Merging MSRE Xe Poisoning Model

[hapfang]

The PR related to #768 is now ready for review and merging.

[GiudGiud]

that's a lot less than PWRs? is that a core size thing or strictly the external removal?

[hapfang]

Indeed, i think it is probably due to the combination of both. the value is also close in a parallel SAM-Griffin modeling study

[GiudGiud]

what is the penalty now with the fixes in the model

[hapfang]

🚀 PR Update: Refactoring & Review Comments Addressed

Hi @GiudGiud,

I have completed the refactoring of the MSRE Xenon poisoning model and addressed your feedback on the input files. The model is now much leaner and fully utilizes the modern [Physics] syntax.

🛠️ Physics Syntax Migration

• Thermal-Hydraulics (th_xe.i): Migrated the [NavierStokesFV] block to the new syntax. It is now split into [Physics]/[NavierStokes]/[Flow] and [FluidHeatTransfer], effectively removing the deprecated add_energy_equation = true parameter.
• Scalar Transport: Replaced manual [Variables] and [FVKernels] for delayed neutron precursors ($c_1–c_6$) and isotopes ($I_{135}, Xe_{135}$) with the [ScalarTransport] system. This handles advection and diffusion automatically and significantly simplifies the input file.

⚙️ Logic & Parameter Adjustments

• Consistent Scaling: I have updated consistent_scaling to 1.0. This ensures a more physical representation of the system while maintaining numerical stability within the new physics blocks.
• Pressure Pinning: I tested the transition from point-value-uo to point-value. However, this change resulted in a noticeable degradation of convergence speed, so I have disregarded this change to maintain model performance.
• Cleanup: Removed redundant parameters in the [MultiApps] block (update_old_solution_when_keeping_solution_during_restore and enable = true).

📝 Documentation & Metadata

• Tagging & Status: Updated the model name to Molten Salt Reactor Experiment MultiPhysics Primary loop Model. Changed the V&V status to "demonstration" and refined transient tags to include ;Xenon.
• Technical Clarity: * Added source attribution for the 16-group cross sections (Mauricio's model).
  • Clarified that cross-section updates account for temperature feedback and DNP advection.
  • Added a section explaining why the $-410$ pcm reactivity penalty is lower than a typical PWR (noting online gas stripping and lower power density).
• Terminology: Standardized all mentions of "Master/Sub-app" to "Parent/Child" and fixed the "Plenum" typos.

The model is now updated and ready for further review. All tests have been verified against the new syntax.

[hapfang]

Hi @GiudGiud ,

I’ve addressed the review comments for PR #788. Below is a summary of the updates and a few notes on why certain refactoring suggestions were not implemented in this iteration.

Completed Updates

• Documentation & Accuracy: * Updated repository links and refined the physics scope in msre_xe.md to explicitly define advection, diffusion, and gaseous fission product removal behavior.

  • Corrected salt property descriptions and added a note regarding the explicit TH-neutronics coupling.
  • Applied the preferred "gaseous fission product removal" terminology throughout.

• Model Simplification:

  • Refined the pump model to a single-direction force (pump_y) and removed unnecessary top boundary symmetry conditions in the FVBCs.
  • Removed redundant vel_x and vel_y fields and their associated transfers since they are not required by Griffin

• Scripts & Testing:

  • Added the Python plotting script for Xe concentration and updated the corresponding results figure.
  • Fixed the tests syntax by updating the executable_pattern to support the BlueCRAB coupled environment.

Technical Notes on Refactoring

There are three areas where I have opted to maintain the current manual implementation:

1. Scalar Transport Refactor: I attempted to use the compact ScalarTransport syntax, but encountered issues with strict list-size mapping and the sequential $^{135}\text{I} \rightarrow ^{135}\text{Xe}$ coupling. I have kept the manual FVKernels to ensure technical accuracy and stability.
2. Convection Kernels: I’ve left the ambient convection kernels as-is for now. Given that the integrated physics-based syntax is still relatively new, I prefer the current stable implementation for this initial VTB release.
3. AuxVars Output: The outputs = 'exodus' feature within FunctorMaterials is not currently supported by the version of BlueCRAB available on the Sawtooth cluster. Thus, I have retained the existing output implementation for now, as I am unable to verify the newer FunctorMaterials output feature within my current testing environment.

The model is now in a stable, cleaned-up state. Please let me know if you have any further suggestions before we move toward merging.

[GiudGiud]

what is the penalty now with the fixes in the model

[GiudGiud]

Refined the pump model to a single-direction force (pump_y) and removed unnecessary top boundary symmetry conditions in the FVBCs.

I'm afraid this might have changed the flow rate? We should check which change impacted the results so much.
I imagine the pump force was selected to meet a certain mass flow rate. We need to have the right mdot.
This is actually fairly easily to tune using the PIDTransientControl btw

[hapfang]

Hi @GiudGiud, thanks for the thorough review. I will probably have to wait until the weekend to get time to address them.

[GiudGiud]

Works for me.
Additional comment now that I have the PR locally is that the files have executable permissions which is odd.

chmod -x <each file>

will fix that

EDIT: fixed in PR

[moosebuild]

Job VTB Documentation, step Sync Docs on 3bca4c2 wanted to post the following:

View the site here

This comment will be updated on new commits.