Cable production quality traceability method and system based on process data flow

Abstract

The application discloses a cable production quality tracing method and system based on process data flow, and relates to the technical field of production quality management.The method comprises the following steps: S1, periodically collecting cable production process data, and performing pretreatment on the cable production process data; S2, generating a cable quality comprehensive evaluation value, constructing a cable quality early warning threshold prediction model, outputting a cable quality early warning threshold, comparing the cable quality comprehensive evaluation value with the cable quality early warning threshold, and generating normal and abnormal production data sets; S3, evaluating the influence of cross-layer process parameter fluctuation, comparing the evaluation result with a fluctuation threshold, and generating an abnormal alarm data set; S4, generating an abnormal report, triggering cable production quality tracing, generating a tracing record, storing the tracing record together with the abnormal report, and forming a production quality tracing archive.The method solves the problem that cross-layer process parameter fluctuation and abnormality are difficult to identify in the prior art, and quality control is not timely.

Description

Technical Field

[0001] This invention relates to the field of production quality management technology, specifically to a method and system for tracing cable production quality based on process data flow. Background Technology

[0002] With the development of modern industrial manufacturing technology, quality control and traceability in the cable production process have become increasingly important. Especially in the entire cable production process, which involves complex data and information flows across multiple production steps and stages, a comprehensive and accurate quality traceability method is needed to ensure that every stage of production meets requirements. The quality of the cable production process is not only affected by each technological step but also by the combined effects of various industrial mechanisms, including the physical properties of materials, heat conduction processes, and pressure control. In recent years, many technical solutions have been proposed to achieve quality monitoring and traceability throughout the entire cable production process, comprehensively analyzing the relationship between various process parameters and industrial mechanisms, thereby improving the accuracy and efficiency of quality control.

[0003] For example, the invention with publication number CN121352625A relates to a method and system for full-process quality traceability management in cable production. The method includes: acquiring dynamic production information of the cable to be tested during the production process; determining the health value of target parameters for the target process in cable production; determining the dynamic coupling weight between two different production parameters in the target process based on the health value of the production parameters and the correlation coefficient between the production parameters; determining the dynamic sensitivity factor of the target parameters based on the health value and the dynamic coupling weight between different parameters, thereby determining the process health index of the target process; and determining the net defect contribution value of the target process based on the process health index of subsequent processes, so as to utilize the sum of the net defect contribution values ​​of multiple processes to perform quality traceability of the cable production process under test.

[0004] For example, the invention with publication number CN116151843A discloses a digital management system for the entire process quality of ultra-high voltage cable joints, including a material data monitoring module, which is connected to a material information module. The material information module is connected to a material library, which is connected to a joint list module. The joint list module is connected to a construction personnel information module and an environmental control module. The joint list module and the environmental control module are connected to a third-party review module. The third-party review module performs third-party review on the content of the construction project, etc. The third-party review module is connected to a database through a blockchain technology data transmission module. The approved project content is transmitted to the database for storage through the blockchain technology data transmission module.

[0005] However, while the aforementioned technical solutions have improved quality traceability in cable production to some extent, some significant shortcomings remain. Existing technologies typically record parameters for each production process separately, such as extruder pressure, temperature, and traction speed, but lack systematic analysis of the interactions between multiple processes and fluctuations in cross-layer parameters. Particularly in the production of double- or multi-layer co-extruded cables, the high sensitivity of interface formation to "interlayer temperature windows, temperature gradient changes, and pressure fluctuations" makes it impossible for quality inspection of a single process to effectively predict and control overall quality problems. Existing methods often fail to achieve real-time quality monitoring and early warning across processes, and are also difficult to accurately trace back to specific production stages and equipment status.

[0006] Therefore, in order to address the above issues, there is an urgent need for a cable production quality traceability method and system based on process data flow. Summary of the Invention

[0007] Technical problems to be solved

[0008] To address the shortcomings of existing technologies, this invention provides a cable production quality traceability method and system based on process data flow, which solves the problem that it is difficult to identify cross-layer process parameter fluctuations and anomalies in cable production, leading to untimely quality control.

[0009] Technical solution

[0010] To achieve the above objectives, the present invention provides the following technical solution: a cable production quality traceability method based on process data flow, comprising: S1, periodically collecting cable production process data, and performing time alignment, dynamic smoothing, anomaly removal, missing data completion, and standardization on the cable production process data to generate preprocessed cable production process data; S2, evaluating cable quality based on the preprocessed cable production process data, generating a comprehensive cable quality evaluation value, constructing a cable quality early warning threshold prediction model, outputting a cable quality early warning threshold, comparing the comprehensive cable quality evaluation value and the cable quality early warning threshold, marking normal and abnormal records, and generating normal and abnormal production datasets; S3, extracting process parameters based on the abnormal production dataset, and evaluating the impact of cross-layer process parameter fluctuations by combining the correlation coefficients of process parameters between different layers of cables, comparing the evaluation results with the fluctuation thresholds to determine whether the production process is normal, and generating an abnormal alarm dataset; S4, reading the real-time abnormal alarm dataset, generating an abnormal report, triggering cable production quality traceability, generating traceability records, and storing the traceability records and abnormal reports together to form a production quality traceability archive.

[0011] Furthermore, the specific steps for periodically collecting cable production process data and performing time alignment, dynamic smoothing, anomaly removal, missing data completion, and standardization on the cable production process data to generate preprocessed cable production process data are as follows: A fixed-width sliding time window is set as one sampling period, and production process data for each layer of cable is periodically collected. The cable production process data includes cable number, collection timestamp, extruder pressure, extrusion section temperature, cooling section temperature, cooling water flow rate, cooling section pressure, interlayer interface pressure, and traction speed. For the collected cable production process data, all data are mapped to the same time axis based on the sampling timestamp to complete time alignment. A weighted moving average algorithm is used to perform smoothing filtering on the cable production process data. A local outlier detection algorithm is used to detect and remove abnormal sampling points in the cable production process data. For missing segments formed after removal, a Lagrange interpolation algorithm is used for interpolation completion. Finally, a maximum-minimum standardization algorithm is used to perform numerical standardization on the cable production process data, outputting the preprocessed cable production process data.

[0012] Furthermore, the specific steps for evaluating cable quality and generating a comprehensive cable quality assessment value based on the preprocessed cable production process data are as follows: Read the preprocessed cable production process data, arrange it in ascending order by sampling timestamp, and concatenate all cable production process data corresponding to the same sampling timestamp into a single sampling record; assign a unique sampling record number to each sampling record; for each sampling record, calculate the average values ​​of cooling section temperature, interlayer interface pressure, cooling water flow rate, traction speed, cooling section pressure, and extruder pressure for all layers of cable; simultaneously, calculate the difference between the cooling section temperature in the current sampling period and the cooling section temperature in the previous sampling period, and divide it by the sampling period time to obtain the cooling section temperature gradient; calculate the difference between the cooling section pressure in the current sampling period and the cooling section pressure in the previous sampling period, and divide it by the sampling period time to obtain the cooling section pressure gradient; based on the average cooling section temperature, average interlayer interface pressure, average cooling water flow rate, average traction speed, average cooling section pressure, average extruder pressure, cooling section temperature gradient, and cooling section pressure gradient, a comprehensive cable quality assessment value is calculated.

[0013] Further, the specific steps for comprehensively calculating the overall cable quality assessment value are as follows: Divide the difference between the average cooling section temperature and the reference cooling temperature by the cooling section temperature, use the resulting ratio as the exponent, and use the temperature influence coefficient as the exponent to perform a power function operation to obtain the temperature difference influence term; Add one to the average interlayer interface pressure, take the natural logarithm, and then multiply by the temperature difference influence term to obtain the temperature and pressure interaction influence term; Divide the product of the average cooling water flow rate and the average traction speed by the square of the difference between the average cooling section pressure and the average extruder pressure, add one, use the resulting ratio as the base, and use the pressure influence coefficient as the exponent to perform an exponential function operation to obtain the cooling traction efficiency influence term; Divide the cooling section temperature gradient by the sum of the cooling section pressure gradient and a constant to obtain the temperature and pressure gradient influence term; Add the above four items together to obtain the overall cable quality assessment value.

[0014] Furthermore, the specific steps for constructing a cable quality early warning threshold prediction model and outputting the cable quality early warning threshold are as follows: A gradient boosting regression algorithm is used to construct the cable quality early warning threshold prediction model. The cable production process data and comprehensive cable quality evaluation values ​​corresponding to the most recent N sample records are read to construct N sample records, which are then trained and arranged according to the sample record number. The previous... The sample records are used as the training set to input into the cable quality early warning threshold prediction model, and the model is trained; then... Each sample record is input into the trained cable quality early warning threshold prediction model, and the output is the cable quality early warning threshold.

[0015] Further, the specific steps for comparing the comprehensive cable quality assessment value and the cable quality early warning threshold, marking normal and abnormal records, and generating normal and abnormal production datasets are as follows: The comprehensive cable quality assessment value and the cable quality early warning threshold are compared in real time. When the comprehensive cable quality assessment value is greater than or equal to the cable quality early warning threshold, the current sampled record is marked as a normal record, and the corresponding sampled record number, cable production process data, comprehensive cable quality assessment value, and cable quality early warning threshold are extracted and stored together to construct a normal production dataset. When the comprehensive cable quality assessment value is less than the cable quality early warning threshold, the current sampled record is marked as an abnormal record, and the corresponding sampled record number, cable production process data, comprehensive cable quality assessment value, and cable quality early warning threshold are extracted and stored together to construct an abnormal production dataset.

[0016] Furthermore, based on the abnormal production dataset, process parameters are extracted, and the correlation coefficients between process parameters of each layer of cable are combined to evaluate the impact of cross-layer process parameter fluctuations. The specific steps are as follows: Read the abnormal production dataset. For each abnormal record, extract the cooling section temperature, interlayer interface pressure, and traction speed corresponding to each layer of cable, and record them as the process parameter set. Use the Pearson correlation coefficient to calculate the correlation coefficient between the process parameters of each layer of cable and the process parameters of the previous layer of cable. Take the absolute value of the difference between the i-th process parameter of the current layer and the i-th process parameter of the previous layer, and divide it by the sum of the absolute value of the i-th process parameter of the previous layer and the minima to obtain the variation range of the i-th process parameter. Multiply the variation range of the i-th process parameter by the correlation coefficient between the current layer and the i-th process parameter of the previous layer to obtain the fluctuation impact value of the i-th process parameter. Add up the fluctuation impact values ​​of all process parameters to obtain the cross-layer process parameter fluctuation impact value of the current layer.

[0017] Further, the specific steps for comparing the evaluation results with the fluctuation threshold to determine whether the production process is normal and to generate an abnormal alarm dataset are as follows: Calculate the cross-layer process parameter fluctuation impact value of all layers of cables within the same abnormal record and compare it with the fluctuation threshold in real time; when the cross-layer process parameter fluctuation impact value is less than or equal to the fluctuation impact value, the production process of the current layer cable is determined to be normal; when the cross-layer process parameter fluctuation impact value is greater than the fluctuation impact value, the production process of the current layer cable is determined to be abnormal, an alarm signal is generated, and the corresponding alarm signal generation time, sampling record number, cable number, process parameter set, and cross-layer process parameter fluctuation impact value are extracted and associated for storage to construct an abnormal alarm dataset.

[0018] Furthermore, the specific steps for reading the real-time anomaly alarm dataset, generating anomaly reports, triggering cable production quality traceability, generating traceability records, and storing the traceability records together with the anomaly reports to form a production quality traceability archive are as follows: Receive the real-time anomaly alarm dataset; extract cable production process data, comprehensive cable quality assessment values, and cable quality early warning thresholds from the K most recent sampled records based on the alarm signal generation time; extract the process parameter set from the K most recent sampling periods based on the cable number, the correlation coefficient between the current layer and the i-th process parameter of the previous layer, and the cross-layer process parameter fluctuation impact value; automatically generate anomaly reports; trigger cable production quality traceability; trace back to the specific production time period, production equipment, production batch, machine number, and work order number by associating with historical cable production process data; generate traceability records; store the traceability records together with the anomaly reports to form a complete production quality traceability archive.

[0019] The second aspect of this invention provides a cable production quality traceability system based on process data flow, comprising: a data acquisition and preprocessing module, a process data analysis and quality assessment module, a cross-layer process parameter fluctuation calculation module, and a cable production quality traceability module. The data acquisition and preprocessing module is used to periodically acquire cable production process data and perform time alignment, dynamic smoothing, anomaly removal, missing data completion, and standardization on the cable production process data to generate preprocessed cable production process data. The process data analysis and quality assessment module is used to assess cable quality based on the preprocessed cable production process data, generate a comprehensive cable quality assessment value, and construct a cable quality early warning threshold prediction model. The system outputs cable quality early warning thresholds, compares the comprehensive cable quality assessment value with the cable quality early warning thresholds, marks normal and abnormal records, and generates normal and abnormal production datasets. The cross-layer process parameter fluctuation calculation module extracts process parameters based on the abnormal production dataset, and, combined with the correlation coefficients of process parameters between different cable layers, assesses the impact of cross-layer process parameter fluctuations. It compares the assessment results with the fluctuation thresholds to determine whether the production process is normal and generates an abnormal alarm dataset. The cable production quality traceability module reads the real-time abnormal alarm dataset, generates an abnormal report, triggers cable production quality traceability, generates traceability records, and stores the traceability records along with the abnormal report to form a production quality traceability archive.

[0020] Beneficial effects

[0021] The present invention has the following beneficial effects:

[0022] (1) A method and system for tracing cable production quality based on process data flow. By comprehensively evaluating the multi-layer process parameters of cable production, a quality early warning method based on the fluctuation of cross-layer process parameters is proposed. This method can accurately identify factors that may lead to interface adhesion risks, such as uneven cooling rate and pressure fluctuation. This provides an innovative solution for real-time quality monitoring and risk identification in cable production.

[0023] (2) The cable production quality traceability method and system based on process data flow, by combining advanced data processing algorithms such as weighted moving average, Lagrange interpolation, and maximum and minimum value standardization, has constructed a more accurate cable quality assessment model, which can effectively eliminate the influence of noise, improve the accuracy of quality assessment, and provide a scientific basis for the accurate prediction of cable quality early warning threshold.

[0024] (3) A cable production quality traceability method and system based on process data flow, by comparing the comprehensive evaluation value of cable quality with the early warning threshold of cable quality in real time, marks each sampling record in the cable production process as abnormal and generates an abnormal alarm dataset, which significantly improves the transparency and traceability of the production process and facilitates rapid response and corrective measures.

[0025] (4) A cable production quality traceability method and system based on process data flow effectively reduces production anomalies caused by quality fluctuations by real-time monitoring and correlation of cross-layer process parameter fluctuations, interface bonding risks and quality assessments, ensuring the stability and consistency of cable production, significantly reducing rework costs and the generation of defective products, and improving overall production efficiency and product qualification rate. Attached Figure Description

[0026] Figure 1 A flowchart of a cable production quality traceability method based on process data flow;

[0027] Figure 2 This is a structural diagram of a cable production quality traceability system based on process data flow.

[0028] Figure 3 This is a chart for judging abnormal records based on the comprehensive evaluation value of cable quality.

[0029] Figure 4 Flowchart for generating cable quality early warning thresholds. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] Please see Figures 1-4 This invention provides a technical solution: a cable production quality traceability method based on process data flow, comprising: S1, periodically collecting cable production process data, and performing time alignment, dynamic smoothing, anomaly removal, missing data completion, and standardization on the cable production process data to generate preprocessed cable production process data; S2, evaluating cable quality based on the preprocessed cable production process data, generating a comprehensive cable quality evaluation value, constructing a cable quality early warning threshold prediction model, outputting a cable quality early warning threshold, comparing the comprehensive cable quality evaluation value and the cable quality early warning threshold, marking normal and abnormal records, and generating normal and abnormal production datasets; S3, extracting process parameters based on the abnormal production dataset, and evaluating the impact of cross-layer process parameter fluctuations by combining the correlation coefficients of process parameters between different layers of cables, comparing the evaluation results with the fluctuation thresholds to determine whether the production process is normal, and generating an abnormal alarm dataset; S4, reading the real-time abnormal alarm dataset, generating an abnormal report, triggering cable production quality traceability, generating traceability records, and storing the traceability records and abnormal reports together to form a production quality traceability archive.

[0032] Specifically, the process of periodically collecting cable production process data and performing time alignment, dynamic smoothing, anomaly removal, missing data completion, and standardization on the data to generate pre-processed cable production process data is as follows: First, a fixed-width sliding time window is set as a sampling period, specifically 10 minutes. Production process data for each layer of cable is collected periodically every 10 minutes. For the specific data collection of each layer of cable, it is ensured that the sensors for each layer are customized and configured according to the specific process and procedure requirements of the production line, thereby ensuring that the quality data of each layer of cable can be collected accurately and in a timely manner, providing detailed data for subsequent process optimization and quality traceability. Within each sampling period, the cable production process data includes cable number, collection timestamp, extruder pressure, extrusion section temperature, cooling section temperature, cooling water flow rate, cooling section pressure, interlayer interface pressure, and traction speed, ensuring that all process data are completely recorded each time data is collected. Data collection is performed in real time through various sensors installed on the production equipment. Specifically: the cable number is automatically recorded by the production line control system to identify each cable. The collection timestamp is recorded by the system's clock, and each sampling occurs at a specific point in the cable production process. Extruder pressure is collected in real time by a pressure sensor installed at the extruder to measure the pressure of the material inside. Extrusion section temperature is collected by a thermocouple temperature sensor installed in the extrusion section to monitor the temperature of the material as it passes through. Cooling section temperature is collected by a temperature sensor installed on the cooling section to detect temperature changes in real time and ensure the stability of the cooling process. Cooling water flow rate is collected by a flow meter installed on the cooling water pipeline to measure the flow rate of the cooling water and ensure the efficiency of the cooling system. Cooling section pressure is collected by a pressure sensor installed at the inlet and outlet of the cooling section to monitor the pressure of the cooling system in real time and ensure that the cooling process meets process requirements. Interlayer interface pressure is collected by a dedicated pressure sensor installed between the layers during cable extrusion to monitor the interface pressure between different material layers and ensure the quality of interface bonding. Traction speed is collected by a speed sensor installed on the traction device to monitor the cable traction speed and ensure that the traction process is smooth and matches other processes. For the collected cable production process data, all data is first mapped to the same time axis based on the sampling timestamp to complete time alignment. Time alignment ensures that data from different sources can be compared and analyzed according to a unified time scale, thereby eliminating errors caused by inconsistent sampling times. Next, a weighted moving average algorithm is used to smooth the cable production process data, aiming to reduce high-frequency noise and improve data stability. This algorithm assigns different weights to each sample value and adjusts the degree of data smoothing according to the time window size, thereby effectively removing abnormal fluctuations and improving data reliability.For detected outlier sampling points, a local outlier detection algorithm is used. First, outliers with significant differences compared to neighboring data points are identified and removed to avoid interference with subsequent data analysis and quality assessment. For missing segments resulting from the removal, Lagrange interpolation is used for interpolation completion. This algorithm estimates missing values ​​using polynomial interpolation of known data points, ensuring data integrity. Finally, a maximum-minimum standardization algorithm is used to numerically standardize the cable production process data. The standardization process first calculates the maximum and minimum values ​​for each data point. By subtracting the minimum value from each data point and dividing by the difference between the maximum and minimum values, the data is scaled to a range of 0 to 1. This helps eliminate the influence between different dimensions, allowing comparisons of data on the same scale. During this standardization process, all process data are converted to dimensionless, ensuring that unit differences between different sensors or measuring equipment do not affect subsequent analysis.

[0033] In this implementation plan, the periodic collection and preprocessing of cable production process data significantly improves the accuracy and consistency of the data. Time alignment, smoothing filtering, anomaly removal, missing data completion, and standardization are performed on the collected process data. This not only ensures the integrity and reliability of the data but also eliminates differences between different data sources and units, allowing for comparison and analysis of various data under the same standard. Furthermore, the dimensionless data after standardization enhances the accuracy of subsequent quality assessment and risk analysis, effectively supporting the correlation and in-depth analysis of cross-layer data. This provides a more accurate basis for quality traceability, optimization adjustments, and anomaly early warning in the cable production process.

[0034] Specifically, the steps for evaluating cable quality and generating a comprehensive cable quality assessment value based on pre-processed cable production process data are as follows: First, read the pre-processed cable production process data, arrange it in ascending order by sampling timestamp, and concatenate all cable production process data corresponding to the same sampling timestamp into a single sampling record. A unique sampling record number is assigned to each record for subsequent tracking and analysis. For each sampling record, calculate the average values ​​of cooling section temperature, interlayer interface pressure, cooling water flow rate, traction speed, cooling section pressure, and extruder pressure for all layers of cable. These average values ​​represent the production process level of each layer of cable and directly affect the stability of cable quality. Temperature, pressure, flow rate, and speed are the most critical process parameters in cable production, directly affecting the cable's structural strength, thermal stability, and final electrical performance. The cooling section temperature controls the cooling rate of the cable material; excessively high or low temperatures will affect the cable's mechanical strength and electrical insulation performance. Interlayer pressure directly determines the bonding strength between different material layers; insufficient pressure can lead to interlayer separation, affecting the cable's long-term stability. Cooling water flow rate and traction speed affect the cable's forming and stretching processes, thus influencing the final physical properties of the cable. Cooling section pressure controls the external pressure distribution during the cooling process. Extruder pressure determines the material's flowability and uniformity during extrusion. These factors collectively affect the cable's final performance. To ensure data accuracy, all collected values ​​undergo rigorous verification and integration. Simultaneously, the difference between the cooling section temperature in the current sampling period and the previous sampling period is calculated and divided by the sampling period time to obtain the cooling section temperature gradient, reflecting the rate of temperature change and directly affecting the cable's thermal stability and final quality. The temperature gradient reflects the rate of temperature change during the cable's cooling process; an excessively large temperature gradient can lead to uneven cooling of the cable's internal materials, thus affecting its physical properties. Similarly, the difference between the cooling section pressure in the current sampling period and the cooling section pressure in the previous sampling period is calculated, and then divided by the sampling period time to obtain the cooling section pressure gradient, revealing the rate of change of cooling pressure. Changes in the pressure gradient can affect the bonding strength between the outer and inner layers of the cable. If the cooling section pressure gradient is too large, it may cause interlayer separation or uneven cooling, leading to unstable overall cable quality. Based on the average cooling section temperature, average interlayer interface pressure, average cooling water flow rate, average traction speed, average cooling section pressure, average extruder pressure, cooling section temperature gradient, and cooling section pressure gradient, these data are calculated using a comprehensive algorithm to obtain a comprehensive cable quality assessment value. Through the comprehensive evaluation of these process data, the impact of various process control parameters on cable quality during cable production can be quantified, providing a reliable quantitative basis for subsequent quality prediction, risk identification, and production process optimization.This comprehensive assessment helps to identify potential quality problems in the cable production process in advance, providing a scientific basis for optimizing production processes and improving cable quality.

[0035] This implementation plan effectively integrates the process parameters and key indicators of each layer of cable by comprehensively evaluating the pre-treated cable production process data. It accurately calculates the average values ​​of variables such as cooling section temperature, interlayer interface pressure, cooling water flow rate, traction speed, cooling section pressure, and extruder pressure, as well as the cooling section temperature gradient and cooling section pressure gradient. This comprehensive data provides a complete and accurate quantitative basis for cable quality assessment. By correlating these process parameters with the comprehensive cable quality assessment value, it is possible to fully reflect quality fluctuations during the production process, enhance the monitoring capability of cable quality, and provide data support for future adjustments and optimizations in the production process, thereby effectively improving the stability and reliability of the production process.

[0036] Specifically, the steps for calculating the comprehensive evaluation value of cable quality are as follows: First, divide the difference between the average cooling section temperature and the reference cooling temperature by the cooling section temperature. Use the resulting ratio as an exponent, and then perform a power function operation using the temperature influence coefficient as the exponent to obtain the temperature difference influence term. The temperature influence coefficient is obtained by analyzing the relationship between cooling section temperature and cable quality in historical cable production process data. Specifically, the least squares regression method is used to fit the relationship between cooling section temperature and finished cable quality to calculate the temperature influence coefficient. This coefficient can quantify the impact of temperature changes on cable quality. Through the nonlinear characteristics of the power function, the impact of temperature changes on quality can be significantly amplified, especially when temperature fluctuations are large, making the contribution of temperature changes to cable quality more prominent. Next, add one to the average interlayer interface pressure, take the natural logarithm, and then multiply by the temperature difference influence term to obtain the temperature and pressure interaction influence term. To avoid excessive influence of large interlayer interface pressure values ​​on quality evaluation, the pressure value is logarithmically transformed. The pressure influence coefficient is obtained by analyzing the relationship between interlayer interface pressure and cable quality in historical cable production process data. Specifically, linear regression analysis is used to calculate the pressure influence coefficient based on the correlation between changes in interlayer interface pressure and cable quality in the cable production process data. By performing a natural logarithmic transformation on the pressure data, the nonlinear impact of pressure data on the final evaluation value can be effectively reduced, while maintaining the smoothness of the temperature and pressure interaction effects. Then, the product of the average cooling water flow rate and the average traction speed is divided by the square of the difference between the average cooling section pressure and the average extruder pressure, plus one. Using the resulting ratio as the base, and with the pressure influence coefficient as the exponent, an exponential function is performed to obtain the cooling traction efficiency influence term. In this part, the complex relationship between cooling water flow rate, traction speed, and pressure is modeled using an exponential function to ensure that the impact of cooling traction efficiency on cable quality is effectively amplified when the cooling section pressure and traction speed are high. The use of the exponential function ensures the sensitivity of this term when cooling and traction efficiencies are low, highlighting the contribution of these process parameters to cable quality. The exponential regression model is used to calculate the cooling traction efficiency influence term, thus ensuring that the nonlinear relationship between the cooling and traction processes is accurately expressed. Next, the temperature gradient of the cooling section is divided by the sum of the pressure gradient of the cooling section and a constant to obtain the temperature and pressure gradient influence term. This term reflects the influence of temperature and pressure changes through the ratio between the rate of temperature change and the rate of pressure change in the cooling section. The constant is used to ensure numerical stability during the calculation process and avoid unreasonable fluctuations in the results due to large temperature and pressure gradients. By calculating the ratio of the temperature gradient to the pressure gradient, the combined effects of temperature and pressure changes in the cooling section on cable quality during cable production can be effectively simulated. Finally, the above four terms are added together to obtain the comprehensive cable quality assessment value.By comprehensively considering factors such as temperature differences, pressure interactions, traction efficiency, and temperature-pressure gradients, nonlinear calculation methods such as power functions and exponential functions are used in cable quality assessment, which can more accurately reflect the complex process interactions in cable production.

[0037] The specific formula for calculating the comprehensive evaluation value of cable quality is as follows:

[0038] ;

[0039] In the formula, This represents the overall quality assessment value of the cable. This represents the average temperature of the cooling section. Indicates the reference cooling temperature. Indicates the temperature effect coefficient. This represents the average pressure at the interlayer interface. This represents the average cooling water flow rate. This represents the average traction speed. This represents the average pressure in the cooling section. Indicates the pressure influence coefficient. This represents the average extruder pressure. Indicates the temperature gradient of the cooling section. This indicates the pressure gradient in the cooling section.

[0040] In this embodiment, Table 1 shows the cable production process data and the corresponding comprehensive cable quality assessment results for five sampling records. Specifically: Sampling record 1: Cooling section temperature 34.5°C, interlayer interface pressure 3.8 rpm, cooling water flow rate 65 kW, traction speed 28 rpm, cooling section pressure 12.2 rpm, extruder pressure 13.6 rpm, cooling section temperature gradient 0.04 rpm, cooling section pressure gradient 0.02 rpm, comprehensive cable quality assessment value 0.240. Sampling record 2: Cooling section temperature 33.2°C, interlayer interface pressure 4.5°C, cooling water flow rate 70 kW, traction speed 30 rpm, cooling section pressure 10.6 rpm, extruder pressure 13.1 rpm, cooling section temperature gradient 0.02 rpm, cooling section pressure gradient 0.01 rpm, comprehensive cable quality assessment value 0.189. Sampling Record 3: Cooling section temperature 31.0℃, interlayer interface pressure 2.9℃, cooling water flow rate 60 L / min, traction speed 26 rpm, cooling section pressure 11.6 L / min, extruder pressure 13.0℃, cooling section temperature gradient -0.03℃, cooling section pressure gradient 0.04℃, overall cable quality assessment value 0.060. Sampling Record 4: Cooling section temperature 36.8℃, interlayer interface pressure 5.2℃, cooling water flow rate 75 L / min, traction speed 32 rpm, cooling section pressure 12.8℃, extruder pressure 14.2℃, cooling section temperature gradient 0.06℃, cooling section pressure gradient 0.03℃, overall cable quality assessment value 0.360. Sampling Record 5: Cooling section temperature 30.5°C, interlayer interface pressure 3.2 ppm, cooling water flow rate 55 liters per second, traction speed 24 km / h, cooling section pressure 11.4 liters per second, extruder pressure 12.5 liters per second, cooling section temperature gradient -0.01, cooling section pressure gradient 0.00, and overall cable quality assessment value 0.067. The above data is used to quantitatively assess the quality during cable production and provides a data basis for subsequent comparison between the overall cable quality assessment value and the cable quality early warning threshold to determine normal and abnormal records. It also supports subsequent process optimization and quality traceability analysis.

[0041] Table 1. Comprehensive Evaluation Values ​​of Cable Quality

[0042]

[0043] like Figure 3The chart shows the comprehensive cable quality assessment values ​​and normal / abnormal judgment results for five sampling records. The bar chart uses different colors to distinguish the record status: blue bars represent normal records, and orange bars represent abnormal records. A black dashed line marks the cable quality warning threshold (0.18), used for threshold comparison of the comprehensive cable quality assessment values ​​for each sampling record. The top of each bar is labeled with its corresponding comprehensive cable quality assessment value, facilitating a direct comparison of the quality assessment levels of different sampling records. Specifically, the comprehensive cable quality assessment values ​​for sampling records 1, 2, and 4 are 0.240, 0.189, and 0.360, respectively, all not less than the cable quality warning threshold, corresponding to blue bars and marked as normal records; the comprehensive cable quality assessment values ​​for sampling records 3 and 5 are 0.060 and 0.067, respectively, both less than the cable quality warning threshold, corresponding to orange bars and marked as abnormal records. Figure 3 By using a visualization method that combines column color partitioning, threshold dashed lines, and evaluation value annotations, the comparison results between the comprehensive evaluation value of cable quality corresponding to the sampling record number and the cable quality early warning threshold are presented in a structured manner. This provides a directly referable data basis for subsequent extraction of process parameters based on abnormal production datasets, assessment of the impact of cross-layer process parameter fluctuations, generation of abnormal alarm datasets, and triggering cable production quality traceability.

[0044] This implementation plan, by combining key process parameters such as cooling section temperature, interlayer interface pressure, cooling water flow rate, and traction speed, accurately quantifies the comprehensive impact of these parameters on cable quality. Using least squares regression and linear regression analysis, the temperature and pressure influence coefficients are precisely obtained. Furthermore, through nonlinear calculations using power and exponential functions, the deeper impact of process parameter changes on cable quality is captured. This multi-dimensional quality assessment method can more comprehensively reflect process changes in cable production and maintain high sensitivity and accuracy under different industry conditions. It provides a scientific basis for quality control and risk warning during production, effectively improving the controllability and stability of the production process.

[0045] Specifically, the steps for constructing a cable quality early warning threshold prediction model and outputting the cable quality early warning threshold are as follows: Figure 4As shown, the process reads the cable production process data and comprehensive cable quality assessment value corresponding to the most recent N sampling records. The data includes cable number, collection timestamp, extruder pressure, cooling section temperature, cooling water flow rate, interlayer interface pressure, traction speed, cooling section pressure, etc., as well as the comprehensive cable quality assessment value. These data serve as input features and target values. The specific range of N is usually determined based on the actual sampling frequency and the time span of the data. Generally, N can be between 100 and 1000, and can be flexibly adjusted according to actual production conditions. N sample records are constructed using this data and arranged according to the sampling record number to ensure the consistency of the training data order and the completeness of the samples. Then, the previous... A set of sample records is used as the training set to input into the cable quality early warning threshold prediction model for training. During training, the gradient boosting regression algorithm calculates the contribution of each feature data in each sample record to the cable quality early warning threshold, and improves the accuracy of the prediction results through a stepwise optimization method. The training process includes model initialization, tree structure establishment, weighted adjustment in each round, calculation and minimization of the loss function, etc., and reduces the prediction error through multiple iterations to ensure high-precision prediction of the cable quality early warning threshold. In the training process, a basic prediction model is first initialized, and the error of each iteration is calculated by least squares regression or mean squared error loss function. In each round, the gradient boosting regression algorithm calculates the residual between the current model and the actual target value, and builds a new decision tree based on the residual. The error is continuously reduced by updating the tree structure and split points. In each iteration, the new decision tree is added to the model with weights, gradually improving the model's prediction accuracy. Appropriate regularization techniques are also used during training to prevent the model from overfitting. By limiting the maximum depth of each decision tree, the minimum number of sample splits, and adjusting hyperparameters such as the learning rate, the model is ensured to generalize well on different datasets. The learning rate controls the model's contribution to the update in each round. An excessively large learning rate may lead to excessively large errors, while an excessively small learning rate may make the training process too slow. To ensure the model's stability and effectiveness, methods such as cross-validation are used to evaluate the model's performance on different datasets, and hyperparameters are further optimized. After training, the subsequent... Each sample record is input into a pre-trained cable quality early warning threshold prediction model, which outputs the cable quality early warning threshold. Through multiple rounds of training and optimization using the gradient boosting regression algorithm, the prediction accuracy of the cable quality early warning threshold can be significantly improved, providing strong data support for quality monitoring and early warning in the cable production process, and ensuring more precise and effective quality control during production.

[0046] In this implementation scheme, the gradient boosting regression algorithm can accurately predict cable quality early warning thresholds, thus providing effective early warnings for quality control during cable production. This method analyzes historical production data and establishes a high-precision prediction model by utilizing the relationship between cable production process data and comprehensive cable quality assessment values. The gradient boosting regression algorithm is continuously optimized through multiple iterations, gradually reducing errors to ensure accurate prediction of the cable quality early warning threshold. By outputting this early warning threshold, potential quality problems during production can be identified in advance, facilitating timely corrective measures and improving production stability and product quality. This high-precision early warning mechanism can significantly reduce quality fluctuations during production, improving overall production efficiency and product qualification rate.

[0047] Specifically, the steps for comparing the comprehensive cable quality assessment value and the cable quality early warning threshold, marking normal and abnormal records, and generating normal and abnormal production datasets are as follows: First, the comprehensive cable quality assessment value and the cable quality early warning threshold are compared in real time. When the comprehensive cable quality assessment value is greater than or equal to the cable quality early warning threshold, the current sampling record is marked as a normal record. At this point, by checking the relationship between the comprehensive cable quality assessment value and the cable quality early warning threshold, it is ensured that the sampling record is only marked as a normal record when the quality assessment value reaches or exceeds the early warning threshold. Next, the corresponding sampling record number, cable production process data (including cooling section temperature, traction speed, extruder pressure, interlayer interface pressure, etc.), the comprehensive cable quality assessment value, and the cable quality early warning threshold are extracted and stored in association. This data will be summarized and stored in the normal production dataset for subsequent quality traceability, analysis, and improvement. Second, when the comprehensive cable quality assessment value is less than the cable quality early warning threshold, the current sampling record is marked as an abnormal record. In this case, the cable quality assessment value fails to meet the predetermined quality standard, therefore the sampling record is identified as abnormal, indicating potential quality hazards in the production process. Next, the corresponding sampling record numbers, cable production process data, comprehensive cable quality assessment values, and cable quality early warning thresholds are extracted and stored together. This data will be stored in the abnormal production dataset. These abnormal records provide direct data support for subsequent quality improvement, helping to identify potential production problems in a timely manner, optimize production processes, and reduce the production of defective products. Through this process, all sampling records are clearly categorized, ensuring accurate separation between normal and abnormal production data, and providing a precise basis for subsequent quality analysis, optimization, and process adjustments.

[0048] This implementation plan effectively distinguishes between normal and abnormal production records by comparing the comprehensive cable quality assessment value and the cable quality early warning threshold in real time, ensuring timely identification of quality issues during the production process. The clear separation of normal and abnormal production data not only facilitates subsequent quality traceability and analysis but also provides precise data for quality optimization and process adjustment. Detailed tracking and storage of each sampling record not only improves the traceability of production data but also enhances quality control capabilities during the production process. This allows potential quality hazards to be detected early, reducing the generation of non-conforming products and promoting continuous improvement of the production process.

[0049] Specifically, the steps for evaluating the impact of cross-layer process parameter fluctuations based on the extraction of process parameters from the abnormal production dataset and the correlation coefficients of process parameters between different cable layers are as follows: First, the abnormal production dataset is read. For each abnormal record, the cooling section temperature, interlayer interface pressure, and traction speed corresponding to each cable layer are extracted and recorded as a set of process parameters. Each record contains data on key process parameters of different cable layers during production. These sets of process parameters comprehensively reflect the various process fluctuations during cable production. Next, to accurately capture the correlation between process parameter changes between different layers during cable production, the Pearson correlation coefficient is used to calculate the correlation between the process parameters of each cable layer and the process parameters of the previous cable layer. When calculating the Pearson correlation coefficient, it is necessary to take sequence data within a time window, not just the value of a single abnormal record. Therefore, the process parameter set for each abnormal record should consider a data sequence within a certain time range, rather than just based on a single sampling point. This means that the process parameter sequence of each cable layer needs to be extracted within a given time window, and the Pearson correlation coefficient between different layers needs to be calculated based on these sequences. In this way, the relationship between the changes in process parameters of different cable layers throughout the entire production cycle can be more accurately reflected. The Pearson correlation coefficient quantifies the linear relationship between process parameters of cables in different layers, providing a reliable mathematical basis for subsequent cross-layer process parameter fluctuation assessment. This process ensures that the correlation between process parameters is accurately captured and provides an effective quantitative tool for assessing the impact of cross-layer process fluctuations. Next, the absolute value of the difference between the i-th process parameter of the current layer and the i-th process parameter of the previous layer is taken and divided by the sum of the absolute value of the i-th process parameter of the previous layer and the minimum term to obtain the variation range of the i-th process parameter. This step effectively measures the difference between the current layer and the previous layer of cable under the same process parameters. The minimum term is a small but non-zero positive real number used to avoid numerical instability caused by division by zero during calculation; its value range is... arrive Subsequently, the variation amplitude of the i-th process parameter is multiplied by the correlation coefficient between the i-th process parameter of the current layer and the i-th process parameter of the previous layer to obtain the fluctuation impact value of the i-th process parameter. This calculation integrates the correlation between the variation amplitude of process parameters and the variation of process parameters between layers, yielding the specific contribution value of each process parameter to cable quality fluctuations. Finally, the fluctuation impact values ​​of all process parameters are summed to obtain the cross-layer process parameter fluctuation impact value of the current layer. This comprehensive calculation allows for a comprehensive assessment of the fluctuation effects of different process parameters on cable production quality, accurately identifies the potential impact of process parameter fluctuations between different layers, and thus provides a reliable basis for quality control and optimization of subsequent production processes.

[0050] The specific formula for calculating the impact value of cross-layer process parameter fluctuations is as follows:

[0051] ;

[0052] In the formula, This indicates the impact value of cross-layer process parameter fluctuations in the current layer. This represents the i-th process parameter of the current layer. This represents the i-th process parameter of the previous layer. This represents the correlation coefficient between the current layer and the i-th process parameter of the previous layer. Indicates a minus term.

[0053] In this implementation plan, by assessing the cross-layer fluctuation impact of process parameters in the abnormal production dataset, the correlation between process parameters of each layer of the cable and the impact of their variation on cable quality can be accurately identified. By quantifying the relationship between process parameters across layers using the Pearson correlation coefficient, and by calculating the variation range and fluctuation impact value of process parameters, the combined effect of process parameter fluctuations during the production of different layers of cables can be comprehensively evaluated. This method provides a quantitative basis for in-depth analysis of the interactions between various process stages in the production process, helps to accurately identify potential process instabilities, and provides scientific support for subsequent quality control and production optimization, ensuring more stable and reliable cable product quality.

[0054] Specifically, the steps for comparing the evaluation results with the fluctuation threshold to determine whether the production process is normal and generating an anomaly alarm dataset are as follows: First, calculate the cross-layer process parameter fluctuation impact value for all layers of cables within the same anomaly record and compare it with the fluctuation threshold in real time. The cross-layer process parameter fluctuation impact value, calculated using the aforementioned method, reflects the degree of fluctuation in process parameters between different layers and can effectively measure the coordination and stability between different process stages in cable production. Compare the cross-layer process parameter fluctuation impact value for all layers of cables in each anomaly record with the set fluctuation threshold to determine whether the production process of each layer of cable meets quality standards. When the cross-layer process parameter fluctuation impact value is less than or equal to the fluctuation impact value, the production process of the current layer of cable is determined to be normal. In this case, the fluctuation of the cable production process is within the normal range, indicating that the changes in process parameters during production have not had a significant impact on cable quality, and therefore the current production process can be considered normal. Conversely, when the cross-layer process parameter fluctuation impact value is greater than the fluctuation impact value, the production process of the current layer of cable is determined to be abnormal. At this time, the fluctuation of process parameters exceeds the normal range, which may indicate potential quality hazards in the production process, and the production process needs further optimization or adjustment. When an anomaly is detected in the cable manufacturing process, an alarm signal is generated to promptly alert relevant personnel to potential quality issues during production. Upon generating an alarm signal, the system extracts and stores the corresponding alarm signal generation time, sampling record number, cable number, process parameter set, and cross-layer process parameter fluctuation impact value. By aggregating and storing this key information, an anomaly alarm dataset is constructed, providing data support for subsequent quality analysis and production process optimization. This dataset not only helps identify problems in the production process but also provides detailed records for quality traceability and problem resolution, contributing to improved controllability and stability of the cable manufacturing process.

[0055] In this implementation scheme, by comparing the impact value of cross-layer process parameter fluctuations with the fluctuation threshold in real time, potential anomalies in the cable production process can be accurately identified. This method can promptly detect fluctuations in the production process that exceed the normal range, providing early warnings for potential quality problems during production. By generating alarm signals and associating them with detailed sampling record numbers, cable numbers, and process parameter sets, this invention achieves rapid location and traceability of abnormal production records. This mechanism significantly improves the controllability of the production process, ensures continuous monitoring of cable quality during production, and provides scientific data for quality improvement.

[0056] Specifically, the steps for reading the real-time anomaly alarm dataset, generating anomaly reports, triggering cable production quality traceability, generating traceability records, and storing the traceability records along with the anomaly reports to form a production quality traceability archive are as follows: First, receive the real-time anomaly alarm dataset. Based on the alarm signal generation time, extract the cable production process data, comprehensive cable quality assessment value, and cable quality early warning threshold from the K most recent sampled records. The specific value of K is typically between 50 and 200, depending on the data acquisition frequency and the length of the production cycle. Second, extract the process parameter set from the K most recent sampling periods based on the cable number, including the correlation coefficient between the current layer and the i-th process parameter of the previous layer, and the cross-layer process parameter fluctuation impact value. This data provides a comprehensive basis for subsequent anomaly analysis. Third, when automatically generating an anomaly report, extract all relevant data from the anomaly alarm dataset, including sampling record number, cable production process data, comprehensive cable quality assessment value, cable quality early warning threshold, process parameter set, correlation coefficient, cross-layer process parameter fluctuation impact value, etc., and perform summary analysis. The report details the specific information of the anomaly. The report helps to quickly identify potential quality problems in the cable production process and guides production personnel to make timely adjustments. Subsequently, cable production quality traceability is triggered. By linking historical cable production process data, the process traces back to the specific production time period, equipment, batch, machine number, and work order number. This step, through in-depth analysis of historical data, ensures that the specific stage at which the problem occurred can be traced, identifying potential root causes and providing data support for production optimization. When generating traceability records, the records include key data such as production time period, equipment information, and operating parameters, helping to quickly locate potential problems in the production process. The traceability records are stored together with anomaly reports to form a complete production quality traceability archive. This archive contains the entire process from cable production process data to anomaly alarms, quality assessments, and production traceability, ensuring the transparency and traceability of the production process and providing a reliable data foundation for subsequent quality improvement, problem solving, and process optimization.

[0057] In this implementation plan, by generating anomaly reports based on real-time anomaly alarm datasets and triggering cable production quality traceability, comprehensive quality monitoring and traceability throughout the production process can be achieved. By precisely correlating cable production process data, comprehensive cable quality assessment values, and cable quality early warning thresholds, and combining historical data for in-depth backtracking, this invention effectively improves the transparency and traceability of the production process. In this process, industrial mechanisms in cable production, such as temperature changes, pressure fluctuations, and material responses, are also incorporated into the quality analysis and traceability framework, ensuring a close link between process parameter changes and quality control. The generated traceability records not only help identify potential quality problems but also accurately pinpoint the stages where problems occur, thus providing strong data support for production optimization. This process provides detailed evidence for subsequent quality analysis, process adjustments, and problem solving, greatly improving the stability and quality controllability of cable production.

[0058] like Figure 2 As shown, the second aspect of this invention provides a cable production quality traceability system based on process data flow, including: a data acquisition and preprocessing module, a process data analysis and quality assessment module, a cross-layer process parameter fluctuation calculation module, and a cable production quality traceability module. The data acquisition and preprocessing module is used to periodically acquire cable production process data and perform time alignment, dynamic smoothing, anomaly removal, missing data completion, and standardization on the cable production process data to generate preprocessed cable production process data. The process data analysis and quality assessment module is used to assess cable quality based on the preprocessed cable production process data, generate a comprehensive cable quality assessment value, and construct a cable quality early warning threshold prediction model. The system outputs cable quality early warning thresholds, compares the comprehensive cable quality assessment value with the cable quality early warning thresholds, marks normal and abnormal records, and generates normal and abnormal production datasets. The cross-layer process parameter fluctuation calculation module is used to extract process parameters based on the abnormal production dataset, and evaluate the impact of cross-layer process parameter fluctuations by combining the correlation coefficients of process parameters between each layer of cables. The evaluation results are compared with the fluctuation thresholds to determine whether the production process is normal, and an abnormal alarm dataset is generated. The cable production quality traceability module is used to read the real-time abnormal alarm dataset, generate abnormal reports, trigger cable production quality traceability, generate traceability records, and store the traceability records together with the abnormal reports to form a production quality traceability archive.

[0059] This implementation plan achieves precise monitoring of the entire cable production process, from data acquisition to quality traceability, by combining data acquisition and preprocessing, process data analysis and quality assessment, cross-layer process parameter fluctuation calculation, and cable production quality traceability. Through systematic data processing and analysis methods, cable quality can be accurately assessed, abnormal fluctuations in the process can be identified in a timely manner, and the source of problems can be quickly located through the traceability mechanism. In this process, industrial mechanisms in cable production, such as heat conduction, pressure control, and material adhesion, are all incorporated into the analysis framework to ensure the accuracy of process control. This not only improves the transparency and traceability of the production process but also strengthens the quality control capabilities of cable production, ensures the stability of product quality, and provides reliable data for optimization and adjustment in the production process.

[0060] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0061] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A cable production quality traceability method based on process data flow, characterized in that, Includes the following steps: S1 periodically collects cable production process data and performs time alignment, dynamic smoothing, anomaly removal, missing data completion and standardization on the cable production process data to generate pre-processed cable production process data; S2. Based on the preprocessed cable production process data, the cable quality is evaluated, a comprehensive cable quality evaluation value is generated, a cable quality early warning threshold prediction model is constructed, a cable quality early warning threshold is output, the comprehensive cable quality evaluation value and the cable quality early warning threshold are compared, normal and abnormal records are marked, and normal and abnormal production datasets are generated. S3: Extract process parameters based on abnormal production datasets, and combine the correlation coefficients of process parameters between each layer of cables to evaluate the impact of cross-layer process parameter fluctuations. Compare the evaluation results with the fluctuation thresholds to determine whether the production process is normal and generate an abnormal alarm dataset. The specific steps for evaluating the impact of cross-layer process parameter fluctuations by extracting process parameters based on abnormal production datasets and combining the correlation coefficients of process parameters between different cable layers are as follows: Read the abnormal production dataset. For each abnormal record, extract the cooling section temperature, interlayer interface pressure, and traction speed corresponding to each layer of cable, and record them as a set of process parameters. Then, use the Pearson correlation coefficient to calculate the correlation coefficient between the process parameters of each layer of cable and the process parameters of the previous layer of cable. Take the absolute value of the difference between the i-th process parameter in the current layer and the i-th process parameter in the previous layer, and divide it by the sum of the absolute value of the i-th process parameter in the previous layer and the minima to obtain the variation range of the i-th process parameter. Then multiply the variation range of the i-th process parameter by the correlation coefficient between the i-th process parameter of the current layer and the i-th process parameter of the previous layer to obtain the fluctuation impact value of the i-th process parameter. Add up the fluctuation impact values ​​of all process parameters to obtain the cross-layer process parameter fluctuation impact value of the current layer; The specific steps for comparing the evaluation results with the fluctuation threshold to determine whether the production process is normal and generating an abnormal alarm dataset are as follows: Calculate the cross-layer process parameter fluctuation impact value of all layers of cables within the same anomaly record and compare it with the fluctuation threshold in real time; when the cross-layer process parameter fluctuation impact value is less than or equal to the fluctuation impact value, the production process of the current layer cable is determined to be normal. When the influence value of cross-layer process parameter fluctuation is greater than the fluctuation influence value, the production process of the current layer cable is determined to be abnormal, an alarm signal is generated, and the corresponding alarm signal generation time, sampling record number, cable number, process parameter set and cross-layer process parameter fluctuation influence value are extracted and stored together to construct an abnormal alarm dataset. S4 reads the real-time abnormal alarm dataset, generates an abnormal report, and triggers cable production quality traceability to generate traceability records. The traceability records are stored together with the abnormal report to form a production quality traceability archive.

2. The cable production quality traceability method based on process data flow according to claim 1, characterized in that: The specific steps for periodically collecting cable production process data and performing time alignment, dynamic smoothing, anomaly removal, missing data completion, and standardization on the cable production process data to generate preprocessed cable production process data are as follows: Set a fixed-width sliding time window as a sampling period to periodically collect the production process data of each layer of cable. The cable production process data includes cable number, collection timestamp, extruder pressure, extrusion section temperature, cooling section temperature, cooling water flow rate, cooling section pressure, interlayer interface pressure, and traction speed. For the collected cable production process data, all data are mapped to the same time axis based on the sampling timestamp to complete time alignment; A weighted moving average algorithm was used to perform smoothing filtering on the cable production process data; A local outlier detection algorithm is used to detect and remove abnormal sampling points in the cable production process data. For the missing segments formed after removal, the Lagrange interpolation algorithm is used to interpolate and complete them. Then, the maximum and minimum value standardization algorithm is used to perform numerical standardization processing on the cable production process data, and the preprocessed cable production process data is output.

3. The cable production quality traceability method based on process data flow according to claim 1, characterized in that: The specific steps for evaluating cable quality based on preprocessed cable production process data and generating a comprehensive cable quality assessment value are as follows: Read the preprocessed cable production process data, sort it in ascending order by sampling timestamp, and combine all cable production process data corresponding to the same sampling timestamp into a single sampling record; and add a unique sampling record number to each sampling record. For each sampling record, calculate the average values ​​of cooling section temperature, interlayer interface pressure, cooling water flow rate, traction speed, cooling section pressure, and extruder pressure for all layers of cable; simultaneously, calculate the difference between the cooling section temperature in the current sampling period and the cooling section temperature in the previous sampling period, and then divide it by the sampling period time to obtain the cooling section temperature gradient; calculate the difference between the cooling section pressure in the current sampling period and the cooling section pressure in the previous sampling period, and then divide it by the sampling period time to obtain the cooling section pressure gradient; The comprehensive evaluation value of cable quality is calculated based on the average temperature of the cooling section, the average pressure of the interlayer interface, the average flow rate of cooling water, the average traction speed, the average pressure of the cooling section, the average pressure of the extruder, the temperature gradient of the cooling section, and the pressure gradient of the cooling section.

4. The cable production quality traceability method based on process data flow according to claim 3, characterized in that: The specific steps for obtaining the comprehensive evaluation value of cable quality through comprehensive calculation are as follows: The temperature difference effect term is obtained by dividing the difference between the average cooling section temperature and the reference cooling temperature by the cooling section temperature, using the resulting ratio as the exponent, and the temperature influence coefficient as the exponent, and performing a power function operation. The temperature difference effect term is obtained by adding one to the average interlayer interface pressure, taking the natural logarithm, and then multiplying it by the temperature difference effect term. The temperature and pressure interaction effect term is obtained by dividing the product of the average cooling water flow rate and the average traction speed by the square of the difference between the average cooling section pressure and the average extruder pressure, plus one, using the resulting ratio as the base, and the pressure influence coefficient as the exponent, and performing an exponential function operation. The cooling traction efficiency effect term is obtained by dividing the cooling section temperature gradient by the sum of the cooling section pressure gradient and a constant. The temperature and pressure gradient effect term is obtained by adding the above four terms together. The comprehensive evaluation value of cable quality is obtained.

5. The cable production quality traceability method based on process data flow according to claim 1, characterized in that: The specific steps for constructing the cable quality early warning threshold prediction model and outputting the cable quality early warning threshold are as follows: A gradient boosting regression algorithm is used to construct a cable quality early warning threshold prediction model. The model is built by reading the cable production process data and comprehensive cable quality assessment values ​​corresponding to the most recent N sample records, constructing N sample records, and arranging them according to the sample record number. The sample records are used as the training set to input into the cable quality early warning threshold prediction model, and the model is trained; then... Each sample record is input into the trained cable quality early warning threshold prediction model, and the output is the cable quality early warning threshold.

6. The cable production quality traceability method based on process data flow according to claim 1, characterized in that: The specific steps for comparing the comprehensive evaluation value of cable quality and the early warning threshold of cable quality, marking normal and abnormal records, and generating normal and abnormal production datasets are as follows: The system compares the comprehensive cable quality assessment value and the cable quality early warning threshold in real time. When the comprehensive cable quality assessment value is greater than or equal to the cable quality early warning threshold, the current sampling record is marked as a normal record. The corresponding sampling record number, cable production process data, comprehensive cable quality assessment value and cable quality early warning threshold are extracted and stored together to construct a normal production dataset. When the comprehensive evaluation value of cable quality is less than the cable quality early warning threshold, the current sampling record is marked as an abnormal record, and the corresponding sampling record number, cable production process data, comprehensive evaluation value of cable quality and cable quality early warning threshold are extracted and stored together to construct an abnormal production dataset.

7. The cable production quality traceability method based on process data flow according to claim 1, characterized in that: The specific steps for reading the real-time anomaly alarm dataset, generating an anomaly report, triggering cable production quality traceability, generating traceability records, and storing the traceability records together with the anomaly report to form a production quality traceability archive are as follows: The system receives real-time anomaly alarm datasets and extracts cable production process data, comprehensive cable quality assessment values, and cable quality early warning thresholds from the K most recent sampled records based on the alarm signal generation time. It also extracts the process parameter set from the K most recent sampling periods based on the cable number, the correlation coefficient between the current layer and the i-th process parameter of the previous layer, and the cross-layer process parameter fluctuation impact value, automatically generating an anomaly report. Furthermore, it triggers cable production quality traceability, retracing back to specific production time periods, equipment, batches, machine numbers, and work order numbers by associating with historical cable production process data, generating traceability records. These traceability records are stored together with the anomaly report to form a complete production quality traceability archive.

8. A cable production quality traceability system based on process data flow, employing the cable production quality traceability method based on process data flow as described in any one of claims 1-7, characterized in that, include: The module includes a data acquisition and preprocessing module, a process data analysis and quality assessment module, a cross-layer process parameter fluctuation calculation module, and a cable production quality traceability module, among which: The data acquisition and preprocessing module is used to periodically acquire cable production process data and perform time alignment, dynamic smoothing, anomaly removal, missing data completion and standardization on the cable production process data to generate preprocessed cable production process data. The process data analysis and quality assessment module is used to assess cable quality based on preprocessed cable production process data, generate a comprehensive cable quality assessment value, construct a cable quality early warning threshold prediction model, output a cable quality early warning threshold, compare the comprehensive cable quality assessment value and the cable quality early warning threshold, mark normal and abnormal records, and generate normal and abnormal production datasets. The cross-layer process parameter fluctuation calculation module is used to extract process parameters based on the abnormal production dataset, and combine the correlation coefficients of process parameters between each layer of cables to evaluate the impact of cross-layer process parameter fluctuations. The evaluation results are compared with the fluctuation threshold to determine whether the production process is normal, and an abnormal alarm dataset is generated. The cable production quality traceability module is used to read the real-time abnormal alarm dataset, generate an abnormal report, trigger cable production quality traceability, generate traceability records, and store the traceability records together with the abnormal reports to form a production quality traceability archive.

Images