Physics-Informed Neural Networks for Two-Phase Flow Simulations: An Integrated Approach with Advanced Interface Tracking Methods

Abstract

Physics-informed neural networks (PINNs) are emerging as one of the promising artificial intelligence approaches for solving complex thermal-hydraulic simulations. A critical challenge in these simulations is an accurate representation of the gas-liquid boundary with different interface tracking methods. While numerous studies in conventional computational fluid dynamics (CFD) have addressed this issue, there remains a notable absence of research within the context of PINNsbased two phase flow simulations. Therefore, this study aims to develop a robust and generic PINNs for two-phase flow by incorporating governing equations and three advanced interface tracking methods—specifically, the Volume of Fluid, Level Set, and Phase-Field methods—into the adaptive residual differential (ARD) PINNs framework that has been previously proposed and validated. To further enhance the performance of the PINNs in simulating two-phase flow, the phase field constraints and the time divide-and-conquer strategies are employed for restricting neural network training within the scope of physical laws. The ARD PINNs then is optimized by minimizing both the residual and loss terms of partial differential equation. By incorporating the three different interface tracking methods, it efficiently handles high-order derivative terms and captures the phase interface. The case of single rising bubble in two-phase flow is simulated to validate the robustness and accuracy of the ARD PINNs. The accuracy of the simulation is compared with the velocity, pressure, and phase field obtained from the CFD simulation. The results indicate that the ARD PINNs coupled with these interface tracking methods offers a satisfactory performance in simulating rising bubble phenomenon.