Neural Ordinary Differential Equations for Data-Driven Reduced Order Modeling of Environmental Hydrodynamics

Abstract

Model reduction for fluid flow simulation continues to be of great interest across a number of scientific and engineering fields. Here, we explore the use of Neural Ordinary Differential Equations, a recently introduced family of continuous-depth, differentiable networks (Chen et al 2018), as a way to propagate latent-space dynamics in reduced order models. We compare their behavior with two classical non-intrusive methods based on proper orthogonal decomposition and radial basis function interpolation as well as dynamic mode decomposition. The test problems we consider include incompressible flow around a cylinder as well as real-world applications of shallow water hydrodynamics in riverine and estuarine systems. Our findings indicate that Neural ODEs provide an elegant framework for stable and accurate evolution of latent-space dynamics with a promising potential of extrapolatory predictions. However, in order to facilitate their widespread adoption for large-scale systems, significant effort needs to be directed at accelerating their training times. This will enable a more comprehensive exploration of the hyperparameter space for building generalizable Neural ODE approximations over a wide range of system dynamics.

arXiv comment10 pages, 6 figures, To Appear in Proceedings of AAAI 2021 Spring Symposium on Combining Artificial Intelligence and Machine Learning with Physics Sciences, March 22-24, 2021

References in corpus (16)

• Discovering governing equations from data: Sparse identification of nonlinear dynamical systems
• Deep learning for universal linear embeddings of nonlinear dynamics
• Data-driven discovery of coordinates and governing equations
• The GNAT method for nonlinear model reduction: effective implementation and application to computational fluid dynamics and turbulent flows
• PDE-Net 2.0: Learning PDEs from Data with A Numeric-Symbolic Hybrid Deep Network
• Neural Ordinary Differential Equations
• Time-series learning of latent-space dynamics for reduced-order model closure
• Comparison of POD reduced order strategies for the nonlinear 2D Shallow Water Equations
• How to train your neural ODE: the world of Jacobian and kinetic regularization
• ANODE: Unconditionally Accurate Memory-Efficient Gradients for Neural ODEs
• Deep convolutional recurrent autoencoders for learning low-dimensional feature dynamics of fluid systems
• Augmented Neural ODEs
• A greedy non-intrusive reduced order model for shallow water equations
• Machine-learning error models for approximate solutions to parameterized systems of nonlinear equations
• Approximation Capabilities of Neural ODEs and Invertible Residual Networks
• Simple Video Generation using Neural ODEs