The accurate determination of discrete element parameters is crucial for ensuring reliable results in simulating the critical post-harvest stages of rice grain (processing, transportation, and storage) with different moisture contents. To determine the discrete element parameters, a physical model of rice grain was constructed by the multi-sphere (MS) modeling approach. Using the repose angle as the evaluation index, the discrete element parameters of rice grain were calibrated and optimized through the Plackett–Burman (PB) test, the steepest climbing test, and the Box–Behnken (BB) test using EDEM software. A moisture content–significance discrete element parameters model was further developed based on a moisture content–repose angle model ($R^2$ = 0.992) and a repose angle–significance discrete element parameters model ($R^2$ = 0.970). The calibration results showed that the relative error between the simulated and actual repose angle did not exceed 3.52%. Meanwhile, the cylinder lifting method and unloading mass flow rate verification were performed. And the results showed that the relative errors of the repose angle and mass flow rate of rice grain did not exceed 2.09% and 7.72%, respectively. The study provides a general and reliable method for determining the parameters of discrete element method simulation for rice grain with different moisture contents.

To solve the problems of poor stability and low monitoring precision in the online detection of rice moisture in the drying tower, we designed an online detection device for rice moisture at the outlet of the drying tower. The structure of a tri-plate capacitor was adopted, and the electrostatic field of the tri-plate capacitor was simulated using COMSOL software. A central composite design of three factors and five levels was carried out with the thickness, spacing, and area of the plates as the influencing factors and the capacitance-specific sensitivity as the test index. This device was composed of a dynamic acquisition device and a detection system. The dynamic sampling device was found to achieve dynamic continuous sampling and static intermittent measurements of rice using a ten-shaped leaf plate structure. The hardware circuit of the inspection system with STM32F407ZGT6 as the main control chip was designed to realize stable communication between the master and slave computers. Additionally, an optimized BP neural network prediction model based on the genetic algorithm was established using the MATLAB software. Indoor static and dynamic verification tests were also carried out. The results showed that the optimal plate structure parameter combination includes a plate thickness of 1 mm, plate spacing of 100 mm, and relative area of 18,000.069 mm² while satisfying the mechanical design and practical application needs of the device. The structure of the BP neural network was 2-90-1, the length of individual code in the genetic algorithm was 361, and the prediction model was trained 765 times to obtain a minimum MSE value of 1.9683 × 10⁻⁵, which was lower than that of the unoptimized BP neural network with an MSE of 7.1215 × 10⁻⁴. The mean relative error of the device was 1.44% under the static test and 2.103% under the dynamic test, which met the accuracy requirements for the design of the device.

To enhance the frictional wear performance of AZ31B magnesium alloy, a nanocrystalline layer was prepared by ultrasonic impact (UI) treatment on magnesium alloy samples, and the effects of different ultrasonic impact times (5 min, 10 min, and 15 min) on the organization and wear resistance of the composite coating were studied. The findings revealed that the maximum thickness of the composite coating was about 50 μm after 10 min of impact time, which was approximately 15 μm higher than that of the MAO coating. The elemental composition of the composite coating was mainly Mg, O, and Si elements, and the phase structure of the coating, mainly MgO and Mg₂SiO₄, was the same before and after UI. The microhardness value gradually decreased in a gradient with the increasing distance from the sample surface. The coating had a lower average coefficient of friction (0.18) and lower wear loss (1.56 mg) for the 15 min impact time. Moreover, a small amount of abrasive and adhesive wear served as the primary modes of wear.

Cracking in paddy during processing and storage can significantly degrade its quality and economic value. This study elucidates the crack formation law in paddy across various drying stages through experiments and simulations that integrate multiple physical fields. A predictive model for crack occurrence, based on the relationship between the probability of cracking and drying time, is developed by introducing a critical moisture evaporation coefficient to delineate safe regio for cracking. The findings indicate a sharp increase in the percentage of drying cracks during the initial rapid drying phase, with continued escalation through the slower drying phase. The predictive model’s coefficient of determination exceeded 0.85, demonstrating its efficacy in forecasting crack progression. The primary driver for crack initiation and growth was identified as the cumulative effect of the moisture gradient, with the yield stress proving inadequate to counteract the stresses induced by this gradient, leading to cracking in paddy grains. These results furnish valuable insights for accurately predicting and managing crack development in paddy.