Combined Effects of Pre-deformation and Natural Aging on Creep Aging Behavior of 2A14 Aluminum Alloy: A Hybrid Constitutive Modeling Approach

This study systematically investigates the combined effects of pre-deformation and natural aging (NA) on the creep aging (CA) behavior of 2A14 aluminum alloy, aiming to optimize the forming precision and mechanical properties of aerospace load-bearing components. Experimental and computational analyses demonstrate that pre-deformation significantly suppresses NA-induced hardening, with higher pre-strain levels (0%, 4%, 8%) accelerating the attainment of peak NA strength. CA tests conducted at varying pre-strain levels revealed that pre-deformation significantly enhances both creep strain and creep rate while mitigating the effects of NA. Microstructural characterization via XRD and TEM demonstrated that pre-deformation promotes phases and increases dislocation density, thereby contributing to improved yield strength. To address the limitations of conventional constitutive models in predicting multi-parameter creep behavior, we developed a hybrid pre-strain and NA coupled (PNC) model integrated with a particle swarm optimization long short-term memory (PSO-LSTM) network. The PSO-LSTM-PNC model achieves superior prediction accuracy and computational efficiency, effectively correcting significant deviations in the initial creep stage and under specific process conditions. This approach bridges the gap between physical mechanisms and data-driven modeling, providing a robust framework for optimizing CA processes under complex thermomechanical conditions.