Akademik Veri Yönetim Sistemi
European Journal of Mechanics, A/Solids, cilt.119, 2026 (SCI-Expanded, Scopus)
The thermomechanical response of polycarbonate (PC) under high strain-rate loading is critical for impact-resistant structural design. In this study, two complementary constitutive modeling strategies are developed and assessed for PC: a phenomenological flow model and a data-driven artificial neural network (ANN) surrogate model. While conventional phenomenological formulations often struggle to represent the strongly coupled effects of strain rate and temperature in dynamically loaded polymers, ANN-based approaches for such applications remain largely unexplored. Both constitutive models are implemented as user-defined material subroutines within a finite element framework. Model verification is carried out using high strain-rate tensile experiments, and validation is performed against low-velocity impact tests. The predictive capability of the proposed formulations is further demonstrated through the simulation of the impact response of an automotive polycarbonate headlamp lens. The results show that both modeling approaches successfully reproduce the experimentally observed strain-softening and strain-hardening behaviors of PC over a broad range of strain rates and temperatures. Notably, the ANN-based model achieves comparable accuracy without relying on explicit constitutive assumptions, highlighting its robustness and flexibility. Overall, the study demonstrates the strong potential of data-driven constitutive modeling as a reliable alternative to traditional phenomenological approaches. The proposed framework provides a practical foundation for integrating ANN-based material models into large-scale finite element simulations, particularly for automotive and aerospace applications where accurate impact prediction is essential.