Evaluation and modeling of physical and physiological damage to wheat seeds under successive impact loadings: mathematical and neural networks modeling

Threshing wheat (Triticum aestivum LO at high speeds is the main reason behind abnormal seedlings and vigor reduction of the seeds. This problem is expected to be severe in headstripper combines with successive impact Ioaclings of stripping and threshing units. The airo of this study was to simulate the effects of impact velocities (IV), number of impact loadings (NL), and seed moisture content (MC) on percentage of physical damage (PPD) and percentage of Ioss in germination (PLG) to wheat seeds. Modeling the correlation between dependent ancI independent variables was performed using mathematical and artificial neural networks (ANN). The result showed that all the three independent variables significantly influenced PPD and PLG (P = 0.01). Increasing the IV from 5 to 30 m [s.sup.-1] caused an increase in PPD and PLG from 0.17 to 35.8% and from 0.37 to 19.9%, respectively. It was found that the seeds with higher MC could better withstand physical and physiological damage than those with Iower MC. With an increase in NL from 1 to 3 times, the mean values of PPD and PLG were increased by 2.9 and 2.6 times, respectively. An ANN model with two hidden layers, trained with a back-propagation algorithm, successfully learned the relationship between the input and output variables. In comparison with regression models, ANN performecI better when predicting PPD and PLG to wheat seeds.