Year: 2020
Author: Fei Shuang, Pan Xiao, Yilong Bai, Fujiu Ke
Communications in Computational Physics, Vol. 28 (2020), Iss. 3 : pp. 1019–1037
Abstract
Molecular statics (MS) based on energy minimization serves as a useful simulation technique to study mechanical behaviors and structures at atomic level. The efficiency of MS, however, still remains a challenge due to the complexity of mathematical optimization in large dimensions. In this paper, the Inertia Accelerated Molecular Statics (IAMS) method is proposed to improve computational efficiency in MS simulations. The core idea of IAMS is to let atoms move to meta positions very close to their final equilibrium positions before minimization starts at a specific loading step. It is done by self-learning from historical movements (atomic inertia effect) without knowledge of external loadings. Examples with various configurations and loading conditions indicate that IAMS can effectively improve efficiency without loss of fidelity. In the simulation of three-point bending of nanopillar, IAMS shows efficiency improvement of up to 23 times in comparison with original MS. Particularly, the size-independent efficiency improvement makes IAMS more attractive for large-scale simulations. As a simple yet efficient method, IAMS also sheds light on improving the efficiency of other energy minimization-based methods.
You do not have full access to this article.
Already a Subscriber? Sign in as an individual or via your institution
Journal Article Details
Publisher Name: Global Science Press
Language: English
DOI: https://doi.org/10.4208/cicp.OA-2019-0157
Communications in Computational Physics, Vol. 28 (2020), Iss. 3 : pp. 1019–1037
Published online: 2020-01
AMS Subject Headings: Global Science Press
Copyright: COPYRIGHT: © Global Science Press
Pages: 19
Keywords: Molecular statics energy minimization local optimization efficiency improvement.
Author Details
-
Intelligent step-length adjustment for adaptive path-following in nonlinear structural mechanics based on group method of data handling neural network
Maghami, Ali
Hosseini, Seyed Mahmoud
Mechanics of Advanced Materials and Structures, Vol. 29 (2022), Iss. 20 P.2895
https://doi.org/10.1080/15376494.2021.1880677 [Citations: 9]