Tobacco leaf curing management control system for stepping type curing barn

Abstract

The invention discloses a tobacco leaf curing management control system for a stepping type curing barn. The stepping type curing barn comprises N + 1 curing barns which are connected in series, the curing barns are connected with a heat pump, temperature and humidity sensors and monitoring systems are arranged in the curing barns, the interiors of the stepping type curing barns are integrally communicated, and the bottoms of the N curing barns are connected through guide rails. The tobacco leaf curing management control system comprises an Internet of Things cloud platform, a master control module and N sub-control modules, the master control module and the sub-control modules both adopt PLCs, a '1 + N' topological structure of one master control plus N sub-control is formed, the PLC, the Internet of Things cloud platform and the intelligent monitoring technology are integrated, and the system is simple in structure, convenient to operate and high in practicability. Complete-period management is achieved for the complete tobacco leaf curing process, digitization and intelligentization are comprehensively achieved, and modern development of tobacco leaf curing can be achieved.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent control of tobacco leaf baking, and specifically relates to a tobacco leaf baking management control system for a step-type baking house. Background Art

[0002] With the improvement of the requirements for baking quality in modern agriculture and industry, the requirements for baking ovens are also getting higher and higher; in the prior art, tobacco leaf baking mostly uses a single baking house and a corresponding control system for baking, and the whole process of baking is realized by setting different temperature and humidity stages and corresponding baking times in a baking house.

[0003] For example, the invention patent with the application publication number [CN106418636A] discloses a tobacco leaf baking device with a mobile combined tobacco baking rack. The mobile hanging tobacco baking rack moves on the track in the baking oven through pulleys, and a baking oven is equipped with multiple mobile hanging tobacco baking racks. This solution improves the tobacco baking efficiency and the quality of tobacco leaf baking by configuring multiple tobacco baking racks in a baking oven; the utility model patent with the publication number [CN220933382U] discloses a remote monitoring control system for a digital tobacco leaf baking house, which realizes the remote monitoring of the environmental information in the tobacco leaf baking house by controlling the sensor devices in the baking house; in addition, the invention patent with the authorization publication number [CN114355857B] discloses an intelligent control system for tobacco leaf baking, including a cloud server and a bulk curing barn terminal, and realizes data communication through Internet of Things technology; this solution can provide recommendations for primary baking processes through data learning and model iteration, and provide services for receiving, processing, and storing Internet of Things data, as well as threshold warning services for key indicators including the temperature, humidity, yellowing degree of tobacco leaves, and water loss during the tobacco leaf baking process.

[0004] However, such a tobacco leaf baking mode does not have a standardized and normalized assembly line production, and the manual labor is large, the overall quality control is difficult, the adaptability of the baking scenario to automation equipment is low, the historical data preservation and comprehensive analysis capabilities are low, and it is not intuitive enough. In addition, the existing baking houses lack energy-saving strategies and are prone to waste energy.

[0005] Therefore, there is an urgent need to propose a step-type baking house and design a corresponding baking control system to establish a multi-functional baking factory with unified standards, scientific management, and high intelligence, so as to empower agricultural production with efficient and intelligent digital technology. Summary of the Invention

[0006] In order to solve the defects of poor scalability of the control system of a single tobacco curing barn equipment and difficulty in realizing mass production in the prior art, a tobacco leaf baking management control system for a step-by-step curing barn is proposed. The curing barn adopts a 1+N control topology structure. By dividing the entire baking process into multiple stages according to the requirements of the baking process, the multiple stages correspond to multiple curing barns, and each stage is realized in a fixed curing barn, so as to realize segmented baking assembly line operation, as well as the mass production and large-scale production of tobacco leaf baking, and improve the product consistency and stability.

[0007] The present invention is realized by adopting the following technical solutions: A tobacco leaf baking management control system for a step-by-step curing barn, the step-by-step curing barn includes N+1 curing barns connected in series, the curing barns are connected to a heat pump, the bottoms of the curing barns are connected through guide rails, an observation window and a touch screen are arranged on the door of each curing barn, a temperature and humidity sensor and two cameras arranged diagonally up and down are arranged inside the curing barn, the interior of the step-by-step curing barn is integrally connected, the first N curing barns are baking areas, and the last curing barn is a moisture regain chamber. A baking basket and a corresponding dragging module are arranged in each curing barn. The dragging module includes a push rod, an air pump, a dragging motor and a chain. The push rod is arranged beside the guide rail and is connected to the air pump. When the air pump is inflated, the push rod is lifted, and when the air is released, the push rod retracts. The push rod is pulled by the dragging motor and the chain. When dragging, the chain drives the push rod on the guide rail to move, and then drives the baking basket to move forward, so as to divide the entire baking process into multiple stages according to the requirements of the baking process, and each stage is realized in a fixed curing barn; a pulley is arranged at the bottom of the baking basket, and the main control module controls the dragging module to drive the pulley to slide along the guide rail. One side of the baking basket is a heat insulation board, and the heat insulation board adopts a polyurethane board, which can not only insulate heat but also achieve the purpose of dividing the whole curing barn. The front and back sides of the curing barn are also made of polyurethane boards in the same way; specifically during work, as the baking state and time of the baking object change, the baking baskets enter different baking areas in sequence from front to back. Different baking areas are set with specific temperature, humidity and baking time parameters according to needs. Each baking area undertakes different tasks. The tobacco leaves start baking from entering the first curing barn and reach the last baking area step by step, that is, the assembly line baking is completed;

[0008] The baking management control system includes an Internet of Things cloud platform, a main control module and N sub-control modules. The main control module and the sub-control modules both adopt PLC controllers. The Internet of Things cloud platform and the main control module are located in the central control room, and the former are both connected to the sub-control modules through a switch. The main control module is used to monitor the working status of each sub-control module and the dragging of the baking basket, and upload data to the Internet of Things cloud platform through the switch; the sub-control module is used to realize the management and control of a single curing barn, and the Internet of Things cloud platform is used to realize data storage, analysis, remote control and abnormal alarm.

[0009] Further, the baking management control system further includes a quality detection module, and the main control module includes a parameter adjustment module and a feedback analysis module;

[0010] The quality inspection module is used to identify the appearance of tobacco leaves and extract key features, including color distribution, surface wrinkling degree, and curling degree of the midrib at the leaf tip, to judge the baking quality of tobacco leaves;

[0011] The feedback analysis module compares the detection results of the current baking room with the target quality parameter library, analyzes the source of deviation, and generates the process parameters that need to be adjusted for the previous baking room based on the difference value;

[0012] The parameter adjustment module calculates the required baking time, temperature and humidity, and heating rate through a dynamic feedback algorithm, which is used as a baking guidance strategy for manual decision-making, and feeds the adjusted parameters back to the sub-control module;

[0013] The effect after each adjustment is verified by the quality inspection module of the subsequent baking room, and the result is fed back to the main control module. By recording and storing relevant data, a closed-loop optimization process is formed;

[0014] Among them, the principle of the parameter adjustment module is as follows:

[0015] (1) Data input: Automatically detect the characteristic values, extract the color and integrity score Q through the neural network auto , combined with the manual judgment score Q manual , and obtain the historical process parameters of the current and previous baking rooms from the process parameter database;

[0016] (2) Deviation analysis: Calculate the deviation value ΔQ between the baking quality of the current baking room and the target parameters m :

[0017] ΔQ m =(ω 1 ·Q auto +ω 2 ·Q manual )-Q target

[0018] Among them, ω 1 , ω 2 are weight factors, and Q target is the target quality parameter;

[0019] (3) Introduce a dynamic regulation model based on Bayesian optimization to reduce the number of adjustment attempts and improve the accuracy;

[0020] The influence function of the parameter deviation of the previous baking room on the current baking room is:

[0021] Q m =f(T m-1 , H m-1 , T m-2 , H m-2) + ∈

[0022] Among them, Q m is the quality score of the current baking area, f is a mapping function, T m-1 is the temperature parameter of the previous baking area with the number m - 1, H m-1 is the humidity parameter of the previous baking area with the number m - 1, T m-2 is the temperature parameter of the previous baking area with the number m - 2, H m-2 is the humidity parameter of the previous baking area with the number m - 1, m is the current baking room, m - 1 and m - 2 are the previous baking rooms, and ∈ is the noise interference term;

[0023] According to the target quality parameter Q target , the new parameter combination is optimized as follows:

[0024] (T′ m-1 , H′ m-1 , T′ m-2 , H′ m-2 ) = argmin‖Q m - Q target ‖

[0025] Among them, T′ m-1 is the optimized temperature setting value of the m - 1 baking area, H′ m-1 is the optimized humidity setting value of the m - 1 baking area, T′ m-2 is the optimized temperature setting value of the m - 2 baking area, H′ m-2 is the optimized humidity setting value of the m - 2 baking area;

[0026] (4) Closed - loop control: Send the optimized new parameter combination to the corresponding sub - control module of the previous baking room. After the next round of baking, re - detect and record the adjustment effect, and update the parameter influence weights in the model.

[0027] Furthermore, to achieve precise control of the temperature and humidity in the baking room and real - time data management, during the baking process, different stages require different temperature - humidity curves and need to be dynamically adjusted according to real - time data. The main control module also designs a PID - fuzzy hybrid control strategy to achieve precise control of temperature and humidity according to the environmental characteristics of different baking rooms, where:

[0028] The PID control adopts a proportional - integral - derivative strategy. By performing fuzzy processing on the real - time error e and the error change rate Δe, combined with the control rule base, a control compensation value is dynamically generated:

[0029] (1) Fuzzy processing: Map the error e and the error change rate Δe to fuzzy sets:

[0030] e ∈ {Large, Medium, Small}

[0031] Δe ∈ {Fast, Slow}

[0032] (2) Construct a fuzzy rule base: If e is Medium and Δe is Fast, then output a medium increment. If e is Large and Δe is Slow, then output a high increment. Defuzzify the fuzzy output value by the weighted average method to generate an available control compensation amount;

[0033] (3) Use the compensation value of fuzzy control to dynamically adjust the PID gain parameters K p , K i , K d , and the adjustment strategy is as follows:

[0034] During the heating-up stage: Dynamically increase K p , speed up the response speed, and at the same time suppress integral saturation by adjusting K i ;

[0035] During the constant-temperature stage: Dynamically decrease K p , increase K d , and enhance stability;

[0036] During the cooling-down stage: Dynamically adjust K i to prevent fluctuations caused by error accumulation;

[0037] Furthermore, the sub-control module generates an adaptive temperature and humidity curve by combining the collected temperature and humidity data, and dynamically adjusts the current temperature and humidity curve by real-time monitoring of the material state and combining with the preset target curve.

[0038] Further, a priority scheduling mechanism is introduced between the main control module and the sub-control modules to ensure the correctness of the control logic;

[0039] Specifically, a master-slave multi-threaded architecture is adopted. The main control module runs multi-threaded tasks, establishes independent communication channels with each sub-control module, realizes high-speed data transmission through the TCP / IP protocol, and designs a double-buffer queue with a distributed data synchronization mechanism between the sub-control modules and the main control module to handle the concurrency and real-time problems of data transmission; The data management of the buffer adopts a timestamp and version number mechanism to ensure data consistency.

[0040] Further, the main control module has a high priority to switch the sub-control modules, and each sub-control module can read and set the baking temperature, humidity and baking time, and generate a real-time baking curve to be displayed on the touch screens of each baking room;

[0041] The sub-control module communicates with the heat pump via 485. The heat pump is connected to the temperature and humidity sensors inside the curing barn. After the heat pump obtains the data measured by the temperature and humidity sensors, it transmits the data to the sub-control module. The sub-control module sets the required temperature, humidity, and baking time and sends them to the heat pump. After the heat pump reads them, it operates through its own setting program.

[0042] Among them, after the heat pump receives the data from the temperature and humidity sensors, it optimizes the data of the temperature and humidity sensors in real time through an improved Kalman filtering algorithm. First, the heat pump obtains real-time temperature and humidity data, and then eliminates the outliers that significantly exceed the normal range; then predicts the current state value, corrects it in combination with the observed value, and outputs the estimated state value; subsequently, compares the filtered data with the target value to generate an error curve; finally, transmits the optimized temperature and humidity data to the sub-control module.

[0043] Furthermore, when generating the baking curve, after the sub-control module reads the temperature, humidity, and baking time, it generates the baking curve by curve fitting and saves the data. By clicking on the user interface to enter the baking curve viewing interface, combined with the trend graph control on the touch screen, the real-time baking curve is displayed.

[0044] Furthermore, the baking management control system further includes a user terminal. The user terminal includes a PC side and a mobile side, which are used to view and operate the curing barn system. Users can view the temperature, humidity, operating parameters, and image data status of each curing barn in real time, and can modify the temperature and humidity parameters and adjust the baking time.

[0045] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0046] This solution adopts a 1+N topology structure, with the main control module for unified management and the sub-control modules for independent operation, which can achieve:

[0047] (1) Pipeline-type baking and batch production: By dividing the entire baking process into multiple stages according to the requirements of the baking process, each stage is implemented in a fixed curing barn, and each curing barn independently controls a specific temperature and humidity stage, realizing segmented baking pipeline operation, as well as batch and large-scale tobacco leaf baking, improving product consistency and stability;

[0048] (2) Full-process dynamic process optimization and intelligent feedback closed-loop: The step-type curing barn uses the baking effect of each batch as a feedback mechanism to achieve closed-loop optimization control, detects whether the process of the current curing barn (numbered m) reaches the expected quality, and transmits the data to the previous curing barns (such as m-1, m-2) for optimization and adjustment, so as to dynamically adjust the process parameters during production, reduce the quality fluctuations caused by parameter deviations in the previous stage, and significantly improve the overall quality of the product;

[0049] (3) Energy conservation and optimization of distributed energy allocation: Each tobacco barn is independently equipped with an air source heat pump, and the power is configured according to the technological requirements of the tobacco baking area. In the front tobacco baking area, the technological requirements are relatively low (the temperature is at a relatively low level), and a low-power heat pump is used for operation to save energy. In the rear tobacco baking area, the technological requirements are relatively high (high temperature demand), and a high-power heat pump is configured to meet the baking needs. The step-by-step design significantly reduces the overall operating cost and energy consumption through stage-by-stage power distribution. Description of the Drawings

[0050] Figure 1 It is a schematic structural diagram of the step-by-step tobacco barn according to an embodiment of the present invention;

[0051] Figure 2 It is a schematic diagram of the baking basket according to an embodiment of the present invention;

[0052] Figure 3 It is a schematic diagram of the framework of the intelligent management and control system of the step-by-step tobacco barn according to an embodiment of the present invention;

[0053] Among them, 1. Tobacco barn; 11. Observation window; 12. Touch screen; 2. Heat pump; 3. Moisture regain chamber; 4. Guide rail; 5. Baking basket; 51. Heat insulation board Detailed Embodiment

[0054] In order to more clearly understand the above-mentioned objects, features and advantages of the present invention, the present invention will be further described below in conjunction with the drawings and embodiments. Many specific details are set forth in the following description in order to fully understand the present invention. However, the present invention may be implemented in other ways different from those described herein. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0055] Embodiment, a tobacco leaf baking management and control system for a step-by-step tobacco barn, relying on the digital step-by-step tobacco barn structure, establishes a matching 1+N topology-based intelligent control system. Specifically, as Figure 1 shown, the step-by-step tobacco barn includes N+1 tobacco barns 1 connected in series. The tobacco barn 1 is connected to the heat pump 2. An observation window 11 and a touch screen 12 are arranged on each tobacco barn door. Temperature and humidity sensors and a monitoring system are arranged inside the tobacco barn 1. The monitoring system includes two high-resolution cameras arranged diagonally up and down. The interior of the step-by-step tobacco barn is integrally connected. The first N tobacco barns are baking areas, and the last tobacco barn is a moisture regain chamber 3. The bottoms of the N tobacco barns are connected through a guide rail 4. The 1+N topology design is specifically for such a step-by-step tobacco barn. Combining Figure 3 , the baking management and control system includes an Internet of Things cloud platform, a main control module and N sub-control modules, and both the main control module and the sub-control modules adopt PLC controllers.

[0056] Among them, the main control module is located in the central control room and is mainly responsible for monitoring the working status of each sub-control module and the dragging of the baking baskets. It realizes the centralized management of the sub-control modules through a switch, and uploads data to the Internet of Things cloud platform through the switch. In this embodiment, the RS485 network cable communication protocol is used for communication to ensure the stability and anti-interference ability of the system during large-scale deployment. The data is uploaded to the Internet of Things cloud platform through the HTTP protocol to ensure the real-time performance and compatibility of the transmission. Each sub-control module is connected to the main control module through an RS485 communication line. The main control module reads the working status of the sub-control module, including whether it is working and whether the communication is normal. Moreover, the main control module has a high priority and can switch the sub-control module on and off. Each sub-control module can read and set the baking temperature and humidity and the baking time, and generate a real-time baking curve for display on the touch screen monitors of each baking room; the sub-control module communicates with the heat pump at the back end through 485, and the heat pump is connected to the temperature and humidity sensors inside the baking room. After the heat pump obtains the data, it transmits it to the sub-control module. The sub-control module can set the required temperature, humidity and baking time and send them to the heat pump. After the heat pump reads them, it will work through its own setting program.

[0057] In this embodiment, the communication and coordination mechanism between the main control module and the sub-control module has specific particularities. In the technical implementation, the communication between the main control module and multiple sub-control modules is a key link, and it is necessary to ensure the real-time performance and reliability of multi-module communication: in the step-type baking room system, the operation data and control instructions of multiple sub-control modules need to be uploaded to the main control module in a timely manner. At the same time, the main control module also needs to issue real-time control instructions. Especially in a complex environment, it is necessary to ensure that there is no delay and no packet loss in communication.

[0058] Here, considering the issues of instruction conflict and priority management, the main control module and the sub-control module may issue conflicting instructions (such as heat pump control) in some cases. Here, it is necessary to introduce a priority scheduling mechanism to ensure the correctness of the control logic. This solution adopts a multi-thread communication architecture, that is, a master-slave multi-thread architecture. The main control module runs multi-thread tasks, establishes independent communication channels with each sub-control module, realizes high-speed data transmission through the TCP / IP protocol, and designs a dual-buffer queue with a distributed data synchronization mechanism between the sub-control module and the main control module to handle the concurrency and real-time performance issues of data transmission. The data management of the buffer can adopt the timestamp and version number mechanism to ensure data consistency. In addition, a redundant communication mechanism is designed. To cope with network failures, a set of backup communication links based on the 485 bus are designed. When the main communication network fails, it can automatically switch to the backup link.

[0059] In addition, the sub-control module is used to manage and control a single baking room, including temperature and humidity control, baking time control, generating real-time baking curves, etc. Among them, when generating the baking curve, after the PLC reads the temperature, humidity and baking time, through internal algorithm processing, by means of curve fitting, the baking curve is generated and the data is saved. When collecting the temperature and humidity parameters, through the designed user interface on the touch screen, click to enter the baking curve viewing interface. A trend graph control is set on the touch screen to display the real-time baking curve. The touch screen uses a capacitive screen, and the interface is set with a home page, a temperature and humidity setting interface, a baking curve viewing interface, etc. The home page includes the current baking temperature and humidity and the set temperature and humidity range, and the baking time. The temperature, humidity and baking time setting interface includes the set range of temperature and humidity, the heating rate, the constant temperature time, etc.

[0060] In a complex environment (such as high humidity and drastic temperature changes), the sensor data may fluctuate or have errors, affecting the execution accuracy of the heat pump. Therefore, in order to optimize the communication between the sub-control module and the heat pump, this embodiment proposes a data filtering and error correction algorithm. After the heat pump receives the sensor data, the improved Kalman filtering algorithm is used to optimize the sensor data in real time, eliminating the noise signal and improving the accuracy of the data. The specific principle is as follows:

[0061] (1) Optimization of the state model and measurement model

[0062] Regarding the sampling values of the temperature and humidity sensors as the measurement values of the system, a state space model is constructed:

[0063] State equation: Describes the changing trend of the temperature and humidity of the system, considering the influence of external disturbances.

[0064] x k+1 =A k x k +B K u k +ω k

[0065] Measurement equation: Describes the relationship between the observed value and the true value of the sensor.

[0066] z k =H k x k +ν k

[0067] Among them, ω k and v k respectively represent the system process noise and the observation noise, A k is the state transition matrix, which describes the changing trend of the temperature and humidity state from time step k to k + 1. Usually, it is established through physical models or empirical formulas, reflecting the internal dynamics of the system; B KControl input matrix, representing control input u k Effect on the system state; x k Represents the state vector of the system, here referring to the states of temperature (T) and humidity (H) in the baking environment; u k Control input, representing an external input or control variable (e.g., the set value or adjustment of the baking equipment) used to affect the system state..

[0068] (2) Noise adaptive adjustment

[0069] Traditional Kalman filtering assumes that the covariance matrix of noise is a fixed value. However, in the actual baking environment, the noise intensity changes with time and environment. The improved algorithm dynamically adjusts the noise covariance matrices Q and R by real-time monitoring of the residual mean to enhance the adaptability of the algorithm. Q represents the covariance matrix of process noise, describing the randomness of the system dynamics. It is related to the noise ω k in the state equation, representing the unpredictable changes in the system. R represents the covariance matrix of observation noise, describing the uncertainty of the measurement noise v k In practical applications, the sensor error changes with the environment, so real-time adjustment is required. The dynamic adjustment of the noise covariance matrices Q and R depends on the residual mean calculated in real time. By monitoring the residuals of the system (the difference between the actual measured value and the predicted value), the intensity of the noise is estimated and the covariance matrix is adaptively adjusted.

[0070] (3) Enhanced robustness

[0071] When there are abnormal fluctuations or mutations in the data, directly using the observed values may introduce large errors. During the baking process, the changes in temperature and humidity should be continuous and stable. However, if the sensor fails or there are external interferences, the observed data may show sudden deviations or fluctuations, which may not reflect the true state of the system. The improved algorithm combines the sliding window method and robust statistics to eliminate or smooth the outliers and improve the anti-interference ability.

[0072] (4) Multi-sensor fusion

[0073] If multiple sensors are configured in a single baking room, the improved algorithm fuses the multi-sensor data with weights to reduce the risk of single-point failure. Fusion strategy:

[0074]

[0075] where ω i is the weight of the i-th sensor, dynamically adjusted based on the real-time signal quality of the sensor.

[0076] (5) Implementation process

[0077] First, the heat pump receives real-time temperature and humidity data from sensors; then it eliminates outliers that are significantly beyond the normal range; next, it predicts the current state value, corrects it by combining the observed value, and outputs the estimated state value; subsequently, it compares the filtered data with the target value to generate an error curve for subsequent optimization of the control algorithm; finally, it transfers the optimized temperature and humidity data to the PLC.

[0078] Considering that traditional baking houses cannot adjust the baking process according to real-time detection results, it is easy to cause the accumulation of previous problems and affect the quality of the final product. Aiming at the quality problems caused by process deviations in the previous baking areas in the pipeline-type step-by-step baking house, this solution realizes dynamic adjustment of the baking time and temperature and humidity parameters in the previous baking areas to ensure the stability of the overall quality of the baking house. The step-by-step baking house can use the numbers of each batch of baked products as a feedback mechanism. For example, it detects whether the process in the current baking area (number m) reaches the expected quality and transfers the data to the previous baking areas (such as m - 1, m - 2) for optimization and adjustment. To dynamically adjust the process parameters during production, reduce the quality fluctuations caused by parameter deviations in the previous stage, and significantly improve the overall quality of the product. The principle of the dynamic feedback control loop is as follows:

[0079] The baking management control system also includes a quality detection module. The main control module includes a parameter adjustment module and a feedback analysis module. The quality detection module uses the trained neural network algorithm to identify the appearance of tobacco leaves and extract key features (such as color distribution, surface wrinkling degree, curling degree of leaf tips and leaf veins, etc.); then there is manual auxiliary observation. The quality and baking degree of tobacco leaves are observed and judged through the monitoring system and the observation window to enhance the detection accuracy; the feedback analysis module compares the detection results of the current numbered baking house (m) with the target quality parameter library, analyzes the source of deviation, and generates the process parameters that need to be adjusted for the previous baking houses (m - 1, m - 2) based on the difference value (such as the characteristics of under-baking or over-baking); the parameter adjustment module calculates the required adjusted baking time, temperature and humidity, and heating rate through the dynamic feedback algorithm, which is used as a baking guidance strategy for manual auxiliary decision-making, and feeds the adjusted parameters back to the sub-control module. The effect of each adjustment is verified by the quality detection module of the subsequent baking house, and the results are fed back to the main control module, and the relevant data are recorded and stored to form a closed-loop optimization process. After long-term operation, the algorithm further improves the accuracy of parameter adjustment through learning and optimization of historical data.

[0080] Among them, the principle of the parameter adjustment module is as follows:

[0081] (1) Data input: Automatically detect the characteristic values, such as the scores of color, integrity, etc. (such as Q auto ) extracted by the neural network. Manual judgment score: Qualitative feedback based on manual observation (such as Q manual) Obtain the historical process parameters of the current and previous baking zones from the process parameter database (such as temperature and humidity set values, time). The process parameter database is a system for storing historical baking process data, which contains various control parameters used during baking. The control parameters include: baking process parameters such as the set temperature and baking time at each baking stage.

[0082] (2) Deviation analysis: Calculate the deviation value between the quality of the current baking zone and the target parameter:

[0083] ΔQ m =(ω 1 ·Q auto +ω 2 ·Q manual )-Q target

[0084] Where ω 1 , ω 2 are weight factors dynamically allocated according to the reliability of automatic detection and manual judgment, and Q target is the target quality parameter.

[0085] (3) Introduce a dynamic regulation model based on Bayesian optimization to reduce the number of adjustment attempts and improve accuracy;

[0086] The influence function of the parameter deviations of the previous baking zones m - 1 and m - 2 on the current baking zone is:

[0087] Q m =f(T m-1 , H m-1 , T m-2 , H m-2 ) + ∈

[0088] Where T is the temperature, H is the humidity, and ∈ is the noise interference term.

[0089] According to the target quality parameter Q target , optimize to obtain a new parameter combination:

[0090] (T′ m-1 , H′ m-1 , T′ m-2 , H′ m-2 ) = argmin‖Q m - Q target ‖

[0091] Where T′ m-1 is the optimized temperature set value of baking zone m - 1, H′ m-1 is the optimized humidity set value of baking zone m - 1, T′ m-2 is the optimized temperature set value of baking zone m - 2, and H′ m-2 is the optimized humidity set value of baking zone m - 2;

[0092] (4) Closed-loop control: Send the optimized new parameters to the corresponding previous baking zone. After the next round of baking, re-detect and record the adjustment effect, and update the parameter influence weights in the model.

[0093] In order to achieve precise control of the temperature and humidity in the baking room and real-time data management, during the baking process, different temperature and humidity curves are required at different stages, and they need to be dynamically adjusted according to real-time data, so as to facilitate optimizing the baking process by analyzing historical data. For this reason, this embodiment proposes a PID-fuzzy hybrid control algorithm. Aiming at the environmental characteristics of different baking rooms, it combines traditional PID control and fuzzy logic control to achieve precise control of temperature and humidity, where:

[0094] PID control adopts the proportional-integral-derivative strategy, and its control output formula is as follows:

[0095]

[0096] where: e(t) is the real-time error of the system, K p , K i , K d are the proportional, integral, and derivative gain parameters respectively. By performing fuzzy processing on the real-time error e and the error change rate Δe, combined with the control rule base, a control compensation value is dynamically generated. The specific principle is as follows:

[0097] (1) Fuzzy processing: Map the error e and the error change rate Δe to fuzzy sets:

[0098] e ∈ {Large, Medium, Small}

[0099] Δe ∈ {Fast, Slow}

[0100] (2) Construct a fuzzy rule base: If e is Medium and Δe is Fast, then output a medium increment. If e is Large and Δe is Slow, then output a high increment. Defuzzify the fuzzy output value by the weighted average method to generate an available control compensation amount.

[0101] (3) Use the compensation value of fuzzy control to dynamically adjust the PID gain parameters K p , K i , K d , and the adjustment strategy is as follows:

[0102] During the heating-up stage: Dynamically increase K p , accelerate the response speed, and at the same time suppress integral saturation by adjusting K i ;

[0103] During the constant-temperature stage: Dynamically decrease K p , increase Kd , enhance stability;

[0104] During the cooling stage: Dynamically adjust K i , prevent fluctuations caused by error accumulation;

[0105] Formula representation:

[0106] K p = K p0 + ΔK p

[0107] K i = K i0 + ΔK i

[0108] K d = K d0 + ΔK d

[0109] Among them, ΔK p , ΔK i , ΔK d is obtained by fuzzy control calculation.

[0110] Furthermore, the sub-control module can generate an adaptive temperature and humidity curve by combining the collected temperature and humidity data. By monitoring the material state (such as humidity and color change) in real time through the system, and combining with the preset target curve, the current temperature and humidity curve can be dynamically adjusted through a machine learning model (such as a neural network or a decision tree).

[0111] The Internet of Things cloud platform is used to implement functions such as data storage, analysis, remote control, and abnormal alarm. It uses a distributed database (such as Hadoop or Elasticsearch) to store real-time data, and combines a data compression algorithm (such as the inter-frame difference method) to improve the storage and query efficiency. The user terminal includes a PC terminal and a mobile terminal, which are used to view and operate the baking system. The user can view the real-time status of each baking room at any time, including temperature and humidity, operating parameters, and image data, and supports modifying temperature and humidity parameters, adjusting the baking time, and archiving data by date and batch for subsequent traceability and analysis; in addition, multiple alarm notification methods can be provided to ensure the timely transmission of abnormal information.

[0112] For the convenience of introduction, this embodiment takes 13 + 1 baking rooms and tobacco leaf baking as an example for introduction. Among them, the first 13 are baking areas, and the last one is a moisture regain chamber. The main control module refers to Figure 2For the PLC framework part, a total of 14 PLC programmable controllers are used in this embodiment. Thirteen of them are distributed in each baking area. A single PLC is connected to a temperature and humidity sensor, a touch screen, a communication network cable, a power supply cable, etc. Among them, on the one hand, the communication network cable transmits signals and working status to the PLC of the main control module, and on the other hand, it communicates with the heat pump and sends control instructions. The PLCs of the 13 sub-control modules respectively control functions such as temperature and humidity reading and setting, baking time regulation, real-time data reading, data saving and downloading, and generation of temperature and humidity curves of a single baking room. The sub-control modules are relatively independent and do not communicate with each other. They operate independently. The hardware (such as PLC, camera, sensor) and software managed by each sub-control module can be replaced and upgraded separately.

[0113] Combined with Figure 1-2 As shown in the figure, a baking basket 5 and a dragging module are arranged inside the step-type baking room. One baking basket is arranged in each baking area, and different baking processes are carried out for the baking baskets in different baking areas. The structure of the baking basket 5 is as shown in Figure 2 As shown in the figure, the bottom of the baking basket 5 is provided with pulleys, and the pulleys are driven by a motor with a reduction gear of the dragging module to slide along the guide rail 4. Specifically, during operation, the main control module controls the action of the dragging module to realize the movement of the baking basket 5 in the step-type baking room. Step-type baking means that as the baking state and time of the tobacco leaves change, they enter different baking areas in sequence from front to back. Each baking area undertakes different tasks. The tobacco leaves start baking from the first one and reach the last one step by step, that is, the baking is completed. For example, the baking temperature in the first baking area is 38 degrees, the temperatures in the second and third baking rooms are 40 degrees, the temperatures in the fourth and fifth baking rooms are 42 degrees, and the temperature in the sixth baking room is 43 degrees, and so on. By setting different temperatures and humidities for different baking rooms, the baking of tobacco leaves is realized. The temperature control is precise, effectively ensuring the baking quality.

[0114] Among them, the design of the dragging module is a relatively mature technical means. For example, a push rod can be arranged beside the guide rail. The push rod is pulled by a dragging motor and a chain. The entire moving process is completed by the dragging module. When it needs to be dragged forward, during dragging, first start the dragging module of the baking basket. The push rod is lifted after being inflated by an air pump and retracted when deflated. After starting, the push rod pushes the pulley at the bottom of the baking basket forward. The dragging instruction is issued by the PLC. The motor power supply is connected to the motor according to the positive rotation phase sequence. The motor rotates forward. The dragging motor pulls the chain according to the set rated power. The chain drives the push rod on the track to move, and then drives the baking basket forward. When retreating, first turn off the air pump, let the push rod fall, click the retreat instruction of the main control module, and reverse the motor by changing the power supply phase sequence to realize retreat.

[0115] In addition, one side of the baking basket 5 is a heat insulation board 51. The heat insulation board 5 is made of a polyurethane board, which can not only insulate heat but also serve the purpose of dividing the overall baking room, so that after the baking basket 5 enters the step-type baking room, it is divided into individual baking rooms; the step-type baking room can serve as N baking areas or a unified large baking area. Each individual baking area can not only undertake the baking tasks at different baking stages but also independently complete the entire baking process; in addition, at the end of this embodiment, the front and rear sides of the baking room are also insulated with polyurethane boards and do not come into contact with the outside world during the baking process. Compared with traditional baking rooms, the design of this solution can achieve standardized baking and refined control, and has strong scalability and compatibility. It can be used for tobacco leaf baking or the baking of other flowers, plants, seeds, etc., and uses a heat pump as the heat source, which is clean and environmentally friendly.

[0116] For the convenience of baking, there are a total of 14 baking rooms in this embodiment, adopting a "1 + 13" topology structure. For example, during the tobacco leaf baking stage, following the "three-stage" baking process, it is divided into 13 baking areas; the design of each individual baking area refers to the construction technical specifications of the bulk curing barn for design, with a length of 6m, a width of 2m, and a height of 3.2m. The main structure of the baking room is spliced with polyurethane boards; a large door and an observation window are added in front of each baking area for convenient equipment maintenance and observation of the changes in the state of tobacco leaves. Overall, a through-rail design is adopted, and a large-torque reduction motor is selected for the dragging motor. The running speed should be relatively slow. As long as the speed is slow, there is no need to consider slow start and pre-braking problems. This baking room has strong scalability and can be used for tobacco leaf baking or as a baking scenario for other baking crops such as flowers, plants, seeds, etc.

[0117] The above are only the preferred embodiments of the present invention and do not limit the present invention in other forms. Any person skilled in the art may use the technical content disclosed above to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A tobacco leaf baking management and control system for a walking-type baking room, characterized in that: The step-type baking room comprises N+1 baking rooms connected in series, the baking rooms are connected to the heat pump, the bottom of the baking rooms are connected through guide rails, an observation window and a touch screen are arranged on each baking room door, a temperature and humidity sensor and two cameras arranged diagonally in the upper and lower directions are arranged in the baking room, the interior of the step-type baking room is connected as a whole, the first N baking rooms are baking areas, and the last baking room is a rehumidification room, and each baking room is provided with a baking basket and a corresponding drag module, so that the entire baking process is divided into multiple stages according to the requirements of the baking process, and each stage is implemented in a fixed baking room; The baking management and control system includes an Internet of Things cloud platform, a main control module and N sub-control modules. The main control module and the sub-control modules both use PLC controllers. The Internet of Things cloud platform and the main control module are located in the central control room. The former are connected to the sub-control modules through switches. The main control module is responsible for monitoring the working status of each sub-control module and the dragging of the baking basket, and uploading the data to the Internet of Things cloud platform through the switch; the sub-control module is used to realize the management and control of a single baking room, and the Internet of Things cloud platform is used to realize data storage, analysis, remote control and abnormal alarm.

2. The tobacco leaf baking management and control system for a walking-type baking room according to claim 1 is characterized in that: A pulley is provided at the bottom of the baking basket, and the main control module controls the drag module to drive the pulley to slide along the guide rail. During specific operation, the baking basket enters different baking areas from front to back in sequence as the baking state and time of the baked objects change. Different baking areas are set with specific temperature, humidity and baking time parameters as needed. Each baking area undertakes different tasks. The tobacco leaves are baked from the time they enter the first baking room, and reach the last baking area step by step, completing the assembly line baking.

3. The tobacco leaf baking management and control system for a walking-type baking room according to claim 1 is characterized in that: The baking management and control system also includes a quality detection module, and the main control module includes a parameter adjustment module and a feedback analysis module; The quality detection module is used to identify the appearance of tobacco leaves and extract key features, including color distribution, surface wrinkles and curling of leaf tips and veins, so as to judge the baking quality of tobacco leaves; The feedback analysis module compares the detection results of the current baking room with the target quality parameter library, analyzes the source of the deviation, and generates the process parameters that need to be adjusted for the previous baking room based on the difference value; The parameter adjustment module calculates the baking time, temperature and humidity, and heating rate that need to be adjusted through a dynamic feedback algorithm, which serves as a baking guidance strategy for manual assistance in manual decision-making, and feeds back the adjustment parameters to the sub-control module; The effect of each adjustment is verified by the quality inspection module of the subsequent baking room, and the results are fed back to the main control module. By recording and storing relevant data, a closed-loop optimization process is formed; Among them, the principle of the parameter adjustment module is as follows: (1) Data input: Automatically detect feature values, extract color and integrity score Q through neural network auto , combined with manual judgment score Q manual , obtain the historical process parameters of the current and previous baking rooms from the process parameter database; (2) Deviation analysis: Calculate the deviation value ΔQ between the current baking quality and the target parameter m : ΔQ m =(ω1·Q auto +ω2·Q manual )-Q target Among them, ω1, ω2 are weight factors, Q tatget is the target quality parameter; (3) Introducing a dynamic control model based on Bayesian optimization to reduce the number of adjustment attempts and improve accuracy; The influence function of the parameter deviation of the previous baking room on the current baking room is: Q m =f(T m-1 ,H m-1 ,T m-2 ,H m-2 )+∈ Among them, Q m The quality score of the current baking area, f is a mapping function, T m-1 is the temperature parameter of the previous baking zone numbered m-1, H m-1 is the humidity parameter of the previous baking zone numbered m-1, T m-2 is the temperature parameter of the previous baking zone numbered m-2, H m-2 is the humidity parameter of the previous baking area numbered m-1, m is the current baking room, m-1 and m-2 are the previous baking rooms, and ∈ is the noise interference term; According to the target quality parameter Q target , optimize to get a new parameter combination: (T′ m-1 ,H′ m-1 ,T′ m-2 ,H′ m-2 )=argmin‖Q m -Q target ‖ Among them, T′ m-1 Optimized temperature setting value of m-1 baking zone, H′ m-1 Optimized humidity setting value of the m-1 baking zone, T′ m-2 Optimized temperature setting value of the m-2 baking zone, H′ m-2 Optimized m-2 baking zone humidity setting value; (4) Closed-loop control: The optimized new parameter combination is sent to the corresponding previous baking room control module. After the next round of baking, the adjustment effect is re-tested and recorded, and the parameter influence weights in the model are updated.

4. The tobacco leaf baking management and control system for a walking-type baking room according to claim 1, characterized in that: The main control module also designs a PID-fuzzy hybrid control strategy to achieve precise control of temperature and humidity according to the environmental characteristics of different baking rooms, including: PID control adopts proportional integral differential strategy, and dynamically generates control compensation value by fuzzy processing the real-time error e and error change rate Δe, combined with the control rule library: (1) Fuzzy processing: Map the error e and the error change rate Δe into fuzzy sets: e∈{Large,Medium,Small} Δe∈{Fast,Slow} (2) Construct a fuzzy rule base: If e is Medium and Δe is Fast, then output a medium increment; if e is Large and Δe is Slow, then output a high increment. Defuzzify the fuzzy output value using the weighted average method to generate a usable control compensation amount; (3) Use the fuzzy control compensation value to dynamically adjust the PID gain parameter K p , K i , K d , the adjustment strategy is as follows: In the heating stage: Dynamically increase K p , speed up the response, and adjust K i Suppress integral saturation; In the constant temperature stage: Dynamically reduce K p , increase K d , enhance stability; In the cooling stage: Dynamically adjust K i , to prevent fluctuations caused by error accumulation; Furthermore, the sub-control module generates an adaptive temperature and humidity curve based on the collected temperature and humidity data, and dynamically adjusts the current temperature and humidity curve by real-time monitoring of the material status and combining it with the preset target curve.

5. The tobacco leaf baking management and control system for a walking-type baking room according to claim 1 is characterized in that: A priority scheduling mechanism is introduced between the main control module and the sub-control module to ensure the correctness of the control logic; Specifically, a master-slave multi-threaded architecture is adopted. The main control module runs multi-threaded tasks, establishes an independent communication channel with each sub-control module, realizes high-speed data transmission through the TCP / IP protocol, and uses a distributed data synchronization mechanism to design a double buffer queue between the sub-control module and the main control module to handle the concurrency and real-time issues of data transmission; the data management of the buffer adopts a timestamp and version number mechanism to ensure data consistency.

6. The tobacco leaf baking management and control system for a walking-type baking room according to claim 1, characterized in that: The main control module has a high priority to switch the sub-control modules, and each sub-control module can read and set the baking temperature, humidity and baking time, and generate a real-time baking curve to be displayed on the touch screen of each baking room; The sub-control module communicates with the heat pump via 485. The heat pump is connected to the temperature and humidity sensor inside the baking room. The heat pump obtains the data measured by the temperature and humidity sensor and transmits it to the sub-control module. The sub-control module sets the required temperature, humidity and baking time and sends them to the heat pump. After reading, the heat pump works through its own setting program. Among them, after the heat pump receives the temperature and humidity sensor data, the temperature and humidity sensor data is optimized in real time through the improved Kalman filtering algorithm. First, the heat pump obtains real-time temperature and humidity data, and then eliminates abnormal values ​​that are obviously beyond the normal range; then predicts the current state value, combines the observed value for correction, and outputs the estimated state value; then compares the filtered data with the target value to generate an error curve; finally, the optimized temperature and humidity data is passed to the sub-control module.

7. The tobacco leaf baking management and control system for a walking-type baking barn according to claim 6, characterized in that: When generating the baking curve, the sub-control module reads the temperature, humidity and baking time, generates the baking curve through the curve fitting method and saves the data. By clicking on the user interface to enter the baking curve viewing interface, combined with the trend chart control on the touch screen, the real-time baking curve is displayed.

8. The tobacco leaf baking management and control system for a walking-type baking barn according to claim 1, characterized in that: The drag module includes a push rod, an air pump, a drag motor and a chain. The push rod is arranged next to the guide rail and is connected to the air pump. The air pump lifts the push rod after being inflated and retracts the push rod when deflated. The push rod is pulled by the drag motor and the chain. During dragging, the chain drives the push rod on the guide rail to move, thereby driving the baking basket forward.

9. The tobacco leaf baking management and control system for a walking-type baking barn according to claim 1, characterized in that: One side of the baking basket is a heat insulation board, which is made of polyurethane board, which can not only insulate but also achieve the purpose of dividing the whole baking room. Similarly, the front and back sides of the baking room are both made of polyurethane board.

10. The tobacco leaf baking management and control system for a walking-type baking barn according to claim 1, characterized in that: The baking management and control system also includes a user terminal, which includes a PC terminal and a mobile terminal for viewing and operating the baking room system. The user can view the temperature and humidity, operating parameters and image data status of each baking room in real time, and can modify the temperature and humidity parameters and adjust the baking time.

Images