Modeling Fiber Orientation in Microstructures Using Statistical Information From Paper Samples
Please login to view abstract download link
Using paper or wood-fiber-based materials as substitutes for established but resource-intensive materials, like plastics or concrete, has the potential to bring forth a more sustainable future. Although these natural materials have been used for thousands of years, their mechanical behavior is still not fully understood. In contrast to other composite materials, paper exhibits larger fluctuations due to its natural origin. The anisotropic behavior of paper and these natural fluctuations stem from the geometric and material properties of its microstructure. Therefore, a multiscale modeling strategy is indispensable. Different approaches for modeling the microstructure of paper exist. Generating these microstructures, i.e., fiber networks, in great detail is cumbersome but necessary in order to model and analyze the material behavior of paper realistically. This has to include the statistical information from real paper samples. Thus, we propose an automatic generation approach for synthetic fiber networks via statistical analysis of real paper samples. In general, this approach allows us to stochastically modify every geometric and material property of the fibers therein. This work focuses on the statistically accurate representation of fiber orientation. Additional focus lies on the interdisciplinary collaboration for producing, analyzing and comparing real samples with synthetically generated networks. For this collaboration, paper samples were produced and then scanned using micro-CT. Next, the statistical distribution of the fiber orientations of the samples was analyzed. We used this analysis to fit the stochastic model of the synthetic fiber network generation. As a last step, we analyzed the generated synthetic networks for their fiber orientation in order to compare them with the produced paper samples, thereby verifying their accurateness. Having these microstructures and a verified process for generating them allows us to represent the material behavior of paper on multiple scales.
