DEM-driven flexible breakable weed-soil multiphase coupling model: Optimizing maize weeding performance for shield-cutting end-effector

Maize inter-row/intra-plant weeding during the 3-5 leaf stage is plagued by excessive operational resistance and insufficient weed clearance rate, which restricts the adoption of mechanical weeding robots. Existing discrete element method (DEM) models predominantly treat weeds as rigid, non-breakable bodies, failing to capture the flexible deformation and fracture behaviors of weeds during mechanical cutting—this critical limitation hinders the precise optimization of weeding blade parameters. To address this gap, a DEM-driven flexible breakable weed-soil coupling model was developed for the Shield-Cutting End-effector for Maize Inter-row/Intra-plant Weeding (SEMIW). Morphological parameters of Calystegia hederacea (the dominant weed in northern Chinese maize fields) were systematically measured, and a heterogeneous interaction model was established by integrating the Bonding V2 contact model. A quadratic rotational orthogonal combination experiment was conducted to analyze the synergistic effects of cutting angle (CA), ground clearance (GC), and rotational velocity (RV) on weeding performance (RWB: resistance on the weeding blade; WCR: weed clearance rate). The results indicated that RWB was significantly influenced by GC and RV (P<0.001), while WCR exhibited an exponential increase with decreasing GC (R²=0.9998). Under the optimized parameters (CA=20°, GC=2.7 mm, RV=60 rpm), the WCR reached 98.66% (a 9.6% improvement) and the RWB was reduced to 1.832 N (a 63.4% reduction). Field validation confirmed a WCR of 91.13%, which fell within the 95% confidence interval of the simulation model, meeting the agronomic requirements (WCR>90%, RWB<5 N). This study provides a reliable parameter optimization framework for the development of intelligent maize weeding equipment.