Cigarette production quality optimization method and system based on multi-factor correlation analysis

Abstract

This application provides a method and system for optimizing cigarette production quality based on multi-factor correlation analysis, relating to the field of production management technology. This application obtains parameter sequences in micro-time periods by performing time-series slicing processing on multi-dimensional production data, capturing short-term parameter fluctuation characteristics. A constructed dynamic coupling correlation matrix characterizes the correlation strength between parameters, and the correlation threshold is dynamically adjusted using an environmental fitness factor, allowing the matrix to update in real time with changes in the production environment. The dynamic deviation is calculated within a sliding time window and compared with the dynamic threshold. Combined with the dynamic coupling correlation matrix, strong coupling relationships are determined, enabling accurate identification of the actual correlation between current parameters when production status continuously changes, and rapid location of key factors causing quality fluctuations. Based on the coupling relationship description information, targeted process parameter adjustment suggestions are generated, achieving rapid response. Therefore, this application can quickly and accurately achieve production status analysis, thereby improving the management effect of cigarette production quality.

Description

Technical Field

[0001] This application relates to the field of production management technology, specifically to a method and system for optimizing cigarette production quality based on multi-factor correlation analysis. Background Technology

[0002] Cigarette production is a complex manufacturing process involving multiple steps such as tobacco processing, rolling, and packaging. Product quality is influenced by a multitude of process parameters, including temperature, moisture content, and filler content. During production, quality indicators such as cigarette draw resistance, weight, end-feed weight, end content, and empty-end rejection rate must be strictly controlled within standard ranges to ensure product stability and consistency. With the continuous improvement of intelligent manufacturing in the cigarette industry, enhancing the accuracy and timeliness of production quality management has become a key focus for the industry.

[0003] In existing technologies, common production quality management methods mainly employ Statistical Process Control (SPC) or offline analysis techniques based on historical data. These methods analyze production data by establishing statistical correlation models between process parameters and quality indicators, or by using mathematical tools such as grey relational analysis and response surface methodology, to identify key factors affecting quality. The basic principle is to treat each process parameter as an analytical object, calculate its correlation with the target quality indicator, and guide process adjustments based on the analysis results.

[0004] However, the cigarette production process involves numerous complex interactions between parameters. For example, in actual production, changes in moisture content often lead to simultaneous fluctuations in cigarette weight, filler content, and draw resistance; changes in temperature and humidity also cause correlated changes in multiple quality indicators. Existing technologies typically establish relatively fixed correlation models based on historical data, where the weights and correlation coefficients of each parameter remain relatively stable after establishment. However, environmental conditions and equipment status in actual production sites are constantly changing, causing the interactions between parameters to change accordingly. When the production site conditions change, the original fixed model needs to re-collect data and re-analyze calculations to adapt to the new production conditions. This process requires a certain time period, and the fixed model may not be able to fully and accurately reflect the actual correlations between parameters at the current moment in a dynamically changing production environment. Therefore, existing technologies are not fast and accurate enough in handling such multi-parameter interactive production scenarios, ultimately resulting in poor production quality management. Summary of the Invention

[0005] This application provides a method and system for optimizing cigarette production quality based on multi-factor correlation analysis, which can quickly and accurately analyze production status, thereby improving the management effect of cigarette production quality.

[0006] Firstly, this application provides a method for optimizing cigarette production quality based on multi-factor correlation analysis, the method comprising: Multidimensional production data from the cigarette production line is collected in real time and processed by time-series slicing to obtain parameter sequences for multiple micro-time periods. The multidimensional production data includes process quality parameters, equipment operating status parameters, and environmental parameters. A dynamic coupling correlation matrix is ​​constructed based on the parameter sequence, and an environmental fitness factor is calculated based on the real-time fluctuation of environmental parameters in multidimensional production data. The environmental fitness factor is used to dynamically adjust the correlation threshold between each parameter pair in the dynamic coupling correlation matrix. The dynamic coupling correlation matrix characterizes the correlation strength between each process quality parameter and equipment operating status parameter in multidimensional production data. Within a preset sliding time window, the dynamic deviation of each process quality parameter in the multidimensional production data from the target value is calculated based on the dynamic coupling correlation matrix. The dynamic deviation is compared with a dynamic threshold that changes with production conditions. When the dynamic deviation of the first parameter continuously exceeds the dynamic threshold, and the correlation between the first parameter and the second parameter is determined to exceed the correlation threshold based on the dynamic coupling correlation matrix, a strong coupling relationship is determined between the first parameter and the second parameter, and the coupling relationship description information between the first parameter and the second parameter is output. Here, the first parameter is the parameter whose dynamic deviation exceeds the standard among the process quality parameters, and the second parameter is the process quality parameter or equipment operating status parameter that is correlated with the first parameter in the multidimensional production data. Based on the coupling relationship description information, process parameter adjustment suggestions for the first parameter and / or the second parameter are generated, and the adjustment suggestions are sent to the control system of the cigarette production line.

[0007] By employing the aforementioned technical solution, time-series slicing is performed on real-time acquired multidimensional production data to obtain parameter sequences within micro-time periods, enabling precise capture of short-term parameter fluctuations during the production process. Based on this, a dynamic coupling correlation matrix is ​​constructed to characterize the correlation strength between various process quality parameters and equipment operating status parameters. By dynamically adjusting the correlation threshold using an environmental adaptability factor, the matrix can be updated in real-time according to changes in the production environment, thus adapting to dynamically changing production conditions. Furthermore, the dynamic deviation of each process quality parameter is calculated within a sliding time window and compared with the dynamic threshold that changes with production conditions. Combined with the dynamic coupling correlation matrix to determine strong coupling relationships, the actual correlation between parameters at the current moment can be accurately identified even when the production status is continuously changing, quickly pinpointing key factors causing quality fluctuations. Based on the accurately identified coupling relationship description information, targeted process parameter adjustment suggestions are generated, enabling rapid response to production anomalies. Therefore, this application can quickly and accurately achieve production status analysis, thereby improving the management effectiveness of cigarette production quality.

[0008] Secondly, this application provides a cigarette production quality optimization system based on multi-factor correlation analysis, the system comprising: The data acquisition module is used to collect multi-dimensional production data in the cigarette production line in real time, and to perform time-series slicing processing on the multi-dimensional production data to obtain parameter sequences of multiple micro time periods. The multi-dimensional production data includes process quality parameters, equipment operating status parameters and environmental parameters. The association matrix construction module is used to construct a dynamic coupling association matrix based on the parameter sequence, and calculate the environmental fitness factor according to the real-time fluctuation of environmental parameters in multidimensional production data. The environmental fitness factor is used to dynamically adjust the association threshold between each parameter pair in the dynamic coupling association matrix. The deviation calculation module is used to calculate the dynamic deviation of each process quality parameter in multidimensional production data relative to the target value based on the dynamic coupling correlation matrix within a preset sliding time window. The coupling determination module is used to compare the dynamic deviation with the dynamic threshold that changes with the production conditions. When the dynamic deviation of the first parameter continues to exceed the dynamic threshold, and the correlation between the first parameter and the second parameter is determined to exceed the correlation threshold based on the dynamic coupling correlation matrix, it is determined that there is a strong coupling relationship between the first parameter and the second parameter, and the coupling relationship description information between the first parameter and the second parameter is output. The adjustment suggestion generation module is used to generate process parameter adjustment suggestions for the first parameter and / or the second parameter based on the coupling relationship description information, and send the adjustment suggestions to the control system of the cigarette production line.

[0009] Thirdly, this application provides a computer storage medium that stores multiple instructions adapted for loading by a processor and executing any of the methods described above.

[0010] Fourthly, this application provides an electronic device including a processor, a memory, and a transceiver. The memory is used to store instructions, the transceiver is used to communicate with other devices, and the processor is used to execute the instructions stored in the memory to cause the electronic device to perform any of the methods described above.

[0011] In summary, the beneficial effects of the technical solution of this application include: By employing the above technical solution, time-series slicing is performed on real-time acquired multidimensional production data to obtain parameter sequences for micro-time periods, enabling precise capture of short-term parameter fluctuation characteristics during the production process. The constructed dynamic coupling correlation matrix characterizes the correlation strength between various process quality parameters and equipment operating status parameters. An environmental adaptability factor is used to dynamically adjust the correlation threshold, allowing the matrix to update in real-time with changes in the production environment, thus adapting to dynamically changing production conditions. By calculating the dynamic deviation of each process quality parameter within a sliding time window and comparing it with the dynamic threshold that changes with production conditions, combined with the dynamic coupling correlation matrix to determine strong coupling relationships, the actual correlation between parameters at the current moment can be accurately identified even when the production site conditions are continuously changing, quickly pinpointing key factors of quality fluctuations. Based on the accurately identified coupling relationship description information, targeted process parameter adjustment suggestions are generated, enabling rapid response to production anomalies. Therefore, this application can quickly and accurately achieve production status analysis, thereby improving the management effectiveness of cigarette production quality. Attached Figure Description

[0012] Figure 1 This is a flowchart illustrating a method for optimizing cigarette production quality based on multi-factor correlation analysis, according to an embodiment of this application. Figure 2 This is a schematic diagram of the architecture of a multi-factor association analysis system provided in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of a cigarette production quality optimization system based on multi-factor correlation analysis according to an embodiment of this application; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0013] Explanation of reference numerals in the attached figures: 400, electronic device; 401, processor; 402, communication bus; 403, user interface; 404, network interface; 405, memory. Detailed Implementation

[0014] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0015] In the description of the embodiments of this application, words such as "illustrative," "for example," or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "illustrative," "for example," or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design solutions. Rather, the use of words such as "illustrative," "for example," or "for example" is intended to present the relevant concepts in a specific manner.

[0016] In the description of the embodiments of this application, the term "multiple" means two or more. For example, multiple systems means two or more systems, and multiple screen terminals means two or more screen terminals. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. The terms "comprising," "including," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0017] Please see Figure 1 This document presents a flowchart illustrating a method for optimizing cigarette production quality based on multi-factor correlation analysis, as provided in an embodiment of this application. This method can be implemented using a computer program, a microcontroller, or run on a cigarette production quality optimization system based on the von Neumann architecture and multi-factor correlation analysis. The computer program can be integrated into the application or run as a standalone utility application. The specific steps of the cigarette production quality optimization method based on multi-factor correlation analysis are described in detail below.

[0018] S101: Real-time acquisition of multi-dimensional production data in the cigarette production line, and time-series slicing of the multi-dimensional production data to obtain parameter sequences of multiple micro-time periods. The multi-dimensional production data includes process quality parameters, equipment operating status parameters and environmental parameters. Among them, multidimensional production data represents comprehensive production information collected from different dimensions during the cigarette production process, and is a data set describing the production status; process quality parameters refer to parameters that directly reflect the quality characteristics of cigarette products, such as cigarette weight, length, circumference, draw resistance, and end-burning amount, which determine whether the final product meets quality standards; equipment operating status parameters are used to represent the real-time working status of production equipment, including indicators reflecting the health and operational stability of equipment such as motor load, speed, vibration frequency, and temperature; environmental parameters refer to environmental condition data of the production workshop, mainly including temperature and humidity, which have a significant impact on the physical properties and processing performance of tobacco materials; time-series slicing processing means dividing the continuously collected data stream into multiple discrete time segments according to the time dimension, with each segment containing data samples within a specific duration; micro-time periods refer to data windows with short time spans formed after slicing processing; parameter sequences are used to represent a set of parameter values ​​arranged in an ordered manner in the time dimension, reflecting the evolution trajectory of parameters over time.

[0019] Specifically, the system collects multi-dimensional production data in real time through a sensor network deployed at key workstations on the production line. This data includes three dimensions: process quality parameters, equipment operating status parameters, and environmental parameters. After preprocessing, the collected raw data undergoes time-series slicing. The system divides the continuous data stream into multiple micro-segments according to a preset time length, with each micro-segment containing all parameter values ​​collected within that segment. Slicing is performed using a sliding method, meaning there is partial overlap between adjacent time segments, which prevents important information from being missed at slice boundaries. After processing, the system organizes the parameter values ​​within each micro-segment into parameter sequences in chronological order. These sequences provide a structured time-series data foundation for subsequent data analysis.

[0020] In some embodiments, real-time data acquisition and time-series slicing processing can be achieved in several ways: Optionally, a distributed acquisition architecture can be adopted, deploying data acquisition devices in different sections of the production line. Each acquisition device independently completes local data acquisition and preliminary processing, and then the data is aggregated to a central processing server. The server performs time alignment on the data from different sections to ensure that data from different sources can be sliced ​​based on a unified time benchmark. Finally, the aligned data is sliced ​​according to a preset micro-time period length to generate a complete parameter sequence. Optionally, a streaming processing method can be adopted, where data acquired by sensors is continuously input into the processing system in the form of a data stream. The system sets a fixed-length sliding window to slice the data stream in real time. Each time the window slides forward by one step, a new micro-time period parameter sequence is generated. This method enables parallel data acquisition and slicing processing, significantly reducing processing latency and allowing the system to respond quickly to changes in production status. It is understood that other methods can also be used to achieve real-time acquisition and time-series slicing processing of multi-dimensional production data, which are not limited here.

[0021] S102: Construct a dynamic coupling correlation matrix based on parameter sequences, and calculate the environmental fitness factor according to the real-time fluctuation of environmental parameters in multidimensional production data. Use the environmental fitness factor to dynamically adjust the correlation threshold between each parameter pair in the dynamic coupling correlation matrix. The dynamic coupling correlation matrix characterizes the correlation strength between each process quality parameter and equipment operating status parameter in multidimensional production data. The dynamic coupling correlation matrix is ​​a mathematical matrix used to describe the correlation strength between parameters in multidimensional production data. Each element in the matrix represents the degree of coupling between two parameters. The correlation strength is used to represent the degree of mutual influence between process quality parameters and equipment operating status parameters. The larger the value, the tighter the correlation between the two parameters. The environmental adaptability factor is an adjustment coefficient calculated based on the real-time fluctuation of environmental parameters. It is used to dynamically correct the correlation threshold in the correlation matrix. The correlation threshold is a critical value for judging whether there is a significant correlation between two parameters. When the correlation strength of the parameter pair exceeds the threshold, it is considered that there is a strong correlation between the two.

[0022] Specifically, the system first constructs a dynamic coupling correlation matrix based on the parameter sequence. By analyzing the changes in various process quality parameters and equipment operating status parameters over time, it calculates the correlation strength between them and fills the corresponding positions in the matrix with the calculation results. After the matrix is ​​constructed, the system monitors the fluctuations of environmental parameters in real time, extracts the current values ​​of environmental parameters such as temperature and humidity, and calculates the fluctuation range of these parameters relative to the baseline values. Based on the real-time fluctuations of environmental parameters, the system calculates an environmental fitness factor, which reflects the degree of influence of the current environmental conditions on the production process. The system uses the calculated environmental fitness factor to adjust the correlation thresholds in the dynamic coupling correlation matrix. When environmental fluctuations are large, the system adjusts the criteria for judging the significance of parameter correlations accordingly, so that the correlation thresholds can adapt to the current environmental state, thereby more accurately reflecting the actual coupling relationship between parameters.

[0023] In some embodiments, the construction of the dynamic coupling correlation matrix and the dynamic adjustment of the correlation threshold can be achieved in several ways: Optionally, a statistical correlation analysis method is used to construct the correlation matrix. Pairwise correlation calculations are performed on each process quality parameter and equipment operating status parameter in the parameter sequence to obtain the correlation coefficient between each parameter pair as the correlation strength. The correlation strengths of all parameter pairs are organized into a matrix to form an initial dynamic coupling correlation matrix. Then, environmental parameter data is collected in real time, and the deviation between the current environmental parameter and the historical average is calculated. The deviation value is normalized and used as an environmental fitness factor. Finally, the correlation is modified according to the magnitude of the environmental fitness factor and a preset adjustment rule. Thresholds can be set, for example, by appropriately increasing the association threshold when the environmental fitness factor increases. Optionally, a time-varying weighting method can be used, which not only calculates the association strength between parameters but also assigns a time decay weight to each association strength, ensuring that recent data contributes more to the association strength than historical data. This constructs a time-sensitive dynamic coupling association matrix. The environmental fitness factor is calculated using fuzzy logic, with temperature and humidity fluctuations as fuzzy input variables. Fuzzy rule reasoning yields the fuzzy output of the environmental fitness factor, which is then defuzzified to obtain a precise factor value. This factor is used to proportionally adjust the association thresholds of each parameter pair in the matrix, achieving adaptive adjustment of the association judgment criteria. It is understood that other methods can also be used to construct the dynamic coupling association matrix and intelligently adjust the association thresholds; this is not limited here.

[0024] S103: Within a preset sliding time window, based on the dynamic coupling correlation matrix, calculate the dynamic deviation of each process quality parameter in the multidimensional production data relative to the target value. Among them, the sliding time window represents a fixed-length time interval used for analysis, which slides forward continuously over time; dynamic deviation refers to the degree of deviation of the process quality parameter from the target value, which comprehensively considers the parameter's deviation and trend of change; the target value is used to represent the ideal value that the process quality parameter should achieve; the deviation term refers to the difference between the current value of the parameter and the target value; the rate of change term represents the rate at which the parameter value changes over time; and the weighting coefficient refers to the weight of the deviation term and the rate of change term in the dynamic deviation calculation.

[0025] Specifically, within a preset sliding time window, the system calculates the dynamic deviation of each process quality parameter based on a dynamic coupling correlation matrix. During the calculation, the system obtains the current value of each process quality parameter and compares it with the corresponding target value to obtain the deviation term. Simultaneously, the system calculates the rate of change of the parameter value to obtain the rate of change term. The calculation of dynamic deviation combines the deviation term and the rate of change term, comprehensively evaluating the degree of parameter deviation by assigning weighting coefficients to each. The system dynamically adjusts the weighting coefficients of the deviation term and the rate of change term according to the current production time, allowing the calculation of dynamic deviation to adapt to the characteristics of different production stages. For example, during the production start-up phase, more attention may be paid to the rate of change, while during the stable production phase, more attention may be paid to the deviation value.

[0026] In some embodiments, dynamic deviation can be calculated in several ways: Optionally, a linear weighting method is used, where the deviation term and the rate of change term are multiplied by their respective weighting coefficients and then summed to obtain the dynamic deviation. The deviation term is calculated using the difference between the current parameter value and the target value, and the rate of change term is calculated by dividing the difference between the current value and the previous value by the time interval. The weighting coefficients are obtained by querying a preset coefficient table based on the production time. Different production stages correspond to different coefficient combinations, and the system selects the appropriate weighting coefficients for calculation based on the current production stage. Optionally, an adaptive weighting method is used, where the system monitors the production status in real time and dynamically determines the weighting coefficients based on the parameter fluctuation characteristics. When the parameter fluctuation is large, the weight of the rate of change term is increased to more sensitively capture the dynamic changes of the parameter. When the parameter is relatively stable, the weight of the deviation term is increased to more accurately assess the deviation from the target value. The adjustment of the weighting coefficients uses a continuous function to achieve a smooth transition, avoiding abrupt changes in the coefficients that could cause jumps in the deviation calculation, ultimately obtaining a dynamic deviation that flexibly reflects the parameter deviation. It is understood that other methods can also be used to accurately calculate the dynamic deviation, which are not limited here.

[0027] S104: Compare the dynamic deviation with the dynamic threshold that changes with production conditions. When the dynamic deviation of the first parameter continuously exceeds the dynamic threshold, and the correlation between the first parameter and the second parameter is determined to exceed the correlation threshold based on the dynamic coupling correlation matrix, it is determined that there is a strong coupling relationship between the first parameter and the second parameter, and the coupling relationship description information between the first parameter and the second parameter is output. Here, the first parameter is the parameter whose dynamic deviation exceeds the standard among the process quality parameters, and the second parameter is the process quality parameter or equipment operating status parameter that is correlated with the first parameter in the multidimensional production data. Among them, the dynamic threshold represents the critical value used to determine whether the parameter deviation exceeds the standard, and this threshold is adjusted according to the production conditions; the first parameter refers to the process quality parameter whose dynamic deviation exceeds the dynamic threshold; continuous exceedance indicates that the dynamic deviation of the parameter remains in an excessive state for a period of time; correlation degree refers to the correlation strength value between two parameters in the dynamic coupling correlation matrix; the second parameter is used to represent the process quality parameter or equipment operating status parameter that is correlated with the first parameter; strong coupling relationship indicates that there is a significant mutual influence between two parameters; coupling relationship description information refers to a detailed description of the strong coupling relationship.

[0028] This step is performed after calculating the dynamic deviation of each parameter to identify parameter pairs with strong coupling relationships. Specifically, the system compares the calculated dynamic deviation with the corresponding dynamic threshold, which varies depending on the current production conditions. When the dynamic deviation of a process quality parameter exceeds its corresponding dynamic threshold, and this exceeding state persists for a certain period, the system identifies this parameter as the first parameter. Subsequently, the system queries the dynamic coupling correlation matrix to find all parameters related to the first parameter and calculates the correlation degree between the first parameter and these parameters. When the correlation degree between the first parameter and a certain parameter exceeds the correlation threshold, the system determines that there is a strong coupling relationship between these two parameters, and this parameter is identified as the second parameter. The system outputs a description of the coupling relationship between the first and second parameters, illustrating the coupling characteristics between the two parameters.

[0029] In some embodiments, the determination of strong coupling relationships can be achieved in several ways: Optionally, a fixed-duration continuous judgment method is adopted, setting a fixed time length as the judgment standard for continuous exceedance. When the dynamic deviation of the first parameter always exceeds the dynamic threshold within the time length, the correlation check is triggered. The system traverses all the associated parameters corresponding to the first parameter in the dynamic coupling correlation matrix, compares the correlation degree between each associated parameter and the first parameter with the correlation threshold, finds all parameters whose correlation degree exceeds the threshold as the second parameter, and generates coupling relationship description information for each pair of first and second parameters. The description information includes parameter name, correlation degree value, exceedance duration, etc. Optionally, a counting judgment method is adopted, counting the number of times the dynamic deviation of the first parameter exceeds the standard within the sliding time window. When the number of exceedances exceeds a preset threshold, it is determined to be a continuous exceedance. Then, each parameter associated with the first parameter is checked. Not only is the correlation degree compared with the correlation threshold, but the state change of the second parameter during the exceedance of the first parameter is also analyzed. If the second parameter also shows abnormal fluctuations, the strong coupling judgment of the two is strengthened. The generated coupling relationship description information includes a comparison of the dynamic change trends of the parameters of both parties. It is understandable that other methods can be used to accurately identify strong coupling relationships, and no specific method is specified here.

[0030] S105: Based on the coupling relationship description information, generate process parameter adjustment suggestions for the first parameter and / or the second parameter, and send the adjustment suggestions to the control system of the cigarette production line.

[0031] Among them, the process parameter adjustment suggestion refers to the parameter adjustment scheme proposed by the system based on the coupling relationship analysis results; the control system refers to the automated control equipment of the cigarette production line.

[0032] This step is executed after identifying strong coupling relationships and generating descriptive information, and is used to translate the analysis results into specific adjustment measures. Specifically, based on the descriptive information of the coupling relationship between the first and second parameters, the system generates process parameter adjustment suggestions for these parameters. Adjustment suggestions can be made for the first parameter, the second parameter, or a coordinated adjustment scheme for both parameters simultaneously. The system determines the appropriate adjustment direction and magnitude based on the current deviation of the parameters and the coupling relationship between them. After generating the adjustment suggestions, the system sends them to the control system of the cigarette production line via a communication interface. Upon receiving the adjustment suggestions, the control system can execute the corresponding parameter adjustment operations.

[0033] In some embodiments, the generation and transmission of adjustment suggestions can be achieved in several ways: Optionally, a rule-based matching method can be used. The system pre-establishes a rule base for process parameter adjustment, which stores adjustment strategies corresponding to different coupling relationship patterns. The system matches the current coupling relationship description information with the patterns in the rule base, finds the most similar rule, and generates specific adjustment suggestions based on the guidance in the rule. The adjustment suggestions are sent to the control system in the form of structured data via an industrial communication protocol. The control system parses the adjustment suggestions and executes parameter modifications. Optionally, an optimization algorithm can be used to calculate the adjustment scheme. The system establishes an optimization objective function based on the current state and target state of the first and second parameters, as well as the coupling relationship between them. A numerical optimization algorithm is used to solve for the optimal parameter adjustment amount, and the optimization result is used as the adjustment suggestion. The adjustment suggestion includes not only the target setpoint of the parameter but also the adjustment priority and the suggested execution time. The system asynchronously sends the adjustment suggestion to the control system through a message queue to ensure the reliability of communication. It is understood that other methods can also be used to achieve intelligent generation and reliable transmission of adjustment suggestions, which are not limited here.

[0034] Please see Figure 2 This is a schematic diagram of the architecture of a multi-factor correlation analysis system provided in this application embodiment. The system adopts a three-layer architecture design, including a data acquisition layer, a dynamic analysis engine layer, and an application layer. The data acquisition layer, as the perception layer, collects real-time parameters such as cigarette draw resistance, weight, end-felling, moisture content, temperature, moisture content, filling value, as well as the empty head rejection rate of the cigarette machine and the moisture content of the finished product through data acquisition units deployed on equipment such as cigarette machines and drying machines. The dynamic analysis engine layer, as the core layer, includes a preprocessing module to clean and standardize the collected data, a coupling calculation module to execute a dynamic weight coupling algorithm to calculate the real-time correlation between various parameters, and a decision generation module to generate process optimization suggestions based on the calculation results. The application layer, as the service layer, provides users with correlation analysis graph display and early warning information display functions through a platform visualization interface. This architecture realizes a complete data flow and analysis processing process from the bottom-level data acquisition to the upper-level application display, supporting dynamic monitoring and intelligent decision-making of multi-parameter coupling relationships in the production process.

[0035] Based on the above embodiments, as an optional implementation method, in S102, the environmental fitness factor is calculated based on the real-time fluctuation of environmental parameters in multidimensional production data, and the correlation threshold between each parameter pair in the dynamic coupling correlation matrix is ​​dynamically adjusted using the environmental fitness factor. Specifically, this can be achieved through the following steps S201-S204.

[0036] S201: Within each sliding time window, acquire the temperature and humidity parameters of the current production workshop as environmental parameters, and calculate the fluctuation range of the temperature and humidity parameters relative to their respective preset reference values ​​within the current sliding time window. Among them, the preset reference value refers to the standard reference value corresponding to the temperature and humidity parameters under normal production conditions; the fluctuation range represents the deviation between the actual measured value of the environmental parameter and the preset reference value within the current sliding time window.

[0037] Specifically, at the start of the sliding time window, the system acquires real-time readings from the temperature and humidity sensors in the production workshop. These readings represent the temperature and humidity parameters corresponding to the current sliding time window. After acquiring the environmental parameters, the system extracts the preset baseline values ​​for the temperature and humidity parameters, respectively. These baseline values ​​are typically pre-set according to the requirements of the cigarette production process, representing the most suitable production environment conditions. The system calculates the difference between the actual measured value of the temperature parameter and the preset baseline value. The absolute value of this difference is the fluctuation range of the temperature parameter. The same method is used to calculate the fluctuation range of the humidity parameter. The fluctuation range is calculated by comparing the average or instantaneous value of the environmental parameters within the current sliding time window with the baseline value. The resulting value directly reflects the degree to which the environmental parameters deviate from the ideal state.

[0038] S202: Compare the fluctuation amplitude with the preset environmental fluctuation threshold. When the fluctuation amplitude of the temperature parameter and / or humidity parameter exceeds the corresponding environmental fluctuation threshold, calculate the environmental adaptability factor based on the degree of deviation between the fluctuation amplitude exceeding the environmental fluctuation threshold and the corresponding preset benchmark value. Among them, the environmental fluctuation threshold refers to the pre-set critical value used to judge whether the fluctuation of environmental parameters is abnormal. Temperature parameters and humidity parameters correspond to their respective environmental fluctuation thresholds; the degree of deviation is used to quantify the specific numerical magnitude of the fluctuation exceeding the environmental fluctuation threshold.

[0039] Specifically, the system compares the calculated temperature fluctuation amplitude with the corresponding environmental fluctuation threshold, and simultaneously compares the humidity fluctuation amplitude with the corresponding environmental fluctuation threshold. When the temperature fluctuation amplitude exceeds its environmental fluctuation threshold, or the humidity fluctuation amplitude exceeds its environmental fluctuation threshold, or both exceed their respective thresholds, the system determines that there is significant environmental fluctuation. At this point, the system extracts the fluctuation amplitude exceeding the threshold and calculates the ratio or difference between this value and the corresponding preset benchmark value to obtain a quantitative index of the deviation. Based on the magnitude of the deviation, the system calculates an environmental fitness factor according to a preset mapping relationship. The greater the deviation, the larger the calculated environmental fitness factor value. This factor typically ranges from zero to one or uses other normalized intervals for subsequent quantitative adjustment of the associated thresholds.

[0040] S203: Based on the environmental fitness factor, determine the target parameter pairs affected by the fluctuating environmental parameters in the dynamic coupling correlation matrix, reduce the static correlation weight of the target parameter pairs determined based on historical statistical data, and correspondingly increase the dynamic monitoring weight of the target parameter pairs determined based on real-time data within the current sliding time window, thereby updating the correlation threshold between the target parameter pairs.

[0041] Among them, the target parameter pair refers to the parameter combination in the dynamic coupling correlation matrix that is affected by the fluctuation of environmental parameters; the static correlation weight represents the correlation strength weight value between parameter pairs obtained based on historical statistical data analysis; and the dynamic monitoring weight represents the correlation strength weight value between parameter pairs calculated based on real-time data within the current sliding time window.

[0042] Specifically, based on the environmental fitness factor, the system searches for target parameter pairs affected by fluctuating environmental parameters in the dynamic coupling correlation matrix. These target parameter pairs are typically combinations of process quality parameters and equipment operating status parameters that have been found to be correlated with temperature or humidity in historical data analysis. The system extracts the current static correlation weight value of the target parameter pair, which reflects the long-term correlation pattern statistically derived from historical data. Then, it reduces the static correlation weight according to the magnitude of the environmental fitness factor, with the reduction proportional to the environmental fitness factor. Simultaneously, the system extracts the dynamic monitoring weight of the target parameter pair, calculated based on real-time data within the current sliding time window. This weight reflects the real-time correlation status of the parameter pair under current environmental conditions, and the system increases the dynamic monitoring weight proportionally to the reduction in the static correlation weight. By reducing the static correlation weight and increasing the dynamic monitoring weight, the system recalculates the comprehensive correlation strength of the target parameter pair and updates the corresponding correlation threshold based on the adjusted comprehensive correlation strength, enabling the correlation threshold to more accurately reflect the actual correlation between parameters under current environmental conditions.

[0043] Based on the above embodiments, as an optional implementation method, the calculation of dynamic deviation includes a deviation term between the current value of the parameter and the target value, and a parameter change rate term. The weighting coefficients of the deviation term and the change rate term are dynamically adjusted with the production time. In step S103, within a preset sliding time window, the dynamic deviation of each process quality parameter in the multidimensional production data relative to the target value is calculated based on the dynamic coupling correlation matrix. This can be specifically achieved through the following steps S301-S303.

[0044] S301: For each process quality parameter i, calculate the dynamic deviation Di(t) of process quality parameter i at time t according to the formula, where X_i(t) is the current sampled value of process quality parameter i at time t, X_target(t) is the target value corresponding to process quality parameter i, ΔX_i(t) is the rate of change of process quality parameter i at time t, α is the weighting coefficient of the deviation term, and β is the weighting coefficient of the rate of change term; The current sampled value refers to the actual measured value of the process quality parameter collected by the sensor at a specific moment; the rate of change represents the speed at which the process quality parameter changes between adjacent moments, and is obtained by dividing the difference between the current sampled value and the previous sampled value by the time interval.

[0045] The first part of the formula is the squared difference between the target value and the current sampled value. Specifically, it is the target value minus the current sampled value, squared, and then multiplied by a deviation weighting coefficient. The target value represents the ideal value that the process quality parameter should achieve, the current sampled value represents the actual measured value of the parameter at a specific moment, and the deviation weighting coefficient is used to adjust the weighting percentage of this part in the overall calculation. The second part is the rate of change term, representing the speed of change of the process quality parameter, multiplied by a rate of change weighting coefficient. The rate of change is obtained by calculating the difference between the current sampled value and the sampled value at the previous moment, and the rate of change weighting coefficient is used to adjust the influence of the rate of change on the overall calculation.

[0046] The purpose of using this formula in this application is to achieve a comprehensive and dynamic assessment of process quality parameters. Traditional quality monitoring only focuses on the deviation of parameters from target values, failing to capture parameter change trends in a timely manner, leading to a delayed response to quality anomalies. This formula, by simultaneously considering the magnitude of deviation and the rate of change, achieves a proactive assessment of parameter status. When parameters, although not yet significantly deviating from the target, change rapidly, the system can identify potential risks in advance. The introduction of two weighting coefficients allows the formula to flexibly adjust the weight ratio of the deviation and rate of change terms according to different production stages and process requirements. In stages with high stability requirements, the weight of the rate of change term is increased; in stages with high accuracy requirements, the weight of the deviation term is increased, thereby improving the adaptability and accuracy of quality monitoring. Specifically, the system acquires the sampled value of the process quality parameter at the current moment and the corresponding target value, calculates the difference between the two, squares the difference to obtain the base value of the deviation term, and then multiplies it by the weighting coefficient of the deviation term to obtain the weighted result of the deviation term. The system simultaneously acquires the sampled values ​​of the process quality parameter at the current moment and the previous moment, calculates the difference between the two sampled values, divides the difference by the time interval between the two samplings to obtain the rate of change, and multiplies the rate of change by the weighting coefficient of the rate of change term to obtain the weighted result of the rate of change term. The system sums the weighted result of the deviation term and the weighted result of the rate of change term to obtain the dynamic deviation of the process quality parameter at the current moment. The weighting coefficients of the deviation term and the rate of change term are obtained from a preset coefficient configuration table based on the current production time. Different production stages correspond to different combinations of weighting coefficients, realizing adaptive adjustment of the dynamic deviation calculation according to production time.

[0047] Based on the above embodiments, as an optional implementation method, step S301, specifically regarding the values ​​of the two weighting coefficients, further includes steps S3011-S3013.

[0048] S3011: At the beginning of each sliding time window, the data fluctuation characteristic values ​​of each process quality parameter within the current sliding time window are statistically analyzed. The data fluctuation characteristic values ​​include the fluctuation frequency and fluctuation amplitude of each process quality parameter within the time span of the current sliding time window. The fluctuation frequency is the number of times the parameter value crosses its mean line within the current sliding time window, and the fluctuation amplitude is the ratio of the range of the parameter value within the current sliding time window to the target value. Among them, the data fluctuation characteristic value refers to the statistical index that characterizes the fluctuation state of the process quality parameter within the sliding time window; the fluctuation frequency represents the number of times the parameter value crosses its average level, which is obtained by counting the number of times the parameter value crosses the mean line within the sliding time window. The mean line is the arithmetic mean of all sampling points of the parameter value within the time window; the fluctuation amplitude represents the proportion of the parameter value's range of variation relative to the target value, which is obtained by calculating the difference between the maximum and minimum values ​​of the parameter value within the sliding time window, i.e., the range, and then dividing the range by the target value.

[0049] Specifically, the system extracts the process quality parameter values ​​at all sampling moments within the current sliding time window and calculates the arithmetic mean of these values ​​as the mean line. The system checks the parameter values ​​of adjacent sampling points one by one, determining whether the parameter value of the previous sampling point is on one side of the mean line and the parameter value of the subsequent sampling point is on the other side. Each such crossover is counted once, and the fluctuation frequency value is obtained after counting all sampling points. Simultaneously, the system iterates through the parameter values ​​of all sampling points within the sliding time window, identifies the maximum and minimum values, calculates the difference between them to obtain the range, extracts the target value corresponding to the parameter, and divides the range by the target value to obtain the fluctuation amplitude value. The system performs the above statistical calculations for each process quality parameter to obtain the fluctuation frequency and fluctuation amplitude of each parameter, forming a complete set of data fluctuation characteristic values.

[0050] S3012: Determine the production condition type of the current production stage based on the data fluctuation characteristic value. When the fluctuation frequency of the process quality parameter is higher than the preset frequency threshold, the current production condition is determined to be a rapid fluctuation condition; when the fluctuation amplitude of the process quality parameter is greater than the preset amplitude threshold and the fluctuation frequency is lower than the frequency threshold, the current production condition is determined to be a slow drift condition. Among them, the production condition type refers to the production state category classified according to the fluctuation characteristics of process quality parameters; the rapid fluctuation condition indicates that the parameter frequently crosses the mean line but the deviation is relatively small; the slow drift condition indicates that the parameter continuously deviates from the target value and the deviation is large but the crossing frequency is low; the frequency threshold and amplitude threshold are the critical values ​​of fluctuation frequency and fluctuation amplitude used to determine the production condition type, respectively.

[0051] This step is executed after obtaining the data fluctuation characteristic values ​​to identify the current production stage's operating condition. Specifically, the system extracts pre-set frequency and amplitude threshold values ​​and compares the fluctuation frequency of each process quality parameter with these thresholds. When the fluctuation frequency of a process quality parameter is higher than the frequency threshold, the system determines that the parameter is in a rapid fluctuation state, indicating that the parameter value oscillates frequently around the mean line, and marks the current production condition as a rapid fluctuation condition. When the fluctuation frequency of a process quality parameter is lower than the frequency threshold, the system further compares the fluctuation amplitude of the parameter with the amplitude threshold. If the fluctuation amplitude is greater than the amplitude threshold, the system determines that the parameter is in a slow drift state, indicating that the parameter value deviates from the target value for a long time and the deviation is large, and marks the current production condition as a slow drift condition. The system independently performs operating condition type determination for each process quality parameter, determining the corresponding production operating condition type for each parameter.

[0052] S3013: Adaptively adjust the values ​​of the weighting coefficients α and β of the corresponding process quality parameters according to the determined production condition type: Under rapid fluctuation conditions, increase the value of the weighting coefficient β of the corresponding process quality parameter and correspondingly decrease the value of the weighting coefficient α, so that the calculation of dynamic deviation focuses on capturing the trend characteristics of the parameter change rate; under slow drift conditions, increase the value of the weighting coefficient α of the corresponding process quality parameter and correspondingly decrease the value of the weighting coefficient β, so that the calculation of dynamic deviation focuses on evaluating the cumulative effect of the parameter deviation amplitude.

[0053] Specifically, when the system identifies a process quality parameter as being in a rapidly fluctuating condition, it extracts the current weighted coefficients for the deviation and rate of change terms. The system increases the weighted coefficient for the rate of change term, with the increase determined by how much the fluctuation frequency exceeds a frequency threshold. Simultaneously, it decreases the weighted coefficient for the deviation term, maintaining a constant sum of the two coefficients or reducing it by a fixed proportion. This adjusted weighting configuration makes the dynamic deviation calculation more focused on the parameter's rate of change and fluctuation trend, improving sensitivity to rapid oscillations and enabling the system to respond promptly to frequent parameter changes. When the system identifies a process quality parameter as being in a slowly drifting condition, it increases the weighted coefficient for the deviation term, with the increase determined by how much the fluctuation amplitude exceeds an amplitude threshold. Simultaneously, it correspondingly decreases the weighted coefficient for the rate of change term. This adjusted weighting configuration makes the dynamic deviation calculation more focused on the degree of deviation between the parameter and the target value, enhancing the ability to assess the accumulation of continuous deviations and enabling the system to effectively identify long-term drift trends in the parameter. The system applies the adjusted weighting coefficients to subsequent dynamic deviation calculations, achieving differentiated processing for different operating conditions.

[0054] Based on the above embodiments, as an optional implementation, step S301, in the process of calculating the dynamic deviation, specifically further includes: During the calculation of dynamic deviation, the preset interference indication parameters in the multidimensional production data are monitored in real time. When the change of the interference indication parameters within the preset monitoring period exceeds the preset interference judgment threshold, the time period corresponding to the preset monitoring period is determined as the instantaneous interference period of the equipment. The interference indication parameters are specific parameters in the equipment operating status parameters that characterize the abnormal state of the equipment. During the period of instantaneous equipment interference, freeze the matrix elements in the dynamic coupling correlation matrix corresponding to the target process quality parameters that are related to the interference indication parameters, and suspend the updating of the correlation analysis model for the target process quality parameters; After the equipment momentary interference period ends, an interpolation algorithm is used to complete the missing data of the target process quality parameters during the momentary interference period. Based on the completed data, the coupling relationship between the target process quality parameters and other parameters during the momentary interference period is recalculated, and the recalculated coupling relationship is updated to the dynamic coupling correlation matrix to maintain the temporal continuity of the dynamic coupling correlation matrix.

[0055] Among them, interference indication parameters refer to specific parameters in the equipment operating status parameters that can characterize the abnormal state of the equipment, including indicators that reflect abnormal operation of the equipment such as equipment vibration amplitude, current fluctuation, and temperature change; preset monitoring cycle refers to the fixed time length for continuous monitoring of interference indication parameters; interference judgment threshold refers to the critical value used to judge whether the change of interference indication parameters has reached the degree of abnormality; equipment instantaneous interference period refers to the time period during which the equipment experiences a short-term abnormal state, resulting in instantaneous fluctuations in production data.

[0056] Specifically, the system synchronously acquires real-time values ​​of pre-specified interference indication parameters from multi-dimensional production data within each sampling period, and several consecutive sampling periods constitute a preset monitoring period. The system extracts all sampled values ​​of the interference indication parameters within the preset monitoring period, identifies the maximum and minimum values, and calculates the difference between them to obtain the variation amplitude within that monitoring period. The system compares the calculated variation amplitude with the interference judgment threshold corresponding to the interference indication parameter. When the variation amplitude exceeds the interference judgment threshold, the system determines that the equipment has experienced a transient abnormal state within that monitoring period. The system records the start and end times of the preset monitoring period, marking this time period as the equipment's transient interference period, and simultaneously marking all sampling points within this period as interference impact data in the data annotation record.

[0057] The system extracts the identifiers of affected interference indication parameters based on the time stamp of the instantaneous equipment interference period. The system queries the dynamic coupling correlation matrix for all process quality parameters correlated with the interference indication parameter. These parameters are identified by elements in the matrix with non-zero correlation values, and are marked as target process quality parameters. The system traverses the dynamic coupling correlation matrix, locating all matrix elements corresponding to the target process quality parameter, including all parameter pairs where the parameter is the first or second parameter, and sets a freeze flag for these matrix elements. During the instantaneous equipment interference period, when executing the periodic update process of the correlation analysis model, the system checks the freeze flag of each matrix element to be updated, skips all correlation calculations and weight adjustments for frozen elements, and keeps the values ​​of these elements unchanged before the interference occurred. This avoids data anomalies caused by equipment malfunctions interfering with the identification and quantification of the true coupling relationship between parameters.

[0058] Once the system detects that the change in the interference indicator parameter has decreased to below the interference judgment threshold and remains stable, it determines that the instantaneous interference period has ended. The system extracts the start and end times of the interference period and locates the sampling data of all target process quality parameters within that period. The system processes this data using an interpolation algorithm, extracting normal data points before and after the interference period as interpolation benchmarks. Based on the continuity characteristics of the time series, it calculates the parameter estimates for each sampling time within the interference period and replaces the original abnormal data within the interference period with these estimates, completing the missing data completion. Based on the completed data sequence, the system re-executes the calculation of the coupling relationship between the target process quality parameters and other parameters, including re-statistically analyzing the synchronous change characteristics of parameter pairs, recalculating the cross-correlation coefficients, and re-evaluating the correlation strength. The system updates the corresponding matrix elements in the dynamic coupling correlation matrix with the recalculated coupling relationship values, unfreezes these elements, restores the matrix to its normal dynamic update process, and ensures that the time series of coupling relationships recorded in the matrix remains complete and continuous.

[0059] S302: Within the current sliding time window, based on the dynamic deviation of each process quality parameter at each sampling moment within the sliding time window, calculate the contribution of each process quality parameter to the preset target quality index. The contribution represents the degree of influence of the deviation of a single process quality parameter on the overall deviation of the target quality index. The contribution rate refers to the degree of influence of the deviation of a single process quality parameter on the overall deviation of the target quality indicator; the target quality indicator represents a comprehensive indicator for measuring the overall quality level of cigarette products, which is determined by multiple process quality parameters.

[0060] This step is executed after calculating the dynamic deviation of each process quality parameter, and is used to quantify the contribution of each parameter to the overall quality deviation. Specifically, the system extracts the dynamic deviation values ​​at all sampling moments within the current sliding time window. For each process quality parameter, it obtains the dynamic deviation sequence of that parameter at each sampling moment within the sliding time window. The system calculates the statistical characteristic values ​​of the parameter's dynamic deviation sequence, including the average, maximum, and standard deviation of the dynamic deviation in the sequence. Based on the parameter's weighting coefficient in the target quality indicator, the system multiplies the parameter's dynamic deviation statistical characteristic value with the weighting coefficient to obtain the parameter's contribution to the target quality indicator deviation. The system normalizes the contributions of all process quality parameters so that the sum of the contributions of each parameter equals one, obtaining the final contribution value for each process quality parameter. The larger the contribution value, the greater the impact of the parameter's deviation on the overall deviation of the target quality indicator.

[0061] S303: When the dynamic deviation of multiple process quality parameters continuously exceeds the dynamic threshold, the multiple process quality parameters with excessive dynamic deviation are sorted based on their contribution. The process quality parameter with the highest contribution is identified as the dominant influencing factor in the current production stage, and the dominant influencing factor is given priority as the first parameter to determine the strong coupling relationship.

[0062] Among them, the dominant influencing factor refers to the process quality parameter that has the most significant impact on the deviation of the target quality index at the current production stage.

[0063] This step is executed after calculating the contribution of each process quality parameter, and is used to identify and prioritize the parameters with the most critical impact on quality. Specifically, the system detects the dynamic deviation of all process quality parameters, identifies parameters whose dynamic deviation continuously exceeds a dynamic threshold, and determines whether the continuous exceedance is based on whether the number of exceedances or the duration of exceedances within a sliding time window meets preset conditions. When multiple process quality parameters are identified as simultaneously meeting the continuous exceedance condition, the system extracts the contribution values ​​of each of these exceedance parameters and sorts the exceedance parameters from highest to lowest contribution. The system selects the process quality parameter with the highest contribution in the sorted results and marks it as the dominant influencing factor in the current production stage. The system designates this dominant influencing factor as the first parameter and prioritizes the determination process of the strong coupling relationship between this parameter and other parameters. That is, it prioritizes querying the correlation value corresponding to this parameter in the dynamic coupling correlation matrix, prioritizes identifying the second parameter strongly coupled with this parameter, and prioritizes generating process parameter adjustment suggestions for this parameter, thereby achieving priority processing and rapid response for the parameters with the greatest impact on quality.

[0064] Based on the above embodiments, as an optional implementation method, before constructing the dynamic coupling correlation matrix based on the parameter sequence in step S102, steps S401-S403 are specifically included.

[0065] S401: Within multiple consecutive sliding time windows, continuously collect motor load data and vibration frequency data from the equipment operating status parameters. Based on the long-term mean drift of the motor load data and the spectral variation characteristics of the vibration frequency data, calculate the equipment fretting wear coefficient of the current production equipment. The equipment fretting wear coefficient characterizes the degree of gradual change in the mechanical performance of the production equipment due to long-term operation. Among them, motor load data refers to the load parameters such as power or current of the motor driving the production equipment during operation; vibration frequency data refers to the frequency distribution data obtained after frequency domain analysis of the vibration signal generated by the equipment during operation; long-term mean drift indicates the degree to which the average level of motor load data deviates from the initial reference value over a long period of time; spectrum change characteristics indicate the change of the frequency distribution pattern of vibration frequency data relative to the spectrum reference under normal equipment conditions; equipment fretting wear coefficient refers to a comprehensive index that quantifies the gradual performance degradation of production equipment caused by long-term operation, such as wear of mechanical parts, increased clearance, and decreased transmission accuracy.

[0066] Specifically, the system collects motor load data and vibration frequency data within each sliding time window and stores this data in a long-term data buffer. The system calculates the average value of the motor load data within the current sliding time window and compares this average value with the baseline motor load value recorded when the equipment was initially put into use. The difference between the two is the long-term mean drift; a positive drift indicates an increase in motor load, reflecting increased mechanical resistance. The system performs spectral analysis on the vibration frequency data, extracting the amplitude distribution of the main frequency components. It compares the current spectrum with the baseline spectrum of the equipment under normal conditions, identifying changes such as shifts in peak positions and increases in high-frequency components. Based on the combination of the long-term mean drift and spectral change characteristics, the system calculates the equipment's fretting wear coefficient according to a preset wear assessment model. This coefficient comprehensively reflects the gradual degradation of the equipment's mechanical performance.

[0067] S402: The equipment fretting wear coefficient is introduced as a long-term drift coupling variable into the dynamic coupling correlation matrix. Correlation elements between the equipment fretting wear coefficient and the process quality parameters affected by the equipment mechanical performance are established in the dynamic coupling correlation matrix. The process quality parameters affected by the equipment mechanical performance include cigarette weight parameters and end-burning amount parameters. Among them, long-term drift coupling variables refer to variables that reflect the long-term performance change trend of equipment and have a coupling relationship with process quality parameters; correlation elements refer to matrix items in the dynamic coupling correlation matrix that indicate the correlation between two parameters; cigarette weight parameter represents the mass of a single cigarette, which is affected by mechanical properties such as filling accuracy and conveying stability; end-drop tobacco amount parameter represents the amount of tobacco that falls off the end of the cigarette, which is affected by mechanical properties such as cutting accuracy and vibration control.

[0068] Specifically, the system introduces the equipment fretting wear coefficient as a new variable dimension into the dynamic coupling correlation matrix, assigning an independent row or column index to this coefficient within the matrix. The system identifies process quality parameters affected by the equipment's mechanical performance, including cigarette weight and end-burning amount; the numerical stability of these parameters directly depends on the equipment's mechanical precision and operational stability. The system establishes correlation elements between the equipment fretting wear coefficient and these process quality parameters in the dynamic coupling correlation matrix, creating corresponding matrix elements for each parameter pair to store the correlation strength value. The system analyzes the changing trends of the equipment fretting wear coefficient and process quality parameters in historical data, calculates the correlation strength between the two, and uses the calculation results as initial correlation strength values ​​to fill in the corresponding matrix elements, thereby achieving a quantitative expression of the coupling relationship between long-term equipment performance degradation and fluctuations in process quality parameters.

[0069] S403: When the cumulative change of the equipment fretting wear coefficient exceeds the preset wear change threshold within the preset long-term observation period, the correlation strength benchmark value of each correlation element related to the mechanical performance of the equipment in the dynamic coupling correlation matrix is ​​corrected according to the cumulative change of the equipment fretting wear coefficient.

[0070] Among them, the cumulative change represents the total change of the equipment fretting wear coefficient over a long-term observation period; the long-term observation period refers to a relatively long time span for continuous monitoring of equipment performance degradation, usually several days to several weeks; the wear change threshold refers to the critical value for determining whether the degree of equipment wear has reached the point where the correlation strength needs to be adjusted; the correlation strength benchmark value refers to the standard correlation strength value of the correlated elements in the dynamic coupling correlation matrix under normal equipment conditions.

[0071] Specifically, the system records the initial value of the equipment's fretting wear coefficient at the start of the long-term observation period, continuously tracks the change of this coefficient during the observation period, and records the current value of the equipment's fretting wear coefficient at the end of the observation period. The system calculates the cumulative change by comparing the current value with the initial value. The system compares the cumulative change with a preset wear change threshold. When the cumulative change exceeds the threshold, it indicates that the mechanical performance degradation of the equipment has reached a significant level. The system extracts all correlation elements related to the mechanical performance of the equipment from the dynamic coupling correlation matrix, including correlation elements between the equipment's fretting wear coefficient and cigarette weight parameters, end-burning amount parameters, etc. Based on the magnitude of the cumulative change, the system calculates a correction coefficient for the correlation strength according to a linear or nonlinear mapping relationship. The system multiplies the baseline correlation strength value of each correlation element by the correction coefficient to obtain the corrected correlation strength, which is then updated to the corresponding matrix elements, allowing the correlation strength to be dynamically adjusted according to the degree of equipment performance degradation.

[0072] Based on the above embodiments, as an optional implementation method, after determining in step S104 that there is a strong coupling relationship between the first parameter and the second parameter, the specific implementation includes steps S501-S503.

[0073] S501: When the dynamic deviation of multiple process quality parameters in the dynamic coupling correlation matrix exceeds their respective dynamic thresholds at the same time, based on the correlation distribution between multiple process quality parameters in the dynamic coupling correlation matrix and other parameters in the multidimensional production data, it is determined whether multiple process quality parameters have a common driving factor. The common driving factor is a single parameter that has a high correlation with multiple process quality parameters that exceed the deviation limit at the same time. Among them, the driving factor refers to the source parameter that affects multiple process quality parameters; the correlation distribution refers to the numerical distribution of the correlation strength between a certain parameter and other parameters in the dynamic coupling correlation matrix; the common driving factor refers to a single parameter in the dynamic coupling correlation matrix that maintains a high correlation with multiple process quality parameters that exceed the deviation limit, and this parameter has a common influence on the fluctuation of these process quality parameters.

[0074] Specifically, the system identifies all process quality parameters whose dynamic deviation exceeds their respective dynamic thresholds and marks these parameters as a set of parameters with excessive deviation. The system extracts the row or column corresponding to each parameter with excessive deviation from the dynamic coupling correlation matrix, obtaining the correlation degree values ​​between this parameter and all other parameters in the multidimensional production data, thus constructing the correlation degree distribution of this parameter. The system performs intersection analysis on the correlation degree distributions of all parameters in the set of parameters with excessive deviation, searching for parameters that exhibit high correlation in all correlation degree distributions. The system sets a criterion for high correlation, traverses each candidate parameter in the multidimensional production data, and checks whether the candidate parameter maintains a correlation degree higher than the criterion with all parameters in the set of parameters with excessive deviation. If the condition is met, the candidate parameter is determined to be a common driving factor. The system counts the number of candidate parameters that meet the condition; when at least one parameter meets the condition, a common driving factor is determined to exist.

[0075] S502: When a common driving factor is identified, the common driving factor is identified as a source of coupling interference. The influence component of the common driving factor on each process quality parameter with excessive deviation is calculated. The influence component caused by the common driving factor is removed from the dynamic deviation of each process quality parameter to obtain the pure deviation of each process quality parameter after removing coupling interference. Among them, the coupled interference source refers to the driving factor identified as having a common impact on multiple process quality parameters; the influence component represents the numerical part of the contribution of the coupled interference source to the dynamic deviation of a specific process quality parameter; the pure deviation refers to the deviation value remaining after removing the influence of the coupled interference source from the dynamic deviation of the process quality parameter, which reflects the deviation caused by the parameter itself or other independent factors.

[0076] Specifically, the system identifies common driving factors as coupling interference sources and records their parameter identifiers and current values. For each process quality parameter with excessive deviation, the system extracts the correlation value between the parameter and the coupling interference source. This correlation reflects the intensity of the coupling interference source's influence on the process quality parameter. The system obtains the deviation or fluctuation of the coupling interference source within the current sliding time window, multiplies this deviation by the correlation between the process quality parameter and the coupling interference source, and obtains the influence component of the coupling interference source on the process quality parameter. The system extracts the current dynamic deviation value of the process quality parameter, subtracts the calculated influence component from this value, and obtains the pure deviation value after removing coupling interference. The system repeats the above influence component calculation and removal operation for each parameter in the set of parameters with excessive deviation to obtain the pure deviation value for each parameter with excessive deviation. These pure deviation values ​​eliminate the interference of common driving factors and more accurately reflect the direct coupling relationship between the parameters.

[0077] S503: Recalculate the correlation between process quality parameters that exceed the standard based on the purity deviation, output the recalculated correlation as the purity correlation path, and generate interpretable coupling rules based on the purity correlation path.

[0078] Among them, the pure correlation path refers to the correlation degree between parameters recalculated based on the pure deviation degree. This correlation degree excludes the confusion effect of common driving factors and reflects the direct coupling relationship between parameters. The interpretable coupling rule refers to the rule expression that describes the causal relationship or correlation pattern between parameters. This rule is extracted based on the pure correlation path and can provide a clear logical basis for process adjustment.

[0079] Specifically, the system re-executes the correlation calculation process between parameter pairs based on the pure deviation values ​​of each parameter exceeding the deviation limit. The system extracts the pure deviation time series of any two parameters exceeding the deviation limit within the current sliding time window, calculates the synchronous change characteristics of the two series, including the consistency of the change direction and the correlation of the change magnitude, and quantifies the correlation between the two parameters based on these characteristics. The system updates the corresponding matrix elements in the dynamic coupling correlation matrix with the recalculated correlation values. These updated correlations constitute the pure correlation paths, reflecting the true coupling strength between parameters after removing common interference. The system analyzes parameter pairs with high correlation in the pure correlation paths, extracts the historical change patterns of these parameter pairs, and identifies coupling patterns such as lead-lag relationships and positive / negative correlation characteristics between parameters. The system transforms the identified coupling patterns into rule form, generating interpretable coupling rules that include conditional judgments and result predictions.

[0080] The following are system embodiments of this application, which can be used to execute the method embodiments of this application. For details not disclosed in the system embodiments of this application, please refer to the method embodiments of the application.

[0081] Please see Figure 3 This illustration shows a schematic diagram of a cigarette production quality optimization system based on multi-factor correlation analysis, provided in an exemplary embodiment of this application. The system can be implemented entirely or partially through software, hardware, or a combination of both. The cigarette production quality optimization system based on multi-factor correlation analysis includes: The data acquisition module is used to collect multi-dimensional production data in the cigarette production line in real time, and to perform time-series slicing processing on the multi-dimensional production data to obtain parameter sequences of multiple micro time periods. The multi-dimensional production data includes process quality parameters, equipment operating status parameters and environmental parameters. The association matrix construction module is used to construct a dynamic coupling association matrix based on the parameter sequence, and calculate the environmental fitness factor according to the real-time fluctuation of environmental parameters in multidimensional production data. The environmental fitness factor is used to dynamically adjust the association threshold between each parameter pair in the dynamic coupling association matrix. The deviation calculation module is used to calculate the dynamic deviation of each process quality parameter in multidimensional production data relative to the target value based on the dynamic coupling correlation matrix within a preset sliding time window. The coupling determination module is used to compare the dynamic deviation with the dynamic threshold that changes with the production conditions. When the dynamic deviation of the first parameter continues to exceed the dynamic threshold, and the correlation between the first parameter and the second parameter is determined to exceed the correlation threshold based on the dynamic coupling correlation matrix, it is determined that there is a strong coupling relationship between the first parameter and the second parameter, and the coupling relationship description information between the first parameter and the second parameter is output. The adjustment suggestion generation module is used to generate process parameter adjustment suggestions for the first parameter and / or the second parameter based on the coupling relationship description information, and send the adjustment suggestions to the control system of the cigarette production line.

[0082] This application also provides a computer storage medium that can store multiple instructions. The instructions are adapted to be loaded and executed by a processor as described in the above embodiments for the method of optimizing cigarette production quality based on multi-factor correlation analysis. For the specific execution process, please refer to the detailed description of the embodiments, which will not be repeated here.

[0083] Please see Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 4 As shown, the electronic device 400 may include: at least one processor 401, at least one network interface 404, user interface 403, memory 405, and at least one communication bus 402.

[0084] The communication bus 402 is used to enable communication between these components.

[0085] The user interface 403 may include a display screen and a camera.

[0086] The network interface 404 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0087] The processor 401 may include one or more processing cores. The processor 401 connects to various parts of the server using various interfaces and lines, and performs various server functions and processes data by running or executing instructions, programs, code sets, or instruction sets stored in memory 405, and by calling data stored in memory 405. Optionally, the processor 401 may be implemented using at least one hardware form of digital signal processing, field-programmable gate array, or programmable logic array. The processor 401 may integrate one or more of the following: a central processing unit (CPU), a graphics processing unit (GPU), and a modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content required for display; and the modem handles wireless communication. It is understood that the modem may also be implemented as a separate chip, without being integrated into the processor 401.

[0088] The memory 405 may include random access memory (RAM) or read-only memory (ROM). Optionally, the memory 405 may include a non-transitory computer-readable medium. The memory 405 may be used to store instructions, programs, code, code sets, or instruction sets. The memory 405 may include a program storage area and a data storage area. The program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch functionality, sound playback functionality, image playback functionality, etc.), instructions for implementing the various method embodiments described above, etc.; the data storage area may store data involved in the various method embodiments described above, etc. Optionally, the memory 405 may also be at least one storage device located remotely from the aforementioned processor 401. Figure 4 As shown, the memory 405, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and an application program for optimizing cigarette production quality based on multi-factor correlation analysis.

[0089] exist Figure 4 In the electronic device 400 shown, the user interface 403 is mainly used to provide an input interface for the user and to obtain the user input data; while the processor 401 can be used to call an application program stored in the memory 405 that is a method for optimizing the quality of cigarette production based on multi-factor correlation analysis. When executed by one or more processors, the electronic device executes one or more methods as described in the above embodiments.

[0090] An electronic device readable storage medium stores instructions that, when executed by one or more processors, cause the electronic device to perform one or more methods as described in the above embodiments.

[0091] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0092] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0093] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the shown or discussed mutual couplings or direct couplings or communication connections may be through some service interfaces; indirect couplings or communication connections between apparatuses or units may be electrical or other forms.

[0094] The units described as separate components may or may not be physically separate. The components shown as units 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 units can be selected to achieve the purpose of this embodiment according to actual needs.

[0095] Furthermore, the functional units in the various embodiments of this application 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 as a software functional unit.

[0096] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, portable hard drives, magnetic disks, or optical disks.

[0097] The above are merely exemplary embodiments of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure. Those skilled in the art will readily conceive of other embodiments of this disclosure upon considering the specification and practical application disclosed herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described in this disclosure.

Claims

1. A method for optimizing cigarette production quality based on multi-factor correlation analysis, characterized in that, The method includes: Multidimensional production data from the cigarette production line is collected in real time, and time-series slicing is performed on the multidimensional production data to obtain parameter sequences for multiple micro-time periods. The multidimensional production data includes process quality parameters, equipment operating status parameters, and environmental parameters. A dynamic coupling correlation matrix is ​​constructed based on the parameter sequence, and an environmental fitness factor is calculated based on the real-time fluctuation of environmental parameters in the multidimensional production data. The correlation threshold between each parameter pair in the dynamic coupling correlation matrix is ​​dynamically adjusted using the environmental fitness factor. The dynamic coupling correlation matrix characterizes the correlation strength between each process quality parameter and equipment operating status parameter in the multidimensional production data. Within a preset sliding time window, based on the dynamic coupling correlation matrix, the dynamic deviation of each process quality parameter in the multidimensional production data relative to the target value is calculated. The dynamic deviation is compared with a dynamic threshold that changes with production conditions. When the dynamic deviation of the first parameter continuously exceeds the dynamic threshold, and the correlation between the first parameter and the second parameter is determined to exceed the correlation threshold based on the dynamic coupling correlation matrix, a strong coupling relationship is determined between the first parameter and the second parameter, and the coupling relationship description information between the first parameter and the second parameter is output. Herein, the first parameter is the parameter whose dynamic deviation exceeds the standard among the process quality parameters, and the second parameter is the process quality parameter or equipment operating status parameter in the multidimensional production data that is correlated with the first parameter. Based on the coupling relationship description information, process parameter adjustment suggestions are generated for the first parameter and / or the second parameter, and the adjustment suggestions are sent to the control system of the cigarette production line.

2. The method according to claim 1, characterized in that, The step of calculating an environmental fitness factor based on real-time fluctuations of environmental parameters in the multidimensional production data, and dynamically adjusting the correlation thresholds between parameter pairs in the dynamic coupling correlation matrix using the environmental fitness factor, includes: Within each sliding time window, the temperature and humidity parameters of the current production workshop are acquired as environmental parameters, and the fluctuation range of the temperature and humidity parameters relative to their respective preset reference values ​​within the current sliding time window is calculated. The fluctuation amplitude is compared with a preset environmental fluctuation threshold. When the fluctuation amplitude of the temperature parameter and / or the humidity parameter exceeds the corresponding environmental fluctuation threshold, the environmental adaptability factor is calculated based on the degree of deviation between the fluctuation amplitude exceeding the environmental fluctuation threshold and the corresponding preset benchmark value. Based on the environmental fitness factor, target parameter pairs affected by the fluctuating environmental parameters are determined in the dynamic coupling correlation matrix. The static correlation weight of the target parameter pairs, determined based on historical statistical data, is reduced, and the dynamic monitoring weight of the target parameter pairs, determined based on real-time data within the current sliding time window, is increased accordingly, thereby updating the correlation threshold between the target parameter pairs.

3. The method according to claim 1, characterized in that, The calculation of the dynamic deviation includes a deviation term between the current value and the target value of the parameter, and a parameter change rate term. The weighting coefficients of the deviation term and the change rate term are dynamically adjusted with production time. Within a preset sliding time window, based on the dynamic coupling correlation matrix, the dynamic deviation of each process quality parameter in the multidimensional production data relative to the target value is calculated, including: For each of the process quality parameters i, according to the formula Calculate the dynamic deviation Di(t) of the process quality parameter i at time t, where X_i(t) is the current sampled value of the process quality parameter i at time t, X_target(t) is the target value corresponding to the process quality parameter i, ΔX_i(t) is the rate of change of the process quality parameter i at time t, α is the weighting coefficient of the deviation term, and β is the weighting coefficient of the rate of change term; Within the current sliding time window, based on the dynamic deviation of each process quality parameter at each sampling time within the sliding time window, the contribution of each process quality parameter to the preset target quality index is calculated. The contribution represents the degree of influence of the deviation change of a single process quality parameter on the overall deviation of the target quality index. When the dynamic deviation of multiple process quality parameters continuously exceeds the dynamic threshold, the multiple process quality parameters with excessive dynamic deviation are sorted based on their contribution. The process quality parameter with the highest contribution is identified as the dominant influencing factor in the current production stage, and the dominant influencing factor is used as the first parameter to perform the determination of the strong coupling relationship.

4. The method according to claim 3, characterized in that, The method further includes: At the beginning of each sliding time window, the data fluctuation characteristic values ​​of each process quality parameter within the current sliding time window are statistically analyzed. The data fluctuation characteristic values ​​include the fluctuation frequency and fluctuation amplitude of each process quality parameter within the time span of the current sliding time window. The fluctuation frequency is the number of times the parameter value crosses its mean line within the current sliding time window, and the fluctuation amplitude is the ratio of the range of the parameter value within the current sliding time window to the target value. Based on the data fluctuation characteristic value, the production condition type of the current production stage is determined. When the fluctuation frequency of the process quality parameter is higher than the preset frequency threshold, the current production condition is determined to be a rapid fluctuation condition. When the fluctuation amplitude of the process quality parameter is greater than the preset amplitude threshold and the fluctuation frequency is lower than the frequency threshold, the current production condition is determined to be a slow drift condition. Based on the determined production condition type, the values ​​of the weighting coefficients α and β of the corresponding process quality parameters are adaptively adjusted: under the rapid fluctuation condition, the value of the weighting coefficient β of the corresponding process quality parameter is increased and the value of the weighting coefficient α is correspondingly decreased, so that the calculation of the dynamic deviation focuses on capturing the trend characteristics of the parameter change rate; under the slow drift condition, the value of the weighting coefficient α of the corresponding process quality parameter is increased and the value of the weighting coefficient β is correspondingly decreased, so that the calculation of the dynamic deviation focuses on evaluating the cumulative effect of the parameter deviation amplitude.

5. The method according to claim 3, characterized in that, The method further includes: During the calculation of the dynamic deviation, the preset interference indication parameter in the multidimensional production data is monitored in real time. When the change of the interference indication parameter within the preset monitoring period exceeds the preset interference judgment threshold, the time period corresponding to the preset monitoring period is determined as the instantaneous interference period of the equipment. The interference indication parameter is a specific parameter in the equipment operating status parameters that characterizes the abnormal state of the equipment. During the instantaneous interference period of the equipment, freeze the matrix elements in the dynamic coupling correlation matrix corresponding to the target process quality parameters that are related to the interference indication parameters, and suspend the updating of the correlation analysis model for the target process quality parameters; After the equipment instantaneous interference period ends, an interpolation algorithm is used to complete the missing data of the target process quality parameter during the equipment instantaneous interference period. Based on the completed data, the coupling relationship between the target process quality parameter and other parameters during the equipment instantaneous interference period is recalculated, and the recalculated coupling relationship is updated to the dynamic coupling correlation matrix to maintain the temporal continuity of the dynamic coupling correlation matrix.

6. The method according to claim 1, characterized in that, Before constructing the dynamic coupling correlation matrix based on the parameter sequence, the method further includes: Within multiple consecutive sliding time windows, motor load data and vibration frequency data in the equipment operating status parameters are continuously collected. Based on the long-term mean drift of the motor load data and the spectral variation characteristics of the vibration frequency data, the equipment fretting wear coefficient of the current production equipment is calculated. The equipment fretting wear coefficient characterizes the degree of gradual change in the mechanical performance of the production equipment due to long-term operation. The equipment fretting wear coefficient is introduced as a long-term drift coupling variable into the dynamic coupling correlation matrix. Correlation elements between the equipment fretting wear coefficient and the process quality parameters affected by the equipment mechanical performance are established in the dynamic coupling correlation matrix. The process quality parameters affected by the equipment mechanical performance include cigarette weight parameters and end-burning amount parameters. When the cumulative change of the equipment's fretting wear coefficient exceeds a preset wear change threshold within a preset long-term observation period, the correlation strength benchmark value of each correlation element related to the equipment's mechanical performance in the dynamic coupling correlation matrix is ​​corrected based on the cumulative change of the equipment's fretting wear coefficient.

7. The method according to claim 1, characterized in that, After determining that there is a strong coupling relationship between the first parameter and the second parameter, the method further includes: When multiple process quality parameters in the dynamic coupling correlation matrix have dynamic deviations that simultaneously exceed their respective dynamic thresholds, the correlation distribution between the multiple process quality parameters in the dynamic coupling correlation matrix and other parameters in the multidimensional production data is used to determine whether the multiple process quality parameters have a common driving factor. The common driving factor is a single parameter that has a high correlation with multiple process quality parameters that have exceeded the deviation thresholds. When the common driving factor is identified, it is identified as a coupling interference source. The influence component of the common driving factor on each process quality parameter with excessive deviation is calculated. The influence component caused by the common driving factor is removed from the dynamic deviation of each process quality parameter to obtain the pure deviation of each process quality parameter after removing coupling interference. Based on the purity deviation, the correlation between the process quality parameters that exceed the deviation limit is recalculated, and the recalculated correlation is output as the purity correlation path. An interpretable coupling rule is generated based on the purity correlation path.

8. A cigarette production quality optimization system based on multi-factor correlation analysis, characterized in that, The system includes: The data acquisition module is used to collect multi-dimensional production data in the cigarette production line in real time, and to perform time-series slicing processing on the multi-dimensional production data to obtain parameter sequences of multiple micro time periods. The multi-dimensional production data includes process quality parameters, equipment operating status parameters and environmental parameters. The correlation matrix construction module is used to construct a dynamic coupling correlation matrix based on the parameter sequence, calculate the environmental fitness factor according to the real-time fluctuation of environmental parameters in the multidimensional production data, and dynamically adjust the correlation threshold between each parameter pair in the dynamic coupling correlation matrix using the environmental fitness factor. The deviation calculation module is used to calculate the dynamic deviation of each process quality parameter in the multidimensional production data relative to the target value based on the dynamic coupling correlation matrix within a preset sliding time window. The coupling determination module is used to compare the dynamic deviation with a dynamic threshold that changes with production conditions. When the dynamic deviation of the first parameter continuously exceeds the dynamic threshold, and the correlation between the first parameter and the second parameter is determined to exceed the correlation threshold based on the dynamic coupling correlation matrix, it is determined that there is a strong coupling relationship between the first parameter and the second parameter, and the coupling relationship description information between the first parameter and the second parameter is output. The adjustment suggestion generation module is used to generate process parameter adjustment suggestions for the first parameter and / or the second parameter based on the coupling relationship description information, and send the adjustment suggestions to the control system of the cigarette production line.

9. A computer storage medium, characterized in that, The computer storage medium stores a plurality of instructions, which are adapted to be loaded by a processor and executed as described in any one of claims 1 to 7.

10. An electronic device, characterized in that, The device includes a processor, a memory, and a transceiver, wherein the memory is used to store instructions, the transceiver is used to communicate with other devices, and the processor is used to execute the instructions stored in the memory to cause the electronic device to perform the method as described in any one of claims 1 to 7.

Images