Construction of the discrete element model for maize ears and verification of threshing simulation

The maize ear model makes it difficult to represent the kernel separation and cob breakage in the maize threshing during discrete element simulation. In this study, a new discrete element model was constructed for the maize ear aggregate using threshing simulation. Hertz-Mindlin (no-slip) and Bonding contact models were combined in the EDEM software. According to the layered structure of the maize cob, a three-layered discrete element model with adhesive bonding was established using a layered modelling and meshing method. The three-point bending mechanical test was carried out, where the bending breaking force and stiffness were taken as evaluation indicators. The simulated bending tests were conducted with the Plackett-Burman, steepest ascent, and Box-Behnken. The bonding parameters between particles were calibrated in each layer of the maize cob discrete element model. A maize ear aggregate discrete element model was established using a horse-tooth-shaped maize kernel prototype and a five-ball bonding kernel-cob connection. The forces of the axial and radial connectivity were simulated and then calibrated for the kernels and cobs. Finally, the maize threshing process was simulated in three threshing mechanisms: trapezoidal tooth, round-headed nail tooth, and rasp bar. The results showed that the bending breaking force and stiffness of the maize cob were simulated as 168.76 N and 13.08 N/mm, respectively, under the optimal bonding parameter combination, with relative errors of -0.12%, and -0.14%, respectively. The relative errors were -12.84% and 13.25%, respectively, for the axial and radial compression of the kernel pedicel strength. The average normal contact forces of the grains on three threshing elements were 12.50, 12.32, and 8.03 N, respectively, in the simulation area of the threshing. The decreasing trend was consistent with the grain breakage rate in the bench tests. Furthermore, the critical contact force was determined to quantify the rate of broken kernels, according to the cumulative frequency curve of the contact force between the kernels and the threshing elements. The simulation rates of unthreshed kernels were 0.15% and 0.35%, respectively, for the trapezoidal tooth and rasp bar, which were lower by 0.07 and 0.25 percentage points than that in the bench test, respectively. The simulation unthreshed rate was 0.37% for the round-headed nail tooth, which was 0.04 percentage points higher than that in the bench test. There was a positively skewed unimodal distribution in the proportion of threshed material along the axis of the drum. Among them, the peak values in the proportion of the simulated discharge grain mass were 1.03, 1.86, and 0.85 percentage points higher than those in the bench tests, respectively. There were similar threshing qualities between threshing simulation and bench tests. The reason was that there were different threshing mechanisms in the various threshing elements. Specifically, the trapezoidal tooth and round-headed nail tooth relied mainly on the impact and strike to the thresh kernels, indicating the lower frequency of the contacts with kernels. The greater force of contact with the kernel was obtained to cause the higher broken kernel rate during effective threshing. By contrast, the lower broken kernels rate was achieved in the rasp bar to thresh kernels. The larger contact area was obtained to rub the ears at a relatively smaller average contact force, leading to the more frequent contact with the kernels. Although there were still differences between the simulations and the actual, the parameters were accurately evaluated in the discrete element model of corn ear aggregate. The maize threshing and separation were simulated to clarify the mechanical characteristics between the kernel and cob. The maize ear model can be expected to evaluate the threshing performance of the threshing and separation using various indicators, such as the kernel force, contact frequency, the rate of unthreshed kernels, and discharged material distribution. This finding can provide a strong reference for the subsequent optimization of threshing and separation.