← Back to all news
arXiv Machine Learning

Towards Efficient Parametric State Estimation in Circulating Fuel Reactors with Shallow Recurrent Decoder Networks

This paper presents a novel Shallow Recurrent Decoder network for efficient parametric state estimation in circulating fuel reactors, specifically applied to the Molten Salt Fast Reactor (MSFR). The model infers the full reactor state—including neutron fluxes, precursor concentrations, temperature, pressure, and velocity—using only three out‑of‑core neutron flux time‑series measurements, while also handling parametric time‑series data to explore different accident scenarios. The approach demonstrates accurate real‑time reconstruction with low training cost and provides uncertainty quantification, making it suitable for monitoring and control within a reactor digital twin.

Related stories

arXiv Machine Learning
Sep 7

Constrained Sensing and Reliable State Estimation with Shallow Recurrent Decoders on a TRIGA Mark II Reactor

The paper introduces Shallow Recurrent Decoder (SHRED) networks as a data‑driven method for accurate state estimation in engineering systems, specifically applied to the TRIGA Mark II research reactor. SHRED maps sparse sensor measurements to the full state space, handling noisy data and requiring minimal training time. The study demonstrates SHRED’s performance using both synthetic CFD data and experimental temperature recordings, achieving low reconstruction errors and showcasing its potential for real‑time monitoring and digital twin development.

By Stefano Riva, Carolina Introini, Jos\`e Nathan Kutz, Antonio Cammi
More like this →
arXiv Machine Learning
Sep 1

The PUR-1 Cyber-Physical Digital Twin

arXiv:2608.30186v1 Announce Type: cross Abstract: Digital twin technologies have the potential to improve operational flexibility and responsiveness capabilities of nuclear systems. To provide decisi...

By Vasileios Theos, Jonah Lau, Konstantinos Gkouliaras, Zachery Dahm, Konstantinos Vasili, Noah Fillgrove, William Richards, True Miller, Brian Jowers, Stylianos Chatzidakis
benchmarks
More like this →
arXiv Machine Learning
Aug 27

Joint Initialization of Flux Networks and Effective Multiplication Factor for Physics-Informed Neural Networks Solving Neutron Diffusion Problems

The paper introduces JI-PINN, a joint initialization strategy for physics-informed neural networks that simultaneously initializes neutron flux and the effective multiplication factor (keff) using a low-resolution approximate solution. By jointly optimizing both quantities under physical constraints, the method achieves significant reductions in computational time—up to nearly 50%—across various benchmark neutron diffusion problems while maintaining accuracy. It also reduces anomalous keff deviations, offering a more robust approach to solving K‑eigenvalue problems with PINNs.

By Qin Hang, Yangdi Yi, Jiayi Li, Xu Wang, Heng Zhang
diffusionbenchmarks
More like this →
arXiv Machine Learning
Sep 22

Bayesian Filtering in Physical Systems via Test-time Trained Flow Matching

The paper introduces the Belief Flow Filter (BFF), a generative filtering framework that encodes the evolving posterior distribution directly into flow matching model weights and updates them via test‑time gradient descent. By avoiding particle representations and Gaussian assumptions, BFF aligns structurally with Bayesian filtering and targets the recursive filtering operator. Empirical results on five physical systems—including chaotic dynamics, sparse observations, and a tokamak plasma estimation task—show that BFF outperforms existing methods in most benchmark metrics.

By Ruiqi Feng, Chongyi Wang, Tao Zhang, Tailin Wu
diffusionbenchmarks
More like this →
arXiv Machine Learning
Sep 4

Equation Recast for Canonical Operator Learning Across Parametric PDEs

The paper introduces equation recast, a method that transforms parametric operator learning into learning a single canonical operator. By analytically deriving parameter-induced variations from the governing equations and incorporating them as effective sources, the approach enables zero‑shot predictions across new parameter regimes and supports extrapolation in multi‑parameter, nonlinear, and singular PDE settings. It also integrates sparse heterogeneous datasets, uses loss of convergence as an internal warning, and demonstrates unification of electron‑temperature data from multiple tokamak geometries in high‑fidelity nuclear fusion simulations.

By Qiyun Cheng, Valentin Duruisseaux, Cesar F. Clauser, Md Hossain Sahadath, Huihua Yang, Shaowu Pan, Nathaniel Ferraro, Anima Anandkumar, Wei Ji, Cristina Rea
More like this →