The different creep-aging forming processes of 2524 aluminum alloy were taken as the research object, and the effects of creep-aging temperature and creep stress on the fatigue-crack propagation properties of the alloy were studied. The research results showed the following under the same sintering time of 9 h, at creep-aging temperatures of 100 °C, 130 °C, 160 °C, and 180 °C, respectively, with an increase in creep-aging temperature: the fatigue-crack propagation rate was promoted, the spacing of fatigue striations increased, and the sizes of dimples decreased while the number was enlarged; this proves that the fatigue property of the alloy was weakened. Compared with the specimens with creep deformation radii of 1000 mm and 1500 mm, the creep deformation stress was the smallest when the forming radius was 1800 mm, with a higher threshold value of fatigue-crack growth in the near-threshold region of fatigue-crack propagation (ΔK ≤ 8 MPa·m^1/2). Under the same fatigue cycle, the specimens under the action of larger creep stress endured a longer fatigue stable-propagation time and a faster fracture speed. Comparing the effect of creep-aging temperature and creep stress, the creep-aging temperature plays a dominant role in the fatigue-crack propagation of creep-aged 2524 aluminum alloy.

Wearable inertial sensors offer a portable means of assessing gait abnormalities associated with neurological and orthopedic disorders. However, intersubject variability, overlapping gait patterns, and class imbalance complicate multiclass recognition. This study presents the Temporal Structural Dual Graph Network (TSDGNet) for pathological gait classification using inertial measurement unit (IMU) signals. Four IMUs positioned on the head, lower back, and both feet provide acceleration and angular velocity signals during walking and turning. The framework combines multiscale temporal feature extraction with graph-based modeling of sequence-level relationships among sensor representations. Two fixed prior graphs encode anatomical and symbiotic relationships, while trainable graph attention and context-guided gated residual refinement adapt sensor-node representations. Class-balanced optimization is incorporated to reduce training bias toward majority classes. Subject-independent evaluation was performed on a public clinical dataset comprising healthy controls and seven neurological or orthopedic disease groups. Comparisons with representative baseline and advanced models showed competitive classification performance. Ablation analyses examined the contributions of temporal modeling, structural relationship learning, and class-balanced optimization. These findings support the potential of combining wearable inertial sensing with temporal and graph-based learning for objective classification of pathological gait.