Article (Scientific journals)
Prediction of the biomass particles through the physics informed neural network.
DARLIK, Fateme; ADHAV, Prasad; PETERS, Bernhard
2022In ECCOMAS Congress 2022 - 8th European Congress on Computational Methods in Applied Sciences and Engineering
Peer reviewed
 

Files


Full Text
Fdarlik.pdf
Publisher postprint (78.43 kB)
Download
Full Text Parts
PREDICTION OF THE BIOMASS PARTICLES THROUGH%0ATHE PHYSICS-INFORMED NEURAL NETWORK%0A.pdf
Author postprint (1.91 MB)
Download

All documents in ORBilu are protected by a user license.

Send to



Details



Keywords :
Data-driven scientific computing; Physics informed neural network; Discrete element method
Abstract :
[en] Woody biomass energy is a kind of renewable energy that contributes to the reduction of greenhouse gas emissions, the creation of healthier forests, and the reduction of wildfire danger. Simulations of biomass combustion, in general, are time-consuming simulations with a large number of input particles. We use a deep hidden physics-based neural network model to predict the behavior of particles throughout the simulation based on the equations of motion to achieve an efficient simulation and reduce the processing effort. We replace discrete element methods with inverse methods, which have the advantage of simulating velocity fields without knowing the simulation's boundary and initial conditions. Reconstruction of the velocity fields is done using a recurrent neural network in conjunction with a physics-based loss function. The proposed model is suitable for modeling problems that involve moving particles in a fixed bed. The number of neurons and activation functions in the artificial neural network are optimized, and the effect of the sampling method and the number of outputs are studied.
Research center :
ULHPC - University of Luxembourg: High Performance Computing
LuXDEM - University of Luxembourg: Luxembourg XDEM Research Centre
Disciplines :
Mechanical engineering
Author, co-author :
DARLIK, Fateme ;  University of Luxembourg > Faculty of Science, Technology and Medicine (FSTM) > Department of Engineering (DoE)
ADHAV, Prasad  ;  University of Luxembourg > Faculty of Science, Technology and Medicine (FSTM) > Department of Engineering (DoE)
PETERS, Bernhard ;  University of Luxembourg > Faculty of Science, Technology and Medicine (FSTM) > Department of Engineering (DoE)
External co-authors :
no
Language :
English
Title :
Prediction of the biomass particles through the physics informed neural network.
Publication date :
2022
Journal title :
ECCOMAS Congress 2022 - 8th European Congress on Computational Methods in Applied Sciences and Engineering
Peer reviewed :
Peer reviewed
Focus Area :
Computational Sciences
FnR Project :
FNR12252781 - Data-driven Computational Modelling And Applications, 2017 (01/09/2018-28/02/2025) - Andreas Zilian
Available on ORBilu :
since 16 April 2023

Statistics


Number of views
135 (15 by Unilu)
Number of downloads
41 (5 by Unilu)

Scopus citations®
 
0
Scopus citations®
without self-citations
0

Bibliography


Similar publications



Contact ORBilu