NCTAM papers, National Congress of Theoretical and Applied Mechanics, Japan
The 64th Japan National Congress for Theoretical and Applied
Session ID : GS4-07
Conference information

GS4
A New Analysis Method of Unsteady Flow Fields Based on a Feature Space Induced by Deep Autoencoder
Noriyasu OMATASusumu SHIRAYAMA
Author information
CONFERENCE PROCEEDINGS FREE ACCESS

Details
Abstract

Data of unsteady flow fields obtained by numerical calculations and experiments becomes enormous, and it is difficult to analyze its spatiotemporal structure. In this paper, we derive a low dimensional feature space using deep autoencoder and propose an analysis method which uses comparison on the space as the core.

Content from these authors
© 2017 The Japan Society of Mechanical Engineers
Previous article
feedback
Top