Jaan-Willem Simon's KTH Solid Mechanics Seminar "Multiscale Analysis of Damage in Thermoplastic Fiber-Reinforced Composites"

Jaan_Simon 2026_10_08.pdf (pdf 302 kB) Abstract. The increasing demand for lightweight structures in transportation and related sectors has intensified the use of fiber-reinforced thermoplastic composites, driven by their short processing cycles, high toughness, and recyclability. However, their macroscopic performance is governed by complex damage-induced softening mechanisms originating in the thermoplastic matrix, which exhibits rate-dependent visco-elastoplastic behavior and large inelastic deformations. Reliable failure prediction remains challenging, particularly for woven thermoplastic composite laminates involving interacting mechanisms across multiple scales, as conventional multiscale approaches lose objectivity once softening occurs. In this contribution, we employ a gradient-enhanced visco-hyperelastic-plastic matrix model within resolved composite microstructures. The gradient extension eliminates the pathological mesh dependence associated with softening, enabling mesh-objective microscale predictions under highly localized deformation. In computational homogenization, however, these localized responses lead to RVE-size-dependent effective behavior, a scale dependence that gradient regularization alone cannot resolve. We address this limitation using the failure zone homogenization scheme, which accounts for the finite extent of the failure zone and mitigates RVE-size sensitivity. By combining mesh-objective microscale modeling with size-objective homogenization, the proposed framework establishes a physically consistent link between matrix-level damage and effective structural behavior, supporting predictive and scalable modeling of thermoplastic composite components under demanding loading conditions.