We present a novel integrated multiscale, multiphysics, and data-driven framework for predictive modeling and process optimization of glass fiber cable composites. Our hybrid model synergistically couples physics-based simulations with machine learning corrections through a regularized monolithic formulation, ensuring consistency with governing equations and experimental data. This coupling significantly reduces predictive uncertainty, achieving up to a 25% improvement in curing kinetics calibration and a 40% decrease in porosity-related defects compared to traditional models, while accurately capturing thermo-chemo-mechanical fields. We validate our numerical simulations against high-fidelity datasets and demonstrate concurrent optimization of stiffness, lightweight performance, and structural durability. Our methodology enables reliable, adaptive modeling and intelligent control of advanced composite manufacturing processes, thereby laying the groundwork for next-generation design and monitoring strategies in aerospace, automotive, and space industries.
Wilba et al. (Wed,) studied this question.