Micro-meso uncertainty propagation in tensile behaviour of C/C-SiC composites with in-situ experimental validation

Composite mechanical behaviour is inherently stochastic, arising from complex damage mechanisms and uncontrolled manufacturing parameters. To tackle this challenge, a novel micro-meso uncertainty propagation framework is proposed to quantify the coupled effects of random defects and probabilistic constituent properties on tensile behaviours of C/C-SiC composites. The microscopic and mesoscopic models with random defects were established using SEM and X-CT characterized parameters. Subsequently, a probabilistic constitutive model was developed to incorporate the uncertainty in constituent properties. To couple this with the random defect distribution, extensive microscopic simulations of fibre bundles were performed. At the mesoscale, spatial random variables following the microscopic property distribution were integrated into geometric model. This integration achieves cross-scale uncertainty propagation. Within the progressive failure analysis, the method predicts both the dispersion of macroscopic mechanical properties and the stochastic evolution of mesoscopic damage. These results were validated through combined standard and in-situ CT tensile tests. The simulation results agree with experiments across three critical aspects: macroscopic mechanical response, surface strain and fracture location, and internal damage evolution. The proposed uncertainty propagation method resolves the coupling of uncertainties across scales and spatial domains in composites. It offers a novel approach for material optimization and structural failure probability assessment.