A method and system for calculating the silk content in a leaf silk wind-separation stem tagging system

Abstract

This invention discloses a method and system for calculating the fiber content of a leaf-tobacco wind-separation and stem-tobacco strip system in the field of tobacco processing parameter detection technology. The method includes acquiring historical operating data and corresponding manually calibrated fiber content values, constructing a training dataset, and fitting parameters to a fiber content calculation model based on the training dataset to obtain a regression model. Subsequently, the operating parameters of the wind-separation system are acquired, and normalized characteristic variables such as wind speed normalization coefficient, stem-tobacco strip production normalization coefficient, and negative pressure normalization coefficient are constructed. These normalized characteristic variables are then input into the regression model to calculate the real-time fiber content. The real-time fiber content is output to the wind-separation system control unit, and the operating parameters of the wind-separation system are adjusted based on the real-time fiber content. This invention enables real-time calculation and dynamic updating of the fiber content in a leaf-tobacco wind-separation and stem-tobacco strip system.

Description

Technical Field

[0001] This invention relates to a method and system for calculating the filament content in a leaf filament wind separation and stem labeling system, belonging to the technical field of tobacco processing parameter detection. Background Technology

[0002] In the cigarette manufacturing process, the leaf air separation process is the core step in removing impurities such as stems and clumps from the leaf shreds. Its working principle utilizes the difference in suspension velocity between stems and qualified leaf shreds; airflow transports the qualified leaf shreds to subsequent processes, while the denser stems settle and are collected. Inevitably, some qualified leaf shreds will be separated along with the stems during air separation. If the separated stems have an excessively high shred content, it will lead to raw material waste and increased production costs; if the shred content is too low, it indicates that the airflow intensity is too high, which may result in some small stems not being effectively separated, affecting the smoking quality of subsequent cigarette products. Therefore, the shred content of the stems is a core indicator for measuring the performance of the air separation system and balancing raw material loss and impurity removal efficiency.

[0003] Currently, the industry commonly uses manual offline sampling to detect the fiber content of stems: samples are periodically taken at the stem outlet of the air separation system, qualified leaf fibers are manually picked out and weighed, and the proportion of leaf fiber weight to the total sample weight is calculated as the fiber content. This method has significant drawbacks: First, it has a strong detection lag, as the sampling, picking, and weighing process usually takes tens of minutes, making it impossible to reflect the real-time operating status of the air separation system and difficult to guide timely adjustments to air separation parameters; second, it has high labor costs, requiring dedicated personnel to conduct sampling and testing periodically, and the picking accuracy is greatly affected by personnel experience and working conditions, resulting in insufficient stability of the test results; third, it cannot achieve full-time coverage, as fluctuations in fiber content during sampling intervals cannot be captured, easily leading to the omission of unqualified operating conditions.

[0004] Although some studies have attempted to identify leaf filaments in stems using online detection methods such as machine vision and near-infrared detection, these methods have high requirements for the detection environment. Stacking and occlusion of leaf filaments and stems can easily cause recognition errors, and the recognition accuracy of a single image is usually less than 80%, which cannot meet the production precision requirements. Near-infrared detection equipment needs to be calibrated regularly with standard samples and is easily affected by fluctuations in leaf filament moisture. The equipment has high deployment costs and is difficult to maintain, making it difficult to promote its application in large-scale production. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the existing technology. In the existing tobacco processing process, the leaf filament air separation and stem tag system has the technical problem that it cannot reflect the operating status of the air separation system in real time and it is difficult to dynamically adjust the operating parameters of the air separation system according to the changes in the filament content.

[0006] To solve the above-mentioned technical problems, the present invention is implemented using the following technical solution:

[0007] In the first aspect, a method for calculating the fiber content of a leaf fiber wind separation stem system is provided, including calculating the adjustment cycle, outputting the real-time fiber content obtained according to the calculated adjustment cycle to the wind separation system control unit, and adjusting the operating parameters of the wind separation system according to the real-time fiber content;

[0008] The calculation of the adjustment cycle includes the following steps:

[0009] Obtain historical operating data and corresponding manually calibrated silk content values, construct a training dataset, and perform parameter fitting on the silk content calculation model based on the training dataset to obtain a regression model;

[0010] Obtain the operating parameters of the air separation system, and construct normalized feature variables based on the operating parameters of the air separation system; the normalized feature variables include the wind speed normalization coefficient, the stem yield normalization coefficient, and the negative pressure normalization coefficient;

[0011] Input the constructed normalized feature variables into the regression model to calculate the real-time silk content;

[0012] in,

[0013] The regression model further includes: setting an update period between adjacent calculation adjustment periods, the update period including the following steps:

[0014] Obtain the calibrated fiber content value of manually sampled data within the currently completed calculation adjustment cycle and the average value of the wind separation system operating parameters for the corresponding time period, and re-input them into the training dataset to refit and update the regression model.

[0015] Furthermore, the construction of the regression model includes:

[0016] A training dataset is constructed based on the collected historical operating data and manually calibrated silk content values, and the parameters of the silk content calculation model are fitted based on the training dataset to obtain a regression model;

[0017] in,

[0018] The regression model uses the least squares method for parameter fitting, and the goodness of fit is... Not less than 0.9.

[0019] Furthermore, the wind speed normalization coefficient Calculate using the following formula:

[0020] ;

[0021] Where v is the real-time wind speed. The preset wind speed for the corresponding blade grade.

[0022] Furthermore, the normalization coefficient of the stem yield Calculate using the following formula:

[0023] ;

[0024] in This refers to the real-time total mass flow rate at the outlet of the stem label. This is the preset theoretical yield benchmark value for the corresponding grade of leaf filaments.

[0025] Furthermore, the negative pressure normalization coefficient Calculate using the following formula:

[0026] ;

[0027] Where p is the real-time negative pressure value of the air separation chamber. Set a rated negative pressure value for the air separation chamber.

[0028] Furthermore, the regression model calculates the real-time silk content using the following formula. :

[0029] ;

[0030] in, This is the wind speed normalization coefficient. The normalization coefficient for stem yield is . is the negative pressure normalization coefficient, w is the real-time value of leaf fiber moisture content, a, b, c, and d are all weighting coefficients, and k is a correction constant.

[0031] Furthermore, the operating parameters of the air separation system include the inlet flow rate of the air separation chamber blades, the moisture content of the blades, the air velocity of the air separation airflow, the total mass flow rate of the stems at the outlet, and the negative pressure value of the air separation chamber.

[0032] Secondly, a leaf-fiber wind-separation stem-tagging system is provided, employing the fiber content calculation method of the leaf-fiber wind-separation stem-tagging system described in the first aspect, including:

[0033] Electronic belt scale: used to collect the flow rate of blades at the inlet of the air separator;

[0034] Online moisture meter: used to collect leaf filament moisture content;

[0035] Wind speed sensor: used to collect wind speed data of airflow in wind-selected air;

[0036] Weighing device: used to collect the total mass flow rate at the stem outlet;

[0037] Negative pressure sensor: used to collect negative pressure values ​​in the air separation chamber;

[0038] The PLC controller is used to connect with electronic belt scales, online moisture meters, wind speed sensors, weighing devices and negative pressure sensors, and to receive the operating parameters of the wind separation system collected by each sensor. It also calls the regression model and executable program stored in the memory to calculate the normalized characteristic variables in order to output the real-time fiber content.

[0039] The damper actuator is used to connect to the PLC controller and adjust the damper opening according to the control commands output by the PLC controller.

[0040] The fan frequency converter control unit is used to connect to the PLC controller and adjust the fan operating frequency according to the control commands output by the PLC controller.

[0041] Furthermore, there are 10 wind speed sensors, which are respectively set at 10 sampling points, and the wind speed of the wind-selected airflow is determined based on the average value of the data from the 10 sampling points.

[0042] The sampling points are evenly distributed across the cross-section of the air outlet of the air separator.

[0043] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0044] 1. This invention constructs a training dataset by acquiring historical operating data and corresponding manually calibrated silk content values, and then fits the parameters of the silk content calculation model based on the training dataset to establish a regression model. By acquiring the operating parameters of the wind-separation system, normalized feature variables such as wind speed normalization coefficient, stem yield normalization coefficient, and negative pressure normalization coefficient are constructed. The normalized feature variables are input into the regression model to calculate the real-time silk content, and the operating parameters of the wind-separation system are adjusted according to the real-time silk content. At the same time, the regression model is refitted and updated by combining the manually sampled calibration silk content values ​​in the current cycle, thereby realizing the real-time calculation and dynamic correction of the silk content of the leaf silk wind-separation stem system.

[0045] 2. This invention does not require additional online detection equipment. It can realize real-time calculation of the fiber content by only using the operating parameters collected by the existing sensors of the air separation system. The deployment cost is low and it is compatible with the air separation system of most existing tobacco processing lines without the need for structural modification of existing equipment. At the same time, the calculation results are highly real-time, and the data update frequency is consistent with the existing sensor acquisition frequency, which can realize the fiber content feedback at the second level. This makes it easy for the control system to adjust parameters such as air separation speed and negative pressure in a timely manner to balance the impurity removal efficiency and raw material loss.

[0046] 3. This invention updates model parameters through periodic manual calibration data iteration, which can effectively adapt to changes in operating conditions such as fluctuations in the characteristics of raw materials and equipment wear, ensuring the long-term accuracy of calculation results. The error between the calculated value and the manually calibrated value can be controlled within 0.5%, meeting the needs of production management and control. Attached Figure Description

[0047] Figure 1 The diagram shown is a schematic diagram of a method for calculating the silk content of a leaf silk wind-separation stem tag system provided in an embodiment of the present invention;

[0048] Figure 2 The diagram shown is a schematic diagram of the module structure of a leaf filament wind-sorting stem tagging system provided in an embodiment of the present invention.

[0049] In the picture:

[0050] 201. Air separation chamber; 202. Electronic belt scale; 203. Online moisture meter; 204. Wind speed sensor; 205. Negative pressure sensor; 206. Weighing device; 207. PLC controller; 208. Memory; 209. Damper actuator. Detailed Implementation

[0051] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments and specific features in the embodiments are detailed descriptions of the technical solution of the present application, rather than limitations thereof. In the absence of conflict, the technical features in the embodiments and embodiments can be combined with each other.

[0052] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0053] Example 1:

[0054] This embodiment provides a method for calculating the fiber content of a leaf fiber wind-separation stem tag system, including calculating the adjustment cycle, which includes the following steps:

[0055] Obtain historical operating data and corresponding manually calibrated silk content values, construct a training dataset, and perform parameter fitting on the silk content calculation model based on the training dataset to obtain a regression model;

[0056] Obtain the operating parameters of the air separation system, and construct normalized feature variables based on the operating parameters of the air separation system; the normalized feature variables include the wind speed normalization coefficient, the stem yield normalization coefficient, and the negative pressure normalization coefficient;

[0057] In this embodiment, the implementation object is the leaf shred air separation system of a cigarette factory's tobacco processing line, and the shred content is calculated for a certain brand of flue-cured tobacco leaves. The specific steps are as follows:

[0058] Pre-constructed training dataset: Historical operating data of the blades of this brand were collected, with a total of 80 valid samples. Each sample corresponds to the average value of the operating parameters of the stable production section for 10 minutes, including the blade flow rate at the inlet of the air classifier, blade moisture content, air classifier air velocity, total mass flow rate at the stem outlet, negative pressure value of the air classifier, and the manual detection filament content calibration value of the stem for each sample.

[0059] The wind speed corresponding to this blade grade is the wind selection standard. The theoretical yield benchmark value for stem cuttings is set at 12 m / s. The rated negative pressure value of the air separation chamber is set at 120 kg / h. Set to -200Pa.

[0060] Input the constructed normalized feature variables into the regression model to calculate the real-time silk content;

[0061] Based on the training dataset, the parameters of the silk content calculation model were fitted using the least squares method, resulting in weight coefficients a=2.1, b=1.5, c=0.3, d=0.8, and a correction constant k=-3.2. The goodness of fit was... =0.93, which meets the accuracy requirements. Therefore, the final calculation model for the fiber content of this grade is obtained:

[0062] ;

[0063] When the air separation system is operating normally, the PLC collects the current operating parameters once per second, including the average wind speed of the air separation airflow v=11.8m / s, the total mass flow rate of the stem outlet m=21kg / h, the inlet leaf moisture content w=12.5, and the negative pressure value of the air separation chamber p=-214Pa. Substituting these parameters into the model, the real-time stem fiber content is calculated as follows:

[0064] ;

[0065] The real-time fiber content is output to the air separation system control unit, and the operating parameters of the air separation system are adjusted according to the real-time fiber content.

[0066] An update cycle is set between adjacent calculation and adjustment cycles, and the update cycle includes the following steps:

[0067] Obtain the calibrated fiber content value of manually sampled data within the currently completed calculation adjustment cycle and the average value of the wind separation system operating parameters for the corresponding time period, and re-input them into the training dataset to refit and update the regression model.

[0068] The preset control range for the fiber content of this grade of blade is 2% to 5%. The current calculated value of 4.2% is within the range, so the parameters of the air separation system remain unchanged. If the calculated value is higher than 5%, the control system will automatically reduce the damper opening to decrease the airflow velocity. If the calculated value is lower than 2%, the damper opening will be automatically increased to increase the airflow velocity, thereby achieving dynamic control of the fiber content.

[0069] In this embodiment, a calibration period of 15 days is set. After the period ends, the filament content of the stem in the period is measured in parallel three times, and the average value is taken as the calibration value. The average value of the wind separation system operating parameters and the calibration value of the corresponding period are added to the training dataset. At the same time, the earliest three historical samples are removed, the model parameters are refitted, and the model iterative update is completed.

[0070] After one month of continuous operation and verification, the average deviation between the fiber content calculated in this embodiment and the value of manual sampling and detection during the same period was 0.28%, and the maximum deviation was 0.47%, which fully meets the accuracy requirements of production control. The single production line reduced the loss of leaf fibers by about 86 kg in a month, resulting in significant economic benefits.

[0071] Specifically, the construction of the regression model includes:

[0072] A training dataset is constructed based on the collected historical operating data and manually calibrated silk content values, and the parameters of the silk content calculation model are fitted based on the training dataset to obtain a regression model;

[0073] in,

[0074] The regression model uses the least squares method for parameter fitting, and the goodness of fit is... Not less than 0.9.

[0075] The wind speed normalization coefficient Calculate using the following formula:

[0076] ;

[0077] Where v is the real-time wind speed. The preset wind speed for the corresponding blade grade.

[0078] The normalization coefficient of the twig production Calculate using the following formula:

[0079] ;

[0080] in This refers to the real-time total mass flow rate at the outlet of the stem label. This is the preset theoretical yield benchmark value for the corresponding grade of leaf filaments.

[0081] The negative pressure normalization coefficient Calculate using the following formula:

[0082] ;

[0083] Where p is the real-time negative pressure value of the air separation chamber. Set a rated negative pressure value for the air separation chamber.

[0084] The regression model calculates the real-time silk content using the following formula. :

[0085] ;

[0086] in, This is the wind speed normalization coefficient. The normalization coefficient for stem yield is . is the negative pressure normalization coefficient, w is the real-time value of leaf fiber moisture content, a, b, c, and d are all weighting coefficients, and k is a correction constant.

[0087] Example 2:

[0088] This embodiment also provides a leaf fiber wind separation stem tagging system, which uses the leaf fiber wind separation stem tagging system fiber content calculation method described in the first aspect;

[0089] The operating parameters of the air separation system include the inlet flow rate of the blades in the air separation chamber 201, the moisture content of the blades, the air velocity of the air separation airflow, the total mass flow rate of the stems at the outlet, and the negative pressure value of the air separation chamber 201.

[0090] Specifically, the leaf filament wind-selection stem labeling system includes:

[0091] Electronic belt scale 202: Used to collect the flow rate of the blades at the inlet of the air separator 201;

[0092] Online moisture meter 203: Used to collect leaf filament moisture content;

[0093] Wind speed sensor 204: Used to collect wind speed of wind-selected airflow;

[0094] Weighing device 206: used to collect the total mass flow rate at the stem outlet;

[0095] Negative pressure sensor 205: Used to collect negative pressure values ​​in the air separation chamber;

[0096] The PLC controller 207 is used to connect with the electronic belt scale 202, online moisture meter 203, wind speed sensor 204, weighing device 206 and negative pressure sensor 205, and to receive the operating parameters of the wind separation system collected by each sensor, and to call the regression model and executable program stored in the memory to calculate the normalized characteristic variables in order to output the real-time fiber content.

[0097] The damper actuator 209 is used to connect to the PLC controller 207 and adjust the damper opening according to the control commands output by the PLC controller 207.

[0098] The fan frequency conversion control unit 210 is used to connect to the PLC controller 207 and adjust the fan operating frequency according to the control commands output by the PLC controller 207.

[0099] Ten wind speed sensors 204 are provided and are respectively set at ten sampling points. The wind speed of the wind-selected airflow is determined based on the average value of the data from the ten sampling points.

[0100] The sampling points are evenly distributed across the cross-section of the air outlet of the air separator.

[0101] In this embodiment, as Figure 2 As shown, an electronic belt scale 202 and an online moisture meter 203 are sequentially installed at the inlet of the air classifier 201; 10 wind speed sensors 204 are evenly distributed at the air outlet of the air classifier 201, and a negative pressure sensor 205 is installed on the side wall; a weighing device 206 is installed at the stem discharge port of the air classifier 201; all the above detection devices are connected to the PLC controller 207, which is associated with a memory 208 containing a calibration model and an executable program. The output end is connected to the damper actuator 209 of the air classifier system and the frequency converter control unit 210 of the fan, and can automatically adjust the air classifier operating parameters according to the calculated fiber content.

[0102] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for calculating the silk content in a leaf silk wind-selection and stem-picking system, characterized in that: This includes calculating the adjustment cycle, outputting the real-time fiber content obtained based on the calculated adjustment cycle to the air separation system control unit, and adjusting the air separation system operating parameters based on the real-time fiber content; The calculation of the adjustment cycle includes the following steps: Obtain historical operating data and corresponding manually calibrated silk content values, construct a training dataset, and perform parameter fitting on the silk content calculation model based on the training dataset to obtain a regression model; Obtain the operating parameters of the air separation system, and construct normalized feature variables based on the operating parameters of the air separation system; the normalized feature variables include the wind speed normalization coefficient, the stem yield normalization coefficient, and the negative pressure normalization coefficient; Input the constructed normalized feature variables into the regression model to calculate the real-time silk content; in, The regression model further includes: setting an update period between adjacent calculation adjustment periods, the update period including the following steps: Obtain the calibrated fiber content value of manually sampled data within the currently completed calculation adjustment cycle and the average value of the wind separation system operating parameters for the corresponding time period, and re-input them into the training dataset to refit and update the regression model.

2. The method for calculating the silk content of the leaf silk wind-selection and stem-picking system according to claim 1, characterized in that, The construction of the regression model includes: A training dataset is constructed based on the collected historical operating data and manually calibrated silk content values, and the parameters of the silk content calculation model are fitted based on the training dataset to obtain a regression model; in, The regression model uses the least squares method for parameter fitting, and the goodness of fit is... Not less than 0.

9.

3. The method for calculating the silk content of the leaf silk wind-selection and stem-picking system according to claim 1, characterized in that, The wind speed normalization coefficient Calculate using the following formula: ； Where v is the real-time wind speed. The preset wind speed for the corresponding blade grade.

4. The method for calculating the silk content of the leaf silk wind-selection and stem-picking system according to claim 1, characterized in that, The normalization coefficient of the twig production Calculate using the following formula: ； in This refers to the real-time total mass flow rate at the outlet of the stem label. This is the preset theoretical yield benchmark value for the corresponding grade of leaf filaments.

5. The method for calculating the silk content of the leaf silk wind-selection and stem-picking system according to claim 1, characterized in that, The negative pressure normalization coefficient Calculate using the following formula: ； Where p is the real-time negative pressure value of the air separation chamber. Set a rated negative pressure value for the air separation chamber.

6. The method for calculating the silk content of the leaf silk wind-selection and stem-picking system according to claim 1, characterized in that, The regression model calculates the real-time silk content using the following formula. : ； in, This is the wind speed normalization coefficient. The normalization coefficient for stem yield is . is the negative pressure normalization coefficient, w is the real-time value of leaf fiber moisture content, a, b, c, and d are all weighting coefficients, and k is a correction constant.

7. The method for calculating the silk content of the leaf silk wind-selection and stem-picking system according to claim 1, characterized in that, The operating parameters of the air separation system include the inlet flow rate of the blades in the air separation chamber, the moisture content of the blades, the air velocity of the air separation airflow, the total mass flow rate of the stems at the outlet, and the negative pressure value of the air separation chamber.

8. A leaf and stem sorting system, characterized in that, The method for calculating the fiber content of the leaf fiber wind separation stem tag system according to any one of claims 1-7 includes: Electronic belt scale (202): Used to collect the flow rate of the blades at the inlet of the air separator (201); Online moisture meter (203): used to collect leaf filament moisture content; Wind speed sensor (204): used to collect wind speed of wind-selected airflow; Weighing device (206): used to collect the total mass flow rate at the stem outlet; Negative pressure sensor (205): Used to collect negative pressure values ​​in the air separation chamber; The PLC controller (207) is used to connect with the electronic belt scale (202), online moisture meter (203), wind speed sensor (204), weighing device (206) and negative pressure sensor (205), and to receive the operating parameters of the wind separation system collected by each sensor, and to call the regression model and executable program stored in the memory to calculate the normalized characteristic variables in order to output the real-time fiber content. The damper actuator (209) is used to connect to the PLC controller (207) and adjust the damper opening according to the control instructions output by the PLC controller (207); The fan frequency conversion control unit (210) is used to connect to the PLC controller (207) and adjust the fan operating frequency according to the control instructions output by the PLC controller (207).

9. The leaf and stem sorting system according to claim 1, characterized in that, Ten wind speed sensors (204) are provided and are respectively set at ten sampling points. The wind speed of the wind-selected airflow is determined based on the average value of the data from the ten sampling points. The sampling points are evenly distributed across the cross-section of the air outlet of the air separator.

Images