NeuralNetFreeEnergy

This free energy class uses a multilayer perceptron with input nodes fed by coupled non-linear variables, and an output node representing a free energy value.

The neural net description is loaded from a user specified file. The supported input formats are a custom format that is written by the supplied pytorch based fitting scripts, and the text format written by the GENANN C++ library.

Evaluates a fitted deep neural network to obtain a free energy and its derivatives with a preset activation function.

Input Parameters

activation_functionSIGMOIDWeights and biases file format
Default:SIGMOID
C++ Type:MooseEnum
Options:SIGMOID, SOFTSIGN, TANH
Controllable:No
Description:Weights and biases file format
blockThe list of blocks (ids or names) that this object will be applied
C++ Type:std::vector<SubdomainName>
Controllable:No
Description:The list of blocks (ids or names) that this object will be applied
boundaryThe list of boundaries (ids or names) from the mesh where this object applies
C++ Type:std::vector<BoundaryName>
Controllable:No
Description:The list of boundaries (ids or names) from the mesh where this object applies
computeTrueWhen false, MOOSE will not call compute methods on this material. The user must call computeProperties() after retrieving the MaterialBase via MaterialBasePropertyInterface::getMaterialBase(). Non-computed MaterialBases are not sorted for dependencies.
Default:True
C++ Type:bool
Controllable:No
Description:When false, MOOSE will not call compute methods on this material. The user must call computeProperties() after retrieving the MaterialBase via MaterialBasePropertyInterface::getMaterialBase(). Non-computed MaterialBases are not sorted for dependencies.
constant_onNONEWhen ELEMENT, MOOSE will only call computeQpProperties() for the 0th quadrature point, and then copy that value to the other qps.When SUBDOMAIN, MOOSE will only call computeQpProperties() for the 0th quadrature point, and then copy that value to the other qps. Evaluations on element qps will be skipped
Default:NONE
C++ Type:MooseEnum
Options:NONE, ELEMENT, SUBDOMAIN
Controllable:No
Description:When ELEMENT, MOOSE will only call computeQpProperties() for the 0th quadrature point, and then copy that value to the other qps.When SUBDOMAIN, MOOSE will only call computeQpProperties() for the 0th quadrature point, and then copy that value to the other qps. Evaluations on element qps will be skipped
debugFalseTabulate the NN to a file for debugging purposes
Default:False
C++ Type:bool
Controllable:No
Description:Tabulate the NN to a file for debugging purposes
declare_suffixAn optional suffix parameter that can be appended to any declared properties. The suffix will be prepended with a '_' character.
C++ Type:MaterialPropertyName
Unit:(no unit assumed)
Controllable:No
Description:An optional suffix parameter that can be appended to any declared properties. The suffix will be prepended with a '_' character.
file_formatWeights and biases file format
C++ Type:MooseEnum
Options:MAGPIE, GENANN
Controllable:No
Description:Weights and biases file format
file_nameData file containing the weights and biasses for a fully connected deep neural network
C++ Type:FileName
Controllable:No
Description:Data file containing the weights and biasses for a fully connected deep neural network
inputsCoupled Variables that are inputs for the neural network
C++ Type:std::vector<VariableName>
Unit:(no unit assumed)
Controllable:No
Description:Coupled Variables that are inputs for the neural network
prop_nameslist of material properties fed from the outputs of the neural network
C++ Type:std::vector<MaterialPropertyName>
Unit:(no unit assumed)
Controllable:No
Description:list of material properties fed from the outputs of the neural network

Optional Parameters

control_tagsAdds user-defined labels for accessing object parameters via control logic.
C++ Type:std::vector<std::string>
Controllable:No
Description:Adds user-defined labels for accessing object parameters via control logic.
enableTrueSet the enabled status of the MooseObject.
Default:True
C++ Type:bool
Controllable:Yes
Description:Set the enabled status of the MooseObject.
implicitTrueDetermines whether this object is calculated using an implicit or explicit form
Default:True
C++ Type:bool
Controllable:No
Description:Determines whether this object is calculated using an implicit or explicit form
seed0The seed for the master random number generator
Default:0
C++ Type:unsigned int
Controllable:No
Description:The seed for the master random number generator
use_displaced_meshFalseWhether or not this object should use the displaced mesh for computation. Note that in the case this is true but no displacements are provided in the Mesh block the undisplaced mesh will still be used.
Default:False
C++ Type:bool
Controllable:No
Description:Whether or not this object should use the displaced mesh for computation. Note that in the case this is true but no displacements are provided in the Mesh block the undisplaced mesh will still be used.

Advanced Parameters

output_propertiesList of material properties, from this material, to output (outputs must also be defined to an output type)
C++ Type:std::vector<std::string>
Controllable:No
Description:List of material properties, from this material, to output (outputs must also be defined to an output type)
outputsnone Vector of output names where you would like to restrict the output of variables(s) associated with this object
Default:none
C++ Type:std::vector<OutputName>
Controllable:No
Description:Vector of output names where you would like to restrict the output of variables(s) associated with this object

Outputs Parameters

prop_getter_suffixAn optional suffix parameter that can be appended to any attempt to retrieve/get material properties. The suffix will be prepended with a '_' character.
C++ Type:MaterialPropertyName
Unit:(no unit assumed)
Controllable:No
Description:An optional suffix parameter that can be appended to any attempt to retrieve/get material properties. The suffix will be prepended with a '_' character.
use_interpolated_stateFalseFor the old and older state use projected material properties interpolated at the quadrature points. To set up projection use the ProjectedStatefulMaterialStorageAction.
Default:False
C++ Type:bool
Controllable:No
Description:For the old and older state use projected material properties interpolated at the quadrature points. To set up projection use the ProjectedStatefulMaterialStorageAction.

Material Property Retrieval Parameters

References

No citations exist within this document.

(tests/neuralnet/genann.i)

[Mesh]
  type = GeneratedMesh
  dim = 2
  nx = 20
  ny = 20
[]

[AuxVariables]
  [./T]
    [./InitialCondition]
      type = FunctionIC
      function = y/2+0.1
    [../]
  [../]
  [./c]
    [./InitialCondition]
      type = FunctionIC
      function = x
    [../]
  [../]
[]

[Materials]
  [./Fn]
    type = NeuralNetFreeEnergy
    file_name = genann.dat
    file_format = GENANN
    inputs     = 'T c'
    prop_names = 'F'
    outputs = exodus
  [../]
[]

[Problem]
  kernel_coverage_check = false
  solve = false
[]

[Executioner]
  type = Steady
[]

[Outputs]
  exodus = true
[]

(tests/neuralnet/magpie.i)

[Mesh]
  type = GeneratedMesh
  dim = 2
  nx = 20
  ny = 20
[]

[AuxVariables]
  [./T]
    [./InitialCondition]
      type = FunctionIC
      function = y/2+0.1
    [../]
  [../]
  [./c]
    [./InitialCondition]
      type = FunctionIC
      function = x
    [../]
  [../]
[]

[Materials]
  [./Fn]
    type = NeuralNetFreeEnergy
    file_name = weights_biases.txt
    file_format = MAGPIE
    inputs     = 'T c'
    prop_names = 'F'
    outputs = exodus
  [../]
[]

[Problem]
  kernel_coverage_check = false
  solve = false
[]

[Executioner]
  type = Steady
[]

[Outputs]
  exodus = true
[]

(examples/deep_neural_network/dnn.i)

[Mesh]
  type = GeneratedMesh
  dim = 2
  nx = 100
  ny = 100
  xmax = 70
  ymax = 70
[]

[AuxVariables]
  [T]
    initial_condition = 0.2
  []
  [diff]
  []
[]

[AuxKernels]
  [diff]
    type = ParsedAux
    variable = diff
    expression = 'cn - cp'
    coupled_variables = 'cn cp'
  []
[]

[Variables]
  # active = 'cp wp'
  [cn]
    [InitialCondition]
      type = RandomIC
      min = 0.45
      max = 0.55
      seed = 1234
      #type = FunctionIC
      #function = cos(x/70*2*pi)/2+0.5
    []
  []
  [wn]
  []

  [cp]
    [InitialCondition]
      type = RandomIC
      min = 0.45
      max = 0.55
      seed = 1234
      #type = FunctionIC
      #function = cos(x/70*2*pi)/2+0.5
    []
  []
  [wp]
  []
[]

[Kernels]
  # active = 'cp_res wp_res timep'
  [cn_res]
    type = ADSplitCHParsed
    variable = cn
    f_name = Fn
    kappa_name = ad_kappa_c
    w = wn
  []
  [wn_res]
    type = ADSplitCHWRes
    variable = wn
    mob_name = ad_M
  []
  [timen]
    type = ADCoupledTimeDerivative
    variable = wn
    v = cn
  []

  [cp_res]
    type = SplitCHParsed
    variable = cp
    f_name = Fp
    kappa_name = kappa_c
    w = wp
  []
  [wp_res]
    type = SplitCHWRes
    variable = wp
    mob_name = M
  []
  [timep]
    type = CoupledTimeDerivative
    variable = wp
    v = cp
  []
[]

[BCs]
  [Periodic]
    [all]
      auto_direction = 'x y'
    []
  []
[]

[Materials]
  # active = 'Fp pfmobility'
  [Fn]
    type = NeuralNetFreeEnergy
    file_name = weights_biases.txt
    file_format = MAGPIE
    activation_function = SOFTSIGN
    inputs = 'T cn'
    prop_names = 'Fn'
    outputs = exodus
    debug = true
  []
  [Fp]
    type = DerivativeParsedMaterial
    property_name = 'Fp'
    expression = 'cp*(1-cp)+0.001*(T*300+400)*(cp*log(cp)+(1-cp)*log(1-cp))'
    coupled_variables = 'cp T'
    derivative_order = 2
    outputs = exodus
  []
  [pfmobility]
    type = GenericConstantMaterial
    prop_names = 'M kappa_c'
    prop_values = '1 0.25'
  []
  [adpfmobility]
    type = ADGenericConstantMaterial
    prop_names = 'ad_M ad_kappa_c'
    prop_values = '1 0.25'
  []
[]

[Preconditioning]
  [SMP]
    type = SMP
    full = true
  []
[]

[Problem]
  kernel_coverage_check = false
[]

[Postprocessors]
  [diff]
    type = ElementL2Difference
    variable = cn
    other_variable = cp
  []
[]

[Executioner]
  type = Transient
  solve_type = 'NEWTON'

  nl_abs_tol = 1e-12
  num_steps = 150
  dt = 5
[]

[Outputs]
  exodus = true
  gnuplot = true
  print_linear_residuals = false
[]

Input Parameters
Input Files
References