A method and system for predicting moisture at the outlet of a cut tobacco dryer

Abstract

The application discloses a kind of prediction method and system of cut tobacco outlet moisture, belong to cut tobacco outlet moisture prediction technical field.Its method includes: obtaining the outside boundary condition of cylinder of cut tobacco machine;The outside boundary condition of cylinder includes historical outside boundary condition and current outside boundary condition;The historical outside boundary condition is input into the fuzzy rule classification model constructed in advance, and current working condition stage is output;The current outside boundary condition and the inside boundary condition of cylinder mechanism model feedback are input into the heat and mass transfer regression algorithm model constructed in advance, and inside heat and mass transfer coefficient of cylinder is output;The current outside boundary condition, current working condition and inside heat and mass transfer coefficient of cylinder are input into the cylinder mechanism model constructed in advance, and the predicted value of cut tobacco machine outlet moisture is output.The application can consider the multiple limiting conditions existing in actual production, and organically combine cut tobacco machine mechanism model and data model, can predict cut tobacco machine outlet moisture under given condition.

Description

Technical Field

[0001] This invention relates to the field of tobacco shred exit moisture prediction technology, and in particular to a method and system for predicting the exit moisture of a tobacco drying machine. Background Technology

[0002] There are currently two main methods for predicting the outlet moisture content of yarn drying machines:

[0003] One method is mechanism modeling and prediction. This type of method is based on thermodynamic principles, which describes the heat and mass transfer process in the wire drying cylinder in the form of physical formulas and summarizes them into mathematical equations for computer solution.

[0004] There are two main approaches to predicting outlet moisture content using mechanistic modeling. One approach ignores the convective mass transfer process of tobacco shreds within the drying drum and the energy changes within the drum, using Fick's second law to characterize the diffusion process and the Arrhenius formula to correct the diffusion coefficient. The other approach establishes convective heat and mass transfer formulas between tobacco shreds, moisture, and hot air within the drying drum and calculates the energy changes during moisture migration. Both approaches neglect or simplify certain factors in terms of mechanism and cannot account for the complex internal conditions of the drying drum during actual operation. On the one hand, theoretical calculations often only predict ideal conditions under given laboratory conditions, and accuracy is difficult to guarantee once conditions are changed. On the other hand, theoretical calculations cannot consider the impact of disturbances such as air leakage and temperature fluctuations generated in actual production. This limits the application of mechanistic models in predicting outlet moisture content in tobacco drying machines.

[0005] Another type is data modeling prediction. This method uses machine learning to extract patterns and experiences from large amounts of data collected from the field or laboratory, and predicts the moisture content of the output based on a given production state. Commonly used methods include Bayesian networks and Long Short-Term Memory networks. Due to the production characteristics and data features of the drying machine, these methods often require collecting a large amount of actual production data and conducting long-term model training to achieve high fitting accuracy. This approach mainly brings two problems: First, the trained model contains a large number of meaningless hyperparameters. These parameters only reflect the mathematical characteristics of the model and have no relation to the actual physical entity, resulting in the model's control methods lacking inputs with clear physical meaning. Second, the trained model is at risk of overfitting. When the drying machine's production conditions change, the drying machine ages, or the drying machine is replaced, the model's accuracy often decreases significantly, requiring re-collection and retraining, which is very time-consuming and labor-intensive. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method and system for predicting the moisture content of tobacco shreds at the outlet of a tobacco drying machine. This method and system can take into account various limiting conditions in actual production and organically combine the mechanism model of the tobacco drying machine with various data models, thereby accurately predicting the moisture content of tobacco shreds at the outlet of the tobacco drying machine under given conditions.

[0007] To achieve the above objectives, the present invention is implemented using the following technical solution:

[0008] On one hand, the present invention provides a method for predicting the moisture content at the outlet of a yarn drying machine, comprising:

[0009] Obtain the outer boundary conditions of the yarn drying machine; the outer boundary conditions include historical outer boundary conditions and current outer boundary conditions.

[0010] The historical outer boundary conditions are input into a pre-built fuzzy rule classification model, and the current operating condition stage is output.

[0011] The current external boundary conditions and the internal boundary conditions fed back from the drying cylinder mechanism model are input into the pre-constructed heat and mass transfer regression algorithm model, and the internal heat and mass transfer coefficient is output.

[0012] The current external boundary conditions, current operating conditions, and internal heat and mass transfer coefficients are input into a pre-built drying cylinder mechanism model, and the predicted value of the moisture content at the outlet of the drying machine is output.

[0013] Optional, also includes:

[0014] Input the historical outer boundary conditions into the moisture residual correction model and output the corrected value of the moisture at the outlet of the drying machine;

[0015] The predicted value of the moisture content at the outlet of the drying machine is corrected using the correction value of the moisture content at the outlet of the drying machine.

[0016] Optionally, the external boundary conditions of the cylinder include tobacco flow rate, hot air temperature, hot air flow rate, and cylinder wall temperature;

[0017] The current operating conditions include the shutdown phase, the beginning phase, the stabilization phase, and the end phase.

[0018] Optionally, the construction of the fuzzy rule classification model includes:

[0019] The shutdown phase, dry start phase, stable phase, and dry end phase are labeled as linearly independent four-dimensional vectors, namely [1, 0, 0, 0], [0, 1, 0, 0], [0, 0, 1, 0], and [0, 0, 0, 1], respectively.

[0020] Construct a two-dimensional vector to distinguish between the beginning and end of the drying process, wherein the two-dimensional vector is [moisture content, cumulative operating time];

[0021] Based on the four-dimensional vector and the two-dimensional vector, the constructed fuzzy rule classification model is obtained.

[0022] Optionally, the processing steps of the heat and mass transfer regression algorithm model include:

[0023] Based on the moisture content of the dried tobacco shreds, calculate the logarithmic curve of the moisture content of the dried tobacco shreds. The formula for calculating the logarithmic curve of the moisture content of the dried tobacco shreds is as follows:

[0024] ;

[0025] ;

[0026] in, The moisture content of dried tobacco shreds; Indicates the dry basis moisture content of tobacco shreds; Indicates the dry equilibrium moisture content of the sample; Indicates the initial moisture content of the tobacco shreds; Indicates the heat and mass transfer coefficient inside the cylinder; Indicates the thickness of the dried product; Indicates drying time;

[0027] Using the SVR regression algorithm, based on the current external boundary conditions and the internal boundary conditions fed back by the drying drum mechanism model, the logarithmic curve of the moisture ratio of the dried tobacco shreds is calculated. The curve is fitted to obtain its slope; based on the slope of the curve, the heat and mass transfer coefficient inside the cylinder is calculated. .

[0028] Optionally, the processing steps of the drying cylinder mechanism model include:

[0029] The relative humidity of hot air is calculated based on the preset interpolation table and hot air temperature. The equilibrium humidity of water vapor in the tobacco film is calculated using the Hendersen correlation and the preset tobacco physical property parameter table.

[0030] The air film water vapor balance density of the hot air is calculated based on the relative humidity of the hot air, and the air film water vapor balance density of the tobacco is calculated based on the air film water vapor balance humidity of the tobacco.

[0031] Based on the current operating conditions, the air film water vapor balance density of the hot air, and the air film water vapor balance density of the tobacco, the predicted value of the moisture content at the outlet of the tobacco drying machine is calculated.

[0032] Optionally, the formulas for calculating the film water vapor equilibrium density of the hot air and the film water vapor equilibrium density of the tobacco are as follows:

[0033] ;

[0034] ;

[0035] ;

[0036] in, , These represent the film water vapor equilibrium density of hot air and tobacco, respectively. , These represent the hot air temperature and the tobacco temperature, respectively. Indicates the quality of the hot air; , These represent the relative humidity of hot air and the equilibrium humidity of water vapor in the tobacco film, respectively. , These represent the average temperature of the hot air and the average temperature of the tobacco, respectively. , Represents the correlation coefficient; This indicates the equilibrium humidity of the air film on the surface of the tobacco.

[0037] Optionally, the formula for calculating the predicted moisture content at the outlet of the yarn drying machine is as follows:

[0038] ;

[0039] ;

[0040] ;

[0041] ;

[0042] ;

[0043] ;

[0044] ;

[0045] in, This indicates the moisture content at the outlet of the drying machine during the drying stage; This indicates the moisture content at the outlet of the drying machine during the drying stage; This indicates the moisture content at the outlet of the drying machine during the stable phase. Indicates the linear velocity of the tobacco shreds; Indicates the rotational speed of the cylinder; Indicates the roller tilt angle; This indicates the residence time of the tobacco unit inside the tube; Indicates the length of the roller; , This indicates the moisture content of tobacco shreds in the initial drying stage and the moisture content of tobacco shreds in the final drying stage. Indicates drying time; Indicates the drying end time; , Represents the correlation coefficient; Indicates the convective mass transfer coefficient; Indicates the heat exchange area; , These represent the water vapor density in the hot air and tobacco under the air film equilibrium state, respectively.

[0046] In a second aspect, the present invention provides a system for predicting the moisture content at the outlet of a yarn drying machine, including a processor and a storage medium;

[0047] The storage medium is used to store instructions;

[0048] The processor is configured to operate according to the instructions to execute the method according to the first aspect.

[0049] Thirdly, the present invention provides a computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the steps of the method described in the first aspect.

[0050] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0051] This invention combines the advantages of mechanistic modeling and data modeling. The drying cylinder mechanism model has a clear and interpretable physical meaning. For the influence of working conditions and the actual diffusion coefficient of tobacco that cannot be considered in the mechanism model, fuzzy rule classification model and heat and mass transfer regression algorithm model are used for fitting respectively. There are one-way and two-way data interactions between the sub-models. The model can be automatically corrected under conditions such as external production data updates or aging of the drying machine. Adaptive updates are achieved through moisture residual correction model, thereby achieving higher prediction accuracy and prediction stability. Attached Figure Description

[0052] Figure 1 The diagram shown is a flowchart of one embodiment of the method for predicting the moisture content at the outlet of the yarn drying machine according to the present invention.

[0053] Figure 2 The figure shown is a discrete modeling schematic diagram of the drying cylinder mechanism model of the present invention in one embodiment;

[0054] Figure 3 The diagram shows a comparison of the predicted moisture content at the outlet of the drying machine in one embodiment of the present invention and the prior art.

[0055] Figure 4 The diagram shows a comparison of the predicted moisture content at the outlet of the drying machine in another embodiment of the present invention and the prior art. Detailed Implementation

[0056] 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 of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0057] The term "and / or" simply describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. Additionally, the character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0058] Example 1

[0059] like Figure 1 As shown, this embodiment introduces a method for predicting the moisture content at the outlet of a tobacco drying machine. It combines the advantages of mechanistic modeling and data modeling, while addressing the shortcomings of each when used alone. The mechanistic model of the drying cylinder has a clear and interpretable physical meaning. For the influence of operating conditions, the actual diffusion coefficient of tobacco, and environmental disturbances of the drying machine that cannot be considered in the mechanistic model, the fuzzy rule classification model, the heat and mass transfer regression algorithm model, and the moisture residual correction model are used for fitting, respectively.

[0060] The method specifically includes the following steps:

[0061] Step 1: Obtain the outer boundary conditions of the tobacco drying machine. The outer boundary conditions include historical outer boundary conditions and current outer boundary conditions. The outer boundary conditions include tobacco flow rate, hot air temperature, hot air flow rate, and drum wall temperature.

[0062] For the data acquisition result processing module, it receives the raw data collected by the sensor and performs certain preprocessing operations such as data filtering and data cleaning to obtain the outer boundary conditions of the cylinder. Then, it inputs the outer boundary conditions of the cylinder required by other models into the corresponding modules.

[0063] Step 2: Input the historical outer boundary conditions into the pre-built fuzzy rule classification model, and output the current operating condition stage, specifically:

[0064] For the fuzzy rule classification model, it mainly obtains the historical outer boundary conditions of the cylinder based on the data acquisition result processing module. After LSTM prediction and fuzzy rule classification, the corresponding operating condition stage is obtained and output to the mechanism model. The current operating condition stage includes the shutdown stage, the dry start stage, the stable stage, and the dry end stage.

[0065] First, time-series data of tobacco moisture content from the inlet to the outlet of the drying cylinder are acquired by sensors, and missing values ​​are filled. Interpolation is used to interpolate discontinuous sampling data caused by equipment failure to generate an equally spaced time series. Then, noise filtering and smoothing are performed to retain the trend characteristics of moisture content change. Finally, the maximum and minimum value method is used to map the moisture content data to the [0,1] interval for normalization.

[0066] The construction of the fuzzy rule classification model includes:

[0067] First, the dry-beginning and dry-end stages are categorized and labeled. The shutdown stage, dry-beginning stage, stable stage, and dry-end stage are labeled as linearly independent four-dimensional vectors: [1, 0, 0, 0], [0, 1, 0, 0], [0, 0, 1, 0], and [0, 0, 0, 1]. This avoids numerical interference between categories and prevents training instability due to numerical magnitude. Then, a two-dimensional vector [moisture content, cumulative running time] containing timestamps is constructed to avoid confusion between the dry-beginning and dry-end stages. Based on the four-dimensional and two-dimensional vectors, the constructed fuzzy rule classification model is obtained.

[0068] Construct a deep learning network containing a bidirectional LSTM layer, which consists of an input layer, an LSTM layer, a Dropout layer, and a fully connected layer connected in sequence.

[0069] Input layer: 2 neurons, corresponding to water content and time features; LSTM layer: the number of neurons was determined to be 32 through grid search; Dropout layer: inactivation rate 0.2; fully connected layer: 4 neurons.

[0070] A variable learning rate strategy is employed during training. In the early stages of network training, a relatively high learning rate is used for iterative updates to quickly approach the minimum point and avoid getting trapped in local optima. When the model residuals cannot be further reduced within consecutive iterations, the learning rate is decreased to precisely reach the minimum point and improve model accuracy. Furthermore, an L2 regularization term with a coefficient λ=0.01 is introduced into the loss function to prevent overfitting during network training.

[0071] The process is transformed into a simulation model. The model takes tobacco moisture content and time as input and classification labels as output: 0 represents the shutdown stage, 1 represents the beginning stage, 2 represents the stable stage, and 3 represents the end stage. The solution strategy is adjusted according to the operating stage. For the beginning and end stages, the flow rate of tobacco in the drum is less than the rated operating condition, leading to an enhanced dehumidification process. This phenomenon is more pronounced at the beginning and end of the drying process. Therefore, an exponential function is used to fit the gain of the dehumidification process over time during the beginning and end stages. Different formulas for calculating the time derivative of moisture content are used at different stages to ensure that the heat and mass transfer characteristics of different operating stages can be accurately matched.

[0072] Step 3: Input the current external boundary conditions and the internal boundary conditions fed back from the drying cylinder mechanism model into the pre-constructed heat and mass transfer regression algorithm model, and output the internal heat and mass transfer coefficients, specifically:

[0073] For the heat and mass transfer regression algorithm model, it calculates the moisture diffusion coefficient of the tobacco at the corresponding position under the given drying cylinder boundary based on the external boundary conditions (hot air temperature, cylinder wall temperature) input by the data acquisition result processing module and the internal boundary conditions (tobacco temperature, tobacco moisture content) calculated by the drying cylinder mechanism model. It then outputs the coefficient to the drying cylinder mechanism model as the internal heat and mass transfer coefficient.

[0074] The moisture content of tobacco can be expressed in two ways: dry basis moisture content and wet basis moisture content, as follows:

[0075] ;

[0076] ;

[0077] in, , These represent the moisture content on a dry basis and the moisture content on a wet basis, respectively. This indicates the mass of moisture in wet tobacco shreds; This indicates the quality of absolutely dry tobacco shreds.

[0078] Moisture ratio refers to the residual moisture content of a material under certain drying conditions, reflecting the speed of tobacco leaf drying. Its calculation formula is as follows:

[0079] ;

[0080] in, The moisture content of dried tobacco shreds; Indicates the dry basis moisture content of tobacco shreds; Indicates the dry equilibrium moisture content of the sample; Indicates the initial moisture content of the tobacco shreds;

[0081] Since the equilibrium moisture content at the end of tobacco drying is smaller than the typical initial moisture content, it can be considered that... =0;

[0082] The drying medium transfers heat to the tobacco shreds via convection. As the tobacco shreds heat up, their moisture evaporates and is transferred to the drying medium, completing the mass transfer process. Assuming the tobacco drying process is thin-layer drying, a differential equation describing the drying process can be established according to Fick's second law. Its analytical solution is:

[0083] ;

[0084] Omitting higher-order terms and taking the logarithm, we obtain the logarithmic curve of the moisture content of the dried tobacco. , represented as:

[0085] ;

[0086] in, Indicates the heat and mass transfer coefficient inside the cylinder; Indicates the thickness of the dried product; Indicates drying time;

[0087] The above formula represents the natural logarithm of the moisture ratio during the tobacco drying process. The relationship between moisture content and drying time t is linear. Using data from different drying machine conditions (i.e., external and internal boundary conditions), the SVR regression algorithm was employed to plot the natural logarithmic drying curve of the moisture ratio during tobacco drying. A fitting process is performed, and the slope is calculated. Based on the slope, the moisture diffusion coefficient, i.e., the heat and mass transfer coefficient inside the cylinder, is obtained. The heat and mass transfer regression algorithm model obtained in this way can calculate the heat and mass transfer coefficients of the tobacco at various points inside the drum, given the known parameters inside and outside the drum of the tobacco drying machine. .

[0088] Step 4: Input the current outer boundary conditions of the cylinder, the current operating conditions, and the heat and mass transfer coefficients inside the cylinder into the pre-constructed drying cylinder mechanism model, and output the predicted value of the moisture content at the outlet of the drying machine, specifically:

[0089] For the drying cylinder mechanism model, in the fusion model, the drying cylinder mechanism model serves as the center of data interaction. It receives external boundary condition data parsed by the data acquisition and processing module, current operating condition data calculated by the fuzzy rule classification model, and internal heat and mass transfer coefficients calculated by the heat and mass transfer regression algorithm model as input. After calculations within the drying cylinder mechanism model, the outlet moisture value is calculated. Finally, the outlet moisture correction value calculated by the moisture residual correction model is superimposed to obtain the final outlet moisture output value. The specific calculation process is as follows:

[0090] like Figure 2 As shown, the model is divided into two computational regions: the tobacco region and the hot air region, using discrete modeling. There are convective heat transfer and convective mass transfer processes between the tobacco and the hot air. At the same time, the cylinder wall also exchanges heat with both the tobacco and the hot air. In addition, the latent heat absorbed by the evaporation of moisture during the dehumidification process of the tobacco must also be considered.

[0091] The dehumidification mechanism of tobacco shreds adopts the two-film theory. The mass transfer flux can be written as the product of the convective mass transfer coefficient and the mass transfer driving force. The mass transfer driving force is expressed as the difference in water vapor density between the hot air on the surface of the tobacco shreds and the air film.

[0092] The calculation of the hot air zone includes the following steps:

[0093] Hot air side mass conservation equation:

[0094] ;

[0095] The energy conservation equation for the hot air side is:

[0096] ;

[0097] ;

[0098] in, , These represent the hot air flow rates entering and exiting the cylinder, respectively. This indicates the flow rate of water vapor evaporating from the tobacco. This represents the evaporation potential of water vapor absorbed by the hot air; This indicates the amount of heat generated by the hot air being heated by the cylinder wall; Indicates the hot air inlet mass flow rate; This indicates the specific heat capacity of air; This indicates the temperature difference between the inlet and outlet within the hot air section volume; This represents the correction factor for convective heat transfer; This represents the convective heat transfer coefficient between the hot air and the cylinder wall; This indicates the convective heat transfer area between the hot air and the cylinder wall; Indicates the number of segments; This represents the temperature of the wall of the i-th volumetric cylinder; , These represent the inlet and outlet temperatures of the hot air within the i-th segment, respectively.

[0099] The calculation of the tobacco shred area includes the following steps:

[0100] Mass conservation equation for tobacco shreds:

[0101] ;

[0102] Energy conservation equation for tobacco:

[0103] ;

[0104] ;

[0105] ;

[0106] in, , These represent the flow rates of tobacco entering and exiting the cylinder, respectively. Indicates the flow rate of tobacco inside the tube; This indicates the specific heat capacity of steam; , These represent the steam temperature in the i-th segment and the steam temperature in the (i+1)-th segment, respectively. This indicates the heat exchanged between steam and tobacco. This indicates the heat exchange between steam and the wall. Indicates wall temperature;

[0107] The dehumidification process of tobacco shreds adopts the two-film theory, in which the time derivative of the moisture content of the tobacco shreds is regarded as the mass transfer flux, which can be expressed as the product of the convective mass transfer coefficient and the mass transfer driving force. According to the two-film theory, the mass transfer driving force can be expressed as the difference between the equilibrium density of water vapor in the air film and the equilibrium density of water vapor in the air film on the surface of the tobacco shreds.

[0108] The processing steps of the drying cylinder mechanism model include:

[0109] The relative humidity of hot air is calculated based on the preset interpolation table and hot air temperature. The equilibrium humidity of water vapor in the tobacco film is calculated using the Hendersen correlation and the preset tobacco physical property parameter table.

[0110] The preset interpolation table is shown in Table 1, and is represented as follows:

[0111] When the hot air temperature is 338.15K, the relative humidity of the hot air is 0.115%.

[0112] When the hot air temperature is 358.15K, the relative humidity of the hot air is 0.050%.

[0113] When the hot air temperature is 378.15K, the relative humidity of the hot air is 0.024%.

[0114] When the hot air temperature is 398.15K, the relative humidity of the hot air is 0.012%.

[0115] When the hot air temperature is 418.15K, the relative humidity of the hot air is 0.007%.

[0116] The preset tobacco shred physical property parameters are shown in Table 2, and are represented as follows:

[0117] When the temperature of the tobacco shreds is 338.15K, the equilibrium humidity of the gas film on the surface of the tobacco shreds is 0.048%.

[0118] When the temperature of the tobacco shreds is 358.15K, the equilibrium humidity of the gas film on the surface of the tobacco shreds is 0.035%.

[0119] When the temperature of the tobacco shreds is 378.15K, the equilibrium humidity of the gas film on the surface of the tobacco shreds is 0.027%.

[0120] When the temperature of the tobacco shreds is 398.15K, the equilibrium humidity of the gas film on the surface of the tobacco shreds is 0.022%.

[0121] When the temperature of the tobacco shreds is 418.15K, the equilibrium humidity of the gas film on the surface of the tobacco shreds is 0.018%.

[0122] Table 1 Preset Interpolation Table

[0123]

[0124] Table 2 Preset Tobacco Properties

[0125]

[0126] The film water vapor equilibrium density of the hot air is calculated based on the relative humidity of the hot air, and the film water vapor equilibrium density of the tobacco is calculated based on the film water vapor equilibrium humidity of the tobacco. The formulas for calculating the film water vapor equilibrium density of the hot air and the film water vapor equilibrium density of the tobacco are as follows:

[0127] ;

[0128] ;

[0129] ;

[0130] in, , These represent the film water vapor equilibrium density of hot air and tobacco, respectively. , These represent the hot air temperature and the tobacco temperature, respectively. Indicates the quality of the hot air; , These represent the relative humidity of hot air and the equilibrium humidity of water vapor in the tobacco film, respectively. , These represent the average temperature of the hot air and the average temperature of the tobacco, respectively. , Represents the correlation coefficient; This indicates the equilibrium humidity of the air film on the surface of the tobacco shreds;

[0131] Based on the current operating condition, the air-film water vapor balance density of the hot air, and the air-film water vapor balance density of the tobacco shreds, the predicted value of the moisture content at the outlet of the tobacco drying machine is calculated. The formula for calculating the predicted value of the moisture content at the outlet of the tobacco drying machine is as follows:

[0132] ;

[0133] ;

[0134] ;

[0135] ;

[0136] ;

[0137] ;

[0138] ;

[0139] in, This indicates the moisture content at the outlet of the drying machine during the drying stage; This indicates the moisture content at the outlet of the drying machine during the drying stage; This indicates the moisture content at the outlet of the drying machine during the stable phase. Indicates the linear velocity of the tobacco shreds; Indicates the rotational speed of the cylinder; Indicates the roller tilt angle; This indicates the residence time of the tobacco unit inside the tube; Indicates the length of the roller; , This indicates the moisture content of tobacco shreds in the initial drying stage and the moisture content of tobacco shreds in the final drying stage. Indicates drying time; Indicates the drying end time; , Represents the correlation coefficient; Indicates the convective mass transfer coefficient; Indicates the heat exchange area; , These represent the water vapor density in the hot air and tobacco under the air film equilibrium state, respectively.

[0140] Step 5: Input the historical outer boundary conditions into the moisture residual correction model, and output the corrected value of the moisture content at the outlet of the drying machine, specifically:

[0141] For the moisture residual correction model, it mainly accepts the historical outer boundary conditions of the cylinder obtained by the data acquisition result processing module, uses time series prediction algorithms such as Transformer to calculate the calculation deviation value of the outlet moisture, outputs the correction value of the outlet moisture of the drying machine, and outputs the correction value to the drying cylinder mechanism model. The predicted value of the outlet moisture of the drying machine is corrected by using the correction value of the outlet moisture of the drying machine.

[0142] This embodiment fully considers the various constraints existing in actual production and organically combines the drying cylinder mechanism model with various data models. A good model architecture is designed to ensure that the data interaction and calculation process between modules are reasonable. The drying cylinder mechanism model provides the data model with inputs with clear physical meaning, while the data model analyzes the influence of environmental and other disturbance factors that cannot be accurately quantified for the mechanism model. The two complement each other.

[0143] Example 2

[0144] Based on Example 1, this example introduces a specific experimental example of a method for predicting the moisture content at the outlet of a yarn drying machine, including:

[0145] Using production data from a cigarette factory as training samples, the method of Example 1 was used to predict the moisture content at the outlet of the tobacco drying machine. Figure 3 The image shows a comparison between the calculated export moisture forecast and the original production batch data.

[0146] from Figure 3 It can be seen that during the entire batch production stage (approximately 8.5 hours), the predicted moisture content at the outlet of the drying machine has a good fit with the data collection results. Furthermore, it can accurately distinguish the current operating stage in the initial production stage (drying head stage) and the final production stage (drying tail stage), with a prediction accuracy of 1.85%.

[0147] To demonstrate the accuracy of the forecasting method in capturing the impact of various complex disturbances in production conditions, Figure 4 The comparison between the predicted moisture content at the outlet of the drying machine and the data collection results during the stable production phase (approximately 8.4 hours) of the entire production batch is shown. It can be seen that the prediction effect is good, with a prediction accuracy of 0.637%.

[0148] Example 3

[0149] This embodiment introduces a system for predicting the moisture content at the outlet of a yarn drying machine, including a processor and a storage medium;

[0150] The storage medium is used to store instructions;

[0151] The processor is configured to operate according to the instructions to execute the method according to embodiment 1 or 2.

[0152] Example 4

[0153] This embodiment describes a computer-readable storage medium storing computer instructions that, when executed by a processor, implement the steps of the method described in Embodiment 1 or 2.

[0154] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0155] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0156] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0157] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0158] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A method for predicting the moisture content at the outlet of a yarn drying machine, characterized in that, include: Obtain the outer boundary conditions of the yarn drying machine; the outer boundary conditions include historical outer boundary conditions and current outer boundary conditions. The historical outer boundary conditions are input into a pre-built fuzzy rule classification model, and the current operating condition stage is output. The current external boundary conditions and the internal boundary conditions fed back from the drying cylinder mechanism model are input into the pre-constructed heat and mass transfer regression algorithm model, and the internal heat and mass transfer coefficient is output. The current outer boundary conditions of the cylinder, the current operating conditions, and the heat and mass transfer coefficients inside the cylinder are input into the pre-built drying cylinder mechanism model, and the predicted value of the moisture content at the outlet of the drying machine is output. The external boundary conditions of the cylinder include tobacco flow rate, hot air temperature, hot air flow rate, and cylinder wall temperature; The current operating condition stages include the shutdown stage, the dry start stage, the stabilization stage, and the dry finish stage; The construction of the fuzzy rule classification model includes: The shutdown phase, dry start phase, stable phase, and dry end phase are labeled as linearly independent four-dimensional vectors, namely [1, 0, 0, 0], [0, 1, 0, 0], [0, 0, 1, 0], and [0, 0, 0, 1], respectively. Construct a two-dimensional vector to distinguish between the beginning and end of the drying process, wherein the two-dimensional vector is [moisture content, cumulative operating time]; Based on the four-dimensional vector and the two-dimensional vector, the constructed fuzzy rule classification model is obtained; The processing steps of the heat and mass transfer regression algorithm model include: Based on the moisture content of the dried tobacco shreds, calculate the logarithmic curve of the moisture content of the dried tobacco shreds. The formula for calculating the logarithmic curve of the moisture content of the dried tobacco shreds is as follows: ； ； in, The moisture content of dried tobacco shreds; Indicates the dry basis moisture content of tobacco shreds; Indicates the dry equilibrium moisture content of the sample; Indicates the initial moisture content of the tobacco shreds; Indicates the heat and mass transfer coefficient inside the cylinder; Indicates the thickness of the dried product; Indicates drying time; Using the SVR regression algorithm, based on the current external boundary conditions and the internal boundary conditions fed back by the drying drum mechanism model, the logarithmic curve of the moisture ratio of the dried tobacco shreds is calculated. The curve is fitted to obtain its slope; based on the slope of the curve, the heat and mass transfer coefficient inside the cylinder is calculated. ; The processing steps of the drying cylinder mechanism model include: The relative humidity of hot air is calculated based on the preset interpolation table and hot air temperature. The equilibrium humidity of water vapor in the tobacco film is calculated using the Hendersen correlation and the preset tobacco physical property parameter table. The air film water vapor balance density of the hot air is calculated based on the relative humidity of the hot air, and the air film water vapor balance density of the tobacco is calculated based on the air film water vapor balance humidity of the tobacco. Based on the current operating conditions, the air film water vapor balance density of the hot air, and the air film water vapor balance density of the tobacco, calculate the predicted value of the moisture content at the outlet of the tobacco drying machine. The formulas for calculating the film water vapor equilibrium density of the hot air and the film water vapor equilibrium density of the tobacco are as follows: ； ； ； in, , These represent the film water vapor equilibrium density of hot air and tobacco, respectively. , These represent the hot air temperature and the tobacco temperature, respectively. Indicates the quality of the hot air; , These represent the relative humidity of hot air and the equilibrium humidity of water vapor in the tobacco film, respectively. , These represent the average temperature of the hot air and the average temperature of the tobacco, respectively. , Represents the correlation coefficient; This indicates the equilibrium humidity of the air film on the surface of the tobacco shreds; The formula for calculating the predicted moisture content at the outlet of the drying machine is as follows: ； ； ； ； ； ； ； in, This indicates the moisture content at the outlet of the drying machine during the drying stage; This indicates the moisture content at the outlet of the drying machine during the drying stage; This indicates the moisture content at the outlet of the drying machine during the stable phase. Indicates the linear velocity of the tobacco shreds; Indicates the rotational speed of the cylinder; Indicates the roller tilt angle; This indicates the residence time of the tobacco unit inside the tube; Indicates the length of the roller; , This indicates the moisture content of tobacco shreds in the initial drying stage and the moisture content of tobacco shreds in the final drying stage. Indicates drying time; Indicates the drying end time; , Represents the correlation coefficient; Indicates the convective mass transfer coefficient; Indicates the heat exchange area; , These represent the film water vapor equilibrium density of hot air and tobacco, respectively.

2. The method for predicting the moisture content at the outlet of a yarn drying machine according to claim 1, characterized in that, Also includes: Input the historical outer boundary conditions into the moisture residual correction model and output the corrected value of the moisture at the outlet of the drying machine; The predicted value of the moisture content at the outlet of the drying machine is corrected using the correction value of the moisture content at the outlet of the drying machine.

3. A system for predicting the moisture content at the outlet of a yarn drying machine, characterized in that, Including processor and storage media; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the method according to any one of claims 1 to 2.

4. A computer-readable storage medium storing computer instructions thereon, characterized in that, When executed by a processor, the computer instructions implement the steps of the method according to any one of claims 1 to 2.

Images