LSTM-PINN: An Hybrid Method for Prediction of Steady-State Electrohydrodynamic Flow

Ze Tao, Ke Xu, Fujun Liu

arXiv:2512.21614·physics.flu-dyn·Published 2025-12-25

Physics-Informed Neural Networks (PINNs) have demonstrated considerable success in solving complex fluid dynamics problems. However, their performance often deteriorates in regimes characterized by steep gradients, intricate boundary conditions, and stringent physical constraints, leading to convergence failures and numerical instabilities. To overcome these limitations, we propose a hybrid framework that integrates Long Short-Term Memory (LSTM) networks into the PINN architecture, enhancing its ability to capture spatial correlations in the steady-state velocity field of a two-dimensional charged fluid under an external electric field. Our results demonstrate that the LSTM-enhanced PINN model significantly outperforms conventional Multilayer Perceptron (MLP)-based PINNs in terms of convergence rate, numerical stability, and predictive accuracy. This innovative approach offers improved computational efficiency and reliability for modeling electrohydrodynamic flows, providing new insights and strategies for applications in microfluidics and nanofluidics.

TopicsScientific Machine Learning & PINNs

Tagsphysics-informed-neural-networks pinns

arXiv categoriesphysics.flu-dyn, math.AP

arXiv abstract pagePDF