TCNEQ) version of PHASTA written in terms of entropy variables. The reader is referred to the following for additional information.

* F. Chalot, T.J.R. Hughes, and F. Shakib, '''"Symmetrization of Conservation Laws with Entropy for High-Temperature Hypersonic Computations,"''' Computing Systems in Engineering, 1(2-4):495–521, 1990.

* J. Pointer, '''"Influence of Interpolation Variables and Discontinuity Capturing Operators on Inviscid Hypersonic Flow Simulations Using a Stabilized Continuous Galerkin Solver,"''' Ph.D. dissertation, University of Colorado, Boulder, CO, 2022.

== Pre-Processing ==
In this section, details of the meshing and model attributes are provided. For simulation cases where the gas is modeled as a single species, the ''scalar'' quantities for boundary and initial conditions are not required. Currently, capability exists to simulate a gas with number of species (''nsp'') ≤ 5. When 2 ≤ ''nsp'' ≤ 5, the ''scalar'' quantities are used to govern the composition of the gas.

=== Meshing ===
Within the Simmodeler utility, the mesh can either be created or loaded from an existing .cas file. Below are steps for loading a mesh from a .cas file:
# Launch Simmodeler (for this example, SimModeler7.0-190604 is used)
# File > Import Discrete Data > (select .cas file to import) > (keep defaults and click OK) > (select YES to keep volume mesh)
# Save .sms and .smd files
# Attributes can now be assigned to the model as normal

=== Boundary Conditions ===
Below are the recognized boundary conditions that can be applied for the current version:
* comp1/comp2/comp3 - Specification of one/two/three components of velocity, [m/s]
* temperature - Specification of translational-rotational temperature, [K]. By default, vibrational temperature is held in equilibrium with this value and nonequilibrium is controlled through simulation inputs.
* surfID - When value is set to 702, the boundary is treated as a slip wall. If using this option, include a boundary layer mesh along the surface to ensure the wall normal direction is accurately computed.
* scalar_1 - Mole fraction of species 2 of the gas
* scalar_2 - Mole fraction of species 3 of the gas
* scalar_3 - Mole fraction of species 4 of the gas
* scalar_4 - Mole fraction of species 5 of the gas
* pressure - Specification of static pressure over a surface, [Pa]
** Used to compute mole fraction of species 1 of the gas with Dalton's Law of partial pressures and subtracting the summation of the other mole fractions from a value of 1
* heat flux - set to zero for adiabatic wall boundary condition

=== Initial Conditions ===
Below are the required initial conditions for the current version:
* initial velocity - Components and magnitude of flow velocity, [m/s]
** If a supersonic outlet condition is used, set such that flow is initialized Mach > 1
* initial temperature - Value used to set translational-rotational temperature, [K]
* initial scalar_1 - Initial value of species 2 mole fraction
* initial scalar_2 - Initial value of species 3 mole fraction
* initial scalar_3 - Initial value of species 4 mole fraction
* initial scalar_4 - Initial value of species 5 mole fraction
* initial pressure - Static pressure of the gas, [Pa]
** For multi-species flows, this value is used in combination with the initial scalar values to compute the mole fraction of species 1

== Simulation Inputs ==

Example of solver.inp file:

<nowiki># ibksiz flmpl flmpr itwmod wmodts dmodts fwr taucfct
# PHASTA Version 1.5 Input File
#
# Basic format is
#
# Key Phrase : Acceptable Value (integer, double, logical, or phrase
# list of integers, list of doubles )
#
#
# ****************** Notes ********************************
# 1. Bulk viscosity set to 0 in common.f (bmu)
# 2. Right now, nsp cannot be anything other than 5
# 3. idiff must be zero for now
# 4. iqtot does nothing for now since source terms are not being handled
# (commented out call to e3source in e3.f)
# 5.

#SOLUTION CONTROL
#{
Equation of State: Compressible
Number of Timesteps: 1
Time Step Size: 3.2339656949004e-08 # cfl = 1 is dt = 7.8702386e-8

Limit Density: 0 0.01 0.1
Limit u1: 0 0. 2.8e3
Limit u2: 0 0 0
Limit u3: 0 0 0
Limit Temperature: 0 230 3500 !also limits vibrational temperature
#}

#OUTPUT CONTROL
#{
Number of Timesteps between Restarts: 250 #replaces nout/ntout
Print Error Indicators: True
Error Indicator Threshold: 0.01
Number of Error Smoothing Iterations: 0 #ierrsmooth
Load and set 3D IC: False
Position Tolerance on IC Load: 1e-7 # sets the tolerance for matching points
#}

#MATERIAL CONTROL
#{
Viscosity: 1.95508431704028e-5 #<<<<< this is required to run (value used if nsp.eq.99)
Thermal Conductivity: 26.6843390135759e-3
Viscous Control: None #Viscous #None

#}

#REACTING FLOW
#{
Number of species: 1 # nsp
#.......currently only allowing 1<=nsp<=5.
#.........and make sure scalars passed from simmodeler are in correct slots
#
Species IDs: 1 # specIDs, length of array must equal
# nsp (IDs numbered in order:
# N2,O2,NO,N,O
# ID:99 for Air molecule
Ref Entropy Conditions: 1e3 230 230 0 #[P0,T0,T0vib,S0]
Ref Entropy Species: 1 2 # uses IDs above (NOT ACTIVE YET)
Ref Entropy Mole Frac: 1 0 0 0 0
Allow reactions: False # ichem = 1 if True
Chemical heat release: False # iqtot = 1 if True
Limit on reaction step: 0.00001 # rlim (limits change in species cs per step)
Tolerance to global time: 1.0 #0.01 # ttol
Temperature threshold: 500 # Tth (below which, reactions ignored)
Reaction solver MIN steps: 5 # nstepmin
Reaction solver MAX steps: 100 # nstepmax
Two Temperature coefficient: 0.5 # qta (Tvib**qta*T**(1-qta))
Exclude vib energy: True # ivib0 = 1 if True
Exclude vib source: True # ivibS0 = 1 if True
Tvib BC Ratio: 1.0 # at any BC with T set, Tvib = tvibBC * T
Vibrational Temperature IC: -1 # set negative value to force Tvib = T
#}

#LINEAR SOLVER
#
Solver Type: GMRES sparse <<< default, but why does it have to be set?
Number of GMRES Sweeps per Solve: 1 # replaces nGMRES
Minimum Number of Iterations per Nonlinear Iteration: 10 # minIters
Number of Krylov Vectors per GMRES Sweep: 100 # replaces Kspace
Tolerance on Momentum Equations: 0.01 # epstol(1), affects etol for Hessenberg problem

#}

#DISCRETIZATION CONTROL
#{
Weak Form: SUPG # alternate is Galerkin only for compressible
Time Integration Rule: First Order # 1st Order sets rinf(1) -1
# Time Integration Rule: Second Order # Second Order sets rinf next
# Time Integration Rho Infinity: 0.5 # rinf(1) Only used for 2nd order
# Tau Matrix: Diagonal-Franca-Entropy #itau=4
Tau Matrix: Matrix-Ent-Adv
# Tau Matrix: Diagonal-Shakib-Entropy
Include Viscous Correction in Stabilization: False # if p=1 idiff=1
# if p=2 idiff=2
Tau Time Constant: 1.0
Tau C Scale Factor: 1.0 # taucfct best value depends
Number of Elements Per Block: 64 #ibksiz

#}

#DISCONTINUITY CAPTURING
#{
Discontinuity Capturing: DC-quadratic # Current Options: DC-mallet, DC-minimum, DC-quadratic, DC-yzbeta
Multiplier for DC factor: 1 # scales DC variable in e3DC
Discontinuity Capturing Scheme: 1 #0: discontinuous, 1: continuous (L2 projection)
Include Source Term in DC: 0 # 1: sets idcSRC to 1
Write DCqpt: 0
#----Parameters for YZBeta DC operator ----
Beta Value: 1 # 1: smoother , 2: sharper, 12: compromise between 1 and 2
YZB Farfield Conditions: 1e5 2119 10 10 300 # [Pressure, X-Vel, Y-Vel, Z-Vel, Temperature]
YZB Farfield Mole Frac: 1 0 0 0 0 # mole fractions at reference condition
# [xN2,xO2,xNO,xN,xO]
# must sum to 1, must be length 5
Include Umod Term: 1 # 0: no, 1: yes
Mach Adjustment Bm Value: 1 # 0: off,1: smoother shock, 2: sharper shock
Mach Adjustment Bj Value: 6 # 0: off,1: smoother shock, 2: sharper shock
Include Time Term in Z: 1 # 0: no, 1: yes
#------------------------------------------

#}

#STEP SEQUENCE
#{
Step Construction : 0 1 0 1
#}</nowiki>

TCNEQ Version

2022-03-17T22:40:42Z

Jopo9795: /* Simulation Inputs */

== Background ==
The following information relates to the use of the thermochemical nonequilibrium (TCNEQ) version of PHASTA written in terms of entropy variables. The reader is referred to the following for additional information.

* F. Chalot, T.J.R. Hughes, and F. Shakib, '''"Symmetrization of Conservation Laws with Entropy for High-Temperature Hypersonic Computations,"''' Computing Systems in Engineering, 1(2-4):495–521, 1990.

* J. Pointer, '''"Influence of Interpolation Variables and Discontinuity Capturing Operators on Inviscid Hypersonic Flow Simulations Using a Stabilized Continuous Galerkin Solver,"''' Ph.D. dissertation, University of Colorado, Boulder, CO, 2022.

== Pre-Processing ==
In this section, details of the meshing and model attributes are provided. For simulation cases where the gas is modeled as a single species, the ''scalar'' quantities for boundary and initial conditions are not required. Currently, capability exists to simulate a gas with number of species (''nsp'') ≤ 5. When 2 ≤ ''nsp'' ≤ 5, the ''scalar'' quantities are used to govern the composition of the gas.

=== Meshing ===
Within the Simmodeler utility, the mesh can either be created or loaded from an existing .cas file. Below are steps for loading a mesh from a .cas file:
# Launch Simmodeler (for this example, SimModeler7.0-190604 is used)
# File > Import Discrete Data > (select .cas file to import) > (keep defaults and click OK) > (select YES to keep volume mesh)
# Save .sms and .smd files
# Attributes can now be assigned to the model as normal

=== Boundary Conditions ===
Below are the recognized boundary conditions that can be applied for the current version:
* comp1/comp2/comp3 - Specification of one/two/three components of velocity, [m/s]
* temperature - Specification of translational-rotational temperature, [K]. By default, vibrational temperature is held in equilibrium with this value and nonequilibrium is controlled through simulation inputs.
* surfID - When value is set to 702, the boundary is treated as a slip wall. If using this option, include a boundary layer mesh along the surface to ensure the wall normal direction is accurately computed.
* scalar_1 - Mole fraction of species 2 of the gas
* scalar_2 - Mole fraction of species 3 of the gas
* scalar_3 - Mole fraction of species 4 of the gas
* scalar_4 - Mole fraction of species 5 of the gas
* pressure - Specification of static pressure over a surface, [Pa]
** Used to compute mole fraction of species 1 of the gas with Dalton's Law of partial pressures and subtracting the summation of the other mole fractions from a value of 1
* heat flux - set to zero for adiabatic wall boundary condition

=== Initial Conditions ===
Below are the required initial conditions for the current version:
* initial velocity - Components and magnitude of flow velocity, [m/s]
** If a supersonic outlet condition is used, set such that flow is initialized Mach > 1
* initial temperature - Value used to set translational-rotational temperature, [K]
* initial scalar_1 - Initial value of species 2 mole fraction
* initial scalar_2 - Initial value of species 3 mole fraction
* initial scalar_3 - Initial value of species 4 mole fraction
* initial scalar_4 - Initial value of species 5 mole fraction
* initial pressure - Static pressure of the gas, [Pa]
** For multi-species flows, this value is used in combination with the initial scalar values to compute the mole fraction of species 1

== Simulation Inputs ==

Example of solver.inp file:

<nowiki># ibksiz flmpl flmpr itwmod wmodts dmodts fwr taucfct
# PHASTA Version 1.5 Input File
#
# Basic format is
#
# Key Phrase : Acceptable Value (integer, double, logical, or phrase
# list of integers, list of doubles )
#
#
# ****************** Notes ********************************
# 1. Bulk viscosity set to 0 in common.f (bmu)
# 2. Right now, nsp cannot be anything other than 5
# 3. idiff must be zero for now
# 4. iqtot does nothing for now since source terms are not being handled
# (commented out call to e3source in e3.f)
# 5.

#SOLUTION CONTROL
#{
Equation of State: Compressible
Number of Timesteps: 1
Time Step Size: 3.2339656949004e-08 # cfl = 1 is dt = 7.8702386e-8

Limit Density: 0 0.01 0.1
Limit u1: 0 0. 2.8e3
Limit u2: 0 0 0
Limit u3: 0 0 0
Limit Temperature: 0 230 3500 !also limits vibrational temperature
#}

#OUTPUT CONTROL
#{
Number of Timesteps between Restarts: 250 #replaces nout/ntout
Print Error Indicators: True
Error Indicator Threshold: 0.01
Number of Error Smoothing Iterations: 0 #ierrsmooth
Load and set 3D IC: False
Position Tolerance on IC Load: 1e-7 # sets the tolerance for matching points
#}

#MATERIAL CONTROL
#{
Viscosity: 1.95508431704028e-5 #<<<<< this is required to run (value used if nsp.eq.99)
Thermal Conductivity: 26.6843390135759e-3
Viscous Control: None #Viscous #None

#}

#REACTING FLOW
#{
Number of species: 1 # nsp
#.......currently only allowing 1<=nsp<=5.
#.........and make sure scalars passed from simmodeler are in correct slots
#
Species IDs: 1 # specIDs, length of array must equal
# nsp (IDs numbered in order:
# N2,O2,NO,N,O
# ID:99 for Air molecule
Ref Entropy Conditions: 1e3 230 230 0 #[P0,T0,T0vib,S0]
Ref Entropy Species: 1 2 # uses IDs above (NOT ACTIVE YET)
Ref Entropy Mole Frac: 1 0 0 0 0
Allow reactions: False # ichem = 1 if True
Chemical heat release: False # iqtot = 1 if True
Limit on reaction step: 0.00001 # rlim (limits change in species cs per step)
Tolerance to global time: 1.0 #0.01 # ttol
Temperature threshold: 500 # Tth (below which, reactions ignored)
Reaction solver MIN steps: 5 # nstepmin
Reaction solver MAX steps: 100 # nstepmax
Two Temperature coefficient: 0.5 # qta (Tvib**qta*T**(1-qta))
Exclude vib energy: True # ivib0 = 1 if True
Exclude vib source: True # ivibS0 = 1 if True
Tvib BC Ratio: 1.0 # at any BC with T set, Tvib = tvibBC * T
Vibrational Temperature IC: -1 # set negative value to force Tvib = T
#}

#LINEAR SOLVER
#
Solver Type: GMRES sparse <<< default, but why does it have to be set?
Number of GMRES Sweeps per Solve: 1 # replaces nGMRES
Minimum Number of Iterations per Nonlinear Iteration: 10 # minIters
Number of Krylov Vectors per GMRES Sweep: 100 # replaces Kspace
Tolerance on Momentum Equations: 0.01 # epstol(1), affects etol for Hessenberg problem

#}

#DISCRETIZATION CONTROL
#{
Weak Form: SUPG # alternate is Galerkin only for compressible
Time Integration Rule: First Order # 1st Order sets rinf(1) -1
# Time Integration Rule: Second Order # Second Order sets rinf next
# Time Integration Rho Infinity: 0.5 # rinf(1) Only used for 2nd order
# Tau Matrix: Diagonal-Franca-Entropy #itau=4
Tau Matrix: Matrix-Ent-Adv
# Tau Matrix: Diagonal-Shakib-Entropy
Include Viscous Correction in Stabilization: False # if p=1 idiff=1
# if p=2 idiff=2
Tau Time Constant: 1.0
Tau C Scale Factor: 1.0 # taucfct best value depends
Number of Elements Per Block: 64 #ibksiz

#}

#DISCONTINUITY CAPTURING
#{
Discontinuity Capturing: DC-quadratic # Current Options: DC-mallet, DC-minimum, DC-quadratic, DC-yzbeta
Multiplier for DC factor: 1 # scales DC variable in e3DC
Discontinuity Capturing Scheme: 1 #0: discontinuous, 1: continuous (L2 projection)
Include Source Term in DC: 0 # 1: sets idcSRC to 1
Write DCqpt: 0
#----Parameters for YZBeta DC operator ----
Beta Value: 1 # 1: smoother , 2: sharper, 12: compromise between 1 and 2
YZB Farfield Conditions: 1e5 2119 10 10 300 # [Pressure, X-Vel, Y-Vel, Z-Vel, Temperature]
YZB Farfield Mole Frac: 1 0 0 0 0 # mole fractions at reference condition
# [xN2,xO2,xNO,xN,xO]
# must sum to 1, must be length 5
Include Umod Term: 1 # 0: no, 1: yes
Mach Adjustment Bm Value: 1 # 0: off,1: smoother shock, 2: sharper shock
Mach Adjustment Bj Value: 6 # 0: off,1: smoother shock, 2: sharper shock
Include Time Term in Z: 1 # 0: no, 1: yes
#------------------------------------------

#}

#STEP SEQUENCE
#{
Step Construction : 0 1 0 1
#}</nowiki>

TCNEQ Version

2022-03-17T19:41:07Z

Jopo9795: /* Pre-Processing */

TCNEQ Version

2022-03-17T19:37:56Z

Jopo9795: /* Pre-Processing */

TCNEQ Version

2022-03-17T19:30:40Z

Jopo9795: /* Initial Conditions */

TCNEQ Version

2022-03-17T19:30:17Z

Jopo9795: /* Initial Conditions */

TCNEQ Version

2022-03-17T19:29:50Z

Jopo9795: /* Boundary Conditions */

TCNEQ Version

2022-03-17T19:28:47Z

Jopo9795: /* Initial Conditions */

TCNEQ Version

2022-03-17T19:24:45Z

Jopo9795: /* Boundary Conditions */

TCNEQ Version

2022-03-17T16:28:25Z

Jopo9795: /* Pre-Processing */

TCNEQ Version

2022-03-17T14:54:16Z

Jopo9795: /* Pre-Processing */

TCNEQ Version

2022-03-14T14:54:35Z

Jopo9795: /* Background */

TCNEQ Version

2022-03-14T14:51:45Z

Jopo9795: /* Background */

PhParAdapt/Simmetrix

2021-11-09T17:55:40Z

Jopo9795: /* Creating Initial Restart and Error Files */

The following page provides details for performing mesh adaptation using the PhParAdapt tool with Simmetrix routines.

== Initial Notes to User ==
* The solution migration feature is currently broken and thus requires other means for transferring the solution to new meshes that are created from phParAdapt (e.g. solution interpolation in Paraview)
* To date, the steps described herein have only been tried with the following restrictions:
** Tetrahedral elements
*** Extruded element types apparently can be used but require removal of any extrusion constraints in the mesh. Below are some incomplete details regarding that process for reference. See [[#Removal_of_Extrusion_Constraint |Removal of Extrusion Constraint]].
** Serial case
** Geometries created using SimModeler7.0-190604
** The version of phParAdapt noted here uses only the 6th entry of the error field for identifying regions for adaptation

== Adaptation Process ==

=== Creating Initial Restart and Error Files ===
Beginning inside of the 1-1-Chef directory,
<nowiki>mkdir 1-1-phParAdapt
cd 1-1-phParAdapt </nowiki>

Create soft links to the geom files above the 1-1-Chef directory,
<nowiki>ln -s ../../geom.smd
ln -s ../../geom.sms
ln -s ../../geom_nat.x_t </nowiki>

Create an <code>adapt.inp</code> file using the template [[#Adapt.inp_(non-adaptation_step) | Adapt.inp (non-adaptation step)]]. Keep <code>adaptFlag</code> set to zero as the purpose of running phParAdapt here is to create restart.<timeStepNumber>.1 file associated with a new mesh format that can be used for adaptation.

Run phParAdapt in this directory, which for this specific example was:

<code>mpirun -np 1 /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi-O 2>&1 | tee phParAdapt.log</code>

If the above step fails, review the 'MeshSim.#.log' file that is produced to identify the error. One error related to analysis attributes has been experienced. The issue was found to be that under the 'Analysis attributes' section, on the 'problem definition' line the name must be set to 'geom'.

After completion, the 1-procs_case and mesh_parts.sms directories should now be present. If a solution has already been generated for a different mesh on the same geometry, it can now be interpolated onto the new restart.<#>.1 file that was generated. If no solution exists, run Phasta using this restart file until the desired solution state has been achieved. Be sure to set <code>Print Error Indicators: True</code> in the solver.inp file so the error fields are saved in the final restart file. Note, error fields are not printed in intermediate restart files. Before proceeding, check that the error fields exist using <code>grep -a " : <" restart.<#>.1</code>

=== Performing the Adaptation Step ===
Within the 1-1-phParAdapt directory,
<nowiki>mkdir A1-phParAdapt
cd A1-phParAdapt
ln -s ../geom.smd
ln -s ../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
ln -s geom.sms parts.sms
ln -s ../1-procs_case/restart.1.1
ln -s restart.1.1 errors.1.1 </nowiki>

In the above step, it is assumed that the solution was interpolated onto restart.1.1. If that was not the case and the solution was advanced until a desired step was reached, replace restart.1.1 & errors.1.1 with restart.<#>.1 & errors.<#>.1 corresponding to the desired solution step number with error fields.

Create an adapt.inp file using [[#Adapt.inp_(adaptation step) |Adapt.inp_(adaptation step)]] as a template. Change the <code>timeStepNumber</code> parameter to correspond to the restart file that was linked into the current directory. Make sure the <code>adaptFlag</code> parameter is set to one and change additional threshold flags to control the result of the refinement.

Run phParAdapt using the mpirun command from the previous section. A directory corresponding to the <code>timeStepNumber</code> parameter should have been created with a mesh_parts.sms folder inside upon successful completion.

=== Generate Restart File for Refined Mesh ===
From within the A1-phParAdapt directory, execute the following (noting to replace <#> with the appropriate time step number):
<nowiki>cd <#>
mkdir 1-1-phParAdapt
cd 1-1-phParAdapt
ln -s ../../../geom.smd
ln -s ../../../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
cp ../../../adapt.inp .</nowiki>

In the above list of commands, the adapt.inp file from the non-adaptation step was copied into the current directory. Run phParAdapt using the same mpirun command as above to generate the 1-procs_case folder and the new restart file with the refined mesh. Again, since the solution migration feature is not working, the restart file will contain the initial solution and no error fields. A previously obtained solution can now be interpolated to the newly created one or PHASTA can be run in this directory to obtain a solution on the new mesh.

In order to perform a second (or more) adaptation step, create a restart file in the 1-procs_case folder in the current directory (either by interpolation or running PHASTA) and make sure it contains the error fields. Create a new directory <code>A2-phParAdapt</code> at this level and repeat the above steps. Be sure to use the time step number moving forward that is associated with the restart file just mentioned.

== File Examples ==

The file examples below are what were used for this specific example and may contain parameter flags that may need to be changed or are not used at all.

=== Adapt.inp (non-adaptation step) ===
<nowiki>numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 0
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.00001
AdaptOption 11
hmax 0.04
hmin 0.000052083
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 1
numTotParts 1
SolutionMigration 0
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 1
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

=== Adapt.inp (adaptation step) ===
<nowiki>eviscScal 0
pgrad 0
ub 1.45
lb 0.70
numSplit 5
sizeRatio 0.35 # confirmed that 0.25 does a 4x ref
numSmooth 0 #confirmed that this gets used
ratioThresh 0.75 # this is h_new/h_orig from isotropic days
localAdapt 1 # makes setLocal force adaptivity only around nodes set size
AnisoSimmetrix 4 # > 0 use Simmetrix, else ours 1 does largest length reduction by sizeRatio, 2 does both does medium and largest reduction by sizeRatio
coarsenMode 0 # disables coarsening see coarsenMode docs for 1, and 2
numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 1
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.5e-4
AdaptOption 11
hmax 1e6
hmin 1e-5
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 0
numTotParts 1
SolutionMigration 1
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 0
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

== Notes on Debugging ==
If debugging of phParAdapt is desired, replace the normal <code>mpirun</code> command with the following:

<code>mpirun -np 1 -tv /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi</code>

where the <code>-O</code> has been removed from the path to the executable to call the build version that was not optimized. Note that this requires phParAdapt to have been built at some point prior with the <code>Debug</code> flag.

== Notes on Folder Structure ==
Below is an overview of the folder structure to reference during the adaptation process:
# [ geom.smd, geom.sms, geom_nat.x_t ]
# 1-1-Chef
## 1-1-phParAdapt
### adapt.inp
### run_phParAdapt.sh
### 1-procs_case
#### geombc.dat.1
#### restart.<#>.1
### mesh_parts.sms
### A1-phParAdapt
#### restart.<#>.1
#### errors.<#>.1
#### [geom.smd, geom.sms, geom_nat.x_t, parts.sms]
#### adapt.inp
#### <#> ( step number directory )
##### mesh_parts.sms
##### 1-1-phParAdapt
###### [geom.smd, geom.sms, geom_nat.x_t]
###### adapt.inp
###### 1-procs_case ( restart files with adapted mesh )

== Removal of Extrusion Constraint ==
Riccardo has used a procedure in the past to remove the extrusion constraint from a mesh that enabled adaptation to be performed on extrusion-type elements.

The following directory contains an example of this process:
<code> /projects/tools/Models/NASAWingBodyJunction/RajMeshFine/Mesh/1-1-phParAdapt/RemExtrusion/ </code>

The executable that is called for removal of the constraint is a part of SCOREC-core, found at <code> <path/to/SCOREC-core/build/dir>/test/rm_extrusion </code> and the source code, in case edits are required, is found at <code> <path/to/SCOREC-core/source/dir>/core/test/rm_extrusion.cc </code>

PhParAdapt/Simmetrix

2021-07-14T16:10:40Z

Jopo9795: /* Creating Initial Restart and Error Files */

The following page provides details for performing mesh adaptation using the PhParAdapt tool with Simmetrix routines.

== Initial Notes to User ==
* The solution migration feature is currently broken and thus requires other means for transferring the solution to new meshes that are created from phParAdapt (e.g. solution interpolation in Paraview)
* To date, the steps described herein have only been tried with the following restrictions:
** Tetrahedral elements
*** Extruded element types apparently can be used but require removal of any extrusion constraints in the mesh. Below are some incomplete details regarding that process for reference. See [[#Removal_of_Extrusion_Constraint |Removal of Extrusion Constraint]].
** Serial case
** Geometries created using SimModeler7.0-190604
** The version of phParAdapt noted here uses only the 6th entry of the error field for identifying regions for adaptation

== Adaptation Process ==

=== Creating Initial Restart and Error Files ===
Beginning inside of the 1-1-Chef directory,
<nowiki>mkdir 1-1-phParAdapt
cd 1-1-phParAdapt </nowiki>

Create soft links to the geom files above the 1-1-Chef directory,
<nowiki>ln -s ../../geom.smd
ln -s ../../geom.sms
ln -s ../../geom_nat.x_t </nowiki>

Create an <code>adapt.inp</code> file using the template [[#Adapt.inp_(non-adaptation_step) | Adapt.inp (non-adaptation step)]]. Keep <code>adaptFlag</code> set to zero as the purpose of running phParAdapt here is to create restart.<timeStepNumber>.1 file associated with a new mesh format that can be used for adaptation.

Run phParAdapt in this directory, which for this specific example was:

<code>mpirun -np 1 /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi-O 2>&1 | tee phParAdapt.log</code>

If the above step fails, review the 'MeshSim.#.log' file that is produced to identify the error. One error related to analysis attributes has been experienced. Removing all attributes other than an initial temperature attribute in Simmodeler and then recreating the soft links in the above step fixed this issue.

After completion, the 1-procs_case and mesh_parts.sms directories should now be present. If a solution has already been generated for a different mesh on the same geometry, it can now be interpolated onto the new restart.<#>.1 file that was generated. If no solution exists, run Phasta using this restart file until the desired solution state has been achieved. Be sure to set <code>Print Error Indicators: True</code> in the solver.inp file so the error fields are saved in the final restart file. Note, error fields are not printed in intermediate restart files. Before proceeding, check that the error fields exist using <code>grep -a " : <" restart.<#>.1</code>

=== Performing the Adaptation Step ===
Within the 1-1-phParAdapt directory,
<nowiki>mkdir A1-phParAdapt
cd A1-phParAdapt
ln -s ../geom.smd
ln -s ../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
ln -s geom.sms parts.sms
ln -s ../1-procs_case/restart.1.1
ln -s restart.1.1 errors.1.1 </nowiki>

In the above step, it is assumed that the solution was interpolated onto restart.1.1. If that was not the case and the solution was advanced until a desired step was reached, replace restart.1.1 & errors.1.1 with restart.<#>.1 & errors.<#>.1 corresponding to the desired solution step number with error fields.

Create an adapt.inp file using [[#Adapt.inp_(adaptation step) |Adapt.inp_(adaptation step)]] as a template. Change the <code>timeStepNumber</code> parameter to correspond to the restart file that was linked into the current directory. Make sure the <code>adaptFlag</code> parameter is set to one and change additional threshold flags to control the result of the refinement.

Run phParAdapt using the mpirun command from the previous section. A directory corresponding to the <code>timeStepNumber</code> parameter should have been created with a mesh_parts.sms folder inside upon successful completion.

=== Generate Restart File for Refined Mesh ===
From within the A1-phParAdapt directory, execute the following (noting to replace <#> with the appropriate time step number):
<nowiki>cd <#>
mkdir 1-1-phParAdapt
cd 1-1-phParAdapt
ln -s ../../../geom.smd
ln -s ../../../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
cp ../../../adapt.inp .</nowiki>

In the above list of commands, the adapt.inp file from the non-adaptation step was copied into the current directory. Run phParAdapt using the same mpirun command as above to generate the 1-procs_case folder and the new restart file with the refined mesh. Again, since the solution migration feature is not working, the restart file will contain the initial solution and no error fields. A previously obtained solution can now be interpolated to the newly created one or PHASTA can be run in this directory to obtain a solution on the new mesh.

In order to perform a second (or more) adaptation step, create a restart file in the 1-procs_case folder in the current directory (either by interpolation or running PHASTA) and make sure it contains the error fields. Create a new directory <code>A2-phParAdapt</code> at this level and repeat the above steps. Be sure to use the time step number moving forward that is associated with the restart file just mentioned.

== File Examples ==

The file examples below are what were used for this specific example and may contain parameter flags that may need to be changed or are not used at all.

=== Adapt.inp (non-adaptation step) ===
<nowiki>numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 0
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.00001
AdaptOption 11
hmax 0.04
hmin 0.000052083
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 1
numTotParts 1
SolutionMigration 0
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 1
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

=== Adapt.inp (adaptation step) ===
<nowiki>eviscScal 0
pgrad 0
ub 1.45
lb 0.70
numSplit 5
sizeRatio 0.35 # confirmed that 0.25 does a 4x ref
numSmooth 0 #confirmed that this gets used
ratioThresh 0.75 # this is h_new/h_orig from isotropic days
localAdapt 1 # makes setLocal force adaptivity only around nodes set size
AnisoSimmetrix 4 # > 0 use Simmetrix, else ours 1 does largest length reduction by sizeRatio, 2 does both does medium and largest reduction by sizeRatio
coarsenMode 0 # disables coarsening see coarsenMode docs for 1, and 2
numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 1
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.5e-4
AdaptOption 11
hmax 1e6
hmin 1e-5
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 0
numTotParts 1
SolutionMigration 1
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 0
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

== Notes on Debugging ==
If debugging of phParAdapt is desired, replace the normal <code>mpirun</code> command with the following:

<code>mpirun -np 1 -tv /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi</code>

where the <code>-O</code> has been removed from the path to the executable to call the build version that was not optimized. Note that this requires phParAdapt to have been built at some point prior with the <code>Debug</code> flag.

== Notes on Folder Structure ==
Below is an overview of the folder structure to reference during the adaptation process:
# [ geom.smd, geom.sms, geom_nat.x_t ]
# 1-1-Chef
## 1-1-phParAdapt
### adapt.inp
### run_phParAdapt.sh
### 1-procs_case
#### geombc.dat.1
#### restart.<#>.1
### mesh_parts.sms
### A1-phParAdapt
#### restart.<#>.1
#### errors.<#>.1
#### [geom.smd, geom.sms, geom_nat.x_t, parts.sms]
#### adapt.inp
#### <#> ( step number directory )
##### mesh_parts.sms
##### 1-1-phParAdapt
###### [geom.smd, geom.sms, geom_nat.x_t]
###### adapt.inp
###### 1-procs_case ( restart files with adapted mesh )

== Removal of Extrusion Constraint ==
Riccardo has used a procedure in the past to remove the extrusion constraint from a mesh that enabled adaptation to be performed on extrusion-type elements.

The following directory contains an example of this process:
<code> /projects/tools/Models/NASAWingBodyJunction/RajMeshFine/Mesh/1-1-phParAdapt/RemExtrusion/ </code>

The executable that is called for removal of the constraint is a part of SCOREC-core, found at <code> <path/to/SCOREC-core/build/dir>/test/rm_extrusion </code> and the source code, in case edits are required, is found at <code> <path/to/SCOREC-core/source/dir>/core/test/rm_extrusion.cc </code>

PhParAdapt/Simmetrix

2021-04-28T22:50:55Z

Jopo9795: /* Initial Notes to User */

The following page provides details for performing mesh adaptation using the PhParAdapt tool with Simmetrix routines.

== Initial Notes to User ==
* The solution migration feature is currently broken and thus requires other means for transferring the solution to new meshes that are created from phParAdapt (e.g. solution interpolation in Paraview)
* To date, the steps described herein have only been tried with the following restrictions:
** Tetrahedral elements
*** Extruded element types apparently can be used but require removal of any extrusion constraints in the mesh. Below are some incomplete details regarding that process for reference. See [[#Removal_of_Extrusion_Constraint |Removal of Extrusion Constraint]].
** Serial case
** Geometries created using SimModeler7.0-190604
** The version of phParAdapt noted here uses only the 6th entry of the error field for identifying regions for adaptation

== Adaptation Process ==

=== Creating Initial Restart and Error Files ===
Beginning inside of the 1-1-Chef directory,
<nowiki>mkdir 1-1-phParAdapt
cd 1-1-phParAdapt </nowiki>

Create soft links to the geom files above the 1-1-Chef directory,
<nowiki>ln -s ../../geom.smd
ln -s ../../geom.sms
ln -s ../../geom_nat.x_t </nowiki>

Create an <code>adapt.inp</code> file using the template [[#Adapt.inp_(non-adaptation_step) | Adapt.inp (non-adaptation step)]]. Keep <code>adaptFlag</code> set to zero as the purpose of running phParAdapt here is to create restart.<timeStepNumber>.1 file associated with a new mesh format that can be used for adaptation.

Run phParAdapt in this directory, which for this specific example was:

<code>mpirun -np 1 /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi-O 2>&1 | tee phParAdapt.log</code>

After completion, the 1-procs_case and mesh_parts.sms directories should now be present. If a solution has already been generated for a different mesh on the same geometry, it can now be interpolated onto the new restart.<#>.1 file that was generated. If no solution exists, run Phasta using this restart file until the desired solution state has been achieved. Be sure to set <code>Print Error Indicators: True</code> in the solver.inp file so the error fields are saved in the final restart file. Note, error fields are not printed in intermediate restart files. Before proceeding, check that the error fields exist using <code>grep -a " : <" restart.<#>.1</code>

=== Performing the Adaptation Step ===
Within the 1-1-phParAdapt directory,
<nowiki>mkdir A1-phParAdapt
cd A1-phParAdapt
ln -s ../geom.smd
ln -s ../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
ln -s geom.sms parts.sms
ln -s ../1-procs_case/restart.1.1
ln -s restart.1.1 errors.1.1 </nowiki>

In the above step, it is assumed that the solution was interpolated onto restart.1.1. If that was not the case and the solution was advanced until a desired step was reached, replace restart.1.1 & errors.1.1 with restart.<#>.1 & errors.<#>.1 corresponding to the desired solution step number with error fields.

Create an adapt.inp file using [[#Adapt.inp_(adaptation step) |Adapt.inp_(adaptation step)]] as a template. Change the <code>timeStepNumber</code> parameter to correspond to the restart file that was linked into the current directory. Make sure the <code>adaptFlag</code> parameter is set to one and change additional threshold flags to control the result of the refinement.

Run phParAdapt using the mpirun command from the previous section. A directory corresponding to the <code>timeStepNumber</code> parameter should have been created with a mesh_parts.sms folder inside upon successful completion.

=== Generate Restart File for Refined Mesh ===
From within the A1-phParAdapt directory, execute the following (noting to replace <#> with the appropriate time step number):
<nowiki>cd <#>
mkdir 1-1-phParAdapt
cd 1-1-phParAdapt
ln -s ../../../geom.smd
ln -s ../../../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
cp ../../../adapt.inp .</nowiki>

In the above list of commands, the adapt.inp file from the non-adaptation step was copied into the current directory. Run phParAdapt using the same mpirun command as above to generate the 1-procs_case folder and the new restart file with the refined mesh. Again, since the solution migration feature is not working, the restart file will contain the initial solution and no error fields. A previously obtained solution can now be interpolated to the newly created one or PHASTA can be run in this directory to obtain a solution on the new mesh.

In order to perform a second (or more) adaptation step, create a restart file in the 1-procs_case folder in the current directory (either by interpolation or running PHASTA) and make sure it contains the error fields. Create a new directory <code>A2-phParAdapt</code> at this level and repeat the above steps. Be sure to use the time step number moving forward that is associated with the restart file just mentioned.

== File Examples ==

The file examples below are what were used for this specific example and may contain parameter flags that may need to be changed or are not used at all.

=== Adapt.inp (non-adaptation step) ===
<nowiki>numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 0
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.00001
AdaptOption 11
hmax 0.04
hmin 0.000052083
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 1
numTotParts 1
SolutionMigration 0
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 1
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

=== Adapt.inp (adaptation step) ===
<nowiki>eviscScal 0
pgrad 0
ub 1.45
lb 0.70
numSplit 5
sizeRatio 0.35 # confirmed that 0.25 does a 4x ref
numSmooth 0 #confirmed that this gets used
ratioThresh 0.75 # this is h_new/h_orig from isotropic days
localAdapt 1 # makes setLocal force adaptivity only around nodes set size
AnisoSimmetrix 4 # > 0 use Simmetrix, else ours 1 does largest length reduction by sizeRatio, 2 does both does medium and largest reduction by sizeRatio
coarsenMode 0 # disables coarsening see coarsenMode docs for 1, and 2
numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 1
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.5e-4
AdaptOption 11
hmax 1e6
hmin 1e-5
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 0
numTotParts 1
SolutionMigration 1
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 0
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

== Notes on Debugging ==
If debugging of phParAdapt is desired, replace the normal <code>mpirun</code> command with the following:

<code>mpirun -np 1 -tv /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi</code>

where the <code>-O</code> has been removed from the path to the executable to call the build version that was not optimized. Note that this requires phParAdapt to have been built at some point prior with the <code>Debug</code> flag.

== Notes on Folder Structure ==
Below is an overview of the folder structure to reference during the adaptation process:
# [ geom.smd, geom.sms, geom_nat.x_t ]
# 1-1-Chef
## 1-1-phParAdapt
### adapt.inp
### run_phParAdapt.sh
### 1-procs_case
#### geombc.dat.1
#### restart.<#>.1
### mesh_parts.sms
### A1-phParAdapt
#### restart.<#>.1
#### errors.<#>.1
#### [geom.smd, geom.sms, geom_nat.x_t, parts.sms]
#### adapt.inp
#### <#> ( step number directory )
##### mesh_parts.sms
##### 1-1-phParAdapt
###### [geom.smd, geom.sms, geom_nat.x_t]
###### adapt.inp
###### 1-procs_case ( restart files with adapted mesh )

== Removal of Extrusion Constraint ==
Riccardo has used a procedure in the past to remove the extrusion constraint from a mesh that enabled adaptation to be performed on extrusion-type elements.

The following directory contains an example of this process:
<code> /projects/tools/Models/NASAWingBodyJunction/RajMeshFine/Mesh/1-1-phParAdapt/RemExtrusion/ </code>

The executable that is called for removal of the constraint is a part of SCOREC-core, found at <code> <path/to/SCOREC-core/build/dir>/test/rm_extrusion </code> and the source code, in case edits are required, is found at <code> <path/to/SCOREC-core/source/dir>/core/test/rm_extrusion.cc </code>

PhParAdapt/Simmetrix

2021-04-28T20:20:52Z

Jopo9795: /* Creating Initial Restart and Error Files */

The following page provides details for performing mesh adaptation using the PhParAdapt tool with Simmetrix routines.

== Initial Notes to User ==
* The solution migration feature is currently broken and thus requires other means for transferring the solution to new meshes that are created from phParAdapt (e.g. solution interpolation in Paraview)
* To date, the steps described herein have only been tried with the following restrictions:
** Tetrahedral elements
*** Extruded element types apparently can be used but require removal of any extrusion constraints in the mesh. Below are some incomplete details regarding that process for reference. See [[#Removal_of_Extrusion_Constraint |Removal of Extrusion Constraint]].
** Serial case
** Geometries created using SimModeler7.0-190626
** The version of phParAdapt noted here uses only the 6th entry of the error field for identifying regions for adaptation

== Adaptation Process ==

=== Creating Initial Restart and Error Files ===
Beginning inside of the 1-1-Chef directory,
<nowiki>mkdir 1-1-phParAdapt
cd 1-1-phParAdapt </nowiki>

Create soft links to the geom files above the 1-1-Chef directory,
<nowiki>ln -s ../../geom.smd
ln -s ../../geom.sms
ln -s ../../geom_nat.x_t </nowiki>

Create an <code>adapt.inp</code> file using the template [[#Adapt.inp_(non-adaptation_step) | Adapt.inp (non-adaptation step)]]. Keep <code>adaptFlag</code> set to zero as the purpose of running phParAdapt here is to create restart.<timeStepNumber>.1 file associated with a new mesh format that can be used for adaptation.

Run phParAdapt in this directory, which for this specific example was:

<code>mpirun -np 1 /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi-O 2>&1 | tee phParAdapt.log</code>

After completion, the 1-procs_case and mesh_parts.sms directories should now be present. If a solution has already been generated for a different mesh on the same geometry, it can now be interpolated onto the new restart.<#>.1 file that was generated. If no solution exists, run Phasta using this restart file until the desired solution state has been achieved. Be sure to set <code>Print Error Indicators: True</code> in the solver.inp file so the error fields are saved in the final restart file. Note, error fields are not printed in intermediate restart files. Before proceeding, check that the error fields exist using <code>grep -a " : <" restart.<#>.1</code>

=== Performing the Adaptation Step ===
Within the 1-1-phParAdapt directory,
<nowiki>mkdir A1-phParAdapt
cd A1-phParAdapt
ln -s ../geom.smd
ln -s ../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
ln -s geom.sms parts.sms
ln -s ../1-procs_case/restart.1.1
ln -s restart.1.1 errors.1.1 </nowiki>

In the above step, it is assumed that the solution was interpolated onto restart.1.1. If that was not the case and the solution was advanced until a desired step was reached, replace restart.1.1 & errors.1.1 with restart.<#>.1 & errors.<#>.1 corresponding to the desired solution step number with error fields.

Create an adapt.inp file using [[#Adapt.inp_(adaptation step) |Adapt.inp_(adaptation step)]] as a template. Change the <code>timeStepNumber</code> parameter to correspond to the restart file that was linked into the current directory. Make sure the <code>adaptFlag</code> parameter is set to one and change additional threshold flags to control the result of the refinement.

Run phParAdapt using the mpirun command from the previous section. A directory corresponding to the <code>timeStepNumber</code> parameter should have been created with a mesh_parts.sms folder inside upon successful completion.

=== Generate Restart File for Refined Mesh ===
From within the A1-phParAdapt directory, execute the following (noting to replace <#> with the appropriate time step number):
<nowiki>cd <#>
mkdir 1-1-phParAdapt
cd 1-1-phParAdapt
ln -s ../../../geom.smd
ln -s ../../../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
cp ../../../adapt.inp .</nowiki>

In the above list of commands, the adapt.inp file from the non-adaptation step was copied into the current directory. Run phParAdapt using the same mpirun command as above to generate the 1-procs_case folder and the new restart file with the refined mesh. Again, since the solution migration feature is not working, the restart file will contain the initial solution and no error fields. A previously obtained solution can now be interpolated to the newly created one or PHASTA can be run in this directory to obtain a solution on the new mesh.

In order to perform a second (or more) adaptation step, create a restart file in the 1-procs_case folder in the current directory (either by interpolation or running PHASTA) and make sure it contains the error fields. Create a new directory <code>A2-phParAdapt</code> at this level and repeat the above steps. Be sure to use the time step number moving forward that is associated with the restart file just mentioned.

== File Examples ==

The file examples below are what were used for this specific example and may contain parameter flags that may need to be changed or are not used at all.

=== Adapt.inp (non-adaptation step) ===
<nowiki>numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 0
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.00001
AdaptOption 11
hmax 0.04
hmin 0.000052083
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 1
numTotParts 1
SolutionMigration 0
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 1
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

=== Adapt.inp (adaptation step) ===
<nowiki>eviscScal 0
pgrad 0
ub 1.45
lb 0.70
numSplit 5
sizeRatio 0.35 # confirmed that 0.25 does a 4x ref
numSmooth 0 #confirmed that this gets used
ratioThresh 0.75 # this is h_new/h_orig from isotropic days
localAdapt 1 # makes setLocal force adaptivity only around nodes set size
AnisoSimmetrix 4 # > 0 use Simmetrix, else ours 1 does largest length reduction by sizeRatio, 2 does both does medium and largest reduction by sizeRatio
coarsenMode 0 # disables coarsening see coarsenMode docs for 1, and 2
numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 1
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.5e-4
AdaptOption 11
hmax 1e6
hmin 1e-5
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 0
numTotParts 1
SolutionMigration 1
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 0
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

== Notes on Debugging ==
If debugging of phParAdapt is desired, replace the normal <code>mpirun</code> command with the following:

<code>mpirun -np 1 -tv /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi</code>

where the <code>-O</code> has been removed from the path to the executable to call the build version that was not optimized. Note that this requires phParAdapt to have been built at some point prior with the <code>Debug</code> flag.

== Notes on Folder Structure ==
Below is an overview of the folder structure to reference during the adaptation process:
# [ geom.smd, geom.sms, geom_nat.x_t ]
# 1-1-Chef
## 1-1-phParAdapt
### adapt.inp
### run_phParAdapt.sh
### 1-procs_case
#### geombc.dat.1
#### restart.<#>.1
### mesh_parts.sms
### A1-phParAdapt
#### restart.<#>.1
#### errors.<#>.1
#### [geom.smd, geom.sms, geom_nat.x_t, parts.sms]
#### adapt.inp
#### <#> ( step number directory )
##### mesh_parts.sms
##### 1-1-phParAdapt
###### [geom.smd, geom.sms, geom_nat.x_t]
###### adapt.inp
###### 1-procs_case ( restart files with adapted mesh )

== Removal of Extrusion Constraint ==
Riccardo has used a procedure in the past to remove the extrusion constraint from a mesh that enabled adaptation to be performed on extrusion-type elements.

The following directory contains an example of this process:
<code> /projects/tools/Models/NASAWingBodyJunction/RajMeshFine/Mesh/1-1-phParAdapt/RemExtrusion/ </code>

The executable that is called for removal of the constraint is a part of SCOREC-core, found at <code> <path/to/SCOREC-core/build/dir>/test/rm_extrusion </code> and the source code, in case edits are required, is found at <code> <path/to/SCOREC-core/source/dir>/core/test/rm_extrusion.cc </code>

PhParAdapt/Simmetrix

2021-04-28T20:19:36Z

Jopo9795: /* Initial Notes to User */

The following page provides details for performing mesh adaptation using the PhParAdapt tool with Simmetrix routines.

== Initial Notes to User ==
* The solution migration feature is currently broken and thus requires other means for transferring the solution to new meshes that are created from phParAdapt (e.g. solution interpolation in Paraview)
* To date, the steps described herein have only been tried with the following restrictions:
** Tetrahedral elements
*** Extruded element types apparently can be used but require removal of any extrusion constraints in the mesh. Below are some incomplete details regarding that process for reference. See [[#Removal_of_Extrusion_Constraint |Removal of Extrusion Constraint]].
** Serial case
** Geometries created using SimModeler7.0-190626
** The version of phParAdapt noted here uses only the 6th entry of the error field for identifying regions for adaptation

== Adaptation Process ==

=== Creating Initial Restart and Error Files ===
Beginning inside of the 1-1-Chef directory,
<nowiki>mkdir 1-1-phParAdapt
cd 1-1-phParAdapt </nowiki>

Create soft links to the geom files above the 1-1-Chef directory,
<nowiki>ln -s ../../geom.smd
ln -s ../../geom.sms
ln -s ../../geom_nat.x_t </nowiki>

Create an <code> adapt.inp </code> using the template [[#Adapt.inp_(non-adaptation_step) | Adapt.inp (non-adaptation step)]]. Keep <code>adaptFlag</code> set to zero as the purpose of running phParAdapt here is to create restart.<timeStepNumber>.1 file associated with a new mesh format that can be used for adaptation.

Run phParAdapt in this directory, which for this specific example was:

<code>mpirun -np 1 /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi-O 2>&1 | tee phParAdapt.log</code>

After completion, the 1-procs_case and mesh_parts.sms directories should now be present. If a solution has already been generated for a different mesh on the same geometry, it can now be interpolated onto the new restart.<#>.1 file that was generated. If no solution exists, run Phasta using this restart file until the desired solution state has been achieved. Be sure to set <code>Print Error Indicators: True</code> in the solver.inp file so the error fields are saved in the final restart file. Note, error fields are not printed in intermediate restart files. Before proceeding, check that the error fields exist using <code>grep -a " : <" restart.<#>.1</code>

=== Performing the Adaptation Step ===
Within the 1-1-phParAdapt directory,
<nowiki>mkdir A1-phParAdapt
cd A1-phParAdapt
ln -s ../geom.smd
ln -s ../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
ln -s geom.sms parts.sms
ln -s ../1-procs_case/restart.1.1
ln -s restart.1.1 errors.1.1 </nowiki>

In the above step, it is assumed that the solution was interpolated onto restart.1.1. If that was not the case and the solution was advanced until a desired step was reached, replace restart.1.1 & errors.1.1 with restart.<#>.1 & errors.<#>.1 corresponding to the desired solution step number with error fields.

Create an adapt.inp file using [[#Adapt.inp_(adaptation step) |Adapt.inp_(adaptation step)]] as a template. Change the <code>timeStepNumber</code> parameter to correspond to the restart file that was linked into the current directory. Make sure the <code>adaptFlag</code> parameter is set to one and change additional threshold flags to control the result of the refinement.

Run phParAdapt using the mpirun command from the previous section. A directory corresponding to the <code>timeStepNumber</code> parameter should have been created with a mesh_parts.sms folder inside upon successful completion.

=== Generate Restart File for Refined Mesh ===
From within the A1-phParAdapt directory, execute the following (noting to replace <#> with the appropriate time step number):
<nowiki>cd <#>
mkdir 1-1-phParAdapt
cd 1-1-phParAdapt
ln -s ../../../geom.smd
ln -s ../../../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
cp ../../../adapt.inp .</nowiki>

In the above list of commands, the adapt.inp file from the non-adaptation step was copied into the current directory. Run phParAdapt using the same mpirun command as above to generate the 1-procs_case folder and the new restart file with the refined mesh. Again, since the solution migration feature is not working, the restart file will contain the initial solution and no error fields. A previously obtained solution can now be interpolated to the newly created one or PHASTA can be run in this directory to obtain a solution on the new mesh.

In order to perform a second (or more) adaptation step, create a restart file in the 1-procs_case folder in the current directory (either by interpolation or running PHASTA) and make sure it contains the error fields. Create a new directory <code>A2-phParAdapt</code> at this level and repeat the above steps. Be sure to use the time step number moving forward that is associated with the restart file just mentioned.

== File Examples ==

The file examples below are what were used for this specific example and may contain parameter flags that may need to be changed or are not used at all.

=== Adapt.inp (non-adaptation step) ===
<nowiki>numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 0
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.00001
AdaptOption 11
hmax 0.04
hmin 0.000052083
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 1
numTotParts 1
SolutionMigration 0
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 1
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

=== Adapt.inp (adaptation step) ===
<nowiki>eviscScal 0
pgrad 0
ub 1.45
lb 0.70
numSplit 5
sizeRatio 0.35 # confirmed that 0.25 does a 4x ref
numSmooth 0 #confirmed that this gets used
ratioThresh 0.75 # this is h_new/h_orig from isotropic days
localAdapt 1 # makes setLocal force adaptivity only around nodes set size
AnisoSimmetrix 4 # > 0 use Simmetrix, else ours 1 does largest length reduction by sizeRatio, 2 does both does medium and largest reduction by sizeRatio
coarsenMode 0 # disables coarsening see coarsenMode docs for 1, and 2
numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 1
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.5e-4
AdaptOption 11
hmax 1e6
hmin 1e-5
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 0
numTotParts 1
SolutionMigration 1
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 0
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

== Notes on Debugging ==
If debugging of phParAdapt is desired, replace the normal <code>mpirun</code> command with the following:

<code>mpirun -np 1 -tv /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi</code>

where the <code>-O</code> has been removed from the path to the executable to call the build version that was not optimized. Note that this requires phParAdapt to have been built at some point prior with the <code>Debug</code> flag.

== Notes on Folder Structure ==
Below is an overview of the folder structure to reference during the adaptation process:
# [ geom.smd, geom.sms, geom_nat.x_t ]
# 1-1-Chef
## 1-1-phParAdapt
### adapt.inp
### run_phParAdapt.sh
### 1-procs_case
#### geombc.dat.1
#### restart.<#>.1
### mesh_parts.sms
### A1-phParAdapt
#### restart.<#>.1
#### errors.<#>.1
#### [geom.smd, geom.sms, geom_nat.x_t, parts.sms]
#### adapt.inp
#### <#> ( step number directory )
##### mesh_parts.sms
##### 1-1-phParAdapt
###### [geom.smd, geom.sms, geom_nat.x_t]
###### adapt.inp
###### 1-procs_case ( restart files with adapted mesh )

== Removal of Extrusion Constraint ==
Riccardo has used a procedure in the past to remove the extrusion constraint from a mesh that enabled adaptation to be performed on extrusion-type elements.

The following directory contains an example of this process:
<code> /projects/tools/Models/NASAWingBodyJunction/RajMeshFine/Mesh/1-1-phParAdapt/RemExtrusion/ </code>

The executable that is called for removal of the constraint is a part of SCOREC-core, found at <code> <path/to/SCOREC-core/build/dir>/test/rm_extrusion </code> and the source code, in case edits are required, is found at <code> <path/to/SCOREC-core/source/dir>/core/test/rm_extrusion.cc </code>

PhParAdapt/Simmetrix

2021-04-28T20:18:04Z

Jopo9795: /* Initial Notes to User */

The following page provides details for performing mesh adaptation using the PhParAdapt tool with Simmetrix routines.

== Initial Notes to User ==
* The solution migration feature is currently broken and thus requires other means for transferring the solution to new meshes that are created from phParAdapt (e.g. solution interpolation in Paraview)
* To date, the steps described herein have only been tried with the following restrictions:
** Tetrahedral elements
*** Extruded element types apparently can be used but require removal of any extrusion constraints in the mesh. Below are some incomplete details regarding that process for reference. See [[#Removal_of_Extrusion_Constraint |Removal of Extrusion Constraint]].
** Serial case
** Geometries created using SimModeler7.0-190626
** Adaptation using 6th entry of the error field

== Adaptation Process ==

=== Creating Initial Restart and Error Files ===
Beginning inside of the 1-1-Chef directory,
<nowiki>mkdir 1-1-phParAdapt
cd 1-1-phParAdapt </nowiki>

Create soft links to the geom files above the 1-1-Chef directory,
<nowiki>ln -s ../../geom.smd
ln -s ../../geom.sms
ln -s ../../geom_nat.x_t </nowiki>

Create an <code> adapt.inp </code> using the template [[#Adapt.inp_(non-adaptation_step) | Adapt.inp (non-adaptation step)]]. Keep <code>adaptFlag</code> set to zero as the purpose of running phParAdapt here is to create restart.<timeStepNumber>.1 file associated with a new mesh format that can be used for adaptation.

Run phParAdapt in this directory, which for this specific example was:

<code>mpirun -np 1 /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi-O 2>&1 | tee phParAdapt.log</code>

After completion, the 1-procs_case and mesh_parts.sms directories should now be present. If a solution has already been generated for a different mesh on the same geometry, it can now be interpolated onto the new restart.<#>.1 file that was generated. If no solution exists, run Phasta using this restart file until the desired solution state has been achieved. Be sure to set <code>Print Error Indicators: True</code> in the solver.inp file so the error fields are saved in the final restart file. Note, error fields are not printed in intermediate restart files. Before proceeding, check that the error fields exist using <code>grep -a " : <" restart.<#>.1</code>

=== Performing the Adaptation Step ===
Within the 1-1-phParAdapt directory,
<nowiki>mkdir A1-phParAdapt
cd A1-phParAdapt
ln -s ../geom.smd
ln -s ../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
ln -s geom.sms parts.sms
ln -s ../1-procs_case/restart.1.1
ln -s restart.1.1 errors.1.1 </nowiki>

In the above step, it is assumed that the solution was interpolated onto restart.1.1. If that was not the case and the solution was advanced until a desired step was reached, replace restart.1.1 & errors.1.1 with restart.<#>.1 & errors.<#>.1 corresponding to the desired solution step number with error fields.

Create an adapt.inp file using [[#Adapt.inp_(adaptation step) |Adapt.inp_(adaptation step)]] as a template. Change the <code>timeStepNumber</code> parameter to correspond to the restart file that was linked into the current directory. Make sure the <code>adaptFlag</code> parameter is set to one and change additional threshold flags to control the result of the refinement.

Run phParAdapt using the mpirun command from the previous section. A directory corresponding to the <code>timeStepNumber</code> parameter should have been created with a mesh_parts.sms folder inside upon successful completion.

=== Generate Restart File for Refined Mesh ===
From within the A1-phParAdapt directory, execute the following (noting to replace <#> with the appropriate time step number):
<nowiki>cd <#>
mkdir 1-1-phParAdapt
cd 1-1-phParAdapt
ln -s ../../../geom.smd
ln -s ../../../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
cp ../../../adapt.inp .</nowiki>

In the above list of commands, the adapt.inp file from the non-adaptation step was copied into the current directory. Run phParAdapt using the same mpirun command as above to generate the 1-procs_case folder and the new restart file with the refined mesh. Again, since the solution migration feature is not working, the restart file will contain the initial solution and no error fields. A previously obtained solution can now be interpolated to the newly created one or PHASTA can be run in this directory to obtain a solution on the new mesh.

In order to perform a second (or more) adaptation step, create a restart file in the 1-procs_case folder in the current directory (either by interpolation or running PHASTA) and make sure it contains the error fields. Create a new directory <code>A2-phParAdapt</code> at this level and repeat the above steps. Be sure to use the time step number moving forward that is associated with the restart file just mentioned.

== File Examples ==

The file examples below are what were used for this specific example and may contain parameter flags that may need to be changed or are not used at all.

=== Adapt.inp (non-adaptation step) ===
<nowiki>numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 0
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.00001
AdaptOption 11
hmax 0.04
hmin 0.000052083
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 1
numTotParts 1
SolutionMigration 0
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 1
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

=== Adapt.inp (adaptation step) ===
<nowiki>eviscScal 0
pgrad 0
ub 1.45
lb 0.70
numSplit 5
sizeRatio 0.35 # confirmed that 0.25 does a 4x ref
numSmooth 0 #confirmed that this gets used
ratioThresh 0.75 # this is h_new/h_orig from isotropic days
localAdapt 1 # makes setLocal force adaptivity only around nodes set size
AnisoSimmetrix 4 # > 0 use Simmetrix, else ours 1 does largest length reduction by sizeRatio, 2 does both does medium and largest reduction by sizeRatio
coarsenMode 0 # disables coarsening see coarsenMode docs for 1, and 2
numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 1
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.5e-4
AdaptOption 11
hmax 1e6
hmin 1e-5
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 0
numTotParts 1
SolutionMigration 1
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 0
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

== Notes on Debugging ==
If debugging of phParAdapt is desired, replace the normal <code>mpirun</code> command with the following:

<code>mpirun -np 1 -tv /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi</code>

where the <code>-O</code> has been removed from the path to the executable to call the build version that was not optimized. Note that this requires phParAdapt to have been built at some point prior with the <code>Debug</code> flag.

== Notes on Folder Structure ==
Below is an overview of the folder structure to reference during the adaptation process:
# [ geom.smd, geom.sms, geom_nat.x_t ]
# 1-1-Chef
## 1-1-phParAdapt
### adapt.inp
### run_phParAdapt.sh
### 1-procs_case
#### geombc.dat.1
#### restart.<#>.1
### mesh_parts.sms
### A1-phParAdapt
#### restart.<#>.1
#### errors.<#>.1
#### [geom.smd, geom.sms, geom_nat.x_t, parts.sms]
#### adapt.inp
#### <#> ( step number directory )
##### mesh_parts.sms
##### 1-1-phParAdapt
###### [geom.smd, geom.sms, geom_nat.x_t]
###### adapt.inp
###### 1-procs_case ( restart files with adapted mesh )

== Removal of Extrusion Constraint ==
Riccardo has used a procedure in the past to remove the extrusion constraint from a mesh that enabled adaptation to be performed on extrusion-type elements.

The following directory contains an example of this process:
<code> /projects/tools/Models/NASAWingBodyJunction/RajMeshFine/Mesh/1-1-phParAdapt/RemExtrusion/ </code>

The executable that is called for removal of the constraint is a part of SCOREC-core, found at <code> <path/to/SCOREC-core/build/dir>/test/rm_extrusion </code> and the source code, in case edits are required, is found at <code> <path/to/SCOREC-core/source/dir>/core/test/rm_extrusion.cc </code>

PhParAdapt/Simmetrix

2021-04-28T20:17:41Z

Jopo9795: /* Initial Notes to User */

The following page provides details for performing mesh adaptation using the PhParAdapt tool with Simmetrix routines.

== Initial Notes to User ==
* The solution migration feature is currently broken and thus requires other means for transferring the solution to new meshes that are created from phParAdapt (e.g. solution interpolation in Paraview)
* To date, the steps described herein have only been tried with the following restrictions:
** Tetrahedral elements
*** Extruded element types apparently can be used but require removal of any extrusion constraints in the mesh. Below are some incomplete details regarding that process for reference. See [[#Removal_of_Extrusion_Constraint |Removal of Extrusion Constraint]].
** Serial case
** Geometries created using SimModeler7.0-190626
** Adaptation using 6th entry of errors field

== Adaptation Process ==

=== Creating Initial Restart and Error Files ===
Beginning inside of the 1-1-Chef directory,
<nowiki>mkdir 1-1-phParAdapt
cd 1-1-phParAdapt </nowiki>

Create soft links to the geom files above the 1-1-Chef directory,
<nowiki>ln -s ../../geom.smd
ln -s ../../geom.sms
ln -s ../../geom_nat.x_t </nowiki>

Create an <code> adapt.inp </code> using the template [[#Adapt.inp_(non-adaptation_step) | Adapt.inp (non-adaptation step)]]. Keep <code>adaptFlag</code> set to zero as the purpose of running phParAdapt here is to create restart.<timeStepNumber>.1 file associated with a new mesh format that can be used for adaptation.

Run phParAdapt in this directory, which for this specific example was:

<code>mpirun -np 1 /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi-O 2>&1 | tee phParAdapt.log</code>

After completion, the 1-procs_case and mesh_parts.sms directories should now be present. If a solution has already been generated for a different mesh on the same geometry, it can now be interpolated onto the new restart.<#>.1 file that was generated. If no solution exists, run Phasta using this restart file until the desired solution state has been achieved. Be sure to set <code>Print Error Indicators: True</code> in the solver.inp file so the error fields are saved in the final restart file. Note, error fields are not printed in intermediate restart files. Before proceeding, check that the error fields exist using <code>grep -a " : <" restart.<#>.1</code>

=== Performing the Adaptation Step ===
Within the 1-1-phParAdapt directory,
<nowiki>mkdir A1-phParAdapt
cd A1-phParAdapt
ln -s ../geom.smd
ln -s ../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
ln -s geom.sms parts.sms
ln -s ../1-procs_case/restart.1.1
ln -s restart.1.1 errors.1.1 </nowiki>

In the above step, it is assumed that the solution was interpolated onto restart.1.1. If that was not the case and the solution was advanced until a desired step was reached, replace restart.1.1 & errors.1.1 with restart.<#>.1 & errors.<#>.1 corresponding to the desired solution step number with error fields.

Create an adapt.inp file using [[#Adapt.inp_(adaptation step) |Adapt.inp_(adaptation step)]] as a template. Change the <code>timeStepNumber</code> parameter to correspond to the restart file that was linked into the current directory. Make sure the <code>adaptFlag</code> parameter is set to one and change additional threshold flags to control the result of the refinement.

Run phParAdapt using the mpirun command from the previous section. A directory corresponding to the <code>timeStepNumber</code> parameter should have been created with a mesh_parts.sms folder inside upon successful completion.

=== Generate Restart File for Refined Mesh ===
From within the A1-phParAdapt directory, execute the following (noting to replace <#> with the appropriate time step number):
<nowiki>cd <#>
mkdir 1-1-phParAdapt
cd 1-1-phParAdapt
ln -s ../../../geom.smd
ln -s ../../../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
cp ../../../adapt.inp .</nowiki>

In the above list of commands, the adapt.inp file from the non-adaptation step was copied into the current directory. Run phParAdapt using the same mpirun command as above to generate the 1-procs_case folder and the new restart file with the refined mesh. Again, since the solution migration feature is not working, the restart file will contain the initial solution and no error fields. A previously obtained solution can now be interpolated to the newly created one or PHASTA can be run in this directory to obtain a solution on the new mesh.

In order to perform a second (or more) adaptation step, create a restart file in the 1-procs_case folder in the current directory (either by interpolation or running PHASTA) and make sure it contains the error fields. Create a new directory <code>A2-phParAdapt</code> at this level and repeat the above steps. Be sure to use the time step number moving forward that is associated with the restart file just mentioned.

== File Examples ==

The file examples below are what were used for this specific example and may contain parameter flags that may need to be changed or are not used at all.

=== Adapt.inp (non-adaptation step) ===
<nowiki>numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 0
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.00001
AdaptOption 11
hmax 0.04
hmin 0.000052083
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 1
numTotParts 1
SolutionMigration 0
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 1
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

=== Adapt.inp (adaptation step) ===
<nowiki>eviscScal 0
pgrad 0
ub 1.45
lb 0.70
numSplit 5
sizeRatio 0.35 # confirmed that 0.25 does a 4x ref
numSmooth 0 #confirmed that this gets used
ratioThresh 0.75 # this is h_new/h_orig from isotropic days
localAdapt 1 # makes setLocal force adaptivity only around nodes set size
AnisoSimmetrix 4 # > 0 use Simmetrix, else ours 1 does largest length reduction by sizeRatio, 2 does both does medium and largest reduction by sizeRatio
coarsenMode 0 # disables coarsening see coarsenMode docs for 1, and 2
numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 1
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.5e-4
AdaptOption 11
hmax 1e6
hmin 1e-5
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 0
numTotParts 1
SolutionMigration 1
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 0
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

== Notes on Debugging ==
If debugging of phParAdapt is desired, replace the normal <code>mpirun</code> command with the following:

<code>mpirun -np 1 -tv /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi</code>

where the <code>-O</code> has been removed from the path to the executable to call the build version that was not optimized. Note that this requires phParAdapt to have been built at some point prior with the <code>Debug</code> flag.

== Notes on Folder Structure ==
Below is an overview of the folder structure to reference during the adaptation process:
# [ geom.smd, geom.sms, geom_nat.x_t ]
# 1-1-Chef
## 1-1-phParAdapt
### adapt.inp
### run_phParAdapt.sh
### 1-procs_case
#### geombc.dat.1
#### restart.<#>.1
### mesh_parts.sms
### A1-phParAdapt
#### restart.<#>.1
#### errors.<#>.1
#### [geom.smd, geom.sms, geom_nat.x_t, parts.sms]
#### adapt.inp
#### <#> ( step number directory )
##### mesh_parts.sms
##### 1-1-phParAdapt
###### [geom.smd, geom.sms, geom_nat.x_t]
###### adapt.inp
###### 1-procs_case ( restart files with adapted mesh )

== Removal of Extrusion Constraint ==
Riccardo has used a procedure in the past to remove the extrusion constraint from a mesh that enabled adaptation to be performed on extrusion-type elements.

The following directory contains an example of this process:
<code> /projects/tools/Models/NASAWingBodyJunction/RajMeshFine/Mesh/1-1-phParAdapt/RemExtrusion/ </code>

The executable that is called for removal of the constraint is a part of SCOREC-core, found at <code> <path/to/SCOREC-core/build/dir>/test/rm_extrusion </code> and the source code, in case edits are required, is found at <code> <path/to/SCOREC-core/source/dir>/core/test/rm_extrusion.cc </code>

PhParAdapt/Simmetrix

2021-04-28T20:17:10Z

Jopo9795: /* Generate Restart File for Refined Mesh */

The following page provides details for performing mesh adaptation using the PhParAdapt tool with Simmetrix routines.

== Initial Notes to User ==
* The solution migration feature is currently broken and thus requires other means for transferring the solution to new meshes that are created from PhParAdapt (e.g. solution interpolation in Paraview)
* To date, the steps described herein have only been tried with the following restrictions:
** Tetrahedral elements
*** Extruded element types apparently can be used but require removal of any extrusion constraints in the mesh. Below are some incomplete details regarding that process for reference. See [[#Removal_of_Extrusion_Constraint |Removal of Extrusion Constraint]].
** Serial case
** Geometries created using SimModeler7.0-190626
** Adaptation using 6th entry of errors field

== Adaptation Process ==

=== Creating Initial Restart and Error Files ===
Beginning inside of the 1-1-Chef directory,
<nowiki>mkdir 1-1-phParAdapt
cd 1-1-phParAdapt </nowiki>

Create soft links to the geom files above the 1-1-Chef directory,
<nowiki>ln -s ../../geom.smd
ln -s ../../geom.sms
ln -s ../../geom_nat.x_t </nowiki>

Create an <code> adapt.inp </code> using the template [[#Adapt.inp_(non-adaptation_step) | Adapt.inp (non-adaptation step)]]. Keep <code>adaptFlag</code> set to zero as the purpose of running phParAdapt here is to create restart.<timeStepNumber>.1 file associated with a new mesh format that can be used for adaptation.

Run phParAdapt in this directory, which for this specific example was:

<code>mpirun -np 1 /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi-O 2>&1 | tee phParAdapt.log</code>

After completion, the 1-procs_case and mesh_parts.sms directories should now be present. If a solution has already been generated for a different mesh on the same geometry, it can now be interpolated onto the new restart.<#>.1 file that was generated. If no solution exists, run Phasta using this restart file until the desired solution state has been achieved. Be sure to set <code>Print Error Indicators: True</code> in the solver.inp file so the error fields are saved in the final restart file. Note, error fields are not printed in intermediate restart files. Before proceeding, check that the error fields exist using <code>grep -a " : <" restart.<#>.1</code>

=== Performing the Adaptation Step ===
Within the 1-1-phParAdapt directory,
<nowiki>mkdir A1-phParAdapt
cd A1-phParAdapt
ln -s ../geom.smd
ln -s ../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
ln -s geom.sms parts.sms
ln -s ../1-procs_case/restart.1.1
ln -s restart.1.1 errors.1.1 </nowiki>

In the above step, it is assumed that the solution was interpolated onto restart.1.1. If that was not the case and the solution was advanced until a desired step was reached, replace restart.1.1 & errors.1.1 with restart.<#>.1 & errors.<#>.1 corresponding to the desired solution step number with error fields.

Create an adapt.inp file using [[#Adapt.inp_(adaptation step) |Adapt.inp_(adaptation step)]] as a template. Change the <code>timeStepNumber</code> parameter to correspond to the restart file that was linked into the current directory. Make sure the <code>adaptFlag</code> parameter is set to one and change additional threshold flags to control the result of the refinement.

Run phParAdapt using the mpirun command from the previous section. A directory corresponding to the <code>timeStepNumber</code> parameter should have been created with a mesh_parts.sms folder inside upon successful completion.

=== Generate Restart File for Refined Mesh ===
From within the A1-phParAdapt directory, execute the following (noting to replace <#> with the appropriate time step number):
<nowiki>cd <#>
mkdir 1-1-phParAdapt
cd 1-1-phParAdapt
ln -s ../../../geom.smd
ln -s ../../../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
cp ../../../adapt.inp .</nowiki>

In the above list of commands, the adapt.inp file from the non-adaptation step was copied into the current directory. Run phParAdapt using the same mpirun command as above to generate the 1-procs_case folder and the new restart file with the refined mesh. Again, since the solution migration feature is not working, the restart file will contain the initial solution and no error fields. A previously obtained solution can now be interpolated to the newly created one or PHASTA can be run in this directory to obtain a solution on the new mesh.

In order to perform a second (or more) adaptation step, create a restart file in the 1-procs_case folder in the current directory (either by interpolation or running PHASTA) and make sure it contains the error fields. Create a new directory <code>A2-phParAdapt</code> at this level and repeat the above steps. Be sure to use the time step number moving forward that is associated with the restart file just mentioned.

== File Examples ==

The file examples below are what were used for this specific example and may contain parameter flags that may need to be changed or are not used at all.

=== Adapt.inp (non-adaptation step) ===
<nowiki>numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 0
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.00001
AdaptOption 11
hmax 0.04
hmin 0.000052083
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 1
numTotParts 1
SolutionMigration 0
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 1
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

=== Adapt.inp (adaptation step) ===
<nowiki>eviscScal 0
pgrad 0
ub 1.45
lb 0.70
numSplit 5
sizeRatio 0.35 # confirmed that 0.25 does a 4x ref
numSmooth 0 #confirmed that this gets used
ratioThresh 0.75 # this is h_new/h_orig from isotropic days
localAdapt 1 # makes setLocal force adaptivity only around nodes set size
AnisoSimmetrix 4 # > 0 use Simmetrix, else ours 1 does largest length reduction by sizeRatio, 2 does both does medium and largest reduction by sizeRatio
coarsenMode 0 # disables coarsening see coarsenMode docs for 1, and 2
numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 1
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.5e-4
AdaptOption 11
hmax 1e6
hmin 1e-5
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 0
numTotParts 1
SolutionMigration 1
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 0
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

== Notes on Debugging ==
If debugging of phParAdapt is desired, replace the normal <code>mpirun</code> command with the following:

<code>mpirun -np 1 -tv /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi</code>

where the <code>-O</code> has been removed from the path to the executable to call the build version that was not optimized. Note that this requires phParAdapt to have been built at some point prior with the <code>Debug</code> flag.

== Notes on Folder Structure ==
Below is an overview of the folder structure to reference during the adaptation process:
# [ geom.smd, geom.sms, geom_nat.x_t ]
# 1-1-Chef
## 1-1-phParAdapt
### adapt.inp
### run_phParAdapt.sh
### 1-procs_case
#### geombc.dat.1
#### restart.<#>.1
### mesh_parts.sms
### A1-phParAdapt
#### restart.<#>.1
#### errors.<#>.1
#### [geom.smd, geom.sms, geom_nat.x_t, parts.sms]
#### adapt.inp
#### <#> ( step number directory )
##### mesh_parts.sms
##### 1-1-phParAdapt
###### [geom.smd, geom.sms, geom_nat.x_t]
###### adapt.inp
###### 1-procs_case ( restart files with adapted mesh )

== Removal of Extrusion Constraint ==
Riccardo has used a procedure in the past to remove the extrusion constraint from a mesh that enabled adaptation to be performed on extrusion-type elements.

The following directory contains an example of this process:
<code> /projects/tools/Models/NASAWingBodyJunction/RajMeshFine/Mesh/1-1-phParAdapt/RemExtrusion/ </code>

The executable that is called for removal of the constraint is a part of SCOREC-core, found at <code> <path/to/SCOREC-core/build/dir>/test/rm_extrusion </code> and the source code, in case edits are required, is found at <code> <path/to/SCOREC-core/source/dir>/core/test/rm_extrusion.cc </code>

PhParAdapt/Simmetrix

2021-04-28T20:08:29Z

Jopo9795: /* Performing the Adaptation Step */

The following page provides details for performing mesh adaptation using the PhParAdapt tool with Simmetrix routines.

== Initial Notes to User ==
* The solution migration feature is currently broken and thus requires other means for transferring the solution to new meshes that are created from PhParAdapt (e.g. solution interpolation in Paraview)
* To date, the steps described herein have only been tried with the following restrictions:
** Tetrahedral elements
*** Extruded element types apparently can be used but require removal of any extrusion constraints in the mesh. Below are some incomplete details regarding that process for reference. See [[#Removal_of_Extrusion_Constraint |Removal of Extrusion Constraint]].
** Serial case
** Geometries created using SimModeler7.0-190626
** Adaptation using 6th entry of errors field

== Adaptation Process ==

=== Creating Initial Restart and Error Files ===
Beginning inside of the 1-1-Chef directory,
<nowiki>mkdir 1-1-phParAdapt
cd 1-1-phParAdapt </nowiki>

Create soft links to the geom files above the 1-1-Chef directory,
<nowiki>ln -s ../../geom.smd
ln -s ../../geom.sms
ln -s ../../geom_nat.x_t </nowiki>

Create an <code> adapt.inp </code> using the template [[#Adapt.inp_(non-adaptation_step) | Adapt.inp (non-adaptation step)]]. Keep <code>adaptFlag</code> set to zero as the purpose of running phParAdapt here is to create restart.<timeStepNumber>.1 file associated with a new mesh format that can be used for adaptation.

Run phParAdapt in this directory, which for this specific example was:

<code>mpirun -np 1 /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi-O 2>&1 | tee phParAdapt.log</code>

After completion, the 1-procs_case and mesh_parts.sms directories should now be present. If a solution has already been generated for a different mesh on the same geometry, it can now be interpolated onto the new restart.<#>.1 file that was generated. If no solution exists, run Phasta using this restart file until the desired solution state has been achieved. Be sure to set <code>Print Error Indicators: True</code> in the solver.inp file so the error fields are saved in the final restart file. Note, error fields are not printed in intermediate restart files. Before proceeding, check that the error fields exist using <code>grep -a " : <" restart.<#>.1</code>

=== Performing the Adaptation Step ===
Within the 1-1-phParAdapt directory,
<nowiki>mkdir A1-phParAdapt
cd A1-phParAdapt
ln -s ../geom.smd
ln -s ../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
ln -s geom.sms parts.sms
ln -s ../1-procs_case/restart.1.1
ln -s restart.1.1 errors.1.1 </nowiki>

In the above step, it is assumed that the solution was interpolated onto restart.1.1. If that was not the case and the solution was advanced until a desired step was reached, replace restart.1.1 & errors.1.1 with restart.<#>.1 & errors.<#>.1 corresponding to the desired solution step number with error fields.

Create an adapt.inp file using [[#Adapt.inp_(adaptation step) |Adapt.inp_(adaptation step)]] as a template. Change the <code>timeStepNumber</code> parameter to correspond to the restart file that was linked into the current directory. Make sure the <code>adaptFlag</code> parameter is set to one and change additional threshold flags to control the result of the refinement.

Run phParAdapt using the mpirun command from the previous section. A directory corresponding to the <code>timeStepNumber</code> parameter should have been created with a mesh_parts.sms folder inside upon successful completion.

=== Generate Restart File for Refined Mesh ===
From within the A1-phParAdapt directory, execute the following (noting to replace <#> with the appropriate time step number):
<nowiki>cd <#>
mkdir 1-1-phParAdapt
cd 1-1-phParAdapt
ln -s ../../../geom.smd
ln -s ../../../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
cp ../../../adapt.inp .</nowiki>

Note that in the above, the adapt.inp file from the non-adaptation step was copied into the current directory.

== File Examples ==

The file examples below are what were used for this specific example and may contain parameter flags that may need to be changed or are not used at all.

=== Adapt.inp (non-adaptation step) ===
<nowiki>numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 0
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.00001
AdaptOption 11
hmax 0.04
hmin 0.000052083
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 1
numTotParts 1
SolutionMigration 0
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 1
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

=== Adapt.inp (adaptation step) ===
<nowiki>eviscScal 0
pgrad 0
ub 1.45
lb 0.70
numSplit 5
sizeRatio 0.35 # confirmed that 0.25 does a 4x ref
numSmooth 0 #confirmed that this gets used
ratioThresh 0.75 # this is h_new/h_orig from isotropic days
localAdapt 1 # makes setLocal force adaptivity only around nodes set size
AnisoSimmetrix 4 # > 0 use Simmetrix, else ours 1 does largest length reduction by sizeRatio, 2 does both does medium and largest reduction by sizeRatio
coarsenMode 0 # disables coarsening see coarsenMode docs for 1, and 2
numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 1
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.5e-4
AdaptOption 11
hmax 1e6
hmin 1e-5
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 0
numTotParts 1
SolutionMigration 1
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 0
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

== Notes on Debugging ==
If debugging of phParAdapt is desired, replace the normal <code>mpirun</code> command with the following:

<code>mpirun -np 1 -tv /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi</code>

where the <code>-O</code> has been removed from the path to the executable to call the build version that was not optimized. Note that this requires phParAdapt to have been built at some point prior with the <code>Debug</code> flag.

== Notes on Folder Structure ==
Below is an overview of the folder structure to reference during the adaptation process:
# [ geom.smd, geom.sms, geom_nat.x_t ]
# 1-1-Chef
## 1-1-phParAdapt
### adapt.inp
### run_phParAdapt.sh
### 1-procs_case
#### geombc.dat.1
#### restart.<#>.1
### mesh_parts.sms
### A1-phParAdapt
#### restart.<#>.1
#### errors.<#>.1
#### [geom.smd, geom.sms, geom_nat.x_t, parts.sms]
#### adapt.inp
#### <#> ( step number directory )
##### mesh_parts.sms
##### 1-1-phParAdapt
###### [geom.smd, geom.sms, geom_nat.x_t]
###### adapt.inp
###### 1-procs_case ( restart files with adapted mesh )

== Removal of Extrusion Constraint ==
Riccardo has used a procedure in the past to remove the extrusion constraint from a mesh that enabled adaptation to be performed on extrusion-type elements.

The following directory contains an example of this process:
<code> /projects/tools/Models/NASAWingBodyJunction/RajMeshFine/Mesh/1-1-phParAdapt/RemExtrusion/ </code>

The executable that is called for removal of the constraint is a part of SCOREC-core, found at <code> <path/to/SCOREC-core/build/dir>/test/rm_extrusion </code> and the source code, in case edits are required, is found at <code> <path/to/SCOREC-core/source/dir>/core/test/rm_extrusion.cc </code>

PhParAdapt/Simmetrix

2021-04-28T20:07:09Z

Jopo9795: /* Notes on Debugging */

The following page provides details for performing mesh adaptation using the PhParAdapt tool with Simmetrix routines.

== Initial Notes to User ==
* The solution migration feature is currently broken and thus requires other means for transferring the solution to new meshes that are created from PhParAdapt (e.g. solution interpolation in Paraview)
* To date, the steps described herein have only been tried with the following restrictions:
** Tetrahedral elements
*** Extruded element types apparently can be used but require removal of any extrusion constraints in the mesh. Below are some incomplete details regarding that process for reference. See [[#Removal_of_Extrusion_Constraint |Removal of Extrusion Constraint]].
** Serial case
** Geometries created using SimModeler7.0-190626
** Adaptation using 6th entry of errors field

== Adaptation Process ==

=== Creating Initial Restart and Error Files ===
Beginning inside of the 1-1-Chef directory,
<nowiki>mkdir 1-1-phParAdapt
cd 1-1-phParAdapt </nowiki>

Create soft links to the geom files above the 1-1-Chef directory,
<nowiki>ln -s ../../geom.smd
ln -s ../../geom.sms
ln -s ../../geom_nat.x_t </nowiki>

Create an <code> adapt.inp </code> using the template [[#Adapt.inp_(non-adaptation_step) | Adapt.inp (non-adaptation step)]]. Keep <code>adaptFlag</code> set to zero as the purpose of running phParAdapt here is to create restart.<timeStepNumber>.1 file associated with a new mesh format that can be used for adaptation.

Run phParAdapt in this directory, which for this specific example was:

<code>mpirun -np 1 /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi-O 2>&1 | tee phParAdapt.log</code>

After completion, the 1-procs_case and mesh_parts.sms directories should now be present. If a solution has already been generated for a different mesh on the same geometry, it can now be interpolated onto the new restart.<#>.1 file that was generated. If no solution exists, run Phasta using this restart file until the desired solution state has been achieved. Be sure to set <code>Print Error Indicators: True</code> in the solver.inp file so the error fields are saved in the final restart file. Note, error fields are not printed in intermediate restart files. Before proceeding, check that the error fields exist using <code>grep -a " : <" restart.<#>.1</code>

=== Performing the Adaptation Step ===
Within the 1-1-phParAdapt directory,
<nowiki>mkdir A1-phParAdapt
cd A1-phParAdapt
ln -s ../geom.smd
ln -s ../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
ln -s geom.sms parts.sms
ln -s ../1-procs_case/restart.1.1
ln -s restart.1.1 errors.1.1 </nowiki>

In the above step, it is assumed that the solution was interpolated onto restart.1.1. If that was not the case and the solution was advanced until a desired step was reached, replace restart.1.1 & errors.1.1 with restart.<#>.1 & errors.<#>.1 corresponding to the desired solution step number with error fields.

Create an adapt.inp file using [[#Adapt.inp_(adaptation step) |Adapt.inp_(adaptation step)]] as a template. Change the <code>timeStepNumber</code> parameter to correspond to the restart file that was linked into the current directory. Make sure the <code>adaptFlag</code> parameter is set to one and change additional threshold flags to control the result of the refinement.

Run phParAdapt with the following command,

<code>mpirun -np 1 /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi-O 2>&1 | tee phParAdapt.log</code>

A directory corresponding to the <code>timeStepNumber</code> parameter should have been created with a mesh_parts.sms folder inside.

=== Generate Restart File for Refined Mesh ===
From within the A1-phParAdapt directory, execute the following (noting to replace <#> with the appropriate time step number):
<nowiki>cd <#>
mkdir 1-1-phParAdapt
cd 1-1-phParAdapt
ln -s ../../../geom.smd
ln -s ../../../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
cp ../../../adapt.inp .</nowiki>

Note that in the above, the adapt.inp file from the non-adaptation step was copied into the current directory.

== File Examples ==

The file examples below are what were used for this specific example and may contain parameter flags that may need to be changed or are not used at all.

=== Adapt.inp (non-adaptation step) ===
<nowiki>numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 0
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.00001
AdaptOption 11
hmax 0.04
hmin 0.000052083
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 1
numTotParts 1
SolutionMigration 0
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 1
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

=== Adapt.inp (adaptation step) ===
<nowiki>eviscScal 0
pgrad 0
ub 1.45
lb 0.70
numSplit 5
sizeRatio 0.35 # confirmed that 0.25 does a 4x ref
numSmooth 0 #confirmed that this gets used
ratioThresh 0.75 # this is h_new/h_orig from isotropic days
localAdapt 1 # makes setLocal force adaptivity only around nodes set size
AnisoSimmetrix 4 # > 0 use Simmetrix, else ours 1 does largest length reduction by sizeRatio, 2 does both does medium and largest reduction by sizeRatio
coarsenMode 0 # disables coarsening see coarsenMode docs for 1, and 2
numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 1
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.5e-4
AdaptOption 11
hmax 1e6
hmin 1e-5
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 0
numTotParts 1
SolutionMigration 1
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 0
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

== Notes on Debugging ==
If debugging of phParAdapt is desired, replace the normal <code>mpirun</code> command with the following:

<code>mpirun -np 1 -tv /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi</code>

where the <code>-O</code> has been removed from the path to the executable to call the build version that was not optimized. Note that this requires phParAdapt to have been built at some point prior with the <code>Debug</code> flag.

== Notes on Folder Structure ==
Below is an overview of the folder structure to reference during the adaptation process:
# [ geom.smd, geom.sms, geom_nat.x_t ]
# 1-1-Chef
## 1-1-phParAdapt
### adapt.inp
### run_phParAdapt.sh
### 1-procs_case
#### geombc.dat.1
#### restart.<#>.1
### mesh_parts.sms
### A1-phParAdapt
#### restart.<#>.1
#### errors.<#>.1
#### [geom.smd, geom.sms, geom_nat.x_t, parts.sms]
#### adapt.inp
#### <#> ( step number directory )
##### mesh_parts.sms
##### 1-1-phParAdapt
###### [geom.smd, geom.sms, geom_nat.x_t]
###### adapt.inp
###### 1-procs_case ( restart files with adapted mesh )

== Removal of Extrusion Constraint ==
Riccardo has used a procedure in the past to remove the extrusion constraint from a mesh that enabled adaptation to be performed on extrusion-type elements.

The following directory contains an example of this process:
<code> /projects/tools/Models/NASAWingBodyJunction/RajMeshFine/Mesh/1-1-phParAdapt/RemExtrusion/ </code>

The executable that is called for removal of the constraint is a part of SCOREC-core, found at <code> <path/to/SCOREC-core/build/dir>/test/rm_extrusion </code> and the source code, in case edits are required, is found at <code> <path/to/SCOREC-core/source/dir>/core/test/rm_extrusion.cc </code>

PhParAdapt/Simmetrix

2021-04-28T20:06:23Z

Jopo9795: /* File Examples */

The following page provides details for performing mesh adaptation using the PhParAdapt tool with Simmetrix routines.

== Initial Notes to User ==
* The solution migration feature is currently broken and thus requires other means for transferring the solution to new meshes that are created from PhParAdapt (e.g. solution interpolation in Paraview)
* To date, the steps described herein have only been tried with the following restrictions:
** Tetrahedral elements
*** Extruded element types apparently can be used but require removal of any extrusion constraints in the mesh. Below are some incomplete details regarding that process for reference. See [[#Removal_of_Extrusion_Constraint |Removal of Extrusion Constraint]].
** Serial case
** Geometries created using SimModeler7.0-190626
** Adaptation using 6th entry of errors field

== Adaptation Process ==

=== Creating Initial Restart and Error Files ===
Beginning inside of the 1-1-Chef directory,
<nowiki>mkdir 1-1-phParAdapt
cd 1-1-phParAdapt </nowiki>

Create soft links to the geom files above the 1-1-Chef directory,
<nowiki>ln -s ../../geom.smd
ln -s ../../geom.sms
ln -s ../../geom_nat.x_t </nowiki>

Create an <code> adapt.inp </code> using the template [[#Adapt.inp_(non-adaptation_step) | Adapt.inp (non-adaptation step)]]. Keep <code>adaptFlag</code> set to zero as the purpose of running phParAdapt here is to create restart.<timeStepNumber>.1 file associated with a new mesh format that can be used for adaptation.

Run phParAdapt in this directory, which for this specific example was:

<code>mpirun -np 1 /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi-O 2>&1 | tee phParAdapt.log</code>

After completion, the 1-procs_case and mesh_parts.sms directories should now be present. If a solution has already been generated for a different mesh on the same geometry, it can now be interpolated onto the new restart.<#>.1 file that was generated. If no solution exists, run Phasta using this restart file until the desired solution state has been achieved. Be sure to set <code>Print Error Indicators: True</code> in the solver.inp file so the error fields are saved in the final restart file. Note, error fields are not printed in intermediate restart files. Before proceeding, check that the error fields exist using <code>grep -a " : <" restart.<#>.1</code>

=== Performing the Adaptation Step ===
Within the 1-1-phParAdapt directory,
<nowiki>mkdir A1-phParAdapt
cd A1-phParAdapt
ln -s ../geom.smd
ln -s ../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
ln -s geom.sms parts.sms
ln -s ../1-procs_case/restart.1.1
ln -s restart.1.1 errors.1.1 </nowiki>

In the above step, it is assumed that the solution was interpolated onto restart.1.1. If that was not the case and the solution was advanced until a desired step was reached, replace restart.1.1 & errors.1.1 with restart.<#>.1 & errors.<#>.1 corresponding to the desired solution step number with error fields.

Create an adapt.inp file using [[#Adapt.inp_(adaptation step) |Adapt.inp_(adaptation step)]] as a template. Change the <code>timeStepNumber</code> parameter to correspond to the restart file that was linked into the current directory. Make sure the <code>adaptFlag</code> parameter is set to one and change additional threshold flags to control the result of the refinement.

Run phParAdapt with the following command,

<code>mpirun -np 1 /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi-O 2>&1 | tee phParAdapt.log</code>

A directory corresponding to the <code>timeStepNumber</code> parameter should have been created with a mesh_parts.sms folder inside.

=== Generate Restart File for Refined Mesh ===
From within the A1-phParAdapt directory, execute the following (noting to replace <#> with the appropriate time step number):
<nowiki>cd <#>
mkdir 1-1-phParAdapt
cd 1-1-phParAdapt
ln -s ../../../geom.smd
ln -s ../../../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
cp ../../../adapt.inp .</nowiki>

Note that in the above, the adapt.inp file from the non-adaptation step was copied into the current directory.

== File Examples ==

The file examples below are what were used for this specific example and may contain parameter flags that may need to be changed or are not used at all.

=== Adapt.inp (non-adaptation step) ===
<nowiki>numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 0
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.00001
AdaptOption 11
hmax 0.04
hmin 0.000052083
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 1
numTotParts 1
SolutionMigration 0
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 1
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

=== Adapt.inp (adaptation step) ===
<nowiki>eviscScal 0
pgrad 0
ub 1.45
lb 0.70
numSplit 5
sizeRatio 0.35 # confirmed that 0.25 does a 4x ref
numSmooth 0 #confirmed that this gets used
ratioThresh 0.75 # this is h_new/h_orig from isotropic days
localAdapt 1 # makes setLocal force adaptivity only around nodes set size
AnisoSimmetrix 4 # > 0 use Simmetrix, else ours 1 does largest length reduction by sizeRatio, 2 does both does medium and largest reduction by sizeRatio
coarsenMode 0 # disables coarsening see coarsenMode docs for 1, and 2
numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 1
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.5e-4
AdaptOption 11
hmax 1e6
hmin 1e-5
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 0
numTotParts 1
SolutionMigration 1
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 0
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

== Notes on Debugging ==
If debugging of phParAdapt is desired, replace the normal <code>mpirun</code> command with the following:

<code>mpirun -np 1 -tv /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi</code>

where the <code>-O</code> has been removed to call the build version that was not optimized. Note that this requires phParAdapt to have been built at some point prior with the <code>Debug</code> flag.

== Notes on Folder Structure ==
Below is an overview of the folder structure to reference during the adaptation process:
# [ geom.smd, geom.sms, geom_nat.x_t ]
# 1-1-Chef
## 1-1-phParAdapt
### adapt.inp
### run_phParAdapt.sh
### 1-procs_case
#### geombc.dat.1
#### restart.<#>.1
### mesh_parts.sms
### A1-phParAdapt
#### restart.<#>.1
#### errors.<#>.1
#### [geom.smd, geom.sms, geom_nat.x_t, parts.sms]
#### adapt.inp
#### <#> ( step number directory )
##### mesh_parts.sms
##### 1-1-phParAdapt
###### [geom.smd, geom.sms, geom_nat.x_t]
###### adapt.inp
###### 1-procs_case ( restart files with adapted mesh )

== Removal of Extrusion Constraint ==
Riccardo has used a procedure in the past to remove the extrusion constraint from a mesh that enabled adaptation to be performed on extrusion-type elements.

The following directory contains an example of this process:
<code> /projects/tools/Models/NASAWingBodyJunction/RajMeshFine/Mesh/1-1-phParAdapt/RemExtrusion/ </code>

The executable that is called for removal of the constraint is a part of SCOREC-core, found at <code> <path/to/SCOREC-core/build/dir>/test/rm_extrusion </code> and the source code, in case edits are required, is found at <code> <path/to/SCOREC-core/source/dir>/core/test/rm_extrusion.cc </code>

PhParAdapt/Simmetrix

2021-04-28T20:01:49Z

Jopo9795: /* Creating Initial Restart and Error Files */

The following page provides details for performing mesh adaptation using the PhParAdapt tool with Simmetrix routines.

== Initial Notes to User ==
* The solution migration feature is currently broken and thus requires other means for transferring the solution to new meshes that are created from PhParAdapt (e.g. solution interpolation in Paraview)
* To date, the steps described herein have only been tried with the following restrictions:
** Tetrahedral elements
*** Extruded element types apparently can be used but require removal of any extrusion constraints in the mesh. Below are some incomplete details regarding that process for reference. See [[#Removal_of_Extrusion_Constraint |Removal of Extrusion Constraint]].
** Serial case
** Geometries created using SimModeler7.0-190626
** Adaptation using 6th entry of errors field

== Adaptation Process ==

=== Creating Initial Restart and Error Files ===
Beginning inside of the 1-1-Chef directory,
<nowiki>mkdir 1-1-phParAdapt
cd 1-1-phParAdapt </nowiki>

Create soft links to the geom files above the 1-1-Chef directory,
<nowiki>ln -s ../../geom.smd
ln -s ../../geom.sms
ln -s ../../geom_nat.x_t </nowiki>

Create an <code> adapt.inp </code> using the template [[#Adapt.inp_(non-adaptation_step) | Adapt.inp (non-adaptation step)]]. Keep <code>adaptFlag</code> set to zero as the purpose of running phParAdapt here is to create restart.<timeStepNumber>.1 file associated with a new mesh format that can be used for adaptation.

Run phParAdapt in this directory, which for this specific example was:

<code>mpirun -np 1 /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi-O 2>&1 | tee phParAdapt.log</code>

After completion, the 1-procs_case and mesh_parts.sms directories should now be present. If a solution has already been generated for a different mesh on the same geometry, it can now be interpolated onto the new restart.<#>.1 file that was generated. If no solution exists, run Phasta using this restart file until the desired solution state has been achieved. Be sure to set <code>Print Error Indicators: True</code> in the solver.inp file so the error fields are saved in the final restart file. Note, error fields are not printed in intermediate restart files. Before proceeding, check that the error fields exist using <code>grep -a " : <" restart.<#>.1</code>

=== Performing the Adaptation Step ===
Within the 1-1-phParAdapt directory,
<nowiki>mkdir A1-phParAdapt
cd A1-phParAdapt
ln -s ../geom.smd
ln -s ../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
ln -s geom.sms parts.sms
ln -s ../1-procs_case/restart.1.1
ln -s restart.1.1 errors.1.1 </nowiki>

In the above step, it is assumed that the solution was interpolated onto restart.1.1. If that was not the case and the solution was advanced until a desired step was reached, replace restart.1.1 & errors.1.1 with restart.<#>.1 & errors.<#>.1 corresponding to the desired solution step number with error fields.

Create an adapt.inp file using [[#Adapt.inp_(adaptation step) |Adapt.inp_(adaptation step)]] as a template. Change the <code>timeStepNumber</code> parameter to correspond to the restart file that was linked into the current directory. Make sure the <code>adaptFlag</code> parameter is set to one and change additional threshold flags to control the result of the refinement.

Run phParAdapt with the following command,

<code>mpirun -np 1 /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi-O 2>&1 | tee phParAdapt.log</code>

A directory corresponding to the <code>timeStepNumber</code> parameter should have been created with a mesh_parts.sms folder inside.

=== Generate Restart File for Refined Mesh ===
From within the A1-phParAdapt directory, execute the following (noting to replace <#> with the appropriate time step number):
<nowiki>cd <#>
mkdir 1-1-phParAdapt
cd 1-1-phParAdapt
ln -s ../../../geom.smd
ln -s ../../../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
cp ../../../adapt.inp .</nowiki>

Note that in the above, the adapt.inp file from the non-adaptation step was copied into the current directory.

== File Examples ==

The file examples below are what were used for this specific example and may contain parameter flags that may need to be changed or are not used at all.

=== Adapt.inp (non-adaptation step) ===
<nowiki>numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 0
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.00001
AdaptOption 11
hmax 0.04
hmin 0.000052083
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 1
numTotParts 1
SolutionMigration 0
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 1
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

=== Adapt.inp (adaptation step) ===
<nowiki>eviscScal 0
pgrad 0
ub 1.45
lb 0.70
numSplit 5
sizeRatio 0.35 # confirmed that 0.25 does a 4x ref
numSmooth 0 #confirmed that this gets used
ratioThresh 0.75 # this is h_new/h_orig from isotropic days
localAdapt 1 # makes setLocal force adaptivity only around nodes set size
AnisoSimmetrix 4 # > 0 use Simmetrix, else ours 1 does largest length reduction by sizeRatio, 2 does both does medium and largest reduction by sizeRatio
coarsenMode 0 # disables coarsening see coarsenMode docs for 1, and 2
numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 1
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.5e-4
AdaptOption 11
hmax 1e6
hmin 1e-5
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 0
numTotParts 1
SolutionMigration 1
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 0
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

== Notes on Folder Structure ==
Below is an overview of the folder structure to reference during the adaptation process:
# [ geom.smd, geom.sms, geom_nat.x_t ]
# 1-1-Chef
## 1-1-phParAdapt
### adapt.inp
### run_phParAdapt.sh
### 1-procs_case
#### geombc.dat.1
#### restart.<#>.1
### mesh_parts.sms
### A1-phParAdapt
#### restart.<#>.1
#### errors.<#>.1
#### [geom.smd, geom.sms, geom_nat.x_t, parts.sms]
#### adapt.inp
#### <#> ( step number directory )
##### mesh_parts.sms
##### 1-1-phParAdapt
###### [geom.smd, geom.sms, geom_nat.x_t]
###### adapt.inp
###### 1-procs_case ( restart files with adapted mesh )

== Removal of Extrusion Constraint ==
Riccardo has used a procedure in the past to remove the extrusion constraint from a mesh that enabled adaptation to be performed on extrusion-type elements.

The following directory contains an example of this process:
<code> /projects/tools/Models/NASAWingBodyJunction/RajMeshFine/Mesh/1-1-phParAdapt/RemExtrusion/ </code>

The executable that is called for removal of the constraint is a part of SCOREC-core, found at <code> <path/to/SCOREC-core/build/dir>/test/rm_extrusion </code> and the source code, in case edits are required, is found at <code> <path/to/SCOREC-core/source/dir>/core/test/rm_extrusion.cc </code>

PhParAdapt/Simmetrix

2021-04-28T19:59:05Z

Jopo9795: /* Performing the Adaptation Step */

The following page provides details for performing mesh adaptation using the PhParAdapt tool with Simmetrix routines.

== Initial Notes to User ==
* The solution migration feature is currently broken and thus requires other means for transferring the solution to new meshes that are created from PhParAdapt (e.g. solution interpolation in Paraview)
* To date, the steps described herein have only been tried with the following restrictions:
** Tetrahedral elements
*** Extruded element types apparently can be used but require removal of any extrusion constraints in the mesh. Below are some incomplete details regarding that process for reference. See [[#Removal_of_Extrusion_Constraint |Removal of Extrusion Constraint]].
** Serial case
** Geometries created using SimModeler7.0-190626
** Adaptation using 6th entry of errors field

== Adaptation Process ==

=== Creating Initial Restart and Error Files ===
Beginning inside of the 1-1-Chef directory,
<nowiki>mkdir 1-1-phParAdapt
cd 1-1-phParAdapt </nowiki>

Create soft links to the geom files above the 1-1-Chef directory,
<nowiki>ln -s ../../geom.smd
ln -s ../../geom.sms
ln -s ../../geom_nat.x_t </nowiki>

Create an <code> adapt.inp </code> using the template [[#Adapt.inp_(non-adaptation_step) | Adapt.inp (non-adaptation step)]]. Keep <code>adaptFlag</code> set to zero as the purpose of running phParAdapt here is to create restart.<timeStepNumber>.1 file associated with a new mesh format that can be used for adaptation.

Run phParAdapt in this directory, which for this specific example was:

<code>mpirun -np 1 /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi</code>

After completion, the 1-procs_case and mesh_parts.sms directories should now be present. If a solution has already been generated for a different mesh on the same geometry, it can now be interpolated onto the new restart.<#>.1 file that was generated. If no solution exists, run Phasta using this restart file until the desired solution state has been achieved. Be sure to set <code>Print Error Indicators: True</code> in the solver.inp file so the error fields are saved in the final restart file. Note, error fields are not printed in intermediate restart files. Before proceeding, check that the error fields exist using <code>grep -a " : <" restart.<#>.1</code>

=== Performing the Adaptation Step ===
Within the 1-1-phParAdapt directory,
<nowiki>mkdir A1-phParAdapt
cd A1-phParAdapt
ln -s ../geom.smd
ln -s ../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
ln -s geom.sms parts.sms
ln -s ../1-procs_case/restart.1.1
ln -s restart.1.1 errors.1.1 </nowiki>

In the above step, it is assumed that the solution was interpolated onto restart.1.1. If that was not the case and the solution was advanced until a desired step was reached, replace restart.1.1 & errors.1.1 with restart.<#>.1 & errors.<#>.1 corresponding to the desired solution step number with error fields.

Create an adapt.inp file using [[#Adapt.inp_(adaptation step) |Adapt.inp_(adaptation step)]] as a template. Change the <code>timeStepNumber</code> parameter to correspond to the restart file that was linked into the current directory. Make sure the <code>adaptFlag</code> parameter is set to one and change additional threshold flags to control the result of the refinement.

Run phParAdapt with the following command,

<code>mpirun -np 1 /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi-O 2>&1 | tee phParAdapt.log</code>

A directory corresponding to the <code>timeStepNumber</code> parameter should have been created with a mesh_parts.sms folder inside.

=== Generate Restart File for Refined Mesh ===
From within the A1-phParAdapt directory, execute the following (noting to replace <#> with the appropriate time step number):
<nowiki>cd <#>
mkdir 1-1-phParAdapt
cd 1-1-phParAdapt
ln -s ../../../geom.smd
ln -s ../../../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
cp ../../../adapt.inp .</nowiki>

Note that in the above, the adapt.inp file from the non-adaptation step was copied into the current directory.

== File Examples ==

The file examples below are what were used for this specific example and may contain parameter flags that may need to be changed or are not used at all.

=== Adapt.inp (non-adaptation step) ===
<nowiki>numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 0
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.00001
AdaptOption 11
hmax 0.04
hmin 0.000052083
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 1
numTotParts 1
SolutionMigration 0
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 1
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

=== Adapt.inp (adaptation step) ===
<nowiki>eviscScal 0
pgrad 0
ub 1.45
lb 0.70
numSplit 5
sizeRatio 0.35 # confirmed that 0.25 does a 4x ref
numSmooth 0 #confirmed that this gets used
ratioThresh 0.75 # this is h_new/h_orig from isotropic days
localAdapt 1 # makes setLocal force adaptivity only around nodes set size
AnisoSimmetrix 4 # > 0 use Simmetrix, else ours 1 does largest length reduction by sizeRatio, 2 does both does medium and largest reduction by sizeRatio
coarsenMode 0 # disables coarsening see coarsenMode docs for 1, and 2
numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 1
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.5e-4
AdaptOption 11
hmax 1e6
hmin 1e-5
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 0
numTotParts 1
SolutionMigration 1
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 0
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

== Notes on Folder Structure ==
Below is an overview of the folder structure to reference during the adaptation process:
# [ geom.smd, geom.sms, geom_nat.x_t ]
# 1-1-Chef
## 1-1-phParAdapt
### adapt.inp
### run_phParAdapt.sh
### 1-procs_case
#### geombc.dat.1
#### restart.<#>.1
### mesh_parts.sms
### A1-phParAdapt
#### restart.<#>.1
#### errors.<#>.1
#### [geom.smd, geom.sms, geom_nat.x_t, parts.sms]
#### adapt.inp
#### <#> ( step number directory )
##### mesh_parts.sms
##### 1-1-phParAdapt
###### [geom.smd, geom.sms, geom_nat.x_t]
###### adapt.inp
###### 1-procs_case ( restart files with adapted mesh )

== Removal of Extrusion Constraint ==
Riccardo has used a procedure in the past to remove the extrusion constraint from a mesh that enabled adaptation to be performed on extrusion-type elements.

The following directory contains an example of this process:
<code> /projects/tools/Models/NASAWingBodyJunction/RajMeshFine/Mesh/1-1-phParAdapt/RemExtrusion/ </code>

The executable that is called for removal of the constraint is a part of SCOREC-core, found at <code> <path/to/SCOREC-core/build/dir>/test/rm_extrusion </code> and the source code, in case edits are required, is found at <code> <path/to/SCOREC-core/source/dir>/core/test/rm_extrusion.cc </code>

PhParAdapt/Simmetrix

2021-04-28T19:48:19Z

Jopo9795: /* Performing the Adaptation Step */

The following page provides details for performing mesh adaptation using the PhParAdapt tool with Simmetrix routines.

== Initial Notes to User ==
* The solution migration feature is currently broken and thus requires other means for transferring the solution to new meshes that are created from PhParAdapt (e.g. solution interpolation in Paraview)
* To date, the steps described herein have only been tried with the following restrictions:
** Tetrahedral elements
*** Extruded element types apparently can be used but require removal of any extrusion constraints in the mesh. Below are some incomplete details regarding that process for reference. See [[#Removal_of_Extrusion_Constraint |Removal of Extrusion Constraint]].
** Serial case
** Geometries created using SimModeler7.0-190626
** Adaptation using 6th entry of errors field

== Adaptation Process ==

=== Creating Initial Restart and Error Files ===
Beginning inside of the 1-1-Chef directory,
<nowiki>mkdir 1-1-phParAdapt
cd 1-1-phParAdapt </nowiki>

Create soft links to the geom files above the 1-1-Chef directory,
<nowiki>ln -s ../../geom.smd
ln -s ../../geom.sms
ln -s ../../geom_nat.x_t </nowiki>

Create an <code> adapt.inp </code> using the template [[#Adapt.inp_(non-adaptation_step) | Adapt.inp (non-adaptation step)]]. Keep <code>adaptFlag</code> set to zero as the purpose of running phParAdapt here is to create restart.<timeStepNumber>.1 file associated with a new mesh format that can be used for adaptation.

Run phParAdapt in this directory, which for this specific example was:

<code>mpirun -np 1 /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi</code>

After completion, the 1-procs_case and mesh_parts.sms directories should now be present. If a solution has already been generated for a different mesh on the same geometry, it can now be interpolated onto the new restart.<#>.1 file that was generated. If no solution exists, run Phasta using this restart file until the desired solution state has been achieved. Be sure to set <code>Print Error Indicators: True</code> in the solver.inp file so the error fields are saved in the final restart file. Note, error fields are not printed in intermediate restart files. Before proceeding, check that the error fields exist using <code>grep -a " : <" restart.<#>.1</code>

=== Performing the Adaptation Step ===
Within the 1-1-phParAdapt directory,
<nowiki>mkdir A1-phParAdapt
cd A1-phParAdapt
ln -s ../geom.smd
ln -s ../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
ln -s geom.sms parts.sms
ln -s ../1-procs_case/restart.1.1
ln -s restart.1.1 errors.1.1 </nowiki>

In the above step, it is assumed that the solution was interpolated onto restart.1.1. If that was not the case and the solution was advanced until a desired step was reached, replace restart.1.1 & errors.1.1 with restart.<#>.1 & errors.<#>.1 corresponding to the desired solution step number with error fields.

Create an adapt.inp file using [[#Adapt.inp_(adaptation step) |Adapt.inp_(adaptation step)]] as a template. Change the <code>timeStepNumber</code> parameter to correspond to the restart file that was linked into the current directory. Change additional parameter flags

== File Examples ==

The file examples below are what were used for this specific example and may contain parameter flags that may need to be changed or are not used at all.

=== Adapt.inp (non-adaptation step) ===
<nowiki>numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 0
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.00001
AdaptOption 11
hmax 0.04
hmin 0.000052083
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 1
numTotParts 1
SolutionMigration 0
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 1
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

=== Adapt.inp (adaptation step) ===
<nowiki>eviscScal 0
pgrad 0
ub 1.45
lb 0.70
numSplit 5
sizeRatio 0.35 # confirmed that 0.25 does a 4x ref
numSmooth 0 #confirmed that this gets used
ratioThresh 0.75 # this is h_new/h_orig from isotropic days
localAdapt 1 # makes setLocal force adaptivity only around nodes set size
AnisoSimmetrix 4 # > 0 use Simmetrix, else ours 1 does largest length reduction by sizeRatio, 2 does both does medium and largest reduction by sizeRatio
coarsenMode 0 # disables coarsening see coarsenMode docs for 1, and 2
numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 1
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.5e-4
AdaptOption 11
hmax 1e6
hmin 1e-5
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 0
numTotParts 1
SolutionMigration 1
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 0
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

== Notes on Folder Structure ==
Below is an overview of the folder structure to reference during the adaptation process:
# [ geom.smd, geom.sms, geom_nat.x_t ]
# 1-1-Chef
## 1-1-phParAdapt
### adapt.inp
### run_phParAdapt.sh
### 1-procs_case
#### geombc.dat.1
#### restart.<#>.1
### mesh_parts.sms
### A1-phParAdapt
#### restart.<#>.1
#### errors.<#>.1
#### [geom.smd, geom.sms, geom_nat.x_t, parts.sms]
#### adapt.inp
#### <#> ( step number directory )
##### mesh_parts.sms
##### 1-1-phParAdapt
###### [geom.smd, geom.sms, geom_nat.x_t]
###### adapt.inp
###### 1-procs_case ( restart files with adapted mesh )

== Removal of Extrusion Constraint ==
Riccardo has used a procedure in the past to remove the extrusion constraint from a mesh that enabled adaptation to be performed on extrusion-type elements.

The following directory contains an example of this process:
<code> /projects/tools/Models/NASAWingBodyJunction/RajMeshFine/Mesh/1-1-phParAdapt/RemExtrusion/ </code>

The executable that is called for removal of the constraint is a part of SCOREC-core, found at <code> <path/to/SCOREC-core/build/dir>/test/rm_extrusion </code> and the source code, in case edits are required, is found at <code> <path/to/SCOREC-core/source/dir>/core/test/rm_extrusion.cc </code>

PhParAdapt/Simmetrix

2021-04-28T19:45:17Z

Jopo9795: /* Creating Initial Restart and Error Files */

The following page provides details for performing mesh adaptation using the PhParAdapt tool with Simmetrix routines.

== Initial Notes to User ==
* The solution migration feature is currently broken and thus requires other means for transferring the solution to new meshes that are created from PhParAdapt (e.g. solution interpolation in Paraview)
* To date, the steps described herein have only been tried with the following restrictions:
** Tetrahedral elements
*** Extruded element types apparently can be used but require removal of any extrusion constraints in the mesh. Below are some incomplete details regarding that process for reference. See [[#Removal_of_Extrusion_Constraint |Removal of Extrusion Constraint]].
** Serial case
** Geometries created using SimModeler7.0-190626
** Adaptation using 6th entry of errors field

== Adaptation Process ==

=== Creating Initial Restart and Error Files ===
Beginning inside of the 1-1-Chef directory,
<nowiki>mkdir 1-1-phParAdapt
cd 1-1-phParAdapt </nowiki>

Create soft links to the geom files above the 1-1-Chef directory,
<nowiki>ln -s ../../geom.smd
ln -s ../../geom.sms
ln -s ../../geom_nat.x_t </nowiki>

Create an <code> adapt.inp </code> using the template [[#Adapt.inp_(non-adaptation_step) | Adapt.inp (non-adaptation step)]]. Keep <code>adaptFlag</code> set to zero as the purpose of running phParAdapt here is to create restart.<timeStepNumber>.1 file associated with a new mesh format that can be used for adaptation.

Run phParAdapt in this directory, which for this specific example was:

<code>mpirun -np 1 /projects/tools/Simmetrix.develop_19/phParAdapt-Sim/phParAdapt-Sim19/bin/x86_64_linux/phParAdapt-parasolid-openmpi</code>

After completion, the 1-procs_case and mesh_parts.sms directories should now be present. If a solution has already been generated for a different mesh on the same geometry, it can now be interpolated onto the new restart.<#>.1 file that was generated. If no solution exists, run Phasta using this restart file until the desired solution state has been achieved. Be sure to set <code>Print Error Indicators: True</code> in the solver.inp file so the error fields are saved in the final restart file. Note, error fields are not printed in intermediate restart files. Before proceeding, check that the error fields exist using <code>grep -a " : <" restart.<#>.1</code>

=== Performing the Adaptation Step ===
Within the 1-1-phParAdapt directory,
<nowiki>mkdir A1-phParAdapt
cd A1-phParAdapt
ln -s ../geom.smd
ln -s ../geom_nat.x_t
ln -s ../mesh_parts.sms geom.sms
ln -s geom.sms parts.sms
ln -s ../1-procs_case/restart.1.1
ln -s restart.1.1 errors.1.1 </nowiki>

In the above step, it is assumed that the solution was interpolated onto restart.1.1. If that was not the case and the solution was advanced until a desired step was reached, replace restart.1.1 & errors.1.1 with restart.<#>.1 & errors.<#>.1 corresponding to the desired solution step number with error fields.

Create an adapt.inp file using [[#Adapt.inp_(adaptation step) |Adapt.inp_(adaptation step)]] as a template.

== File Examples ==

The file examples below are what were used for this specific example and may contain parameter flags that may need to be changed or are not used at all.

=== Adapt.inp (non-adaptation step) ===
<nowiki>numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 0
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.00001
AdaptOption 11
hmax 0.04
hmin 0.000052083
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 1
numTotParts 1
SolutionMigration 0
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 1
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

=== Adapt.inp (adaptation step) ===
<nowiki>eviscScal 0
pgrad 0
ub 1.45
lb 0.70
numSplit 5
sizeRatio 0.35 # confirmed that 0.25 does a 4x ref
numSmooth 0 #confirmed that this gets used
ratioThresh 0.75 # this is h_new/h_orig from isotropic days
localAdapt 1 # makes setLocal force adaptivity only around nodes set size
AnisoSimmetrix 4 # > 0 use Simmetrix, else ours 1 does largest length reduction by sizeRatio, 2 does both does medium and largest reduction by sizeRatio
coarsenMode 0 # disables coarsening see coarsenMode docs for 1, and 2
numberSolutionVars 6
numberErrorVars 10
refWeights 1 1 1 1 1 1 1 1 1 1
refThreshold 0.01
globalP 1
timeStepNumber 1
numelX 0
NSFTAG -1
ensa_dof 6
attributeFileName geom.smd
meshFileName geom.sms
modelFileName geom_nat.x_t
Idirection 0
BYPASS 0
zScale 0
adaptFlag 1
errorName 0
SONFATH 0
lStart 0
rRead 6
rStart 0
AdaptStrategy 5
AdaptFactor 0.5e-4
AdaptOption 11
hmax 1e6
hmin 1e-5
multipleRestarts 0
Periodic 0
prCD 0
timing 0
wGraph 0
phastaVersion 1.9.5
old_format 0
FortFormFlag 0
outputFormat binary
CUBES 0
internalBCNodes 0
version UNKNOWN
WRITEASC 0
phastaIO 0
numTotParts 1
SolutionMigration 1
DisplacementMigration 0
isReorder 0
isMCOPI 0
isBLAdapt 0
isThickAdapt 0
isSizeLimit 0
MaxLimitFact 2
MinLimitFact 2
rho 1.225
mu 1.7825e-5
dwalMigration 0
buildMapping 1 </nowiki>

== Notes on Folder Structure ==
Below is an overview of the folder structure to reference during the adaptation process:
# [ geom.smd, geom.sms, geom_nat.x_t ]
# 1-1-Chef
## 1-1-phParAdapt
### adapt.inp
### run_phParAdapt.sh
### 1-procs_case
#### geombc.dat.1
#### restart.<#>.1
### mesh_parts.sms
### A1-phParAdapt
#### restart.<#>.1
#### errors.<#>.1
#### [geom.smd, geom.sms, geom_nat.x_t, parts.sms]
#### adapt.inp
#### <#> ( step number directory )
##### mesh_parts.sms
##### 1-1-phParAdapt
###### [geom.smd, geom.sms, geom_nat.x_t]
###### adapt.inp
###### 1-procs_case ( restart files with adapted mesh )

== Removal of Extrusion Constraint ==
Riccardo has used a procedure in the past to remove the extrusion constraint from a mesh that enabled adaptation to be performed on extrusion-type elements.

The following directory contains an example of this process:
<code> /projects/tools/Models/NASAWingBodyJunction/RajMeshFine/Mesh/1-1-phParAdapt/RemExtrusion/ </code>

The executable that is called for removal of the constraint is a part of SCOREC-core, found at <code> <path/to/SCOREC-core/build/dir>/test/rm_extrusion </code> and the source code, in case edits are required, is found at <code> <path/to/SCOREC-core/source/dir>/core/test/rm_extrusion.cc </code>

PhParAdapt/Simmetrix

2021-04-28T19:33:50Z

Jopo9795: /* Creating Initial Restart and Error Files */

PhParAdapt/Simmetrix

2021-04-28T19:11:20Z

Jopo9795: /* Initial Notes to User */

PhParAdapt/Simmetrix

2021-04-28T19:10:10Z

Jopo9795: /* File Examples */

PhParAdapt/Simmetrix

2021-04-28T19:09:19Z

Jopo9795: /* Creating Initial Restart and Error Files */

PhParAdapt/Simmetrix

2021-04-28T19:07:36Z

Jopo9795: /* File Examples */

PhParAdapt/Simmetrix

2021-04-28T19:03:01Z

Jopo9795: /* Notes on Folder Structure */

PhParAdapt/Simmetrix

2021-04-28T19:00:29Z

Jopo9795: /* Initial Notes to User */

PhParAdapt/Simmetrix

2021-04-28T18:59:19Z

Jopo9795: /* Initial Notes to User */

PhParAdapt/Simmetrix

2021-04-28T18:58:43Z

Jopo9795: /* Initial Notes to User */

PhParAdapt/Simmetrix

2021-04-28T18:58:19Z

Jopo9795: /* Initial Notes to User */

PhParAdapt/Simmetrix

2021-04-28T18:55:24Z

Jopo9795: /* = Creating Initial Restart and Error Files */

PhParAdapt/Simmetrix

2021-04-28T18:55:03Z

Jopo9795: /* Initial Notes to User */

PhParAdapt/Simmetrix

2021-04-28T18:41:57Z

Jopo9795: /* Notes on Folder Structure */

PhParAdapt/Simmetrix

2021-04-28T18:36:29Z

Jopo9795: /* Notes on Folder Structure */

PhParAdapt/Simmetrix

2021-04-28T18:26:58Z

Jopo9795: /* Initial Notes to User */

PhParAdapt/Simmetrix

2021-04-28T18:18:49Z

Jopo9795: /* Initial Notes to User */

PhParAdapt/Simmetrix

2021-04-28T18:18:22Z

Jopo9795: /* Initial Notes to User */