Title: | Flow-Field Prediction in Periodic Domains Using a Convolution Neural Network with Hypernetwork Parametrization |
Authors: | Bublík, Ondřej Heidler, Václav Pecka, Aleš Vimmr, Jan |
Citation: | BUBLÍK, O. HEIDLER, V. PECKA, A. VIMMR, J. Flow-Field Prediction in Periodic Domains Using a Convolution Neural Network with Hypernetwork Parametrization. International Journal of Applied Mechanics, 2023, roč. 15, č. 2, s. 1-20. ISSN: 1758-8251 |
Issue Date: | 2023 |
Publisher: | World Scientific |
Document type: | článek article |
URI: | 2-s2.0-85147495139 http://hdl.handle.net/11025/52987 |
ISSN: | 1758-8251 |
Keywords in different language: | blade cascade;compressible fluid flow;convolutional neural network;hypernetwork;periodic boundary condition |
Abstract in different language: | This paper deals with flow field prediction in a blade cascade using the convolution neural network. The convolutional neural network (CNN) predicts density, pressure and velocity fields based on the given geometry. The blade cascade is modeled as a single interblade channel with periodic boundary conditions. In this paper, an algorithm that enforces periodic boundary conditions onto the CNN is presented. The main target of this study is to parametrize the CNN model depending on the Reynolds number. A new parametrization approach based on a so-called hypernetwork is employed for this purpose. The idea of this approach is that when the Reynolds number is modified, the hypernetwork modifies the weights of the CNN in such a way that it produces flow fields corresponding to that particular Reynolds number. The concept of the hypernetwork-based parametrization is tested on the problem of a compressible fluid flow through a blade cascade with variable blade profiles and Reynolds numbers. |
Rights: | Plný text je přístupný v rámci univerzity přihlášeným uživatelům © World Scientific Publishing Europe Ltd. |
Appears in Collections: | Články / Articles (NTIS) Články / Articles (KME) OBD |
Files in This Item:
File | Size | Format | |
---|---|---|---|
NN1.pdf | 4,56 MB | Adobe PDF | View/Open Request a copy |
Please use this identifier to cite or link to this item:
http://hdl.handle.net/11025/52987
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.