Porosity is the key factor affecting a medium’s tortuosity, effective evaporation area coefficient, and ventilation resistance, and further affects the drying efficiency, energy consumption, and drying uniformity in the drying process. To reveal the dynamic variation characteristics of porosity in paddy flow layer, an air convection drying apparatus was established and a mathematical porosity model was established based on response surface methodology. The reliability of the model was verified through EDEM–Fluent coupled digital simulation and experiments. The research results show that under different paddy flow rates v_d (0.01 m/s, 0.03 m/s, 0.05 m/s), different moisture contents M_c (14% w.b., 23% w.b., 32% w.b.), different wind speeds v_w (0.4 m/s, 0.6 m/s, 0.8 m/s), and different layer thicknesses L (100 mm, 150 mm, 200 mm), the porosity values obtained by the porosity measurement device range from 39.562% to 46.006%. The relative errors between the actual values (ε_r), the simulation values (ε_s), and the predicted values (ε_p) are all within ±1%. Moreover, the obtained mathematical porosity model has high reliability (R² = 0.968). The Conclusions provide an analysis method for dynamic change characteristic parameters and basic data for the dynamic change of porosity to reduce drying energy consumption, improve the drying power coefficient, and enhance drying quality.

In this study, we aimed to address the lack of systematic research on key collision dynamics parameters (elastic restitution coefficient) in the full mechanization of rapeseed operations, which hinders the development of precision agriculture. In this present work, the restitution coefficient of rapeseed was systematically investigated, and a predictive model (R² = 0.959) was also established by using Box–Behnken design response surface methodology (BBD-RSM). The results show that the collision restitution coefficient varies in the range of 0.539–0.649, with the key influencing factors ranked as follows: moisture content (M_c) > material layer thickness (L) > drop height (H). The EDEM simulation methodology was adopted to validate the experimental results, and the results show that there is a minimal relative error (−1% < δ < 1%) between the measured and simulated rebound heights, indicating that the established model shows a reliable prediction performance. Moreover, by comprehensively analyzing stress, strain, and energy during the collision process between rapeseed and Q235 steel, it can be concluded that the process can be divided into five stages—free fall, collision compression, collision recovery, rebound oscillation, and rebound stabilization. The maximum stress (1.19 × 10⁻² MPa) and strain (6.43 × 10⁻⁶ mm) were observed at the beginning of the collision recovery stage, which can provide some theoretical and practical basis for optimizing and designing rapeseed machines, thus achieving the goals of precise control, harvest loss reduction, and increased yields.

In response to the subjective issues, inconsistent quality standards, high labor intensity and low sorting efficiency during the drying process of green pepper, an improved YOLOv11 algorithm was proposed for quality detection. A multi-scale edge enhancement module (MEEM) is introduced into the backbone network, replacing the original basic C3K2 module with C3K2-MEEM to enhance the extraction of detailed features in images of dried green Sichuan pepper and prevent missed detections, false detections, and boundary confusion. The LRSA module is integrated into the 10th layer of the backbone network to improve the clarity of the tumor-like texture of the Sichuan pepper and reduce the influence of impurities, automatically allocating attention based on feature similarity to preserve local information. In the neck layer, the DPCF module is added to the FPN+PAN feature fusion stage to achieve multi-scale feature collaboration, meeting the detection requirements of dried green Sichuan pepper. The results show that the accuracy recall rate, mean average precision, and model size of the improved MLD-YOLOv11 algorithm are 92.1%, 96.6%, 95.6%, and 11.06 MB, respectively. Compared with the training results of the original YOLOv11 model, the average accuracy of the improved model has increased by 2.2 percentage points, and GFLOPs have definitely decreased by 2 G, with parameter reduction of approximately 3.10%. Compared with other mainstream models, the MLD-YOLOv11 model has significant advantages in terms of mean average precision, model size, and floating point operations per second, making it more suitable for industrial applications and providing an efficient, accurate, and lightweight solution for the quality detection of dried green Sichuan pepper.