SolutionChangeNormFixedPoint

Computes the L2 norm of the change in a variable between consecutive fixed-point (Picard) iterations:

Intended for use as a staggered-scheme convergence criterion in conjunction with PostprocessorConvergence and the multiapp_fixed_point_convergence Executioner parameter.

Usage

The postprocessor defaults to execute_on = MULTIAPP_FIXED_POINT_CONVERGENCE, which causes it to evaluate after each sub-app solve and transfer but before the convergence decision is made.

commentnote

When using this postprocessor with PostprocessorConvergence, the parameters max_iterations and converge_at_max_iterations belong to the [Convergence] block, not the [Executioner] block.

Example Input File Syntax


[AuxVariables]
  [d]
  []
[]

[Postprocessors]
  [d_fp_change]
    type = SolutionChangeNormFixedPoint
    variable = d
  []
[]

[Convergence]
  [fp_conv]
    type = PostprocessorConvergence
    postprocessor = d_fp_change
    tolerance = 1e-4
    max_iterations = 20
    converge_at_max_iterations = false
  []
[]

[Executioner]
  type = Transient
  multiapp_fixed_point_convergence = fp_conv
[]

Input Files