Automatic control system for cutter door pressure of tobacco cutter

Abstract

This invention discloses an automatic control system for the knife gate pressure of a shredder, comprising a data acquisition module, an automatic control module, a feedback adjustment module, and an adaptive module. The data acquisition module includes a data acquisition unit and a data processing unit. The automatic control module includes a model building unit, a model prediction unit, and an automatic control unit. The feedback adjustment module includes an effect detection unit and a feedback adjustment unit. The adaptive module includes a model self-learning unit. This invention can effectively realize the automatic control and adjustment of the knife gate pressure of a shredder, shorten the knife gate pressure adjustment time, solve the problem of uneven shredding width, reduce sliver slippage and sliver formation, and improve the stability of the quality of the shredded shreds, providing practical guidance for the shredding production process.

Description

Technical Field

[0001] This invention relates to the field of automatic control of tobacco processing equipment, and more specifically to an automatic control system for the pressure of the knife gate of a tobacco shredder. Background Technology

[0002] Shredding is a crucial process in the tobacco processing workshop. Its main task is to cut tobacco sheets that meet the incoming material requirements into shreds of a certain width to satisfy subsequent processing needs. The quality of the shredded tobacco has varying degrees of impact on the tobacco structure, physical properties of cigarettes, cigarette smoke composition, and sensory quality of cigarettes.

[0003] The shredder mainly consists of two parts: the feeding mechanism and the shredder gate. The feeding mechanism includes a gate clamping device and upper and lower copper busbar chains. The shredder gate includes the cutter head, sharpening system, and other components. The standard pressure for the gate is 17kN-21kN. This pressure is usually set as a fixed value based on equipment operating capacity and work experience. However, due to uneven moisture content and flow fluctuations in the incoming material, the quality of the shredded tobacco can vary. If the moisture content is too low, many tobacco shreds or broken pieces will exceed the standard width requirements; if the moisture content is too high, the tobacco will easily stick together, making it difficult to separate and resulting in clumps.

[0004] Currently, referring to Figure 1 The setting and adjustment of the knife gate pressure of the shredder mainly rely on the operator's work experience and sensory evaluation. The initial knife gate pressure is set based on the experience accumulated over a long period of time for the production grade. After the shredder starts shredding, the quality of the shredded blades, broken pieces, or capillary filaments is evaluated by sensory evaluation, and the knife gate pressure is adjusted accordingly. Because the shredder has a large production flow and produces many grades every day, adjusting the knife gate pressure by manual experience not only has a lag, but also has a certain degree of error in sensory evaluation.

[0005] In view of the above, the present invention will establish an automatic control system and method for the knife gate pressure of a shredder, which predicts the knife gate pressure based on the moisture and temperature of the incoming material, and detects and provides feedback on the quality of the shredded leaves, thereby realizing the automatic control and adjustment of the knife gate pressure of the shredder. Summary of the Invention

[0006] To overcome the above problems, this invention provides an automatic control system for the knife gate pressure of a shredder, which can effectively shorten the adjustment time of the knife gate pressure, solve the problem of uneven shredding width, reduce shredding and stripping, improve the stability of the quality of the shredded shreds, and have practical guiding significance for the shredding production process.

[0007] The technical solution adopted by this invention to solve its technical problem is as follows:

[0008] Automatic control system for knife gate pressure of shredder, including

[0009] The data acquisition module includes a data acquisition unit and a data processing unit. The data acquisition unit is used to screen and acquire process parameters that affect the moisture content and temperature at the hot air leaf outlet. The data processing unit is used to standardize the acquired data.

[0010] The automatic control module includes a model building unit, a model prediction unit, and an automatic control unit. The model building unit is used to build a prediction model for the moisture content and temperature at the hot air leaf moistening outlet based on the collected data. The model prediction unit is used to predict the moisture content and temperature at the hot air leaf moistening outlet based on the established prediction model and production data, and to provide control data to the automatic control unit. The automatic control unit is used to build an automatic control model for the knife gate pressure based on the incoming material status, and to control the knife gate pressure based on the established control model and control data.

[0011] The feedback adjustment module includes an effect detection unit and a feedback adjustment unit. The effect detection unit is used to detect the quality of the cut leaf filaments, and the feedback adjustment unit generates feedback adjustment data based on the detection results and feeds it back to the automatic control unit.

[0012] As an improvement to the above technical solution, the system also includes an adaptive module, including a model self-learning unit. The prediction model and control model self-learn and adjust according to changes in the raw materials of the production grade, the parameters of the shredder, etc., and continuously adjust the prediction and control accuracy of the model to meet the production requirements of the shredder.

[0013] As an improvement to the above technical solution, the process parameters screened and collected by the data acquisition unit include the moisture content at the loose rehydration inlet, the temperature at the loose rehydration inlet, the total water volume of loose rehydration, the moisture content at the loose rehydration outlet, the temperature at the loose rehydration outlet, the moisture content at the blade feeding inlet, the moisture content at the blade feeding outlet, the temperature at the blade feeding outlet, the moisture content at the hot air moistening inlet, the temperature at the hot air moistening inlet, and the opening degree of the hot air moistening compensating steam valve.

[0014] As an improvement to the above technical solution, the hot air leaf-humidifying outlet moisture and outlet temperature prediction model is a neural network prediction model, which is constructed by using the process parameters selected by the data acquisition unit as the model input layer and the hot air leaf-humidifying outlet moisture and outlet temperature as the model output layer.

[0015] The neural network prediction model is set with a training target of 0.05, a training speed of 0.01, and a maximum number of steps of 100.

[0016] As an improvement to the above technical solution, the automatic control model for the blade gate pressure obtains the optimal control parameters for the blade gate pressure through a response surface methodology test. This response surface methodology test sets the blade filament qualification rate as the dependent variable y, and sets the hot air leaf moistening outlet moisture, outlet temperature, and blade gate pressure as factors x, conducting a three-factor, two-level test.

[0017] The response surface methodology uses the blade pass rate as the response variable. The control model is set to include the main effects of all factors and the second-order interaction effects. The response surface methodology (three factors and two levels) is used to conduct a variance analysis on the blade pass rate of hot air leaf humidification outlet temperature, outlet moisture and knife gate pressure. That is, a variance analysis of factors on the blade pass rate is conducted to obtain the regression equations of outlet temperature (x1), outlet moisture (x2) and knife gate pressure (x3) on the blade pass rate (y).

[0018] Based on the regression equation, a contour plot of the response surface is drawn, and suitable control parameters are obtained through response surface optimization.

[0019] The regression equation is:

[0020]

[0021] As an improvement to the above technical solution, the effect detection unit includes a hyperspectral detector installed at the lifting belt of the shredder, which performs quality detection on the shredded filaments after cutting using a hyperspectral filament state verification model.

[0022] The hyperspectral leaf filament state verification model verifies the leaf filament state based on spectral data through imaging scanning, spectral data correction and extraction, and spectral characteristic analysis.

[0023] The imaging scan performs spectral scanning on the cut leaf filaments using a line scanning method. The leaf filament samples (including standard width leaf filaments, strip leaf filaments, parallel strip leaf filaments, and capillary filaments, etc.) are placed on a moving platform and moved along the belt direction. A hyperspectral detector continuously scans and records a line on the leaf filament sample image to achieve real-time online non-destructive monitoring of the cut leaf filaments.

[0024] The spectral data correction and extraction are used to perform black-and-white correction on the spectral images of the leaf filament samples;

[0025] The spectral characteristic analysis is used to extract the spectral and texture features of the leaf filament samples, and to match the hyperspectral data of the samples with the corresponding sample types (standard width leaf filaments, strip leaf filaments, parallel strip leaf filaments, and capillary filaments) and to determine the leaf filament qualification rate.

[0026] As an improvement to the above technical solution, the feedback adjustment data includes:

[0027] When the qualified rate of tobacco leaves is ≥85.0%, the knife gate pressure setting is reasonable; when the qualified rate of tobacco leaves is <85.0%, the knife gate pressure setting is unreasonable: when the amount of tobacco leaves in the combined strips is greater than the amount of tobacco leaves in the loose strips, the knife gate pressure should be reduced by at least 1kN; when the amount of tobacco leaves in the combined strips is less than the amount of tobacco leaves in the loose strips, the knife gate pressure should be increased by at least 1kN.

[0028] The beneficial effects of this invention are as follows:

[0029] This invention predicts the moisture and temperature of the incoming blades during the hot air moistening process before shredding, establishing a control model for the incoming blades and blade pressure based on the moisture and temperature of the incoming material. Physical testing of the blade quality is performed at the shredder outlet to provide feedback on the shredder blade pressure, enabling automatic control and adjustment of the shredder blade pressure. This effectively shortens the adjustment time for the shredder blade pressure, solves the problem of uneven shredding width, reduces blade slippage and sliver formation, and improves the stability of the shredded blade quality, providing practical guidance for the shredding production process.

[0030] The automatic control model for knife gate pressure of the present invention obtains the optimal control parameters for knife gate pressure through response surface experiments. Only a small number of experiments are needed to obtain the optimal control parameters for the target blade filament qualification rate. This can effectively improve the automation level of knife gate pressure control and adjustment of the shredder and increase the qualification rate of the shredder blades after cutting.

[0031] This invention can effectively realize the automatic control and adjustment of the knife gate pressure of the shredder, shorten the adjustment time of the knife gate pressure, solve the problem of uneven shredding width, reduce the running and shaving of the shreds, improve the stability of the quality of the shredded leaves, and have practical guiding significance for the shredding production process. Attached Figure Description

[0032] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0033] Figure 1 This is a schematic diagram of a traditional knife gate pressure control process;

[0034] Figure 2 This is a schematic diagram of the knife gate pressure control process of the present invention;

[0035] Figure 3 This is a schematic diagram of the system modules of the present invention;

[0036] Figure 4 A schematic diagram of the structure of the hot air leaf humidification outlet moisture and outlet temperature prediction model;

[0037] Figure 5 This is a schematic diagram of the results of variance analysis based on the filament qualification rate;

[0038] in Figure 5(a), (b), (c), and (d) are respectively: normal probability plot, histogram, fitted value plot, and residual plot of the leaf filament qualification rate;

[0039] Figure 6 This is a contour plot of the leaf fiber qualification rate versus outlet moisture and outlet temperature in Example 3.

[0040] Figure 7 This is a contour plot of the blade wire qualification rate versus blade pressure and outlet temperature in Example 3;

[0041] Figure 8 This is a contour plot of the blade filament qualification rate, blade pressure, and outlet moisture content in Example 3.

[0042] For ease of understanding, Figure 6 , 7 The color order from left to right shown corresponds to the order from top to bottom of the leaf filament pass rate interval in the upper right corner of the figure; Figure 8 The color intensity from bottom to top shown corresponds to the leaf filament pass rate interval in the upper right corner of the figure from top to bottom. Detailed Implementation

[0043] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments and accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0044] Example 1

[0045] Reference Figures 2-3 Automatic control system for knife gate pressure of shredder, including

[0046] The data acquisition module includes a data acquisition unit and a data processing unit. The data acquisition unit is used to screen and acquire process parameters that affect the moisture content and temperature at the hot air leaf outlet. The data processing unit is used to standardize the acquired data to remove the influence of data dimensions.

[0047] The automatic control module includes a model building unit, a model prediction unit, and an automatic control unit. The model building unit is used to build a prediction model for the moisture content and temperature at the hot air leaf moistening outlet based on the collected data. The model prediction unit is used to predict the moisture content and temperature at the hot air leaf moistening outlet based on the established prediction model and production data, and to provide control data to the automatic control unit. The automatic control unit is used to build an automatic control model for the knife gate pressure based on the incoming material status, and to control the knife gate pressure based on the established control model and control data.

[0048] The feedback adjustment module includes an effect detection unit and a feedback adjustment unit. The effect detection unit is used to detect the quality of the cut leaf filaments, and the feedback adjustment unit generates feedback adjustment data based on the detection results and feeds it back to the automatic control unit.

[0049] The adaptive module includes a model self-learning unit. The prediction model and control model self-learn and adjust according to changes in raw material grades and shredder parameters, continuously adjusting the model prediction and control accuracy to meet the production requirements of the shredder.

[0050] In this embodiment, the process parameters screened and collected by the data acquisition unit include the moisture content at the loose rehydration inlet, the temperature at the loose rehydration inlet, the total water volume for loose rehydration, the moisture content at the loose rehydration outlet, the temperature at the loose rehydration outlet, the moisture content at the blade feeding inlet, the moisture content at the blade feeding outlet, the temperature at the blade feeding outlet, the moisture content at the hot air moistening inlet, the temperature at the hot air moistening inlet, and the opening degree of the hot air moistening compensating steam valve.

[0051] In this embodiment, refer to Figure 4 The hot air leaf-humidifying outlet moisture and outlet temperature prediction model is a neural network prediction model, which is constructed by using the process parameters selected by the data acquisition unit as the model input layer and the hot air leaf-humidifying outlet moisture and outlet temperature as the model output layer.

[0052] The neural network prediction model is set with a training target of 0.05, a training speed of 0.01, and a maximum number of steps of 100.

[0053] In this embodiment, the automatic control model for the blade gate pressure obtains the optimal control parameters for the blade gate pressure through a response surface methodology test. The response surface methodology test sets the blade filament qualification rate as the dependent variable y, and sets the hot air leaf moistening outlet moisture, outlet temperature, and blade gate pressure as factors x, and conducts a three-factor, two-level test.

[0054] The response surface methodology uses the blade pass rate as the response variable. The control model is set to include the main effects of all factors and the second-order interaction effects. The response surface methodology (three factors and two levels) is used to conduct a variance analysis on the blade pass rate of hot air leaf humidification outlet temperature, outlet moisture and knife gate pressure. That is, a variance analysis of factors on the blade pass rate is conducted to obtain the regression equations of outlet temperature (x1), outlet moisture (x2) and knife gate pressure (x3) on the blade pass rate (y).

[0055] Based on the regression equation, a contour plot of the response surface is drawn, and suitable control parameters are obtained through response surface optimization.

[0056] The regression equation is:

[0057]

[0058] In this embodiment, the effect detection unit includes a hyperspectral detector installed at the lifting belt of the shredder. It performs quality detection on the shredded filaments after cutting using a hyperspectral filament state verification model. The detection items include qualified filaments, slivers, and run-off filaments.

[0059] The hyperspectral leaf filament state verification model verifies the leaf filament state based on spectral data through imaging scanning, spectral data correction and extraction, and spectral characteristic analysis.

[0060] The imaging scan performs spectral scanning on the cut leaf filaments using a line scanning method. The leaf filament samples (including standard width leaf filaments, strip leaf filaments, parallel strip leaf filaments, and capillary filaments, etc.) are placed on a moving platform and moved along the belt direction. A hyperspectral detector continuously scans and records a line on the leaf filament sample image to achieve real-time online non-destructive monitoring of the cut leaf filaments.

[0061] The spectral data correction and extraction are used to perform black-and-white correction on the spectral images of the leaf filament samples;

[0062] The spectral characteristic analysis is used to extract the spectral and texture features of the leaf filament samples, and to match the hyperspectral data of the samples with the corresponding sample types (standard width leaf filaments, strip leaf filaments, parallel strip leaf filaments, and capillary filaments) and obtain the leaf filament qualification rate.

[0063] In this embodiment, the feedback adjustment data includes:

[0064] When the qualified rate of tobacco leaves is ≥85.0%, the knife gate pressure setting is reasonable; when the qualified rate of tobacco leaves is <85.0%, the knife gate pressure setting is unreasonable: when the amount of tobacco leaves in the combined strips is greater than the amount of tobacco leaves in the loose strips, the knife gate pressure should be reduced by at least 1kN; when the amount of tobacco leaves in the combined strips is less than the amount of tobacco leaves in the loose strips, the knife gate pressure should be increased by at least 1kN.

[0065] Example 2

[0066] Reference Figure 2 Automatic control method for knife gate pressure of shredder, including

[0067] S1 Data Acquisition

[0068] The process parameters affecting the moisture content and temperature at the hot air leaf outlet were screened and collected, and the collected data were standardized to remove the influence of data dimensions.

[0069] S2 Establish a prediction model

[0070] A prediction model for the moisture content and temperature at the outlet of hot air leaf humidification was established using historical production data. Based on the established prediction model and real-time production data, the moisture content and temperature at the outlet of hot air leaf humidification were predicted.

[0071] S3 establishes a control model

[0072] An automatic control model for the knife gate pressure based on the incoming material status is established, and the automatic control of the knife gate pressure is performed based on the established control model and S2 prediction data; when the moisture and temperature of the incoming material blades change, the knife gate pressure of the shredder will be adjusted accordingly.

[0073] S4 Feedback Adjustment

[0074] The quality of the cut leaf shreds is inspected. Based on the inspection results, it is determined whether the knife gate pressure setting is reasonable and meets the shredding requirements. Feedback adjustment suggestions are generated and fed back to S3 for optimization.

[0075] S5 model self-learning

[0076] The prediction and control models learn and adjust themselves based on changes in raw material grade (shredding width process standard), shredder parameters, etc., continuously adjusting the model prediction and control accuracy to meet the shredder production requirements.

[0077] In step S1, the data acquisition, screening, and acquisition of process parameters include the moisture content at the loosening and rehydration inlet, the temperature at the loosening and rehydration inlet, the total water volume for loosening and rehydration, the moisture content at the loosening and rehydration outlet, the temperature at the loosening and rehydration outlet, the moisture content at the blade feeding inlet, the moisture content at the blade feeding outlet, the temperature at the blade feeding outlet, the moisture content at the hot air moistening inlet, the temperature at the hot air moistening inlet, and the opening degree of the hot air moistening compensating steam valve.

[0078] In step S2, the prediction model for the moisture content and temperature at the hot air leaf humidification outlet is a neural network prediction model, which is constructed by using the process parameters selected by the data acquisition unit as the model input layer and the moisture content and temperature at the hot air leaf humidification outlet as the model output layer.

[0079] The neural network prediction model has a training target of 0.05, a training speed of 0.01, and a maximum number of steps of 100; the hidden layer activation function is hyperbolic tangent, and the output layer activation function is identity.

[0080] In step S3, the automatic control model for knife gate pressure obtains the optimal control parameters for knife gate pressure through response surface methodology. The response surface methodology sets the blade qualification rate as the dependent variable y, and sets the hot air leaf outlet moisture, outlet temperature, and knife gate pressure as factors x, and conducts a three-factor, two-level experiment.

[0081] The response surface methodology uses the blade qualification rate as the response variable. The control model is set to include the main effects of all factors and the second-order interaction effects. The response surface methodology (three factors and two levels) is used to conduct variance analysis on the blade qualification rate of hot air leaf humidification outlet temperature, outlet moisture and knife gate pressure. The regression equations of outlet temperature (x1), outlet moisture (x2) and knife gate pressure (x3) on the blade qualification rate (y) are obtained.

[0082] Based on the regression equation, a contour plot of the response surface is drawn, and suitable control parameters are obtained through response surface optimization.

[0083] In step S4, the quality of the cut leaf filaments is tested using a hyperspectral leaf filament state verification model:

[0084] The hyperspectral leaf filament state verification model verifies the leaf filament state based on spectral data through imaging scanning, spectral data correction and extraction, and spectral characteristic analysis.

[0085] The imaging scan performs spectral scanning on the cut leaf filaments using a line scanning method. The leaf filament samples (including standard width leaf filaments, strip leaf filaments, parallel strip leaf filaments, and capillary filaments, etc.) are placed on a moving platform and moved along the belt direction. A hyperspectral detector continuously scans and records a line on the leaf filament sample image to achieve real-time online non-destructive monitoring of the cut leaf filaments.

[0086] The spectral data correction and extraction are used to perform black-and-white correction on the spectral images of the leaf filament samples;

[0087] The spectral characteristic analysis is used to extract the spectral and texture features of the leaf filament samples, and to match the hyperspectral data of the samples with the corresponding sample types (standard width leaf filaments, strip leaf filaments, parallel strip leaf filaments, and capillary filaments) and obtain the leaf filament qualification rate.

[0088] In step S4, the feedback adjustment suggestions include:

[0089] When the qualified rate of tobacco leaves is ≥85.0%, the knife gate pressure setting is reasonable; when the qualified rate of tobacco leaves is <85.0%, the knife gate pressure setting is unreasonable: when the amount of tobacco leaves in the combined strips is greater than the amount of tobacco leaves in the loose strips, the knife gate pressure is reduced by n×1kN; when the amount of tobacco leaves in the combined strips is less than the amount of tobacco leaves in the loose strips, the knife gate pressure is increased by n×1kN; where n is a natural number ≥1. This step is repeated until the qualified rate of tobacco leaves is ≥85.0%.

[0090] Example 3

[0091] Establishment of an automatic control model for knife gate pressure.

[0092] A response surface methodology experiment was conducted using a three-factor, two-level approach. Specifically, the blade pass rate was set as the dependent variable y, and the hot air moisture content at the leaf outlet, outlet temperature, and knife gate pressure were set as factors x. There were 6 center points.

[0093] Examples of parameter settings are shown in Table 1:

[0094] Table 1

[0095] Serial Number parameter low level High level 1 Outlet temperature (°C) 30.0 40.0 2 Export moisture content (%) 21.1 22.1 3 Knife gate pressure (kN) 17 21

[0096] The parameters for the response surface methodology are shown in Table 2.

[0097] Table 2

[0098] Serial Number Outlet temperature (°C) Export moisture content (%) Knife gate pressure (kN) Leaf fiber qualification rate (%) 1 30 21.1 19 85.4 2 40 21.1 19 86.3 3 30 22.1 19 85.9 4 40 22.1 19 86.7 5 30 21.6 17 86.9 6 40 21.6 17 87.2 7 30 21.6 21 86.2 8 40 21.6 21 87.5 9 35 21.1 17 85.6 10 35 22.1 17 86.7 11 35 21.1 21 86.0 12 35 22.1 21 86.2 13 35 21.6 19 85.9 14 35 21.6 19 86.4 15 35 21.6 19 86.6 16 35 21.6 19 86.9

[0099] Using the leaf filament qualification rate as the response variable, the model was designed to include the main effects of all factors and second-order interaction effects, as referenced... Figure 5 Response surface methodology (three-factor, two-level) experiments were conducted, and variance analysis was performed on the effect of hot air humidification outlet temperature, outlet moisture content, and blade gate pressure on the blade qualification rate. The results are shown in Table 3.

[0100] Table 3

[0101]

[0102]

[0103] The model term p-value is 0.014, which is less than 0.05, indicating that the current model is generally effective.

[0104] The regression equations for the blade yield (y) obtained through response surface methodology were obtained from the hot air humidification outlet temperature (x1), outlet moisture content (x2), and blade gate pressure (x3):

[0105]

[0106] The main effects of hot air humidifier outlet temperature and knife gate pressure in the regression equation are significant, and the R-sq value of the regression equation error as a percentage of the total error is 80.05%, indicating that the model is good.

[0107] Based on the regression equation, a contour plot of the response surface can be drawn, and the optimal control parameters can be obtained through response surface optimization.

[0108] Figures 6-8 The figure shows the contour plot of the response surface when the gate pressure is 19kN, the outlet moisture content is 21.6%, and the outlet temperature is 35℃. According to the response surface optimizer, the optimal process parameters are: outlet moisture content 22.1%, outlet temperature 43℃, and gate pressure 21kN. At this time, the blade qualification rate can reach 87.5%.

[0109] It should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An automatic control system for the blade pressure of a shredder, characterized in that: include The data acquisition module includes a data acquisition unit and a data processing unit. The data acquisition unit is used to screen and acquire process parameters that affect the moisture content and temperature at the hot air leaf outlet. The data processing unit is used to standardize the acquired data. The automatic control module includes a model building unit, a model prediction unit, and an automatic control unit. The model building unit is used to build a prediction model for the moisture content and temperature at the hot air leaf moistening outlet based on the collected data. The model prediction unit is used to predict the moisture content and temperature at the hot air leaf moistening outlet based on the established prediction model and production data, and to provide control data to the automatic control unit. The automatic control unit is used to build an automatic control model for the knife gate pressure based on the incoming material status, and to control the knife gate pressure based on the established control model and control data. The feedback adjustment module includes an effect detection unit and a feedback adjustment unit. The effect detection unit is used to detect the quality of the cut leaf filaments, and the feedback adjustment unit generates feedback adjustment data based on the detection results and feeds it back to the automatic control unit. The automatic control model for the knife gate pressure obtains the optimal control parameters for the knife gate pressure through response surface methodology; this response surface methodology sets the blade wire qualification rate as the dependent variable. The hot air leaf humidification outlet moisture, outlet temperature, and blade gate pressure were set as factors. A three-factor, two-level experiment was conducted. The response surface methodology experiment used the blade filament qualification rate as the response variable. A control model was set that included the main effects and second-order interaction effects of all factors. A variance analysis of the factor versus the blade filament qualification rate was performed to obtain the outlet temperature. Export moisture and knife gate pressure Leaf filament qualification rate The regression equation; Based on the regression equation, a contour plot of the response surface is drawn, and the control parameters are obtained through response surface optimization. The regression equation is: ； The effect detection unit includes a hyperspectral detector installed at the lifting belt of the shredder, which performs quality detection on the shredded leaf filaments after cutting through a hyperspectral leaf filament state verification model. The feedback adjustment data includes: When the qualified rate of tobacco leaves is ≥85.0%, the knife gate pressure setting is reasonable; when the qualified rate of tobacco leaves is <85.0%, the knife gate pressure setting is unreasonable: when the amount of tobacco leaves in the bundled strips is greater than the amount of tobacco leaves in the loose strips, the knife gate pressure needs to be reduced; when the amount of tobacco leaves in the bundled strips is less than the amount of tobacco leaves in the loose strips, the knife gate pressure needs to be increased.

2. The automatic control system for knife gate pressure of the shredder according to claim 1, characterized in that: It also includes an adaptive module, including a model self-learning unit. The prediction model and control model learn and adjust themselves based on changes in the raw materials and the parameters of the shredder, continuously adjusting the model prediction and control accuracy to meet the production requirements of the shredder.

3. The automatic control system for knife gate pressure of the shredder according to claim 1, characterized in that: The process parameters screened and collected by the data acquisition unit include the moisture content at the loose rehydration inlet, the temperature at the loose rehydration inlet, the total water volume for loose rehydration, the moisture content at the loose rehydration outlet, the temperature at the loose rehydration outlet, the moisture content at the blade feeding inlet, the moisture content at the blade feeding outlet, the temperature at the blade feeding outlet, the moisture content at the hot air moistening inlet, the temperature at the hot air moistening inlet, and the opening degree of the hot air moistening compensating steam valve.

4. The automatic control system for knife gate pressure of the shredder according to claim 1, characterized in that: The prediction model for the moisture content and temperature at the hot air leaf humidification outlet is a neural network prediction model. It is constructed by using the process parameters selected by the data acquisition unit as the model input layer and the moisture content and temperature at the hot air leaf humidification outlet as the model output layer.

5. The automatic control system for knife gate pressure of the shredder according to claim 1, characterized in that: The hyperspectral leaf filament state verification model verifies the leaf filament state based on spectral data through imaging scanning, spectral data correction and extraction, and spectral characteristic analysis. The imaging scan performs spectral scanning on the cut leaf filaments using a line scanning method. The leaf filament sample is placed on a moving platform and moved along the belt direction. A hyperspectral detector continuously scans and records a line on the leaf filament sample image, enabling real-time online non-destructive monitoring of the cut leaf filaments. The spectral data correction and extraction are used to perform black-and-white correction on the spectral images of the leaf filament samples; The spectral characteristic analysis is used to extract the spectral and texture features of the leaf filament samples, and to match the hyperspectral data of the samples with the corresponding sample types to obtain the leaf filament qualification rate.

Images