PRCT-Net: A Transfer-Learning Physics-Informed Convolutional Transformer for Fast Surrogate Prediction of Confined Dual-Cylinder Three-Dimensional Free-Surface Two-Phase Flow
Journal of Fluids Engineering · 01.11.2026 · 148 (11)
DOI: 10.1115/1.4072333 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Fluids Engineering, online erschienen am 30.07.2026. Verfasst von Xiao Zhang und 4 weiteren Autor:innen. Die Arbeit nennt 34 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Wave-induced free-surface two-phase flows around cylindrical obstacles are commonly encountered in many engineering contexts and provide a useful canonical configuration for evaluating data-driven flow-field prediction methods. In this study, we focus on a geometrically confined dual-cylinder traveling-wave configuration as a benchmark problem for three-dimensional gas–liquid two-phase flow prediction. The results show that A transfer-learning Physics-informed convolutional transformer (PRCT)-net improves computational efficiency while maintaining high prediction accuracy for the tested dual-cylinder traveling-wave cases. This study should be interpreted as a controlled benchmark for a confined dual-cylinder traveling-wave two-phase-flow configuration. The results demonstrate that PRCT-net enables efficient adaptation and interpolation/extrapolation within the investigated cylinder-spacing range. However, this validation does not establish universal generalization to different wave velocities, cylinder diameters, or domain geometries, which require further systematic investigation.
Zitieren
APA
Zhang, X., Zhao, Q., Zeng, X., Qin, M., & Liu, H. (2026). PRCT-Net: A Transfer-Learning Physics-Informed Convolutional Transformer for Fast Surrogate Prediction of Confined Dual-Cylinder Three-Dimensional Free-Surface Two-Phase Flow. Journal of Fluids Engineering, 148(11). https://doi.org/10.1115/1.4072333
Vancouver
Zhang X, Zhao Q, Zeng X, Qin M, Liu H. PRCT-Net: A Transfer-Learning Physics-Informed Convolutional Transformer for Fast Surrogate Prediction of Confined Dual-Cylinder Three-Dimensional Free-Surface Two-Phase Flow. Journal of Fluids Engineering. 2026;148(11). doi:10.1115/1.4072333
BibTeX
@article{Zhang2026,
author = {Zhang, Xiao and Zhao, Qingkai and Zeng, Xinyao and Qin, Mengqi and Liu, Haoran},
title = {PRCT-Net: A Transfer-Learning Physics-Informed Convolutional Transformer for Fast Surrogate Prediction of Confined Dual-Cylinder Three-Dimensional Free-Surface Two-Phase Flow},
journal = {Journal of Fluids Engineering},
year = {2026},
volume = {148},
number = {11},
doi = {10.1115/1.4072333}
}
EUVIMED