NuSHRED - Compressed Data

NuSHRED — Datasets for Shallow Recurrent Decoder applications to nuclear reactor systems

Compressed simulation datasets supporting the GitHub repository NuSHRED (Shallow Recurrent Decoder for Nuclear Reactor Applications).

These data accompany the application of the Shallow Recurrent Decoder (SHRED) method to several nuclear reactor systems, including out-core and in-core state estimation, verification & validation on an experimental facility, and multi-fidelity learning across low- and high-fidelity models.

Repository: https://github.com/ERMETE-Lab/NuSHRED
Download: datasets can be retrieved with Code/download_datasets.py in the repository.

Each item is provided as a compressed archive (*.zip) and corresponds to one subfolder after extraction.

[MSFR] — Molten Salt Fast Reactor (MSFR), accidental scenario Unprotected Loss Of Fuel Flow (ULOFF)

• Parametric transients from multi-physics OpenFOAM simulations (2D axisymmetric mesh)
• Includes the single-transient reconstruction case used in Paper 1
• Used in Papers 1, 2, and 5 (MSFR multi-fidelity case)

[DYNASTY] — DYNASTY experimental facility

• RELAP5 model data: single transient (reconstruction & prediction) and parametric transients
• Used in Paper 3

[TRIGA] — TRIGA Mark II reactor

• Single transient from a CFD model (Introini et al., 2018)
• Reconstruction mode
• Used in Paper 4

[LRA-neutronics] — LRA 2D benchmark reactor

• Neutronics snapshots from multigroup diffusion and point-kinetics models
• Used in Paper 5 (neutronics multi-fidelity case)

[RDA] — Non-linear reaction–diffusion–advection system

• Multiple chemical species
• High-fidelity PDE model and low-fidelity ODE (lumped) model
• Used in Paper 5 (reaction–diffusion–advection multi-fidelity case)

Legacy naming (previous Zenodo release)

Related publications

• P1 — Robust state estimation from partial out-core measurements (Progress in Nuclear Energy, 2025)
• P2 — Parametric state estimation in circulating fuel reactors (arXiv:2503.08904)
• P3 — SHRED on the DYNASTY experimental facility (arXiv:2503.08907)
• P4 — Constrained sensing on a TRIGA Mark II reactor (arXiv:2510.12368)
• P5 — Multi-fidelity learning with SHRED (arXiv:2606.05202)