Mechano-Physiological Coupling Damage in the Abscission Zone of Table Grapes Under Combined Tensile–Bending Loading and Prediction of Berry Detachment Rate

Due to the fragility of table grapes, berries are prone to detachment during harvesting, transportation, and postharvest handling. This study aimed to clarify the effects of mechanical loading on the mechanical–physiological coupling process of berry detachment and to establish a predictive model for berry detachment percentage. In this study, the postharvest detachment mechanism of ‘Kyoho’ grapes was investigated by applying different levels of combined tensile–bending loads to the berry abscission zone. The results showed that combined tensile–bending stress accelerated the deterioration of visual quality and caused a marked decline in nutritional and flavor-related components. Furthermore, this stress altered respiration rate and endogenous hormone levels. In addition, combined tensile–bending loading increased the activities of cell wall-degrading enzymes, enhanced cell membrane permeability, and disturbed reactive oxygen species metabolism, with antioxidant enzyme activities in the abscission zone reaching their peaks earlier. Correlation analysis identified key physical and physiological indicators closely associated with berry detachment percentage. Physical–physiological indicators, mechanical loading level, and storage time were used as input variables, whereas berry detachment percentage was used as the output variable. Three predictive models, including multiple linear regression (MLR), back-propagation neural network (BPNN), and genetic algorithm-optimized BP neural network (GA-BP), were developed to predict berry detachment percentage. The results showed that the GA-BP model achieved higher prediction accuracy than the MLR and BPNN models, with R2 = 0.9968 and RMSE = 1.5807. This study provides important insights into the mechanical–physiological coupling process underlying postharvest berry detachment in table grapes under mechanical loading. Moreover, the developed model may provide a useful tool for evaluating berry detachment risk during postharvest storage and transportation, thereby helping to improve quality management and extend shelf life.