A high-frequency high-speed transmission cable forming control system

Abstract

The present application relates to the technical field of industrial automation control, and in particular to a high-frequency high-speed transmission cable forming control system, comprising: according to the cable production process, the production line is divided into multiple forming intervals, the initial process parameters of each forming interval are obtained, the abnormal aperture of the cable is taken as a trigger condition, the initial process parameters of the cable are monitored, and the influence parameter set when the cable is continuously extruded is determined; according to the influence parameter set of the current cable, the deviation state is deduced, the deviation abnormality of the influence parameter set is obtained; the deviation abnormality of the influence parameter set is taken as an index, the fluctuation proportion at the corresponding moment is combined, the adjustment factor of the collaborative control is determined; the position point when the aperture is abnormal is associated according to the adjustment factor of the collaborative control, the abnormal aperture section is configured, and the corresponding initial process parameters are adjusted according to the numerical mutation of the abnormal aperture section. The precision and efficiency of cable production are realized, and the cost and time of production are reduced.

Description

Technical Field

[0001] This invention relates to the field of industrial automation control technology, specifically a high-frequency, high-speed transmission cable forming control system. Background Technology

[0002] As a fundamental support for modern industry, the cable industry is widely used in power transmission, communications, automobile manufacturing, construction, and many other fields. Its electrical performance is highly dependent on properties such as the size and continuity of the core aperture inside the cable. In the extrusion molding process of cored cables, it is necessary to ensure product quality while reducing production costs, taking into account fluctuations in process parameters.

[0003] For example, Chinese Patent Publication No. CN119871852A discloses a method and system for measuring and controlling material loss during cable extrusion, relating to the field of industrial automation control technology. The method includes: acquiring the current material extrusion rate of the extruder and the outer diameter data of the produced cable, wherein the material extrusion rate is obtained based on a high-precision, high-speed weighing device, and the outer diameter data is measured in real time by a diameter measuring instrument; calculating the current actual material loss value based on the material extrusion rate and outer diameter data; and adjusting the screw speed of the extruder and the traction speed of the traction machine for the cable based on the actual material loss value and outer diameter data, so that the actual material loss value and outer diameter data are respectively within corresponding set ranges.

[0004] For example, Chinese Patent Publication No. CN119017669A discloses an injection molding process control method, system, and storage medium. The control method includes the following steps: preheating the mold to an initial temperature by controlling the heating timing zone, the initial temperature being determined by the heating temperature on the control setting interface; maintaining the mold temperature for a stable time, the stabilization time being timed by the injection delay timing in the process display area; injecting molten material into the mold by controlling the injection pressure and speed, the injection parameters being determined by the setting interface, and adjusting the injection pressure and speed in real time; maintaining the injection pressure for a holding time, the holding time being timed by the holding pressure timing zone, and monitoring the pressure changes inside the mold in real time during the holding pressure process.

[0005] Existing technologies use material extrusion volume and outer diameter data to integrally calculate outer diameter deviation and determine material loss during cable production; or they use temperature and pressure control during mold production to determine overall process control. However, existing technologies tend to rely on single-parameter independent closed-loop control, failing to consider the synergistic effects of process parameters in different forming intervals. This makes it impossible to locate abnormal aperture positions based on the correlation of multiple parameters, leading to reduced cable production efficiency and insufficient yield. Summary of the Invention

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a high-frequency and high-speed transmission cable forming control system, including: a data acquisition module, used to divide the production line into multiple forming sections according to the cable production process, and obtain the initial process parameters of each forming section, including the die angle, vacuum degree, linear speed, segment temperature value and extrusion speed when the cable is extruded.

[0007] The parameter monitoring module is used to monitor the initial process parameters of the cable based on the abnormality of the cable's aperture as a trigger condition, and to determine the set of influencing parameters when the cable is continuously extruded.

[0008] The deviation identification module is used to deduce the deviation status based on the current set of influence parameters of the cable, determine the mapping relationship between each parameter in the set of influence parameters and the aperture anomaly, and obtain the deviation anomaly degree of the set of influence parameters.

[0009] The collaborative control module is used to take the process parameters corresponding to each forming interval as the control target, use the deviation anomaly degree of the parameter set as the index, and combine the fluctuation ratio at the corresponding time to determine the adjustment factor of the collaborative control.

[0010] The parameter feedback module is used to configure abnormal aperture segments based on the adjustment factor of the collaborative control, the location point of the abnormal aperture, and the initial process parameters of the corresponding forming interval are adjusted according to the numerical change of the abnormal aperture segment.

[0011] The beneficial effects of this invention are as follows: First, by dividing the production line into multiple forming sections, this invention acquires initial process parameters such as mold angle, vacuum degree, linear speed, segment temperature, and extrusion speed in real time. Using orifice anomalies as a trigger condition, it reverse-engineers the target parameters according to the process flow, compares the time / location correspondence between parameter fluctuations and orifice anomalies, preliminarily determines the influencing parameters, and then forms a set of influencing parameters based on the priority of anomaly type. This clarifies the influencing parameters related to orifice anomalies and combines them according to the dimensions of the currently acquired parameters, avoiding the problem of difficulty in distinguishing parameter combinations during parameter backtracking, and providing a data foundation for subsequent data processing.

[0012] II. This invention calculates residual values ​​based on the range of values ​​of influencing parameters within a unit of time as a verification benchmark. These residual values ​​are combined to form multiple candidate item sets. Association rule mining is performed on candidate item sets at any given time, and data that meets the rules is identified as deviation states. Specifically, during association rule mining, consecutive residual time points are merged into abnormal periods. Parameters within these abnormal periods are paired according to upstream and downstream process parameters. After time-series and process constraints, candidate item sets are filtered by frequency of occurrence, gradually expanding to obtain frequent itemsets. Finally, association rules are generated, and the mapping between parameters and aperture anomalies is completed. This quantifies the impact of various industrial parameters on aperture anomalies under combination, realizing the association expression of multi-parameter anomalies and providing a foundation for subsequent data tracing and deviation anomaly quantification.

[0013] Third, this invention calculates the fluctuation ratio of influencing parameters, obtains a trend coefficient by curve fitting of the fluctuation ratio, and combines the mapping relationship between parameters and aperture anomalies to calculate the similarity of the trend coefficient. The final similarity value is used as the deviation anomaly degree. This achieves a quantitative interpretation of anomalies in multi-parameter combinations. Subsequently, a hierarchical retrieval is performed based on the deviation anomaly degree threshold. The final control target is determined by the least common subset of the initial control targets. Based on the absolute value of the difference between the current deviation anomaly degree and the average deviation anomaly degree, a corresponding adjustment factor is configured. This hierarchical retrieval and subset configuration method reduces conflicts in control targets under multi-parameter combinations, further restricts the goals and processes of coordinated control, and improves the accuracy and rationality of cable data processing.

[0014] Fourth, this invention clearly displays the location points corresponding to the adjustment factors in the parameter feedback module, combines continuous location points into abnormal aperture segments, clusters each abnormal segment into equal-length sub-segments, and determines the adjustment direction of process parameters based on the adjustment factors corresponding to the clusters, thus completing the initial adjustment of process parameters. This achieves the binding of process adjustment actions with actual abnormal data, improving the accuracy and processing efficiency for cable segment control. Attached Figure Description

[0015] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0016] Figure 1 This is a system framework diagram of a high-frequency, high-speed transmission cable forming and control system.

[0017] Figure 2 This is a flowchart illustrating the parameter monitoring module of a high-frequency, high-speed transmission cable forming control system.

[0018] Figure 3 This is a flowchart illustrating the deviation identification module of a high-frequency, high-speed transmission cable forming control system.

[0019] Figure 4This is a flowchart illustrating the collaborative control module of a high-frequency, high-speed transmission cable forming control system.

[0020] Figure 5 This is a flowchart illustrating the parameter feedback module of a high-frequency, high-speed transmission cable forming control system. Detailed Implementation

[0021] The embodiments of the present invention are described in detail below. The embodiments described below are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. Where specific techniques or conditions are not specified in the embodiments, they shall be performed in accordance with the techniques or conditions described in the literature in the art or in accordance with the product manual.

[0022] See Figure 1 A high-frequency, high-speed transmission cable forming control system includes: a data acquisition module, a parameter monitoring module, a deviation identification module, a collaborative control module, and a parameter feedback module; wherein, the output end of the data acquisition module is connected to the parameter monitoring module, the output end of the parameter monitoring module is connected to the deviation identification module, the output end of the deviation identification module is connected to the collaborative control module, and the output end of the collaborative control module is connected to the parameter feedback module.

[0023] The data acquisition module is used to divide the production line into multiple forming sections according to the cable production process and obtain the initial process parameters of each forming section. These initial process parameters include the die angle, vacuum degree, linear speed, segment temperature value and extrusion speed when the cable is extruded.

[0024] The parameter monitoring module is used to monitor the initial process parameters of the cable based on the abnormality of the cable's aperture as a trigger condition, and to determine the set of influencing parameters when the cable is continuously extruded.

[0025] The deviation identification module is used to deduce the deviation status based on the current set of influence parameters of the cable, determine the mapping relationship between each parameter in the set of influence parameters and the aperture anomaly, and obtain the deviation anomaly degree of the set of influence parameters.

[0026] The collaborative control module is used to take the process parameters corresponding to each forming interval as the control target, use the deviation anomaly degree of the parameter set as the index, and combine the fluctuation ratio at the corresponding time to determine the adjustment factor of the collaborative control.

[0027] The parameter feedback module is used to configure abnormal aperture segments based on the adjustment factor of the collaborative control, the location point of the abnormal aperture, and the initial process parameters of the corresponding forming interval are adjusted according to the numerical change of the abnormal aperture segment.

[0028] In the specific implementation scenario of this invention, the produced cable is a cored cable, and the control status during the cable extrusion process is determined by real-time monitoring of aperture data. Its forming process is divided into multiple continuous stages, including a melting stage, a forming stage, a shaping stage, and a cooling stage. The system uses the aperture anomaly at the final output end as the primary trigger condition, traces back to the initial process parameters in each process flow, and dynamically adjusts the preset parameters based on the relative degree of anomaly and their synergistic effects, thereby obtaining the optimal parameter configuration for the current production scenario.

[0029] In the data acquisition module, the method for determining the initial process parameters during segmented extrusion further includes: using the aperture during the extrusion process as the verification benchmark parameter; when an aperture abnormality is detected, synchronous combination is performed according to the extrusion sequence of the cable; the range of values ​​for mold angle, vacuum degree, linear speed, segment temperature value, and extrusion speed corresponding to each segment is statistically analyzed; and the data is grouped according to the forming interval to which the initial process parameters belong, serving as the basic dataset for the current analysis.

[0030] During the actual data acquisition process, the vacuum level is collected in real time by a vacuum pressure sensor installed in the forming cavity; temperature data is collected in segments by multiple temperature sensors installed in the extruder body, die head, mold, and shaping section; linear speed is obtained in real time by the traction machine encoder; output parameters such as the diameter and concentricity of the formed cable are collected in real time by a laser online diameter gauge or a vision inspection device to determine the forming quality; the mold angle is measured by the mold cone angle or the exit angle; and the extrusion speed is checked by the screw speed.

[0031] To adapt to small-batch trial production environments, the system segments and calibrates the cables according to fixed lengths, for example, with each 1 meter as a calibration segment. It records the aperture deviation value at each segment and the uniformity of the aperture in continuous distribution. At the same time, it determines the initial process parameters corresponding to the generation of the segment and synchronizes them to the corresponding forming interval to analyze the impact of parameters such as vacuum degree, line speed, and temperature on the extrusion quality of the cable.

[0032] Among the selected initial process parameters, vacuum level has a direct impact on cable quality. It is typically controlled within the range of -0.08 MPa to -0.1 MPa to remove air bubbles and volatile substances from the insulation material during extrusion, ensuring that the insulation layer forms a uniform and stable core pore structure after cooling. Insufficient vacuum level can easily lead to residual air bubbles, resulting in an irregular structure with inconsistent pore sizes and uneven distribution in the insulation layer, thus reducing cable quality. Conversely, excessive vacuum level may cause excessive shrinkage of the insulation material, leading to core collapse or excessively small pore sizes, thereby affecting the dielectric constant and overall structural stability.

[0033] Four parameters—die angle, linear speed, segment temperature, and extrusion speed—jointly influence the tensile and packing behavior of the material and are key factors requiring verification. The system checks the temperatures of each segment in the melting, forming, sizing, and cooling sections, as well as parameters such as the die angle during die sizing, to determine if there are any abnormalities in the initial configuration. It then analyzes the relative impact of these abnormalities in the parameter combinations and provides feedback to adjust the current equipment configuration to achieve the optimal parameter combination.

[0034] In one embodiment of the present invention, the parameter monitoring module is used to perform regression verification on aperture anomalies identified after cable production is completed. The aperture anomalies refer not only to deviations in aperture size, but also to various other anomalies such as channel blockage, channel deformity, channel inhomogeneity, and deviations in the concentricity of the channel and conductor. All of these anomaly data serve as trigger conditions to determine the process flow segments requiring parameter monitoring.

[0035] like Figure 2 As shown, the parameter monitoring module, when determining the set of influencing parameters during the continuous extrusion process of the cable, further includes the following implementation method: First, based on the process flow corresponding to each forming section, the currently collected initial process parameters are reverse-engineered to extract multiple sets of target parameters. The target parameters are the initial process parameters corresponding to each process flow, and the currently configured process parameters are combined according to a preset time sequence logic. Specifically, the die angle determines the shape of the material flow channel, which directly affects the formation of the core pores; the linear speed is a key parameter affecting the cable extrusion and traction balance; the segment temperature value affects the stability of the pore structure; and the extrusion speed directly determines the insulation layer thickness and pore density. Based on the time period and process sequence corresponding to each parameter in the current process, a set of target parameters is constructed in a time-continuous manner in the order of die angle → linear speed → segment temperature value → extrusion speed → vacuum degree, and divided into multiple sets of target parameter combinations according to the actual values ​​of each parameter in the current production process.

[0036] Secondly, based on the temporal sequence of aperture anomaly triggering, the correspondence between the occurrence time of the aperture anomaly and the fluctuation time of each target parameter is compared. If the fluctuation time of a certain target parameter corresponds to the location of the aperture anomaly, then that target parameter is preliminarily identified as an influencing parameter.

[0037] Since aperture anomalies are detected as output data after the cable has completed its shaping process, the system needs to trace back to the location of the aperture anomaly according to the normal cable processing flow to determine its corresponding production time point. If the target parameter at that time point fluctuates, the corresponding parameter is considered as an influencing parameter leading to aperture anomalies such as aperture deviation and channel deformity.

[0038] Finally, based on the aperture anomaly type corresponding to the initially determined influencing parameters, a matching analysis is performed, and the priority order of each influencing parameter during the matching process is combined to form the final output set of influencing parameters.

[0039] Furthermore, firstly, for the data portion of the currently introduced set of influencing parameters that exhibits numerical fluctuations, a cross-correlation matrix is ​​constructed by using the set of influencing parameters as input data. After performing eigenvalue decomposition on the cross-correlation matrix, the priority weight corresponding to each influencing parameter is determined according to the ratio of the eigenvalue of each parameter in the set of influencing parameters to the total eigenvalue. Then, based on the magnitude of the priority weight, a priority is assigned to each influencing parameter (e.g., linear velocity, segmented temperature value, etc.) from high to low, and it is associated with and stored in relation to the aperture anomaly type corresponding to the numerical fluctuation, forming a priority-related database.

[0040] Based on this, the currently identified impact parameters and aperture anomaly types are used as search conditions to search from priority-related databases, obtaining the search sets corresponding to the impact parameters and the search sets corresponding to the aperture anomaly types respectively. The intersection of the two sets is selected as the current matching search data set, and the priority covered in this set is determined as the matching priority of the impact parameters in the current scenario.

[0041] The essence of parameter matching analysis lies in identifying the priority parameters that cause pore size anomalies. For example, when the overall pore size is too large or too small, vacuum degree is prioritized and set as the highest priority, followed by analysis of other influencing parameters. When the pore size fluctuates continuously, linear speed and segmented temperature values ​​are prioritized. When a local abrupt change occurs in the pore size, abrupt change in extrusion speed is prioritized. The above examples only illustrate the priority matching relationships under some types of pore size anomalies. In actual processing, based on the specific type of the current pore size anomaly, the priority database is queried to determine the priority matching and subsequent matching parameters, and these parameters are used as the current output set of influencing parameters. The priority values ​​of the matching are determined according to the query results of the pore size anomaly analysis scenario, corresponding to the priority weight of each influencing parameter.

[0042] In one embodiment of the present invention, when it is necessary to combine the value range of the target parameter to screen the set of influencing parameters that conforms to the current scenario, the fluctuation ratio corresponding to the initial process parameter is further introduced. The fluctuation ratio is used to perform correlation analysis on each parameter in the set of influencing parameters, and a parameter mapping relationship under the deviation state derivation is established, thereby quantifying the degree of deviation anomaly of each influencing parameter.

[0043] like Figure 3As shown, the deviation identification module, when deriving the deviation state, further includes the following implementation method: using the standard value range of any parameter in the set of influencing parameters within a unit time as a reference benchmark, wherein the unit time can be set according to production needs, for example, selecting 1 minute as the unit; simultaneously, the currently checked cable length can also be adjusted according to the actual production process. Based on this, the residual value corresponding to each influencing parameter is calculated. This residual value is used to characterize the degree of deviation between the actual fluctuation value of the parameter and the preset predicted value, thereby reflecting the difference between the current operating state and the ideal operating state.

[0044] The specific calculation method for the currently identified residual value is as follows: Based on the standard range of the influencing parameter within a unit of time, the moving average of the influencing parameter within that unit of time is calculated. The difference between the moving average corresponding to the data fluctuation moment and the preset predicted value is then calculated. The resulting difference is the residual value of the influencing parameter at the current moment. The subsequent fluctuation ratio is defined as the percentage of the residual value relative to the predicted value.

[0045] The predicted value is selected by determining it based on the average parameter value of the corresponding device under normal operating conditions in historical data.

[0046] The residual values ​​of each influencing parameter are combined to form multiple candidate sets; association rule mining is performed on the candidate sets at any time, and the data that satisfies the association rule is regarded as the current inferred deviation state; when combining residual values, the residual values ​​of all parameters are first sorted by absolute value, and the parameter value of each parameter is normalized so that the fluctuation of the parameter belongs to the range of -1 to 1.

[0047] The candidate set of combinations represents data combinations that satisfy both temporal and process constraints. For example, a combination of upstream process parameters (melting zone temperature, extrusion speed, melt pressure) → midstream process parameters (forming zone vacuum degree, die angle) → downstream process parameters (setting zone temperature, cooling water temperature) requires that the upstream and downstream parameters occur within corresponding time windows. For instance, setting a time window W only mines abnormal upstream parameters between t−W and t, and abnormal downstream parameters between t and t+W. Furthermore, the process constraints indicate a combination rule where a residual occurs upstream followed by a residual downstream.

[0048] If we take into account the priority of the parameter set at the input, we can also use the data corresponding to each priority level as the upstream parameter to be identified first, and determine the downstream parameter that appears in the residual together with it, so as to complete the combined output of the residual terms.

[0049] When performing association rule mining on the candidate set at any given time, the implementation method also includes: traversing the time points where residuals appear for any influencing parameter, and merging the time points where residuals appear consecutively into an abnormal time period.

[0050] During abnormal periods, the influencing parameters are paired according to the composition of upstream and downstream parameters to extract all data combinations. Then, the frequency of occurrence of data combinations is filtered based on the parts of the data combinations that satisfy the time and process constraints. The filtered data is regarded as the candidate set. At this time, the frequency of occurrence is equivalent to the support in association rule mining, that is, the frequency of the corresponding parameters appearing at the same time. For any two parameters, under the process and time constraints from upstream to downstream parameters, data combinations with higher occurrence frequencies are selected, such as combinations with an occurrence frequency ≥ 10%.

[0051] Gradually expand the number of data items in the candidate itemset and simultaneously filter them according to their confidence level to obtain frequent itemsets.

[0052] Association rules are generated from frequent itemsets, and the influence parameters corresponding to the association rules are synchronized to the aperture anomaly to complete the setting of the mapping relationship.

[0053] The progressively expanding number of data items represents increasing from 2 to 3 data items, and so on, to continuously mine association rules on the set of parameters influencing the current input. The confidence level used represents the probability of an outcome occurring after a given condition is met; for example, if a combination of candidate options is high extrusion speed + low linear speed, then its confidence level represents the support of high extrusion speed + low linear speed / the support of high extrusion speed, to determine its relative overall probability of occurrence. Confidence level filtering applies not only to data with 2 data items but also to data with 3 data items, retaining data with a confidence level greater than 75% to select the most frequently occurring values ​​in the corresponding scenario. Here, 75% is only used as an example to illustrate the point of obtaining a larger amount of data.

[0054] The implementation method of obtaining the deviation anomaly degree of the set of influencing parameters in the deviation identification module also includes: calculating the fluctuation ratio corresponding to the influencing parameter, performing time-series curve fitting on the fluctuation ratio of each influencing parameter, and determining the trend coefficient of the influencing parameter at each time based on the slope value of the fitted curve at the corresponding time. When fitting the curve, the fluctuation ratio of each parameter is regarded as the vertical axis and its corresponding timestamp is regarded as the horizontal axis, thereby recording the relative tendency at each time.

[0055] When calculating the trend coefficient, Empirical Mode Decomposition (EMD) can be used. The decomposition process can be represented as follows: Identify all local extrema (peaks / troughs) of the fitted curve; fit the upper and lower envelopes using cubic splines; calculate the average of the envelopes, subtract the original sequence to obtain the first-order mode vector, and repeat this process until the remaining data of the fitted curve satisfies monotonicity. This yields multiple sets of mode vectors and the trend term. The trend coefficient is then represented as the average slope of the trend term.

[0056] By treating the trend coefficient as a multi-factor variable, and utilizing the mapping relationship between each parameter in the influencing parameter set and the aperture anomaly, the similarity is calculated using the trend coefficients corresponding to the influencing parameters, and the output similarity value is regarded as the deviation anomaly degree of the influencing parameter set.

[0057] Since the mapping relationship of the set of influencing parameters is based on association rules, it can include multiple parameters with different dimensions. At this time, the similarity calculation will be based on the similarity between the current trend coefficient and historical data. The similarity of each influencing parameter can be calculated using the Pearson correlation coefficient. The average of the similarities of the multiple influencing parameters corresponding to the current mapping relationship is taken as the deviation anomaly of the current output.

[0058] The larger the deviation anomaly value, the more similar the current residual-induced operating state is to the orifice anomaly operating state, indicating a more serious deviation from the normal situation. It is necessary to simultaneously correct the parameters of multiple steps to reduce the risk of orifice anomaly. If the value is smaller, it means that the correlation with the current orifice anomaly operating state is lower. It is necessary to check the state according to the real-time fluctuation ratio to determine the interference situation under multi-parameter collaborative control.

[0059] In one embodiment of the present invention, in the collaborative control module, based on the deviation situation corresponding to the above-mentioned set of influencing parameters, the fluctuation amplitude when the deviation occurs is emphasized, and the fluctuation ratio is regarded as the current abnormal change ratio in the form of absolute value; then, according to the value of the deviation anomaly degree, the deviation interference under real-time production is further reflected.

[0060] like Figure 4 As shown, the implementation of the collaborative control module also includes: dividing the deviation anomaly range into multiple deviation anomaly intervals; if the current deviation anomaly is greater than a preset threshold, matching the corresponding deviation anomaly interval; retrieving parameters according to the fluctuation ratio within each deviation anomaly interval; and extracting preliminary control targets from the matched historical cases. The preset threshold can be set according to the average deviation anomaly in historical data to identify scenarios with large and small deviations. The deviation anomaly interval can be divided into three or more intervals proportionally to the current value range to determine the parameters that need to be controlled within each interval when the deviation anomaly value is too large, by narrowing the search range.

[0061] When retrieving parameters based on the fluctuation ratio within each deviation anomaly range, the implementation method also includes: performing a fast search based on the deviation anomaly range and the type of aperture anomaly to obtain a list of historical cases; the list of historical cases will use the type of specific aperture anomaly + deviation anomaly as the search range to determine the historical cases that can be solved in the current scenario.

[0062] Using the fluctuation ratio of each influencing parameter within the deviation anomaly range as the retrieval feature, a similarity matching retrieval is performed on the historical case list to determine a subset of cases that meet the matching requirements; introducing the fluctuation ratio involves selecting cases with the same fluctuation ratio as the current parameter, which can be determined using a B+ tree retrieval method.

[0063] Initial process parameter ranges and cable batches are introduced. Cases matching the current scenario are selected from a subset of cases, and the parameters corresponding to those cases are used as the initial control targets. Finally, cases within the same scenario are restricted according to the current initial process parameter range and cable batches to derive the control targets corresponding to the current scenario from these cases.

[0064] If the deviation anomaly is not greater than the preset threshold, the parameters are retrieved based on the fluctuation ratio at the corresponding time to extract the preliminary control target; if it is not greater than the preset threshold, it means that the deviation anomaly is small at this time, and the parameters can be retrieved according to the current scenario. The retrieval method is the same as the retrieval method for the deviation anomaly interval mentioned above.

[0065] When the initial control target corresponds to multiple sets of influencing parameters, the least common subset of the initial control targets is regarded as the current output control target; if the initial control target corresponds to only one set of influencing parameters, the corresponding initial control target is regarded as the current control target.

[0066] Based on the forming range corresponding to the control target, and combined with the absolute value of the difference between the deviation anomaly of the current control target and the average deviation anomaly, a corresponding adjustment factor is configured. The average deviation anomaly is calculated based on the deviation anomalies of historical data belonging to the same scenario as the current control target, and its average value is used as a benchmark to characterize the degree of deviation of the current control target from the normal operating state.

[0067] The adjustment factor is defined as the absolute value of the difference between the deviation anomaly and the average deviation anomaly, and its adjustment range is determined according to the deviation ratio of the control target. After considering the differences in the impact of different forming intervals on the process, a coefficient value corresponding to the forming interval is further introduced. This coefficient value is used to characterize the influence weight of different forming intervals, and its value range can be calibrated based on historical data, for example, set between 0.5 and 0.7. Since the deviation anomaly is essentially a measure of trend similarity, after calculating its deviation from the average scenario, combined with the forming interval coefficient, the adjustment factor value applicable to the current interval can be obtained, specifically expressed as the product of the absolute value of the difference and the forming interval coefficient.

[0068] Furthermore, considering the differences in parameter types, corresponding weighting coefficients can be set according to the ratio of the fluctuation ratio of each parameter to the corresponding data combination, and the adjustment factor can be weighted to obtain the weighted adjustment factor as the final output.

[0069] The adjustment factor specifies the intensity of parameter adjustment. For example, if the current line speed adjustment factor is 0.03, it means that the line speed setting will be adjusted by ±3%, and the adjustment direction is determined by the type of corresponding aperture anomaly. Alternatively, the adjustment factors of all parameters can be sorted according to their values ​​to determine the parts that need to be adjusted first in the current scenario.

[0070] In one embodiment of the present invention, the parameter feedback module is used to associate the adjustment factor with the aperture anomaly location, and to achieve a combined output of the adjustment point and the anomaly location by means of location point association.

[0071] Specifically, such as Figure 5 As shown, the parameter feedback module includes the following implementation: displaying the location points corresponding to each adjustment factor, and combining continuously distributed location points into abnormal aperture segments; wherein, adjacent location points with a distance less than a preset threshold (e.g., 1 meter) are considered continuous and merged.

[0072] For each abnormal aperture segment, its numerical mutation region is identified, and it is divided into several sub-segments of equal length for cluster analysis. The equal-length sub-segment clustering refers to dividing the abnormal aperture segment into equal lengths (e.g., 1 meter). If the segment length is 5 meters, it is divided into 5 equal-length sub-segments.

[0073] For each abnormal aperture segment, the adjustment direction of the initial process parameters is determined by using the adjustment factor corresponding to each sub-segment within the cluster. The initial process parameters are then adjusted according to the determined adjustment direction, and the process is continuously verified until the production results meet expectations.

[0074] During the clustering process, clustering algorithms such as K-means are used to calculate the data of each sub-segment. The average and standard deviation of the adjustment factor of each sub-segment, or the characteristic value of aperture anomaly, are used as the clustering feature vectors to cluster each sub-segment. By calculating the silhouette coefficient of the cluster, the clustering result corresponding to the largest silhouette coefficient is selected as the optimal cluster. The larger the silhouette coefficient, the denser the cluster structure. This cluster is the main data object analyzed in the current scenario.

[0075] For each cluster, the average value within the cluster is used as the basis for adjustment direction; the abnormal apertures within each cluster are processed in descending order of the average value within the cluster, and the final adjustment direction is determined by combining convolution operation.

[0076] When performing adjustment direction analysis, interactive convolution operations are performed based on the currently extracted adjustment factors to verify the adjustment direction. Specifically, the method for determining the adjustment direction of the initial process parameters includes: performing convolution operations on adjacent abnormal aperture segments based on the adjustment factors corresponding to the current abnormal aperture segment, and using the vector direction obtained from the convolution interaction as the current output adjustment direction. During the convolution operation, a third-order sliding window convolution kernel is used, and the convolution dimension is selected as a combination of the feature value corresponding to the aperture anomaly and the magnitude of the adjustment factor; the current abnormal aperture segment is divided into three units of equal length, and the convolution values ​​of its front, middle, and rear units with the corresponding units of adjacent abnormal segments are calculated to verify the matching degree between the parameter adjustment direction and the aperture change. If the convolution value is positive, it indicates a match; if it is negative, it indicates a mismatch. Only when all convolution values ​​are positive is the currently calculated vector direction output as the adjustment direction; otherwise, the adjustment direction needs to be re-determined according to the single abnormal aperture segment processing method.

[0077] The feature value corresponding to the aperture anomaly can be selected according to actual needs. For example, the aperture deviation value can be used as the feature value, or a quantitative index used to characterize the degree of anomaly under different anomaly types can be used to achieve a quantitative description of the convolution calculation target.

[0078] In addition, if the parameters corresponding to the adjustment factors involved in adjacent abnormal segments (such as die angle, vacuum degree, linear speed, segment temperature value, extrusion speed, etc.) only appear in isolation in a certain segment, then the single abnormal aperture segmentation method is adopted, that is, the value of the adjustment factor at the corresponding time is used as the content of convolution calculation.

[0079] When an anomalous aperture segment has no adjacent segments, a stepped convolution operation is performed on that single anomalous aperture segment. The stepped convolution refers to dividing a single anomalous aperture segment into multiple equal-length units and performing the convolution operation by moving one unit at a time in a sliding manner. After all adjustment factors have completed the convolution interaction, principal component analysis is performed on the convolution interaction results, and the resulting principal component directions are used as the adjustment direction of the current output.

[0080] For cases where there is only a single anomalous aperture segment, since the data sources are similar, the convolution output of the previous segment can be used as the convolution input of the next segment for recursive calculation. The convolution vector of each unit of equal length can be extracted by principal component analysis, and its principal component direction can be used as the adjustment direction of the current convolution interaction.

[0081] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention, which are still covered within the protection scope of the present invention.

Claims

1. A high-frequency, high-speed transmission cable forming control system, characterized in that, include: The data acquisition module is used to divide the production line into multiple forming sections according to the cable production process and obtain the initial process parameters of each forming section. These initial process parameters include the die angle, vacuum degree, linear speed, segment temperature value and extrusion speed when the cable is extruded. The parameter monitoring module is used to monitor the initial process parameters of the cable based on the abnormality of the cable aperture, and to determine the set of influencing parameters when the cable is continuously extruded. The deviation identification module is used to deduce the deviation status based on the current set of influence parameters of the cable, determine the mapping relationship between each parameter in the set of influence parameters and the aperture anomaly, and obtain the deviation anomaly degree of the set of influence parameters. Other methods for deriving deviation states include: Using the standard range of any parameter in the set of influencing parameters within a unit of time as the verification benchmark, determine the residual value corresponding to each influencing parameter; The residual values ​​of each influencing parameter are combined to form multiple candidate sets; association rule mining is performed on the candidate sets at any time, and the data that satisfies the association rules are regarded as the current deduced deviation state. Other methods for obtaining the deviation anomaly degree of the influencing parameter set include: Calculate the fluctuation ratio corresponding to the influencing parameter, fit a time series curve to the fluctuation ratio of each influencing parameter, and determine the trend coefficient of the influencing parameter at each time point based on the slope value of the fitted curve at the corresponding time point. By using the trend coefficient as a multi-factor variable, the mapping relationship between each parameter in the influencing parameter set and the aperture anomaly is utilized. The similarity is calculated using the trend coefficient corresponding to the influencing parameter, and the output similarity value is regarded as the deviation anomaly degree of the influencing parameter set. The collaborative control module is used to take the process parameters corresponding to each forming interval as the control target, use the deviation anomaly degree of the parameter set as the index, and combine the fluctuation ratio at the corresponding time to determine the adjustment factor of the collaborative control. The implementation methods of the collaborative control module also include: The deviation anomaly is divided into multiple deviation anomaly intervals based on the range of deviation anomaly values. If the current deviation anomaly is greater than the preset threshold, the corresponding deviation anomaly interval is matched, and the parameters are retrieved according to the fluctuation ratio within each deviation anomaly interval. The preliminary control target is extracted from the matched historical cases. If the deviation anomaly is not greater than the preset threshold, the parameter is retrieved based on the fluctuation ratio at the corresponding time to extract the preliminary control target. When there are multiple sets of influencing parameters corresponding to the initial control target, the least common subset of the initial control targets is regarded as the current output control target; Based on the forming range corresponding to the control target, and combined with the absolute value of the difference between the deviation anomaly of the current control target and the average deviation anomaly, the adjustment factor is configured. The parameter feedback module is used to configure abnormal aperture segments based on the adjustment factor of the collaborative control, the location point of the abnormal aperture, and the initial process parameters of the corresponding forming interval are adjusted according to the numerical change of the abnormal aperture segment. The implementation methods of the parameter feedback module also include: Display the location points corresponding to each adjustment factor, and combine consecutive location points into abnormal aperture segments; For each anomalous aperture segment, the numerical abrupt change is clustered into equal-length sub-segments. For each abnormal aperture segment, the adjustment direction of the initial process parameters is determined by using the adjustment factor corresponding to the cluster; based on the obtained adjustment direction, the initial process parameters are adjusted.

2. The high-frequency, high-speed transmission cable forming control system according to claim 1, characterized in that, When determining the initial process parameters for segmented extrusion in the data acquisition module, the implementation methods also include: Using the extrusion aperture as the verification parameter, when an abnormality occurs in the aperture, the extrusion sequence of the cable is combined simultaneously to statistically analyze the range of values ​​for die angle, vacuum degree, linear speed, segment temperature, and extrusion speed at each segment. The data is then grouped according to the molding interval corresponding to the initial process parameters as the data for the current analysis.

3. The high-frequency, high-speed transmission cable forming control system according to claim 1, characterized in that, When determining the set of influencing parameters during continuous cable extrusion in the parameter monitoring module, the implementation methods also include: Based on the process flow corresponding to each forming section, reverse the current initial process parameters and extract multiple sets of target parameters. Based on the time sequence of aperture anomaly triggering, compare the time of aperture anomaly with the fluctuation of each target parameter. If there is a correspondence between the time of target parameter fluctuation and the location of aperture anomaly, the corresponding target parameter will be preliminarily identified as the influencing parameter. Matching analysis is performed based on the pore size anomaly type corresponding to the influencing parameters, and the set of influencing parameters is considered as the output based on the priority of the matched parameters.

4. The high-frequency, high-speed transmission cable forming control system according to claim 1, characterized in that, When performing association rule mining on a candidate set at any given time, the implementation methods also include: Iterate through the time points when residuals appear for any influencing parameter, and merge the time points when residuals appear consecutively into an abnormal period. The impact parameters during the abnormal period are matched in pairs according to the composition of upstream and downstream parameters. All data combinations are extracted, and the frequency of occurrence of data combinations is filtered according to the part of the data combination that satisfies the time constraints and process constraints. The filtered data is regarded as the candidate set. Gradually expand the number of data items in the candidate itemset and simultaneously filter them according to the confidence level of the candidate itemset to obtain frequent itemsets; Association rules are generated from frequent itemsets, and the influence parameters corresponding to the association rules are synchronized to the aperture anomaly to complete the setting of the mapping relationship.

5. The high-frequency, high-speed transmission cable forming control system according to claim 1, characterized in that, When retrieving parameters based on the fluctuation ratio within each deviation anomaly interval, the implementation methods also include: Quickly retrieve a list of historical cases by analyzing the deviation anomaly range and the type of aperture anomaly. Using the fluctuation ratio of each influencing parameter within the deviation anomaly range as the retrieval feature, a similarity matching retrieval is performed on the historical case list to determine a subset list of cases that meet the matching requirements. Introduce initial process parameter ranges and cable batches, select cases that match the current scenario from the case subset list, and use the corresponding parameters in the cases as the current initial control targets.

6. The high-frequency, high-speed transmission cable forming control system according to claim 1, characterized in that, When determining the direction of adjustment for initial process parameters, the implementation methods also include: Based on the adjustment factor of the current abnormal aperture segment, convolution operation is performed on adjacent abnormal aperture segments, and the direction of their convolutional interaction vector is regarded as the adjustment direction of the current output. When there are no adjacent abnormal aperture segments, a stepwise convolution operation is performed on a single abnormal aperture segment, and after all adjustment factors have completed the convolution interaction, the principal component direction of the convolution interaction result is used as the adjustment direction of the current output.

Images