Stochastic Modeling of Composite Interfaces: Sensitivity to Spatial Correlation and Bayesian Identification from Standard Fracture Tests

2026-06-11

AI Summary

This study addresses the challenge of reliability assessment in composite structures arising from interfacial uncertainties, particularly concerning stiffener debonding. A high-fidelity modeling framework is proposed that integrates spatially correlated random fields with stochastic finite element methods. Interfacial variability in laminated composites is characterized using a covariance kernel, and—leveraging approximate Bayesian computation (ABC) for the first time—key statistical parameters, especially the correlation length, are inferred from standard Mode I/II fracture test data. The correlation length is identified as the dominant factor governing interfacial uncertainty. This approach enables accurate representation of scattering effects at the macroscopic scale, offering a robust pathway for uncertainty quantification and virtual testing in damage-tolerant composite aircraft structures.

Abstract

To enable a numerical handling of uncertainties in composite structures, this work presents a stochastic finite-element framework aimed at improving the reliability assessment of aerospace composites, with particular attention to stiffener debonding. By representing interface variability between laminate parts with spatially correlated random fields, the method aims at considering scattering effect at a higher scale of simulation and testing. A parametric study carried out on standardized Mode I and Mode II fracture tests reveals that the correlation length is the primary driver of observed variability, while the regularity of the covariance kernel has only a marginal impact. To guarantee industrial relevance, we demonstrate that this key parameter can be extracted from experimental fracture data using an Approximate Bayesian Computation approach. The proposed methodology therefore offers a robust route to high-fidelity virtual testing and to the predictive management of uncertainties in the design of damage-tolerant composite airframes.

Problem

• stochastic modeling
• composite interfaces
• spatial correlation
• uncertainty quantification
• fracture tests

Innovation

• stochastic finite-element
• spatial correlation
• Bayesian identification
• composite interfaces
• fracture tests