Data-Driven Fluid Flow Simulations by Using Convolutional Neural Network

Autor:	Kazuhiko Kakuda, Wataru Okaniwa, Yuto Morimasa, Shinichiro Miura, Tomoyuki Enomoto
Rok vydání:	2020
Předmět:	Physics::Fluid Dynamics Smoothed-particle hydrodynamics Computer science Free surface Hash function Fluid dynamics Construct (python library) Convolutional neural network Algorithm ComputingMethodologies_COMPUTERGRAPHICS Data-driven Parametric statistics
Zdroj:	Computational and Experimental Simulations in Engineering ISBN: 9783030646899
DOI:	10.1007/978-3-030-64690-5_2
Popis:	In this paper, we present the data-driven fluid flow simulations using the deep CNN (Convolutional Neural Network) with the parametric softsign activation functions. To simulate the fluid flow problems, the particle-method approach based on SPH (Smoothed Particle Hydrodynamics) is used herein. The GPU-implementation consists mainly of the search for neighboring particles in the locally uniform grid cell using hash function. We construct significantly the deep CNN architectures with novel activation functions, so-called parametric softsign. Numerical results demonstrate the workability and validity of the present approach through the dam-breaking fluid flow simulations with free surface.
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_________::1cbc5d5a129ed2e2c5fa108a3b2079be https://doi.org/10.1007/978-3-030-64690-5_2 Zobrazit plný text záznamu