Calibration and optimization of discrete element method parameter of <i>Pinus koraiensis</i> seeds

The development of food processing equipment for Pinus koraiensis seeds has been limited by insufficient data on physical and contact properties. This study obtained dimensional and mechanical characteristics of the seeds through experiments and theoretical analysis. A 3D morphological model was reconstructed via imaging, and a bonded multi-sphere discrete element model (DEM) was created using a virtual–physical transformation filling method. Contact parameters between seeds and steel were calibrated through combined tests, yielding a restitution coefficient of 0.327, static friction coefficient of 0.260, and rolling friction coefficient of 0.140. Key factors affecting repose angle were identified using Plackett–Burman design and optimized via the steepest ascent method. A Box–Behnken design produced optimal inter-seed contact parameters: restitution 0.334, static friction 0.340, and rolling friction 0.021. The minimal relative error between simulated and experimental repose angles was 1.52%. DEM accuracy was validated further, showing that over 131 sub-spheres kept simulation error below 5% while ensuring computational efficiency. This study offers a reliable DEM and calibrated parameters for simulating cleaning, conveying, screening, feeding, and shell-breaking of P. koraiensis seeds, and serves as a reference for modeling other irregularly shaped agricultural materials.