Solid and 3D beam finite element models for the nonlinear elastic analysis of helical strands within a computational homogenization framework. (December 2021)

• Record Type: Journal Article
• Authors: Ménard, Fabien; Cartraud, Patrice
• Is Part Of: Computers & structures. Volume 257(2021)
• Publication Date: 2021-12
• Subjects: Homogenization -- Helical symmetry -- Strand -- Cable -- Bending -- Contact -- Finite element method; Structural engineering -- Data processing -- Periodicals; Electronic data processing -- Structures, Theory of -- Periodicals; 624.171
• DOI: 10.1016/j.compstruc.2021.106675
• Languages: English
• ISSNs: 0045-7949

Abstract

Highlights: Helical strands analysis is addressed rigorously with homogenization theory. Strand response is non-linear due to contact interactions between its components. Different models with solid and beam finite elements are used. Numerical results are compared to those coming from analytical models. Abstract: This paper proposes a computational approach for studying the overall behaviour and local stress state of strand-type structures. This method is based on the homogenization theory of periodic beamlike structures, with the local problem posed on the strand axial period being solved using the finite element method. This approach fully utilises the strand's helical symmetry, thus minimising the size of the computational domain. Consequently, accounting for geometric complexity and contact interactions, which are of paramount importance for bending loads, is more straightforward. The numerical model mesh size can also be reduced thanks to the use of beam elements, and one objective of this paper is to assess the accuracy of such a model in comparison with solid element models and analytical results. These comparisons are performed on both single-layer and multi-layer strands. Results demonstrate the capability of the proposed computational approach to accurately capture the nonlinear bending behaviour stemming from the stick-slip transition as well as local stress distributions. As for the beam model, it apparently offers a very good compromise between accuracy and numerical efficiency.