A method, storage medium and system for controlling filter rod draw resistance and circumference of a forming machine

Abstract

This invention relates to a method, storage medium, and system for controlling the suction resistance and circumference of filter rods in a molding machine. It includes a closed-loop control method for automatic adjustment of the suction resistance and an automatic adjustment of the circumference of filter rods in a molding machine. This method uses a comprehensive testing platform to inspect the products produced by the molding machine. Based on the inspection data, algorithms, and database analysis, corresponding adjustment commands are sent to the molding machine for adjustment. This replaces the previous manual measurement on the comprehensive testing platform, where manual adjustment switches are adjusted based on the measurement results. This achieves intelligent control of the circumference and suction resistance process quality of a single molding machine. This invention realizes the digitalization and intelligentization of filter rod production in molding machines, improves the process quality processing capabilities of filter rods in molding machines, enhances the intelligent manufacturing level of the equipment, and reduces the labor intensity of production.

Description

Technical Field

[0001] This invention relates to the field of cigarette filter rod production, specifically to a method, storage medium, and system for controlling the suction resistance and circumference of filter rods in a forming machine. Background Technology

[0002] Filter rod suction resistance and circumferential stability are key to stabilizing cigarette product quality. Currently, the industry lacks mature technologies for quality control of filter rod suction resistance in forming machines, relying instead on traditional manual adjustment methods. This results in common problems such as high labor intensity, inconsistent quality control levels, and difficulty in achieving homogenized filter rod production control. This project aims to combine in-depth research and analysis of big data from existing forming machine filter rod online detection with research on operators' experience in quality control adjustments. Through mathematical model building and algorithm development, an artificial intelligence fuzzy PID algorithm model and control strategy for filter rod suction resistance and circumferential stability will be established. This will achieve a breakthrough in intelligent control technology for the single-machine circumferential and suction resistance processes of forming machines, improving the quality processing capabilities of forming machines, enhancing the homogenization level of filter rod production, and increasing the digitalization and intelligent manufacturing capabilities of existing forming machine equipment.

[0003] For a long time, filter rod production has relied on the molding machine operators to analyze the online detection data of filter rod "weight, circumference, and suction resistance", the suction resistance quality trend, and make a comprehensive judgment based on the production operation experience. The online suction resistance quality control of filter rods has been carried out by manual adjustment, which has resulted in common industry problems such as large differences in the quality control level of human personnel and difficulty in controlling the quality uniformity of filter rod production process.

[0004] Existing control technologies, in practical use, require continuous manual monitoring of the machines, necessitating significant human intervention and increasing employee workload. Manual operation can also be relatively slow, limiting production speed and output. In contrast, automated production lines typically achieve higher production efficiency. However, manual control is susceptible to human error, potentially leading to inconsistent product quality. Furthermore, manual control may not be as effective as automated systems in real-time tracking and monitoring of key indicators during production. Delayed detection of potential quality issues during production results in inconsistent product quality, increased production costs, and wasted human resources.

[0005] CN115153084A discloses a control method for monitoring and distributing filter rod pressure drop for precise control of cigarette draw resistance. The method includes: during filter rod production, a host computer controls an online filter rod detection device to periodically sample and detect the filter rods. The detected filter rod pressure drop index information is combined with the filter rod's identity information to form filter rod information. This filter rod information is stored in both a filter rod tray information recorder and the host computer. After receiving the filter rod pressure drop value information required by each cigarette machine from the host computer, the filter rod storage warehouse searches the filter rod information recorder in the filter rod tray, finding N filter rod trays that meet the requirements, where N is a natural number. Based on the management regulations of the filter rod storage warehouse, the qualified filter rod trays are then delivered to the corresponding cigarette machine's filter rod dispensing device.

[0006] While the aforementioned existing technologies disclose precise control of cigarette draw resistance, they do not explain the closed-loop control of the filter rod circumference, nor do they explain the interaction between draw resistance and circumference, resulting in poor product quality stability. Summary of the Invention

[0007] The purpose of this invention is to test the products produced by the molding machine using a comprehensive testing platform. Based on the test data, the system analyzes the data using algorithms and databases, and sends corresponding adjustment instructions to the molding machine for adjustment. This replaces the previous manual measurement on the comprehensive testing platform, which involved adjusting the corresponding manual control switches based on the measurement results. This achieves intelligent control of the quality of the molding machine's circumference and suction resistance processes.

[0008] Based on the circumference and suction resistance characteristics of the molding machine, fuzzy control rules are designed, and a fuzzy PID control law is constructed to achieve closed-loop control. The fuzzy PID control law can dynamically adjust control parameters based on real-time sensor feedback data to achieve target quality and efficiency. To further improve the fuzzy control effect, an artificial intelligence self-learning component is added, enabling the system to optimize the control algorithm based on historical and real-time data, continuously learn control experience, and adaptively improve control performance. The system can continuously monitor the circumference and suction resistance data of the product, and adaptively adjust the control algorithm in real time according to the error situation, maintaining the stability and accuracy of closed-loop control. Thus, the system continuously optimizes and improves the algorithm technology based on actual production conditions and feedback information, improving the production efficiency of the molding machine and product quality.

[0009] Specifically, the technical solution of the present invention is as follows:

[0010] A closed-loop control method for automatic adjustment of suction resistance of filter rods in a molding machine includes the following steps:

[0011] S11. During the filter rod production process, the host computer controls an electronic balance to periodically sample and detect the weight of the tow bundles. The detected weight data is sent to the control program, which, according to pre-set rules, combines the detected tow bundle weight information with the filter rod's identity information to form filter rod information. In this application, the filter rod's identity information must at least include the filter rod production machine number, filter rod raw material number, applicable cigarette specifications, and other identity information, and may also include relevant control parameter information during the production process. The control system judges the detected tow bundle weight information to determine the current state information of the tow bundle (the tow bundle is generally divided into three states: bundle head, bundle middle, and bundle tail), so as to adjust the corresponding control parameter information during the filter rod production process. After the above information is combined to form the filter rod information, it is stored in the main controller for subsequent use.

[0012] S12. After the control system determines the status information of the filament bundle, if the filament bundle is in the state of the beginning or end of the bundle, it will improve the stability of the data information; if the filament bundle is in the state of the middle of the bundle, it will reduce the stability of the data information.

[0013] S13. During the tow packaging process, the comprehensive testing platform collects data from four filter rods online at set intervals. First, the measured filter rod data is analyzed by intelligent judgment. Abnormal suction resistance data is discarded, and the filter rods are re-collected for repeated measurement. The difference between the average of the four filter rod data and the standard value is analyzed and calculated. The standard value has been determined in advance. If the difference is less than the defined standard value, it is considered normal data. If the difference is greater than the standard value, it is analyzed in conjunction with the algorithm database. An adjustment parameter is calculated by the suction resistance control algorithm model and fed back to the opening parameter of the forming machine for adjustment.

[0014] S14. Abnormal Data Removal: To ensure the accuracy of the feedback data, linear regression analysis of the suction resistance measurement data is performed to remove abnormal data that does not conform to the expected pattern. This abnormal data will not be used as the basis for system analysis and calculation. The suction resistance of the filter rod is calculated using the following formula; data exceeding the predicted value within a specified range are considered abnormal. ( (W represents the absorption resistance coefficient, W represents the weight of a single sample, C represents the sample circumference, and L represents the sample length)

[0015] Ordinary 100mm: (y represents suction resistance, (Indicates the suction resistance coefficient).

[0016] S15, The steps of the suction resistance control algorithm model are as follows:

[0017] Experimental Design and Data Collection:

[0018] Prepare samples with different tow bundle heights, ensuring that these samples are consistent in other process parameters (such as tow material, tobacco formulation, etc.) in order to focus on the effect of tow bundle height on draw resistance;

[0019] These samples were tested using a suction resistance testing device, and the suction resistance value of each sample was recorded.

[0020] Meanwhile, precise measuring tools (such as electronic scales and altimeters) are used to collect the weight and height data of each sample's tow bundle in real time;

[0021] Data Analysis:

[0022] The collected data were organized and the relationship between the bundle height and the suction resistance was analyzed. Statistical software (such as SPSS, Excel, etc.) can be used to perform correlation analysis, regression analysis, etc., to determine the specific impact of the bundle height on the suction resistance.

[0023] Visualization tools such as scatter plots and line graphs can be used to help intuitively understand the changing trend between the height of the tow bundle and the suction resistance value;

[0024] Establish a mathematical model:

[0025] Based on the data analysis results, a mathematical model is established to describe the relationship between the height of the tow bundle and the suction resistance value. This model can be a linear regression model, a multinomial regression model, or other suitable models, depending on the characteristics of the data and the analysis results.

[0026] To validate the model and ensure it can accurately predict the suction resistance at different bundle heights, methods such as cross-validation and leave-one-out validation can be used to evaluate the model's generalization ability.

[0027] Model application and optimization:

[0028] Applying the established mathematical model to actual production and adjusting the bundle height according to the required draw resistance value helps to achieve more precise draw resistance control and improve the quality stability of tobacco products.

[0029] Regularly evaluate and optimize the model to adapt to changes in production processes and market demands. This can be achieved by collecting new data and updating model parameters.

[0030] 1. Real-time data acquisition and monitoring system

[0031] First, establish a real-time data acquisition and monitoring system to collect data on tow package weight and filter rod suction resistance. This can be achieved by installing sensors and instruments at key locations on the production line; these sensors and instruments need to be highly accurate and stable to ensure the accuracy and reliability of the collected data.

[0032] 2. Data Analysis and Model Building

[0033] The collected data needs to be analyzed to study the impact of different bundle heights on suction resistance. This can be achieved through statistical analysis, regression analysis, and other methods. Based on the data analysis, a mathematical model describing the relationship between bundle height and suction resistance value should be established. This model can be a mathematical formula or a machine learning model to predict the suction resistance value at different bundle heights.

[0034] 3. Software Analysis and Intelligent Control Algorithms

[0035] By using specialized software to analyze the variation pattern of filter rod suction resistance during the production process of monofilament bundles, we can understand the changing trend of suction resistance at different production stages and the key factors affecting suction resistance. Based on these analysis results, we can introduce an intelligent fuzzy PID control algorithm model into the production process. This algorithm can intelligently and precisely adjust the opening speed ratio of each section of the filter rod in the "head, middle, and surrounding" segments of the production rod based on real-time suction resistance data and bundle height data. Through closed-loop control, we can ensure that the suction resistance value remains within the set range during the production process, thereby improving the homogeneity of filter rod suction resistance production.

[0036] S16. The steps for constructing the fuzzy PID control algorithm model are as follows:

[0037] First, real-time online data of filter rod weight, circumference, and suction resistance of the forming machine are collected, along with parameters such as real-time weight of the tow package, ambient temperature and humidity, start-up time, number of consecutive samplings, and current opening roller speed ratio. Using the collected parameters, a fuzzy logic control model for the circumference and suction resistance of the forming machine is established, including fuzzification, construction of a fuzzy rule base, and fuzzy inference mechanism. Based on the control requirements of the circumference and suction resistance of the forming machine, a suitable fuzzy quantization method and fuzzy rules are designed to fuzzify the input and output variables and establish a fuzzy rule base.

[0038] Secondly, a fuzzy inference mechanism is designed, which uses a fuzzy rule base to infer fuzzy output based on the current fuzzy input variables.

[0039] Finally, through fuzzy defuzzification, the fuzzy output is transformed into specific control parameters, which are used as inputs to the PID controller to control the circumference and suction resistance of the molding machine.

[0040] S17. The circumference and suction resistance adjustment control strategy is calculated and output through the fuzzy PID controller. The suction resistance control strategy realizes the supply roller coefficient ( (V1 represents the feed roller speed of the filter rod forming machine, VKDF represents the filter strip speed of the forming machine), the speed ratio of the opening roller (pre-stretching roller ( (V1 represents the feed roller speed of the filter rod forming machine, V0 represents the pre-stretching roller speed), stretching roller ( (V2 represents the speed of the stretching roller, V1 represents the speed of the feed roller of the filter rod forming machine) and the supply roller ( (V3 represents the supply roller speed, V2 represents the extension roller speed) Parameters are intelligently written.

[0041] The automatic closed-loop control method for adjusting the circumference of the filter rod in the molding machine according to the present invention includes the following steps:

[0042] S21. During the filter rod production process, the host computer controls an electronic balance to periodically sample and detect the weight of the tow bundles. The detected weight data is sent to the control program, which, according to pre-set rules, combines the detected tow bundle weight information with the filter rod's identity information to form filter rod information. In this application, the filter rod's identity information must at least include the filter rod production machine number, filter rod raw material number, and applicable cigarette specifications, and may also include relevant control parameter information during the production process. The control system judges the detected tow bundle weight information to determine the current state of the tow bundle (a tow bundle is generally divided into three states: bundle head, bundle middle, and bundle tail), so as to adjust the corresponding control parameter information during filter rod production. After combining the above information to form filter rod information, it is stored in the main controller for subsequent use.

[0043] S22. After the control system determines the status information of the filament bundle, if the filament bundle is in the state of the beginning or end of the bundle, it will improve the stability of the data information; if the filament bundle is in the state of the middle of the bundle, it will reduce the stability of the data information.

[0044] S23. During the tow packaging process, the comprehensive testing platform collects measurements of 20 filter rods online at set intervals. First, the measured filter rod data is analyzed using intelligent judgment to calculate abnormal circumferential data for the 20 filter rods. By comparing the data before and after, it is determined whether the filter rods are within the standard range. Filter rods exceeding the standard range are discarded. For those within the set range, the average value is calculated, and the difference between the average and the standard value is analyzed. The standard value is predetermined. If the difference is less than the adjustment level, it is considered normal data. If the difference is greater than the adjustment level, it is analyzed using a fuzzy algorithm database to calculate an adjustment parameter, which is then fed back to the forming machine for adjustment. The parameter value is obtained through the fuzzy algorithm and transmitted to the forming machine. In subsequent operating cycles, the first filter rod's circumferential value in the array is removed, and new filter rod circumferential data is added while abnormal circumferential data is removed. The new circumferential average data is then subjected to fuzzy judgment, combined with the filter rod parameter information and the forming machine parameter information, to determine and modify the data.

[0045] S24. Divide the difference between the measured mean and the standard value into four thresholds, namely the upper and lower limits ±¾ threshold and the upper and lower limits ±½ threshold. The measurement data between the standard value of the circumference and the upper and lower limits ±½ threshold are not used as the basis for adjustment. Based on the measurement data of the four thresholds, the background big data analysis is combined with the digital intelligent working model of the molding machine expert to give the molding machine the corresponding feedback signal for circumference adjustment.

[0046] S25. The steps for establishing the expert digital intelligent working model of the molding machine are as follows:

[0047] First, collect and organize the operational experience data of excellent machine operators. This can be achieved by interviewing machine operators, observing their actual operation, and recording relevant data. The collected data can include operating steps, techniques, precautions, etc.

[0048] After collecting the data, we need to label and classify it, categorizing each operational experience according to different tags, such as different car models, different scenarios, etc. This will help to build an expert experience database and accurately match problems in the future.

[0049] Based on the collected data, we can build an expert digital intelligence experience base, which can be modeled in the form of knowledge graphs or semantic networks to better organize and retrieve data.

[0050] After establishing an expert digital intelligent experience database, we can develop an intelligent question-and-answer system that allows users to obtain relevant operational experience by asking questions. This system can use natural language processing and machine learning technologies to analyze the semantics and keywords of user questions, match the most relevant operational experience, and provide detailed answers.

[0051] After establishing the expert digital intelligent experience base and intelligent question-and-answer system, we need to continuously iterate and improve it. This includes constantly updating the content of the experience base, adding more operational experience data, and optimizing the system to improve the accuracy of question matching and the quality of answers.

[0052] S26. The fuzzy PID control algorithm model is used to determine the circumferential adjustment range. Under the condition of stable filter rod suction resistance, the circumferential adjustment control strategy is output. The circumferential control strategy adds two operating variables, M4021.3 and M4021.2, to the system and writes intelligent parameters through parallel circumferential adjustment plus and minus buttons.

[0053] The beneficial effects of this invention include:

[0054] This invention combines in-depth research and analysis of existing online detection big data for filter rods in molding machines with research on the quality control experience and methods of production personnel. Through mathematical model building and algorithm development, it establishes an artificial intelligence fuzzy PID control algorithm model and intelligent control strategy method for filter rod suction resistance and circumference. This achieves a breakthrough in intelligent control technology for single-machine process quality of molding machine equipment, improves the processing quality of filter rods in the molding machine, promotes the improvement of intelligent manufacturing capabilities of existing molding machine equipment, and increases the value of fixed assets. Through intelligent closed-loop control of suction resistance and circumference process quality in filter rod production, it realizes the digitalization and intelligentization of filter rod production in molding machines, improves the level of intelligent manufacturing capabilities of equipment, and reduces the labor intensity of production. Attached Figure Description

[0055] Figure 1 : Structure diagram of filter rod suction resistance / circumference control system of the molding machine. In the diagram: 100-control system, 200-comprehensive test bench, 300-filter rod molding machine.

[0056] Figure 2 : Quantitative relationship diagram of the filter rod suction resistance / circumferential control system model of the forming machine. In the diagram: 5 represents the driven wheel, 6 represents the driving wheel; 0 represents the pre-stretching mechanism, 1 represents the pre-stretching roller group (including the driving wheel 6 and the driven wheel 5), 2 represents the stretching roller group (including the driving wheel 6 and the driven wheel 5), 3 represents the smoothing roller group (including the driving wheel 6 and the driven wheel 5), and 4 represents the supply roller outputting the filament to the forming machine; the arrows I, II, III, and IV in the diagram represent the direction of movement of the filament or roller; V0, V1, V2, V3, and VKDF in the diagram represent the system input speed, pre-stretching roller speed, stretching roller speed, supply roller speed, and supply roller opening speed, respectively.

[0057] Figure 3 Structure diagram of intelligent fuzzy PID control system for filter rod suction resistance / circumference of molding machine.

[0058] Figure 4 Flowchart of algorithm calculation for filter rod suction resistance / circumference control system of molding machine.

[0059] Figure 5 Flowchart of intelligent fuzzy PID control law for filter rod suction resistance / circumference of molding machine. Detailed Implementation

[0060] The following are specific embodiments of the present invention, and the technical solutions of the present invention will be further described in conjunction with the accompanying drawings. However, the present invention is not limited to these embodiments.

[0061] like Figure 2As shown, by collecting real-time online data on the weight, circumference, and suction resistance of the filter rods in the forming machine, as well as parameters such as the real-time weight of the tow bundle, ambient temperature and humidity, start-up time, number of continuous samplings, and current opening roller speed ratio, the fuzzy PID controller calculates and outputs a circumference and suction resistance adjustment control strategy. The suction resistance control strategy realizes the supply roller coefficient ( (V1 represents the feed roller speed of the filter rod forming machine, VKDF represents the filter strip speed of the forming machine), the speed ratio of the opening roller (pre-stretching roller ( (V1 represents the feed roller speed of the filter rod forming machine, V0 represents the pre-stretching roller speed), stretching roller ( (V2 represents the speed of the stretching roller, V1 represents the speed of the feed roller of the filter rod forming machine) and the supply roller ( (V3 represents the supply roller speed, V2 represents the extension roller speed) Intelligent closed-loop adjustment of parameters; circumferential control strategy realizes online adjustment of circumferential closed-loop control parameters by adding / subtracting;

[0062] The expert digital intelligent working model for the molding machine is established based on the operational experience and methods of outstanding molding machine operators. This model incorporates the quality control experience of these operators into the control algorithm, creating an intelligent adjustment control model for filter rod circumference and suction resistance. The expert digital intelligent fuzzy control algorithm model performs real-time intelligent analysis of quality data, processing and calculating the data to intelligently optimize control system parameters and output accurate control strategies. This intelligently optimizes the precision of the quality control strategy for the filter rod molding machine's circumference and suction resistance processes, achieving a homogenized production process capability for filter rod suction resistance and circumference quality. The intelligent quality control technology for the molding machine process combines fuzzy PID algorithms with artificial intelligence self-learning. The algorithm technology of circumferential and suction resistance closed-loop control enables the algorithm to adapt to different data and environments. The system adopts a comprehensive operation combining fuzzy PID algorithm and intelligent self-learning. The fuzzy controller is mainly composed of three functional modules: fuzzification, fuzzy inference engine, and precision, as well as a knowledge base (including database and rule base). Fuzzy PID control takes the deviation e and the change of deviation ec as input, and uses fuzzy control rules to adjust the PID parameters online to meet the different requirements of different deviations e and deviation increments ec on the PID parameters. The control strategy realizes intelligent closed-loop control of the circumferential and suction resistance single-machine process quality of the molding machine, reducing the labor intensity of personnel, improving production efficiency, product quality, and enhancing the intelligent manufacturing level of molding machine equipment.

[0063] The steps for artificial intelligence self-learning are as follows:

[0064] First, real-time data during the operation of the molding machine is collected, including process parameters of the molding machine's circumference and suction resistance, environmental conditions, etc. The collected data is preprocessed and features are extracted to facilitate subsequent machine learning model training.

[0065] Secondly, by utilizing historical and real-time data, machine learning models, such as neural networks and decision trees, are established to predict the optimal control parameters for the circumference and suction resistance of the molding machine.

[0066] Secondly, historical data is used to train the model, and the model parameters are continuously adjusted based on real-time data to adapt to different process changes and environmental conditions.

[0067] Finally, based on the prediction results of the machine learning model, the parameters of the fuzzy PID controller are optimized to achieve automatic adjustment of the circumference and suction resistance of the molding machine.

[0068] like Figure 1 , Figure 2 and Figure 5 As shown, the integrated test bench collects data from four filter rods online for measurement. The measured suction resistance data is analyzed. If the data deviation is large, it is discarded as abnormal data, and the filter rods are re-collected for measurement. When all four filter rods are within the deviation, the difference between the mean and the standard value is analyzed. If the difference is less than the defined standard value, it is considered normal data. If the difference is greater than the standard value, it is analyzed in conjunction with the algorithm database to calculate an adjustment parameter, which is then fed back to the opening parameter of the molding machine for adjustment.

[0069] A fuzzy PID control algorithm model is adopted, and a reference adjustment method for the fuzzy PID controller and a parameter self-tuning method are designed. The algorithm model development aims to collect real-time online data on the weight, circumference, and suction resistance of the filter rods in the forming machine, as well as parameters such as the real-time weight of the yarn bundle, ambient temperature and humidity, start-up time, number of consecutive samplings, and the current opening roller speed ratio. The fuzzy PID controller will then calculate and output circumference and suction resistance adjustment control strategies. The suction resistance control strategy will achieve the desired adjustment of the supply roller coefficient (…). (V1 represents the feed roller speed of the filter rod forming machine, VKDF represents the filter strip speed of the forming machine), the speed ratio of the opening roller (pre-stretching roller ( (V1 represents the feed roller speed of the filter rod forming machine, V0 represents the pre-stretching roller speed), stretching roller ( (V2 represents the speed of the stretching roller, V1 represents the speed of the feed roller of the filter rod forming machine) and the supply roller ( (V3 represents the supply roller speed, V2 represents the extension roller speed) Parameters are intelligently written;

[0070] The suction resistance data is inversely proportional to the opening roller, extension roller, and supply roller of the forming machine. Adjusting the suction resistance data of the forming machine is achieved by adjusting the rotational speed of these three rollers. The system adjustment method involves analyzing the difference ΔD between the average suction resistance and the standard value. Based on this difference, the adjustment range is divided into six levels. Furthermore, based on the relationship between the difference and the speed adjustment amount, the speed adjustment amount is also divided into two areas: V1vKDF±0.001, V2vKDF±0.002, and V3vKDF±0.003, corresponding to the roller speed adjustment buttons on the forming machine. Through calculation and analysis, the adjustment data can be provided to the operator for adjustment.

[0071] Through tracking and analysis of the suction resistance data, the value of parameter V1vKDF shows an inverse trend with the suction resistance value. During system adjustment, the area with a deviation value less than 45 Pa was designated as the adjustment-free zone. The adjustment zone was divided into six levels: 45 < ΔD ≤ 85, 85 < ΔD ≤ 135, 135 < ΔD, and the lower limit -85 < ΔD ≤ -45, -135 < ΔD ≤ -85, and ΔD ≤ -135. When the deviation value exceeded the adjustment-free zone, V1vKDF, V2vKDF, and V3vKDF were adjusted in tandem. During the adjustment process, the adjustment amount ΔV of V1vKDF was directly proportional to the deviation value ΔD. The adjustment relationship is as follows:

[0072] Deviation value:

[0073] Adjustment amount:

[0074] The adjustment amounts ΔV for the three upper and lower regions of the standard value are V1vKDF±0.001, V2vKDF±0.002, and V3vKDF±0.003, respectively. During the adjustment process, in order to avoid unnecessary abnormalities caused by excessive adjustment in a single instance, the maximum adjustment amount in a single instance is limited to ±0.003.

[0075] like Figure 1 , Figure 3 and Figure 4 As shown, the integrated testing platform collects measurements from 20 filter rods online and analyzes the circumferential data. If the data deviation is large, it is discarded as abnormal data, and the filter rods are re-collected for measurement. When all 20 filter rods are within the deviation range, the difference between the mean and the standard value is analyzed. If the difference is less than the adjustment level, it is considered normal data. If the difference is greater than the adjustment level, it is analyzed in conjunction with the algorithm database to calculate an adjustment parameter, which is then fed back to the molding machine for adjustment.

[0076] The difference between the planned measurement mean and the standard value is divided into four thresholds: the upper and lower limits ±¾ threshold and the ±½ threshold. The measurement data between the standard value of the circumference and the upper and lower limits ±½ threshold are not used as the basis for adjustment. Based on the measurement data of the four thresholds, the background big data analysis, combined with the expert intelligent automatic working model, gives the forming machine the corresponding feedback signal for circumference adjustment.

[0077] A fuzzy PID control algorithm model is adopted to calculate the reference adjustment method of the fuzzy PID controller. The algorithm model is developed by collecting real-time online data of filter rod weight, circumference, and suction resistance of the molding machine, and then processing the data through the fuzzy PID controller. It determines the circumference adjustment range and outputs the circumference adjustment control strategy under the condition of stabilizing the filter rod suction resistance. The circumference control strategy adds two operation variables, M4021.3 and M4021.2, to the system. The widening variable M4021.3 controls the widening of the channel, and the narrowing variable M4021.2 controls the narrowing of the channel. The circumference adjustment plus and minus buttons are connected in parallel and written into the intelligent parameters.

[0078] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for controlling the suction resistance and circumference of a filter rod in a molding machine, characterized in that: This includes a closed-loop control method for automatic adjustment of the suction resistance of the filter rod in a molding machine and a closed-loop control method for automatic adjustment of the circumference of the filter rod in a molding machine. The automatic closed-loop control method for adjusting the suction resistance of the filter rod in the molding machine includes the following steps: S11. During the filter rod production process, the host computer controls the electronic balance to periodically sample and detect the weight of the tow package. The detected weight data is sent to the control program. According to pre-set rules, the detected tow package weight information is combined with the filter rod's identity information to form the filter rod information. The filter rod's identity information must include at least the filter rod production machine number, the filter rod raw material number, the applicable cigarette specification information, and the corresponding control parameter information during the production process. The control system judges the detected tow package weight information to determine the current status of the tow package so as to adjust the corresponding control parameter information during the filter rod production process. The filter rod information is stored in the main controller for subsequent use. S12. After the control system determines the status information of the filament bundle, if the filament bundle is in the state of the beginning or end of the bundle, it will improve the stability of the data information; if the filament bundle is in the state of the middle of the bundle, it will reduce the stability of the data information. S13. During the tow packaging process, the comprehensive testing platform collects data from four filter rods online at set intervals. First, the measured filter rod information is analyzed through intelligent judgment. Abnormal suction resistance data is eliminated, and the filter rods are re-collected for repeated measurement. The difference between the average value of the four filter rod information and the standard value is analyzed and calculated. The standard value has been determined in advance. If the difference is less than the defined standard value, it is considered normal data. If the difference is greater than the standard value, it is analyzed in conjunction with the algorithm database. An adjustment parameter is calculated through the suction resistance control algorithm model and fed back to the opening parameter of the forming machine for adjustment. S14. Abnormal Data Removal: To ensure the accuracy of the feedback data, linear regression analysis of the suction resistance measurement data is performed to remove abnormal data that does not conform to the expected pattern. This abnormal data will not be used as the basis for system analysis and calculation. The suction resistance of the filter rod is calculated using the following formula; data exceeding the predicted value within a specified range are considered to be abnormal. ,in: W represents the absorption resistance coefficient, C represents the weight of a single sample, and L represents the sample circumference. Ordinary 100mm: Where: y represents suction resistance, Indicates the suction resistance coefficient; S15. The circumference and suction resistance adjustment control strategies are calculated and output through a fuzzy PID controller. The suction resistance control strategy realizes the supply roller coefficient. Intelligent writing of the speed ratio parameter of the loosening roll, wherein the speed ratio of the loosening roll includes the speed ratio coefficient of the pre-stretching roll. , extension roller speed ratio coefficient and the ratio coefficient of the supply roller speed Where V0, V1, V2, V3, and VKDF represent the system input speed, pre-stretching roller speed, stretching roller speed, supply roller speed, and supply roller opening speed, respectively. The closed-loop control method for automatic circumferential adjustment of the filter rod in the molding machine includes the following steps: S21. During the filter rod production process, the host computer controls an electronic balance to periodically sample and detect the weight of the tow bundle. The detected weight data is sent to the control program. According to pre-set rules, the detected tow bundle weight information is combined with the filter rod's identity information to form filter rod information. The filter rod's identity information must include at least the filter rod production machine number, the filter rod raw material number, the applicable cigarette specification information, and the corresponding control parameter information during the production process. The control system judges the detected tow bundle weight information to determine the current status of the tow bundle so as to adjust the corresponding control parameter information during the filter rod production process. The filter rod information is stored in the main controller for subsequent use. S22. After the control system determines the status information of the filament bundle, if the filament bundle is in the state of the beginning or end of the bundle, it will improve the stability of the data information; if the filament bundle is in the state of the middle of the bundle, it will reduce the stability of the data information. S23. During the tow packaging process, the comprehensive testing platform collects measurements of 20 filter rods online at set intervals. First, the measured filter rod information is analyzed through intelligent judgment to calculate the abnormal circumference data of the 20 filter rods. By comparing the data before and after, it is determined whether the filter rods are within the standard range. Filter rods that exceed the standard range are removed. For those within the set range, the average value is calculated, and the difference between the average value and the standard value is analyzed. The standard value has been determined in advance. If the difference is less than the adjustment level, it is considered normal data. If the difference is greater than the adjustment level, it is analyzed in conjunction with the fuzzy algorithm database to calculate an adjustment parameter, which is then fed back to the forming machine for adjustment. The parameter value is obtained through the fuzzy algorithm and transmitted to the forming machine. In subsequent operating cycles, the circumference value of the first filter rod in the array is first removed, and new filter rod circumference data is added. At the same time, abnormal circumference data is removed. The new circumference average data is fuzzy judged and modified in conjunction with the filter rod information and the forming machine parameter value. S24. Divide the difference between the measured mean and the standard value into four thresholds, namely the upper and lower limits ±¾ threshold and the upper and lower limits ±½ threshold. The measurement data between the standard value of the circumference and the upper and lower limits ±½ threshold are not used as the basis for adjustment. Based on the measurement data of the four thresholds, the background big data analysis is combined with the digital intelligent working model of the molding machine expert to give the molding machine the corresponding feedback signal for circumference adjustment. S25. Establish an expert digital intelligent working model, use a fuzzy PID control algorithm model to determine the circumferential adjustment range, and output a circumferential adjustment control strategy under the condition of stable filter rod suction resistance.

2. The method for controlling the suction resistance and circumference of a filter rod in a molding machine according to claim 1, characterized in that, The suction resistance control algorithm model includes experimental design and data collection: Prepare samples with different bundle heights, ensuring that these samples are consistent across other process parameters, in order to focus on the effect of bundle height on suction resistance; These samples were tested using a suction resistance testing device, and the suction resistance value of each sample was recorded. Meanwhile, precise measuring tools were used to collect real-time data on the weight and height of the tow bundles for each sample.

3. The method for controlling the suction resistance and circumference of a filter rod in a molding machine according to claim 2, characterized in that, The suction resistance control algorithm model also includes data analysis: The collected data were organized and analyzed to determine the relationship between the height of the tow bundle and the suction resistance. Statistical software was used to perform correlation and regression analysis to determine the specific impact of the tow bundle height on the suction resistance.

4. The method for controlling the suction resistance and circumference of a filter rod in a molding machine according to claim 3, characterized in that, The suction resistance control algorithm model also includes: Visualization tools such as scatter plots and line graphs help to intuitively understand the changing trend between the height of the tow bundle and the suction resistance value.

5. The method for controlling the suction resistance and circumference of a filter rod in a molding machine according to claim 4, characterized in that, The suction resistance control algorithm model also includes the establishment of a mathematical model: Based on the data analysis results, a mathematical model is established to describe the relationship between the height of the tow bundle and the suction resistance value. This model can be a linear regression model, a multinomial regression model, or other suitable model, depending on the characteristics of the data and the analysis results. The model was validated to ensure that it could accurately predict the suction resistance value at different bundle heights. Cross-validation and leave-one-out validation methods were used to evaluate the model's generalization ability.

6. The method for controlling the suction resistance and circumference of a filter rod in a molding machine according to claim 1, characterized in that, The suction resistance control algorithm model also includes model application and optimization: Applying the established mathematical model to actual production and adjusting the bundle height according to the required draw resistance value helps to achieve more precise draw resistance control and improve the quality stability of tobacco products. The model is regularly evaluated and optimized to adapt to changes in production processes and market demands. Continuous optimization of the model is achieved by collecting new data and updating model parameters.

7. The method for controlling the suction resistance and circumference of a filter rod in a molding machine according to claim 1, characterized in that, The steps for establishing the expert digital intelligence working model are as follows: First, real-time online data of filter rod weight, circumference, and suction resistance of the forming machine are collected, along with real-time weight of the tow package, ambient temperature and humidity, start-up time, number of consecutive samplings, and current opening roller speed ratio. Using the collected parameters, a fuzzy logic control model for the circumference and suction resistance of the forming machine is established, including fuzzification, construction of a fuzzy rule base, and fuzzy inference mechanism. Based on the control requirements of the circumference and suction resistance of the forming machine, a suitable fuzzy quantization method and fuzzy rules are designed to fuzzify the input and output variables and establish a fuzzy rule base. Secondly, a fuzzy inference mechanism is designed, which uses a fuzzy rule base to infer fuzzy output based on the current fuzzy input variables. Finally, through fuzzy defuzzification, the fuzzy output is transformed into specific control parameters, which are used as inputs to the PID controller to control the circumference and suction resistance of the molding machine.

8. A method for controlling the suction resistance and circumference of a filter rod in a molding machine according to any one of claims 1-7, characterized in that, The steps for establishing the expert digital intelligent working model of the molding machine are as follows: First, collect and organize the operational experience data of outstanding machine operators. This is achieved by interviewing machine operators, observing their actual operations, and recording relevant data. The collected data includes operating steps, techniques, and precautions. After collecting the data, it is labeled and classified, and each operational experience is classified according to different tags. This will help to build an expert digital intelligence experience library and accurately match problems in the future. Based on the collected data, an expert digital intelligence experience base is established. This experience base is modeled in the form of knowledge graphs or semantic networks to better organize and retrieve data. After establishing an expert digital intelligent experience database, an intelligent question-and-answer system was developed to enable users to obtain relevant operational experience by asking questions. The system uses natural language processing and machine learning technologies to analyze the semantics and keywords of user questions, match the most relevant operational experience, and provide detailed answers. After establishing an expert digital intelligent experience base and an intelligent question-and-answer system, continuous iteration and improvement are carried out. This includes constantly updating the content of the experience base, adding more operational experience data, and optimizing the system to improve the accuracy of question matching and the quality of answers.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, The computer program is executed by a processor to implement the steps of the closed-loop control method for automatic adjustment of suction resistance and circumference of a molding machine filter rod as described in any one of claims 1-8.

10. A closed-loop control system for automatic adjustment of filter rod suction resistance and circumference in a molding machine, comprising a molding machine, a comprehensive testing platform, and a control system, characterized in that, The control system includes a computer-readable storage medium as described in claim 9, the control system is electrically connected to the integrated test bench, and the control system is electrically connected to the filter rod forming machine.

Images