A partitioned formulation of the stabilized explicit finite element contactimpact algorithm with reciprocal mass matrices

0549233 - ÚT 2022 RIV GR eng A - Abstrakt
Kopačka, Ján - González, J.A. - Kolman, Radek - Park, K.-Ch.
A partitioned formulation of the stabilized explicit finite element contactimpact algorithm with reciprocal mass matrices.
COMPDYN 2021. Atény: National Technical University of Athens, 2021. ISBN 978-618-85072-5-8.
[International Conference on Computational Methods in Structural Dynamics and Earthquake Engineering /8./. 28.06.2021-30.06.2021, Atény]
Grant CEP: GA ČR(CZ) GA19-14237S
Institucionální podpora: RVO:61388998
Klíčová slova: contact-impact * bipenalty method * explicit time integration * reciprocal mass matrix * localized lagrange multipliers
Obor OECD: Applied mechanics
https://2021.compdyn.org/

This contribution deals with an accurate yet powerful methodology of contact-impact treatment for explicit finite element analysis. It stands on four pillars. First, the classic penalty treatment of contact constraints is well known to negatively affect the size of the critical time step. By using the bipenalty method [1], which in addition to potential energy also penalizes kinetic energy, it is possible to maintain a critical time step of the non-penalized system. In addition, oscillations of contact forces are reduced by using a predictor-corrector form of an explicit time integration scheme. Second, for uncoupling the interface terms from the free body equations, a method of localized Lagrange multipliers [2] is employed to formulate the contact problem in a partitioned manner. Third, the sparse inverse mass matrices in the free body equations of motion are directly constructed [3]. And finally, an element-by-element mass matrix scaling technique that allows the extension
of the time integration step [4] is adopted. Thanks to these four techniques, it was possible to increase the performance of the explicit contact algorithm while improving the stability of the numerical solution, as shown by solving several numerical examples.
Trvalý link: http://hdl.handle.net/11104/0327462