Title: Open boundary modeling in molecular dynamics with machine learning
Authors: Neumann, Philipp 
Wittmer, Niklas
Language: en
Subject (DDC): DDC::000 Informatik, Informationswissenschaft, allgemeine Werke
DDC::500 Naturwissenschaften und Mathematik
Issue Date: 2020
Document Type: Conference Object
Journal / Series / Working Paper (HSU): Lecture notes in computer science 
Volume: 12142
Page Start: 334
Page End: 347
Conference: 20th International Conference on Computational Science ICCS 2020 
Abstract: 
Molecular-continuum flow simulations combine molecular dynamics (MD) and computational fluid dynamics for multiscale considerations. A specific challenge in these simulations arises due to the “open MD boundaries” at the molecular-continuum interface: particles close to these boundaries do not feel any forces from outside which results in unphysical behavior and incorrect thermodynamic pressures. In this contribution, we apply neural networks to generate approximate boundary forces that reduce these artefacts. We train our neural network with force-distance pair values from periodic MD simulations and use this network to later predict boundary force contributions in non-periodic MD systems. We study different training strategies in terms of MD sampling and training for various thermodynamic state points and report on accuracy of the arising MD system. We further discuss computational efficiency of our approach in comparison to existing boundary force models.
Organization Units (connected with the publication): High Performance Computing 
URL: https://api.elsevier.com/content/abstract/scopus_id/85087284627
ISBN: 9783030504328
ISSN: 03029743
DOI: 10.1007/978-3-030-50433-5_26
Appears in Collections:Publications of the HSU Researchers

Show full item record

CORE Recommender

SCOPUSTM   
Citations

2
checked on May 24, 2022

Google ScholarTM

Check

Altmetric

Altmetric


Items in openHSU are protected by copyright, with all rights reserved, unless otherwise indicated.