Spatio-temporal graph neural network for multi-step prediction of cylinder water entry flow fields
- 1 November 2025
- journal article
- Published by AIP Publishing in Physics of Fluids
- Vol. 37 (11)
Abstract
This study introduces a Spatio-Temporal Graph Neural Network (ST-GNN) for the multi-step prediction of highly dynamic cylinder water entry flow fields. The proposed model integrates a Graph Neural Network (GNN) to capture complex spatial dependencies on unstructured meshes with a Long Short-Term Memory (LSTM) network to model temporal evolution. We generate a dataset from numerical simulations across different cases, varying both Froude numbers and initial entry angles. Extensive ablation studies are conducted to systematically optimize the model's architecture and hyperparameters, revealing that a single-step prediction strategy with an input sequence of eight timesteps yields the most robust performance, significantly reducing error accumulation in long-term prediction. The optimized model achieves high accuracy (R2>0.97) and successfully predicts key physical phenomena during water entry. Furthermore, generalization tests on unseen cases demonstrate the model's excellent interpolation capability within the training data range and reveal the challenges of extrapolation, particularly for the complex phase interface. Crucially, while the multi-step prediction of the velocity field degrades due to error accumulation, the inferred prediction of the water volume fraction remains remarkably stable. This indicates that the model has learned not just the instantaneous spatial correlations between the fields, but more importantly, has captured the underlying temporal dynamics that govern the evolution of the phase interface.Keywords
Funding Information
- National Natural Science Foundation of China (52325108)
- National Natural Science Foundation of China (52431010)
This publication has 48 references indexed in Scilit:
- Mesh-based super-resolution of fluid flows with multiscale graph neural networksComputer Methods in Applied Mechanics and Engineering, 2025
- Hybrid LSTM+CNN architecture for unsteady flow predictionMaterials Today Communications, 2023
- Dynamics analysis of high-speed water entry of axisymmetric body using fluid-structure-acoustic coupling methodJournal of Fluids and Structures, 2022
- Recent progress of machine learning in flow modeling and active flow controlChinese Journal of Aeronautics, 2022
- Numerical simulation of the effect of waves on cavity dynamics for oblique water entry of a cylinderJournal of Hydrodynamics, 2020
- A deep-learning-based surrogate model for data assimilation in dynamic subsurface flow problemsJournal of Computational Physics, 2020
- Active flow control using machine learning: A brief reviewJournal of Hydrodynamics, 2020
- Numerical investigations on the oblique water entry of high-speed projectilesApplied Mathematics and Computation, 2019
- Prediction of aerodynamic flow fields using convolutional neural networksComputational Mechanics, 2019
- Large eddy simulation and experimental investigation on the cavity dynamics and vortex evolution for oblique water entry of a cylinderApplied Ocean Research, 2018