Learning Interpretable and Thermodynamically Stable Partial Differential Equations

09/28/2020
by   Juntao Huang, et al.
0

In this work, we develop a method for learning interpretable and thermodynamically stable partial differential equations (PDEs) based on the Conservation-dissipation Formalism of irreversible thermodynamics. As governing equations for non-equilibrium flows in one dimension, the learned PDEs are parameterized by fully-connected neural networks and satisfy the conservation-dissipation principle automatically. In particular, they are hyperbolic balance laws. The training data are generated from a kinetic model with smooth initial data. Numerical results indicate that the learned PDEs can achieve good accuracy in a wide range of Knudsen numbers. Remarkably, the learned dynamics can give satisfactory results with randomly sampled discontinuous initial data although it is trained only with smooth initial data.

READ FULL TEXT

Authors

page 1

page 2

page 3

page 4

08/16/2021

Partial differential equations on hypergraphs and networks of surfaces: derivation and hybrid discretizations

We introduce a general, analytical framework to express and to approxima...
04/16/2021

Finite Difference Nets: A Deep Recurrent Framework for Solving Evolution PDEs

There has been an arising trend of adopting deep learning methods to stu...
02/25/2021

SPINN: Sparse, Physics-based, and Interpretable Neural Networks for PDEs

We introduce a class of Sparse, Physics-based, and Interpretable Neural ...
04/28/2022

GCN-FFNN: A Two-Stream Deep Model for Learning Solution to Partial Differential Equations

This paper introduces a novel two-stream deep model based on graph convo...
05/20/2019

ExaHyPE: An Engine for Parallel Dynamically Adaptive Simulations of Wave Problems

ExaHyPE ("An Exascale Hyperbolic PDE Engine") is a software engine for s...
06/17/2022

Local Characteristic Decomposition Based Central-Upwind Scheme

We propose novel less diffusive schemes for conservative one- and two-di...
This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.