ANN-based Modeling of Sunflower Threshing Process and Defining the Optimal Operation Point for Separation Efficiency

Harvesting is one of the most important field operations in sunflower production. Seed damage and low separation efficiency are the top concerns of harvesting sunflower. In this study, a threshing cylinder with rubber teeth and a concave for harvesting sunflower were designed and evaluated. The variable parameters were threshing cylinder speed (TCS), threshing space (TS) and moisture content (MC) of sunflower head. Azargol variety was used to evaluate the threshing unit. The tests were performed at three cylinder speed levels (280, 380 and 480 rpm), two threshing spaces (8 and 10 cm) and two moisture content of sunflower head based on the crop condition (20% and 45% wet basis). An ANN model was developed to predict the amount of materials in each part of the concave. Results showed that the sunflower seeds had no damage during the threshing process and the presented model could predict the amount of materials in each part of the concave with a regression coefficient R2=0.95. Based on the ANN model, with a decrease in MC and TS, and an increase in TCS, the separation efficiency was increased. Furthermore, optimal parameters for the threshing unit which were suggested by Design Expert software to maximize the separation efficiency were 18% w.b, 450 rpm and 10.5 cm for MC, TSC, and TS, respectively and in this condition separation efficiency was determined to be 94.92%.