Fluid Flow Modelling Using Physics-Informed Convolutional Neural Network in Parametrised Domains

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F49777513%3A23520%2F23%3A43970327" target="_blank" >RIV/49777513:23520/23:43970327 - isvavai.cz</a>

Výsledek na webu

<a href="https://www.tandfonline.com/doi/epdf/10.1080/10618562.2023.2260763?needAccess=true" target="_blank" >https://www.tandfonline.com/doi/epdf/10.1080/10618562.2023.2260763?needAccess=true</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1080/10618562.2023.2260763" target="_blank" >10.1080/10618562.2023.2260763</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Fluid Flow Modelling Using Physics-Informed Convolutional Neural Network in Parametrised Domains

Popis výsledku v původním jazyce

We design and implement a physics-informed convolutional neural network (CNN) to solve fluid flow problems on a parametrised domain. The goal is to compare the effectiveness of training based solely on CFD-generated training data with purely physics-informed training and training based on a combination of both. We consider the problem of incompressible fluid flow in a convergent-divergent channel with variable wall shape. A speciality of the designed neural network is the incorporation of the boundary condition directly in the CNN. A physics-informed CNN that uses a non-Cartesian mesh poses a challenge when evaluating partial derivatives. We propose a gradient layer that pproximates the first and second partial derivatives by finite differences using Lagrange interpolation. Our analysis shows that the convergence of purely physics-informed training is slow. However, using a small training dataset in combination with physics-informed training can achieve results comparable to physics-uninformed training with a considerably larger training dataset.

Název v anglickém jazyce

Fluid Flow Modelling Using Physics-Informed Convolutional Neural Network in Parametrised Domains

Popis výsledku anglicky

We design and implement a physics-informed convolutional neural network (CNN) to solve fluid flow problems on a parametrised domain. The goal is to compare the effectiveness of training based solely on CFD-generated training data with purely physics-informed training and training based on a combination of both. We consider the problem of incompressible fluid flow in a convergent-divergent channel with variable wall shape. A speciality of the designed neural network is the incorporation of the boundary condition directly in the CNN. A physics-informed CNN that uses a non-Cartesian mesh poses a challenge when evaluating partial derivatives. We propose a gradient layer that pproximates the first and second partial derivatives by finite differences using Lagrange interpolation. Our analysis shows that the convergence of purely physics-informed training is slow. However, using a small training dataset in combination with physics-informed training can achieve results comparable to physics-uninformed training with a considerably larger training dataset.

Druh

J_imp - Článek v periodiku v databázi Web of Science

CEP obor

—

OECD FORD obor

20302 - Applied mechanics

Projekt

<a href="/cs/project/GA21-31457S" target="_blank" >GA21-31457S: Použití neuronových sítí pro rychlou predikci proudového pole v úlohách interakce tekutiny s tělesem</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2023

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název periodika

INTERNATIONAL JOURNAL OF COMPUTATIONAL FLUID DYNAMICS

ISSN

1061-8562

e-ISSN

1029-0257

Svazek periodika

37

Číslo periodika v rámci svazku

1

Stát vydavatele periodika

GB - Spojené království Velké Británie a Severního Irska

Počet stran výsledku

15

Strana od-do

67-81

Kód UT WoS článku

001075311200001

EID výsledku v databázi Scopus

2-s2.0-85173600041