Probabilistic Characteristics Study of Tensile Properties of Bamboo Inter-Node Material Based on Random Field Theory

Driven by the low-carbon transformation in the construction industry, Moso bamboo has emerged as a promising green material. However, a critical gap exists in current structural design theories: they predominantly rely on homogeneous assumptions, failing to capture the inherent spatial variability and coupled strength–stiffness degradation of bamboo. Experimental results reveal a distinct longitudinal gradient, where the top section (5–8 m) exhibits approximately 15% higher average tensile strength and 12% higher average elastic modulus compared to the bottom section (1–3 m). This oversight severely compromises the accuracy of structural reliability evaluations. To address this, this study pioneers a high-precision digital representation method by developing a novel three-dimensional (3D) anisotropic bivariate coupled random field model. Experimental and stochastic finite element simulations demonstrate that the model accurately replicates spatial variations, maintaining relative errors for primary statistical indicators strictly below 1% while robustly capturing the bivariate coupling characteristics. Crucially, by integrating non-parametric probability box (P-box) theory, the macroscopic tensile resistance of full-scale bamboo members is rigorously bounded within a definitive statistical interval. Furthermore, the extracted Interval Skewness Ratio generally remains greater than 1.0 (averaging 1.38), providing robust quantitative proof of an asymmetric structural degradation governed by the brittle weakest-link failure mechanism. By effectively eliminating non-physical sample generation associated with traditional univariate models, this research makes a significant contribution to the field, providing a rigorous digital twin framework and theoretical foundation for the advanced non-probabilistic safety design of modern bamboo structures.