Heating rate effects on creep aging behavior of 7B50 aluminum alloy and deep learning enhanced constitutive modeling

This study investigates creep deformation and age strengthening of 7B50 aluminum alloy under different heating rates. The experimental results demonstrate that an elevated heating rate facilitates the accumulation of dislocation density, thereby accelerating the creep process and resulting in greater creep strain, concurrently postponing the initiation of steady-state creep deformation. Conversely, decreasing the heating rate facilitates the nucleation and growth of larger η′ phase particles with increased volume fraction, thereby providing stronger pinning effects against dislocation movement, consequently enhancing material strength. A heating rate-strain (HRS) constitutive model was developed, and machine learning was employed to predict the deviations between HRS model predictions and experimental data. Notably, a genetic algorithm-optimized long short-term memory (GA-LSTM) approach demonstrated superior accuracy in predicting these deviations compared to artificial neural network (ANN) and recurrent neural network (RNN) methods. By incorporating corrected GA-LSTM-HRS predictions, we established a coupled macro-micro constitutive framework integrating relative dislocation density, precipitate size, and volume fraction to describe material strength evolution. The model demonstrates excellent agreement with experimental observations. These findings, combined with the developed modeling framework, provide both theoretical insights and computational tools for precise creep age forming of components under diverse heating conditions.