Multiscale modeling of the anisotropic elastic-plastic behavior of paper

点击翻译按钮获取中文摘要

Paper is gaining more and more importance in numerous technically relevant applications, in particular in the packaging industry, because it is a renewable raw material and easily recyclable. Even though paper has been in use for around 2000 years, the material behavior and its modeling are still not sufficiently understood. The challenge in the material modeling of paper is its multiscale nature. The intrinsic microstructure, which consists of a fiber network, has a significant effect on the macroscale material response. To tackle this challenge, a numerical modeling strategy is to be developed in the current project. This strategy aims at an efficient description of the material on the structural scale while simultaneously accounting for the effects of the network scale. Thereby, the influence of single parameters and effects on the macroscopic behavior can be isolated and investigated, which is hardly possible with an experimental setup.The work plan is divided into several work packages. First, representative volume elements (RVE) are generated for the fiber networks. With these, the variation of apparent properties on the macroscale due to statistical distributions of geometrical data such as fiber length and material properties of single fibers can be analyzed. Then, these RVE are implemented into the FE2 framework based on the finite element method. In order to perform the corresponding simulations efficiently with reasonable computational cost, the proper orthogonal decomposition technique is applied to reduce the model dimensions. Finally, the modeling strategy is validated against experimental results from the creasing test.The result of this project is a simulation tool, that is enables the description of the material behavior of paper in an efficient and robust manner while simultaneously accounting for the microscale effects. Moreover, a deeper understanding is generated on how the different effects on the fiber and network level interact, which is impossible with purely experimental investigations.