Reproducing results of TRISO multiphysics tutorial

[nancysemwal]

Hello.

I am following the TRISO gas cooled compact multiphysics tutorial and want to reproduce the temperature plot. I have only run the openmc-nekrs simulation, not the openmc-THM one.

Here is the temperature plot in the tutorial versus the plot generated by me:

I have 2 questions regarding the differences between the 2 plots:

1. The tutorial is getting a peak fuel temp > 1100 K, while mine lies somewhere between 1000-1100 K. Is a difference of ~100 K acceptable for this problem setup? I am still developing my understanding of the physics involved and would guess that in a separate run, maybe I will also get a similar peak temperature. Am I right?

2. My plot for the wall temp is not at all matching with the tutorial's plot. For the wall temp, I have plotted the nek_temp variable since the tutorial states:

Next, we define a number of auxiliary variables - power will receive the heat source from OpenMC, nek_temp will receive the wall temperature from NekRS, nek_bulk_temp will receive the volumetric fluid temperature from NekRS

But I don't get the same plot profile.

Since I get correct plot profile & min max values for bulk and graphite temperatures, so I'm guessing I should get the other 2 correct as well.

Can anybody provide some insight?

[aprilnovak]

Hi @nancysemwal -- the setup is completely fine, but I'm glad you are checking this detail! About a year ago, we changed the value of thermal conductivity used in this tutorial (as well as the other TRISO tutorials in Cardinal) because we realized it was lower than it should be. We were lazy though, and didn't update any of the figures associated with the tutorials -- what you are seeing is expected.

For your second question, the nek_temp variable is indeed the wall temperature, but it is azimuthally defined around the circumference of the cooling channel. Now that I'm looking at these input files, I think they actually do not contain the input file syntax you'd need to plot that quantity, solely as a function of z (basically, you want to integrate nek_temp around the circumference of the cooling channel, and then plot that quantity as a function of z). These plots were made as part of a journal article, and are only loosely tied to the tutorial -- sorry about that!

To plot the wall temperature from Nek, as a function of z, please try

[UserObjects]
  [average_wall_temp_axial]
    type = LayeredAverage
    variable = nek_temp
    direction = z
    num_layers = ${num_layers_for_plots}
    block = ${fuel_blocks}
  []
[]

[VectorPostprocessors]
  [wall_temp_avg]
    type = SpatialUserObjectVectorPostprocessor
    userobject = average_wall_temp_axial
  []
[]

[nancysemwal]

Okay, I understand what you're trying to say. Thank you for your input.

Just to confirm my understanding :

1. For the required plot, I have added these 2 [UserObjects] and [VectorPostprocessors] blocks to the openmc_nek.i file. Or should I have added it to the solid_nek.i file?
2. I think these 2 above changes would require me to run the whole simulation again from the beginning. Or just for merely plotting the wall_temp axially like I want, I can make use of the the various output files from the already completed simulations (the various .e and nek .f* files)? Maybe taking tips from the restart_nek_and_moose tutorial?

[aprilnovak]

1. You'll want to add them in whichever file nek_temp exists as a variable (the solid_nek.i file).
2. You're correct, if you just add these userobjects as-is you would need to re-run the case. Alternatively, if you have the exodus output files, you could accomplish the same thing by loading in the existing Exodus file output. I think you'll want to use SolutionUserObject to read in a field variable from an output file. You'll basically read nek_temp from the output file into a "living" variable via SolutionUserObject, and then you can perform the same integration by adding

[UserObjects]
  [average_wall_temp_axial]
    type = LayeredAverage
    variable = nek_temp
    direction = z
    num_layers = ${num_layers_for_plots}
    block = ${fuel_blocks}
  []
[]

[VectorPostprocessors]
  [wall_temp_avg]
    type = SpatialUserObjectVectorPostprocessor
    userobject = average_wall_temp_axial
  []
[]

Please let me know if that works, or if more specific instructions would help!

[nancysemwal]

Thanks. I have decided to re-run the simulation and explore the [SolutionUserObject] path later.

However, when I add the above [UserObjects] and [VectorPostprocessors] to the solid_nek.i file to get the axial wall temp, I get this error:

The 'block' parameter of the object 'average_wall_temp_axial' must be a subset of the 'block' parameter of the variable 'nek_temp':
       Object 'average_wall_temp_axial': 2, 1002
       Variable 'nek_temp': 101

I can understand this error to some extent, because the input file also contains this :

[AuxVariables]
  [nek_temp]
    block = ${density_blocks}
  []
 []

which states that nek_temp is defined with density_blocks.

So I think in average_wall_temp_axial [UserObjects], instead of block = ${fuel_blocks}, it should be block = ${density_blocks}?

Or am I getting it all wrong?