Calibration of the contact parameters for soybean bonded particle model based on discrete element method

To effectively retrieve simulation contact parameters of soybean bonded particles for an efficient gas-solid two-phase flow coupling simulation of the pneumatic seed-metering device, the angle of repose and the angle of stacking based on the physical seed-piling test were evaluated. The Plackett-Burman and steepest ascent tests were ratified to simplify the simulation analysis of soybean bonded particles, identifying critical influencing factors and centroids. The Box-Behnken response surface test was then implemented to pinpoint optimal values for saliency factors, confirming the universality of the calibrated contact parameters for soybean bonded particles across varying fraction particle sizes. The results indicated that the effect of the static friction coefficient of soybean-plexiglass on the angle of repose was exceedingly significant, and that of both static and rolling friction coefficients of soybean-soybean was generally prominent. While it was abundantly clear that both the static and rolling friction coefficients of soybean-soybean supremely affected the angle of stacking. The Box-Behnken response surface test yielded ideal results: 0.0678 for the static friction coefficient of soybean-soybean, 0.2453 for the static friction coefficient of soybean-plexiglass, and 0.0079 for the rolling friction coefficients of soybean-soybean, culminating in the angle of repose of 28.32 degrees and the angle of stacking of 28.76 degrees. The maximal error between simulated and measured values of the angle of repose and the angle of stacking of soybean bonded particles constructed with various fraction particle sizes was estimated to be 1.59 %, implying superior generality of the calibrated contact parameters. The insights of this investigation can be effectively applied to the coupling simulation analysis of the pneumatic soybean seed-metering device's operations and serve as a reference for other researchers developing particle models for discrete element simulations using the bonded particle method.