CALIBRATION OF DISCRETE ELEMENT SIMULATION PARAMETERS FOR BUCKWHEAT SEEDS

To address the challenge of parameter accuracy in discrete element method (DEM) simulations of buckwheat seeds within a seed metering device, this study characterized the physical properties of buckwheat seeds and subsequently calibrated the simulation model parameters. Initially, physical experiments were conducted to determine the triaxial dimensions, angle of repose, and static friction coefficients of buckwheat seeds against stainless steel surfaces. A three-dimensional model of the buckwheat seeds was then generated using Inventor and it was imported into EDEM to simulate the seed stacking process, with the angle of repose quantified via image processing techniques. Employing a Plackett-Burman design, initial parameters were screened, identifying the static friction coefficient between buckwheat seeds, the rolling friction coefficient between buckwheat seeds, and the static friction coefficient between buckwheat seeds and stainless steel as significant factors influencing the angle of repose. The optimal range for these significant parameters was determined through a steepest ascent experiment, and a second-order regression model was developed using Box-Behnken experimental results to optimize the angle of repose and the identified parameters. The optimized parameter set comprised a static friction coefficient between buckwheat seeds and stainless steel of 0.448, a rolling friction coefficient between buckwheat seeds of 0.038, and a static friction coefficient between buckwheat seeds of 0.372. Validation through both simulations and physical experiments revealed a relative error of 1.08%, confirming the reliability of the calibrated parameters for simulating buckwheat seed sowing machinery.