Quantitative evaluation method and system for grain impact damage during corn machine harvesting and threshing

Abstract

The application discloses a kind of corn machine harvest threshing when kernel collision damage quantitative evaluation method and system, it is related to the field of agricultural engineering and granular mechanics technique.First, build the kernel damage quantitative evaluation model of fusion apparent-intrinsic characteristics, capture the real damaged state of kernel.Second, select multiple groups of corn kernel samples, carry out collision test under different moisture content, collision intensity and collision frequency and other conditions, obtain the cumulative plastic dissipation energy of each stage, apparent-intrinsic damage characteristic parameter, based on the model built the damage probability under each condition, provide data support for kernel collision damage evolution modeling.Combining collision condition parameters, cumulative plastic dissipation energy and damage probability, establish the kernel collision damage evolution constitutive equation under wide-range moisture content, reflect the single kernel collision damage evolution law.Finally, kernel collision damage evolution constitutive equation is embedded into discrete element simulation software, realize the quantitative analysis of single kernel collision damage in corn threshing link.

Description

Technical Field

[0001] This invention relates to the fields of agricultural engineering and particle mechanics, and in particular to a method and system for quantitatively evaluating kernel collision damage during corn harvesting and threshing. Background Technology

[0002] During the mechanized harvesting and threshing of corn, the high-frequency collisions between kernels and threshing elements, as well as between kernels and kernels, and between kernels and ears, are the main causes of kernel damage and quality decline.

[0003] Existing methods for assessing grain damage are mostly limited to macroscopic-level statistical analysis of visible damage or to average damage rate prediction models based on particle swarm scale. For example, Chinese patent application CN202511656674.1 discloses a "mechanism- and data-driven method for predicting the evolution of grain damage rate during wheat and corn harvesting and threshing," which includes obtaining response data on the change of wheat or corn grain damage rate over time under different threshing conditions based on bench tests; constructing a physical model of grain damage rate during wheat and corn threshing based on discrete element simulation analysis results to obtain response data on the change of grain damage rate over time under different threshing conditions; constructing a training dataset for the prediction model of grain damage rate during wheat and corn harvesting and threshing; and establishing a mechanism- and data-driven method for predicting the evolution of grain damage rate during wheat and corn harvesting and threshing. Essentially, it establishes a predictive model for the evolution of grain damage rate under different threshing conditions (drum speed, concave plate gap, and feed rate).

[0004] However, the accuracy of single-kernel damage probability assessment directly affects the quantitative analysis results of corn kernel damage during threshing. The single-kernel damage probability model of the aforementioned patent ignores the cumulative effect of collision damage or the latent damage of kernels. In particular, when obtaining kernel damage data through threshing simulation, it cannot directly provide quantitative analysis results of single-kernel collision damage. Summary of the Invention

[0005] Based on this, the present invention provides a quantitative evaluation method for kernel collision damage during corn harvesting and threshing, which comprehensively considers the cumulative effect of collision damage or the latent damage of kernels, and provides quantitative analysis results of single kernel collision damage.

[0006] This invention provides a method for quantitatively evaluating kernel collision damage during corn harvesting and threshing, comprising:

[0007] Obtain multiple samples of corn kernels;

[0008] The sample corn kernels were subjected to a collision test based on preset collision condition parameters. The cumulative plastic dissipation energy, maximum normal contact force and characteristic parameters were recorded at each collision frequency. The collision condition parameters included the collision incident velocity, the number of collisions and the kernel moisture content. The characteristic parameters included the overt damage characteristic parameters and the latent damage characteristic parameters.

[0009] Based on the aforementioned feature parameters, a pre-established single-kernel collision damage quantification evaluation method is used to obtain the damage probability of the sample corn kernels under different collision conditions.

[0010] The constitutive equation for the evolution of grain collision damage over a wide range of moisture content is constructed by configuring the grain moisture content, the maximum normal contact force, the collision incident velocity, the number of collisions, and the cumulative plastic dissipation energy as input variables and the damage probability as the objective function.

[0011] The constitutive equation is used to quantitatively evaluate the kernel collision damage during the corn threshing process.

[0012] In some embodiments, the feature parameters are obtained based on the following steps:

[0013] Multi-collision energy level impact tests were conducted on the sample corn kernels with different moisture contents to separate the obviously damaged kernels from the sample corn kernels; the obviously damaged kernels included the sample corn kernels that were broken and had apparent cracks.

[0014] Extract the dominant damage characteristic parameters of the dominantly damaged grains;

[0015] In the remaining seemingly intact corn kernels, kernels with hidden damage were separated based on a high-intensity cold light source transmission test.

[0016] Extract the latent damage characteristic parameters of the latently damaged grains.

[0017] In some embodiments, the step of obtaining the damage probability of the sample corn kernels under different collision conditions based on the feature parameters and using a pre-established single-kernel collision damage quantification evaluation method includes:

[0018] Obtain the sensitive feature parameter from among the multiple feature parameters;

[0019] Obtain the standard deviation and correlation coefficient between the aforementioned sensitive feature parameters;

[0020] Based on the standard deviation and the correlation coefficient, obtain the optimal weight vector of the feature parameter;

[0021] Based on the explicit damage feature parameters and the optimal weight vector, the explicit damage failure probability is obtained;

[0022] Based on the latent damage feature parameters and the optimal weight vector, the cumulative probability of latent damage is obtained;

[0023] The damage probability is obtained based on the failure probability of the explicit damage and the cumulative probability of the implicit damage.

[0024] In some embodiments, the sensitive feature parameters are obtained based on the following steps:

[0025] Obtain the nonlinear correlation between each of the aforementioned characteristic parameters and the threshing energy level;

[0026] Based on the aforementioned nonlinear correlation, the feature parameters are divided into low-sensitivity feature parameters and sensitive feature parameters;

[0027] Remove the aforementioned low-sensitivity feature parameters;

[0028] The sensitive feature parameters are retained.

[0029] In some embodiments, the probability of overt damage failure is obtained based on the following steps:

[0030] Obtain the normalized eigenvalues ​​of the overt damage feature parameters;

[0031] An exponential penalty activation function is introduced, and the probability of failure of the explicit damage is obtained based on the normalized feature value and the optimal weight vector.

[0032] In some embodiments, the explicit damage feature parameters are configured to include crack length, crack aperture, grain breakage rate, and two-dimensional fractal dimension of the crack, and the implicit damage feature parameters are configured to include three-dimensional fractal dimension, micropore volume fraction, and damage volume fraction.

[0033] In some embodiments, the constitutive equation is constructed based on the following steps:

[0034] A penalized fitness function is constructed based on multiple physical constraints; the multiple physical constraints are configured to include model constitutive morphology constraints, physical dimension constraints, and evolutionary mechanism constraints.

[0035] The collision condition parameters, the cumulative plastic dissipation energy, and the maximum normal contact force are configured as input variables, the damage probability is configured as the objective function, and the constitutive equation is constructed based on the penalized fitness function.

[0036] In some embodiments, the constitutive morphological constraints of the model include limiting the polynomial and exponential functions to be basis functions that fit the nonlinear evolution characteristics of material damage; the physical dimensional constraints include a combination of dimensional consistency constraints and redundant feature elimination; the evolution mechanism constraints include embedding an extreme value penalty term for irreversible damage and ensuring that the constitutive equation satisfies a preset interval.

[0037] In some embodiments, the method further includes:

[0038] The constitutive equation is embedded in discrete element simulation analysis software to achieve real-time quantitative analysis of the collision damage of kernels during corn threshing.

[0039] Accordingly, the present invention also provides a quantitative evaluation system for kernel collision damage during corn harvesting and threshing, used to execute the quantitative evaluation method for kernel collision damage during corn harvesting and threshing described in any one of the above-mentioned methods, including:

[0040] The first acquisition module is configured to acquire multiple samples of corn kernels;

[0041] The second acquisition module is configured to conduct a collision test on the sample corn kernels based on preset collision condition parameters, and record the cumulative plastic dissipation energy, maximum normal contact force and characteristic parameters at each collision frequency; the collision condition parameters include the collision incident velocity, the number of collisions and the kernel moisture content, and the characteristic parameters include the overt damage characteristic parameters and the latent damage characteristic parameters.

[0042] The third acquisition module is configured to acquire the damage probability of the sample corn kernels under different collision conditions based on the feature parameters and using a pre-established single kernel collision damage quantification evaluation method.

[0043] The constitutive equation construction module is configured to take the grain moisture content, the maximum normal contact force, the impact incident velocity, the number of impacts and the cumulative plastic dissipation energy as input variables, and the damage probability as the objective function to construct a constitutive equation for the evolution of grain impact damage over a wide range of moisture content.

[0044] The evaluation module is configured to quantitatively evaluate kernel collision damage during corn threshing based on the constitutive equation.

[0045] Compared with the prior art, the technical solution provided by this invention has the following advantages:

[0046] The present invention provides a method for quantitatively evaluating kernel collision damage during corn harvesting and threshing. First, a quantitative evaluation model for kernel collision damage integrating explicit and implicit characteristics is constructed to simultaneously consider both surface and internal implicit damage, accurately capturing the true damage state of the kernels. Second, multiple sets of corn kernel samples are selected, and kernel collision tests are conducted under various working conditions including different moisture contents, collision intensities, and collision times. Cumulative plastic dissipation energy, explicit damage characteristic parameters, and implicit damage characteristic parameters are collected at each stage. Based on the aforementioned quantitative evaluation model for kernel damage, the damage probability of corn kernels under each experimental condition is obtained, providing basic data for subsequent kernel collision damage evolution modeling. Combining collision condition parameters, cumulative plastic dissipation energy, and damage probability, a constitutive equation for kernel collision damage evolution under a wide range of moisture contents is constructed to reflect the evolution law of single-kernel collision damage. Finally, the kernel collision damage evolution constitutive equation is embedded into discrete element simulation software using an API interface to achieve quantitative analysis of single-kernel collision damage during corn threshing, thereby completing an overall evaluation of kernel collision damage during the corn threshing process.

[0047] It is understood that, compared with the prior art, the quantitative evaluation system for kernel collision damage during corn harvesting and threshing provided in this embodiment of the invention includes all the technical features and technical effects of the quantitative evaluation method for kernel collision damage during corn harvesting and threshing described above, which will not be repeated here. Attached Figure Description

[0048] Figure 1 This is a flowchart illustrating a method for quantitatively evaluating kernel collision damage during corn harvesting and threshing provided by the present invention.

[0049] Figure 2 This is another flowchart illustrating the quantitative evaluation method for kernel collision damage during corn harvesting and threshing provided by the present invention. Detailed Implementation

[0050] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to provide a thorough understanding of the present invention. However, the present invention can be practiced in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0051] This invention provides a method for quantitatively evaluating kernel collision damage during corn harvesting and threshing. (See attached document.) Figure 1 , Figure 1 This is a flowchart illustrating a method for quantitatively evaluating kernel collision damage during corn harvesting and threshing provided by the present invention.

[0052] This invention provides a method for quantitatively evaluating kernel collision damage during corn harvesting and threshing, comprising the following steps S100 to S500.

[0053] Step S100: Obtain multiple samples of corn kernels.

[0054] Specifically, the sample corn kernels should be selected from representative varieties, and the sample corn kernels should be consistent in terms of geometric size, maturity, and initial moisture content to eliminate the interference of individual differences on the results of subsequent collision tests. These sample corn kernels will serve as the physical carrier for establishing damage assessment benchmarks, and the sample size should meet statistical requirements to ensure the robustness of the subsequent model construction.

[0055] Step S200: Based on the preset collision condition parameters, conduct a collision test on the sample corn kernels and record the cumulative plastic dissipation energy, maximum normal contact force and characteristic parameters at each collision frequency; the collision condition parameters include the collision incident velocity, the number of collisions and the kernel moisture content, and the characteristic parameters include the overt damage characteristic parameters and the latent damage characteristic parameters.

[0056] In this embodiment, the impact test can be conducted on a pneumatic single-kernel impact test bench or a similar device, simulating the mechanical environment of the actual threshing process by precisely controlling the impact conditions. The kernel moisture content refers to a wide-range moisture content, covering the typical operating range of mechanized corn harvesting and threshing, such as 24% to 32%, rather than a single fixed value. This wide-range setting aims to capture the nonlinear effects of moisture content changes on the viscoelasticity and damage sensitivity of the sample corn kernels, ensuring the applicability of the evaluation method at different harvest stages.

[0057] The incident velocity of the impact determines the kinetic energy input level at the moment of impact, while the maximum normal contact force reflects the peak stress response of the sample corn kernels during the impact. During the experiment, data was simultaneously collected using a high-speed camera and a high-frequency force sensor to calculate the cumulative plastic dissipation energy after the nth impact. Cumulative plastic dissipation energy is a key intermediate variable characterizing the nature of damage driving; its physical meaning is the total energy consumed by irreversible mechanisms such as plastic deformation and microcrack propagation during multiple impacts, and it can usually be obtained by integrating the kinetic energy difference before and after the impact.

[0058] Alternatively, the cumulative plastic dissipation energy can be obtained using the following formula:

[0059] (1)

[0060] in, To accumulate plastic dissipation energy, For the quality of the corn kernels in the sample, Let the collision incident velocity be... This refers to the rebound speed.

[0061] In some embodiments, the feature parameters are obtained based on the following steps: conducting a multi-collision energy level impact test based on a wide-range moisture content to separate overtly damaged kernels from the sample corn kernels; overtly damaged kernels include broken and apparent cracked sample corn kernels; extracting overt damage feature parameters from the overtly damaged kernels; and separating latently damaged kernels from the remaining seemingly intact sample corn kernels based on a high-intensity cold light source transmission test; extracting latent damage feature parameters from the latently damaged kernels.

[0062] Specifically, this embodiment uses a multi-collision energy level alternating impact test to simulate the complex load spectrum that sample corn kernels are subjected to during the threshing process, so that sample corn kernels with different degrees of damage can be fully displayed, thereby pre-treating the sample corn kernels.

[0063] First, visible damaged grains that have undergone macroscopic breakage or surface cracking are separated using visual recognition or machine vision algorithms. Since the structural integrity of visible damaged grains has been compromised, physically separating them first ensures the accuracy of visible feature extraction and avoids the mixing of visible and latent damage signals, thus purifying the sample background for subsequent accurate identification of latent damage.

[0064] Subsequently, non-destructive testing was conducted on the remaining seemingly intact corn kernels using a high-intensity cold light source transmission test. The use of a cold light source avoided secondary thermal damage to moisture-sensitive biological tissues caused by thermal radiation, ensuring the stability of the corn kernel sample's state during the testing process. Under a high-intensity transmitted light field, latent defects such as microcracks and micropores with uneven density inside the corn kernel sample would form specific grayscale textures due to the differences in the attenuation coefficients of X-rays or visible light, thereby achieving effective separation and marking of kernels with latent damage.

[0065] In some embodiments, the explicit damage characteristic parameters are configured to include crack length, crack aperture, grain breakage rate, and crack two-dimensional fractal dimension, while the implicit damage characteristic parameters are configured to include three-dimensional spatial fractal dimension, micropore volume fraction, and damage volume fraction.

[0066] Specifically, the explicit damage characteristic parameters focus on describing the geometric failure morphology of the sample corn kernel surface. Among them, crack length and crack aperture characterize the scale of macroscopic crack propagation; kernel breakage rate characterizes the degree of loss of overall structural integrity. It is particularly noteworthy that the collision cracks on the sample corn kernel surface often exhibit complex branching and meandering morphologies. The two-dimensional fractal dimension of the crack can quantitatively characterize the spatial filling capacity and complexity of this surface crack topology. The higher the value, the more developed the crack network and the more complex the energy dissipation mode on the surface.

[0067] In contrast, latent damage characteristic parameters focus on the degradation and evolution of the mesoscopic structure within the sample corn kernels. The three-dimensional fractal dimension measures the three-dimensional connectivity and roughness of the internal microcrack network, i.e., the penetration depth and uniformity of damage distribution in three-dimensional space. The micropore volume fraction reflects the proportion of micropores generated by stress concentration within the endosperm tissue, indicating a decrease in material density. The damage volume fraction integrates the total volume effect of microcracks and micropores, characterizing the degree of weakening of the effective load-bearing cross-section.

[0068] Through the synergistic effect of the aforementioned explicit and implicit feature parameters, this embodiment achieves a holographic digital description of collision damage in sample corn kernels. It should be understood that although this embodiment lists the specific parameters mentioned above, in other embodiments, some parameter combinations can be selected, or other physical quantities capable of equivalently characterizing macroscopic and mesoscopic damage features can be introduced, depending on the specific evaluation accuracy requirements, as long as they can cover cross-scale information from surface geometric failure to internal structural degradation.

[0069] Step S300: Based on the feature parameters, use the pre-established single-kernel collision damage quantification evaluation method to obtain the damage probability of sample corn kernels under different collision conditions.

[0070] Specifically, by weighted fusion or specific mathematical mapping of explicit and implicit damage feature parameters, the damage probability of each sample corn kernel is calculated to reflect the current damage state of the sample corn kernel, thus providing a standardized target variable for the subsequent construction of constitutive equations.

[0071] Please see Figure 2 , Figure 2 This is another flowchart illustrating the quantitative evaluation method for kernel collision damage during corn harvesting and threshing provided by the present invention. It should be noted that the single-kernel collision damage quantitative evaluation method in this embodiment is pre-established using formulas (3), (4), and (2), which can be understood as pre-establishing a kernel damage quantitative evaluation model that integrates explicit and implicit features. When obtaining the damage probability D, it can be obtained directly from formula (2), or sequentially through formulas (3), (4), and (2). When a quantitative evaluation method for single-kernel collision damage is established in advance, the quantitative evaluation method for kernel collision damage during corn harvesting and threshing provided in this embodiment of the invention can also be understood as including steps S1 to S5: S1, constructing a kernel damage quantitative evaluation model that integrates explicit and implicit features; S2, obtaining sample corn kernels with different moisture content gradients and conducting kernel collision tests under different collision intensities and collision times; S3, recording the collision incident velocity, collision intensity, cumulative plastic dissipation energy and corresponding damage probability of sample corn kernels under different collision frequencies; S4, constructing a constitutive equation for the evolution of kernel collision damage over a wide range of moisture content; S5, based on the AIP interface and in conjunction with discrete element simulation software, constructing a quantitative analysis system for single-kernel collision damage in the corn threshing process, and executing the quantitative evaluation method for kernel collision damage during corn harvesting and threshing through this analysis system. The principles of steps S1 to S5 are the same as those of steps S100 to S500, which are obtained sequentially through formula (3), formula (4) and formula (2).

[0072] In some embodiments, step S300 includes: obtaining sensitive feature parameters among multiple feature parameters; obtaining the standard deviation and correlation coefficient among sensitive feature parameters; obtaining the optimal weight vector of the feature parameters based on the standard deviation and correlation coefficient; obtaining the failure probability of overt damage based on the overt damage feature parameters and the optimal weight vector; obtaining the cumulative probability of latent damage based on the latent damage feature parameters and the optimal weight vector; and obtaining the damage probability based on the failure probability of overt damage and the cumulative probability of latent damage.

[0073] Specifically, this embodiment employs the CRITIC (Criteria Importance Through Intercriteria Correlation) weighting method to determine the optimal weight vector. Based on the statistical characteristics of the data itself, the CRITIC method offers stronger objectivity and reproducibility. Two dimensions are considered in obtaining the optimal weight vector: first, the standard deviation of the feature parameter, i.e., the degree of variation of the feature parameter among different damage samples; a larger standard deviation indicates a stronger ability of the feature to distinguish damage levels; second, the correlation coefficient between feature parameters, i.e., the degree of information overlap between feature parameters. If a feature parameter is highly positively correlated with other feature parameters, it contains less unique information and should be assigned a lower weight. By combining these two dimensions, the CRITIC method can adaptively decouple collinearity interference between feature parameters, generating a set of optimal weight vectors that reflect both individual discrimination and overall information independence. Finally, based on the determined weight vector, the failure probability of explicit damage and the probability of implicit damage are calculated and fused to obtain the damage probability.

[0074] Optionally, the damage probability is obtained based on the following formula:

[0075] (2)

[0076] in, For the probability of damage, This represents the probability of failure due to overt damage. This represents the probability of latent damage.

[0077] In some embodiments, sensitive feature parameters are obtained based on the following steps: obtaining the nonlinear correlation between each feature parameter and the degranulation energy level; classifying the feature parameters into low-sensitivity feature parameters and sensitive feature parameters based on the nonlinear correlation; eliminating low-sensitivity feature parameters; and retaining sensitive feature parameters.

[0078] Specifically, the damage characteristics generated by the sample corn kernels during collision are numerous, but not all characteristics can sensitively reflect changes in collision energy. For example, some geometric features may saturate at low-energy collisions or only appear at high-energy collisions, exhibiting complex nonlinear responses. This embodiment preferably uses the Maximum Information Coefficient (MIC) algorithm to measure the nonlinear correlation between each feature parameter and the threshing energy level. By setting a reasonable MIC threshold, features with weak correlation to the threshing energy level and low variability can be identified as low-sensitivity feature parameters and eliminated. This step is essentially a signal-to-noise ratio improvement process, thereby ensuring that the input variables used in subsequent modeling are all effective factors truly driven by the collision damage mechanism, avoiding model overfitting or wasted computational resources due to the introduction of irrelevant variables.

[0079] In some embodiments, the failure probability of overt damage is obtained based on the following steps: obtaining the normalized feature value of the overt damage feature parameter; introducing an exponential penalty activation function, and obtaining the failure probability of overt damage based on the normalized feature value and the optimal weight vector.

[0080] Specifically, the microcracks on the surface of the sample corn kernels have a limited impact on the overall structural integrity before reaching a critical size; however, once the crack length or aperture exceeds a certain critical threshold, the sample corn kernels will undergo catastrophic unstable propagation, leading to instantaneous breakage. Therefore, this embodiment constructs an exponentially penalized activation function in the following form: first, the explicit feature parameters such as crack length and crack aperture are normalized, and then they are substituted into the exponential function form.

[0081] Optionally, the probability of failure due to overt damage is obtained based on the following formula:

[0082] (3)

[0083] in, This represents the probability of failure due to overt damage. The optimal weight vector, is the normalized eigenvalue of the j-th overt damage feature parameter.

[0084] Optionally, the cumulative probability of latent damage is obtained based on the following formula:

[0085] (4)

[0086] in, As latent damage accumulates, The optimal weight vector, It is the normalized eigenvalue of the i-th latent damage feature parameter.

[0087] In step S400, the grain moisture content, maximum normal contact force, collision incident velocity, number of collisions and cumulative plastic dissipation energy are configured as input variables, and the damage probability is configured as the objective function to construct the constitutive equation for the evolution of grain collision damage over a wide range of moisture content.

[0088] In this embodiment, the construction of the constitutive equation is essentially a mathematical modeling process that establishes a mapping relationship between input and output variables. The input variables include the aforementioned collision condition parameters (wide-range moisture content, collision velocity, and number of collisions), as well as the accumulated plastic dissipation energy and the maximum normal contact force. The output variable is the damage probability obtained in step S300. It should be understood that although this embodiment mainly describes the overall framework for constructing the constitutive equation, in actual implementation, this construction process is not a simple numerical fitting, but rather a constrained optimization process that incorporates the physical laws of material damage evolution to ensure that the generated equation can truly reflect the nonlinear evolution of sample corn kernel damage with energy accumulation under wide-range moisture content conditions.

[0089] In some embodiments, the constitutive equation is constructed based on the following steps: a penalized fitness function is built based on multiple physical constraints; the multiple physical constraints are configured to include model constitutive morphology constraints, physical dimension constraints, and evolution mechanism constraints; the collision condition parameters, the cumulative plastic dissipation energy, and the maximum normal contact force are configured as input variables, the damage probability is configured as the objective function, and the constitutive equation is constructed based on the penalized fitness function.

[0090] Specifically, the constitutive equation construction in this embodiment is not a traditional pure data fitting, but a symbolic regression optimization process deeply guided by prior physical knowledge. To overcome the shortcomings of purely data-driven models that easily produce results that violate physical common sense (such as the damage probability decreasing with increasing energy, or predicted values ​​exceeding the probability domain), this invention innovatively transforms the fundamental axioms of material damage mechanics into a mathematically penalized fitness function, making it an intrinsic constraint when the algorithm searches the solution space. This means that the final generated constitutive equation not only numerically approximates the experimental samples but also structurally conforms to the physical laws governing the damage evolution of the corn kernels in the samples.

[0091] In some embodiments, the constitutive morphological constraints of the model include limiting the polynomial and exponential functions to basis functions that fit the nonlinear evolution characteristics of material damage; the physical dimensional constraints include a combination of dimensional consistency constraints and redundant feature elimination; the evolution mechanism constraints include embedding an extreme value penalty term for irreversible damage and ensuring that the constitutive equation satisfies a preset interval.

[0092] Specifically, at the constitutive morphological constraint level of the model, this invention limits the basis function library of the symbolic regression algorithm to only include combinations of polynomial and exponential functions. In the initial stage of a collision, damage grows slowly and approximately linearly with energy accumulation; as microcracks initiate and propagate, the damage enters an accelerated evolution phase; and near complete failure, the damage rate tends to saturate. Therefore, purely linear or polynomial basis functions are difficult to accurately fit the entire domain, while purely exponential functions are difficult to capture the initial linear response. Therefore, limiting the combination of polynomial and exponential functions ensures, from a mathematical structure perspective, that the generated equations possess the ability to describe the nonlinear evolution of the entire damage process, while avoiding non-physical oscillations caused by overfitting.

[0093] At the level of physical dimension constraint logic, this invention enforces dimensionless processing when constructing input variables, thereby eliminating the impact of differences in units of different physical quantities on the stability of numerical calculations, ensuring the dimensional homogeneity of the constitutive equations, and giving the undetermined coefficients in the equations clear physical meaning rather than being purely mathematical fitting parameters. Simultaneously, the algorithm automatically eliminates redundant feature combinations that cannot form dimensionless groups during the search process, i.e., feature combinations without physical meaning, thereby improving the simplicity and generalization ability of the model.

[0094] At the level of evolutionary mechanism constraints, two key types of extreme value penalty terms are embedded in the penalty-type fitness function. The first type is a monotonic penalty term, used to enforce the satisfaction of... That is, the damage probability D must increase with the cumulative plastic dissipation energy. The value increases monotonically with increasing value. During algorithm iteration, if a candidate equation calculates a negative derivative value at any sampling point, a huge penalty value will be added to its fitness score, forcing the equation to be eliminated in natural selection. The second type is the range boundary penalty term, used to force the expression to meet the specified value. 0 represents that the corn kernels in the sample are completely undamaged, and 1 represents that the corn kernels in the sample are completely damaged. When the output value of the candidate equation exceeds this probability definition range, a penalty mechanism will also be triggered.

[0095] As a preferred result of the above-mentioned multi-physical constraint-guided modeling, this embodiment provides an example of a constitutive equation for the evolution of grain collision damage over a wide range of moisture content obtained through constraint optimization:

[0096] (5);

[0097] Where D is the probability of damage (0≤D≤1). For wide-range moisture content, Here, n is the reference moisture content used for dimensionless conversion, and n is the number of collisions. Let be the cumulative plastic dissipation energy after n collisions. For reference dissipation energy, For the maximum normal contact force, Let the collision incident velocity be... For reference to collision power or for power conservation terms used to ensure dimensional consistency, , , , , and All of these are dimensionless model parameters determined through constraint optimization.

[0098] It should be understood that this equation is not a pre-set empirical formula, but rather a product of the adaptive evolution of the symbolic regression algorithm under the aforementioned triple physical constraints. The structure of this equation profoundly reflects the physical mechanism of damage evolution: where... It characterizes the fatigue cumulative effect of damage with increasing number of collisions, reflecting the historical dependence of damage; Taking into account the coupling effect of energy dissipation and instantaneous impact intensity, the exponential function form of the outer layer indicates the nonlinear saturation characteristics of damage from quantitative to qualitative change. This example fully demonstrates that the multi-physical constraint modeling method proposed in this invention can effectively uncover the physical constitutive relationships hidden behind complex experimental data, providing an evaluation method with both theoretical depth and engineering practicality for the accurate quantification of collision damage in sample corn kernels.

[0099] Step S500: Quantitatively evaluate the kernel collision damage during the corn threshing process based on the constitutive equation.

[0100] Specifically, once the constitutive equation is constructed, it can be applied to the quantitative analysis of kernel damage during the corn threshing process. By simply obtaining the parameters, cumulative plastic dissipation energy, and maximum normal contact force under the corn threshing collision condition, and substituting them into the constitutive equation, the corresponding corn damage probability can be calculated in real time, thereby achieving a rapid and quantitative assessment of the degree of single kernel collision damage during the corn threshing process.

[0101] In some embodiments, the method for quantitatively evaluating kernel collision damage during corn harvesting and threshing further includes: embedding constitutive equations into discrete element simulation software to achieve real-time quantitative analysis of collision damage to the corn kernels under test.

[0102] Specifically, the constitutive equations constructed in the aforementioned embodiments are first compiled into a dynamic link library file or a user-defined script module. Then, using the open application programming interface (EDEMAPI) provided by the discrete element simulation software, this module is registered as a callable function within the solver. This allows the constitutive equations to run as a native component of the simulation engine, rather than an external post-processing tool, thus ensuring computational efficiency and data synchronization. It should be understood that although this embodiment uses discrete element simulation software as an example, this method is equally applicable to finite element analysis software or multibody dynamics simulation platforms in other embodiments, as long as the platform supports loading user-defined material models or contact models.

[0103] Through the above steps S100 to S500, this embodiment establishes a complete quantitative evaluation method for collision damage of maize kernels. With cumulative plastic dissipation energy and cross-scale characteristic parameters as the link, it not only ensures the physical authenticity of the evaluation results, but also achieves an accurate description of the damage evolution law under wide-range moisture content conditions.

[0104] Accordingly, the present invention also provides a quantitative evaluation system for kernel collision damage during corn harvesting and threshing, comprising: a first acquisition module configured to acquire multiple sample corn kernels; a second acquisition module configured to conduct collision tests on the sample corn kernels based on preset collision condition parameters, recording the cumulative plastic dissipation energy, maximum normal contact force, and characteristic parameters at each collision frequency; the collision condition parameters include collision incident velocity, number of collisions, and kernel moisture content, and the characteristic parameters include explicit damage characteristic parameters and implicit damage characteristic parameters; a third acquisition module, based on the characteristic parameters, using a pre-established single-kernel collision damage quantitative evaluation method, to acquire the damage probability of sample corn kernels under different collision conditions; a constitutive equation construction module configured to configure kernel moisture content, maximum normal contact force, collision incident velocity, number of collisions, and cumulative plastic dissipation energy as input variables, and damage probability as the objective function, to construct a constitutive equation for the evolution of kernel collision damage over a wide range of moisture content; and an evaluation module configured to quantitatively evaluate kernel collision damage during corn threshing based on the constitutive equation.

[0105] It is understood that, compared with the prior art, the quantitative evaluation system for kernel collision damage during corn harvesting and threshing provided in this embodiment of the invention includes all the technical features and technical effects of the quantitative evaluation method for kernel collision damage during corn harvesting and threshing described above, which will not be repeated here.

[0106] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0107] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for quantitatively evaluating kernel collision damage during corn harvesting and threshing, characterized in that, include: Obtain multiple samples of corn kernels; The sample corn kernels were subjected to a collision test based on preset collision condition parameters. The cumulative plastic dissipation energy, maximum normal contact force and characteristic parameters were recorded at each collision frequency. The collision condition parameters included the collision incident velocity, the number of collisions and the kernel moisture content. The characteristic parameters included the overt damage characteristic parameters and the latent damage characteristic parameters. Based on the aforementioned feature parameters, a pre-established single-kernel collision damage quantification evaluation method is used to obtain the damage probability of the sample corn kernels under different collision conditions. The constitutive equation for the evolution of grain collision damage over a wide range of moisture content is constructed by configuring the grain moisture content, the maximum normal contact force, the collision incident velocity, the number of collisions, and the cumulative plastic dissipation energy as input variables and the damage probability as the objective function. The constitutive equation is used to quantitatively evaluate the kernel collision damage during the corn threshing process.

2. The method for quantitatively evaluating kernel collision damage during corn harvesting and threshing according to claim 1, characterized in that, The feature parameters are obtained based on the following steps: Multi-collision energy level impact tests were conducted on the sample corn kernels with different moisture contents to separate the obviously damaged kernels from the sample corn kernels; the obviously damaged kernels included the sample corn kernels that were broken and had apparent cracks. Extract the dominant damage characteristic parameters of the dominantly damaged grains; In the remaining seemingly intact corn kernels, kernels with hidden damage were separated based on a high-intensity cold light source transmission test. Extract the latent damage characteristic parameters of the latently damaged grains.

3. The method for quantitatively evaluating kernel collision damage during corn harvesting and threshing according to claim 1, characterized in that, Based on the aforementioned feature parameters, the steps for obtaining the damage probability of the sample corn kernels under different collision conditions using a pre-established single-kernel collision damage quantification evaluation method include: Obtain the sensitive feature parameter from among the multiple feature parameters; Obtain the standard deviation and correlation coefficient between the aforementioned sensitive feature parameters; Based on the standard deviation and the correlation coefficient, obtain the optimal weight vector of the feature parameter; Based on the explicit damage feature parameters and the optimal weight vector, the explicit damage failure probability is obtained; Based on the latent damage feature parameters and the optimal weight vector, the cumulative probability of latent damage is obtained; The damage probability is obtained based on the failure probability of the explicit damage and the cumulative probability of the implicit damage.

4. The method for quantitatively evaluating kernel collision damage during corn harvesting and threshing according to claim 3, characterized in that, The sensitive feature parameters are obtained based on the following steps: Obtain the nonlinear correlation between each of the aforementioned characteristic parameters and the threshing energy level; Based on the aforementioned nonlinear correlation, the feature parameters are divided into low-sensitivity feature parameters and sensitive feature parameters; Remove the aforementioned low-sensitivity feature parameters; The sensitive feature parameters are retained.

5. The method for quantitatively evaluating kernel collision damage during corn harvesting and threshing according to claim 3, characterized in that, The probability of overt damage failure is obtained based on the following steps: Obtain the normalized eigenvalues ​​of the overt damage feature parameters; An exponential penalty activation function is introduced, and the probability of failure of the explicit damage is obtained based on the normalized feature value and the optimal weight vector.

6. The method for quantitatively evaluating kernel collision damage during corn harvesting and threshing according to claim 1, characterized in that, The explicit damage characteristic parameters are configured to include crack length, crack aperture, grain breakage rate, and crack two-dimensional fractal dimension, while the implicit damage characteristic parameters are configured to include three-dimensional fractal dimension, micropore volume fraction, and damage volume fraction.

7. The method for quantitatively evaluating kernel collision damage during corn harvesting and threshing according to claim 1, characterized in that, The constitutive equation is constructed based on the following steps: A penalized fitness function is constructed based on multiple physical constraints; the multiple physical constraints are configured to include model constitutive morphology constraints, physical dimension constraints, and evolutionary mechanism constraints. The collision condition parameters, the cumulative plastic dissipation energy, and the maximum normal contact force are configured as input variables, the damage probability is configured as the objective function, and the constitutive equation is constructed based on the penalized fitness function.

8. The method for quantitatively evaluating kernel collision damage during corn harvesting and threshing according to claim 7, characterized in that, The constitutive morphological constraints of the model include limiting the polynomials and exponential functions to be basis functions that fit the nonlinear evolution characteristics of material damage; the physical dimensional constraints include a combination of dimensional consistency constraints and redundant feature elimination. The evolutionary mechanism constraint includes embedding an extreme value penalty term for irreversible damage and ensuring that the constitutive equation satisfies a preset interval.

9. The method for quantitatively evaluating kernel collision damage during corn harvesting and threshing according to claim 1, characterized in that, The method further includes: The constitutive equation is embedded in discrete element simulation software to achieve real-time quantitative analysis of the collision damage of kernels during corn threshing.

10. A quantitative evaluation system for kernel collision damage during corn harvesting and threshing, used to execute the quantitative evaluation method for kernel collision damage during corn harvesting and threshing as described in any one of claims 1 to 9, characterized in that, include: The first acquisition module is configured to acquire multiple samples of corn kernels; The second acquisition module is configured to conduct a collision test on the sample corn kernels based on preset collision condition parameters, and record the cumulative plastic dissipation energy, maximum normal contact force and characteristic parameters at each collision frequency; the collision condition parameters include the collision incident velocity, the number of collisions and the kernel moisture content, and the characteristic parameters include the overt damage characteristic parameters and the latent damage characteristic parameters. The third acquisition module is configured to acquire the damage probability of the sample corn kernels under different collision conditions based on the feature parameters and using a pre-established single kernel collision damage quantification evaluation method. The constitutive equation construction module is configured to take the grain moisture content, the maximum normal contact force, the impact incident velocity, the number of impacts and the cumulative plastic dissipation energy as input variables, and the damage probability as the objective function to construct a constitutive equation for the evolution of grain impact damage over a wide range of moisture content. The evaluation module is configured to quantitatively evaluate kernel collision damage during corn threshing based on the constitutive equation.

Images