Multiscale quasicontinuum approaches for beam lattices
English
Beex, Lars[University of Luxembourg > Faculty of Science, Technology and Communication (FSTC) > Engineering Research Unit >]
Peerlings, Ron[Eindhoven University of Technology > Mechanical Engineering > > Associate Professor]
Geers, Marc[Eindhoven University of Technology > Mechanical Engineering > > Professor]
Kerfriden, Pierre[Cardiff University > School of Engineering > > Senior Lecturer]
Bordas, Stéphane[University of Luxembourg > Faculty of Science, Technology and Communication (FSTC) > Engineering Research Unit >]
Heaney, Claire[Cardiff University > School of Engineering > > Research Associate]
Os, van, Koen[Philips Research > Intelligent Textiles > > Scientist]
Jul-2014
25
No
No
International
5th International Conference on Computational Methods
from 28-7-2014 to 30-7-2014
Cambridge
United Kingdom
[en] Quasicontinuum method ; Multiscale ; beam lattice ; beam ; Network model ; Discrete model
[en] The quasicontinuum (QC) method was originally developed to reduce the computational efforts of large-scale atomistic (conservative) lattice computations. QC approaches have an intrinsically multiscale character, as they combine fully resolved regions in which discrete lattice events can occur, with coarse-grained regions in which the lattice model is interpolated and integrated (summed in QC terminology). In previous works, virtual-power-based QC approaches were developed for dissipative (i.e. non-conservative) lattice computations which can for instance be used for fibrous materials. The virtual-power-based QC approaches have focused on dissipative spring/truss networks, but numerous fibrous materials can more accurately be described by (planar) beam networks. In this presentation, different QC approaches for planar beam lattices are introduced. In contrast to spring/truss lattices, beam networks include not only displacements but also rotations which need to be incorporated in the QC method, resulting in a mixed formulation. Furthermore, the presentation will show that QC approaches for planar beam lattices require higher-order interpolations to obtain accurate results, which also influences the numerical integration (summation in QC terminology). Results using different interpolations and types of integration will be shown for multiscale examples.