PE packaging line pinch roller heating temperature control method and system and storage medium

Abstract

The invention relates to the technical field of intelligent temperature control, and discloses a PE packaging line pinch roller heating temperature control method and system, and a storage medium, and the method comprises the steps: taking a process variable and an environment variable in high-dimensional time sequence data as nodes, taking a dynamic causal relationship as an edge, and constructing a packaging line dynamic causal structure diagram; taking the pinch roller temperature as a target node, screening a causal influence relationship to obtain a local network, and tracing a key path to obtain a precursor variable; mapping a precursor variable dynamic behavior mode, and generating a pinch roller temperature feed-forward compensation strategy; analyzing the feed-forward compensation strategy and the real-time state vector of the precursor variable to obtain a predictive compensation amount; based on the predictive compensation amount, a pinch roller temperature control instruction is cooperatively decided, and a comprehensive control instruction is generated; according to the execution effect of the comprehensive control instruction and the topology consistency, the strategy applicability is judged, and a reconstruction trigger signal is obtained; according to the invention, the temperature control operation efficiency of the pinch roller of the PE packaging line can be improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent temperature control technology, and in particular to a method, system, and storage medium for controlling the heating temperature of the pressure rollers in PE packaging lines. Background Technology

[0002] PE packaging lines play a crucial role in the plastic packaging industry, and the temperature control of their pressure rollers directly affects the sealing strength, appearance quality, and production efficiency of the packaged products. Traditional control methods typically employ feedback-based PID control mechanisms, maintaining temperature stability by monitoring the pressure roller temperature in real time and adjusting the heating device power. Furthermore, some advanced control strategies introduce multivariate regulation and adaptive algorithms to cope with disturbances during production line operation. However, these methods largely rely on linear relationships between local variables and fail to fully consider the complex dynamic correlations between process and environmental variables in high-dimensional time-series data, resulting in inadequate performance of the control model when dealing with nonlinear and time-delayed disturbances.

[0003] In existing technologies, pressure roller temperature control methods have two main problems: First, due to the failure to systematically model the dynamic causal relationship between process variables and environmental variables, the control strategy cannot accurately capture the time delay characteristics and influence paths of multi-variable interactions, resulting in a disconnect between compensation actions and actual disturbances; second, the lack of a feedforward compensation mechanism based on causal topology means that control commands rely solely on historical data and real-time feedback, making it difficult to predict the changing trends of dynamic behavior patterns, resulting in system response lag and decreased stability. Therefore, how to improve the operating efficiency of pressure roller temperature control in PE packaging lines has become an urgent problem to be solved. Summary of the Invention

[0004] This disclosure provides a method, system, device, storage medium, and computer program for controlling the heating temperature of the pressure rollers in a PE packaging line.

[0005] In a first aspect, this disclosure provides a method for controlling the heating temperature of the pressure rollers in a PE packaging line, including: S1. Using process variables and environmental variables in the high-dimensional time series dataset as nodes, and the dynamic causal relationship between the process variables and the environmental variables as edges, construct a dynamic causal structure graph of encapsulation lines. S2. Based on the pressure roller temperature as the target node, the dynamic causal structure diagram is filtered for causal influence relationships to obtain the local causal network of the pressure roller temperature, and the key causal paths in the local causal network are traced to obtain the precursor variables of the pressure roller temperature. S3. Perform rule mapping on the dynamic behavior patterns in the precursor variables to obtain the feedforward compensation strategy for the pressure roller temperature. S4. Perform feedforward manipulation variable analysis on the feedforward compensation strategy and the real-time state vector in the precursor variables to obtain the predictive compensation amount of the pressure roller temperature. S5. Based on the predictive compensation amount, make collaborative decisions on the control command for the pressure roller temperature to obtain the comprehensive control command for the pressure roller temperature; S6. Based on the execution effect of the integrated control command and the topological consistency of the dynamic causal structure graph, the applicability of the feedforward compensation strategy is determined to obtain the reconstruction trigger signal of the feedforward compensation strategy.

[0006] In a preferred embodiment, the step of constructing a dynamic causal structure graph of encapsulation lines, using process variables and environmental variables in a high-dimensional time-series dataset as nodes and the dynamic causal relationships between the process variables and the environmental variables as edges, includes: The original timing data of the packaging line is obtained, and the original timing data is coordinated to obtain the high-dimensional timing data of the packaging line. Using preset variable function classification rules, the operating condition variables in the high-dimensional time series data are divided to obtain the process variable group and environmental variable group of the packaging line; The process variable group and the environment variable group are node-based to obtain the node set of the encapsulation line; Time-delay correlation analysis is performed on the interaction of variables in the node set to obtain a dynamic correlation measure between the process variables and the environmental variables; By performing causal-guided inference on the dynamic correlation metric, the directed edge set of the encapsulation line is obtained; The node set and the directed edge set are topologically reconstructed to obtain the dynamic causal structure graph of the encapsulation line.

[0007] In a preferred embodiment, the dynamic causal structure graph is screened for causal relationships based on the pressure roller temperature as the target node to obtain a local causal network of the pressure roller temperature. Then, key causal paths within this local causal network are traced to obtain the precursor variables of the pressure roller temperature, including: Using the pressure roller temperature as the target node, direct causal edge extraction is performed on the dynamic causal structure graph to obtain the direct association set of the pressure roller temperature; In the dynamic causal structure graph, multi-order causal backtracking is performed starting from the end node of the causal edge in the direct association set to obtain the indirect causal path cluster of the pressure roller temperature; Using the directly associated set as points and the indirect causal path cluster as edges, a complete causal network for the pressure roller temperature is constructed. The path strength of each causal path in the complete causal network is quantified to obtain the causal influence weight of the complete causal network. Based on the causal influence weights, the path saliency of all paths in the complete causal network is screened to obtain the key causal path of the pressure roller temperature; By tracing the key causal path, the precursor variable of the pressure roller temperature is obtained.

[0008] In a preferred embodiment, the step of performing rule mapping on the dynamic behavior patterns in the precursor variables to obtain the feedforward compensation strategy for the pressure roller temperature includes: Based on the empirical rules of the historical optimal control process of the packaging line, the basic compensation rules for the pressure roller temperature are obtained; The predecessor variables are dynamically evolved and their behavior trajectory sequences are obtained by tracking their dynamic evolution. By mining the pattern features of the behavioral trajectory sequence, the saliency features of the dynamic behavioral pattern are obtained; The saliency features are matched and mapped with the basic compensation rules to obtain candidate compensation rules for the pressure roller temperature; A multi-objective optimization decision is made on the candidate compensation rules to obtain the feedforward compensation strategy for the pressure roller temperature.

[0009] In a preferred embodiment, feedforward manipulator variable analysis is performed on the feedforward compensation strategy and the real-time state vector in the precursor variables to obtain the predictive compensation amount for the pressure roller temperature, including: The feedforward compensation strategy is deconstructed to obtain the control rule fragments of the feedforward compensation strategy; Based on the control rule fragment, variable selection is performed on the real-time state vector in the predecessor variables to obtain a subset of key variables of the real-time state vector. Deviation identification is performed on the subset of key variables to obtain the real-time deviation value of the subset of key variables; The real-time deviation value and the control rule segment are dynamically weighted and fused to obtain the predictive compensation amount for the pressure roller temperature.

[0010] In a preferred embodiment, the real-time deviation value and the control rule segment are dynamically weighted and fused to obtain a predictive compensation amount for the pressure roller temperature, wherein the calculation formula for the predictive compensation amount is as follows: ; In the formula, The predictive compensation amount, This represents the total number of predecessor variables contained in the subset of key variables. For the first The dynamic weighting coefficients of the aforementioned precursor variables, For the first The real-time deviation values ​​of the aforementioned precursor variables. The absolute magnitude of the real-time deviation value. For the first The time constants of the aforementioned precursor variables The differential gain coefficient, This indicates that the expression within the parentheses is subjected to time differentiation. It is an exponential function.

[0011] In a preferred embodiment, the step of collaboratively deciding on the control command for the pressure roller temperature based on the predictive compensation amount to obtain a comprehensive control command for the pressure roller temperature includes: The predictive compensation amount is discretized in a distributed manner to obtain the influence distribution of the pressure roller temperature; Based on the influence distribution, the control command for the pressure roller temperature is weighted by multiple objectives to obtain the weighted decision criterion for the pressure roller temperature. Based on the weighted decision criterion, the control command for the pressure roller temperature is subjected to multi-objective trade-offs to obtain the trade-off control command for the pressure roller temperature. Based on the dynamic operating margin of the packaging line, the stability of the control command after the trade-off is evaluated, and the enhanced command for the pressure roller temperature is obtained. The enhanced command and the real-time feedback control command for the pressure roller temperature are fused and optimized to obtain a comprehensive control command for the pressure roller temperature.

[0012] In a preferred embodiment, the stability determination of the applicability of the feedforward compensation strategy based on the execution effect of the integrated control command and the topological consistency of the dynamic causal structure graph, to obtain the reconstruction trigger signal of the feedforward compensation strategy, includes: The integrated control commands are applied to the packaging line to extract the control performance of the packaging line. The operating characteristics of the pressure roller temperature are obtained; Based on the dynamic causal structure graph, the causal topological coupling analysis of the operational characteristics is performed to obtain the causal consistency coefficient of the feedforward compensation strategy; Based on the causal consistency coefficient, the effectiveness of the feedforward compensation strategy is determined, and the applicability of the feedforward compensation strategy is determined. Based on the applicability decision, the stability state of the feedforward compensation strategy is determined, and the reconstruction trigger signal of the feedforward compensation strategy is obtained.

[0013] To address the aforementioned problems, the present invention also provides a heating temperature control system for the pressure roller of a PE packaging line, comprising: The dynamic causal structure graph construction module is used to construct a dynamic causal structure graph with process variables and environmental variables in a high-dimensional time series dataset as nodes and the dynamic causal relationship between the process variables and the environmental variables as edges. The causal screening and precursor variable acquisition module is used to screen the causal influence relationship of the dynamic causal structure diagram based on the pressure roller temperature as the target node, so as to obtain the local causal network of the pressure roller temperature, and trace the key causal path in the local causal network to obtain the precursor variable of the pressure roller temperature. The feedforward compensation strategy generation module is used to perform rule mapping on the dynamic behavior patterns in the precursor variables to obtain the feedforward compensation strategy for the pressure roller temperature. The predictive compensation analysis module is used to analyze the feedforward manipulation variables of the feedforward compensation strategy and the real-time state vector in the precursor variables to obtain the predictive compensation amount of the pressure roller temperature. The integrated control command decision module is used to make collaborative decisions on the control command for the pressure roller temperature based on the predictive compensation amount, so as to obtain the integrated control command for the pressure roller temperature. The compensation strategy stability determination module is used to determine the applicability of the feedforward compensation strategy based on the execution effect of the integrated control command and the topological consistency of the dynamic causal structure graph, and to obtain the reconstruction trigger signal of the feedforward compensation strategy.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention constructs a dynamic causal structure diagram to systematically analyze the time-delay correlation and causal path between process variables and environmental variables, accurately identifying the precursor variables affecting the pressure roller temperature and their behavioral patterns. Based on this, the generated feedforward compensation strategy can predict disturbance trends and analyze predictive compensation amounts, thereby significantly improving the timeliness and accuracy of temperature control and overcoming the response lag problem caused by neglecting the dynamic interaction of variables in traditional methods.

[0015] 2. This invention, through a collaborative decision-making mechanism for integrated control commands and a strategy applicability determination mechanism, enables the system to dynamically evaluate the consistency between control effects and causal topology, and trigger strategy reconfiguration accordingly. This mechanism ensures that the feedforward compensation strategy always matches the real-time state of the production line, effectively avoiding control failures caused by changes in operating conditions, and improving the system's stability and adaptability in complex operating environments. Attached Figure Description

[0016] The present disclosure will be described in more detail below based on embodiments and with reference to the accompanying drawings: Figure 1 The flowchart shows the process of the pressure roller heating temperature control method of the PE packaging line according to Embodiment 1 of the present invention; Figure 2 The diagram shows the functional block diagram of the pressure roller heating temperature control system of the PE packaging line according to Embodiment 2 of the present invention. Detailed Implementation

[0017] To enable those skilled in the art to better understand the technical solutions of this disclosure, and to fully understand and implement the process of how this disclosure applies technical means to solve technical problems and achieve corresponding technical effects, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, not all embodiments. The embodiments of this disclosure and the various features within them can be combined with each other without conflict, and the resulting technical solutions are all within the protection scope of this disclosure. All other embodiments obtained by those skilled in the art based on the embodiments of this disclosure without creative effort should fall within the protection scope of this disclosure.

[0018] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0019] Example 1 Figure 1 This is a schematic flowchart illustrating the method for controlling the heating temperature of the pressure rollers in a PE packaging line according to an embodiment of this disclosure. Figure 1 As shown, a smart device control method includes: S1. Using process variables and environmental variables in the high-dimensional time series dataset as nodes, and the dynamic causal relationship between the process variables and the environmental variables as edges, construct a dynamic causal structure graph of encapsulation lines. In this embodiment of the invention, the step of constructing a dynamic causal structure graph of encapsulation lines, using process variables and environmental variables in a high-dimensional time-series dataset as nodes and the dynamic causal relationship between the process variables and the environmental variables as edges, includes: The original timing data of the packaging line is obtained, and the original timing data is coordinated to obtain the high-dimensional timing data of the packaging line. Using preset variable function classification rules, the operating condition variables in the high-dimensional time series data are divided to obtain the process variable group and environmental variable group of the packaging line; The process variable group and the environment variable group are node-based to obtain the node set of the encapsulation line; Time-delay correlation analysis is performed on the interaction of variables in the node set to obtain a dynamic correlation measure between the process variables and the environmental variables; By performing causal-guided inference on the dynamic correlation metric, the directed edge set of the encapsulation line is obtained; The node set and the directed edge set are topologically reconstructed to obtain the dynamic causal structure graph of the encapsulation line.

[0020] This process involves collecting raw, time-varying data generated during the actual production operation of the PE packaging line. This data covers key information throughout the entire packaging line operation process, including but not limited to the operating speed of the pressure rollers, the output power of the heating device, the material transfer speed, real-time temperature feedback of the heating zone, and equipment operating current. A systematic review of this raw time-series data is conducted. Professional verification methods are used to identify and eliminate abnormal fluctuations and invalid records. For missing data points, appropriate supplementation is performed based on the time-series variation patterns and characteristics of adjacent valid data. Simultaneously, the recording format and time sampling scale of all data are standardized to ensure synchronization across different data types. Through this process of review, verification, supplementation, and format standardization, a high-dimensional time-series dataset covering key parameters of the packaging line operation and ensuring reliable data quality is ultimately generated.

[0021] In advance, based on the working principle, production process requirements, and actual control needs of the PE packaging line, a clear and executable variable function classification standard is formulated. This standard clearly defines the functional attributes and classification basis of different variables. Based on this pre-set classification standard, all operating condition variables contained in the high-dimensional time series data are identified and classified one by one. Variables that directly participate in the production operation process of the packaging line and can be adjusted through equipment control are classified into the process variable group, such as pressure roller speed, heating device power, and material supply speed. External variables that affect the operating status of the packaging line but are not directly controlled by the production process are classified into the environmental variable group, such as temperature, humidity, and power grid voltage fluctuations in the workshop. Through this classification operation, the process variable group and environmental variable group of the packaging line are obtained respectively.

[0022] For the process variable group and environmental variable group obtained by classification, each variable is independently identified and defined in detail. Each variable is given unique identification information, and the specific physical meaning, data collection range, numerical change law and other characteristics of each variable are clarified. Each variable is transformed into an independent basic unit that can be used for network structure analysis. This independent unit is a node. All the identified and defined process variable nodes and environmental variable nodes together constitute the node set of the encapsulation line.

[0023] This study delves into the interaction processes of variables corresponding to each node in the node set during the operation of the packaging line, focusing on exploring the time delay characteristics of the mutual influence between different variables. By continuously tracking the numerical changes of variables, the study statistically analyzes the time interval between a change in one variable and a corresponding change in another related variable, while quantifying the strength of this influence. For example, it analyzes how long it takes for the heating temperature of the pressure roller to change after an increase in ambient temperature, and the magnitude of this temperature change. Through this analysis, quantitative results that comprehensively reflect the strength of the correlation and time delay between variables are obtained; these results constitute a dynamic correlation measure between process variables and environmental variables.

[0024] Based on the dynamic correlation measurement results obtained above, and combined with the production operation logic, process flow, and actual interaction patterns between variables of the PE packaging line, the directionality of the influence relationships between variables is determined. It clearly distinguishes which variable's change causes changes in other variables, and which variable's change is a result of being influenced by other variables, thus determining the specific direction of the causal relationship between variables. Each pair of variables with a causal relationship is represented by edges with clear directions; the starting end of the edge corresponds to the causal variable node, and the ending end corresponds to the result variable node. All such directed edges are integrated together to form the directed edge set of the packaging line.

[0025] Following the causal relationships defined by the directed edge set, the nodes in the node set are connected in an orderly manner through their corresponding directed edges. During this connection process, the hierarchical relationships between variables are fully considered, ensuring that the variable node at the front of the causal chain correctly points to the subsequent variable node it influences through directed edges, thus constructing a network structure that completely and clearly reflects the dynamic causal relationships between all variables. Through this topology integration and reconstruction operation, the dynamic causal structure diagram of the PE encapsulation line is finally formed.

[0026] The beneficial effects include ensuring the integrity, accuracy, and consistency of the original time-series data, providing a reliable data foundation for subsequent causal relationship construction, avoiding interference from poor-quality data in the analysis results, achieving accurate classification of operating condition variables, clarifying the functional attributes of variables, avoiding analytical confusion caused by mixed variables, laying a clear foundation for node-based processing and causal analysis, transforming variables into independent and identifiable node units, forming an ordered set of nodes, providing a carrier for constructing a network of relationships between variables, facilitating the intuitive representation of variable interactions, capturing the time lag characteristics and strength differences of relationships between variables, avoiding biases in relationship judgments caused by ignoring time lags, providing accurate evidence for inferring causal relationships, clearly defining the direction of causal relationships between variables, avoiding misjudging relationships and causality or confusing directions, and the resulting directed edge set providing core edge structure support for the dynamic causal structure graph, integrating nodes and directed edges into an intuitive dynamic causal structure graph, clearly showing the causal paths and hierarchical relationships between variables, improving the efficiency and accuracy of subsequent screening of key paths and tracing precursor variables.

[0027] S2. Based on the pressure roller temperature as the target node, the dynamic causal structure diagram is filtered for causal influence relationships to obtain the local causal network of the pressure roller temperature, and the key causal paths in the local causal network are traced to obtain the precursor variables of the pressure roller temperature. In this embodiment of the invention, the step of filtering causal relationships in the dynamic causal structure graph based on the pressure roller temperature as the target node to obtain a local causal network of the pressure roller temperature, and tracing the key causal paths in the local causal network to obtain the precursor variables of the pressure roller temperature, includes: Using the pressure roller temperature as the target node, direct causal edge extraction is performed on the dynamic causal structure graph to obtain the direct association set of the pressure roller temperature; In the dynamic causal structure graph, multi-order causal backtracking is performed starting from the end node of the causal edge in the direct association set to obtain the indirect causal path cluster of the pressure roller temperature; Using the directly associated set as points and the indirect causal path cluster as edges, a complete causal network for the pressure roller temperature is constructed. The path strength of each causal path in the complete causal network is quantified to obtain the causal influence weight of the complete causal network. Based on the causal influence weights, the path saliency of all paths in the complete causal network is screened to obtain the key causal path of the pressure roller temperature; By tracing the key causal path, the precursor variable of the pressure roller temperature is obtained.

[0028] Clearly define the core position of the pressure roller temperature in the dynamic causal structure diagram, and take it as the target object for analyzing influencing factors. Comprehensively review all causal connections in the dynamic causal structure diagram, and determine whether the endpoint of each causal connection matches the node corresponding to the pressure roller temperature. Filter out all causal connections whose endpoints are exactly the pressure roller temperature node; these filtered causal connections together constitute the direct association set of pressure roller temperature.

[0029] In the constructed dynamic causal structure diagram, the terminal nodes of each causal connection in the directly related set are first accurately identified. These terminal nodes are the preceding nodes that directly affect the pressure roller temperature node. Using each terminal node as the initial search point, the search proceeds in the reverse direction of the causal connection. First, the first-level preceding nodes and their corresponding causal connections that directly affect the terminal node are identified. Then, using the first-level preceding nodes as the new starting point, the search continues in reverse to find the second-level preceding nodes and their corresponding causal connections that affect them, progressing step by step in this manner. Considering the actual operating conditions of the packaging line and the tightness of the data correlation, the number of backtracking levels is reasonably set. All causal paths obtained through multi-level reverse searches are collected to form an indirect causal path cluster for the pressure roller temperature.

[0030] Each node involved in a causal connection within the directly related set serves as the core support point for constructing the network. Each causal path in the indirect causal path cluster acts as a link between the core node and the indirectly affected nodes. The core node and its corresponding indirectly affected node are connected in an orderly manner, strictly following the actual causal relationship of each path, while maintaining the original connections between directly related nodes. Through this combination, a clear and complete causal network for the pressure roller temperature is ultimately formed, encompassing both direct and indirect influences.

[0031] Based on historical operating data of the packaging line, the required control precision of the pressure roller temperature, and the actual correlation between various variables, an evaluation standard system is established, including key dimensions such as correlation frequency, influence magnitude, and duration. For each causal path in the complete causal network, its specific performance under each evaluation dimension is analyzed. A corresponding weight is assigned according to the importance of each evaluation dimension, and then a specific numerical value is calculated for each causal path. This numerical value is the causal influence weight of that causal path, which directly reflects the intensity of its influence on the pressure roller temperature.

[0032] Based on the actual production needs of the packaging line, the priority of pressure roller temperature control, and historical control experience, a reasonable screening threshold is determined. This threshold effectively distinguishes between paths that have a significant impact on pressure roller temperature and those that have a less significant impact. The causal influence weight of each causal path in the complete causal network is compared one by one with the set screening threshold. Causal paths with weights greater than or equal to the screening threshold are selected. These selected paths are the critical causal paths that have a significant impact on pressure roller temperature. The specific structure and associated node information of each critical causal path are recorded in detail.

[0033] For each selected critical causal path, starting from the target node of the pressure roller temperature, the variables corresponding to each node on the path are searched one by one along the reverse causal direction of the path. This includes variables that are directly related as well as variables that are indirectly related through multiple levels. During the search, it is ensured that no variable on the path is missed. All variables involved in all critical causal paths are comprehensively collected, and duplicate variables are removed. These collected and deduplicated variables are the precursor variables that can directly or indirectly affect the pressure roller temperature. The specific position of each precursor variable in the critical causal path and its relationship with other variables are clearly recorded.

[0034] The beneficial effects include: quickly focusing on the elements directly affecting the pressure roller temperature, eliminating interference from irrelevant causal connections, improving the efficiency of causal relationship screening, laying a precise foundation for subsequent analysis, comprehensively exploring potential paths indirectly affecting the pressure roller temperature, avoiding the omission of important indirect factors, making the analysis of pressure roller temperature influencing factors more complete, integrating direct and indirect influencing factors, constructing a structured and complete causal network, improving the systematicness and orderliness of pressure roller temperature influencing factor analysis, transforming abstract causal influences into comparable values, providing objective and accurate basis for screening key causal paths, improving the scientificity and accuracy of screening, accurately focusing on the key paths that play a dominant role in pressure roller temperature, eliminating interference from secondary paths, reducing subsequent workload, improving the accuracy and efficiency of influencing factor analysis, accurately identifying all key precursor variables affecting pressure roller temperature, providing a clear basis for formulating feedforward compensation strategies, and improving the pertinence and effectiveness of pressure roller temperature control.

[0035] S3. Perform rule mapping on the dynamic behavior patterns in the precursor variables to obtain the feedforward compensation strategy for the pressure roller temperature. In this embodiment of the invention, the step of performing rule mapping on the dynamic behavior patterns in the precursor variables to obtain the feedforward compensation strategy for the pressure roller temperature includes: Based on the empirical rules of the historical optimal control process of the packaging line, the basic compensation rules for the pressure roller temperature are obtained; The predecessor variables are dynamically evolved and their behavior trajectory sequences are obtained by tracking their dynamic evolution. By mining the pattern features of the behavioral trajectory sequence, the saliency features of the dynamic behavioral pattern are obtained; The saliency features are matched and mapped with the basic compensation rules to obtain candidate compensation rules for the pressure roller temperature; A multi-objective optimization decision is made on the candidate compensation rules to obtain the feedforward compensation strategy for the pressure roller temperature.

[0036] The criteria for determining the historical optimal control process of the packaging line were defined, selecting the operating period where the pressure roller temperature fluctuation was within the minimum range and the packaged product yield reached the highest level as the historical optimal control process. Comprehensive data was collected during these periods, including pressure roller temperature adjustment parameters, environmental conditions, packaging line operating load, and the working status of relevant components. The collected information was systematically analyzed, and recurring control logic and operating procedures were extracted. These logics and procedures, proven effective in practice, constitute the empirical rules of the historical optimal control process. These empirical rules were then categorized and integrated according to control scenarios, influencing factors, and other dimensions to form a set of universally applicable and practical pressure roller temperature basic compensation rules, providing an initial basis for the construction of subsequent compensation strategies.

[0037] For the identified precursor variable of the pressure roller temperature, a reasonable data acquisition time interval is set to ensure timely capture of changes in the variable. Real-time values ​​of each precursor variable are continuously acquired at the set intervals, and the corresponding time points are recorded. During the acquisition process, the changing trends of each precursor variable are closely monitored, including different states such as rising, falling, and remaining stable, as well as the rate of change. The values ​​of each precursor variable at different time points and their corresponding states are arranged chronologically to form a complete and continuous sequence of precursor variable behavior, fully presenting the dynamic change process of the precursor variable.

[0038] A thorough analysis of the resulting precursor variable behavior trajectory sequences is conducted, focusing on the magnitude of variable value changes (i.e., the difference between values ​​at adjacent time points); frequency of change (i.e., the number of times the value changes significantly); duration (i.e., the length of time the variable maintains a certain trend or stable state); and the temporal patterns and numerical characteristics of peaks and troughs. By comparing trajectory sequences under different time periods and scenarios, features that significantly influence the temperature change of the pressure roller, recur in multiple runs, or exhibit highly representative trends are identified. These selected features with key influence are the salient features of the dynamic behavior pattern.

[0039] First, the basic compensation rules are broken down to clarify the core adaptation conditions, such as the applicable scenarios, triggering conditions, and corresponding adjustment methods for each rule. The previously extracted saliency features of dynamic behavior patterns are then compared and analyzed with the adaptation conditions of each basic compensation rule to determine whether the saliency feature matches the applicable scenario and triggering requirements of a particular basic compensation rule. Basic compensation rules that highly match the adaptation conditions with the saliency features are then selected. These selected basic compensation rules, which match the current behavior pattern of the precursor variable, collectively constitute the candidate compensation rules for the pressure roller temperature.

[0040] The pressure roller temperature feedforward compensation strategy has several clearly defined objectives, including improving the control accuracy of the pressure roller temperature, accelerating the response speed of temperature adjustment, ensuring the overall stability of the packaging line, and reducing energy consumption. For each candidate compensation rule, its performance under each objective is evaluated. For example, one rule may excel in control accuracy but have a slightly slower response speed, while another rule may have advantages in both response speed and stability. The advantages and disadvantages of each candidate compensation rule are comprehensively weighed, considering both performance under a single objective and the balance between multiple objectives. Finally, one or more candidate compensation rules that exhibit the best overall performance across multiple objectives are selected and integrated to form a pressure roller temperature feedforward compensation strategy that can meet various control requirements.

[0041] The beneficial effects are as follows: the basic compensation rules are derived from the historical optimal control experience of the packaging line, ensuring the reliability and practicality of the rules; after sorting and integration, the structure is clear and the applicable scenarios are well-defined, laying an effective foundation for subsequent matching with significant features. Continuously collecting precursor variable data at reasonable intervals can comprehensively capture its dynamic evolution process, avoiding misjudgment of behavior patterns due to missing or lagging data. The resulting behavior trajectory sequence provides continuous and systematic support for subsequent mining of pattern features. Focusing on the core information of the trajectory sequence to extract significant features can effectively eliminate irrelevant data and interference information. Clear significant features make subsequent matching with basic compensation rules more targeted, avoiding rule adaptation bias. By accurately comparing the significant features with the adaptation conditions of basic compensation rules, rules that fit the current variable behavior pattern can be selected, inapplicable rules can be eliminated, the scope of subsequent optimization decisions can be narrowed, invalid interference can be reduced, and decision-making efficiency and accuracy can be improved. Multi-objective optimization decisions take into account the requirements of control precision, response speed, and operational stability, avoiding the limitations of single-objective orientation; the integrated feedforward compensation strategy can achieve optimal balance in multiple dimensions, can cope with complex packaging line operation scenarios, and achieve efficient and precise control of pressure roller temperature.

[0042] S4. Perform feedforward manipulation variable analysis on the feedforward compensation strategy and the real-time state vector in the precursor variables to obtain the predictive compensation amount of the pressure roller temperature. In this embodiment of the invention, the feedforward compensation strategy and the real-time state vector in the precursor variables are analyzed by feedforward manipulation variables to obtain the predictive compensation amount for the pressure roller temperature, including: The feedforward compensation strategy is deconstructed to obtain the control rule fragments of the feedforward compensation strategy; Based on the control rule fragment, variable selection is performed on the real-time state vector in the predecessor variables to obtain a subset of key variables of the real-time state vector. Deviation identification is performed on the subset of key variables to obtain the real-time deviation value of the subset of key variables; The real-time deviation value and the control rule segment are dynamically weighted and fused to obtain the predictive compensation amount for the pressure roller temperature.

[0043] The real-time deviation value and the control rule segment are dynamically weighted and fused to obtain the predictive compensation amount for the pressure roller temperature. The specific formula for calculating the predictive compensation amount is as follows: ; In the formula, The predictive compensation amount, This represents the total number of predecessor variables contained in the subset of key variables. For the first The dynamic weighting coefficients of the aforementioned precursor variables, For the first The real-time deviation values ​​of the aforementioned precursor variables. The absolute magnitude of the real-time deviation value. For the first The time constants of the aforementioned precursor variables The differential gain coefficient, This indicates that the expression within the parentheses is subjected to time differentiation. It is an exponential function.

[0044] First, comprehensively analyze the overall control logic of the feedforward compensation strategy, clarifying core elements such as the direction of temperature regulation, applicable operating conditions, and the sequential relationship of each regulation step. Based on different control objectives, applicable scenarios, and specific types of adjustment actions, systematically break down the complete feedforward compensation strategy. This decomposes the originally coherent strategy into multiple independent, individually applicable, and logically complete specific rule units. These decomposed rule units are the control rule fragments. During the decomposition process, it is crucial to ensure that each rule fragment clearly reflects a specific control logic, without omitting key regulation requirements or exhibiting logical duplication.

[0045] Each control rule segment is analyzed individually to clarify the temperature control requirements, the regulation targets to be achieved, and the specific requirements for related variables. Based on these clarified requirements, all variables included in the real-time state vector of the precursor variables are screened and analyzed one by one. The correlation between each variable and the corresponding control rule segment is determined, and the direct impact and significance of the variable on the pressure wheel temperature regulation effect are assessed. Variables that are highly correlated with the control rule segment, can directly affect the pressure wheel temperature regulation, and have a significant impact are selected. These selected variables are integrated to form a key variable subset of the real-time state vector.

[0046] First, collect the historical optimal operating data for each variable in the key variable subset during the long-term stable operation of the packaging line. Combined with the design standards and temperature control targets of the packaging line, determine a stable and reasonable standard baseline value for each key variable. Then, collect the current actual operating value of each variable in the key variable subset using real-time monitoring equipment. Compare and analyze the current actual value of each variable with the predetermined standard baseline value one by one, accurately calculating the difference between the two. This difference is the real-time deviation value of that key variable.

[0047] Taking into account the importance and priority of each control rule segment in the overall temperature control, and considering the relative influence of each variable in the key variable subset on the pressure roller temperature, a corresponding weight value is assigned to the real-time deviation value of each key variable. This weight value is not fixed but adaptively adjusted based on the real-time operating conditions of the packaging line, load changes, environmental fluctuations, etc., to ensure that the weight accurately reflects the influence of variables and rules under the current operating conditions. Then, the real-time deviation value of each key variable is multiplied by its corresponding dynamic weight. Based on the specific requirements of the control rule segment, all calculation results are integrated and calculated, comprehensively considering the influence of various factors, to finally derive a predictive compensation amount that can accurately adjust the pressure roller temperature.

[0048] In the formula for calculating the predictive compensation amount This is the desired predictive compensation amount for the pressure roller temperature, used for pressure roller temperature adjustment to match actual deviation requirements. This represents the total number of predecessor variables in the subset of key variables, derived from previous variable selection results. For the first The dynamic weighting coefficients of each precursor variable are determined by the variable's influence on the pressure roller temperature, the importance of the control rules, and real-time operating conditions, and are used to highlight the role of key factors. For the first The real-time deviation value of each precursor variable, obtained by comparing the actual value of the variable with the standard benchmark value, is the core data for calculating the compensation amount. Positive and negative values ​​indicate the direction of deviation, while large and small values ​​indicate the degree of deviation. This represents the absolute magnitude of the deviation, measuring only the magnitude of the deviation and avoiding directional interference in the calculation. For the first The time constants of the precursor variables reflect the response speed to changes in deviation and regulate the pace of change in the influence of deviation. As an exponential function, it performs non-linear adjustment on the impact of deviation, weakening the extreme effects of excessive deviation and making the compensation amount change more smoothly. This is the differential gain coefficient, used to adjust the impact of the dynamic correction on the compensation amount, and is set according to operating condition fluctuations. It is a time differential operation, calculating the rate of change of the expression within parentheses to capture the dynamic effects of operating conditions. It is in the order of variables, for each variable The results are summed to obtain the total static impact. The overall formula first calculates the total static compensation impact of each variable, then adds the dynamic correction amount, and finally obtains the predictive compensation amount, which takes into account both real-time static deviation and dynamic changes in operating conditions.

[0049] The beneficial effects are as follows: splitting the feedforward compensation strategy into control rule segments reduces redundant interference and improves the pertinence and efficiency of subsequent variable selection and deviation calculation. Key variables are selected according to control rules, and irrelevant and weakly influential variables are excluded, reducing data processing volume and avoiding interference, laying the foundation for accurate deviation calculation. Real-time deviation is obtained by comparing the actual value of the variable with the standard benchmark value, providing accurate basic data, clarifying the compensation target, avoiding insufficient or excessive compensation, dynamically adjusting weights and integrating data, so that the predictive compensation amount fits the real-time operating conditions, improving the accuracy and timeliness of compensation, and helping to stabilize the pressure roller temperature. This formula integrates the static influence and dynamic trend of key variables, balances the role of various factors, and ensures that the compensation amount calculation is scientific and accurate, closely matching the actual operating needs of the packaging line.

[0050] S5. Based on the predictive compensation amount, make collaborative decisions on the control command for the pressure roller temperature to obtain the comprehensive control command for the pressure roller temperature; In this embodiment of the invention, the step of making collaborative decisions on the control command for the pressure roller temperature based on the predictive compensation amount to obtain a comprehensive control command for the pressure roller temperature includes: The predictive compensation amount is discretized in a distributed manner to obtain the influence distribution of the pressure roller temperature; Based on the influence distribution, the control command for the pressure roller temperature is weighted by multiple objectives to obtain the weighted decision criterion for the pressure roller temperature. Based on the weighted decision criterion, the control command for the pressure roller temperature is subjected to multi-objective trade-offs to obtain the trade-off control command for the pressure roller temperature. Based on the dynamic operating margin of the packaging line, the stability of the control command after the trade-off is evaluated, and the enhanced command for the pressure roller temperature is obtained. The enhanced command and the real-time feedback control command for the pressure roller temperature are fused and optimized to obtain a comprehensive control command for the pressure roller temperature.

[0051] Based on the actual operating scenarios of the PE packaging line pressure rollers, operating condition ranges covering different operating states are divided, and the temperature control dimension corresponding to each range is clearly defined. The predictive compensation amount is broken down according to the operating condition range and control dimension, so that each sub-compensation amount accurately corresponds to the specific control scenario, clearly presenting the impact of different sub-components on the pressure roller temperature, forming the influence distribution of the pressure roller temperature.

[0052] Based on the distribution of influence, the core objectives of pressure roller temperature control are defined, including temperature stability, control response, energy saving and consumption reduction, and equipment protection. According to the magnitude of the control effect of each sub-influence, a corresponding importance percentage is assigned to each control objective, with higher-influence objectives assigned a higher percentage, forming a weighted decision-making benchmark for control command decisions.

[0053] Various candidate control commands for pressure roller temperature control are collected. Based on a weighted decision-making criterion, the degree to which each candidate command satisfies each control objective is evaluated, and the overall effect is quantified by combining the proportion of objective importance. Following priority ranking, the satisfaction of high-proportion objectives is guaranteed first, while taking into account the basic needs of other objectives, and the control command with the best overall effect is selected.

[0054] Real-time monitoring of the PE packaging line's operating status collects key data such as equipment load, energy supply, and component tolerance to determine the current tolerable range of control fluctuations and load redundancy, i.e., dynamic operating margin. The balanced control commands are then substituted into this margin for simulation analysis to determine if there are any instability risks during execution. Commands are adjusted and optimized to address these risks, resulting in enhanced commands adapted to the operating conditions.

[0055] The system acquires both enhanced control commands and feedback control commands reflecting the real-time temperature status of the pressure roller. It compares the key elements of these two commands, such as control direction, intensity, and timing, to identify conflicting and overlapping control measures. For conflicting measures, a reasonable control scheme is determined; for overlapping measures, the control intensity is optimized; and the system integrates the forward-looking nature of enhanced commands with the immediacy of feedback commands to form a comprehensive control command.

[0056] The beneficial effects include: refining the analysis of influencing factors, enabling precise matching of compensation amounts with control scenarios, avoiding control deviations, improving the pertinence and accuracy of decision-making, ensuring that weight allocation aligns with actual impacts, preventing key objectives from being overlooked, guaranteeing a scientific and reasonable decision-making benchmark, enhancing the effectiveness of control strategies, achieving a balance between multiple control objectives, avoiding the one-sidedness of single-objective-oriented commands, improving the overall quality of pressure roller temperature control, proactively mitigating operational mismatch risks, ensuring stable and safe command execution, enhancing the adaptability of commands to real-time operating conditions, improving reliability, integrating the advantages of both types of commands, compensating for the shortcomings of single commands, achieving a balance between temperature fluctuation prediction and real-time response, and improving control accuracy and efficiency.

[0057] S6. Based on the execution effect of the integrated control command and the topological consistency of the dynamic causal structure graph, the applicability of the feedforward compensation strategy is determined by stability, and the reconstruction trigger signal of the feedforward compensation strategy is obtained. In this embodiment of the invention, the stability determination of the applicability of the feedforward compensation strategy based on the execution effect of the integrated control command and the topological consistency of the dynamic causal structure graph, to obtain the reconstruction trigger signal of the feedforward compensation strategy, includes: The integrated control commands are applied to the packaging line to extract the control performance of the packaging line. The operating characteristics of the pressure roller temperature are obtained; Based on the dynamic causal structure graph, the causal topological coupling analysis of the operational characteristics is performed to obtain the causal consistency coefficient of the feedforward compensation strategy; Based on the causal consistency coefficient, the effectiveness of the feedforward compensation strategy is determined, and the applicability of the feedforward compensation strategy is determined. Based on the applicability decision, the stability state of the feedforward compensation strategy is determined, and the reconstruction trigger signal of the feedforward compensation strategy is obtained.

[0058] The packaging line control execution unit receives comprehensive control commands and adjusts the operating status of components such as the power of the pressure roller heating device and the speed of the transmission mechanism according to the commands to ensure that the commands are accurately applied to the packaging line. During execution, data such as the actual temperature of the pressure roller, the rate of temperature change, and the temperature difference change are continuously collected within a preset time to extract control effectiveness; then, this data is analyzed to determine the patterns of temperature stability and fluctuations, thus forming the operating characteristics of the pressure roller temperature.

[0059] Based on the constructed dynamic causal structure diagram, the dynamic causal structure diagram clarifies the correlation and tightness of variables affecting the pressure roller temperature. For the operating characteristics, the source of influence of the variable corresponding to each characteristic is identified and matched with the variable correlation relationship in the structure diagram. The degree of fit between the variable effect reflected by the characteristic and the preset causal relationship is analyzed. The degree of fit is converted into a numerical value that intuitively reflects the consistency between the causal relationship of the strategy and the actual control, namely the causal consistency coefficient.

[0060] The causal consistency coefficient is divided into multiple level ranges, with each range corresponding to a specific strategy effectiveness level. The coefficient is assigned to the corresponding range. Combining the packaging line production process, product quality, and temperature control stability requirements, the strategy corresponding to the coefficient is analyzed to determine whether it meets the temperature control requirements and can cope with variable changes. The applicability decision is then made to determine whether the strategy is currently suitable for continued application.

[0061] For different applicability decisions, corresponding stability judgment criteria are established: suitable for use is stable, needs adjustment to be critically stable, and needs to be replaced is unstable. The current state of the strategy is defined by comparing it with the criteria; the strategy can continuously and stably control the temperature in the current and expected production cycle, and whether there is a risk of temperature runaway; a signal is generated based on the judgment result: if stable, no reconstruction is needed; if critical or unstable, reconstruction is required.

[0062] The beneficial effects include ensuring accurate implementation of instructions and comprehensive performance data, providing a solid foundation for subsequent strategy applicability assessments, avoiding judgment biases caused by insufficient data support, ensuring logical consistency in analysis and avoiding directional deviations by relying on existing dynamic cause-effect structure diagrams, providing reference standards for strategy applicability assessments by quantification coefficients, improving the objectivity and accuracy of results, reducing subjective errors, combining judgment intervals with production standards to make performance assessments systematic and avoid subjectivity, accurately reflecting the actual value of the strategy in the decision-making process, providing a clear basis for subsequent strategy adjustments or reconstruction, ensuring the practicality of the judgment, and ensuring that the decision-making results directly correspond to the judgment standards, thus ensuring accurate stability assessments; the reconstruction trigger signal accurately guides strategy optimization, promptly handles unstable strategies, avoids temperature control failures, and ensures continuous production on the packaging line and stable product quality.

[0063] Example 2 like Figure 2 As shown in the figure, this embodiment also provides a functional block diagram of the pressure roller heating temperature control system for the PE packaging line.

[0064] The pressure roller heating temperature control system 100 of the PE packaging line described in this embodiment can be installed in an electronic device. Depending on the functions implemented, the pressure roller heating temperature control system 100 of the PE packaging line may include a dynamic causal structure graph construction module 101, a causal screening and precursor variable acquisition module 102, a feedforward compensation strategy generation module 103, a predictive compensation quantity analysis module 104, a comprehensive control command decision module 105, and a compensation strategy stability determination module 106. The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and are stored in the memory of the electronic device.

[0065] In this embodiment, the functions of each module / unit are as follows: The dynamic causal structure graph construction module 101 is used to construct a dynamic causal structure graph with process variables and environmental variables in the high-dimensional time series dataset as nodes and the dynamic causal relationship between the process variables and the environmental variables as edges. The causal screening and precursor variable acquisition module 102 is used to screen the causal influence relationship of the dynamic causal structure diagram based on the pressure roller temperature as the target node, so as to obtain the local causal network of the pressure roller temperature, and trace the key causal path in the local causal network to obtain the precursor variable of the pressure roller temperature. The feedforward compensation strategy generation module 103 is used to perform rule mapping on the dynamic behavior patterns in the precursor variables to obtain the feedforward compensation strategy for the pressure roller temperature. The predictive compensation analysis module 104 is used to analyze the feedforward manipulator variables of the feedforward compensation strategy and the real-time state vector in the precursor variables to obtain the predictive compensation amount of the pressure roller temperature. The integrated control command decision module 105 is used to make collaborative decisions on the control command for the pressure roller temperature based on the predictive compensation amount, so as to obtain the integrated control command for the pressure roller temperature. The stability determination module 106 of the compensation strategy is used to determine the applicability of the feedforward compensation strategy based on the execution effect of the integrated control command and the topological consistency of the dynamic causal structure graph, and to obtain the reconstruction trigger signal of the feedforward compensation strategy.

[0066] In detail, each module in the pressure roller heating temperature control system 100 of the PE packaging line described in the embodiments of the present invention adopts the same technical means as the pressure roller heating temperature control method of the PE packaging line described in Embodiment 1 and Embodiment 2, and can produce the same technical effect, which will not be repeated here.

[0067] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0068] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0069] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0070] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0071] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for controlling the heating temperature of the pressure roller in a PE packaging line, characterized in that, The method includes: S1. Using process variables and environmental variables in the high-dimensional time series dataset as nodes, and the dynamic causal relationship between the process variables and the environmental variables as edges, construct a dynamic causal structure graph of encapsulation lines. S2. Based on the pressure roller temperature as the target node, the dynamic causal structure diagram is filtered for causal influence relationships to obtain the local causal network of the pressure roller temperature, and the key causal paths in the local causal network are traced to obtain the precursor variables of the pressure roller temperature. S3. Perform rule mapping on the dynamic behavior patterns in the precursor variables to obtain the feedforward compensation strategy for the pressure roller temperature. S4. Perform feedforward manipulation variable analysis on the feedforward compensation strategy and the real-time state vector in the precursor variables to obtain the predictive compensation amount of the pressure roller temperature. S5. Based on the predictive compensation amount, make collaborative decisions on the control command for the pressure roller temperature to obtain the comprehensive control command for the pressure roller temperature; S6. Based on the execution effect of the integrated control command and the topological consistency of the dynamic causal structure graph, the applicability of the feedforward compensation strategy is determined to obtain the reconstruction trigger signal of the feedforward compensation strategy.

2. The method for controlling the heating temperature of the pressure rollers in a PE packaging line as described in claim 1, characterized in that, The method of constructing a dynamic causal structure graph with encapsulated lines, using process variables and environmental variables in a high-dimensional time-series dataset as nodes and the dynamic causal relationships between the process variables and environmental variables as edges, includes: The original timing data of the packaging line is obtained, and the original timing data is coordinated to obtain the high-dimensional timing data of the packaging line. Using preset variable function classification rules, the operating condition variables in the high-dimensional time series data are divided to obtain the process variable group and environmental variable group of the packaging line; The process variable group and the environment variable group are node-based to obtain the node set of the encapsulation line; Time-delay correlation analysis is performed on the interaction of variables in the node set to obtain a dynamic correlation measure between the process variables and the environmental variables; By performing causal-guided inference on the dynamic correlation metric, the directed edge set of the encapsulation line is obtained; The node set and the directed edge set are topologically reconstructed to obtain the dynamic causal structure graph of the encapsulation line.

3. The method for controlling the heating temperature of the pressure rollers in a PE packaging line as described in claim 1, characterized in that, The process involves using the pressure roller temperature as the target node, filtering causal relationships in the dynamic causal structure graph to obtain a local causal network for the pressure roller temperature, and tracing key causal paths within this local causal network to obtain the antecedent variables of the pressure roller temperature, including: Using the pressure roller temperature as the target node, direct causal edge extraction is performed on the dynamic causal structure graph to obtain the direct association set of the pressure roller temperature; In the dynamic causal structure graph, multi-order causal backtracking is performed starting from the end node of the causal edge in the direct association set to obtain the indirect causal path cluster of the pressure roller temperature; Using the directly associated set as points and the indirect causal path cluster as edges, a complete causal network for the pressure roller temperature is constructed. The path strength of each causal path in the complete causal network is quantified to obtain the causal influence weight of the complete causal network. Based on the causal influence weights, the path saliency of all paths in the complete causal network is screened to obtain the key causal path of the pressure roller temperature; By tracing the key causal path, the precursor variable of the pressure roller temperature is obtained.

4. The method for controlling the heating temperature of the pressure rollers in a PE packaging line as described in claim 1, characterized in that, The step of mapping the dynamic behavior patterns in the precursor variables to obtain the feedforward compensation strategy for the pressure roller temperature includes: Based on the empirical rules of the historical optimal control process of the packaging line, the basic compensation rules for the pressure roller temperature are obtained; The predecessor variables are dynamically evolved and their behavior trajectory sequences are obtained by tracking their dynamic evolution. By mining the pattern features of the behavioral trajectory sequence, the saliency features of the dynamic behavioral pattern are obtained; The saliency features are matched and mapped with the basic compensation rules to obtain candidate compensation rules for the pressure roller temperature; A multi-objective optimization decision is made on the candidate compensation rules to obtain the feedforward compensation strategy for the pressure roller temperature.

5. The method for controlling the heating temperature of the pressure rollers in a PE packaging line as described in claim 1, characterized in that, By performing feedforward manipulation variable analysis on the feedforward compensation strategy and the real-time state vector in the precursor variables, a predictive compensation amount for the pressure roller temperature is obtained, including: The feedforward compensation strategy is deconstructed to obtain the control rule fragments of the feedforward compensation strategy; Based on the control rule fragment, variable selection is performed on the real-time state vector in the predecessor variables to obtain a subset of key variables of the real-time state vector. Deviation identification is performed on the subset of key variables to obtain the real-time deviation value of the subset of key variables; The real-time deviation value and the control rule segment are dynamically weighted and fused to obtain the predictive compensation amount for the pressure roller temperature.

6. The method for controlling the heating temperature of the pressure rollers in a PE packaging line as described in claim 5, characterized in that, The real-time deviation value and the control rule segment are dynamically weighted and fused to obtain the predictive compensation amount for the pressure roller temperature. The specific formula for calculating the predictive compensation amount is as follows: ； In the formula, The predictive compensation amount, This represents the total number of predecessor variables contained in the subset of key variables. For the first The dynamic weighting coefficients of the aforementioned precursor variables, For the first The real-time deviation values ​​of the aforementioned precursor variables. The absolute magnitude of the real-time deviation value. For the first The time constants of the aforementioned precursor variables The differential gain coefficient, This indicates that the expression within the parentheses is subjected to time differentiation. It is an exponential function.

7. The method for controlling the heating temperature of the pressure rollers in a PE packaging line as described in claim 1, characterized in that, The step of making collaborative decisions on the control command for the pressure roller temperature based on the predictive compensation amount to obtain a comprehensive control command for the pressure roller temperature includes: The predictive compensation amount is discretized in a distributed manner to obtain the influence distribution of the pressure roller temperature; Based on the influence distribution, the control command for the pressure roller temperature is weighted by multiple objectives to obtain the weighted decision criterion for the pressure roller temperature. Based on the weighted decision criterion, the control command for the pressure roller temperature is subjected to multi-objective trade-offs to obtain the trade-off control command for the pressure roller temperature. Based on the dynamic operating margin of the packaging line, the stability of the control command after the trade-off is evaluated, and the enhanced command for the pressure roller temperature is obtained. The enhanced command and the real-time feedback control command for the pressure roller temperature are fused and optimized to obtain a comprehensive control command for the pressure roller temperature.

8. The method for controlling the heating temperature of the pressure rollers in a PE packaging line as described in claim 1, characterized in that, The applicability of the feedforward compensation strategy is determined by the stability assessment of the execution effect of the integrated control command and the topological consistency of the dynamic causal structure graph, resulting in the reconstruction trigger signal of the feedforward compensation strategy, including: The integrated control commands are applied to the packaging line to extract the control performance of the packaging line. The operating characteristics of the pressure roller temperature are obtained; Based on the dynamic causal structure graph, the causal topological coupling analysis of the operational characteristics is performed to obtain the causal consistency coefficient of the feedforward compensation strategy; Based on the causal consistency coefficient, the effectiveness of the feedforward compensation strategy is determined, and the applicability of the feedforward compensation strategy is determined. Based on the applicability decision, the stability state of the feedforward compensation strategy is determined, and the reconstruction trigger signal of the feedforward compensation strategy is obtained.

9. A temperature control system for the pressure rollers of a PE packaging line, characterized in that, include: The dynamic causal structure graph construction module is used to construct a dynamic causal structure graph with process variables and environmental variables in a high-dimensional time series dataset as nodes and the dynamic causal relationship between the process variables and the environmental variables as edges. The causal screening and precursor variable acquisition module is used to screen the causal influence relationship of the dynamic causal structure diagram based on the pressure roller temperature as the target node, so as to obtain the local causal network of the pressure roller temperature, and trace the key causal path in the local causal network to obtain the precursor variable of the pressure roller temperature. The feedforward compensation strategy generation module is used to perform rule mapping on the dynamic behavior patterns in the precursor variables to obtain the feedforward compensation strategy for the pressure roller temperature. The predictive compensation analysis module is used to analyze the feedforward manipulation variables of the feedforward compensation strategy and the real-time state vector in the precursor variables to obtain the predictive compensation amount of the pressure roller temperature. The integrated control command decision module is used to make collaborative decisions on the control command for the pressure roller temperature based on the predictive compensation amount, so as to obtain the integrated control command for the pressure roller temperature. The compensation strategy stability determination module is used to determine the applicability of the feedforward compensation strategy based on the execution effect of the integrated control command and the topological consistency of the dynamic causal structure graph, and to obtain the reconstruction trigger signal of the feedforward compensation strategy.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the method according to any one of claims 1 to 8.

Images