New Paper on ArXiv — Autoregressive Modelling for 3D CFRP Microstructure Generation
Published:
New Paper Published on ArXiv
Our new preprint is now available on ArXiv! This work presents a comprehensive framework for processing, modelling, and generating statistically representative 3D fiber microstructures from experimental X-ray μCT observations.
About the Paper
The study introduces an integrated framework that combines advanced statistical modelling with physical generation strategies for fiber composite microstructures. Key contributions include:
- An analytical slice-segment ellipse-intersection method for extracting per-fiber in-plane and out-of-plane misalignment profiles from X-ray μCT data
- A stochastic autoregressive model capturing slice-wise misalignment distributions through copula-based in-plane dependence and latent autoregressive continuity
- Bayesian optimization for model calibration, achieving deviations generally below 10% from original statistical descriptors
- A physical generation strategy using Delaunay-based neighbourhood construction and ellipse-based contact resolution to ensure non-overlapping, geometrically admissible microstructures
The framework successfully generates approximately 2,400 synthetic fibers while preserving strong statistical fidelity to the original X-ray μCT data, providing a scalable route for generating simulation-ready fiber composite microstructures for virtual testing and analysis.
Paper Details
- Title: Autoregressive Modelling and Synthetic Generation of High-Fidelity, Statistically Equivalent 3D Microstructures for As-Manufactured Misalignments in Fiber-Reinforced Composites
- Authors: Mohamad A. Raja, Clemens Dransfeld, Boyang Chen
- Submitted: June 18, 2026
- Category: Computational Engineering, Finance, and Science (cs.CE)