Method and device for optimizing structure of cut stem sorting mechanism

Abstract

The invention discloses a cut stem sorting mechanism structure optimization method and device, and the method comprises the steps: obtaining a numerical calculation model of a plurality of drilling plate structure parameters based on a three-dimensional model of a cut stem sorting mechanism; based on the structure parameters of the multiple drilling plates, a first orthogonal experiment design space is constructed in combination with the wide spacing level, and sensitive parameters are determined with the average speed, the speed uniformity index and the pressure loss after airflow passes through the drilling plates as evaluation indexes; on the basis of the sensitive parameters, a second orthogonal experiment design space is constructed in combination with the narrow spacing level, and target parameters are determined with the particle interception capacity, the speed uniformity index and the pressure loss as evaluation indexes; and training the target neural network model based on a numerical calculation result of the numerical calculation model of the target parameter as training data to obtain an optimal solution of the target parameter. According to the method, the optimal solution of the target parameters can be utilized, and the drilling scheme of the drilling plate in the cut stem sorting mechanism can be optimized.

Description

Technical Field

[0001] This application relates to the field of stem sorting mechanisms, and in particular to a method and apparatus for optimizing the structure of a stem sorting mechanism. Background Technology

[0002] With economic development and improved living standards, consumers have higher expectations for the quality of various products. In fields such as tobacco and tea, they demand higher purity, better taste, and better appearance. Therefore, the performance of stem sorting mechanisms needs continuous optimization. Stem sorting mechanisms contain multiple functional modules, among which the perforated plate plays a role in streamlining airflow and maintaining internal cleanliness. However, its structural parameters present a design contradiction: smaller perforations are beneficial for maintaining internal cleanliness, but the higher airflow velocity causes flow field chaos, hindering the sorting of materials and impurities, and significantly increasing local energy loss; larger perforations are beneficial for streamlining airflow, but also lead to a large accumulation of particles at the bottom, obstructing airflow. The stem sorting mechanism has a complex structure, and disassembling the equipment using experimental analysis methods is time-consuming. Therefore, an effective numerical calculation method is needed to save costs and optimize the perforated plate structure. Furthermore, while optimizing parameters can generally be obtained by constructing an orthogonal experimental design space, it is impossible to optimize structural parameters within a continuous input space.

[0003] The closest existing technology is likely to be the use of orthogonal experimental design combined with CFD (Computational Fluid Dynamics) simulation to optimize the sorting mechanism structure. The typical workflow is: determine influencing parameters → design orthogonal experimental tables → perform CFD simulations → analyze results → select the best parameter combination from the orthogonal table. However, existing methods do not consider the particulate phase, have extremely high computational costs, limit the number of experiments, and orthogonal experiments can only be selected from preset discrete level combinations, making it impossible to establish a continuous mapping relationship between parameters and performance, and difficult to accurately predict and optimize parameter combinations that have not been simulated. Summary of the Invention

[0004] This application provides a method and apparatus for optimizing the structure of a stem sorting mechanism, with the aim of tailoring an optimal drilling scheme for a perforated plate (i.e., a drilled plate) in the stem sorting device.

[0005] To achieve the above objectives, this application provides the following technical solution:

[0006] A method for optimizing the structure of a stem sorting mechanism includes:

[0007] Numerical calculation models for multiple borehole plate structural parameters are obtained based on the three-dimensional model of the filament sorting mechanism.

[0008] Based on multiple borehole plate structural parameters and combined with wide-spacing horizontals, a first orthogonal experimental design space is constructed, and the average velocity, velocity uniformity index, and pressure loss of the airflow after passing through the borehole plate are used as evaluation indicators to determine sensitive parameters.

[0009] Based on the aforementioned sensitive parameters and combined with the narrow spacing level, a second orthogonal experimental design space is constructed, and the target parameters are determined using particle retention capacity, velocity uniformity index, and pressure loss as evaluation indicators.

[0010] The numerical calculation results of the numerical calculation model based on the target parameters are used as training data to train the target neural network model to obtain the optimal solution of the target parameters; the target neural network model is used to characterize the mapping relationship between the value of the target parameters and the performance of the borehole plate.

[0011] Optionally, based on multiple drill plate structural parameters and combined with a wide-spacing horizontal plane, a first orthogonal experimental design space is constructed, and the average velocity, velocity uniformity index, and pressure loss of the airflow after passing through the drill plate are used as evaluation indicators to determine sensitive parameters, including:

[0012] Based on the aforementioned three-dimensional model, and combined with computational fluid dynamics, a turbulence model of the filament sorting mechanism is generated; the turbulence model is used to simulate the flow field information of the internal channels of the filament sorting mechanism.

[0013] Based on multiple drill plate structural parameters and combined with wide-spacing horizontals, a first orthogonal experimental design space is constructed.

[0014] In the first orthogonal experimental design space, the turbulence model is run to obtain the first simulation result; the first simulation result includes the flow field information of the internal channel of the wire sorting mechanism under different values ​​of the various drill plate structural parameters;

[0015] From the first simulation results, the average velocity, velocity uniformity index, and pressure loss of the airflow after passing through the perforated plate are extracted;

[0016] Using the average velocity of the airflow after passing through the perforated plate, the velocity uniformity index, and the pressure loss as evaluation indicators, sensitive parameters are selected from the various structural parameters of the perforated plate.

[0017] Optionally, sensitive parameters are selected from the various structural parameters of the perforated plate, using the average velocity of the airflow after passing through the perforated plate, the velocity uniformity index, and the pressure loss as evaluation indicators, including:

[0018] The first evaluation index is obtained by weighting and summing the average velocity of the airflow after passing through the perforated plate, the velocity uniformity index, and the pressure loss.

[0019] Based on the first evaluation index, at least one sensitive parameter is selected from the various drill plate structural parameters; wherein the first evaluation index of the sensitive parameter is higher than that of other drill plate structural parameters.

[0020] Optionally, based on the aforementioned sensitive parameters and combined with the narrow spacing level, a second orthogonal experimental design space is constructed, and the target parameters are determined using particle retention capacity, velocity uniformity index, and pressure loss as evaluation indicators, including:

[0021] Based on the turbulence model of the filament sorting mechanism, a two-phase coupled numerical calculation model is generated by combining the discrete element method. The two-phase coupled numerical calculation model is used to simulate the force information and flow field information of the internal channel of the filament sorting mechanism. The force information includes the force of the fluid on the particles. The fluid is used to characterize air, and the particles are used to characterize materials.

[0022] Based on the aforementioned sensitive parameters, and combined with the narrow-spacing level, a second orthogonal experimental design space is constructed.

[0023] In the second orthogonal experimental design space, the two-phase coupled numerical calculation model is run to obtain the second simulation results; the second simulation results include the force information and flow field information of the internal channel of the filament sorting mechanism under different values ​​of multiple sensitive parameters.

[0024] From the second simulation results, particle retention capacity, velocity uniformity index, and pressure loss are extracted; the particle retention capacity is equal to... In the formula Characterizes the particle mass passing through the drilled plate. Characterizes the mass of particles retained by the drill plate;

[0025] Using the particle retention capacity, the velocity uniformity index, and the pressure loss as evaluation indicators, target parameters are selected from the various sensitive parameters.

[0026] Optionally, using the particle retention capacity, the velocity uniformity index, and the pressure loss as evaluation indicators, target parameters are selected from the various sensitive parameters, including:

[0027] The second evaluation index is obtained by weighting and summing the particle retention capacity, the velocity uniformity index, and the pressure loss.

[0028] Based on the second evaluation index, at least one target parameter is selected from the various sensitive parameters; wherein the second evaluation index of the target parameter is higher than that of the other sensitive parameters.

[0029] Optionally, the target neural network model includes a residual network and a particle swarm optimization module; the residual network is used to fit the mapping relationship; the particle swarm optimization module is used to obtain the optimal values ​​of the target parameters in the continuous space based on the mapping relationship using a particle swarm optimization algorithm.

[0030] Optionally, the target parameters include opening shape parameters and position parameters; the opening shape parameters include the length and width of the rectangular hole; the position parameters include the interval between the rectangular hole and the first rectangular hole, and the interval between the rectangular hole and the second rectangular hole; the first rectangular hole is another rectangular hole that is adjacent to the rectangular hole in the horizontal direction; the second rectangular hole is another rectangular hole that is adjacent to the rectangular hole in the vertical direction.

[0031] A device for optimizing the structure of a stem sorting mechanism includes:

[0032] The computational model determination unit is used to obtain a numerical calculation model of multiple drill plate structural parameters based on the three-dimensional model of the wire sorting mechanism.

[0033] The sensitive parameter screening unit is used to construct a first orthogonal experimental design space based on multiple borehole plate structural parameters and a wide-spacing horizontal plane, and to determine the sensitive parameters by using the average velocity, velocity uniformity index and pressure loss of the airflow after passing through the borehole plate as evaluation indicators.

[0034] The target parameter screening unit is used to construct a second orthogonal experimental design space based on the sensitive parameters and the narrow spacing level, and to determine the target parameters using particle retention capacity, velocity uniformity index and pressure loss as evaluation indicators.

[0035] The machine model training unit is used to train the target neural network model based on the numerical calculation results of the numerical calculation model of the target parameters as training data to obtain the optimal solution of the target parameters; the target neural network model is used to characterize the mapping relationship between the value of the target parameters and the performance of the borehole plate.

[0036] A storage medium comprising a stored program, wherein the program is executed by a processor to perform the described filament sorting mechanism structure optimization method.

[0037] An electronic device includes: a processor, a memory, and a bus; the processor and the memory are connected via the bus.

[0038] The memory is used to store the program, and the processor is used to run the program, wherein the program is executed by the processor to perform the structure optimization method of the filament sorting mechanism.

[0039] The technical solution provided in this application obtains numerical calculation models of multiple perforated plate structural parameters based on a three-dimensional model of the filament sorting mechanism. Based on these parameters and a wide-spacing horizontal plane, a first orthogonal experimental design space is constructed, and the average velocity, velocity uniformity index, and pressure loss of the airflow after passing through the perforated plate are used as evaluation indicators to determine sensitive parameters. Based on these sensitive parameters and a narrow-spacing horizontal plane, a second orthogonal experimental design space is constructed, and the particle retention capacity, velocity uniformity index, and pressure loss are used as evaluation indicators to determine target parameters. The numerical calculation results of the numerical calculation model based on the target parameters are used as training data to train a target neural network model to obtain the optimal solution for the target parameters. This application can optimize the perforation scheme of the perforated plate in the filament sorting mechanism using the optimal solution for the target parameters. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 A flowchart illustrating a method for optimizing the structure of a stem sorting mechanism provided in this application embodiment;

[0042] Figure 2 A flowchart illustrating another method for optimizing the structure of a stem sorting mechanism provided in this application embodiment;

[0043] Figure 3 A flowchart illustrating another method for optimizing the structure of a stem sorting mechanism provided in this application embodiment;

[0044] Figure 4 A schematic diagram of the structure optimization device for a stem sorting mechanism provided in this application embodiment;

[0045] Figure 5 A simplified three-dimensional model schematic diagram of a stem sorting mechanism provided in this application embodiment;

[0046] Figure 6 A schematic diagram of the airflow trajectory in the internal channel of a stem sorting mechanism provided in this application embodiment;

[0047] Figure 7 This is a schematic diagram of a drill plate structure provided in an embodiment of this application;

[0048] Figure 8 A schematic diagram of flow field information for a stem sorting mechanism provided in an embodiment of this application;

[0049] Figure 9 A schematic diagram of the fluid-particle calculation results output by a two-phase coupled numerical calculation model provided in this application embodiment;

[0050] Figure 10 This is a schematic diagram of the architecture of a Res-FCNet neural network provided in an embodiment of this application. Detailed Implementation

[0051] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0052] In this application, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. The terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0053] The technical terms used in the embodiments of this application are as follows.

[0054] Stem and shred sorting mechanism: a specialized device used to separate tobacco stems and shreds.

[0055] Drill plate: A key component in the sorting mechanism, featuring holes of a specific arrangement and shape to streamline airflow, trap impurities, and prevent clogging.

[0056] CFD (Computational Fluid Dynamics): Simulates the flow state of fluids (such as air) using numerical methods.

[0057] DEM (Discrete Element Method): A numerical method for simulating the motion and collisions of discrete particles (such as tobacco stems and tobacco shreds).

[0058] CFD-DEM Coupling: Combining CFD and DEM simulations for accurate simulation of the interaction between fluids and particles (such as drag force and lift).

[0059] Orthogonal experimental design: a multi-factor experimental design method that efficiently analyzes the influence of multiple factors on the results using a small number of experiments.

[0060] ResFCNet (Residual Fully Connected Network): A deep learning model with good learning capabilities.

[0061] PSO (Particle Swarm Optimization): A smart optimization algorithm that simulates the foraging behavior of flocks of birds and is used to find the optimal solution in a continuous space.

[0062] Proxy model: Using a computationally inexpensive mathematical model (such as a trained neural network) to replace a computationally expensive simulation or experimental process.

[0063] Velocity uniformity index: an indicator for evaluating the uniformity of airflow distribution.

[0064] Pressure loss: The pressure difference before and after a fluid flows through a channel or component, reflecting the amount of energy consumed.

[0065] Particle trapping capability: The ability of a drill plate to intercept and prevent impurity particles from passing through.

[0066] like Figure 1 The diagram shown is a flowchart illustrating a method for optimizing the structure of a stem sorting mechanism according to an embodiment of this application, including the following steps.

[0067] S101: A numerical calculation model for obtaining multiple borehole plate structural parameters based on a three-dimensional model of a wire sorting mechanism.

[0068] Among them, the structure of the wire sorting mechanism can be simplified and a corresponding three-dimensional model can be established. Based on the three-dimensional model, a numerical calculation model of multiple drill plate structural parameters can be obtained.

[0069] In some examples, the stem sorting mechanism is modeled in 3D using SolidWorks to obtain... Figure 5 The figure shows a simplified 3D model of a filament sorting mechanism with local fine structures removed. The main structures are air inlet 1, perforated plate 2, impurity outlet 3, feed inlet 4, air outlet 5, and material outlet 6.

[0070] In a possible implementation, the airflow trajectory within the internal channel of the filament sorting mechanism is as follows: Figure 6As shown, material containing impurities enters the wire sorting mechanism through the feed inlet 4. Under the action of airflow, due to the difference in physical properties of different particles, the lighter material particles leave the channel from the material outlet 6. The denser impurities settle to the perforated plate 2 in the airflow. At this time, the perforated plate plays a role in preventing impurity particles from accumulating at the bottom and hindering the airflow. The impurity particles are carried away from the channel by the spiral mechanism at the impurity outlet 3.

[0071] S102: Based on multiple borehole plate structural parameters and combined with wide-spacing horizontals, a first orthogonal experimental design space is constructed, and the average velocity, velocity uniformity index and pressure loss of the airflow after passing through the borehole plate are used as evaluation indicators to determine sensitive parameters.

[0072] Among them, the first orthogonal experimental design space can be regarded as a wide-spacing orthogonal experiment, which includes all the structural parameters of the drill plate that may affect the airflow distribution in the wide-spacing orthogonal experiment, and selects the sensitive parameters that affect the performance of the drill plate.

[0073] Optionally, based on multiple borehole plate structural parameters and combined with a wide-spacing horizontal plane, a first orthogonal experimental design space is constructed. The average velocity, velocity uniformity index, and pressure loss of the airflow after passing through the borehole plate are used as evaluation indicators to determine the implementation process of the sensitive parameters. (See [link to relevant documentation]). Figure 2 The steps shown are accompanied by corresponding explanations.

[0074] S103: Based on sensitive parameters and combined with narrow spacing levels, a second orthogonal experimental design space is constructed, and the target parameters are determined using particle retention capacity, velocity uniformity index and pressure loss as evaluation indicators.

[0075] Among them, the second orthogonal experimental design space can be regarded as a narrow-spacing orthogonal experiment, which includes all sensitive parameters and selects the target parameters that have the most significant impact on the performance of the drill plate.

[0076] Optionally, based on sensitive parameters and combined with narrow-spacing levels, a second orthogonal experimental design space is constructed. The particle trapping capacity, velocity uniformity index, and pressure loss are used as evaluation indicators to determine the process for achieving the target parameters. (See [link to relevant documentation]). Figure 3 The steps shown are accompanied by corresponding explanations.

[0077] Optionally, the target parameters include opening shape parameters and position parameters; the opening shape parameters include the length and width of the rectangular hole; the position parameters include the interval between the rectangular hole and the first rectangular hole, and the interval between the rectangular hole and the second rectangular hole; the first rectangular hole is another rectangular hole that is adjacent to the rectangular hole in the horizontal direction; the second rectangular hole is another rectangular hole that is adjacent to the rectangular hole in the vertical direction.

[0078] See in some examples Figure 7As shown, the length of the rectangular hole can be set to d1, the width can be set to d2, the interval between the rectangular hole and the first rectangular hole can be set to L1, and the interval between the rectangular hole and the second rectangular hole can be set to L2.

[0079] In a possible implementation, the objective parameter in the discrete space can be denoted as: .

[0080] S104: The numerical calculation results of the numerical calculation model based on the target parameters are used as training data to train the target neural network model in order to obtain the optimal solution of the target parameters.

[0081] Among them, the target neural network model is used to characterize the mapping relationship between the numerical values ​​of the target parameters and the performance of the borehole plate.

[0082] It should be noted that while the optimal parameter combinations in the discrete space were obtained through the analysis of narrow-spacing horizontal orthogonal experiments, it is impossible to set the number of factor levels to a large value under realistic conditions. To further enhance the ability of numerical calculation data to infer optimal parameter combinations, an optimization method based on a neural network model is proposed. This method can more effectively learn complex nonlinear mappings, has stronger generalization ability, and provides a fast and accurate evaluation tool for subsequent continuous optimization.

[0083] Optionally, the target neural network model includes a residual network and a particle swarm optimization module; the residual network is used to fit the mapping relationship; the particle swarm optimization module is used to obtain the optimal values ​​of the target parameters in the continuous space based on the mapping relationship through the particle swarm optimization algorithm.

[0084] In some examples, the residual network can be a Res-FCNet neural network, with a network structure such as... Figure 10 As shown, the network contains two hidden layers. Each time a neuron passes through a hidden layer, Dropout is used to randomly deactivate 20% of the neurons to reduce the risk of overfitting during the learning process and improve the robustness of predictions. Residual connections are introduced to avoid gradient vanishing in deep networks, which is beneficial for fully learning from the data and improving prediction ability. The dataset is randomly selected with 80% as the training set and 20% as the test set. The learning rate is set to 0.0001, the optimizer is the Adam algorithm, and iterative training is performed for 100,000 epochs based on the MSE (mean squared error) loss. The results show that the training and test sets... The values ​​are all above 0.95, indicating that the target neural network model fits the training data well and has good predictive ability.

[0085] In a possible implementation, the performance of the drill plate in the residual network can be set to... Specifically, it can be expressed as formula (6).

[0086] (6)

[0087] In formula (6), Represents the activation function. All of these are hyperparameters.

[0088] In some examples, based on the mapping relationship between the target parameters and the performance of the drill plate, the optimal values ​​of the target parameters in the continuous space are obtained through PSO. Specifically, the positions of the particle swarm are first randomly generated within the ranges of d1∈[9,11] mm, d2∈[8,10.5] mm, and L1, L2∈[4,6] mm. and speed The objective function is The position and velocity are updated according to formula (7).

[0089] (7)

[0090] In formula (7), Represents inertia weight. Represents learning factors, Represents a random number within the interval [0,1]. Representative particles The optimal position of the individual, and , Represents the global optimal position of the group, until The change converges to 0.01, yielding the optimal solution for the objective parameters in the continuous space. .

[0091] Combination Figures 2-3 The core process of the filament sorting mechanism structure optimization method shown in the embodiments of this application can be summarized as follows: model simplification and parameter determination → wide-spacing orthogonal experiment (CFD) → sensitive parameter screening → narrow-spacing orthogonal experiment (CFD-DEM) → discrete optimal solution → construction of ResFCNet surrogate model → PSO continuous space optimization → continuous optimal solution. This application aims to automatically find the optimal solution of the perforated plate structure parameters that best balances the three conflicting objectives of "airflow uniformity," "impurity retention capacity," and "system pressure loss (energy consumption)" within a continuous parameter space through the combination of accurate CFD-DEM simulation, neural network surrogate model, and intelligent optimization algorithm. This method can then be extended to the entire sorting mechanism and other similar sorting devices.

[0092] The structure optimization method for the stem sorting mechanism shown in the embodiments of this application can achieve the following beneficial effects.

[0093] (1) High efficiency: Computer simulation and algorithm optimization replace a large number of time-consuming physical experiments and trial and error, greatly shortening the design cycle.

[0094] (2) Low cost: Significantly reduces the economic costs required to manufacture physical prototypes and conduct experimental tests.

[0095] (3) High precision: CFD-DEM coupled simulation ensures the accuracy of the physical model, and ResFCNet-PSO ensures the globality and accuracy of the optimization search. The results obtained are better than those of traditional methods.

[0096] (4) Strong universality: The framework of this method is universal and can be applied to the sorting of stems.

[0097] In the embodiments of this application, the introduction of CFD-DEM two-phase flow coupled numerical simulation into the structural optimization of the wire sorting mechanism can accurately simulate the interaction between airflow and particles, making the evaluation of key performance such as the "retention capacity" of the perforated plate more accurate and reliable, far exceeding that of single CFD analysis. This provides an effective means for structural optimization design in this field. By combining the trained ResFCNet model with the PSO algorithm, automatic global optimization search is realized in the continuous design space, breaking through the limitation of traditional orthogonal experiments that can only find the best in discrete level combinations. A truly global optimal solution is found, and a multi-objective evaluation function that comprehensively considers "airflow uniformity", "impurity retention capacity" and "pressure loss (energy consumption)" is established. This function is then transformed into a single objective for optimization through weighting and other methods, directly solving the inherent contradictions in the design of the perforated plate.

[0098] The processes shown in S101-S104 above can optimize the drilling scheme of the perforated plate in the filament sorting mechanism by utilizing the optimal solution of the target parameters.

[0099] like Figure 2 The diagram shown is a flowchart illustrating another method for optimizing the structure of a stem sorting mechanism provided in this application, including the following steps.

[0100] S201: Based on a three-dimensional model and combined with computational fluid dynamics, a turbulence model of the filament sorting mechanism is generated.

[0101] Among them, the turbulence model is used to simulate the flow field information of the internal channels of the wire sorting mechanism.

[0102] In some examples, the 3D model is imported into Workbench Fluent to calculate the flow field distribution inside the filament sorting mechanism, setting the fluid domain physical parameters, turbulence model, and calculation method. Specifically, the mesh is set to 5mm, the expansion layer mesh height is 0.5mm, the mesh refinement for the perforated plate 2 section is 1mm, the RNGk-ε model is used for turbulence, the standard wall function is used for near-wall treatment, the SIMPLE algorithm is used for calculation, and the momentum equation, turbulent kinetic energy, and other discretization schemes are all based on the second-order upwind scheme to realize the calculation of flow field information.

[0103] S202: Based on multiple drill plate structural parameters and combined with wide-spacing horizontals, a first orthogonal experimental design space is constructed.

[0104] In the first orthogonal experimental design space, CFD calculations were performed on openings of different shapes (e.g., triangles, circles, and squares). Circles and squares showed better uniformity of airflow after passing through the perforated plate. Secondly, considering the airflow area and processing performance, squares were chosen as more suitable.

[0105] In some examples, all parameters affecting the flow field at the local location of the borehole plate are considered; see [link to relevant documentation]. Figure 7 The main structural characteristics of the perforated plate shown include the opening shape parameters d1 and d2, the position parameters L1 and L2, and in addition, the opening type, the distances a1, a2, a3, and a4 between the four sides of the perforated plate and the channel, and the thickness of the perforated plate are also considered. In order to obtain the sensitive parameters that affect the flow field distribution, a wide-spacing horizontal orthogonal experimental design space is constructed, with the parameter fluctuation range being 30% of the standard value. Three levels are selected, for example: the design range of d1 and d2 is 10±3 mm, and the design range of L1 and L2 is 5±1.5 mm.

[0106] In a possible implementation, the flow field information of the filament sorting mechanism can be found in [reference needed]. Figure 8 As shown.

[0107] S203: Run the turbulence model in the first orthogonal experimental design space to obtain the first simulation results.

[0108] The first simulation result includes the flow field information of the internal channel of the wire sorting mechanism under different values ​​of the structural parameters of each drill plate.

[0109] In some examples, fluid simulation software can be used to run a turbulence model to obtain initial simulation results.

[0110] S204: Extract the average velocity, velocity uniformity index, and pressure loss of the airflow after passing through the perforated plate from the first simulation results.

[0111] Among them, the post-processing module of fluid simulation software (such as Fluent software) can be used to extract the average velocity, velocity uniformity index and pressure loss of the airflow after passing through the perforated plate from the first simulation results.

[0112] In some examples, it can be calculated Figure 8 The velocity uniformity index of the section 1 shown is beneficial for the formation of a suitable flow field for the separation of impurities and materials when the flow field in this part is stable. Specifically, the formula for calculating the velocity uniformity index can be found in formula (1).

[0113] (1)

[0114] In formula (1), Represents the number of cross-sectional elements calculated. Representing the Each unit corresponds to a speed. Represents the standard deviation of velocity. Represents the average velocity of the cross section. It represents the velocity uniformity index.

[0115] In some examples, it can be calculated Figure 8 The average velocity of the cross section 2 shown is such that a higher airflow velocity at this location is beneficial for preventing impurities from falling. Specifically, the formula for calculating the average velocity of the airflow after passing through the perforated plate can be found in formula (2).

[0116] (2)

[0117] In formula (2), Represents the number of cross-sectional elements calculated. Representing the Each unit corresponds to a speed. Representing the The area corresponding to each unit This represents the average velocity of the airflow after passing through the perforated plate.

[0118] In some examples, the pressure difference between the air inlet and the air outlet can be calculated. This part has a large capacity loss, and it is necessary to evaluate the impact of the perforated plate on the overall performance of the wire sorting mechanism. Specifically, the formula for calculating the pressure loss can be found in formula (3).

[0119] (3)

[0120] In formula (3), Represents pressure loss, This represents the pressure applied when the material passes through a short distance before reaching the borehole plate. This represents the pressure a short distance after the material leaves the drill plate.

[0121] S205: Sensitive parameters are selected from various perforated plate structural parameters by using the average velocity, velocity uniformity index, and pressure loss of the airflow after passing through the perforated plate as evaluation indicators.

[0122] Optionally, the process of selecting sensitive parameters from various borehole plate structural parameters by using the average velocity, velocity uniformity index, and pressure loss of the airflow after passing through the borehole plate as evaluation indicators can be as follows: a first evaluation indicator is obtained by weighted summation of the average velocity, velocity uniformity index, and pressure loss of the airflow after passing through the borehole plate; at least one sensitive parameter is selected from various borehole plate structural parameters based on the first evaluation indicator; wherein the first evaluation indicator of the sensitive parameter is higher than that of other borehole plate structural parameters.

[0123] In some examples, the impact of each borehole plate structural parameter on the evaluation index was subsequently analyzed using a weighted comprehensive scoring method to screen out sensitive parameters. Considering the large differences in the absolute values ​​of average velocity, velocity uniformity index, and pressure loss, normalized values ​​can be calculated. The higher the average velocity and velocity uniformity index, the better, and the closer the normalized value is to 1. The opposite is true for pressure loss.

[0124] In a possible implementation, the average speed, speed uniformity index and pressure loss are weighted and summed to obtain the calculation process of the first evaluation index, which can be seen in formula (4).

[0125] (4)

[0126] In formula (4), Represents the primary evaluation indicator. Represents the normalized velocity uniformity index. Represents the normalized average velocity. Represents the normalized pressure loss. Represents the preset weights, and .

[0127] The above S201-S205 can utilize the first orthogonal experimental design space to screen out sensitive parameters from various borehole plate structural parameters.

[0128] like Figure 3 The diagram shown is a flowchart illustrating another method for optimizing the structure of a stem sorting mechanism provided in this application, including the following steps.

[0129] S301: Based on the turbulence model of the filament sorting mechanism, a two-phase coupled numerical calculation model is generated by combining the discrete element method.

[0130] Among them, the two-phase coupled numerical calculation model is used to simulate the force information and flow field information of the internal channel of the filament sorting mechanism; the force information includes the force of the fluid on the particles; the fluid is used to characterize air, and the particles are used to characterize materials.

[0131] In some examples, to reduce the workload in determining sensitive parameters, the average velocity obtained through fluid dynamics calculations indirectly represents the particle trapping capacity of the borehole plate. Furthermore, based on this, a CFD-DEM two-phase flow coupled numerical calculation model is built to accurately simulate the particle-fluid-structure interaction. The particle trapping capacity is then used to replace the average airflow velocity to recalculate the weighted evaluation index. (i.e., the second evaluation index) makes the second evaluation index closer to the actual physical process, and the results are more reliable.

[0132] It should be noted that during the coupling process, the fluid simulation software is used to calculate the fluid field and transfer the drag force, lift force and other forces of the fluid on the particles to the discrete element method simulation software (such as EDEM software). The discrete element method simulation software is used to calculate the motion and collision of the particles and transfer the volume fraction and momentum exchange of the particles on the fluid to the fluid simulation software. The states of the particles and the fluid in each time step in the fluid simulation software and the discrete element method simulation software are updated synchronously, and the two solvers are integrated into a unified numerical calculation framework.

[0133] In a possible implementation, the fluid simulation software calculation method can be set to the Phase Coupled SIMPLE algorithm, the spatial discretization method in the discrete element method simulation software can be kept at the system default, the transient solution format can be set to Bounded Second Order Implicit, the particle contact model in the discrete element method simulation software can be the Hertz-Mindlin model, and the calculation time step can be set to 0.84% ​​of the Rayleigh time step, with a value of 10⁻⁶ s.

[0134] In some examples, the fluid-particle calculation results output by the two-phase coupled numerical calculation model can be found in [reference needed]. Figure 9 As shown.

[0135] S302: Based on sensitive parameters and combined with narrow-spacing levels, a second orthogonal experimental design space is constructed.

[0136] In order to obtain sensitive parameters, a second orthogonal experimental design space with narrow spacing was constructed based on the results of the wide-spacing horizontal orthogonal experiment.

[0137] In a possible implementation, the range of the sensitive parameters is d1∈[9,11] mm, d2∈[8,10.5] mm, L1, L2∈[4,6] mm, and 9 levels are selected.

[0138] S303: In the second orthogonal experimental design space, run the two-phase coupled numerical calculation model to obtain the second simulation results.

[0139] The second simulation results include force and flow field information of the internal channels of the filament sorting mechanism under different values ​​of multiple sensitive parameters.

[0140] S304: Extract the particle retention capacity, velocity uniformity index, and pressure loss from the second simulation results.

[0141] Among them, the particle retention capacity is equal to In the formula Characterizes the particle mass passing through the drilled plate. Characterizes the mass of particles retained by the drill plate.

[0142] S305: Using particle retention capacity, velocity uniformity index and pressure loss as evaluation indicators, target parameters are selected from various sensitive parameters.

[0143] Optionally, the process of selecting target parameters from various sensitive parameters using particle retention capacity, velocity uniformity index, and pressure loss as evaluation indicators can be as follows: a second evaluation indicator is obtained by weighted summation of particle retention capacity, velocity uniformity index, and pressure loss; at least one target parameter is selected from various sensitive parameters based on the second evaluation indicator; wherein the second evaluation indicator of the target parameter is higher than that of other sensitive parameters.

[0144] In some examples, the calculation process of the second evaluation index can be found in formula (5).

[0145] (5)

[0146] In formula (5), This represents the second evaluation indicator. This represents the normalized particle retention capacity.

[0147] The procedures shown in S301-S305 above can utilize the second orthogonal experimental design space to select target parameters from various sensitive parameters.

[0148] like Figure 4 The diagram shown is a schematic of the structure optimization device for a stem sorting mechanism provided in this application embodiment, including the following units.

[0149] The computational model determination unit 100 is used to obtain a numerical calculation model of multiple drill plate structural parameters based on the three-dimensional model of the wire sorting mechanism.

[0150] The sensitive parameter screening unit 200 is used to construct a first orthogonal experimental design space based on multiple borehole plate structural parameters and a wide-spacing horizontal plane, and to determine the sensitive parameters by using the average velocity, velocity uniformity index and pressure loss of the airflow after passing through the borehole plate as evaluation indicators.

[0151] Optionally, the sensitive parameter screening unit 200 is specifically used for: generating a turbulence model of the wire sorting mechanism based on a three-dimensional model and combined with computational fluid dynamics; using the turbulence model to simulate the flow field information of the internal channel of the wire sorting mechanism; constructing a first orthogonal experimental design space based on multiple perforated plate structural parameters and combined with wide-spacing horizontals; running the turbulence model in the first orthogonal experimental design space to obtain the first simulation results; the first simulation results include the flow field information of the internal channel of the wire sorting mechanism under different values ​​of the perforated plate structural parameters; extracting the average velocity, velocity uniformity index, and pressure loss of the airflow after passing through the perforated plate from the first simulation results; using the average velocity, velocity uniformity index, and pressure loss of the airflow after passing through the perforated plate as evaluation indicators to screen sensitive parameters from the various perforated plate structural parameters.

[0152] Optionally, the sensitive parameter screening unit 200 is specifically used to: obtain a first evaluation index by weighted summation of the average velocity, velocity uniformity index and pressure loss after the airflow passes through the drill plate; and screen at least one sensitive parameter from each drill plate structural parameter according to the first evaluation index; wherein the first evaluation index of the sensitive parameter is higher than that of other drill plate structural parameters.

[0153] The target parameter screening unit 300 is used to construct a second orthogonal experimental design space based on sensitive parameters and combined with narrow spacing levels, and to determine the target parameters using particle retention capacity, velocity uniformity index and pressure loss as evaluation indicators.

[0154] Optionally, the target parameter filtering unit 300 is specifically used for: generating a two-phase coupled numerical calculation model based on the turbulence model of the stem sorting mechanism and combined with the discrete element method; the two-phase coupled numerical calculation model is used to simulate the force information and flow field information of the internal channel of the stem sorting mechanism; the force information includes the force of the fluid on the particles; the fluid is used to characterize air, and the particles are used to characterize materials; based on the sensitive parameters and combined with the narrow spacing level, a second orthogonal experimental design space is constructed; in the second orthogonal experimental design space, the two-phase coupled numerical calculation model is run to obtain the second simulation results; the second simulation results include the force information and flow field information of the internal channel of the stem sorting mechanism under different values ​​of multiple sensitive parameters; from the second simulation results, the particle retention capacity, velocity uniformity index, and pressure loss are extracted; the particle retention capacity is equal to... In the formula Characterizes the particle mass passing through the drilled plate. The quality of particles retained by the borehole plate is characterized; the particle retention capacity, velocity uniformity index and pressure loss are used as evaluation indicators to select target parameters from various sensitive parameters.

[0155] Optionally, the target parameter screening unit 300 is specifically used to: obtain a second evaluation index by weighted summation of particle retention capacity, velocity uniformity index and pressure loss; and select at least one target parameter from various sensitive parameters according to the second evaluation index; wherein the second evaluation index of the target parameter is higher than that of other sensitive parameters.

[0156] Optionally, the target parameters include opening shape parameters and position parameters; the opening shape parameters include the length and width of the rectangular hole; the position parameters include the interval between the rectangular hole and the first rectangular hole, and the interval between the rectangular hole and the second rectangular hole; the first rectangular hole is another rectangular hole that is adjacent to the rectangular hole in the horizontal direction; the second rectangular hole is another rectangular hole that is adjacent to the rectangular hole in the vertical direction.

[0157] The machine model training unit 400 is used to train the target neural network model based on the numerical calculation results of the numerical calculation model of the target parameters as training data to obtain the optimal solution of the target parameters; the target neural network model is used to characterize the mapping relationship between the numerical value of the target parameters and the performance of the drill plate.

[0158] Optionally, the target neural network model includes a residual network and a particle swarm optimization module; the residual network is used to fit the mapping relationship; the particle swarm optimization module is used to obtain the optimal values ​​of the target parameters in the continuous space based on the mapping relationship through the particle swarm optimization algorithm.

[0159] Each of the units shown above can optimize the drilling scheme of the perforated plate in the filament sorting mechanism by utilizing the optimal solution of the target parameters.

[0160] This application also provides a computer-readable storage medium including a stored program, wherein the program executes the above-described method for optimizing the structure of the filament sorting mechanism provided in this application.

[0161] This application also provides an electronic device, including a processor, a memory, and a bus. The processor and the memory are connected via the bus. The memory is used to store a program, and the processor is used to run the program. During program execution, the above-described method for optimizing the structure of the filament sorting mechanism provided in this application is executed.

[0162] While several specific implementation details are included in the foregoing discussion, these should not be construed as limiting the scope of this application. Certain features described in the context of individual embodiments may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented individually or in any suitable sub-combination in multiple embodiments.

[0163] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A method for optimizing the structure of a stem sorting mechanism, characterized in that, include: Numerical calculation models for multiple borehole plate structural parameters are obtained based on the three-dimensional model of the filament sorting mechanism. Based on multiple borehole plate structural parameters and combined with wide-spacing horizontals, a first orthogonal experimental design space is constructed, and the average velocity, velocity uniformity index, and pressure loss of the airflow after passing through the borehole plate are used as evaluation indicators to determine sensitive parameters. Based on the aforementioned sensitive parameters and combined with the narrow spacing level, a second orthogonal experimental design space is constructed, and the target parameters are determined using particle retention capacity, velocity uniformity index, and pressure loss as evaluation indicators. The numerical calculation results of the numerical calculation model based on the target parameters are used as training data to train the target neural network model to obtain the optimal solution of the target parameters; the target neural network model is used to characterize the mapping relationship between the value of the target parameters and the performance of the borehole plate.

2. The method according to claim 1, characterized in that, Based on multiple borehole plate structural parameters and a wide-spacing horizontal plane, a first orthogonal experimental design space is constructed. The average velocity, velocity uniformity index, and pressure loss of the airflow after passing through the borehole plate are used as evaluation indicators to determine sensitive parameters, including: Based on the aforementioned three-dimensional model, and combined with computational fluid dynamics, a turbulence model of the filament sorting mechanism is generated; the turbulence model is used to simulate the flow field information of the internal channels of the filament sorting mechanism. Based on multiple drill plate structural parameters and combined with wide-spacing horizontals, a first orthogonal experimental design space is constructed. In the first orthogonal experimental design space, the turbulence model is run to obtain the first simulation result; the first simulation result includes the flow field information of the internal channel of the wire sorting mechanism under different values ​​of the various drill plate structural parameters; From the first simulation results, the average velocity, velocity uniformity index, and pressure loss of the airflow after passing through the perforated plate are extracted; Using the average velocity of the airflow after passing through the perforated plate, the velocity uniformity index, and the pressure loss as evaluation indicators, sensitive parameters are selected from the various structural parameters of the perforated plate.

3. The method according to claim 2, characterized in that, Using the average velocity of the airflow after passing through the perforated plate, the velocity uniformity index, and the pressure loss as evaluation indicators, sensitive parameters are selected from the various structural parameters of the perforated plate, including: The first evaluation index is obtained by weighting and summing the average velocity of the airflow after passing through the perforated plate, the velocity uniformity index, and the pressure loss. Based on the first evaluation index, at least one sensitive parameter is selected from the various drill plate structural parameters; wherein the first evaluation index of the sensitive parameter is higher than that of other drill plate structural parameters.

4. The method according to claim 1, characterized in that, Based on the aforementioned sensitive parameters and combined with the narrow spacing level, a second orthogonal experimental design space is constructed. Particle retention capacity, velocity uniformity index, and pressure loss are used as evaluation indicators to determine target parameters, including: Based on the turbulence model of the filament sorting mechanism, a two-phase coupled numerical calculation model is generated by combining the discrete element method. The two-phase coupled numerical calculation model is used to simulate the force information and flow field information of the internal channel of the filament sorting mechanism. The force information includes the force of the fluid on the particles. The fluid is used to characterize air, and the particles are used to characterize materials. Based on the aforementioned sensitive parameters, and combined with the narrow-spacing level, a second orthogonal experimental design space is constructed. In the second orthogonal experimental design space, the two-phase coupled numerical calculation model is run to obtain the second simulation results; the second simulation results include the force information and flow field information of the internal channel of the filament sorting mechanism under different values ​​of multiple sensitive parameters. From the second simulation results, particle retention capacity, velocity uniformity index, and pressure loss are extracted; the particle retention capacity is equal to... In the formula Characterizes the particle mass passing through the drilled plate. Characterizes the mass of particles retained by the drill plate; Using the particle retention capacity, the velocity uniformity index, and the pressure loss as evaluation indicators, target parameters are selected from the various sensitive parameters.

5. The method according to claim 4, characterized in that, Using the particle retention capacity, the velocity uniformity index, and the pressure loss as evaluation indicators, target parameters are selected from the various sensitive parameters, including: The second evaluation index is obtained by weighting and summing the particle retention capacity, the velocity uniformity index, and the pressure loss. Based on the second evaluation index, at least one target parameter is selected from the various sensitive parameters; wherein the second evaluation index of the target parameter is higher than that of the other sensitive parameters.

6. The method according to claim 1, characterized in that, The target neural network model includes a residual network and a particle swarm optimization module; the residual network is used to fit the mapping relationship; the particle swarm optimization module is used to obtain the optimal values ​​of the target parameters in the continuous space based on the mapping relationship using a particle swarm optimization algorithm.

7. The method according to any one of claims 1-6, characterized in that, The target parameters include opening shape parameters and position parameters; the opening shape parameters include the length and width of the rectangular hole; the position parameters include the interval between the rectangular hole and the first rectangular hole, and the interval between the rectangular hole and the second rectangular hole; the first rectangular hole is another rectangular hole that is adjacent to the rectangular hole in the horizontal direction; the second rectangular hole is another rectangular hole that is adjacent to the rectangular hole in the vertical direction.

8. A structural optimization device for a stem sorting mechanism, characterized in that, include: The computational model determination unit is used to obtain a numerical calculation model of multiple drill plate structural parameters based on the three-dimensional model of the wire sorting mechanism. The sensitive parameter screening unit is used to construct a first orthogonal experimental design space based on multiple borehole plate structural parameters and a wide-spacing horizontal plane, and to determine the sensitive parameters by using the average velocity, velocity uniformity index and pressure loss of the airflow after passing through the borehole plate as evaluation indicators. The target parameter screening unit is used to construct a second orthogonal experimental design space based on the sensitive parameters and the narrow spacing level, and to determine the target parameters using particle retention capacity, velocity uniformity index and pressure loss as evaluation indicators. The machine model training unit is used to train the target neural network model based on the numerical calculation results of the numerical calculation model of the target parameters as training data to obtain the optimal solution of the target parameters; the target neural network model is used to characterize the mapping relationship between the value of the target parameters and the performance of the borehole plate.

9. A storage medium, characterized in that, The storage medium includes a stored program, wherein the program is executed by a processor to perform the filament sorting mechanism structure optimization method according to any one of claims 1-7.

10. An electronic device, characterized in that, include: Processor, memory, and bus; The processor and the memory are connected via the bus; The memory is used to store a program, and the processor is used to run the program, wherein the program is executed by the processor to perform the filament sorting mechanism structure optimization method according to any one of claims 1-7.

Images