3D microstructural statistical modeling for effective property prediction of heterogeneous materials
Year: 2026
Type: article
Abstract: This work presents a novel Bayesian framework that integrates the Mean Intercept Length (MIL) method and the Mori–Tanaka (M–T) homogenization model to predict the effective elastic properties of heterogeneous materials. The proposed approach enables morphological characterization and uncertainty quantification directly from microstructural data, linking stereological measurements to micromechanical prediction based on Approximate Bayesian Computations (ABC) Rejection Sampling. The methodology is first assessed using synthetic three-dimensional microstructures with different degrees of inclusion alignment, demonstrating that it accurately captures the imposed anisotropy and yields predictions consistent with Finite Element Method (FEM) simulations. Subsequently, the framework is applied to a real microstructure of nodular cast iron (GGG40) reconstructed from an open-access micro-CT dataset, where the predicted stiffness components and Bayesian confidence intervals encompass the experimental elastic moduli. The novelty of this work lies in the integration of morphology-based and micromechanical models within a Bayesian formulation, enabling both accurate property prediction and rigorous uncertainty quantification. The results confirm that the proposed framework consistently captures the morphological anisotropy and effective stiffness of heterogeneous materials, offering a statistically robust and computationally efficient alternative for micromechanical characterization.
Source: Composite Structures
Language: English
Topic: Composite Material Mechanics
Subfield: Mechanics of Materials
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $4,660
Award: G4190914463
OpenAlex ID: W7160237807