CFD-DEM coupling simulation modeling method of corn kernel for air-assisted precision seed metering device

Abstract

The application relates to the technical field of agricultural engineering simulation modeling, and discloses a CFD-DEM coupling simulation modeling method for a corn kernel of an air-suction type precision seed metering device, which comprises the following steps: acquiring the shape quantity proportion, main size and mapping function of target batch kernels, and determining the main size probability distribution; acquiring physical and contact parameters, and calibrating the rolling friction coefficient in the contact parameters; establishing a multi-ball model local coordinate system to calculate multi-ball geometric data; calling multi-ball geometric generation rules, establishing a discrete element rigid multi-ball kernel template and a discrete element template library; verifying the template, and feeding back and correcting when the error is out of limit; forming discrete element model data files from the verified data, and outputting to a CFD-DEM coupling platform to generate a corn kernel group for simulation calculation. The application restores the size variation distribution characteristics of the kernel group, guarantees the consistency of the geometric shape and macro mechanical characteristics of the simulation, and improves the accuracy of the simulation calculation of the seed metering device.

Description

Technical Field

[0001] This invention relates to the field of agricultural engineering simulation modeling technology, specifically to a CFD-DEM coupled simulation modeling method for corn kernels using an air-suction precision seed metering device. Background Technology

[0002] Air-suction precision seed metering devices utilize negative pressure airflow to adsorb grains and complete the seeding operation. CFD-DEM coupled simulation methods are commonly used to numerically calculate and structurally analyze the multiphase flow process of airflow and grains. However, existing methods for constructing discrete element simulation models of corn grains often use single or a small number of fixed multi-sphere templates to represent the entire corn grain population, neglecting the morphological diversity and specific size probability distribution patterns within a real corn batch. This results in discrepancies between the simulated population and the actual macroscopic material distribution characteristics.

[0003] In the geometric generation process of discrete element multi-sphere models, existing technologies often use a uniform sphere filling algorithm without taking into account the specific external contour differences of corn kernels to formulate corresponding space filling and coordinate allocation rules. Furthermore, they lack a validity check and feedback mechanism for the three-dimensional envelope size and the overlapping state of adjacent spheres, which leads to contour distortion and volume calculation errors in the constructed single-grain model.

[0004] Existing calibration methods for determining contact parameters required for discrete element simulation are relatively simplistic, often relying on empirical values ​​or obtaining parameters solely through single angle-of-repose tests. They lack a closed-loop feedback correction process that integrates various verification scenarios, such as material discharge conditions, resulting in the obtained rolling friction coefficient failing to objectively reflect the mechanical contact characteristics under complex mass accumulation and flow states. These deficiencies in the geometric modeling mechanism and parameter acquisition process lead to deviations in the model data ultimately input to the coupling platform, limiting the reference value of the simulation analysis results for the air-suction seed metering device. Summary of the Invention

[0005] The technical problem solved by this invention is that the existing discrete element simulation modeling method for maize kernels is difficult to fully characterize the morphological diversity and geometric size distribution characteristics of kernel groups, and the generation rules of multi-sphere models lack specificity. The acquisition and calibration process of physical and contact parameters is not perfect, which leads to deviations in the model data of CFD-DEM coupled simulation of air suction precision seed metering device, making it difficult to meet the needs of seed metering device simulation analysis.

[0006] To address the above problems, the present invention provides the following technical solution: This invention provides a CFD-DEM coupled simulation modeling method for maize kernels using a pneumatic precision seed metering device, comprising the following steps: Obtain the morphological quantity ratio, main size, and mapping function of subordinate sizes relative to the main size of corn kernels in the target batch, and determine the probability distribution of the main size. Obtain the physical and contact parameters of the target batch of corn kernels, and calibrate the rolling friction coefficient in the contact parameters; Establish a local coordinate system for the multisphere model, calculate the multisphere geometric data of the target batch of corn kernels, and perform a validity check; Based on the proportion of morphological quantity, the probability distribution of main size, the mapping function, and the multi-sphere geometric data, the corresponding multi-sphere geometric generation rules are constructed and called. Discrete element rigid multi-sphere seed templates and discrete element template libraries are established by combining physical parameters and contact parameters. The discrete element rigid multi-sphere seed template is validated. When the error exceeds the preset limit, feedback correction is performed until the error is within the preset limit. The validated data is organized into a discrete element model data file, which is then output to the CFD-DEM coupling platform to generate a corn kernel cluster for simulation calculations of the air-suction precision seed metering device.

[0007] Furthermore, in the steps of obtaining the morphological quantity proportions, main size, and subordinate size mapping functions of the target batch of corn kernels relative to the main size, and determining the probability distribution of the main size, obtaining the morphological quantity proportions, main size, and subordinate size mapping functions of the target batch of corn kernels relative to the main size includes: Based on their external contours, the target batch of corn kernels was divided into dent, frustum, conical and quasi-spherical types, and the proportion of each type in the target batch of corn kernels was calculated. The overall height, main body height, maximum main body width, maximum main body thickness, and the width and thickness of the transition section at the tip of the embryo are measured as candidate dimensions. Calculate the correlation coefficient between each candidate size and the other sizes, and determine the size with the largest average correlation coefficient as the main size of the corresponding shape, and the other candidate sizes as subordinate sizes; By performing regression fitting using measured sample data, the function parameters of the mapping function between the subordinate dimension and the principal dimension are obtained.

[0008] Furthermore, physical parameters include moisture content, grain density, Young's modulus, and Poisson's ratio, while contact parameters include the collision recovery coefficient, static friction coefficient, and rolling friction coefficient. In the step of obtaining the physical parameters and contact parameters of the target batch of corn kernels, and calibrating the rolling friction coefficient in the contact parameters, calibrating the rolling friction coefficient in the contact parameters includes: A fixed funnel discharge method was used to measure the angle of repose and the height of the pile. A fixed outlet hopper was used to conduct a discharge test and record the discharge time and the mass discharged per unit time. Establish a discrete element calibration model and input the obtained physical parameters, collision recovery coefficient, and static friction coefficient; Establish a parameter combination table for the rolling friction coefficient, conduct repose angle simulation and material discharge simulation for each group of parameters, and determine the final calibrated rolling friction coefficient based on the error.

[0009] Furthermore, a local coordinate system for the multi-sphere model is established, the multi-sphere geometric data of the target batch of corn kernels are calculated, and a validity check is performed, specifically including: Set the center of the top cutting plane of the grain as the local coordinate origin, and establish a local coordinate system for the multi-sphere model with the grain width direction as the x-axis, the grain height direction as the y-axis, and the grain thickness direction as the z-axis. The positive direction of the y-axis points from the top of the grain to the tip of the embryo. The number of y-axis filling layers is determined based on the target contour dimensions, geometric dimension error limits, and the number of constituent spheres. Based on different grain shapes and the number of filling layers, the radius of the constituent spheres, the coordinates of the sphere centers, and the rigid aggregation relationship between the constituent spheres within the same grain are calculated and combined to generate multi-sphere geometric data of various shapes.

[0010] Furthermore, the multisphere geometric data is generated in the following manner: For the preset tooth-shaped and frustum-shaped types, the main body area is divided into multiple layers along the y-axis. The width and thickness of each layer gradually transition from the maximum size of the main body to the transition section size of the tip of the embryo. Based on this, the radius and center coordinates of the constituent sphere are calculated. For the preset spherical cone shape, multiple layers are set along the y-axis, and each layer is set with three spheres of equal radius arranged along the x-axis; For the preset spherical shape, a central sphere is set at the center of the three-dimensional bounding box, and six axial auxiliary spheres with the same radius as the central sphere are arranged in the positive and negative x-axis, positive and negative y-axis and positive and negative z-axis directions, respectively.

[0011] Further, a validity check is performed, specifically including: Check the three-dimensional envelope size, constituent sphere radius, overlap state of adjacent constituent spheres and model volume of the four preset grain morphologies, and check the position, width and thickness of the embryo tip transition section of the dentate, frustum and spherical cone types; When a non-positive constituent sphere radius is detected, it is recalculated. When adjacent constituent spheres do not overlap, the distance between their centers is adjusted. When the three-dimensional envelope size error exceeds the geometric size error limit or the model volume error exceeds the volume error limit, the number of filling layers, the position of the sphere center, or the constituent sphere radius is adjusted.

[0012] Furthermore, based on the proportion of morphological quantities, the probability distribution of principal dimensions, the mapping function, and the multi-sphere geometric data, corresponding multi-sphere geometric generation rules are constructed and invoked. Discrete element rigid multi-sphere seed templates and a discrete element template library are established by combining physical parameters and contact parameters, specifically including: The grain shape is determined based on the proportion of different shapes, and the main size is generated based on the probability distribution of the main size of the corresponding shape. The main size is then substituted into the mapping function to calculate all subordinate sizes. The corresponding multi-sphere geometry generation rules are called to calculate the number of constituent spheres, the radius of the constituent spheres, and the coordinates of the center of the spheres. In the discrete element simulation platform, a discrete element rigid multi-sphere seed template is established by combining according to the rigid aggregation relationship. The main size range, main size probability distribution, mapping function, multi-sphere geometry generation rules, physical parameters, contact parameters, and template call weights of various types of grains are stored in the discrete element template library.

[0013] Furthermore, the discrete element rigid multi-sphere grain template was validated, including geometric consistency validation, angle of repose validation, and material feeding process validation. Geometric consistency verification compares the overall width, overall thickness, overall height, single grain volume, and single grain mass of the measured grains with those of the discrete element rigid multi-sphere grain template; When the error in geometric consistency verification exceeds the preset limit, the number of filling layers, the radius of the constituent spheres, and the position of the sphere center are readjusted. The angle of repose was verified by comparing the angle of repose and the stacking height of the actual grain pile and the simulated grain pile. The discharge process was verified by comparing the discharge time and the discharge mass per unit time of the actual test and the simulation test. When the error in the repose angle verification or the material discharge process verification exceeds the preset limit, the rolling friction coefficient in the contact parameters is recalibrated.

[0014] Furthermore, the output content of the discrete element model data file includes: The composition of various grain templates includes the number of constituent spheres, sphere center coordinates, constituent sphere radius, rigid aggregation relationship, local coordinate direction, single grain mass, single grain volume, morphological quantity ratio, main size probability distribution, grain material parameters, seed metering device boundary material parameters, and contact parameters. After reading the discrete element model data file, the CFD-DEM coupling platform determines the grain morphology according to the proportion of morphological quantity, generates the main size according to the probability distribution of the main size of the corresponding morphology, calculates the subordinate size through the mapping function, and calls the corresponding multi-sphere geometry generation rule to establish a discrete element rigid multi-sphere grain template. The generated corn grain group is used for simulation calculation.

[0015] Furthermore, when establishing a discrete element rigid multi-sphere seed template: The volume of a single grain is the union volume of all constituent spheres within the same grain, and overlapping regions are calculated only once; The mass of a single grain is determined based on the volume of a single grain and the grain density in the physical parameters. The position of the center of mass and the moment of inertia are calculated based on the mass distribution of the union of the constituent spheres.

[0016] This invention provides a CFD-DEM coupled simulation modeling method for maize kernels using an air-suction precision seed metering device. It offers the following advantages: 1. This invention classifies target batches of corn kernels into dent, frustum, conical, and quasi-spherical shapes according to their external contours. It extracts the size with the largest average correlation coefficient as the principal size and establishes a mapping function between subordinate sizes and the principal size. Combined with the probability distribution of the principal size, it calculates the three-dimensional size of each type of kernel. This parameterized morphology construction method based on statistical and regression analysis restores the polymorphism and size distribution characteristics of the actual corn kernel population and improves the fit between the discrete element simulation population and the real kernel population at the macroscopic geometric level.

[0017] 2. This invention establishes multi-sphere geometric generation rules for the three-dimensional contour differences of different morphological grains. By setting a local coordinate system and performing layered filling calculations according to morphological features, multi-sphere geometric data is generated. Based on checking the radii of the constituent spheres, the overlap state, and the three-dimensional envelope size, the geometric consistency of a single grain is verified. When the error exceeds the limit, closed-loop feedback is triggered to readjust the center position of the spheres and the radii of the constituent spheres. The above process overcomes the problem of shape distortion when using conventional fixed templates to characterize multimorphic crop varieties, and ensures the geometric consistency of the discrete element rigid multi-sphere grain template.

[0018] 3. This invention constructs a discrete element calibration model of mechanical contact parameters by conducting repose angle measurement and material feeding experiments. It uses a parameter combination table to perform multi-condition simulation comparison and error feedback on the rolling friction coefficient. The calibrated contact parameters and multi-sphere geometry generation rules are stored in a template library. Finally, the data is packaged and output as a discrete element model data file suitable for the CFD-DEM coupling platform. This mechanism standardizes the acquisition and verification process of simulation parameters, enabling the corn kernel group input to the coupling simulation to have reliable group mechanical contact characteristics, thus meeting the data requirements of fluid-structure interaction calculation of the air suction seed metering device. Attached Figure Description

[0019] Figure 1 This is a system framework diagram of the present invention; Figure 2 This is a flowchart of the method of the present invention; Figure 3 This is a flowchart of the target grain classification and geometric parameterization process of the present invention; Figure 4 This is a flowchart of the physical and contact parameter acquisition and calibration process of the present invention; Figure 5 This is a flowchart illustrating the generation process of four types of multi-sphere geometric data for seeds according to the present invention; Figure 6 This is a flowchart illustrating the construction of the parameterized template library, model validation, and model data output of the present invention. Figure 7This is a schematic diagram of the construction of the discrete element rigid multisphere model of corn kernels according to the present invention. (a) is a schematic diagram of the geometric parameters of corn kernels and the filling relationship of the constituent spheres, and (b) is a schematic diagram of four types of corn kernel DEM models. Detailed Implementation

[0020] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] See attached document Figure 1 This invention provides a CFD-DEM coupled simulation modeling system for corn kernels using an air-suction precision seed metering device. The system includes modules for geometric parameter acquisition and parameterization, physical and contact parameter acquisition and calibration, multi-sphere geometric data generation, discrete element template library construction, model verification and feedback correction, and model data output.

[0022] The geometric parameter acquisition and parameterization module receives the morphological classification results and geometric size measurement data of the target batch of corn kernels. This module determines the principal size based on the correlation between the various sizes of kernels of the same type, establishes a mapping function between subordinate sizes and the principal size, and statistically analyzes the proportion of each type of kernel morphology and the probability distribution of the principal size.

[0023] The physical and contact parameter acquisition and calibration module measures the moisture content, kernel density, Young's modulus, and Poisson's ratio of corn kernels, and obtains the collision recovery coefficient and static friction coefficient of the contact pairs between kernels and the contact pairs between kernels and the seed metering device boundary. Based on this, the rolling friction coefficients of the two types of contact pairs are calibrated through the angle of repose test and the fixed outlet hopper discharge test.

[0024] The multi-sphere geometric data generation module establishes a local coordinate system based on the grain morphology and geometric feature dimensions, calculates the number of constituent spheres, the radius of the constituent spheres, and the coordinates of the center of the spheres for dentate, frustum, spherical, and quasi-spherical grains, determines the rigid aggregation relationship between the constituent spheres within the same grain, and checks the geometric validity of the generated results.

[0025] The Discrete Element Template Library (DEM) module generates the primary size of the grains according to the probability distribution of the primary size, and then calculates the subordinate sizes through a mapping function. After completing the size calculation, it calls the corresponding multi-sphere geometry generation rules to calculate the number of constituent spheres, the radius of the constituent spheres, and the coordinates of the sphere centers, thus establishing a DEM rigid multi-sphere grain template. This module writes the material parameters and contact parameters into the DEM model, and stores the primary size ranges of the four types of grains, mapping parameters, multi-sphere geometry generation rules, and template call weights into the DEM template library.

[0026] The model validation and feedback correction module performs geometric consistency validation, angle of repose validation, and material feeding process validation on the discrete element model. When the error in geometric consistency validation exceeds the preset limit, the number of filling layers, the radius of the constituent spheres, and the position of the sphere center are readjusted. When the errors in angle of repose validation and material feeding process validation exceed the preset limits, the rolling friction coefficients of the two types of contact pairs are recalibrated. After the parameters are adjusted, the grain template is rebuilt and validated again.

[0027] The model data output module organizes the verified component sphere data, morphological quantity ratios, principal size probability distributions, material parameters, and contact parameters, and forms discrete element model data files according to the data reading format of the CFD-DEM coupling platform.

[0028] See attached document Figure 2 This invention provides a CFD-DEM coupled simulation modeling method for maize kernels using an air-suction precision seed metering device, comprising the following steps: S100 involves morphological classification and geometric parameterization of the target batch of corn kernels. Kernels are categorized into dent, frustum, conical, and quasi-spherical types based on their external contours. The placement, measurement reference, and coordinate orientation of each type of kernel are standardized. The overall height, main body height, maximum main body width, maximum main body thickness, and the width and thickness of the embryo tip transition section are measured. The principal dimensions are determined based on the correlation between the dimensions of kernels within the same category. A mapping function between each subordinate dimension and the principal dimension is established, and the probability distribution of the principal dimensions for each type of kernel is determined based on the measured principal dimension data.

[0029] S200: Obtain the physical and contact parameters of corn kernels and calibrate the rolling friction coefficient. Measure the moisture content, kernel density, Young's modulus, and Poisson's ratio of the target batch of kernels to obtain the collision recovery coefficient between kernels, the collision recovery coefficient between kernels and the seed metering device boundary, the static friction coefficient between kernels, and the static friction coefficient between kernels and the seed metering device boundary. Using the angle of repose, stacking height, discharge time, and mass discharged per unit time as criteria, calibrate the rolling friction coefficient between kernels and the rolling friction coefficient between kernels and the seed metering device boundary.

[0030] S300 generates multi-sphere geometric data for four types of maize kernels. A local coordinate system is established for the multi-sphere model. Based on geometric features, the radii, center coordinates, and rigid aggregation relationships of the constituent spheres for dent, frustum, conical, and quasi-spherical kernels are calculated. After calculation, the three-dimensional envelope dimensions, constituent sphere radii, overlap of adjacent constituent spheres, and model volume are checked. For dent, frustum, and conical kernels, the position, width, and thickness of the embryo tip transition section are also checked. If the error exceeds the preset limit, the number of filling layers, constituent sphere radii, and center positions are adjusted, and the multi-sphere geometric data is regenerated.

[0031] S400 establishes a discrete element template library and a discrete element model. Based on the grain morphology and principal size probability distribution, it generates the principal size of a single grain. Substituting the principal size into a mapping function calculates all subordinate sizes. Then, it calls the multi-sphere geometry generation rules for the corresponding morphology in S300 to calculate the number of constituent spheres, their radii, and their center coordinates.

[0032] The grain material parameters, seed metering device boundary material parameters, and contact parameters are written into the discrete element simulation platform to establish a discrete element rigid multi-sphere grain template. The main size range, main size probability distribution, subordinate size mapping parameters, multi-sphere geometric generation rules, material parameters, contact parameters, and template call weights of the four types of grains are stored in the discrete element template library.

[0033] S500 is used to verify and correct the discrete element model. Grain samples not involved in parameter fitting and calibration are selected for geometric consistency verification, angle of repose verification, and discharge process verification. Geometric consistency verification compares the overall width, thickness, height, single grain volume, and single grain mass of the measured grains with those of the grain template; for dent, frustum, and conical grains, it also compares the main body height, embryo tip transition section width, and embryo tip transition section thickness. Angle of repose verification compares the angle of repose and stacking height of the actual grain pile with those of the simulated grain pile. Discharge process verification compares the discharge time and mass discharged per unit time of the actual experiment with those of the simulation experiment. If any verification result exceeds the preset error limit, the corresponding step is returned to adjust the parameters, and then the model is rebuilt and verified again until all errors are within the limits.

[0034] The S600 process organizes and outputs discrete element model data for corn kernels. The output includes the number of constituent spheres, sphere center coordinates, sphere radii, rigid aggregation relationships, local coordinate directions, single kernel mass, single kernel volume, morphological quantity ratio, principal size probability distribution, kernel material parameters, seed metering device boundary material parameters, and contact parameters for each type of kernel template. This output data is used by the CFD-DEM coupling platform for the air-suction high-speed precision seed metering device.

[0035] See attached document Figure 3 Step S100 includes the following sub-steps: S110. Sampling and morphological classification of the target batch of corn kernels. Before classification, kernels are placed with their ventral sides facing the same direction, and the shape of the kernel top, the main cross-section, the embryo tip outline, and the three-dimensional dimensional relationship are used as the criteria for classification. Kernels with a concave top and both sides of the main body tapering towards the embryo tip are classified as dent type; kernels with a quadrilateral main cross-section and a side profile tapering towards the embryo tip are classified as frustum type; kernels with a nearly circular main cross-section and a continuously tapering outer profile towards the embryo tip are classified as conical type; kernels with a three-dimensional dimensional relationship within the spherical type range and a continuously smooth outer profile are classified as spherical type. After classification, the quantity of each type of kernel is recorded, and the proportion of each type in the target batch is calculated.

[0036] The spherical shape classification range is determined based on the statistical distribution of the three-dimensional size ratios of the target batch of grains. All samples are classified using the same set of contour classification criteria.

[0037] S120 defines the geometric dimensions of various types of kernels and completes the measurement. For dent, frustum, and conical kernels, the top tangential plane of the kernel is used as the starting point for height measurement, and the lowest point of the embryo tip is used as the ending point. The distance between the two is defined as the total height. The section where the main body contour begins to enter the embryo tip contraction segment is defined as the embryo tip transition section, and the distance from the top tangent plane of the kernel to this section is defined as the main body height. .

[0038] The maximum width of the main area is defined as The maximum thickness of the main region is defined as The width of the transition section at the tip of the embryo is defined as The thickness of the transition section at the tip of the embryo is defined as .

[0039] For spherical grains, a three-dimensional bounding box is constructed according to a uniform arrangement. The maximum envelope size along the x-axis is defined as the overall width. The maximum envelope size in the z-axis direction is defined as the total thickness. The maximum envelope size along the y-axis is defined as the total height. .

[0040] Measure the size of each grain individually according to the same posture, measurement reference and measurement direction, and record the grain number, morphological category and various size data.

[0041] S130, perform correlation analysis on the geometric dimension data and determine the main dimension. For the same shape, calculate the correlation coefficient between each candidate dimension and the other dimensions, then calculate the average correlation coefficient corresponding to the candidate dimension, and determine the dimension with the largest average correlation coefficient as the main dimension of the shape.

[0042] The thickness of the transition section at the germ tip of dentate kernels As the main dimensions, frustum-shaped and conical kernels are characterized by their main height. As the main size, the spherical kernel is characterized by its overall thickness. As the main size.

[0043] After the main dimensions are determined, the correlation coefficient is used. It represents the degree of correlation between each subordinate dimension and its corresponding master dimension.

[0044] S140, Establish the subordinate dimension mapping function. For each type of shape and each subordinate dimension, perform regression fitting using measured sample data to obtain the function parameters of the subordinate dimension relative to the main dimension. The mapping function takes the following form: ; In the formula, This indicates the subordinate dimension, in mm; This indicates the main dimension of the shape, in mm; Indicates the first regression coefficient; This represents the second regression coefficient. Each subordinate dimension corresponds to a set of regression coefficients.

[0045] S150, determine the proportion of different morphologies and the probability distribution of major sizes. Statistically determine the proportion of each type of grain in the target batch sample, and perform distribution tests on the major size data for each type of grain. If the major size data conforms to a normal distribution, a truncated normal distribution with defined upper and lower bounds is used; if the major size data does not conform to a normal distribution, a discrete probability table is established based on the measured frequencies.

[0046] When generating the model, the grain shape is first determined based on the proportion of different shapes, then the main size is generated based on the probability distribution of the main size of the shape, and then the main size is substituted into the mapping function to calculate the various subordinate sizes of the same grain.

[0047] The generated dimensions should meet the following requirements. Greater than , Not less than And all dimensions are greater than zero. For spherical grains, the generated... , and All values ​​should be greater than zero. If the dimensional data does not meet the above constraints, discard the data set and regenerate it.

[0048] See attached document Figure 4 Step S200 includes the following sub-steps: S210, Determine the moisture content of the target batch of corn kernels. A predetermined weight of kernel samples is randomly selected from the target batch, divided into equal portions, crushed and ground, and then measured using a moisture detection device. The average value of all measurement results is taken as the moisture content of the target batch.

[0049] Moisture content is used to indicate the test status of physical and contact parameters. Each parameter set corresponds to a defined moisture content range. If the moisture content of the target batch exceeds this range, S200 is re-executed.

[0050] S220, the grain density was determined using the liquid displacement method. First, the mass of the empty distillation flask was weighed, followed by the mass of the flask filled with pure water. Then, a predetermined number of grains were added to the empty flask and weighed. Next, the flask containing the grains was filled with pure water, and air bubbles were expelled before weighing. The water temperature, grain quantity, and soaking time were kept consistent across all groups of experiments.

[0051] Grain density is calculated using the following formula: ; In the formula, This indicates grain density, expressed in g / cm³. 3 ; Indicates the number of independent trials; Indicates the test number; This indicates the mass of the empty distillation flask, in grams. This indicates the total mass of the distillation flask and the grains, in grams. This indicates the mass of the distillation flask filled with pure water, expressed in grams. This indicates the total mass of the distillation flask after adding the seeds and filling it with pure water, expressed in grams. This indicates the density of pure water at the test temperature, expressed in g / cm³. 3 .

[0052] S230 uses a biplane compression test to determine Young's modulus and Poisson's ratio. Individual kernels are placed flat in a uniform orientation at the center of the compression platform, with the loading direction aligned with the kernel thickness direction. After the loading head moves to contact the kernel surface, it compresses downwards at a constant speed, while simultaneously recording the compressive force, axial displacement, and lateral deformation.

[0053] Loading is stopped after the compressive force versus displacement curve reaches its first peak, and the stable elastic segment from the initial contact point to the first peak is selected. Poisson's ratio is calculated based on the ratio of transverse contraction strain to axial compressive strain within this elastic segment, and Young's modulus is calculated using the Hertzian contact relationship based on the compressive force and axial deformation. ; ; ; In the formula, This represents Young's modulus, expressed in Pa. This represents the compressive force corresponding to the data point within the stable elastic segment, expressed in N. Indicates Poisson's ratio; This represents the axial deformation corresponding to the same data point. Represents the shape factor in Hertzian contact relationships; This represents the principal radius of curvature of the contact area between the grain and the pressing plane; This represents the principal radius of curvature of the contact area between the grain and the upper pressing plane; Indicates the overall height; Indicates the maximum width of the main body; This indicates the maximum thickness of the main body.

[0054] Length parameters in the compression test were uniformly converted to meters (m) before calculation. For each seed, continuous data points within the stable elastic range were selected to calculate Young's modulus, and the average value of the calculation results for multiple seeds was used as the target batch parameter. The Young's modulus and Poisson's ratio of the seed metering device boundary material were obtained through testing corresponding material samples.

[0055] S240, Determination of the collision recovery coefficient. The collision recovery coefficient between kernels is determined using a simple pendulum test. Two kernels of the same shape and main size range are selected and fixed at the ends of equal-length flexible lines, so that the two kernels are at the same horizontal position and in contact with each other in a static state. One side of the kernels is lifted to a set height and then released, and the maximum height reached by the two kernels after the collision is recorded.

[0056] The collision recovery coefficient between kernels is calculated using the formula for the collision recovery coefficient of a simple pendulum: ; In the formula, Indicates the collision recovery coefficient between grains; Indicates the number of independent trials; Indicates the first The height at which the grains were released in this experiment; This indicates the maximum height reached by the first seed after the collision. This indicates the maximum height reached by the second seed after the collision. All heights use the same unit of length.

[0057] The collision recovery coefficient between the grain and the seed metering device boundary was determined using a free-fall test. Individual grains were fixed above the boundary material of the seed metering device in a uniform posture, and the vertical distance from the release point to the material surface was recorded. After releasing the grain, the maximum rebound height after the grain collided with the boundary material was recorded.

[0058] The collision recovery coefficient between the grain and the boundary of the seed metering device is calculated using the formula for the collision recovery coefficient of free fall: ; In the formula, This represents the collision recovery coefficient between the grain and the boundary of the seed metering device; Indicates the number of independent trials; Indicates the first Release height of this test; Indicates the first The maximum rebound height of this test. All heights use the same unit of length.

[0059] When the seed metering device contains multiple contact materials, samples of the corresponding materials are prepared respectively, the collision recovery coefficient between the grain and various boundary materials is measured, and the measurement results are written into the contact parameter matrix respectively.

[0060] S250, the static friction coefficient is determined by an inclined plane test. To determine the static friction coefficient between the grain and the seed metering device boundary, a single grain is placed on the surface of the boundary material of the seed metering device, the inclination angle of the inclined plane is slowly increased, and the critical inclination angle at which the grain begins to slide continuously is recorded.

[0061] When determining the static friction coefficient between grains, grains of the same batch are fixed on the opposing surfaces of two rigid substrates, so that the surfaces of the upper and lower layers of grains are in direct contact. The inclination angle of the inclined plane is slowly increased, and the critical inclination angle at which the upper substrate begins to slide continuously is recorded.

[0062] The static friction coefficient is calculated using the formula for the static friction coefficient at the critical tilt angle: ; In the formula, This represents the static friction coefficient of the corresponding contact pair; in the static friction test between grains, it is denoted as... In the static friction test between the grain and the boundary of the seed metering device, it was recorded as ; Indicates the number of independent trials; Indicates the first The critical tilt angle measured in this experiment, in degrees; This represents the tangent value of the critical tilt angle.

[0063] When the seed metering device contains multiple contact materials, the static friction coefficient between the grain and the boundary material of each type of seed metering device is measured.

[0064] S260, calibrated rolling friction coefficient. The rolling friction coefficient between grains is denoted as... The coefficient of rolling friction between the grain and the boundary of the seed metering device is denoted as . .

[0065] The target batch of grains was discharged into a horizontal receiving plate using a fixed funnel, and the angle of repose and stack height were measured after the grain pile stabilized. A discharge test was conducted using a fixed outlet hopper, and the discharge time and mass discharged per unit time were recorded. Based on the structural dimensions, feed mass, feed location, and boundary materials of the two actual tests, discrete element calibration models were established, and the parameters obtained from S220 to S250 were input.

[0066] Establish within the set parameter range and The parameter combination table was used to perform repose angle simulation and material discharge simulation for each group of parameters. Parameter combinations where the repose angle error, stacking height error, discharge time error, and discharge mass error per unit time were all within preset limits were retained, and the errors corresponding to each group of parameters were compared.

[0067] Based on the selected parameter combinations, the parameter intervals were narrowed, a new parameter combination table was established, and simulations were conducted. The rolling friction coefficients between grains and between grains and the seed metering device boundary were determined based on the refined calibration parameters: angle of repose, stacking height, discharge time, and discharge quality error per unit time. The grain samples used in the calibration experiment were no longer used for model validation in S500.

[0068] See attached document Figure 5 Step S300 includes the following sub-steps: S310, Establish the local coordinate system and common constraints for the multi-sphere model. Set the center of the top tangent plane of the grain as the origin of the local coordinate system, with the grain width direction as the x-axis, the grain height direction as the y-axis, and the grain thickness direction as the z-axis. The positive direction of the y-axis points from the top of the grain to the tip of the embryo.

[0069] The number of y-axis filling layers is determined based on the target contour dimensions, geometric dimension error limits, and the number of constituent spheres. The constituent spheres within the same grain overlap and combine into a discrete element particle according to a rigid aggregation relationship. No internal contact calculations are performed between the constituent spheres.

[0070] S320 generates multi-sphere geometric data for dentate grains. The main region is divided along the y-axis into... The layers, each with a width and thickness gradually transitioning from the maximum dimension of the main body to the transition section dimension at the tip of the embryo. Layer width, thickness, constituent sphere radius, interlayer spacing, and embryo tip sphere radius are calculated using the following formula: ; In the formula, ; Indicates the first Layer width; Indicates the first Layer thickness; Indicates the first The sphere's radius is composed of layers; Indicates the radius of the first-layer sphere; Indicates the radius of the sphere that makes up the last layer; Indicates the distance between the centers of adjacent main layers; This indicates the radius of the embryonic apex. All length parameters are in mm.

[0071] The spherical coordinates of each layer and the spherical coordinates of the embryo tip are generated according to the following formula: ; The plus and minus signs respectively represent the first and second signs. Each layer consists of two constituent spheres, with each symbol value corresponding to a sphere's center coordinate. When the lateral offset is zero, the layer retains only one central sphere. When the lateral offset is less than zero, the layer sets one central sphere, with its radius being the smaller of half the layer's width and half its thickness.

[0072] After the constituent spheres of each layer and the embryo tip spheres are combined according to a rigid polymerization relationship, they form the geometric data of the multi-sphere dentate grain.

[0073] S330 generates the geometric data for the multi-sphere structure of the frustum-shaped grain. The frustum-shaped grain uses the same layer width and thickness calculation relationships along the y-axis as in S320. Substituting the radii of the first and last layer constituent spheres obtained in this step into the interlayer spacing calculation relationships in S320, we obtain... radius of the embryonic apex According to the overall height With the height of the main body The difference is determined. The radius and coordinates of the sphere formed by the layers are generated according to the following formula: ; In the formula, Pick ; and They represent the first The calculation relationship for layer width and thickness is the same as that in S320; Indicates the first The sphere's radius is composed of layers; Indicates the distance between the centers of adjacent main layers; This indicates the radius of the embryonic apex. All length parameters are in mm.

[0074] The positive and negative signs in the x and z coordinates are combined to obtain values, with each combination corresponding to a sphere center coordinate. Four corner spheres are generated in the same layer. When the offset in a certain coordinate direction is zero, the spheres with overlapping centers are merged. After the spheres in each layer are combined with the embryo tip spheres according to a rigid aggregation relationship, a frustum-shaped multi-sphere geometric data of the grain is formed.

[0075] S340 generates multi-sphere geometry data for spherical conical grains. The spherical conical grains are set along the y-axis. Each layer consists of three spheres of equal radius. The width, thickness, sphere radius, interlayer spacing, and embryonic tip radius of each layer are calculated using the following formula: ; In the formula, ; Indicates the first Layer width; Indicates the first Layer thickness; Indicates the first The sphere's radius is composed of layers; Indicates the radius of the first-layer sphere; Indicates the radius of the sphere that makes up the last layer; Indicates the distance between the centers of adjacent main layers; This indicates the radius of the embryonic apex. All length parameters are in mm.

[0076] No. The first layer The spherical coordinates of the constituent spheres and the spherical coordinates of the embryonic tip are generated according to the following formula: ; In the formula, Take values ​​1 to 3. The three constituent spheres are arranged along the x-axis, with the middle constituent sphere located at the center of the x-axis and the two outer constituent spheres symmetrical about the y-axis. After the constituent spheres of each layer are combined with the embryo tip sphere according to a rigid aggregation relationship, they form the spherical cone-shaped multisphere geometric data of the grain.

[0077] S350 generates multi-sphere geometry data for spherical grains. A spherical grain consists of a central sphere and six axial auxiliary spheres.

[0078] The center of the central sphere is located at the center of the 3D bounding box. In the local coordinate system established in S310, this center is located at the midpoint of the overall height direction. The radius of the central sphere is taken as... , , The smallest of the three and a half dimensions. Six axial auxiliary spheres with the same radius as the central sphere are arranged in the positive x-axis, negative x-axis, positive y-axis, negative y-axis, positive z-axis, and negative z-axis, respectively.

[0079] The distance between the center of the auxiliary sphere and the center of the central sphere along the x-axis is taken as the total width. Half of the radius of the central sphere minus the radius of the central sphere; the distance between the center of the auxiliary sphere and the center of the central sphere along the z-axis is taken as the total thickness. Half of the radius of the central sphere; the distance between the center of the auxiliary sphere and the center of the central sphere along the y-axis is taken as the total height. Half of the radius of the central sphere minus the radius of the central sphere. Auxiliary spheres along each axis overlap with the central sphere, and spheres whose centers coincide are merged.

[0080] When all the constituent spheres are combined according to a rigid aggregation relationship, they form a spherical grain multi-sphere geometric data.

[0081] S360 performs a validity check on the multi-sphere geometric data. For all four types of kernels, the three-dimensional envelope dimensions, constituent sphere radii, overlap of adjacent constituent spheres, and model volume are checked. For dentate, frustum-shaped, and conical kernels, the position, width, and thickness of the embryo tip transition section are also checked.

[0082] When a non-positive constituent sphere radius is detected, recalculate the cross-sectional dimensions of that layer and the constituent sphere radius. If adjacent constituent spheres do not overlap, adjust the distance between their centers. If the three-dimensional envelope dimension error exceeds the geometric dimension error limit, adjust the number of infill layers and the arrangement of the sphere centers for that shape. If the model volume error exceeds the volume error limit, adjust the constituent sphere radii and recheck.

[0083] Once all inspection errors are within the preset limits, the system outputs the number of constituent spheres, sphere center coordinates, constituent sphere radii, local coordinate directions, and rigid aggregation relationships.

[0084] See attached document Figure 6 Steps S400, S500, and S600 include the following sub-steps: S410 generates the geometric parameters of a single grain. The grain morphology is determined based on the proportion of different morphologies in the target batch, and then the principal size is generated based on the probability distribution of the principal size for that morphology. The principal size is then substituted into the mapping function established in S140 to calculate all dependent sizes of the grain.

[0085] After completing the dimension calculations, check according to the constraints in S150. If the dimension data does not meet the constraints, discard the data set and regenerate the master dimension.

[0086] S420 invokes the corresponding multi-sphere geometry generation rule. The dentate grain uses the generation rule in S320, the frustum-shaped grain uses the generation rule in S330, the spherical-conical grain uses the generation rule in S340, and the near-spherical grain uses the generation rule in S350.

[0087] For different primary dimensions under the same shape, the subordinate dimensions, the radius of the constituent spheres, and the coordinates of the sphere centers are calculated separately, without using a uniform scaling ratio.

[0088] S430, establish a discrete element rigid multi-sphere seed template. Import the number of constituent spheres, their center coordinates, radii, and rigid aggregation relationships into the discrete element simulation platform. Using the local coordinate system as a reference, combine all constituent spheres within the same seed into a single discrete element particle.

[0089] The volume of a single grain is taken as the union volume of all constituent spheres, with overlapping regions calculated only once. The mass of a single grain is determined based on its volume and density, while the position of its center of mass and moment of inertia are calculated based on the mass distribution of the constituent spheres.

[0090] S440, Configure material properties and contact parameters. Write the grain density, Young's modulus, and Poisson's ratio into the grain material properties, and write the Young's modulus and Poisson's ratio of the seed metering device boundary material into the boundary material properties.

[0091] The Hertz-Mindlin no-slip contact model was used for both the inter-kernel contact pairs and the kernel-seed metering device boundary contact pairs. , , , , and Write the corresponding contact parameter matrices. All parameters are converted according to the unit system used by the discrete element simulation platform.

[0092] Moisture content serves as a status indicator for the parameter set. If the moisture content of the target batch exceeds the corresponding range of the parameter set, S200 is re-executed.

[0093] S450, establish a discrete element template library. Organize the main size range, main size probability distribution, subordinate size mapping parameters, and multi-sphere geometry generation rules for dentate, frustum, conical, and quasi-spherical grains according to their morphology, and record the material parameters, contact parameters, and calling weights corresponding to each type of parameterized template.

[0094] The parameterized template data of the four types of grains are stored in the discrete element template library, and a category number is assigned to each template type. When generating a single grain, the grain shape is determined based on the proportion of different shapes, and then the main size is generated based on the probability distribution of the main size of the corresponding shape. The main size is substituted into the mapping function to calculate the subordinate sizes, and the corresponding multi-sphere geometry generation rules are called to calculate the constituent sphere data.

[0095] For different primary dimensions under the same shape, the subordinate dimensions, the radius of the constituent spheres, and the coordinates of the sphere centers are calculated separately, without using a uniform scaling ratio.

[0096] S510, Establish an independent validation sample set. Extract grains from the target batch that were not involved in the S100 parameter fitting and S200 parameter calibration as model validation samples. After completing morphological classification, record the geometric dimensions and weight of each grain.

[0097] S520, conduct geometric consistency verification. Based on the morphology and principal dimensions of the verification sample, generate corresponding discrete element rigid multi-sphere grain templates, and compare the overall width, overall thickness, overall height, single grain volume, and single grain mass of the measured grains with those of the templates. For dentate, frustum-shaped, and conical grains, the principal height also needs to be compared. , width of transition section of embryo tip and the thickness of the transition section of the embryo tip .

[0098] If a single geometric error exceeds a preset limit, return to step S300 to adjust the number of fill layers, the radius of the constituent spheres, and the position of the sphere center. After adjustment, re-execute steps S400 and S520.

[0099] S530: Conduct repose angle verification. Using the same fixed funnel and horizontal receiving plate as the S260 calibration test, conduct actual tests using independent verification samples and establish a discrete element verification model under the same conditions. Obtain the repose angle and stacking height of the actual grain pile and the simulated grain pile respectively, and calculate the error between the two.

[0100] If either the angle of repose error or the stacking height error exceeds the preset limit, return to S260 for recalibration. and After the parameters are adjusted, re-execute S400 and S530.

[0101] S540, Conduct material discharge process verification. An actual material discharge test was conducted on an independent verification sample using a fixed outlet hopper. Discharge time and mass discharged per unit time were recorded. A discrete element verification model with identical structural dimensions, material mass, and material discharge location was established.

[0102] If either the discharge time error or the discharge mass error per unit time exceeds the preset limit, return to S260 for recalibration. and After the parameters are adjusted, re-execute S400 and S540.

[0103] S550, forming the final model parameters. After all errors in geometric consistency verification, angle of repose verification, and material feeding process verification are within preset limits, the multi-sphere geometric parameters, material parameters, and contact parameters are fixed to form the discrete element model parameter set for the target batch of corn kernels.

[0104] S610, organize multi-sphere geometric data. Organize the sphere generation rules, sphere center coordinate calculation relationships, sphere radius calculation relationships, local coordinate directions, rigid aggregation relationships, centroid position calculation results, and rotational inertia calculation results of four types of parameterized templates according to the grain morphology.

[0105] S620, Organize and statistically analyze distribution data. Summarize the morphological quantity proportions, main size probability distributions, main size ranges, and template call weights for the four types of grains in the target batch.

[0106] S630, compile material and contact parameters. Summarize grain moisture content, grain density, grain Young's modulus, grain Poisson's ratio, Young's modulus of the seed metering device boundary material, and Poisson's ratio of the seed metering device boundary material. , , , , and .

[0107] S640, Create a discrete element model data file. Following the data reading format of the CFD-DEM coupling platform, write the multi-sphere geometric data, statistical distribution data, material parameters, and contact parameters into the model data file, and assign category numbers to the toothed, frustum-shaped, spherical, and near-spherical parametric templates respectively.

[0108] S650 outputs discrete element model data. The model data file is transferred to the CFD-DEM coupling platform. This platform determines the grain morphology according to the proportion of morphological quantities, generates the principal dimension according to the principal dimension probability distribution of the corresponding morphology, calculates the subordinate dimensions through a mapping function, and then calls the corresponding multi-sphere geometry generation rules to establish a discrete element rigid multi-sphere grain template.

[0109] The generated corn kernel clusters were used for simulation calculations of the seed supply, negative pressure adsorption, seed carrying, seed cleaning, and seed placement processes of the air-suction precision seed metering device.

[0110] Specific application examples: This embodiment uses Zhengdan 958 maize kernels as the object, statistically analyzing the kernel morphology and geometric dimensions to establish a mapping relationship between primary and secondary dimensions; measuring moisture content, kernel density, Young's modulus, collision recovery coefficient, and static friction coefficient, and determining the rolling friction coefficient through DEM calibration experiments; and establishing rigid multi-sphere kernel templates of dentate, frustum, conical, and quasi-spherical shapes based on the geometric characteristics of various kernel types. This results in a discrete element model of maize kernels that includes the proportion of morphological quantities, size distribution, material parameters, and contact parameters, used for CFD-DEM coupled simulation of a pneumatic high-speed precision seed metering device.

[0111] In this embodiment, Indicates the overall height of the grain. This indicates the main height of the cross section from the top of the grain to the embryo tip; Indicates the maximum width of the main body. Indicates the maximum thickness of the main body; Indicates the width of the transition section at the tip of the embryo. This indicates the thickness of the transition section at the tip of the embryo.

[0112] Grain morphology classification and geometric size measurement. 1000 kernels were randomly selected from Zhengdan 958 maize kernels. Based on the shape of the kernel top, the main cross-section, and the outline of the embryo tip, the samples were divided into dent type, frustum type, conical type, and quasi-spherical type. The number of kernels in each type was counted.

[0113] One hundred kernels from each type were selected and measured individually using electronic vernier calipers. Measurements were taken from dent, frustum, and conical kernels. , , , , and ; Overall width of spherical grains measured Overall height and overall thickness During the measurement process, the grains were kept in the same orientation, and the measurement reference and direction were maintained. The proportion of different grain shapes and their average size are shown in Table 1.

[0114] Table 1. Percentage of kernel morphology and average kernel size in Zhengdan 958 corn variety

[0115] Note: "—" indicates that this dimension is not used for this shape. Dent grains accounted for 73.04%, frustum-shaped grains for 9.69%, conical grains for 8.11%, and quasi-spherical grains for 9.16%. When generating the simulated grain population, the number of each type of grain was determined according to the above proportions.

[0116] Establish a mapping relationship between primary and secondary dimensions. Perform correlation analysis on the geometric dimensions of kernels of the same morphology. For dent kernels, use the embryo tip transition section thickness as an example. As the main size, frustum-shaped and conical kernels are characterized by their main height. As the main size, the spherical kernel is characterized by its overall thickness. As the main size.

[0117] Subordinate dimensions are calculated using the subordinate dimension mapping formula: ; The regression parameters for each type of grain are shown in Table 2.

[0118] Table 2. Mapping parameters of kernel size for Zhengdan 958 maize

[0119] A probability distribution is established based on the measured data of the main dimensions of various types of grains. When generating a single grain, the main dimension is first extracted from the corresponding probability distribution, and then the main dimension is substituted into the subordinate dimension mapping formula to calculate the remaining dimensions of the grain.

[0120] The correlation coefficients of all items in Table 2 are greater than 0.90, and the regression results can be used for parameterized generation of grains of different sizes under the same morphology.

[0121] Basic physical parameters were determined. For moisture content determination, 25g of corn kernels were randomly selected from the target batch and divided into 5 equal portions. The kernels in each portion were crushed and ground into powder, and the moisture content was measured using a MW series halogen moisture analyzer.

[0122] The five moisture contents were 12.2%, 12.7%, 11.9%, 12.5%, and 13.1%, with an average of 12.48% and a standard deviation of 0.41%. This embodiment uses a moisture content of 12.48% as the test state for both physical and contact parameters.

[0123] Grain density was determined using the liquid displacement method. For each test, 20 grains were taken, and the following measurements were taken sequentially: the mass of the empty distillation flask, the mass of the distillation flask filled with pure water, the total mass of the distillation flask and grains, and the total mass after adding grains and filling with pure water. The water temperature, grain quantity, and soaking time were kept consistent across all groups of tests, for a total of 5 independent tests.

[0124] Grain density is calculated using the formula for grain density calculation based on the liquid drainage method: ; The parameters measured by the drainage method are used as the grain density in the discrete element model, denoted as . .

[0125] Young's modulus was determined using a biplane compression test to measure Young's modulus and Poisson's ratio. Individual grains were placed flat in a uniform orientation at the center of the universal compressor test platform, with the loading direction aligned with the grain thickness direction. After the indenter moved to contact the grain surface, it was applied downwards at a speed of 5 mm / min, while simultaneously recording the compressive force, axial displacement, and lateral deformation.

[0126] An elastic deformation segment was selected from the initial contact point to before the first peak force. Poisson's ratio was calculated based on the ratio of lateral strain to axial strain within the elastic deformation segment, and Young's modulus was calculated using the Hertzian contact relationship based on the compressive force and axial deformation. Five grains were selected for repeated experiments, and the average value of the calculated results for each grain was used as the model parameters. The test results of the basic physical parameters are shown in Table 3.

[0127] Table 3. Test results of basic physical parameters of Zhengdan 958 corn kernels

[0128] According to Table 3, the grain density is taken as 1.235 g / cm³. 3 The moisture content was taken as 12.48%, and the Young's modulus was taken as 62.34 MPa. The Poisson's ratio was calculated using the results of a biplane compression test.

[0129] Collision recovery coefficient determination. Collision recovery coefficient between kernels. A simple pendulum test was used for the determination. A collision test sample was composed of corn kernels. Two kernels were fixed to the ends of equal-length nylon threads, ensuring they were at rest, at the same horizontal level, and in contact with each other. One side of the kernels was then lifted and released. A high-speed camera recorded the maximum height reached by the two kernels after the collision, obtaining 10 sets of valid test data.

[0130] Collision recovery coefficient between the grain and the photosensitive resin boundary The free-fall test was used for determination. Individual seeds were adsorbed onto the end of a vacuum conduit, and the vertical distance from the seed release point to the surface of the photosensitive resin plate was recorded. After the vacuum pump was turned off, the seeds fell freely and collided with the photosensitive resin plate. The maximum rebound height was recorded using a high-speed camera. A total of 10 independent tests were completed. The test results are shown in Table 4.

[0131] Table 4. Test results of collision recovery coefficient of Zhengdan 958 maize kernels

[0132] According to Table 4, the collision recovery coefficient between grains... The collision recovery coefficient between the seed and the photosensitive resin boundary is set to 0.5923. The value is 0.5322. The photosensitive resin plate is the boundary material sample of the seed metering device used in this embodiment, so 0.5322 is written into the contact parameter between the seed and the photosensitive resin boundary.

[0133] Static friction coefficient determination. The static friction coefficient between the grain and the photosensitive resin boundary was determined using the inclined plane method. The photosensitive resin plate was fixed on the inclined plane apparatus, and a single grain was placed on the surface of the photosensitive resin plate. The inclination angle of the inclined plane was slowly increased, and the critical inclination angle at which the grain began to slide continuously was recorded.

[0134] To determine the static friction coefficient between kernels, an equal number of corn kernels were adhered to the opposing surfaces of two photosensitive resin plates, ensuring contact between the upper and lower layers of kernels. The inclination angle of the inclined plane was slowly increased, and the critical inclination angle at which the upper plate began to slide continuously was recorded. Both experiments were repeated 10 times. The experimental results are shown in Table 5.

[0135] Table 5. Test results of static friction coefficient of Zhengdan 958 corn kernels

[0136] According to Table 5, the static friction coefficient between grains The coefficient of static friction between the seed and the photosensitive resin boundary is 0.4268. Take 0.6271.

[0137] Rolling friction coefficient calibration. The rolling friction coefficient between grains is denoted as... The coefficient of rolling friction between the seed and the photosensitive resin boundary is denoted as . The two rolling friction coefficients were determined using the DEM calibration method.

[0138] Before calibration, the grain density and Young's modulus from Table 3, the collision recovery coefficient from Table 4, and the static friction coefficient from Table 5 were written into the discrete element model and remained unchanged during the calibration process. and Set the parameters to be calibrated, divide the parameter levels within the set range, and establish parameter combinations.

[0139] Zhengdan 958 corn kernels were discharged into a horizontal receiving plate using a fixed funnel. After the kernel pile stabilized, the angle of repose and the pile height were measured. A discrete element model was established based on the funnel size, funnel height, feeding mass, and receiving plate material used in the actual experiment. Simulations were performed on each parameter combination, and the angle of repose and pile height of the simulated kernel pile were recorded.

[0140] Discharge tests were conducted using a fixed outlet hopper, and the discharge time and mass discharged per unit time were recorded. A discrete element calibration model was established based on the actual hopper structure, outlet size, feed mass, and photosensitive resin boundary material, and discharge simulations were carried out using the same parameter combinations.

[0141] The actual test results were compared with the simulation results, and parameter combinations in which the errors of the angle of repose, stacking height, discharge time, and discharge mass per unit time were all within preset limits were retained. Based on the selected parameter combinations, the parameter intervals were narrowed, and new parameter combinations were established and calibrated. The final parameters were determined based on the refined calibration results. and And write the two parameters into the corresponding contact parameters.

[0142] Construction of a rigid multi-spherical seed template. (See attached document.) Figure 7 During modeling, the center of the tangent plane at the top of the grain is used as the local coordinate origin. An x-axis is established along the grain width, a y-axis along the grain height, and a z-axis along the grain thickness. The positive direction of the y-axis points from the top of the grain to the tip of the embryo. The constituent spheres are arranged according to the outer contour of the grain, with adjacent constituent spheres overlapping. Constituent spheres within the same grain are combined into a discrete element particle according to a rigid aggregation relationship.

[0143] For dent kernels, five layers of main constituent spheres are arranged along the kernel height, with an embryo tip sphere placed at the embryo tip position. The cross-sectional width of each layer in the main constituent area is determined by... Towards Transition, cross-sectional thickness from Towards Transition. Calculate the radius and center position of the constituent spheres based on the cross-sectional dimensions of each layer, so that the outer envelope of the multi-sphere model corresponds to the outline of the dentate grain.

[0144] For frustum-shaped kernels, three layers of constituent spheres are arranged along the height of the kernel. The corner constituent spheres are arranged in each layer according to the cross-sectional width and thickness at the corresponding position. The outer envelope of the constituent spheres shrinks from the maximum cross-section of the main body to the transition cross-section of the embryo tip, and an embryo tip sphere is placed at the embryo tip position.

[0145] For conical kernels, three layers of constituent spheres are arranged along the kernel height, with three constituent spheres of equal radius in each layer. The three constituent spheres are arranged along the width direction, with the middle constituent sphere located at the local coordinate center and the two symmetrically arranged constituent spheres on both sides. The width and thickness of each layer gradually change from the main area to the embryo tip transition section, and an embryo tip sphere is placed at the embryo tip position.

[0146] For spherical grains, a multi-sphere model is constructed using a central sphere and six axial auxiliary spheres. The central sphere is located at the center of the 3D bounding box, and the six axial auxiliary spheres are arranged in the positive and negative directions of the x, y, and z axes, respectively. The model is based on the overall width. Overall height and overall thickness Determine the center position of each auxiliary sphere so that the three-dimensional envelope dimensions of the multi-sphere model are 8.07 mm, 9.07 mm and 7.38 mm, respectively.

[0147] Different sizes of seeds of the same morphology are not scaled using a uniform ratio. When generating a single seed, the main size is first generated based on the probability distribution of the main size of the corresponding morphology, and then the subordinate size mapping formula is used to calculate the remaining geometric dimensions. Finally, the radius and center coordinates of the constituent sphere are calculated.

[0148] After arranging the constituent spheres, all constituent spheres within the same grain are combined into a single rigid particle. No internal contact calculations are performed between the constituent spheres. The volume of a single grain is determined by the union volume of all constituent spheres; overlapping areas of the constituent spheres are calculated only once; and the mass of a single grain is calculated based on the union volume and the grain density.

[0149] Material parameters, contact parameters, and discrete element template library configuration. After completing the multi-sphere geometry generation rules for the four types of seeds, the material parameters and contact parameters are written into the discrete element model.

[0150] The grain density is taken as 1.235 g / cm³. 3 The moisture content was taken as 12.48%, the Young's modulus as 62.34 MPa, and the Poisson's ratio was calculated using the results of a biplane compression test.

[0151] Collision recovery coefficient between grains The collision recovery coefficient between the seed and the photosensitive resin boundary is set to 0.5923. Take 0.5322; coefficient of static friction between grains The coefficient of static friction between the seed and the photosensitive resin boundary is 0.4268. Take 0.6271; coefficient of rolling friction between grains The rolling friction coefficient between the grain and the photosensitive resin boundary The calibration results were used.

[0152] All parameters were converted according to the unit system of the discrete element simulation platform and written into the seed material properties, photosensitive resin material properties and corresponding contact parameters respectively.

[0153] The discrete element template library stores the main size range, main size probability distribution, subordinate size mapping parameters, multi-sphere geometric generation rules, material parameters, and contact parameters of dentate, frustum, spherical, and quasi-spherical grains according to their grain morphology.

[0154] The weighting of the four types of grains is based on the proportion of the number of morphologies shown in Table 1. The weighting of dent grains is 73.04%, truncated pyramidal grains is 9.69%, conical grains is 8.11%, and quasi-spherical grains is 9.16%.

[0155] Simulated grain cluster generation. When generating the simulated grain cluster, the shape of a single grain is first determined according to the calling weights of the four grain types, and then the main size is generated from the main size probability distribution of the corresponding shape.

[0156] Substitute the primary dimension into the corresponding subordinate dimension mapping function to calculate all subordinate dimensions of the grain. After completing the dimension constraint check, call the multi-sphere geometry generation rule for the corresponding shape to calculate the number of constituent spheres, the radius of the constituent spheres, and the coordinates of the sphere centers, and write the material parameters and contact parameters into the grain model.

[0157] Repeat the process of determining grain morphology, generating primary size, calculating secondary size, generating multi-sphere data, and writing parameters until the number of grains reaches the simulation set value.

[0158] Model checking, validation, and data output. After completing the grain model construction, check the overall width, overall height, overall thickness, constituent sphere radius, overlap of adjacent constituent spheres, and single grain volume for each type of model. For dent, frustum, and conical grains, also check the main body height, embryo tip transition section width, and embryo tip transition section thickness.

[0159] If the radius of the constituent spheres is not greater than zero, recalculate the cross-sectional dimensions and radii of the corresponding layers. If adjacent constituent spheres do not overlap, adjust the spacing between their centers. If the three-dimensional envelope error of the model exceeds a preset limit, adjust the number of constituent spheres and their center positions, and regenerate the corresponding model.

[0160] From the target batch, select additional kernels that were not involved in size fitting and parameter calibration as validation samples. Geometric consistency validation compares the validation samples with the kernel model in terms of overall width, overall height, overall thickness, single kernel volume, and single kernel mass. For dent, frustum, and conical kernels, simultaneously compare the body height, embryo tip transition section width, and embryo tip transition section thickness.

[0161] Independent samples were used to verify the angle of repose, comparing the angle of repose and stacking height of actual and simulated grain piles. A fixed outlet hopper was used to verify the discharge process, comparing the discharge time and discharge mass per unit time of actual and simulated experiments.

[0162] If the geometric consistency verification error exceeds the preset limit, adjust the radius and center position of the constituent spheres of the corresponding shape. If the repose angle verification and material feeding process verification errors exceed the preset limits, recalibrate. and After adjusting the parameters, rebuild the model and verify it until all errors are within the preset limits.

[0163] After verification, the multi-sphere geometric parameters, material parameters, and contact parameters of the four types of kernels were fixed to form a discrete element model of Zhengdan 958 corn kernels.

[0164] The output data includes the multi-sphere geometric generation rules of four types of grains, the number of constituent spheres, the radius of the constituent spheres, the coordinates of the sphere center, the direction of the local coordinates, the rigid aggregation relationship, the range of the main size, the probability distribution of the main size, the mapping parameters of the subordinate size, the proportion of the number of morphologies, the material parameters, and the contact parameters.

[0165] See attached document Figure 7 When constructing a multi-sphere model of corn kernels, a local coordinate system is first established with the center of the top tangent plane of the kernel as the origin, where the x-axis, y-axis, and z-axis represent the width, height, and thickness of the kernel, respectively. Regarding macroscopic geometric dimensions... Indicates the overall height of the grain. Indicates the height of the main body; and These represent the maximum width and maximum thickness of the main area, respectively. and This represents the width and thickness of the transition section at the tip of the embryo. Regarding the multi-sphere geometry construction parameters, the model divides the main body into multiple layers along the y-axis for a layer-by-layer transition, where... and Representing the first The cross-sectional width and thickness of the layer and Represents the width and thickness of the final layer cross section; Indicates the first The radius of the constituent spheres inside the layer. This represents the distance between the centers of the spheres of adjacent main layers, while This refers to the radius of the germinal tip sphere located at the bottom of the model. Based on the above geometric dimension definitions and multi-sphere generation rules, dent, frustum, quasi-spherical, and conical maize kernel DEM models are finally obtained. The generated kernel models are used for CFD-DEM coupled simulation of the seed supply, negative pressure adsorption, seed carrying, seed cleaning, and seed placement processes of the air-suction high-speed precision seed metering device. The detailed dimensional parameter definition process is as follows: Figure 7 As shown in (a), the four types of grain DEM models generated in the end are as follows: Figure 7 As shown in (b).

Claims

1. A CFD-DEM coupled simulation modeling method for maize kernels using an air-suction precision seed metering device, characterized in that, Includes the following steps: Obtain the morphological quantity ratio, main size, and subordinate size mapping function of corn kernels in the target batch relative to the main size, and determine the main size probability distribution of the main size; Obtain the physical parameters and contact parameters of the target batch of corn kernels, and calibrate the rolling friction coefficient in the contact parameters; Establish a local coordinate system for the multi-sphere model, calculate the multi-sphere geometric data of the target batch of corn kernels, and perform a validity check; Based on the proportion of the number of morphologies, the probability distribution of the main size, the mapping function, and the multi-sphere geometric data, construct and call the corresponding multi-sphere geometric generation rules, and combine the physical parameters and the contact parameters to establish a discrete element rigid multi-sphere seed template and a discrete element template library; The discrete element rigid multi-sphere seed template is validated. When the error exceeds a preset limit, feedback correction is performed until the error is within the preset limit. The validated data is organized into a discrete element model data file, which is then output to the CFD-DEM coupling platform to generate a corn kernel cluster for simulation calculations of the air-suction precision seed metering device.

2. The CFD-DEM coupled simulation modeling method for maize kernels using a pneumatic precision seed metering device according to claim 1, characterized in that, The process of obtaining the mapping function of the morphological quantity ratio, main size, and subordinate size of the target batch of corn kernels relative to the main size specifically includes: The target batch of corn kernels is divided into dent type, frustum type, conical type and quasi-spherical type according to its external contour, and the proportion of each type in the target batch of corn kernels is calculated. The overall height, main body height, maximum main body width, maximum main body thickness, and the width and thickness of the transition section at the tip of the embryo are measured as candidate dimensions. Calculate the correlation coefficient between each candidate size and the remaining sizes, and determine the size with the largest average correlation coefficient as the main size of the corresponding shape, and the remaining candidate sizes as the subordinate sizes; By performing regression fitting using measured sample data, the function parameters of the mapping function of the subordinate dimension relative to the master dimension are obtained.

3. The CFD-DEM coupled simulation modeling method for maize kernels using a pneumatic precision seed metering device according to claim 1, characterized in that, The physical parameters include moisture content, grain density, Young's modulus, and Poisson's ratio; the contact parameters include the collision recovery coefficient, the static friction coefficient, and the rolling friction coefficient. The calibration of the rolling friction coefficient in the contact parameters specifically includes: A fixed funnel discharge method was used to measure the angle of repose and the height of the pile. A fixed outlet hopper was used to conduct a discharge test and record the discharge time and the mass discharged per unit time. Establish a discrete element calibration model and input the obtained physical parameters, the collision recovery coefficient, and the static friction coefficient; Establish a parameter combination table for the rolling friction coefficient, conduct repose angle simulation and material discharge simulation for each group of parameters, and determine the final calibrated rolling friction coefficient based on the error.

4. The CFD-DEM coupled simulation modeling method for maize kernels using a pneumatic precision seed metering device according to claim 1, characterized in that, The process of establishing a local coordinate system for the multi-sphere model, calculating the multi-sphere geometric data of the target batch of corn kernels, and performing a validity check specifically includes: Set the center of the top cutting plane of the grain as the local coordinate origin, and establish the local coordinate system of the multi-sphere model with the grain width direction as the x-axis, the grain height direction as the y-axis, and the grain thickness direction as the z-axis. The positive direction of the y-axis points from the top of the grain to the embryo tip. The number of y-axis filling layers is determined based on the target contour dimensions, geometric dimension error limits, and the number of constituent spheres. Based on different grain shapes and the number of filling layers, the radius of each constituent sphere, the coordinates of its center, and the rigid aggregation relationship between each constituent sphere within the same grain are calculated and combined to generate the geometric data of the multi-spheres of various shapes.

5. The CFD-DEM coupled simulation modeling method for maize kernels using a pneumatic precision seed metering device according to claim 4, characterized in that, The multi-sphere geometric data is generated in the following manner: For the tooth-shaped and frustum-shaped types, the main body area is divided into multiple layers along the y-axis. The width and thickness of each layer gradually transition from the maximum size of the main body to the transition section size of the tip of the embryo. Based on this, the radius of the constituent sphere and the coordinates of the sphere center are calculated. For a spherical cone shape, multiple layers are arranged along the y-axis, and each layer consists of three spheres of equal radius arranged along the x-axis; For spherical shapes, a central sphere is set at the center of the three-dimensional bounding box, and six axial auxiliary spheres with the same radius as the central sphere are arranged in the positive and negative directions of the x-axis, the positive and negative directions of the y-axis, and the positive and negative directions of the z-axis, respectively.

6. The CFD-DEM coupled simulation modeling method for maize kernels using a pneumatic precision seed metering device according to claim 4, characterized in that, The validity check specifically includes: Examine the three-dimensional envelope size, constituent sphere radius, overlap state of adjacent constituent spheres, and model volume of the four types of grain morphology, and examine the position, width, and thickness of the embryo tip transition section of the dentate, frustum, and spherical cone types; When a non-positive constituent sphere radius is detected, it is recalculated. When adjacent constituent spheres do not overlap, the distance between their centers is adjusted. When the three-dimensional envelope size error exceeds the geometric size error limit or the model volume error exceeds the volume error limit, the number of filling layers, the center position of the spheres, or the constituent sphere radius is adjusted.

7. The CFD-DEM coupled simulation modeling method for maize kernels using a pneumatic precision seed metering device according to claim 1, characterized in that, The step of constructing and calling corresponding multi-sphere geometry generation rules based on the proportion of the number of morphologies, the probability distribution of the main size, the mapping function, and the multi-sphere geometric data, and establishing a discrete element rigid multi-sphere seed template and a discrete element template library in combination with the physical parameters and the contact parameters, specifically includes: The seed morphology is determined based on the proportion of the morphological quantities, the main size is generated based on the probability distribution of the main size of the corresponding morphology, and the main size is substituted into the mapping function to calculate all subordinate sizes; The corresponding morphology of the multi-sphere geometry generation rules is called to calculate the number of constituent spheres, the radius of the constituent spheres, and the coordinates of the center of the spheres. The discrete element rigid multi-sphere seed template is then established in the discrete element simulation platform according to the rigid aggregation relationship. The main size range of various types of grains, the probability distribution of the main size, the mapping function, the multi-sphere geometry generation rule, the physical parameters, the contact parameters, and the template call weight are stored in the discrete element template library.

8. The CFD-DEM coupled simulation modeling method for maize kernels using a pneumatic precision seed metering device according to claim 3, characterized in that, The model verification includes geometric consistency verification, angle of repose verification, and material layout process verification. The geometric consistency verification compares the measured grain size with the discrete element rigid multi-sphere grain template in terms of overall width, overall thickness, overall height, single grain volume, and single grain mass. When the error of the geometric consistency verification exceeds the preset limit, the number of filling layers, the radius of the constituent spheres, and the position of the sphere center are readjusted. The angle of repose verification compares the angle of repose and stacking height of the actual grain pile and the simulated grain pile; the discharge process verification compares the discharge time and discharge mass per unit time of the actual test and the simulation test. When the error of the repose angle verification or the material discharge process verification exceeds the preset limit, the rolling friction coefficient in the contact parameters is recalibrated.

9. The CFD-DEM coupled simulation modeling method for maize kernels using a pneumatic precision seed metering device according to claim 1, characterized in that, The output content of the discrete element model data file includes: The composition of various grain templates includes the number of constituent spheres, sphere center coordinates, constituent sphere radius, rigid aggregation relationship, local coordinate direction, single grain mass, single grain volume, the proportion of the aforementioned morphological quantities, the probability distribution of the aforementioned main dimensions, grain material parameters, seed metering device boundary material parameters, and the aforementioned contact parameters. After reading the discrete element model data file, the CFD-DEM coupling platform determines the grain morphology according to the proportion of the morphological quantity, generates the main size according to the probability distribution of the main size of the corresponding morphology, calculates the subordinate size through the mapping function, and calls the corresponding multi-sphere geometry generation rule to establish the discrete element rigid multi-sphere grain template. The generated corn grain group is used for simulation calculation.

10. The CFD-DEM coupled simulation modeling method for maize kernels using a pneumatic precision seed metering device according to claim 7, characterized in that, When establishing the discrete element rigid multi-sphere seed template: The volume of a single grain is the union volume of all constituent spheres within the same grain, and overlapping regions are calculated only once; The mass of a single grain is determined based on the volume of the single grain and the grain density in the physical parameters. The position of the center of mass and the moment of inertia are calculated based on the mass distribution of the union of the constituent spheres.

Images