Mixed-Dimensional Coupling for Parallel Partitioned Nonlinear Finite-Element Analysis
File(s)ASCE_CCE16 - BAI_GAJ.pdf (2.21 MB)
Accepted version
Author(s)
Izzuddin, BA
Jokhio, GA
Type
Journal Article
Abstract
This paper presents a novel mixed-dimensional coupling method for parallel partitioned analysis, allowing efficient nonlinear structural analysis with selective deployment of differently dimensioned elements. With this method, a structural domain is decomposed into partitions that are processed in parallel using a recently developed dual super-element partitioning approach. Individual partitions consist exclusively of reduced dimensioned (1D) or higher dimensioned (3D continuum) elements, and coupling between 1D and 3D partitions is introduced at the boundary of 3D partitions, enabling a reduction in the communication overhead. Several examples are provided to verify the proposed approach and demonstrate its benefits for nonlinear structural analysis.
Date Issued
2016-10-12
Date Acceptance
2016-07-28
Citation
Journal of Computing in Civil Engineering, 2016, 31 (3)
ISSN
1943-5487
Publisher
American Society of Civil Engineers
Journal / Book Title
Journal of Computing in Civil Engineering
Volume
31
Issue
3
Copyright Statement
© 2016 American Society of Civil Engineers
Subjects
Science & Technology
Technology
Computer Science, Interdisciplinary Applications
Engineering, Civil
Computer Science
Engineering
Partitioned modeling
Parallel processing
Mixed-dimensional coupling
Nonlinear structural analysis
High-performance computing.
LAGRANGE MULTIPLIERS
MASONRY STRUCTURES
0905 Civil Engineering
Building & Construction
Publication Status
Published
Article Number
04016062