Calibration method of cigarette moisture measuring instrument

Abstract

The invention discloses a verification method of a cigarette moisture measuring instrument. The verification method comprises the following steps: extracting multi-dimensional production data and constructing a feature matrix; selecting features with significant correlation through a mutual information method, and deleting features with low mutual information values to form an optimized feature set; dividing a data set through stratified sampling, and setting a fixed random seed; building a Light GBM model based on a gradient boosting decision tree, and optimizing model parameters through grid search; performing system deployment on the model, and automatically collecting factors such as perfuming outlet moisture, environment temperature and humidity, storage time and the like; and performing moisture value prediction by using the model, comparing an off-line moisture meter measurement value with a moisture prediction value, and rechecking and checking the moisture meter when the off-line moisture meter measurement value exceeds a threshold value. The method is used for real-time prediction of the off-line moisture meter and verification of the detection precision, so that the precision of the off-line moisture meter is improved, the verification period is shortened, and the verification efficiency is improved.

Description

Technical Field

[0001] This invention relates to the field of cigarette moisture content measuring equipment, and more specifically to a calibration method for a cigarette moisture measuring instrument. Background Technology

[0002] Moisture content in finished cigarettes is a crucial indicator affecting cigarette quality, directly influencing sensory quality, physical properties, and storage safety. Therefore, strict moisture content control is essential during cigarette production to ensure consistent quality. Currently, during finished cigarette release inspection, the testing laboratory uses the 2-hour oven drying method according to national standards to determine moisture content. The principle of this method is that the moisture content is the ratio of the weight loss to the original weight of the sample when dried to constant weight under specified conditions. This method relies heavily on manual inspection, is time-consuming, has low efficiency, and is susceptible to human error. Furthermore, the results only reflect the average level of the cigarette and cannot accurately reflect the distribution of tobacco throughout the cigarette.

[0003] The cigarette moisture and density distribution measuring instrument uses microwave resonance technology, which allows for direct measurement without any sample treatment, avoiding the influence of human factors on moisture content. The instrument can also detect the moisture and density values ​​of each cross-section of the cigarette and perform data analysis on any segment as needed. This method offers short detection time, high measurement accuracy, and good stability, significantly improving work efficiency. It also provides accurate and reliable data for controlling the uniformity of cigarette moisture and density, thereby improving the quality of cigarette manufacturing.

[0004] Currently, when changing cigarette cabinets or batches, a cigarette moisture and density distribution meter is used to test the moisture content of the cigarettes. This can quickly and effectively reflect the moisture content of the cigarettes, preventing the finished cigarettes from having excessive moisture content, which would affect their intrinsic quality. To ensure the validity of the data, a comparison is made weekly with the oven drying method, and adjustments are made promptly if the moisture content deviates. However, with such a calibration cycle, it is difficult to detect deviations in the moisture meter, and the calibration efficiency is also relatively low.

[0005] Therefore, this application aims to establish a calibration method and system for a cigarette moisture meter. Summary of the Invention

[0006] In view of this, the present invention provides a calibration method for a cigarette moisture meter, which is used to verify the real-time prediction and detection accuracy of an offline moisture meter, so as to improve the accuracy of the offline moisture meter, shorten the calibration cycle, and improve the calibration efficiency.

[0007] The technical solution adopted by the present invention to solve the above-mentioned technical problems is as follows: Calibration methods for cigarette moisture analyzers, including S1 Data Preparation S11 Data Collection Extract multi-dimensional production data from industrial databases, including shifts, work groups, process control capability indicators for different grades, moisture content at flavoring outlets, offline moisture meter measurements, storage cabinets, ambient temperature, ambient humidity, and storage time. Constructing the feature matrix and target vector ; S12 Key Feature Screening By using mutual information, highly correlated features are selected, dimensionality is reduced, and features with low mutual information values ​​are removed. To form an optimized feature set :

[0008]

[0009] in, Indicates an indicator function, This indicates that the dataset contains The observed value is less than or equal to this value. Indicates quantile level; Key characteristics after screening include: brand process control capability indicators, flavoring outlet moisture content, ambient temperature, ambient humidity, and storage time; S13 Data Preprocessing The dataset was partitioned using stratified sampling, with the training and test sets divided in a 7:3 ratio, and a fixed random seed was set. (

[0010] S2 Model Construction and Training Construct a Light GBM model based on gradient boosting decision trees; Input factors include the moisture content at the flavoring outlet, storage time, and ambient temperature and humidity; The output is the predicted value from an offline moisture meter; S21 gradient boosting The model prediction value for the t-th iteration is:

[0011] The objective function of the algorithm is , The number of leaf nodes. Leaf weights:

[0012] loss function Optimization objective:

[0013] in, Indicates the first The predicted output value of the wheel model, Indicates sample The true value, Indicates the first Weak learner (decision tree) Indicates the learning rate; S22 Splitting Gain When splitting a decision tree node, the feature and split point with the largest gain are selected; the gain formula is:

[0014] in, Indicates sample First derivative of the loss function Indicates sample Second derivative of the loss function Represents the regularization coefficient; S3 model parameter tuning The parameters involved in optimizing the model parameters through grid search include: Number of leaf nodes (num_leaves): Controls the complexity of the tree; the theoretical optimal value is [value missing]. ; Number of iterations (n_estimators): The number of decision trees, which affects model capacity; Learning rate: scales the contribution of each tree to prevent overfitting; The optimized model parameters are set as follows: number of leaf nodes -15; number of trees -30; learning rate -0.1. S4 Model Deployment and Validation The model is deployed systematically and integrated with the production statistics system to automatically collect factors such as moisture content at the fragrance outlet, ambient temperature and humidity, and storage time. Using models to predict moisture content; Compare the offline moisture meter readings with the predicted moisture values; when the error is ≥ ±0.5%, an alarm will be triggered, and the moisture meter will be retested and calibrated.

[0015] The beneficial effects of this invention are as follows: This invention constructs a high-precision moisture meter monitoring system, achieving: real-time moisture prediction, automatic detection of instrument deviation, and dynamic calibration reminders. Its core advantages are: 1. Improve the real-time performance of verification: By connecting the model with the production statistics system, relevant parameters can be automatically collected and moisture values ​​can be predicted in real time, changing the traditional fixed-cycle verification mode and realizing dynamic monitoring.

[0016] 2. Improve measurement accuracy: By comparing the model's predicted value with the offline moisture meter's measured value, an alarm is triggered and a re-inspection is initiated when the error exceeds ±0.5%. This allows for rapid detection and calibration of instrument deviations, reducing inaccurate moisture detection of cigarettes due to instrument errors.

[0017] 3. Improved verification efficiency: The automated data collection and prediction process reduces manual intervention and shortens the verification cycle. At the same time, the model parameters are optimized through grid search, ensuring the high accuracy of the prediction model and providing reliable support for the stable control of cigarette moisture. Ultimately, this ensures the stability of the sensory quality, physical properties and storage safety of cigarettes. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention, 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 the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of the method flow provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the system structure provided in an embodiment of the present invention; Figure 3 , Figure 4 These are schematic diagrams comparing the model's predicted values ​​with the actual values ​​before and after model parameter optimization; Figure 5 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

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

[0021] Furthermore, the following description is for illustrative purposes and not for limitation, and sets forth specific details such as particular system structures and techniques to provide a thorough understanding of the embodiments of the invention. However, those skilled in the art will understand that the invention can be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted to avoid unnecessary detail that could obscure the description of the invention.

[0022] The first embodiment of the present invention relates to a calibration method for a cigarette moisture meter, including... S1 Data Preparation S11 Data Collection Multi-dimensional production data of tobacco flavoring process from January to April 2025 were extracted from the industrial database, including shift, work group, process control capability index (CPK) of brand, flavoring outlet moisture (current measurement value), offline moisture meter measurement value (for comparison after prediction), tobacco storage cabinet, ambient temperature (°C), ambient humidity (%RH), and storage time (hours). Data size: 150 records, 9-dimensional features; Constructing the feature matrix and target vector :

[0023]

[0024] in, Indicates the number of samples. Indicates the feature dimension.

[0025] S12 Key Feature Screening By selecting highly correlated features using mutual information, dimensionality can be reduced.

[0026] Remove features with low mutual information values This includes shifts, work groups, and wire storage cabinets, forming an optimized feature set. :

[0027]

[0028] in, Indicates an indicator function, This indicates that the dataset contains The observed value is less than or equal to this value. Indicates quantile level; The key characteristics selected after screening include: brand process control capability index (CPK), flavoring outlet moisture, ambient temperature, ambient humidity, and storage time.

[0029] S13 Data Preprocessing The dataset was partitioned using stratified sampling, with the training and test sets divided in a 7:3 ratio, and a fixed random seed was set. (

[0030] S2 Model Construction and Training Build a Light GBM model based on Gradient Boosting Decision Tree (GBDT); Input factors include the grade process control capability index, flavoring outlet moisture content, storage time, and ambient temperature and humidity; The output is the predicted value from an offline moisture meter; S21 gradient boosting The model prediction value for the t-th iteration is:

[0031] The objective function of the algorithm is , The number of leaf nodes. Leaf weights:

[0032] loss function Optimization objective:

[0033] in, Indicates the first The predicted output value of the wheel model, Indicates sample The true value, Indicates the first Weak learner (decision tree) Indicates the learning rate; S22 Splitting Gain When splitting a decision tree node, the feature and split point with the largest gain are selected; the gain formula is:

[0034] in, Indicates sample First derivative of the loss function Indicates sample Second derivative of the loss function Represents the regularization coefficient; S23 model training results Model training is performed using the training set, and the model's predicted values ​​are compared with the actual values, for example... Figure 3 As shown; Model score result: 0.9430097232045492.

[0035] S3 model parameter tuning Optimizing model parameters using Grid Search CV involves the following parameters: Number of leaf nodes (num_leaves): Controls the complexity of the tree; the theoretical optimal value is [value missing]. ; Number of iterations (n_estimators): The number of decision trees, which affects model capacity; Learning rate: scales the contribution of each tree to prevent overfitting;

[0036]

[0037] The optimization objective is to maximize the coefficient of determination. :

[0038] in, For evaluating the model's performance metrics, Indicates sample The mean; The optimized model parameters are set as follows: number of leaf nodes -15; number of trees -30; learning rate -0.1. The model score after parameter tuning is: 0.9530230768551811; Comparison of predicted and actual values ​​after model parameter tuning Figure 4 As shown.

[0039] S4 Model Deployment and Validation The model is deployed systematically and integrated with the production statistics system to automatically collect factors such as moisture content at the fragrance outlet, ambient temperature and humidity, and storage time. Using a model to predict the moisture content of cigarettes; Compare the measured values ​​of the offline moisture meter with the predicted moisture values; when the error is ≥ ±0.5%, an alarm will be triggered, a re-inspection will be performed, and the offline moisture meter will be calibrated according to the relevant metrology instrument operating procedures.

[0040] The second embodiment of the present invention relates to a calibration system for a cigarette moisture meter, referring to... Figure 2 ,include The data collection unit extracts multi-dimensional production data from the industrial database, including shift, work group, process control capability indicators of grade, moisture content at the flavoring outlet, offline moisture meter measurement, storage cabinet, ambient temperature, ambient humidity, and storage time, to construct a feature matrix. The feature selection unit is used to select highly correlated features and delete features with low mutual information values ​​to form an optimized feature set by using the mutual information method. The data preprocessing unit divides the dataset by stratified sampling, proportionally dividing the training and test sets, and setting a fixed random seed; The model building and training unit is used to build a Light GBM model based on gradient boosting decision trees. The input factors include the moisture content at the flavoring outlet, storage time, and ambient temperature and humidity. The output is the predicted value from an offline moisture meter. The model parameters are optimized through grid search, and the parameters involved include the number of leaf nodes, the number of iterations, and the learning rate. The model deployment and verification unit is used to deploy the model in the system and connect with the production statistics system to automatically collect factors such as moisture content at the fragrance outlet, ambient temperature and humidity, and storage time. At the same time, it uses the model to predict the moisture value and compares the measured value of the offline moisture meter with the predicted value. When the error exceeds the threshold, an alarm is triggered, and the moisture meter is re-inspected and verified.

[0041] Figure 5 This is a schematic diagram of the electronic device 30 provided in an embodiment of the present invention. Figure 5 As shown, the electronic device 30 of this embodiment includes: a processor 31, a memory 32, and a computer program 33 stored in the memory 32 and executable on the processor 31, such as a calibration program for a cigarette moisture meter. When the processor 31 executes the computer program 33, it implements the steps in the various cigarette moisture meter calibration method embodiments described above, for example... Figure 1 The steps shown. Alternatively, when processor 31 executes computer program 33, it implements the functions of each module in the above-described device embodiments, for example... Figure 2 The unit module shown.

[0042] For example, computer program 33 may be divided into one or more modules / units, one or more of which are stored in memory 32 and executed by processor 31 to complete the present invention. One or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of computer program 33 in electronic device 30.

[0043] Electronic device 30 can be a desktop computer, laptop, handheld computer, cloud server, or other computing device. Electronic device 30 may include, but is not limited to, a processor 31 and a memory 32. Those skilled in the art will understand that... Figure 5 This is merely an example of electronic device 30 and does not constitute a limitation on electronic device 30. It may include more or fewer components than shown, or combine certain components, or different components. For example, electronic device 30 may also include input / output devices, network access devices, buses, etc.

[0044] The processor 31 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0045] The memory 32 can be an internal storage unit of the electronic device 30, such as a hard disk or RAM of the electronic device 30. The memory 32 can also be an external storage device of the electronic device 30, such as a plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card, or FlashCard. Furthermore, the memory 32 can include both internal and external storage units of the electronic device 30. The memory 32 is used to store computer programs and other programs and data required by the electronic device 30. The memory 32 can also be used to temporarily store data that has been output or will be output.

[0046] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0047] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0048] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0049] In the embodiments provided by this invention, it should be understood that the disclosed devices / electronic devices and methods can be implemented in other ways. For example, the device / electronic device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0050] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0051] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0052] If integrated modules / units are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0053] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A calibration method for a cigarette moisture meter, characterized in that: include S1 Data Preparation S11 Data Collection Extract multi-dimensional production data from industrial databases, including shifts, work groups, process control capability indicators for different grades, moisture content at flavoring outlets, offline moisture meter measurements, storage cabinets, ambient temperature, ambient humidity, and storage time. Constructing the feature matrix and target vector : in, Indicates the number of samples. Indicates feature dimension; S12 Key Feature Screening By selecting highly correlated features using mutual information, dimensionality can be reduced. Remove features with low mutual information values To form an optimized feature set : in, Indicates an indicator function, This indicates that the dataset contains The observed value is less than or equal to this value. Indicates quantile level; Key characteristics after screening include: brand process control capability indicators, flavoring outlet moisture content, ambient temperature, ambient humidity, and storage time; S13 Data Preprocessing Divide the training and test sets proportionally and set a fixed random seed: （ S2 Model Construction and Training Construct a Light GBM model based on gradient boosting decision trees; Input factors include the moisture content at the flavoring outlet, storage time, and ambient temperature and humidity; The output is the predicted value from an offline moisture meter; S21 gradient boosting The model prediction value for the t-th iteration is: The objective function of the algorithm is , The number of leaf nodes. Leaf weights: loss function Optimization objective: in, Indicates the first The predicted output value of the wheel model, Indicates sample The true value, Indicates the first Weak learner Indicates the learning rate; S22 Splitting Gain When splitting a decision tree node, the feature and split point with the largest gain are selected; the gain formula is: in, Indicates sample First derivative of the loss function Indicates sample Second derivative of the loss function Represents the regularization coefficient; S3 model parameter tuning The model parameters are optimized through grid search, including the number of leaf nodes, the number of iterations, and the learning rate. The optimization objective is to maximize the coefficient of determination. : in, For evaluating the model's performance metrics, Indicates sample The mean; S4 Model Deployment and Validation The model is deployed systematically and integrated with the production statistics system to automatically collect factors such as moisture content at the fragrance outlet, ambient temperature and humidity, and storage time. Using models to predict moisture content; Compare the offline moisture meter readings with the predicted moisture values; when the error is ≥ ±0.5%, an alarm will be triggered, and the moisture meter will be retested and calibrated.

2. The calibration method for the cigarette moisture meter according to claim 1, characterized in that: The optimized model parameters are set as follows: number of leaf nodes -15; number of trees -30; learning rate -0.

1.

3. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 2.

4. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 2.

Images