The Optimized Design of Soil-Touching Parts of a Greenhouse Humanoid Weeding Shovel Based on Strain Sensing and DEM-ADAMS Coupling Simulation.

To overcome the shortcomings of plowing and rotary tillage, a human-like weeding shoveling machine was designed. The machine's various moving rods were analyzed using Matlab R2019b(9.7.0.1190202) software to determine the appropriate entry and cutting conditions, as well as non-cutting conditions. It was concluded that a θ2 of 90° was optimal for cutting the soil and that the shoveling depth was suitable for greenhouse weeding. The Adams and DEM coupled discrete element simulation system was developed for this machine and was used to analyze the rotating shaft torque and shovel bending moment. A strain measurement system based on strain gauges was designed to measure the rotating shaft torque and shovel bar bending moment. A bending moment and torque measurement system was designed to perform field measurement tests for comparison with simulation results. The simulation system's rotating shaft had an average torque error of 6.26%, while the shovel rod's bending moment had an average error of 5.43%. The simulation accuracy was within the acceptable error range. Table U8 (81 × 44) of the Uniform Design of the Mixing Factor Level for the Homogeneous Virtual Simulation Test includes eight levels of forward machine speed ranging from 0.1 to 0.45 m/s and four levels of output shaft speed ranging from 90 to 165 r/min. Crank lengths were set at four levels ranging from 155 to 185 mm, while shovel lengths were set at four levels ranging from 185 to 230 mm. Four types of shovel shapes were proposed, including pointed curved shovels, pointed straight shovels, straight-edged curved shovels, and straight-edged straight shovels. A mathematical model was created via a regression analysis of the results of coupled simulation tests to establish the relationship between shaft torque and shovel rod bending moment, tool advance speed, shaft speed, crank length, tool length, and tool shape. The model was used to determine the optimum working parameters.