Scenarios for using the multiapp and transfer system

[lvkas521424]

Check these boxes if you have followed the posting rules.

• Q&A General is the most appropriate section for my question
• I have consulted the posting Guidelines on the Discussions front page
• I have searched the Discussions forum and my question has not been asked before
• I have searched the MOOSE website and the documentation does not answer my question
• I have formatted my post following the posting guidelines (screenshots as a last resort, triple back quotes around pasted text)

Question

Hello developers,
Currently I am using moose framework to access two external programs, and I want to complete the moose framework through the "tight" coupling of the two programs, through the official website I learned that multi-app and transfer is mainly used in some loosely coupled or decoupled scenarios. So I am curious, for the tightly coupled scenario of two external programs, can I continue to use the multi-app and transfer system, for example, when I iterate in picard, since I am not sure how many times I iterate before convergence, at this time I should be in the input card how to set up the execution_time of the transfer, or should I use another function module to accomplish my current picard iteration I'm looking forward to your answer.
Very much looking forward to your answers.

[GiudGiud]

hello

through the official website I learned that multi-app and transfer is mainly used in some loosely coupled or decoupled scenarios

multiapps can do tight coupling too. You just need to set fixed point iteration parameters to the executioner. Notably change the max number of fixed point iterations.

since I am not sure how many times I iterate before convergence, at this time I should be in the input card how to set up the execution_time of the transfer

there are a variety of convergence criteria available, notably using a postprocessor value which is a very flexible option

[lvkas521424]

Hello, I tried the code as I understand it, but I encountered some errors

`void
ReactorCouplingControl::executeNeutronics()
{
  // 检查中子学多应用程序是否存在
  if (!_fe_problem.hasMultiApp(_neutronics_app_name))
  {
    mooseWarning("ReactorCouplingControl: 找不到中子学多应用程序 '", _neutronics_app_name, "'");
    return;
  }
    mooseInfo("_calc_type 1： ", _calc_type);
  // 方案2 - 获取指定的 MultiApp 并单独执行它

  _fe_problem.execMultiApps(_from_neutronics_transfers);
  
  // 从中子学应用接收数据
    mooseInfo("_calc_type 2： ", _calc_type);
  _fe_problem.execMultiAppTransfers(_from_neutronics_transfers, MultiAppTransfer::FROM_MULTIAPP);
    mooseInfo("_calc_type 3： ", _calc_type);
}`

but when i execute the mooseapp, it tell me :

`
ReactorCouplingControl: 执行第一个燃耗步特殊处理


*** Info ***
/home/lvjiahui/projects/mooseprojects/neutronic.i:2.3:
The following information occurred in the Control 'coupling_control' of type ReactorCouplingControl.

_calc_type 1： 1


*** ERROR ***
Unable to locate the desired execute flag (from_scapn), the supplied execution flag was likely not registered.

Stack frames: 17
0: libMesh::print_trace(std::ostream&)
1`

https://github.com/lvkas521424/mooseprojects.git are all my codes and input files, please also correct me.

[GiudGiud]

You are supposed to be passing execution flags to these execMultiApp routines but it seems you are passing an unrelated string, the name of a transfer

[lvkas521424]

I modify _fe_problem.execMultiApps(_from_neutronics_transfers); to _fe_problem.execMultiApps(EXEC_CUSTOM);，and my input file is:

[MultiApps]
  [neutronics]
    type = FullSolveMultiApp
    app_type = MooseprojectsApp       
    input_files = 'scapn_input.i'     
    execute_on = 'CUSTOM'  
  []
[]

[Transfers]
  [from_scapn]
    type = MultiAppCopyTransfer
    from_multi_app = neutronics
    source_variable = 'power_density'
    variable = 'power_density'
    execute_on = 'CUSTOM' 
  []
  
  [to_scapn]
    type = MultiAppCopyTransfer
    to_multi_app = neutronics
    source_variable = 'temperature'
    variable = 'temperature'
    execute_on = 'CUSTOM'  
  []
[]

But it still doesn't work，

*** ERROR ***
Unable to locate the desired execute flag (from_scapn), the supplied execution flag was likely not registered.

Stack frames: 17

[GiudGiud]

It s likely from another call.

You have multiple routines here that expect execute on flags

[lvkas521424]

``excues me,
First of all thank you very much, after I managed to schedule my multiapps using the Execute flag, I found that everything worked. However, I noticed that my appliciation didn't end successfully after I performed all the operations, but it prompted me with

free(): invalid pointer 
Aborted (core dumped)

I suspect that there is something wrong with my input card.
yours lvkas
this is my inputcard:

[Controls]
  [coupling_control]
    type = ReactorCouplingControl
    
    calc_type = NEUTRONICS
    
    burn_step = 1            # 更新为1，与代码默认值一致
    max_burn_steps = 2      # 减少为2，只执行一次燃耗步
    
    neutronics_app = neutronics 
    thermal_app = thermal
    #from_neutronics_transfers = from_neutronics
    execute_on = 'timestep_begin'   # 在每个时间步开始时执行Control
  []
[]

[MultiApps]
  [neutronics]
    type = NeutronicsMultiApp
    app_type = MooseprojectsApp
    input_files = 'scapn_input.i'
    mesh_dims = '10 10 10'           # 要与 ReactorCouplingControl 一致
    power_var_name = 'power_density'  # 要与变量名一致
    temperature_var_name = 'temperature'  # 要与变量名一致
    execute_on = 'NEUTRONIC PRENEUTRONIC CORNEUTRONIC'     # 每个时间步开始时执行
  []
  
  [thermal]
    type = ThermalMultiApp
    app_type = MooseprojectsApp
    input_files = 'thermal_input.i'
    mesh_dims = '10 10 10'           # 保持一致
    power_var_name = 'power_density'  # 要与变量名一致
    temperature_var_name = 'temperature'  # 要与变量名一致
    execute_on = 'THERMAL'            # 使用标准执行标志
  []
[]

# [Transfers]
#   [from_neutronics]
#     type = ReactorTransfer
#     from_multi_app = neutronics
#     source_variable = 'power_density'
#     variable = 'power_density'
#     execute_on = 'timestep_end'       # 每个时间步结束时执行
#     transfer_group = 'from_neutronics'    # 分组名称，用于代码中引用
#   []
# []

# 基本网格和材料设置
[Mesh]
  type = GeneratedMesh
  dim = 3
  nx = 10
  ny = 10
  nz = 10
[]

[Variables]
  [dummy]
    family = LAGRANGE
    order = FIRST
  []
[]

[Kernels]
  [dummy]
    type = Diffusion
    variable = dummy
  []
[]

[Problem]
  type = FEProblem
  verbose_multiapps = true
  solve = False
  kernel_coverage_check = false
  material_coverage_check = false
[]

# 用于数据传输的辅助变量
[AuxVariables]
  [temperature]
    family = MONOMIAL
    order = CONSTANT
    initial_condition = 600.0
  []
  [power_density]
    family = MONOMIAL
    order = CONSTANT
    initial_condition = 0.0
  []
[]

[Executioner]
  type = Transient
  
  # 每个时间步对应一个燃耗步
  start_time = 0
  end_time = 1            # 最大燃耗步数
  dt = 1.0                # 固定时间步长
  
  num_steps = 5           # 最大步数，对应最大燃耗步数
  
  # 求解设置
  solve_type = 'PJFNK'
  nl_abs_tol = 1e-6
  nl_rel_tol = 1e-6
  l_tol = 1e-6
  l_max_its = 100
  nl_max_its = 50
[]

[Outputs]
  exodus = true
  console = true
[]

this is sub's input:

[Mesh]
  [generated]
    type = GeneratedMeshGenerator
    dim = 3
    nx = 10
    ny = 10
    nz = 10
  []
[]

[AuxVariables]
  [power_density]
    family = MONOMIAL
    order = CONSTANT
    initial_condition = 0.0
  []
  
  [temperature]
    family = MONOMIAL
    order = CONSTANT
    initial_condition = 300.0
  []
[]

[Problem]
  type = FEProblem
  solve = false  # 禁用 MOOSE 内置求解器，所有计算由外部处理
[]

[Executioner]
  type = Transient
  num_steps = 1  # 可以减少到1步，因为实际计算在外部完成
  dt = 1.0
[]

[Outputs]
  exodus = true
  console = true
[]

[GiudGiud]

this error is a crash, it cant come from the input card.
It comes from a problem in your code. Likely due to an error in memory management, which is odd because you are not supposed to do that yourself in modern C++

So use the debugging instructions to get the line of code where it crashes, and examine the creation of those objects
https://mooseframework.inl.gov/application_development/debugging.html