A simple and efficient lagrange multiplier based mixed finite element for gradient damage. (June 2023)

• Record Type: Journal Article
• Authors: Riesselmann, J.; Balzani, D.
• Is Part Of: Computers & structures. Volume 281(2023)
• Publication Date: 2023-06
• Subjects: Gradient damage -- Mixed finite elements -- Higher order gradients; Structural engineering -- Data processing -- Periodicals; Electronic data processing -- Structures, Theory of -- Periodicals; 624.171
• DOI: 10.1016/j.compstruc.2023.107030
• Languages: English
• ISSNs: 0045-7949

Abstract

Highlights: The novel FE formulation describes the evolution criterion with a simple Lagrange multiplier method. Formulation does not require stabilization or penalty parameters. No cross-element information is required, enabling a straightforward implementation. Computing time of damage calculations is in the order of purely elastic calculations. Numerical examples show robustness which outperforms competing formulations. Abstract: A novel finite element formulation for gradient-regularized damage models is presented which allows for the robust, efficient, and mesh-independent simulation of damage phenomena in engineering and biological materials. The paper presents a Lagrange multiplier based mixed finite element formulation for finite strains. Thereby, no numerical stabilization or penalty parameters are required. On the other hand, no additional degrees of freedom appear for the Lagrange multiplier which is achieved through a suitable FE-interpolation scheme allowing for static condensation. In contrast to competitive approaches from the literature with similar efficiency, the proposed formulation does not require cross-element information and thus, a straightforward implementation using standard element routine interfaces is enabled. Numerical tests show mesh-independent solutions, robustness of the solution procedure for states of severe damage and under cyclic loading conditions. It is demonstrated that the computing time of the gradient damage calculations exceeds the one of purely elastic computations only by an insignificant amount. Furthermore, an improved convergence behavior compared to alternative approaches is shown.