A full-line asynchronous synchronous control system and method for a biscuit production line

Abstract

This application relates to the field of biscuit production technology, providing a fully asynchronous synchronous control system and method for a biscuit production line. The system includes a data acquisition module, a data processing module, a central control module, and an execution module. Dynamic adaptive adjustment is achieved through a closed-loop control process. The data acquisition module collects real-time data on the speed of each conveyor belt segment, biscuit density, and posture. The data processing module constructs and iteratively optimizes a speed control prediction model based on historical and real-time data to calculate the optimal speed for each segment. The central control module generates segmented acceleration commands for smooth rotation in the cooling section, gradual speed commands for smooth transitions between processes, and dynamic adjustment commands for coordinating the density across the entire line, based on the model output. The execution module responds to the commands with precise speed adjustment. This application achieves global synchronous matching of speed and biscuit density across multiple asynchronous conveyor belts, significantly improving production yield and efficiency while reducing material and energy waste.

Description

Technical Field

[0001] This application relates to the field of biscuit production technology, specifically to an asynchronous synchronous control system and method for a biscuit production line. Background Technology

[0002] A typical biscuit production line consists of a front-end feeding system, a rolling system, a pressing system, an intermediate baking system, an oil spraying system, and a rear-end cooling system and packaging workshop, all connected sequentially. Conveyor belts transfer biscuit dough or finished biscuits between these systems. Because different processes have varying requirements for biscuit processing, the conveyor belts for each system require different operating parameters. This results in varying transport speeds and distribution densities of biscuits in different production areas, leading to a series of production challenges that severely impact production yield and increase material and energy waste. Specifically, these include: (1) Challenges in Attitude Control of the Cooling Section: Existing cooling sections mostly employ a bending and flipping conveyor belt technology. This requires the biscuit to accelerate to a specific speed at the first bend, fly horizontally to the vortex to flip, and then land on the second conveyor belt. The second conveyor belt repeats this process to ensure the biscuit flies smoothly to the third conveyor belt. This structure consists of baffles and cantilever arms. However, existing technology cannot precisely control the speed of the cooling section conveyor belt: too high a speed can easily cause the biscuit to break or crack, reducing yield; too low a speed will prevent the biscuit from reaching the bend wall, causing it to fall directly to the next conveyor belt, resulting in attitude disorder. Ultimately, this leads to an irregular distribution of the biscuit's front and back sides after leaving the cooling section, such as... Figure 1 As shown, this significantly increases the difficulty of the packaging process and the complexity of the production steps.

[0003] (2) Challenges in connecting the cooling section and the packaging workshop: After cooling, the biscuits are transported at high speed, while the conveyor belt in the packaging workshop runs at low speed to adapt to the packaging operation. The speed difference between the two can easily lead to problems such as biscuit flipping and collision during the connection process, further reducing the product yield and becoming one of the main reasons for the current yield loss. How to accurately control the speed of the connection section so that the biscuits enter the flipping machine with the appropriate posture and speed has become a key pain point for the existing production line.

[0004] (3) The problem of coordinating density and speed of the entire production line: When the biscuit transfer density between each process is too high, it will lead to insufficient cooling capacity in the later stage, shortage of manpower in the packaging workshop, production congestion and material waste; when the density is too low, it will cause fuel waste in the baking workshop and material waste in the middle stage spraying process. In the existing technology, this coordination work relies on experienced workers to manually adjust, and the adjustment process takes 10-20 minutes. During this period, the material waste accounts for more than 3% of the daily yield, which seriously reduces production efficiency. Summary of the Invention

[0005] To help solve the above-mentioned technical problems, this application provides an asynchronous synchronous control system and method for the entire biscuit production line.

[0006] An asynchronous-synchronous control system for a biscuit production line, comprising a data acquisition module, a data processing module, a central control module, and an execution module connected in sequence: The data acquisition module is used to collect basic experience data, experimental data, and real-time production data collected by sensors in biscuit production. The data processing module is used to perform statistical analysis on the experimental data to determine the initial confidence interval, and to calculate the multiple regression equation based on the data within the interval to establish an initial control model; it is also configured to compress and optimize the accumulated real-time production data, and to update the parameters of the model using the optimized data. The central control module is used to generate speed adjustment commands based on the predicted speed output by the initial control model, for controlling the cooling section flipping, process connection, and adjusting the overall line density. The execution module is used to receive the instructions and adjust the running speed of the corresponding conveyor belt.

[0007] The data processing module is used to compress and optimize the accumulated real-time production dataset using the T-distribution method.

[0008] A method for full-line asynchronous-synchronous control of a biscuit production line, wherein the control system comprises a data acquisition module, a data processing module, a central control module, and an execution module, includes the following steps: S1: The data acquisition module collects basic experience data on biscuit production and conducts experiments based on this data to obtain experimental data. The data processing module performs statistical analysis on the conveyor belt speed data in the experimental data, determines the initial confidence interval, and calculates a multiple regression equation based on the experimental data within the initial confidence interval to establish an initial control model for predicting the optimal operating speed of the conveyor belt in each process section, including the cooling section. S2: During the continuous production of biscuits, the data acquisition module collects real-time production data through sensors, the data processing module inputs the real-time production data into the initial control model, and the central control module generates and issues speed adjustment commands based on the predicted speed output by the model. These commands are used to control the biscuit turning process in the cooling section, control the connection process from the cooling section outlet to the packaging process entrance, and adjust the distribution density of biscuits throughout the line. The execution module responds to the commands to adjust the running speed of the corresponding conveyor belt. S3: During the continuous production of biscuits, newly collected real-time production data is continuously fed back to the data processing module. The data processing module compresses and optimizes the accumulated real-time dataset related to biscuit production, and uses the compressed and optimized data to update the coefficients of the multiple regression equation in the initial control model, and performs parameter iteration and optimization of the model.

[0009] The biscuit production line includes processes such as feeding, rolling, pressing, baking, cooling, and packaging. S1 includes: The basic experience data includes the conveyor belt speed range, biscuit density threshold, and critical flipping speed range required for flipping in the cooling section, which are determined by human experience for at least one type of biscuit. The experimental data is obtained by conducting experiments in the cooling section at multiple different conveyor belt speeds set according to the basic experience data, including biscuit bending posture data and distribution density data.

[0010] S1 includes: A nonlinear statistical table with biscuit diameter, thickness, and mass as variables and conveyor belt speed as the dependent variable was constructed. An orthogonal array was used to calculate the multiple regression equation to establish a multiple regression initial control model for predicting the optimal operating speed of the conveyor belt in each process section, including the cooling section.

[0011] S1 includes: The statistical distribution is a chi-square distribution, and the multiple regression equation upon which the initial regulation model is based is: Where K is the comprehensive control coefficient, m is the biscuit mass, S is the biscuit surface area, V1 is the conveyor belt speed, and e is a constant term determined based on empirical data.

[0012] S2 includes: Real-time production data includes the physical parameters of the biscuits, the speed of the conveyor belts in each process, and the distribution density of the biscuits.

[0013] S2 includes: For the flipping process that includes multiple bends in the cooling section, instructions are generated based on the critical flipping speed output by the model to independently adjust the running speed of each conveyor belt in two adjacent conveyor belts. The biscuit obtains the required speed to fly over each bend based on these instructions. For the connection process between the cooling section outlet and the packaging process inlet, an instruction is generated to continuously reduce the speed of the connecting section conveyor belt from the end speed value of the cooling section to the speed value of the packaging process inlet; Based on the real-time feedback of biscuit distribution density, the conveyor belt speed of at least one of the feeding, rolling, pressing, or baking processes is dynamically adjusted to maintain the biscuit density throughout the entire line within a preset threshold range.

[0014] S3 includes: The T-distribution method is used to compress and optimize the accumulated real-time dataset related to biscuit production.

[0015] S3 includes: During the iterative optimization process, the confidence interval of the initial control model is gradually reduced. Through iterative optimization, the confidence interval is gradually reduced from the initial 0.9 to 0.3.

[0016] In summary, this application provides an asynchronous synchronous control method and system for a biscuit production line. In the biscuit production process, through a data-driven dynamic adaptive adjustment mechanism, the speed of each section of the conveyor belt is precisely controlled, so as to achieve smooth biscuit rotation in the cooling section, speed and density matching between each process, reduce yield loss, reduce material and energy waste, and improve production efficiency. Attached Figure Description

[0017] Figure 1 The accompanying drawings are related to the background technology. Figure 2 This is a schematic diagram of the cooling section turning process in this application; Figure 3 A schematic diagram showing the current desired speed for this application; Figure 4 This is a schematic diagram illustrating the evolution of the expected velocity parameters in the future of this application; Figure 5 This is a flowchart illustrating an asynchronous-synchronous control method for a biscuit production line according to this application. Detailed Implementation

[0018] The present application will be further described below with reference to the accompanying drawings. The structure and principle of the present application are very clear to those skilled in the art. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.

[0019] The control system of this application includes: Data acquisition module: Used to collect basic production line data, experimental data, and real-time production data, specifically including: (1) Basic experience data: Collect production experience data of front-line workers for different types of biscuits (such as Wan Nian Xiang Cong Biscuits, Le Zhi Chickpea Crisps, etc.), including the conveyor belt speed range suitable for each process, biscuit density threshold, critical speed for turning in the cooling section, etc. (2) Experimental Data: By comparing basic empirical data with theoretical data, a linear confidence interval based on the chi-square distribution was constructed (confidence value set at 0.9). Ten speed data points were selected from the middle range as experimental parameters and assigned to corresponding sections of the production line. Skilled workers were arranged to judge the distribution density of the biscuits (too dense / too sparse) under each speed parameter. At the same time, time-lapse photography was used to capture the posture data of the biscuits bending at different speeds in the cooling section (stability, flipping success rate, damage, etc.). Figure 2 As shown; (3) Real-time production data: Through sensors (speed sensor, density sensor, attitude sensor) installed at the bends of the conveyor belts, cooling sections, and the junctions between cooling and packaging in each process, the real-time data collection is obtained on the conveyor belt running speed, biscuit distribution density, biscuit attitude parameters, and physical parameters such as the diameter, thickness, and mass of the biscuits.

[0020] Data processing module: This is used to process and model various types of collected data, generating conveyor belt speed control parameters for each process. The specific steps are as follows: (1) Data integration and verification: Integrate basic experience data, experimental data and real-time production data, and remove abnormal data; (2) Model building: Based on the integrated data, a nonlinear relationship statistical table of biscuit diameter, thickness, mass and production line speed is constructed. Combined with empirical formulas and relevant parameters, orthogonal arrays are used to calculate the multiple regression equation and establish a speed control prediction model to estimate the optimal operating speed of each section of the conveyor belt under different production conditions. (3) Data optimization: The dataset is compressed using a T-distribution to form a database with two distinct ends, thereby optimizing data storage and retrieval efficiency and further improving the accuracy of the speed regulation prediction model; (4) Dynamic iteration: As the product line becomes richer and production data accumulates, the newly collected real-time data is continuously input into the prediction model to adjust the speed, thereby achieving dynamic iterative optimization of the model and gradually reducing the confidence interval from 0.9 (e.g., ...). Figure 3 (As shown) Reduced to 0.3 (as shown) Figure 4 As shown in the figure, this ensures that the control parameters are accurately matched with the actual production conditions.

[0021] Central control module: As the core of the system, it is used for data distribution, instruction generation, and coordination control. Specific functions include: (1) Receive the optimal speed control parameters output by the data processing module, and combine them with the real-time production data of each process (such as baking temperature, oil spraying flow rate, and packaging efficiency) to generate speed adjustment instructions for each section of the conveyor belt. (2) For the scenario of bending and flipping in the cooling section, the acceleration instructions of the first and second conveyor belts are accurately generated based on the critical flipping speed parameters output by the data processing module to ensure that the biscuit can fly smoothly to the bend to complete the flipping and maintain a stable posture. (3) For the connection between the cooling section and the packaging workshop, a gradual speed adjustment command is generated to smoothly transition the biscuits from high-speed transport to low-speed packaging, avoiding problems such as flipping and collision. (4) Coordinate the operation of each system in real time, and dynamically adjust the conveyor belt speed of the front-end feeding, rolling, pressing and the intermediate baking and spraying system based on the real-time feedback data of biscuit density, so as to ensure that the biscuit density of the entire production line is within the optimal threshold range.

[0022] Execution module: It includes drive units, speed controllers, and auxiliary execution components (such as baffle adjustment mechanisms) for each section of the conveyor belt, which are used to receive speed adjustment commands from the central control module, accurately adjust the running speed of the corresponding conveyor belt, and realize the asynchronous and synchronous coordinated operation of each process.

[0023] This application provides a fully asynchronous synchronous control method for a biscuit production line. The method is applied to a biscuit production line including feeding, rolling, pressing, baking, cooling, and packaging processes, and is executed by a control system comprising a data acquisition module, a data processing module, a central control module, and an execution module. The method includes the following steps: S1.1 Collect basic experience data and experimental data on biscuit production; the basic experience data includes the conveyor belt speed range, biscuit density threshold, and critical flipping speed range required for flipping in the cooling section, which are determined by human experience for at least one type of biscuit; the experimental data is obtained by conducting experiments in the cooling section at multiple different conveyor belt speeds set according to the basic experience data, including biscuit bending posture data and distribution density data. S1.2 Perform chi-square distribution analysis on the conveyor belt speed data in the experimental data to determine its initial confidence interval; S1.3. Based on the experimental data within the initial confidence interval, an initial control model is established using multiple regression analysis with biscuit diameter, thickness, and mass as inputs to predict the optimal operating speed of the conveyor belt in each process segment, including the cooling section. S2. During the operation of the biscuit production line, real-time production data including biscuit physical parameters, conveyor belt speeds of each process, and biscuit distribution density are collected by sensors and input into the initial control model. The central control module generates and issues speed adjustment commands based on the predicted speed output by the model to control the biscuit turning process in the cooling section, control the connection process from the cooling section outlet to the packaging process inlet, and adjust the biscuit distribution density of the entire line. The execution module responds to the commands to adjust the running speed of the corresponding conveyor belt. S3. During the continuous production of biscuits, newly collected real-time production data is continuously fed back to the data processing module, and the following optimization process is executed: S3.1. The accumulated real-time dataset related to biscuit production is compressed and optimized using the T-distribution method; S3.2. Update the coefficients of the multiple regression equation in the initial control model using the compressed and optimized data to achieve parameter iteration and optimization of the model.

[0024] based on Figure 5The asynchronous synchronous control method for a biscuit production line described in this application is implemented through a continuously running closed-loop control process, which is as follows: First, basic data acquisition and model planning are performed, that is, basic experience data and theoretical data of the production line are collected and integrated to construct the confidence interval of the initial speed parameters, and experimental schemes are planned accordingly; on this basis, experimental data acquisition is carried out, and experiments are conducted according to the planned speed parameters to obtain the distribution density of biscuits and the turning posture data of the cooling section under different operating conditions; then, a regression model is established, and the basic data and experimental data are integrated to construct and initialize a multivariate regression model for predicting the optimal operating speed of each section of the conveyor belt; after entering the formal production stage, the system... The system's execution module responds to and provides production status feedback: Based on the speed parameters output by the model, it controls the production line operation and collects real-time production status data such as conveyor belt speed, biscuit density, and posture through sensors. This real-time data enters the data acquisition and processing stage, where it is integrated and verified with historical data. The central control module then analyzes this data and dynamically generates and issues speed adjustment commands to achieve speed and density matching between processes. Simultaneously, the system continuously iterates and optimizes the model, accumulating production feedback data and using methods such as the T-distribution to compress and optimize the dataset. This data is then used to update the parameters of the speed control prediction model, gradually narrowing its prediction confidence interval and continuously improving control accuracy. This closed-loop process achieves dynamic adaptive and coordinated control of the production line's speed and density.

[0025] In this embodiment, encoders are installed on the drive shafts or driven shafts of the conveyor belts that require precise speed control, such as in the cooling section, baking section, and packaging connection section, to measure the shaft rotation angle. To overcome the cumulative errors caused by wear and uneven pitch of the conveyor belt links and to achieve accurate tracking of the biscuit position, a high-precision speed and position detection scheme is preferably adopted.

[0026] Specifically, referring to existing technology (e.g., CN104220348A), two encoder shafts equipped with sprockets are set at adjacent positions on the conveyor belt to be tested. The sprockets on the drive shafts mesh with the links of the conveyor belt or the auxiliary drive chain. By measuring the rotation angles of the two shafts separately using two encoders and calculating their relative rotation difference, the instantaneous pitch of the conveyor belt links between the two sprockets can be determined. Combined with the number of links passing per unit time, the instantaneous precise linear velocity of that section of the conveyor belt can be calculated, with an accuracy far exceeding that of traditional methods relying solely on a single drive shaft encoder. This precise speed data is transmitted to the data acquisition module in real time.

[0027] To achieve accurate acquisition of the operating speed of a single conveyor belt, this application cites prior art methods for determining the precise linear velocity of a conveyor belt by measuring the instantaneous pitch of the conveyor belt using a dual-encoded sprocket shaft or an optical sensor. It should be specifically noted that the invention and technical solutions disclosed in the cited document are entirely focused on solving the localized and fundamental problem of high-precision measurement of the speed and position of a single conveying unit.

[0028] The core inventive point and the technical problem to be solved in this application is how to use real-time data from multiple processes and types of sensors on the production line to build and iteratively optimize a speed control prediction model through a unique data processing module, and generate dynamic coordination instructions through a central control module, so as to achieve global and adaptive synchronization of speed and biscuit distribution density of each independent and asynchronous process segment on the entire production line.

[0029] Specifically, this application addresses the system-level coordination challenges arising from multiple independently driven sections and varying process requirements in a biscuit production line, rather than the measurement accuracy problem of a single conveyor belt. This application proposes a complete closed-loop control system architecture encompassing data acquisition, processing, central control, and execution, with a data-driven adaptive control model at its core. The speed measurement scheme in the cited document is merely one feasible implementation within the "data acquisition module" of this application and does not involve any control logic related to multi-section coordination, model prediction, or dynamic iteration. The ultimate effect achieved by this application is an improvement in the overall production efficiency, a result of emergent optimization, far beyond what can be directly achieved by improving the measurement accuracy of a single link. Therefore, even with the cited document, those skilled in the art will not gain any systematic technical inspiration on how to construct a system capable of achieving dynamic synchronous control of asynchronous processes across the entire production line. Regarding density detection, photoelectric sensor arrays or vision inspection devices, such as industrial cameras, are installed across the width of the conveyor belt at key stages such as after rolling, before baking, and after cooling. By counting the number of cookies blocking the light beam per unit time or identifying the number of cookies within the statistical field of view through image processing algorithms, and combining this with the accurately measured conveyor belt speed, the linear distribution density of the cookies can be calculated in real time.

[0030] Regarding attitude detection, high-speed vision sensors or time-lapse photography equipment are installed at the bending and flipping points of the cooling section and at the junction with the packaging workshop. These devices are installed perpendicular to the conveyor belt plane to continuously capture the movement trajectory of the cookies in the air or their state when they land on the conveyor belt. Through image processing and analysis, attitude parameters such as the success rate of cookie flipping, flight angle offset, and whether collisions or stacking have occurred can be obtained.

[0031] After receiving speed adjustment commands from the central control module, the execution module controls the speed by adjusting the rotational speed of the drive motors of each conveyor belt segment. The drive motors are preferably equipped with servo drivers or frequency converters to achieve precise and rapid speed response. For specific areas in the cooling section where the biscuits need to be flipped horizontally, the drive motors and transmission systems must possess good dynamic response characteristics to execute the segmented acceleration commands issued by the central control module. The conveyor belt itself can be a chain-plate conveyor belt or a flat belt; its drive method (such as sprocket drive or friction roller drive) is well-known in the art and will not be described further here.

[0032] Real-time data collected by all sensors is transmitted to the data processing module via an industrial bus or analog / digital I / O module. The central control module processes the information and generates speed commands, which are then sent to the speed control actuators of each conveyor belt, such as frequency converters, through the same communication link, thus forming a closed-loop control system that enables asynchronous and synchronous dynamic regulation of speed and density across the entire line.

[0033] In summary, this application has the following advantages: (1) Basic data collection and model initialization: Production experience data of Wannianqing scallion biscuits were collected, including the critical turning speed range of the first conveyor belt in the cooling section (1.2-1.5 m / s), the critical turning speed range of the second conveyor belt (1.1-1.4 m / s), and the optimal biscuit density of the entire production line (20-25 pieces / m³). The empirical data was compared with the theoretically calculated speed-density relationship data, and a 0.9 confidence interval of the chi-square distribution was constructed. Ten speed points were selected equally for the experiment. The bending posture of the biscuits at each speed point was recorded by time-lapse photography, and the workers judged the rationality of the density and collected experimental data. Based on the experimental data, the multiple regression equation was calculated using an orthogonal array: Where K is the comprehensive control coefficient, m is the biscuit mass, S is the biscuit surface area, V1 is the conveyor belt speed, and e is a constant term determined based on empirical data, thus completing the model initialization; (2) Real-time control operation: After the production line is started, the data acquisition module collects the speed of each section of the conveyor belt, the density of the biscuits, the physical parameters of the biscuits and the posture data of the cooling section in real time, and transmits them to the data processing module; the data processing module calculates the optimal speed of each section through the optimized multiple regression model and outputs it to the central control module; the central control module generates adjustment instructions: the speed of the first section of the cooling section is adjusted to 1.35m / s and the speed of the second section of the cooling section is adjusted to 1.28m / s to ensure that the biscuits are turned over smoothly; the speed of the cooling section outlet conveyor belt gradually decreases from 1.2m / s to 0.3m / s to match the speed of the packaging workshop conveyor belt; according to the real-time data of the biscuit density (such as the detection of a density of 30 pieces / m, which exceeds the optimal threshold), the central control module instructs the speed of the front-end feeding system conveyor belt to decrease from 0.8m / s to 0.6m / s, and at the same time adjusts the speed of the baking system conveyor belt from 1.0m / s to 0.8m / s, so that the biscuit density is restored to 22 pieces / m; (3) Model iteration optimization: Continuously collect real-time data of different production batches, use T-distribution to compress the dataset, update the parameters of the multivariate regression model, and gradually narrow the confidence interval to 0.3 to improve the control accuracy.

Claims

1. A fully asynchronous synchronous control system for a biscuit production line, characterized in that, It includes a data acquisition module, a data processing module, a central control module, and an execution module, which are connected in sequence: The data acquisition module is used to collect basic experience data, experimental data, and real-time production data collected by sensors in biscuit production. The data processing module is used to perform statistical analysis on the experimental data to determine the initial confidence interval, calculate the multiple regression equation based on the data within the interval, establish the initial control model, compress and optimize the accumulated real-time production data, and update the parameters of the model using the optimized data. The central control module is used to generate speed adjustment commands based on the predicted speed output by the initial control model, for controlling the cooling section flipping, process connection, and adjusting the overall line density. The execution module is used to receive the instructions and adjust the running speed of the corresponding conveyor belt.

2. The system according to claim 1, characterized in that, The data processing module is used to compress and optimize the accumulated real-time production dataset using the T-distribution method.

3. A method for full-line asynchronous-synchronous control of a biscuit production line, characterized in that, The process is executed by a control system comprising a data acquisition module, a data processing module, a central control module, and an execution module, and includes the following steps: S1: The data acquisition module collects basic experience data on biscuit production and conducts experiments based on this data to obtain experimental data. The data processing module performs statistical analysis on the conveyor belt speed data in the experimental data, determines the initial confidence interval, and calculates a multiple regression equation based on the experimental data within the initial confidence interval to establish an initial control model for predicting the optimal operating speed of the conveyor belt in each process section, including the cooling section. S2: During the biscuit production process, the data acquisition module collects real-time production data through sensors, the data processing module inputs the real-time production data into the initial control model, and the central control module generates and issues speed adjustment commands based on the predicted speed output by the model. These commands are used to control the biscuit turning process in the cooling section, control the connection process from the cooling section outlet to the packaging process entrance, and adjust the biscuit distribution density throughout the line. The execution module responds to the commands to adjust the running speed of the corresponding conveyor belt. S3: During the biscuit production process, newly collected real-time production data is continuously fed back to the data processing module. The data processing module compresses and optimizes the accumulated real-time dataset related to biscuit production, and uses the compressed and optimized data to update the coefficients of the multiple regression equation in the initial control model, and performs parameter iteration and optimization of the model.

4. The method according to claim 3, characterized in that, The biscuit production line includes processes such as feeding, rolling, pressing, baking, cooling, and packaging. S1 includes: The basic experience data includes the conveyor belt speed range, biscuit density threshold, and critical flipping speed range required for flipping in the cooling section, which are determined by human experience for at least one type of biscuit. The experimental data is obtained by conducting experiments in the cooling section at multiple different conveyor belt speeds set according to the basic experience data, including biscuit bending posture data and distribution density data.

5. The method according to claim 3, characterized in that, S1 includes: A nonlinear statistical table with biscuit diameter, thickness, and mass as variables and conveyor belt speed as the dependent variable was constructed. An orthogonal array was used to calculate the multiple regression equation to establish a multiple regression initial control model for predicting the optimal operating speed of the conveyor belt in each process section, including the cooling section.

6. The method according to claim 3, characterized in that, S1 includes: The statistical distribution is a chi-square distribution, and the multiple regression equation upon which the initial regulation model is based is: Where K is the comprehensive control coefficient, m is the biscuit mass, S is the biscuit surface area, V1 is the conveyor belt speed, and e is a constant term determined based on empirical data.

7. The method according to claim 3, characterized in that, S2 includes: Real-time production data includes the physical parameters of the biscuits, the speed of the conveyor belts in each process, and the distribution density of the biscuits.

8. The method according to claim 3, characterized in that, S2 includes: For the flipping process that includes multiple bends in the cooling section, instructions are generated based on the critical flipping speed output by the model to independently adjust the running speed of each conveyor belt in two adjacent conveyor belts. The biscuit obtains the required speed to fly over each bend based on these instructions. For the connection process between the cooling section outlet and the packaging process inlet, an instruction is generated to continuously reduce the speed of the connecting section conveyor belt from the end speed value of the cooling section to the speed value of the packaging process inlet; Based on the real-time feedback of biscuit distribution density, the conveyor belt speed of at least one of the feeding, rolling, pressing, or baking processes is dynamically adjusted to maintain the biscuit density throughout the entire line within a preset threshold range.

9. The method according to claim 3, characterized in that, S3 includes: The T-distribution method is used to compress and optimize the accumulated real-time dataset related to biscuit production.

10. The method according to claim 3, characterized in that, S3 includes: During the iterative optimization process, the confidence interval of the initial control model is gradually reduced. Through iterative optimization, the confidence interval is gradually reduced from the initial 0.9 to 0.3.

Images