Code · Zenodo · September 23, 2026
Artificial-viscosity schedules for Buckley–Leverett PINNs: code and reproducibility archive
Saltanbek Talapedenovich MUKHAMBETZHANOV, Yerzhan Kenzhebek, Samson Dawit Bekele, Saida Tastanova, Dagmawi Lemma, Timur Imankulov
Abstract
Software and experimental evidence accompanying Separating Artificial-Viscosity Annealing from the L-BFGS Viscosity in Physics-Informed Neural Networks for Two-Phase Flow. Includes training configurations, saved model weights, predictions, per-run measurements, finite-volume reference fields, and scripts to reconstruct all 15 figures and 18 tables. Separate exploratory open-outlet and boundary-conditioned experiment blocks use 15 paired seeds. Fixed-seed GPU training is not bitwise reproducible; the archive records the measured repeatability limitation. Funded by the Committee of Science of the Ministry of Science and Higher Education of the Republic of Kazakhstan, grant AP23489431.
More by Samson Dawit Bekele
- Separating Artificial-Viscosity Annealing from the L-BFGS Viscosity in Physics-Informed Neural Networks for Two-Phase FlowBig Data and Cognitive Computing · 2026
- Identifiability and Conditioning of Flux-Parameter Inversion for the Buckley–Leverett EquationMathematics · 2026
- Physics-Informed Neural Networks for Coupled Non-Isothermal Two-Phase Flow in Porous Media: Training Pathologies, Remedies, and the Role of Gradient ConflictMachine Learning and Knowledge Extraction · 2026
- Reference dataset and MRST simulation script for "Entropy-Satisfying Physics-Informed Neural Networks for the Two-Dimensional Buckley–Leverett Equation"Zenodo · 2026