Latent-space Physics: Towards Learning the Temporal Evolution of Fluid Flow

TL;DR

This paper introduces a novel deep learning approach using LSTM networks to predict the temporal evolution of fluid flow pressure fields within latent spaces, achieving significant speed-ups over traditional methods.

Contribution

The paper presents the first successful prediction of dense 3D+time physics functions in latent spaces using neural networks, enabling faster fluid flow simulations.

Findings

Neural network-based simulation is over 100 times faster than traditional pressure solvers.

The method effectively predicts complex liquid and buoyancy flow simulations.

Latent space predictions maintain high accuracy in fluid dynamics modeling.

Abstract

We propose a method for the data-driven inference of temporal evolutions of physical functions with deep learning. More specifically, we target fluid flows, i.e. Navier-Stokes problems, and we propose a novel LSTM-based approach to predict the changes of pressure fields over time. The central challenge in this context is the high dimensionality of Eulerian space-time data sets. We demonstrate for the first time that dense 3D+time functions of physics system can be predicted within the latent spaces of neural networks, and we arrive at a neural-network based simulation algorithm with significant practical speed-ups. We highlight the capabilities of our method with a series of complex liquid simulations, and with a set of single-phase buoyancy simulations. With a set of trained networks, our method is more than two orders of magnitudes faster than a traditional pressure solver. Additionally, we present and discuss a series of detailed evaluations for the different components of our algorithm.