Stochastic sensitivity analysis of micromechanical properties of unidirectional layer with natural fibers - hydrogen pipeline application

The primary objective of this study is to predict the sensitivity of the mechanical properties in unidirectional layers of bio-composites. Four bio-sourced fibers were investigated: PALF (Pineapple leaf fiber), Date palm, Alfa, and Hemp fiber. In order to conduct this sensitivity analysis, three models were implemented. The first model utilizes analytical approaches based on four theoretical methods: Halpin-Tsai, Chamis, Hashin-Rosen, and the Rule of Mixtures (ROM). The second model is based on the finite element method (FEM) applied to the unit cell of bio-layer composites. After validating the analytical and numerical results with experimental data, a coupled methodology combining probabilistic stochastic analysis and FEM was employed to determine the sensitivity of all mechanical properties of the natural fiber-based layers. These bio-layers are designed for use in manufacturing the layers and skins of sandwich composite hydrogen transportation pipelines. The probabilistic stochastic approach, utilizing the Monte Carlo method, enables the execution of the sensitivity analysis. The results indicate that uncertainties in the micromechanical characteristics of fibers, the fiber volume fraction, and the fibers' radius are the most significant factors affecting the stiffness of the layer or the multi-layered structure in longitudinal and transversal directions. These findings provide valuable insights for designers to estimate the volume ratio and select the types of components or fibers to manufacture reliable and economically viable Bio-Composites layers with high resistance and quality.