MultiApp transfer variable to initial condition of subapp

[garcs2]

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 everyone,

I am working on making a neutronics-heat transfer-thermohydraulics coupled simulation, using Griffin as the neutronics solver and heat transfer-thermohydraulics (HT-TH) being solved with the base modules in MOOSE. I have Griffin and HT-TH running successfuly separately. The HT-TH is solved using TransientMultiApp where HT acts as the main app and TH is the sub-app. Ideally, the workflow would look like the following:

For clarity, the FullSolveMultiApp executes on timestep_end so that Griffin solves first, then transfers the power density to HT-TH where HT solves first. My question is then: how do I designate the initial condition of HT to be the transferred power density? Since the power density is a scalar, I would like to use ConstantIC but a value is required as input. Below is the relevant portion of my input file, with the IC block not filled in since I don't know what to do:

[Variables]
  [T]
  []
[]
[AuxVariables]
  [fluid_temp]
  initial_condition = '${inlet_T}'
  []
  [flux]
    family = MONOMIAL
    order = CONSTANT
  []
  [power_density]
    family = L2_LAGRANGE
    order = First
    block = ${fuel_blocks}
  []
[]
[Kernels]
  [diffusion]
    type = HeatConduction
    variable = T
  []
  [source]
    type = CoupledForce
    variable = T
    v = power_density
    block = ${fuel_blocks}
  []
[]

[BCs]
  [pin_outer]
    type = MatchedValueBC
    variable = T
    v = fluid_temp
    boundary = 'fluid_solid_interface'
  []
[]

[ICs]
  [power_density]
    type = ConstantIC
    variable = power_density
    value = ???
    block = ${fuel_blocks}
  []
[]

[GiudGiud]

hello

Instead of using a ConstantIC, if you want a value from another application, you should transfer it.
Transfers can be executed on INITIAL.

but here you should be good already?

the FullSolveMultiApp executes on timestep_end so that Griffin solves first, then transfers the power density to HT-TH where HT solves first

since Griffin executes first and sends its power field to the HT solve, the HT solve has the updated power density already for its first solve?

Guillaume

[garcs2]

I tried MULTIAPP_FIXED_POINT_END and I'm still not getting the transfer to pass on properly, I also tried different configurations between TIMESTEP_END and MULTIAPP_FIXED_POINT_END between the main app and sub app and haven't gotten the transfers to function the way that I would like to. Given this issue I'm considering having the scheme shift in the other direction, where HT-TH solve first (with the power density set in HT) and then pass that information onto Griffin. I figure this would work as when I explicitly provide an IC HT-TH solve just fine, but I'm still curious why the transfer isn't solving the way that I was originally intending.

[GiudGiud]

The root cause is that Griffin has two options for nesting multiphysics fixed point iterations and only one is behaving as expected with regards to execution flags.

You should try switching the multiapps iteration in the Richardson executioner.

[garcs2]

Okay so I played with the Richardson executioner to allow for more iterations for Griffin to effectively solve first before transferring. So I'm now seeing that Griffin completes and then the transfer occurs:

Postprocessor Values:
+----------------+----------------+----------------+------------------+----------------+
| time           | eigenvalue     | griffin_power  | integrated_power | power_scaling  |
+----------------+----------------+----------------+------------------+----------------+
|   0.000000e+00 |   0.000000e+00 |   6.000000e+05 |     6.000000e+05 |   6.095066e+28 |
|   1.000000e+00 |   1.029158e+00 |   6.000000e+05 |     6.000000e+05 |   4.614618e+15 |
+----------------+----------------+----------------+------------------+----------------+

  Currently Outputting
    Outputting Step
      Backing Up Application...Backed up MultiApp ... Griffin_htm

      Finished Backing Up Application                                                    [�[33m 31.00 s�[39m] [�[33m13949 MB�[39m]
    Finished Outputting Step                                                             [�[33m 31.72 s�[39m] [�[33m13949 MB�[39m]
  Finished Outputting                                                                    [�[33m 32.58 s�[39m] [�[33m13949 MB�[39m]
Still Executing�[36m
Transfers on MULTIAPP_FIXED_POINT_END To MultiApps�[39m:
----------------------------------------------------------------------------
|         Name         |            Type            |  From  |     To      |
----------------------------------------------------------------------------
| to_htm_power_density | MultiAppProjectionTransfer | Parent | Griffin_htm |
----------------------------------------------------------------------------

  Currently Executing Transfers
    Transferring variables through projection............                                [�[33m 71.00 s�[39m] [�[33m14524 MB�[39m]
�[36mTransfers on MULTIAPP_FIXED_POINT_END Are Finished
�[39m
�[36m
No Transfers on MULTIAPP_FIXED_POINT_END Between MultiApps
�[39m
�[36m
Executing MultiApps on MULTIAPP_FIXED_POINT_END�[39m
  Finished Executing Transfers                                                           [�[33m 71.01 s�[39m] [�[33m14524 MB�[39m]
Warning, Exodus files cannot have titles longer than 80 characters.  Your title will be truncated.
�[36mGriffin_htm0: �[39mOutputting�[36mGriffin_htm0: �[39m
�[36mGriffin_htm0: �[39m  Outputting Step                                                          [�[33m 10.75 s�[39m] [�[33m14882 MB�[39m]
�[36mGriffin_htm0: �[39m
�[36mGriffin_htm0: �[39mTime Step 0, time = 0
�[36mGriffin_htm0: �[39m
�[36mGriffin_htm0: �[39mPostprocessor Values:
�[36mGriffin_htm0: �[39m+----------------+----------------+----------------+----------------+----------------+----------------+----------------+----------------+
�[36mGriffin_htm0: �[39m| time           | flux_integral  | htm_Tfuel      | htm_Tref       | max_fuel_T     | max_power      | min_power      | power_source   |
�[36mGriffin_htm0: �[39m+----------------+----------------+----------------+----------------+----------------+----------------+----------------+----------------+
�[36mGriffin_htm0: �[39m|   0.000000e+00 |   0.000000e+00 |   0.000000e+00 |   0.000000e+00 |   0.000000e+00 |   0.000000e+00 |   0.000000e+00 |   0.000000e+00 |
�[36mGriffin_htm0: �[39m+----------------+----------------+----------------+----------------+----------------+----------------+----------------+----------------+

So it seems that despite Griffin executing and finishing first completely, the transfer is still not being done correctly.

[garcs2]

Hi Guillaume, it turns out the primary issue was of meshing location not matching exactly. I thought I had resolved this issue but clearly not. After making an adjustment the transfer works. Thank you for your patience and consistent replies. I will mark the earlier answer as resolved regarding the IC's.

[GiudGiud]

ok I'm glad