Using Discrete Element Modeling to Evaluate Material Distribution Using Spinner-disc Spreaders

Abstract

Due to rising costs of inorganic fertilizers, more crop producers are turning to organic fertilizers as a soil amendment and fertilizer source. The inherent physical variability of litter makes it difficult to achieve uniform distribution across the swath of application. A new model approach, known as discrete element modeling (DEM), has been developed and used to simulate bulk material conveyance. Therefore, research was conducted to evaluate the ability of a DEM to simulate the conveyance and distribution of broiler litter from a commercially available spinner-disc spreader. The objectives were to: 1) Utilize discrete element modeling software to establish a working 3-D simulation for a typical litter spreader and establish assumed and measured properties of uniform, plastic BB’s and broiler litter. 2) Calibrate the simulation model through comparisons of simulated particle trajectory to experimental field tests. 3) Validate the simulation model for a litter spreader by comparing the DEM results to empirical testing of conveyance and distribution. Results for classifying broiler litter properties indicated that the bulk density increased as moisture content increased. The median particle size distribution (d50) ranged from 0.8-mm at 18% MC to 7.2-mm at 24% MC. Model setup and calibration indicated that the coefficient of static friction, energy density coefficient (cohesion), and the coefficient of restitution impacted the simulated spread pattern by controlling the transverse landing location, longitudinal landing location, and concentration of the spread pattern, respectively. While calibrating the DEM model, limitations were discovered. The minimum particle size in which the model could successfully simulate was 6-mm diameter, mostly attributed to computational power, number of particles, and the size of the domain in which the simulation took place. This discovery limited the model’s ability to simulate litter particles and spread patterns that reflected experimental results (R2= -0.41) since the range of particle sizes for litter was predominately less than 6-mm. The DEM model performed well simulating patterns for plastic BBs (R2=0.85). Model validation for plastic BBs demonstrated that the DEM model lost its predictive ability to estimate the spread pattern when changing spinner-disc speed or drop location and the simulated spread patterns did not reflect those achieved from field experiments. When the disc speed was increased from 400 to 600-rpm, the coefficient of determination decreased to 0.30 from 0.85. Therefore, the model would require calibration for a new spreader hardware setup. It is hopeful that DEM modeling can be used in to support spinner-disc spreader design in the future.