본문 바로가기

Physics-guided deep learning surrogate model with graph attention for long-term radionuclide transport prediction in deep geological repositories

문서 유형
연구출판물 · 학술지 논문
등재
SCIEScopus
저널 순위
JCR Q1Scopus Q1
피인용
Web of Science 4Scopus 4
FWCI (Scopus)
4.75
JCR IF
JCR 2024 11.3
ISSN
0304-3894

저자

  1. Jeong, Dae Seong제1저자Korea Univ, Future & Fus Lab Architectural Civil & Environm En, Seoul 02841, South Korea
  2. Lee, JinukUlsan Natl Inst Sci & Technol, Dept Civil Urban Earth & Environm Engn, Ulsan 44919, South Korea
  3. 표종철Pusan Natl Univ, Dept Environm Engn, Busan 46241, South Korea
  4. Baek, Sang-SooYeungnam Univ, Dept Environm Engn, Gyongsan 38541, South Korea
  5. Kim, Jin HwiKorea Univ, Future & Fus Lab Architectural Civil & Environm En, Seoul 02841, South Korea
  6. Jeong, Mi-SeonKorea Radioact Waste Agcy, HLW Technol Dev Inst, 174 Gajeong Ro, Daejeon 34129, South Korea
  7. Yun, HyungjuKorea Radioact Waste Agcy, HLW Technol Dev Inst, 174 Gajeong Ro, Daejeon 34129, South Korea
  8. Cho, Kyung Hwa교신저자Korea Univ, Sch Civil Environm & Architectural Engn, Seoul 02841, South Korea

초록

To ensure millennial isolation of high-level waste, Deep geological repositories (DGR) safety assessments depend on process-based simulators such as Parallel flow and reactive transport model (PFLOTRAN). Yet, without feasible long-term monitoring, the high computational burden of these models becomes a bottleneck for iterative scenario testing and policy decisions. To overcome this, we developed a graph attention-based physics-guided deep learning (GAT-PGDL) surrogate that embeds decay–diffusion–sorption equations. U-238 and Th-230 transport was simulated for 5000 years in DGRs, with release commencing at year 2000 across ten monitoring nodes. On a single-node workstation, PFLOTRAN requires ∼5760 min per scenario, whereas the GAT-PGDL trains once (∼94 min) and infers in seconds, delivering ∼61 × per-scenario speedup. To assess reliability and interpretability, we employed split conformal prediction for uncertainty quantification, perturbation-based sensitivity analysis, and feature importance analysis. The GAT-PGDL produced 95 % prediction intervals and highlighted sorption and bulk density as key transport controls. For generalization, performance was compared with a data-driven surrogate under scenarios with altered material properties and earlier release times, where the GAT-PGDL outperformed the data-driven model, maintaining R² and NSE > 0.98. These results establish GAT-PGDL as a fast, accurate, and physically reliable surrogate for long-term DGR safety assessments.

키워드

Deep geological repository safety assessmentPFLOTRANPhysics-guided deep learningRadionuclide transportSurrogate modeling

식별자

Scopus EID
2-s2.0-105024757841
WoS UID
001644497800001