Numerical simulation and optimisation of depodding performance in a fresh green soybean harvester

A self-propelled beater drum harvester was designed in this study. A systematic analysis of the missed harvest and damage rates during the harvest process was conducted through material mechanical performance testing combined with Workbench and engineering discrete element method engineering discrete element method (EDEM) simulation software. Workbench was used to validate the compression performance of green soybean pods, and the results were highly consistent with the experimental data. A dynamic model for the beater depodding performance was established, and an orthogonal regression experiment was conducted to investigate the effects of drum rotating speed (A₁), cutting platform height (B₁), and forward speed (C₁). The results showed that the models for missed harvest and damage rates were highly significant (p < 0.0001). Using EDEM software, different operational conditions and combinations of factors were simulated to conduct an in-depth study of the missed harvest rate (Y₁), damage rate (Y₂), and total loss rate (Y_T) the harvester. The maximum prediction error between the simulation results and experimental validation was 9.78 %. To minimise total loss, the parameters were optimised, yielding the following optimal combination: Y_T of 10.8 %, Y₁ of 6.9 %, and Y₂ of 3.9 %. The corresponding C₁ was 0.67 m s⁻¹, A₁ was 305 r. min⁻¹, and B₁ was 40 mm. The errors in the simulation analysis, model prediction, and experimental validation were all within 9.8 %. This study provides scientific guidance for optimising the design of fresh green soybean harvesters.