Latent Space Subdivision: Stable and Controllable Time Predictions for Fluid Flow

We propose an end‐to‐end trained neural network architecture to robustly predict the complex dynamics of fluid flows with high temporal stability. We focus on single‐phase smoke simulations in 2D and 3D based on the incompressible Navier‐Stokes (NS) equations, which are relevant for a wide range of...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Computer graphics forum Jg. 39; H. 8; S. 15 - 25
Hauptverfasser: Wiewel, S., Kim, B., Azevedo, V. C., Solenthaler, B., Thuerey, N.
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Oxford Blackwell Publishing Ltd 01.12.2020
Schlagworte:
Artificial neural networks
CCS Concepts
Compression tests
Computer architecture
Computing methodologies → Neural networks
Designers
Dynamic stability
Energy management
Fluid dynamics
Fluid flow
Incompressible flow
Neural networks
Photovoltaic cells
Physical simulation
ISSN:0167-7055, 1467-8659
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:We propose an end‐to‐end trained neural network architecture to robustly predict the complex dynamics of fluid flows with high temporal stability. We focus on single‐phase smoke simulations in 2D and 3D based on the incompressible Navier‐Stokes (NS) equations, which are relevant for a wide range of practical problems. To achieve stable predictions for long‐term flow sequences with linear execution times, a convolutional neural network (CNN) is trained for spatial compression in combination with a temporal prediction network that consists of stacked Long Short‐Term Memory (LSTM) layers. Our core contribution is a novel latent space subdivision (LSS) to separate the respective input quantities into individual parts of the encoded latent space domain. As a result, this allows to distinctively alter the encoded quantities without interfering with the remaining latent space values and hence maximizes external control. By selectively overwriting parts of the predicted latent space points, our proposed method is capable to robustly predict long‐term sequences of complex physics problems, like the flow of fluids. In addition, we highlight the benefits of a recurrent training on the latent space creation, which is performed by the spatial compression network. Furthermore, we thoroughly evaluate and discuss several different components of our method.
Bibliographie:realFlow
This work was supported by the ERC Starting Grant
https://ge.in.tum.de/publications/2020-lssubdiv-wiewel/
(StG‐2015637014) and the Swiss National Science Foundation (grant no. 200021_168997). Source code and video
ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0167-7055
1467-8659
DOI:10.1111/cgf.14097