Scale-bridging characterization and modeling of statistically inhomogeneous Moso bamboo

This study presents a multiscale framework for investigating the radial distribution of mesostructural characteristics and the resulting macroscopic linear elastic behavior of the statistically inhomogeneous culm wall of moso bamboo. The proposed framework, comprising characterization, reconstruction, and homogenization, enables spatially resolved predictions of macroscopic mechanical properties based on mesostructural input. A set of micromechanical descriptors for statistically inhomogeneous particle structures is defined using indicator and parameter functions. These descriptors are applied to bamboo cross-sections through advanced image processing, allowing for the extraction of spatially resolved features and differentiation between vascular bundle types. A reconstruction algorithm generates local surrogate models that preserve the prescribed mesostructural characteristics using optimized tessellation and area fraction control. Resolution and representativity studies ensure the validity of the reconstructed models across radial positions. Finally, classical mean-field and FFT-based homogenization methods are applied to compute the radial distribution of effective linear elastic properties. The presented framework enhances the understanding of structure–property relationship in bamboo and establishes a foundation for future bio-inspired or bamboo-based material design.