Development of contact algorithm for three-dimensional numerical manifold method

• Main Authors: He, L., Zhao, Z. Y., An, Xinmei
• Other Authors: School of Civil and Environmental Engineering
• Format: Article
• Language: English
• Published: 2013
• Subjects: DRNTU::Engineering::Civil engineering::Geotechnical
• Online Access: https://hdl.handle.net/10356/100326; http://hdl.handle.net/10220/17880
• Institution: Nanyang Technological University

This paper customizes a contact detection and enforcing scheme to fit the three-dimensional (3-D) numerical manifold method (NMM). A hierarchical contact system is established for efficient contact detection. The mathematical mesh, a unique component in the NMM, is utilized for global searching of possible contact blocks and elements, followed by the local searching to identify primitive hierarchies. All the potential contact pairs are then transformed into one of the two essential entrance modes: point-to-plane and crossing-lines modes, among which real contact pairs are detected through a unified formula. The penalty method is selected to enforce the contact constraints, and a general contact solution procedure in the 3-D NMM is established. Because of the implicit framework, an open-close iteration is performed within each time step to determine the correct number of contact pairs among multi-bodies and to achieve complete convergence of imposed contact force at corresponding position. The proposed contact algorithm extensively utilizes most of the original components of the NMM, namely, the mathematical mesh/cells and the manifold elements, as well as the external components associated with contacts, such as the contact body, the contact facet and the contact vertex. In particular, the utilization of two mutually approaching mathematical cells is efficient in detecting contacting territory, which makes this method particularly effective for both convex and non-convex bodies. The validity and accuracy of the proposed contact algorithm are verified and demonstrated through three benchmark problems.