Construction and parameter calibration of the nonlinear elastoplastic discrete element model for no-tillage soil compaction

No-tillage technology has been widely used for the sustainable development in the modern agriculture. But the perennial no-tillage has caused great changes in the soil mechanics. The elastoplastic mechanical properties of no-tillage soil can also lead to the unclear interaction mechanism between agricultural machinery and soil. A Discrete Element Method (DEM) can be expected to serve as an important technique to clarify the interaction. However, two great challenges still remain on the selection of the soil particle contact model and the calibration of the parameters. In this study, a nonlinear elastoplastic contact model was proposed for the no-tillage soil compaction using the Extended DEM (EDEM). The no-tillage loam was collected from the two cropping wheat - maize regions in North China. The parameters were calibrated under the compaction damage caused by external forces in farmland. A sensitivity analysis was designed using a Plackett-Burman technique, in order to select the key parameters, including the coefficient of soil particle recovery, the coefficient of static friction between soil particles, the surface energy of soil particles, as well as the plastic deformation ratio of soil particles to particles. Taking the amount of subsidence as the evaluation index, a quadratic polynomial regression model was established using the Box-Behnken test. The significance parameters were then optimized, where the actual subsidence of 6.2 mm was taken as the target value obtained from the soil plate test under the condition of 12.7 kN load. An optimal combination of parameters was obtained as follows: the recovery coefficient of soil particles was 0.47, the coefficient of static friction between soil particles and particles was 0.62, the surface energy of soil particles was 6.12 J/m2, and the plastic deformation ratio of soil particles was 0.41. In this case, the soil DEM model was obtained for the long-term no-tillage. Five repeated tests were set for the pressure plate settlement test, where the model settlement amounts were 7.19, 6.50, 5.55, 6.62, and 6.07, respectively. The relative error was 2.84%, compared with the actual measurement. The results showed that there was a smaller subsidence error of DEMs simulation and laboratory test. A field test was performed on the soil particle stress transfer, further to calibrate the obtained parameters. It was found that there was little difference between the simulated and the measured parameters, where the relative error was within 9.21%. The determination coefficient R2 between the two fitting curves was 0.91, indicating a higher fitting similarity. Consequently, higher accuracy and reliability were achieved for the parameter calibration of no-tillage soil. This finding can provide promising technical support for the rapid construction of the soil DEM model under no-tillage mode for the optimal agricultural equipment.