Learnable-Differentiable Finite Volume Solver for Accelerated Simulation of Flows

Abstract

Simulation of fluid flows is crucial for modeling physical phenomena like meteorology, aerodynamics, and biomedicine. Classical numerical solvers often require fine spatiotemporal grids to satisfy stability, consistency, and convergence conditions, leading to substantial computational costs. Although machine learning has demonstrated better efficiency, they typically suffer from issues of interpretability, generalizability, and data dependency. Hence, we propose a learnable and differentiable finite volume solver, called LDSolver, designed for efficient and accurate simulation of fluid flows on spatiotemporal coarse grids. LDSolver comprises two key components: (1) a differentiable finite volume solver, and (2) an learnable module providing equivalent approximation for fluxes (derivatives and interpolations), and temporal error correction on coarse grids. Even with limited training data (e.g., only a few trajectories), our model could accelerate the simulation while maintaining a high accuracy with superior generalizability. Experiments on different flow systems (e.g., Burgers, decaying, forced and shear flows) show that LDSolver achieves state-of-the-art performance, surpassing baseline models with notable margins.

19 pages, 12 figures, accepted at KDD 2025 (ACM SIGKDD Conference on Knowledge Discovery and Data Mining)

References in corpus (14)

• DeepONet: Learning nonlinear operators for identifying differential equations based on the universal approximation theorem of operators
• PhyGeoNet: Physics-Informed Geometry-Adaptive Convolutional Neural Networks for Solving Parameterized Steady-State PDEs on Irregular Domain
• PDE-Net 2.0: Learning PDEs from Data with A Numeric-Symbolic Hybrid Deep Network
• Learning data driven discretizations for partial differential equations
• Physics-Informed Multi-LSTM Networks for Metamodeling of Nonlinear Structures
• Dedalus: A Flexible Framework for Numerical Simulations with Spectral Methods
• PhyCRNet: Physics-informed Convolutional-Recurrent Network for Solving Spatiotemporal PDEs
• Deep transfer operator learning for partial differential equations under conditional shift
• Multi-resolution partial differential equations preserved learning framework for spatiotemporal dynamics
• Learned discretizations for passive scalar advection in a 2-D turbulent flow
• Multiwavelet-based Operator Learning for Differential Equations
• Physics informed deep learning for computational elastodynamics without labeled data
• Thermodynamics-informed neural networks for physically realistic mixed reality
• Hypersolvers: Toward Fast Continuous-Depth Models