[Seminar] Multiscale modeling and simulation of carbon fiber-reinforced composites by Prof. Tomonaga Okabe (Tohoku University)

This paper presents a comprehensive multiscale modeling approach for predicting the mechanical behavior and failure of carbon fiber-reinforced composites (CFRPs). CFRPs, known for their exceptional strength-to-weight ratio, are widely used in aerospace and other high-performance applications. However, their anisotropic and hierarchical structure complicates experimental evaluation, prompting the need for predictive computational models. The proposed method integrates multiple simulation techniques across four scales—atomic/molecular, microscopic, mesoscopic, and macroscopic. At the atomic level, quantum chemical reaction-path calculations using the GRRM method and molecular dynamics (MD) simulations are employed to model the curing process and evaluate mechanical properties such as stiffness and strength of epoxy resins. These simulations generate realistic crosslinked network structures of the matrix, which are essential for accurate material characterization. At the microscopic scale, finite element analysis (FEA) of periodic unit cells (PUCs) is used to evaluate the behavior of unidirectional (UD) lamina, incorporating matrix properties derived from the molecular scale. For compressive behavior, kink band models are applied, while tensile failure is predicted using a spring element model (SEM) and Monte Carlo simulations, reflecting the probabilistic nature of fiber fractures. These models collectively yield mechanical properties and failure criteria for the UD lamina. At the mesoscopic level, the properties of the lamina are used to simulate entire laminates under structural loads using FEA and extended finite element methods (XFEM) to capture complex failure phenomena such as delamination and open-hole tension/compression (OHT/OHC) behavior. The full four-scale integration enables the bottom-up prediction of laminate failure behavior from fundamental molecular structure without reliance on empirical testing. Each scale of the model has been validated against experimental results, demonstrating strong agreement and confirming the accuracy and reliability of the approach. This multiscale modeling framework provides a powerful tool for virtual testing and design of CFRPs, offering significant potential for accelerating material development and reducing experimental costs.