DEM-Based Parameter Calibration of Soils with Varying Moisture Contents in Southern Xinjiang Peanut Cultivation Zones

To address the insufficient adaptability of imported peanut harvesting equipment’s soil-engaging components to the specific soil conditions in Xinjiang, this study conducted Discrete Element Method (DEM)-based calibration of soil mechanical parameters using field soil samples with 1–20% moisture content from typical peanut cultivation areas in southern Xinjiang. Through the EDEM simulation platform, a comprehensive approach integrating the Hertz–Mindlin with the JKR adhesion model and Hertz–Mindlin with the Bonding model was employed to systematically calibrate nine key parameters: coefficient of restitution, static friction coefficient, rolling friction coefficient, JKR surface energy, normal/tangential stiffness per unit area, critical normal/tangential force, and soil bonding disk radius. Adopting static angle of repose (SAOR) and unconfined compressive force (UCF) as dual-response indicators, a hybrid experimental design strategy combining Central Composite Design (CCD), Plackett–Burman (PB) screening, and Box–Behnken Design (BBD) optimization was implemented. Regression models for SAOR and UCS were established, yielding six sets of soil parameters optimized for different moisture conditions through parameter optimization. Field validation demonstrated the following: ≤3.27% error in SAOR, ≤1.46% error in UCF, and ≤5.05% error in drawbar resistance validation for field digging shanks. Experimental results confirm that the model demonstrates strong prediction accuracy for soils in typical peanut harvesting regions of southern Xinjiang, thereby providing key parameter references for the future self-developed, highly adaptive soil-engaging components with drag reduction optimization in peanut harvesters for the Xinjiang region.