{"paper":{"title":"Graph Neural ODE Digital Twins for Control-Oriented Reactor Thermal-Hydraulic Forecasting Under Partial Observability","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"A graph neural network with neural ODE dynamics forecasts reactor thermal-hydraulic states accurately at locations without sensors and adapts to real data.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Akzhol Almukhametov, Doyeong Lim, Rui Hu, Yang Liu","submitted_at":"2026-04-08T16:58:14Z","abstract_excerpt":"Real-time supervisory control of advanced reactors requires accurate forecasting of plant-wide thermal-hydraulic states, including locations where physical sensors are unavailable. Meeting this need calls for surrogate models that combine predictive fidelity, millisecond-scale inference, and robustness to partial observability. In this work, we present a physics-informed message-passing Graph Neural Network coupled with a Neural Ordinary Differential Equation (GNN-ODE) to addresses all three requirements simultaneously. We represent the whole system as a directed sensor graph whose edges encod"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"The GNN-ODE surrogate achieves an average MAE of 0.91 K at 60 s and 2.18 K at 300 s for uninstrumented nodes, with R² up to 0.995 for missing-node state reconstruction; after fine-tuning on 30 experimental sequences the learned flow-dependent heat-transfer scaling recovers a Reynolds-number exponent consistent with established correlations.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the directed sensor graph and topology-guided initializer faithfully encode the true hydraulic connectivity and that the physics-informed message passing plus Neural ODE can generalize from simulation transients to real experimental data without introducing systematic bias in the recovered constitutive relation.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A GNN-ODE digital twin forecasts reactor thermal-hydraulic states under partial observability, achieving low error on held-out transients and recovering a physical heat-transfer correlation during sim-to-real adaptation.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A graph neural network with neural ODE dynamics forecasts reactor thermal-hydraulic states accurately at locations without sensors and adapts to real data.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"74d7f00bd1701c8bf013fb27a5a452544c0d60ef1dfc0fb48f9f26e6b2950c18"},"source":{"id":"2604.07292","kind":"arxiv","version":2},"verdict":{"id":"f5efb809-a2ac-416d-853b-30cdb3fb96c2","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T18:25:01.542534Z","strongest_claim":"The GNN-ODE surrogate achieves an average MAE of 0.91 K at 60 s and 2.18 K at 300 s for uninstrumented nodes, with R² up to 0.995 for missing-node state reconstruction; after fine-tuning on 30 experimental sequences the learned flow-dependent heat-transfer scaling recovers a Reynolds-number exponent consistent with established correlations.","one_line_summary":"A GNN-ODE digital twin forecasts reactor thermal-hydraulic states under partial observability, achieving low error on held-out transients and recovering a physical heat-transfer correlation during sim-to-real adaptation.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the directed sensor graph and topology-guided initializer faithfully encode the true hydraulic connectivity and that the physics-informed message passing plus Neural ODE can generalize from simulation transients to real experimental data without introducing systematic bias in the recovered constitutive relation.","pith_extraction_headline":"A graph neural network with neural ODE dynamics forecasts reactor thermal-hydraulic states accurately at locations without sensors and adapts to real data."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.07292/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":2,"snapshot_sha256":"c9b150c366d480179cb032223bd6576c1b587cff8cd56a76a94fb782dbb9d8cb"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}