Self-adaptive control method for quality fluctuation in wire extrusion process

Abstract

The invention provides a self-adaptive control method for quality fluctuation in a wire extrusion process. The self-adaptive control method comprises the following steps: acquiring the thickness of an insulating layer, the thickness of a sheath layer, the outer diameter size and surface quality data at an outlet of an extruder, and performing digital signal processing to obtain a standardized thickness uniformity index; performing classification judgment according to the standardized thickness uniformity index, and if the standardized thickness uniformity index deviates from a preset threshold value, marking the standardized thickness uniformity index as a quality fluctuation event; determining a fluctuation type according to the quality fluctuation event; according to the optimized temperature set value, parameters of an extruder control system are updated, and the operation state is fed back; determining the adjustment effectiveness according to the comparison between the operation state and an initial standardized thickness uniformity index; after the adjustment is confirmed to be effective, collecting the batch diameter of the raw material and environment temperature and humidity data and fusing to obtain a comprehensive environment influence factor; and judging whether the traction speed needs to be subjected to secondary adjustment or not according to the comprehensive environmental influence factor, obtaining a final stable extrusion process parameter group, and transmitting the final stable extrusion process parameter group back to a control system.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of information technology, and in particular to a wire extrusion process quality fluctuation adaptive control method. BACKGROUND

[0002] The plug power cord is an indispensable connecting component of household appliances, industrial equipment and lighting products, and its manufacturing quality is directly related to electrical safety and use reliability. With the accelerated replacement of electrical products and the continuous growth of market demand for high-quality power cords, achieving efficient, stable and consistent continuous production has become a core issue that the wire and cable industry urgently needs to solve.

[0003] Traditional plug power cord production usually completes the processes of conductor twisting, insulation extrusion, sheath extrusion, cutting and pressure welding, plug assembly, etc. on multiple independent devices, and relies on manual handling or simple conveyor belt connection between processes. Although this method can basically meet the production needs, it has obvious defects: once a process fluctuates, the entire production chain is difficult to respond in time, resulting in a large number of products being discovered after the problem accumulates; at the same time, the information between processes is completely cut off, and the adjustment of the front extrusion process parameters can only be completed manually according to the experience of the operator, and cannot be corrected according to the actual quality detected by the rear detection, resulting in adjustment lag, repeated trial and error and material waste. The deeper problem is that the extrusion molding process itself is highly sensitive to many factors. Small changes in copper wire diameter, insulation flowability, environmental temperature and humidity between batches of raw materials will directly affect the thickness uniformity, outer diameter size and surface quality of the insulation layer and the sheath layer. In traditional production, these key quality indicators can only be known by off-line sampling measurement after extrusion is completed, and even if thin thickness or bubbles, scratches and other defects are found, hundreds of kilometers of unqualified cables have already been produced. At this time, it is too late to trace back and adjust the front extruder temperature, screw speed or traction speed, and it is difficult to quickly pull the product quality back to the qualified range.

[0004] Therefore, how to realize real-time and accurate perception of the quality of the insulation layer and the sheath layer throughout the continuous production process of the plug power cord, and automatically trace back to the key parameters of the front extrusion process for targeted adjustment according to the detected quality deviation, has become a key problem to improve production efficiency, reduce the unqualified rate and ensure product consistency. SUMMARY

[0005] The present application provides a wire extrusion process quality fluctuation adaptive control method, mainly comprising:

[0006] The thickness of the insulation layer, the thickness of the sheath layer, the outer diameter size and the surface quality data at the outlet of the extruder are collected, and a standardized thickness uniformity index is obtained through digital signal processing; classification judgment is performed according to the standardized thickness uniformity index, and if it deviates from the preset threshold, it is marked as a quality fluctuation event; the fluctuation type is determined according to the quality fluctuation event; the fuzzy logic control algorithm is used to adjust the related parameters of the extruder temperature, screw speed and traction speed according to the fluctuation type, and the optimized temperature setting value is obtained; the extruder control system parameters are updated according to the optimized temperature setting value, and the running state is fed back; the adjustment effectiveness is confirmed by comparing the running state with the initial standardized thickness uniformity index; when the adjustment effectiveness is confirmed, the raw material batch diameter and environmental temperature and humidity data are collected and fused to obtain a comprehensive environmental impact factor; whether secondary adjustment of the traction speed is needed is judged according to the comprehensive environmental impact factor, and the final stable extrusion process parameter group is obtained and fed back to the control system. Further, the thickness of the insulation layer, the thickness of the sheath layer, the outer diameter size and the surface quality data at the outlet of the extruder are collected, and a standardized thickness uniformity index is obtained through digital signal processing, including: the insulation layer thickness data, the sheath layer thickness data, the outer diameter size data and the surface quality original data are obtained through a sensor array; the original data are filtered to obtain filtered thickness sequence data, outer diameter sequence data and surface quality sequence data; the filtered thickness sequence data and outer diameter sequence data are amplified in amplitude to obtain enhanced thickness change sequence and outer diameter change sequence; the filtered surface quality sequence data are marked for defect position and recorded for defect type to obtain a surface defect record sequence; the thickness deviation value, the outer diameter deviation value and the defect mark information of the corresponding position are extracted from the enhanced thickness change sequence, the outer diameter change sequence and the surface defect record sequence, and the thickness deviation value is corrected according to the pre-set weight relationship to obtain the thickness deviation sequence corrected by defects; the thickness deviation sequence corrected by defects is normalized to calculate the standardized thickness uniformity index. Further, the classification judgment according to the standardized thickness uniformity index includes: temperature influence data at the outlet of the extruder are obtained and filtered to obtain filtered temperature sequence data; the filtered temperature sequence data are corrected for deviation value sequence to obtain corrected temperature deviation sequence; the corrected temperature deviation sequence and the preset thickness uniformity index are combined to calculate the normalized index calculation value by weighted average; the normalized index calculation value is input into a support vector machine model for classification analysis to obtain a classification result; if the classification result deviates from the preset threshold, it is marked as a quality fluctuation event and triggers the subsequent fluctuation type determination process.Further, the quality fluctuation event is determined according to the fluctuation type, and a fuzzy logic control algorithm is used to adjust the extruder temperature, screw speed, and traction speed related parameters, including: obtaining the historical data sequence corresponding to the fluctuation type; deviation calculation is performed on the historical data sequence to obtain a temperature deviation correction sequence; the temperature deviation correction sequence and the historical screw speed adjustment value are weighted and fused to obtain a traction speed optimization parameter; the fuzzy logic control algorithm is used to fuzzy conversion on the traction speed optimization parameter and inference through the preset rule base to obtain the preliminary set value; the preliminary set value and the historical extruder parameter monitoring data are fused to obtain the optimized temperature set value. Further, the extruder control system parameters are updated according to the optimized temperature set value and the running state is fed back, including: updating the temperature parameters in the extruder control system according to the optimized temperature set value to obtain the updated temperature parameter sequence; obtaining the traction speed linkage data corresponding to the updated temperature parameter sequence to determine the running state index; real-time sensing features are extracted from the running state index and matched; the matched features are fused to form the updated extrusion process running state; the updated extrusion process running state is fed back to the real-time sensing database for storage. Further, the adjustment effectiveness is confirmed by comparing the running state with the initial standardized thickness uniformity index, including: extracting the updated extrusion process running state and the initial standardized thickness uniformity index from the real-time sensing database to form comparison data; obtaining the linkage traction speed verification sequence according to the speed related part in the comparison data; determining the quality fluctuation trend according to the verification sequence; if the quality fluctuation trend shows that the fluctuation is reduced, the thickness uniformity index and the fluctuation trend data are fused to obtain the parameter adjustment verification sequence; the adjustment confirmation basis is obtained by comparing the index according to the parameter adjustment verification sequence; the adjustment confirmation signal is generated according to the adjustment confirmation basis. Further, when the adjustment effectiveness is confirmed, the raw material batch diameter and environmental temperature and humidity data are collected and fused to obtain a comprehensive environmental impact factor, including: activating the environmental factor monitoring module according to the adjustment confirmation signal; obtaining raw material batch diameter data and environmental temperature and humidity data from the environmental factor monitoring module; using data fusion technology to fuse the raw material batch diameter data and environmental temperature and humidity data to obtain a preliminary environmental data set; batch variation evaluation is performed on the preliminary environmental data set, and the adjustment confirmation signal is fused to obtain a comprehensive environmental impact sequence; the comprehensive environmental impact factor is obtained by extracting the fusion result from the comprehensive environmental impact sequence.Further, the judging whether the traction speed needs to be adjusted again according to the comprehensive environmental influence factor comprises: comparing the comprehensive environmental influence factor with a preset threshold to obtain a judgment result; if the judgment result shows that the preset threshold is exceeded, a fuzzy logic control algorithm is used to finely adjust the current traction speed value to obtain an adjusted speed value; and the adjusted speed value is fused with the raw material batch diameter data and the environmental temperature and humidity data to determine a final stable extrusion process parameter group. Further, the obtaining of the final stable extrusion process parameter group and the feedback to the control system comprises: feeding back the final stable extrusion process parameter group to an extruder control system; updating the current running parameters of the extruder according to the final stable extrusion process parameter group; and enabling the extruder to continuously run according to the updated parameter group to realize continuous production adjustment. Further, the standardized thickness uniformity index, the quality fluctuation event, the fluctuation type, the optimized temperature setting value, the comprehensive environmental influence factor and the final stable extrusion process parameter group are all recycled in the closed-loop control process to realize continuous self-adaptive adjustment of the extrusion process quality.

[0007] The technical scheme provided by the embodiment of the application can have the following beneficial effects:

[0008] The application discloses an intelligent quality closed-loop control method for a cable extrusion process, which comprises the following steps: collecting, by a sensor array, data of thickness uniformity, outer diameter and surface quality of an insulating layer sheath layer in real time, inputting the data into a support vector machine after standardization by digital signal processing, classifying abnormalities, accurately identifying thickness fluctuation events and determining fluctuation types; then, a fuzzy logic control algorithm is used to fuzz and infer historical data of temperature, screw speed and traction speed, an optimized temperature setting value is output to realize first closed-loop adjustment, and an updated state is stored in a real-time perception database; after adjustment effectiveness is confirmed through comparison, a comprehensive environmental influence factor is generated by further fusing raw material batch characteristics and environmental temperature and humidity data, secondary fine adjustment of traction speed is triggered when the factor exceeds a threshold, and finally, a stable and continuous parameter optimization group is formed to feed back to a control system. The application solves the problem that thickness uniformity is easily disturbed by multiple coupling factors such as temperature, speed, raw materials and environment in the cable extrusion process, resulting in quality fluctuation, realizes full-link intelligent closed-loop control from abnormal perception, type diagnosis, parameter self-adaptive adjustment to environmental compensation, and greatly improves the thickness consistency of the cable outer sheath and production stability. BRIEF DESCRIPTION OF DRAWINGS

[0009] Figure 1 A flowchart of a wire rod extrusion process quality fluctuation self-adaptive control method of the application. DETAILED DESCRIPTION

[0010] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application will be described in detail below with reference to the drawings and specific embodiments.

[0011] As Figure 1 The wire extrusion process quality fluctuation adaptive control method of the embodiment can specifically include the following steps.

[0012] S101, collecting the thickness uniformity data, outer diameter size data and surface quality data of the insulation layer and the sheath layer at the outlet of the extruder through a sensor array, and filtering and amplifying the collected raw data to obtain a standardized thickness uniformity index by using digital signal processing technology.

[0013] The sensor array obtains the insulation layer thickness data, sheath layer thickness data, outer diameter size data and surface quality raw data from the outlet of the extruder. The data is filtered by using digital signal processing technology to obtain filtered thickness sequence data, outer diameter sequence data and surface quality sequence data. The filtered thickness sequence data and outer diameter sequence data are amplified to obtain enhanced thickness change sequence and outer diameter change sequence; at the same time, the filtered surface quality sequence data is marked with defect position and recorded with defect type to obtain a surface defect record sequence. The thickness deviation value, outer diameter deviation value and corresponding position defect mark information are extracted from the enhanced thickness change sequence, outer diameter change sequence and surface defect record sequence, the thickness deviation value is corrected according to the pre-set weight relationship, and a thickness deviation sequence corrected by defects is obtained. The thickness deviation sequence corrected by defects is normalized to calculate a standardized thickness uniformity index.

[0014] In one embodiment, the sensor array collects relevant data of the insulation layer and the sheath layer at the outlet of the extruder.

[0015] Specifically, the sensor array is arranged at a fixed position of the outlet of the extruder, including a plurality of laser thickness measuring sensors and optical imaging sensors, for real-time monitoring of the insulation layer and the sheath layer of the cable. The laser thickness measuring sensor calculates the thickness value by emitting a laser beam and receiving a reflected signal, while the optical imaging sensor captures the surface image to evaluate the quality. These sensors are arranged in an array form to ensure coverage of the entire circumference of the cable, so as to collect the thickness uniformity data, outer diameter size data and surface quality data. This arrangement can achieve comprehensive monitoring of the cable production process and improve the comprehensiveness of the data. Further, the collected raw data often contains noise and weak signals, and therefore digital signal processing technology is used for processing.

[0016] Exemplarily, the digital signal processing technique includes two main steps: filtering and amplification. First, low-pass filter is used to remove high-frequency noise, such as finite impulse response filter, to smooth the original thickness data. The specific process is to input the collected signal into the filter, and the filter retains the useful signal according to the preset cutoff frequency and eliminates the interference, so as to obtain more stable thickness uniformity data. Then, the filtered signal is amplified, and the operational amplifier circuit is used to enhance the signal strength to a processable level. This processing ensures that the data accurately reflects the actual state of the cable.

[0017] Preferably, in the cable insulation layer extrusion scene, the acquisition of the standardized thickness uniformity index is based on the processed data.

[0018] It should be noted that the standardization process involves converting the thickness value into a relative uniformity index, such as calculating the standard deviation of the thickness value and normalizing it to the range of 0 to 1.

[0019] Specifically, the average value of the thickness data collected from multiple sensor points is calculated, and then the deviation of each point is calculated, and the deviation value is divided by the average value to obtain the uniformity score. In this way, the standardized index is obtained, which is used for subsequent quality evaluation. This index helps to identify uneven problems in the extrusion process, such as insulation defects caused by thickness fluctuations.

[0020] In one possible implementation, for the outer diameter size data of the sheath layer, the sensor array continues the laser measurement in the previous thickness data collection, and an ultrasonic sensor is used to supplement. The ultrasonic sensor emits sound waves and measures the echo time to calculate the outer diameter size. This combined use is particularly effective in high-speed extrusion environments, as ultrasonic waves are sensitive to changes in material density and can provide complementary data. When processing, the ultrasonic signal is also subjected to digital filtering, such as median filtering to remove isolated noise points, and then the signal is amplified to enhance the resolution. Finally, the standardized outer diameter index is calculated by comparing the actual size with the preset tolerance, ensuring that the cable meets the specification requirements.

[0021] For example, in the collection and processing of surface quality data, optical sensors capture image data, and digital signal processing techniques include edge detection filtering. Edge detection filtering highlights surface defects, such as cracks or bubbles, through a convolution operation. The specific process is to convert image data into a grayscale matrix I, apply a Sobel operator to calculate horizontal gradient Gx and vertical gradient Gy, get gradient amplitude G equal to the square root of the square of Gx plus the square of Gy, then filter out non-defect noise with G less than the threshold value of 0.1, and finally amplify the contrast to highlight the problem area. This processing obtains a standardized surface quality index, such as a defect density score, for quantifying the appearance quality of the cable. Further, in another embodiment, the sensor array integrates an infrared thermal imaging sensor for monitoring the temperature distribution during extrusion to affect thickness uniformity. The infrared sensor collects thermal image data, and digital signal processing smoothes thermal noise through Gaussian filtering, then amplifies the thermal signal to calculate the temperature gradient. The standardized index is based on a mapping of temperature to thickness, such as converting temperature deviation to a thickness uniformity correction value through an empirical formula C = k * AT, where C is the correction value, AT is the temperature deviation, and k is an empirical coefficient of 0.02. The calculation process is to first calculate AT equal to the actual temperature minus the target temperature, then multiply by k to get C. This method is applied in multi-layer cable extrusion to improve overall monitoring accuracy.

[0022] It can be understood that the flexibility of the above processing techniques is reflected in parameter adjustment.

[0023] For example, the cutoff frequency of the filter can be set according to the type of cable material. For a polyethylene insulation layer, the frequency is set lower to preserve low-frequency thickness variation signals. This adjustment ensures that the processing adapts to different production scenarios.

[0024] In one embodiment, the entire data processing flow is deployed in an embedded processor to achieve real-time feedback. The collected data is converted into digital after analog-to-digital conversion, then enters the processor to execute filtering and amplification algorithms, and outputs standardized indices to the display system. This implementation reduces human intervention and improves production efficiency.

[0025] Specifically, for the standardization of thickness uniformity data, multi-point sampling averaging is involved. The sensor array collects hundreds of points per second, and the variance is calculated as the uniformity index after processing. This index plays a key role in cable quality control, enabling early detection of problems caused by deviations in extruder parameters. Finally, through these processes, the standardized thickness uniformity index supports the optimization of cable production, such as adjusting the extrusion speed to maintain uniformity, thereby ensuring product reliability and consistency.

[0026] S102, according to the standardized thickness uniformity index, a 5-dimensional input vector (including mean, variance, etc.) is constructed, which is input into a support vector machine model for classification analysis. The model is supervised trained with historical data set, and radial basis function is selected as kernel function. If the classification result shows that the thickness uniformity index deviates from the preset threshold value 0.05, it is marked as a quality fluctuation event. The event marker triggers the subsequent data flow parameter adjustment module to determine the fluctuation type.

[0027] Temperature influence data is obtained from the outlet of the extruder, and digital signal processing technology is used to filter the temperature influence data to obtain filtered temperature sequence data. The filtered temperature sequence data is corrected for bias value sequence, and the corrected temperature bias sequence is obtained by calculating the difference between each point in the sequence and the average value and applying a correction coefficient. The correction coefficient is based on the experience calibration of historical production data, and the value is 0.85. The calculation is based on minimizing the variance of the bias. According to the corrected temperature bias sequence and the preset thickness uniformity index, the normalized index calculation value is calculated by weighted average, wherein the weight is 0.6 of the thickness uniformity index and 0.4 of the bias sequence, and the preset thickness uniformity index range is 0-1. The multi-dimensional vector composed of the normalized index calculation value, thickness standard deviation and outer diameter bias value is input into a support vector machine model, wherein the support vector machine model takes the multi-dimensional vector as input and outputs a classification label to obtain a classification analysis result. If the classification analysis result deviates from the preset threshold value, it is judged as a quality fluctuation marker. The quality fluctuation marker triggers the event trigger mechanism, sends a signal to the data flow adjustment, and determines the fluctuation type according to the signal content.

[0028] In one embodiment, the process of inputting the standardized thickness uniformity index into the support vector machine model for classification analysis first needs to understand the basic principles of the support vector machine model. Support vector machine is a supervised learning algorithm used for classification tasks, which separates different classes of data points by constructing a hyperplane in the feature space. In the quality control of cable extrusion production, this model is trained to distinguish between normal and abnormal states of thickness uniformity index.

[0029] Specifically, the model receives standardized indicators as input vectors, which are numerical values obtained from previous data processing steps, such as normalized scores of thickness standard deviation. During training, a historical dataset is used, which includes samples of thickness data with known normal and abnormal classes. The model learns the optimal hyperplane to maximize the separation between different classes. This method ensures classification accuracy, and in practical applications, after inputting new collected indicators, the model calculates their position on one side of the hyperplane and outputs the classification result. Further, if the classification result shows that the thickness uniformity indicator deviates from the preset threshold, it is marked as a quality fluctuation event. Here, the preset threshold is a numerical range set based on cable specification standards, for example, for a polyethylene insulation layer, the threshold may be set to a uniformity score below 0.8. After classification analysis, if the model judges that the indicator falls into the abnormal category, i.e., deviates from the threshold, the system will automatically generate an event marker. This marker contains a timestamp, indicator value, and classification label, which records the specific time and degree of fluctuation. This marking mechanism helps to quickly respond to production problems and avoid manual intervention.

[0030] Exemplarily, in the cable sheath layer extrusion scene, the classification analysis of the support vector machine model can be combined with multi-dimensional indicators. The standardized thickness uniformity indicator not only includes the thickness standard deviation, but also can be extended to the outer diameter deviation value. The model maps the input data to a high-dimensional space through a kernel function, such as a radial basis function, to achieve nonlinear classification. The input process involves feeding the normalized indicator vector into the model, and the model calculates the decision function value. If the value exceeds the threshold boundary, it is considered normal, otherwise an abnormal alarm is triggered. This extension enhances the model's ability to recognize complex fluctuation patterns.

[0031] It should be noted that when the event marker triggers the subsequent data flow to the parameter adjustment module to determine the fluctuation type, the event marker acts as a trigger signal to transmit relevant data such as indicator details and classification results to the adjustment module. The parameter adjustment module is a rule-based system that analyzes incoming data to identify fluctuation types, such as uneven melting due to excessive extruder temperature or material distribution problems caused by screw speed fluctuations. The module matches the most likely type by comparing event data with a preset pattern library, for example, if the indicators show periodic deviations, it is classified as a speed-related fluctuation.

[0032] In one possible implementation, for a multi-layer cable extrusion production line, the classification of the support vector machine model employs an online learning mode. After the model is trained through initial offline supervised learning, the parameters are continuously updated in production to adapt to the variations of different batches of cables. When the input standardized indicators are standardized, the system calculates the classification probability in real time, and if the probability indicates a deviation from the threshold value, such as a deviation of more than 10%, the event is marked. Subsequently, the data flow to the adjustment module, which uses a decision tree algorithm to assist in determining the type, for example, the decision tree input is the event data marked by the support vector machine, including the temperature and pressure correlation, and the output is the fluctuation type, such as thermal distribution fluctuation. This way improves the adaptability of the system.

[0033] Preferably, after determining the fluctuation type, the parameter adjustment module can generate a feedback signal, which includes specific adjustment instructions such as a temperature decrease of 5 degrees Celsius, in the form of a JSON structure, and sends it to the extruder controller to achieve a closed-loop adjustment, such as automatically modifying parameters to stabilize production.

[0034] For example, in the insulation layer extrusion, if the type is confirmed as unstable material supply, the module will recommend adjusting the speed of the feed pump. This determination process is based on the logical chain of data flow, ensuring continuity from classification to adjustment.

[0035] It can be understood that in another embodiment, the classification analysis of the support vector machine model is optimized for thickness indicators related to surface quality. This model is a binary classification model used to distinguish between quality pass and fail. The model input is a joint vector including thickness uniformity TU (thickness standard deviation) and surface defect density DD (defects per unit area), and the output is a class label 0 (normal) or 1 (abnormal). If the classification result deviates from the threshold value, such as the defect score (abnormal probability) is higher than 0.2 (based on historical data statistics), it is marked as a composite quality fluctuation event. Event marking triggers data flow to the adjustment module, which determines the type through pattern matching, such as thickness variation caused by surface roughness, and further supports precise calibration of production parameters. This integrated method is effective in high-speed extrusion environments, improving the accuracy of overall quality monitoring.

[0036] Specifically, the determination of fluctuation type involves multi-step analysis in the parameter adjustment module. First, the module receives event marking data and extracts key features such as deviation amplitude and frequency. Then, compare these features with the historical case library, for example, if the deviation is randomly distributed, it is classified as random noise type; if it is trended, it is a systematic problem. Finally, output the type label to guide the operator or automated system to make adjustments. This detailed analysis ensures accurate diagnosis of fluctuations and reduces defect rates in cable production.

[0037] For example, in the implementation of the sheath layer thickness control, the support vector machine model can be configured as a multi-class classification to distinguish between slight, moderate and severe deviations from the threshold. After inputting the indicators, the model outputs a class, and if it is a severe deviation, a high-priority event is marked, triggering the adjustment module quickly. The module then calculates the adjustment amplitude based on the type, such as an extrusion pressure anomaly, to ensure that production returns to normal. Further, the entire process can be implemented in an industrial control system, with the support vector machine model running on an embedded computing unit, and the event marking and data stream being transmitted in real time. This deployment provides a reliable quality assurance mechanism in the field of cable manufacturing.

[0038] S103, after obtaining the fluctuation type, the historical data of the extruder temperature, screw speed and pulling speed are processed by fuzzy logic control algorithm, and if the processed data matches the current fluctuation type, the optimized temperature setting value is output.

[0039] The historical data sequence corresponding to the fluctuation type classification is obtained from the extruder operation record, and the temperature deviation correction sequence is obtained by using the deviation calculation method for the historical data sequence, wherein the deviation calculation method obtains the temperature deviation correction sequence by multiplying the difference between each point of the sequence and the average value by a correction coefficient, and the correction coefficient c is dynamically determined based on the historical data standard deviation σ, c = 1 / σ. For the temperature deviation correction sequence, the screw speed adjustment value obtained from the historical data sequence is combined to determine the pulling speed optimization parameter by a weighted fusion method, wherein the weighted fusion method sums the weights of the temperature deviation correction sequence and the screw speed adjustment value, and the weights w1 and w2 are dynamically calculated based on the Pearson correlation coefficient, w1 + w2 = 1. The fuzzy logic control algorithm is used to perform a fuzzy conversion process on the pulling speed optimization parameter, wherein the fuzzy logic control algorithm maps the input variable to a fuzzy set, executes inference judgment through a pre-set rule base, and if the inference result exceeds a matching degree threshold, a preliminary setting value is obtained by a center average defuzzification method, the membership function adopts a triangular form, the inference method is Mamdani, and the rule base example includes outputting an optimization parameter if the temperature deviation is high and the speed adjustment is large. According to the fusion of the preliminary setting value and the extruder parameter monitoring data obtained from the historical data sequence, if the fusion deviation is lower than a pre-set threshold, a quality fluctuation response adjustment sequence is determined, wherein the fusion is calculated by averaging the preliminary setting value and the extruder parameter monitoring data. The optimized setting value generation result is obtained by fusing the quality fluctuation response adjustment sequence and the parameter history obtained from the historical data sequence, wherein the fusion is obtained by weighting the quality fluctuation response adjustment sequence and the parameter history.

[0040] In one embodiment, after obtaining the fluctuation type, the system employs a fuzzy logic control algorithm to process historical data of extruder temperature, screw speed, and haul-off speed. This algorithm is based on fuzzy set theory, which converts uncertain data into processable fuzzy variables.

[0041] Specifically, the fuzzy logic control algorithm first defines fuzzy sets of input variables, for example, dividing temperature historical data into fuzzy subsets such as low, medium, and high temperature, and calculating the degree to which each data point belongs to these subsets through membership functions. Membership functions usually take the form of triangles or trapezoids, quantifying historical data, for example, calculating membership degrees step by step within a set range, thus achieving fuzzy processing. This process ensures the conversion of data from precise numerical values to fuzzy descriptions, facilitating subsequent matching. Further, the fuzzy-processed data are matched with the current fluctuation type. Fluctuation types can include temperature instability or speed deviation, and the system reasons through a fuzzy rule base, for example, a rule such as "if the temperature fuzzy value is high and the screw speed fuzzy value is high, then match the heat distribution fluctuation type." The judgment process involves a fuzzy reasoning engine, which aggregates the outputs of multiple rules to calculate the overall matching degree. If the matching degree reaches a preset standard, such as the sum of membership degrees exceeding a certain limit, the match is confirmed. This matching mechanism, when applied in cable insulation layer extrusion, takes into account the time sequence characteristics of historical data, ensuring the accuracy of the judgment.

[0042] Illustratively, in the production scenario of cable sheath layer, the fuzzy logic control algorithm's fuzzy processing is extended to multi-parameter fusion. Historical data such as the temperature sequence of the past hour are first normalized and then input into the fuzzy processing module. The module defines membership functions, for example, the low-speed subset of haul-off speed corresponds to the range where the speed is below the standard value, and generates a fuzzy vector by calculating the membership degree of each data point. Subsequently, the system compares this vector with the current fluctuation type, such as the pattern vector corresponding to thickness fluctuation, to assess the similarity. If the distance between the vectors is less than a threshold, it is judged as a match. This extended method handles the multiple variables in production.

[0043] It should be noted that after judging the match, the system outputs the optimized temperature set value. The output is based on the de-fuzzy process, for example, using the barycentric method to calculate the weighted average of fuzzy output, generating a specific temperature value, such as deriving a 5-degree reduction in set adjustment from the fuzzy high temperature state.

[0044] In one possible implementation, for multi-layer cable extrusion, this output is combined with screw speed history (screw speed data sequence collected in real time by sensors and stored), ensuring that the set value adapts to different layer thickness requirements.

[0045] Specifically, the implementation of the fuzzy logic control algorithm involves the construction of a rule base. The rule base is extracted from historical fluctuation events (abnormal changes recorded in production, such as temperature or speed deviation), for example, the type of fluctuation caused by previous high temperature corresponds to the rule "high temperature input leads to reduced temperature output". The algorithm input includes variables such as temperature T, screw speed S, etc., which are processed by a triangular fuzzy function; the output is the optimized set value, such as the adjusted speed. The construction process first analyzes the event parameters (such as deviation amplitude exceeding 5%), then generates rules to ensure the applicability of the algorithm in cable production.

[0046] Preferably, in another embodiment, the fluctuation type is first obtained through data analysis, such as uneven speed. Then, the fuzzy processing is performed on the real-time data stream, which is defined as the system collecting data from the recent production cycle, continuously fuzzing temperature and speed, for example, updating the membership every minute. The matching judgment uses a dynamic threshold, which is calculated based on the historical fluctuation standard deviation, for example, set to 0.8. If the fuzzy data matching degree exceeds the threshold and matches the fluctuation type, the temperature setting is immediately output, such as adjusting to 90% of the standard value. This real-time method is suitable for high-speed extrusion environment.

[0047] It can be understood that the logic of the whole process starts from obtaining the type and gradually outputs to ensure the continuity of parameter optimization.

[0048] For example, in cable conductor layer extrusion, the algorithm processes the pulling speed history, fuzzes it into slow or fast subsets, and outputs the temperature value after matching to stabilize the melting state.

[0049] In one embodiment, the fuzzy logic control algorithm can be configured in an adaptive mode, with its initial rule base based on standard production data. The adaptive mechanism monitors the rule matching frequency during processing. If the matching rate exceeds 80%, the membership function boundary is updated, for example, by linear interpolation to widen the medium temperature subset range, specifically, the new boundary is equal to the weighted average of the original boundary, where the weight w is 0.2 (w represents the adjustment factor). When the output optimized temperature is calculated, the cumulative historical deviation d (d is the cumulative average of the difference between the actual temperature and the target temperature) is calculated, and the progressive adjustment value a (a is equal to d multiplied by the gain factor 0.1) is generated accordingly. This configuration reflects flexibility in variable batch production.

[0050] S104, update the temperature parameter in the extruder control system through the optimized temperature set value, obtain the updated running state of the extrusion process, and feed the running state back to the real-time perception database for storage for subsequent cycles.

[0051] The temperature parameters in the extruder control system are updated by the optimized temperature set value, and an updated temperature parameter sequence is obtained. For the updated temperature parameter sequence, the pulling speed linkage data of the extrusion process is obtained, and the running state index is determined. Real-time sensing features are extracted from the running state index, and if the feature value matches the preset threshold, it is fused into the updated running state of the extrusion process. The updated running state of the extrusion process is fed back to the real-time sensing database for storage, and a storage sequence for subsequent recycling is obtained.

[0052] In an embodiment, the temperature parameters in the extruder control system are updated by the optimized temperature set value. This process first involves transmitting the optimized value to the input interface of the control system.

[0053] Specifically, the control system usually includes a temperature adjustment module. After receiving the optimized set value, the module adjusts the working state of the heating element. For example, in the production of cable insulation layer extrusion, the temperature is modified from the original set value to the new value to adapt to the demand after matching the fluctuation type. This update ensures uniform heat distribution in the extrusion process and avoids uneven material melting problems.

[0054] It should be noted that the update mechanism is based on standard industrial protocols such as Modbus protocol to achieve instant writing of parameters, so that the extruder applies the new temperature in the next production cycle. Further, the updated running state of the extrusion process is obtained, which is determined by the real-time data collected by the sensor network.

[0055] For example, in the cable sheath layer extrusion scene, the running state includes a combination of temperature actual value, screw speed and pulling speed. The system calculates the deviation value of these parameters, for example, compares the difference between the updated temperature and the set value to generate a state vector. This vector reflects the stability of the extrusion process, such as a deviation less than a preset threshold indicating a normal state. The running state acquisition process emphasizes multi-parameter fusion to ensure comprehensive evaluation of the production environment response.

[0056] Preferably, the running state is fed back to the real-time sensing database for storage. This database is designed to support high-frequency data writing, such as using a time series database such as InfluxDB, and the storage fields include timestamp, temperature parameter and state index (such as running mode, fault code).

[0057] In a possible implementation, the feedback process is realized through an API interface. The system packages and uploads the state data, and the database maintains a historical record table for subsequent recycling. This storage mechanism facilitates data retrieval by the algorithm in the next iteration, forming a closed-loop optimization.

[0058] It can be understood that in the conductor insulation layer extrusion environment in cable production, the application of feedback cycle is extended to multi-machine cooperation. Specifically, after updating the temperature parameter, the running state is not only stored in the local module of the real-time perception database, but also synchronized to the cloud perception system (a cloud-based remote monitoring platform that synchronizes with the local module data in real time), allowing remote monitoring. This extension handles the scaling needs of the production line and ensures data consistency. In another embodiment, the feedback of the running state includes an abnormality detection mechanism, such as obtaining the state by presetting a threshold, such as a temperature fluctuation exceeding plus or minus 5 degrees Celsius. If the updated state shows that it is out of range, an alarm is triggered and detailed logs are stored. This mechanism is practical in high-speed extrusion production and supports fault diagnosis through continuous storage. Further, the logic of the entire updating and feedback process starts with parameter adjustment and gradually moves to data storage, ensuring iterative optimization of extrusion parameters.

[0059] For example, in multi-layer cable extrusion, the update of the optimized temperature setting value is combined with the layer thickness requirement, and the running state feedback updates the corresponding field in the database, promoting the maintenance of production efficiency.

[0060] S105, extract the comparison data of the updated running state and the initial standardized thickness uniformity index from the real-time perception database. If the comparison data indicates that the quality fluctuation is reduced, confirm that the parameter adjustment is effective, and generate an adjustment confirmation signal through effective confirmation.

[0061] Extract the comparison data of the updated running state and the initial standardized thickness uniformity index from the real-time perception database. Obtain a linkage traction speed verification sequence from the traction speed difference sequence in the comparison data, which is used to verify the linkage effect of the traction speed, and determine the quality fluctuation trend according to the fluctuation part in the verification sequence. For the quality fluctuation trend, if the comparison data indicates that the fluctuation is reduced, fuse the thickness uniformity index and the fluctuation trend data to generate a parameter adjustment verification sequence, which is obtained by fusing the thickness uniformity index TU and the fluctuation trend data FT, i.e. PAVS=(TU+FT) / 2. Use the parameter adjustment verification sequence to judge the effectiveness index by comparing the indexes in the verification sequence, i.e. the sequence difference is less than the threshold 0.05, which is considered effective, and obtain the adjustment confirmation basis, which contains the effective sequence part. Through the adjustment confirmation basis, generate an adjustment confirmation signal by fusing the effective part in the basis and storing it in the control system.

[0062] In one embodiment, the comparison data of the updated running state and the initial standardized thickness uniformity index is extracted from the real-time perception database. This process first involves querying specific fields in the database.

[0063] Specifically, the real-time perception database stores various parameters of the extrusion process, for example, in the production of cable insulation layer, the running state includes the actual value of temperature and the thickness measurement data, and the initial standardized thickness uniformity index refers to the standard deviation value of thickness distribution calculated by statistical method, which is obtained at the initial stage of production for benchmark comparison. When extracting, the system uses SQL query statement to select the data set matching the time stamp to ensure the timeliness and accuracy of the comparison data. This extraction mechanism supports high-frequency access and facilitates real-time verification of the effect of parameter adjustment. Further, if the comparison data shows that the quality fluctuation is reduced, it is confirmed that the parameter adjustment is effective, and this confirmation is based on the preset threshold judgment.

[0064] Exemplarily, in the cable sheath layer extrusion scene, the quality fluctuation is quantified by calculating the difference between the updated thickness uniformity index and the initial index, for example, if the difference shows that the standard deviation is reduced by more than 10%, it is considered that the fluctuation is reduced. This judgment process emphasizes data fusion, and the system performs vector comparison between the running state vector and the initial index to generate a fluctuation coefficient, thereby objectively evaluating the effectiveness of the adjustment.

[0065] It should be noted that the definition of fluctuation reduction is derived from business standards, such as reducing the thickness deviation from the initial 0.5 mm to below 0.3 mm, which helps to maintain the uniformity of the extruded material and avoid product defects.

[0066] Preferably, an adjustment confirmation signal is generated through effective confirmation, which is used as the output of the feedback mechanism for subsequent production control.

[0067] In one possible implementation, after the confirmation is effective, the system generates a digital signal, such as a binary flag bit, which is transmitted to the extruder control system. In the cable conductor layer extrusion environment, this signal triggers parameter solidification to ensure that the optimized value obtained from the fluctuation reduction confirmation, i.e., the adjusted extrusion parameter, is applied in multiple cycles. The generation process includes signal encoding and verification to prevent transmission errors: first, the binary flag bit is converted into a self-synchronous signal using Manchester encoding as input data, then a CRC check value is calculated and attached to the encoded data as output for transmission; the receiving end verifies the CRC value to confirm that there is no error, and this mechanism reflects the reliability on high-speed production lines.

[0068] It can be understood that in another embodiment, the extraction of the comparison data is combined with abnormal filtering, such as excluding data points affected by sensor noise. Specifically, the comparison data is derived from the initial thickness deviation data, and the extraction process is to obtain an effective comparison set by screening out abnormal points from the deviation data; the calculation of the initial standardized thickness uniformity index involves the ratio of the average thickness and the deviation, that is, the formula U = A / D, where U is the uniformity index, A is the arithmetic mean of the thickness of all points, and D is the standard deviation of the thickness. In the extraction, the system first applies a Gaussian filtering algorithm to smooth the data (input is the original thickness data, and output is the smoothed data without noise), and then performs comparison. This extended mode handles complex production fluctuations and ensures the accuracy of the confirmation. Further, the logic of the entire comparison and confirmation process starts from data extraction and gradually proceeds to signal generation, forming a verification chain.

[0069] For example, in the process of extruding cable multi-layer insulation material, i.e. multi-layer cable extrusion, if the quality fluctuation, i.e. the standard deviation σ of the production parameter, is reduced, the adjustment confirmation signal generated by the system can be synchronized to the cloud database to support remote quality fluctuation monitoring based on the signal. This application is extended to large-scale production lines, and data consistency is maintained through real-time data uploading and distributed backup mechanisms.

[0070] In an embodiment, the judgment of the validity of the confirmation includes multi-index fusion, i.e. generating a comprehensive fluctuation index I combining temperature T and thickness D data, the calculation formula is I = 0.6 × (T fluctuation rate) + 0.4 × (D fluctuation rate), where T fluctuation rate is the standard deviation divided by the average value of temperature, and D fluctuation rate is the standard deviation divided by the average value of thickness. Exemplarily, if the index I is lower than the threshold value 0.5 showing overall quality improvement, the signal generation process (monitoring system output quality signal) activates the alarm suppression mechanism (blocks alarm trigger through logic gate), avoiding unnecessary interruption. This fusion improves the robustness of parameter adjustment in cable extrusion business, specifically by feeding back the index I to optimize the extrusion speed and pressure parameters in real time, reducing noise interference and ensuring more stable and reliable adjustment.

[0071] Specifically, the comparison data is stored back to the database to form a historical record, facilitating the tracking of adjustment effects. In the cable insulation layer scenario, this storage includes quantitative values of the degree of fluctuation reduction, supporting subsequent analysis. Further, through the generated adjustment confirmation signal, the system can automatically iterate the optimization cycle.

[0072] For example, in the production of sheath layers, the reference index refers to the standard value of the diameter of the raw material, and the reference index is updated after signal confirmation, promoting continuous improvement.

[0073] Step S106, according to the adjustment confirmation signal to activate the environmental factor monitoring module to collect raw material batch diameter and environmental temperature and humidity data, using data fusion technology to fuse the collected data to obtain a preliminary environmental data set, and then evaluate the batch variation of the data set and fuse the adjustment confirmation signal to obtain the comprehensive environmental impact factor.

[0074] According to the adjustment confirmation signal to activate the environmental factor monitoring module, raw material batch diameter and environmental temperature and humidity data are obtained from the module. Data fusion technology is used to fuse the raw material batch diameter and environmental temperature and humidity data to obtain a preliminary environmental data set. For the preliminary environmental data set, batch variation assessment is used to fuse the adjustment confirmation signal, where batch variation assessment refers to assessing the degree of variation associated with raw material batch diameter fluctuations and environmental temperature and humidity. The process is to extract the diameter fluctuation part and the temperature and humidity correlation part from the data set, calculate the sample coefficient of variation CV, and fuse the signal. The CV calculation formula is CV=(SD / Mean) x 100, where SD is the standard deviation and Mean is the average value. The fusion signal uses a weighted average method with weights of 0.6 and 0.4 to determine the comprehensive environmental impact sequence. The fusion result is obtained from the comprehensive environmental impact sequence, and the average value of the sequence elements is calculated to obtain the comprehensive environmental impact factor.

[0075] In one embodiment, the activation of the environmental factor monitoring module is based on the receipt of the adjustment confirmation signal.

[0076] Specifically, when the system receives the adjustment confirmation signal, the signal first undergoes a verification process to ensure its validity.

[0077] For example, the signal may originate from an upstream production control system indicating the need for adjustment of the raw material processing flow. Through this verification, the module is activated and begins to perform data collection tasks. This activation mechanism ensures the synchronization of the monitoring process with production adjustments, enabling real-time response in the field of raw material batch processing. Further, the activated environmental factor monitoring module collects raw material batch diameter data. In the context of raw material batch processing, such as a production line for metal wire, the module uses a laser diameter gauge to non-contact measure the diameter of the raw material in the batch. The specific process includes placing the raw material on the measurement path, and the laser sensor scans and records the diameter value while collecting data from multiple points to calculate the average diameter. This collection method is suitable for continuous production environments and can obtain accurate data without interrupting the process.

[0078] Preferably, the module simultaneously collects environmental temperature and humidity data. In the same raw material processing field, for example, a plastic product forming production line, temperature and humidity sensors are integrated in the monitoring module. The sensors monitor the temperature and humidity values of the production workshop in real time, for example, by acquiring data through a thermistor and a capacitive humidity sensor, respectively. These data acquisition processes emphasize accuracy control to reflect the influence of the environment on the characteristics of the raw materials, such as humidity that can cause a slight change in diameter.

[0079] In one possible implementation, data fusion technology is used to combine the collected data with the adjustment confirmation signal.

[0080] Specifically, data fusion first involves preprocessing of multi-source data, such as normalizing diameter data and applying a median filter algorithm to humidity and temperature data to remove noise. The algorithm inputs the original humidity and temperature data sequence, replaces each point in the sequence with the median value in the neighborhood, and outputs the denoised data. Then, these data are integrated with the adjustment confirmation signal, which may contain adjustment parameters such as target diameter values. The fusion technology can be based on a weighted fusion method, in which the diameter data is assigned a weight of 0.4, the humidity and temperature data are assigned a weight of 0.3, and the signal parameters are assigned a weight of 0.3. The preliminary fusion value is calculated by weighted summation, and further optimized by applying a Kalman filter to handle data uncertainty. The filter inputs the preliminary fusion value and the measurement noise covariance matrix Q = 0.1 (process noise) and R = 0.05 (measurement noise), and the process includes a prediction step to calculate the state estimate and covariance, and an update step to fuse the measurement value to minimize uncertainty. The output is the optimized fusion value. This fusion process takes into account the interaction between environmental variables, such as high temperature that can amplify diameter deviation, to generate a more accurate integrated value. In raw material batch quality control, this technology ensures the consistency and reliability of the data, effectively supporting production adjustment decisions.

[0081] For example, in another embodiment, the fusion process can introduce a neural network model for assistance. Specifically, data collected from environmental sensors (such as temperature, humidity) and signals (such as vibration frequency) are input into a multi-layer perceptron model, which includes an input layer, a hidden layer, and an output layer, using a ReLU activation function, for a total of 3 layers. First, historical production records (such as the relationship between diameter variation and temperature and humidity for past batches) are preprocessed, including normalization and feature extraction. Then, the model is trained using the backpropagation algorithm, with a learning rate of 0.01, a batch size of 32, and 100 training rounds, to learn the relationship between environmental factors and diameter. The model outputs a fused vector representing the comprehensive impact.

[0082] It should be noted that this method is suitable for complex production environments, such as chemical raw material processing production lines, where environmental fluctuations are large, and adaptive fusion is achieved through the model.

[0083] It can be understood that, through the above fusion, a comprehensive environmental impact factor is obtained. The factor is a quantitative index, for example, in the form of a numerical value representing the overall impact of the environment on the diameter of the raw material. In implementation, the factor calculation is based on the linear combination of the fused data, for example, factor value = diameter deviation * temperature and humidity coefficient + signal adjustment weight, where the coefficient is determined by empirical data. Such a factor plays a role in raw material batch optimization and can guide subsequent production parameter adjustment. Further.

[0084] In one embodiment, the application of the comprehensive environmental impact factor is extended to the batch screening scenario.

[0085] For example, in the processing of textile fiber raw materials, the factor is used to assess whether the batch meets the quality standard, and if the factor exceeds the threshold, an alarm is triggered. This extension demonstrates the versatility of the technology without changing the core field.

[0086] In one embodiment, the activation and collection of the data collection module can integrate wireless transmission technology, which is used to activate the collection of environmental parameters and real-time transmission of data. Specifically, the system generates an adjustment confirmation signal based on the collected data and sends it to the module through a wireless network, and the module responds in real time and adjusts the collection data. This way improves the flexibility of the system in large production workshops.

[0087] Preferably, the data fusion process fuses the input environmental sensor data (such as temperature, humidity) through a weighted average algorithm, outputs a comprehensive environmental impact factor, and can be visually displayed, for example, by presenting numerical values and trend charts through a production monitoring interface, for easy analysis by operators. In the field of raw material batch processing, such a comprehensive environmental impact factor is generated by the formula E = 0.4T + 0.3H + 0.3O (where E is the factor, T is the temperature, H is the humidity, and O is other parameters), which can achieve accurate monitoring of the production process, thereby reducing batch defects caused by environmental factors.

[0088] S107, if the comprehensive environmental impact factor exceeds the preset threshold, a fuzzy logic control algorithm is applied to the secondary fine-tuning of the traction speed to obtain a final stable extrusion process parameter set, and the parameter set is returned to the extruder control system to realize continuous production adjustment.

[0089] The comprehensive environmental impact factor is obtained, and a threshold value is used to determine whether the factor exceeds the threshold value, obtaining a determination result. If the determination result shows that the threshold value is exceeded, a fuzzy logic control algorithm is applied for secondary fine-tuning of the traction speed, with the current traction speed deviation and the environmental factor deviation value as inputs, and the adjusted speed value as output, obtaining the adjusted speed value. The raw material batch diameter data and the environmental temperature and humidity data are fused from the adjusted speed value to determine the final stable extrusion process parameter set. The specific fusion process uses a weighted average method for calculation, i.e. P = (0.5*V + 0.3*D + 0.2*(T + H) / 2), where P is the final parameter set value, V is the adjusted speed value, D is the diameter data, T is the temperature data, and H is the humidity data. The final stable extrusion process parameter set is fed back to the extruder control system to realize continuous production adjustment.

[0090] In one embodiment, when the system obtains the comprehensive environmental impact factor, a threshold value determination process is first performed. Specifically, this process involves comparing the factor value with a preset threshold value, which is set based on historical production data. For example, in a plastic raw material extrusion production line, the threshold value is set to a dimensionless deviation index range of 0.5 to 1.0, which is calculated from the mean of the standard deviation of environmental variables on product diameter stability to reflect the potential impact. The judgment logic uses a simple comparison operation. If the factor exceeds the threshold value, the subsequent adjustment mechanism is triggered. This judgment ensures that the system responds sensitively to environmental fluctuations and can identify batches that need intervention in a timely manner in the plastic raw material extrusion process. In this way, the system avoids unnecessary adjustments and only activates the algorithm when the environmental impact is significant. Further, if the determination result shows that the comprehensive environmental impact factor exceeds the preset threshold value, a fuzzy logic control algorithm is applied for secondary fine-tuning of the traction speed. Fuzzy logic control algorithm is a control method for handling uncertainty and fuzzy information, which is based on the principle of converting input variables into fuzzy sets and outputting control decisions through rule-based reasoning. In the plastic raw material extrusion scenario, the algorithm first defines input variables such as the current traction speed deviation and the environmental factor deviation value, which are mapped to membership functions that describe the degree to which variables belong to low, medium, and high fuzzy levels. Then, the algorithm establishes a fuzzy rule base, for example, a rule may state that if the environmental factor is high and the speed deviation is medium, the fine-tuning amplitude is small. These rules are derived from production experience accumulation. Then, the rule is evaluated by the reasoning engine, and the fuzzy output is generated by combining all applicable rules. Finally, the defuzzification process is performed to convert the fuzzy output into an accurate traction speed adjustment value, such as a specific percentage increase or decrease. This algorithm is particularly suitable for handling nonlinear changes caused by the environment in the extrusion process, and can simulate the flexibility of human decision-making. In one possible implementation, the algorithm is integrated into a control module that processes data in real time to achieve fine-tuning.

[0091] In one possible implementation, the algorithm is integrated in a control module, processing data in real-time to implement the fine-tuning.

[0092] Preferably, the inputs of the fuzzy logic control algorithm include the synthetic environmental impact factor and the current extrusion parameters.

[0093] In particular, in a plastic product forming production line, the input variables can be further extended to temperature deviations or humidity effects, which are quantified and input to the algorithm. The design of the membership functions adopts a triangular or trapezoidal form, for example, for the draw speed deviation, the membership function defines a "low deviation" as a range of 0 to 5%, and a "high deviation" as above 10%. Through this function, the algorithm quantifies the uncertainty. The construction of the rule base involves expert knowledge, for example, in the processing of raw material batches, the number of rules can reach 20, covering different environmental scenarios. The reasoning process uses the min-max method, that is, the minimum value of the rule premise is taken as the activation, and then the maximum value of the outputs of multiple rules is taken. Defuzzification adopts the center of gravity method to calculate the weighted average of the fuzzy output to obtain the accurate fine-tuning amount. This detailed process ensures the accuracy of the algorithm in the control of the extrusion speed.

[0094] For example, in another embodiment, the secondary fine-tuning is adjusted for a specific raw material batch. In rubber extrusion production, if the synthetic environmental impact factor exceeds the threshold of 0.5, the algorithm first calculates the reference value of the current draw speed, such as based on the initial set speed of 10 meters per minute, and then applies the fuzzy output to generate a fine-tuning increment, for example, an increase of 2% to compensate for the diameter expansion deviation caused by environmental factors such as temperature. Through multiple iterations, the algorithm verifies the fine-tuning effect until the speed stabilizes. This fine-tuning emphasizes continuity, avoiding sudden changes that affect batch quality.

[0095] It can be understood that through the secondary fine-tuning of the fuzzy logic control algorithm, a final stable set of extrusion process parameters is obtained. The parameter set includes key values such as adjusted draw speed, extrusion pressure and temperature, forming a complete control set. In the field of raw material batch processing, for example, in the extrusion production line of fiber materials, the generation of the parameter set involves integrating the fine-tuning results with other parameters, for example, if the speed fine-tuning is to reduce 1 meter per minute, the pressure is adjusted accordingly to maintain uniformity. The algorithm ensures the stability of the parameter set, which is confirmed by the simulation verification cycle to be free of oscillation. This process highlights the role of the algorithm in optimizing extrusion parameters. Further, in one implementation, the generation of the parameter set can introduce a feedback mechanism.

[0096] In particular, the algorithm simulates the extrusion process after fine-tuning to check whether the parameter set causes the diameter deviation to be within an acceptable range, and if not stable, the fine-tuning step is repeated. This feedback is widely used in the processing of plastic raw materials, and can handle variability between batches.

[0097] In one embodiment, the resulting set of parameters is sent back to the extruder control system.

[0098] In particular, the sending back process is implemented through an internal communication interface, for example using Modbus protocol to transmit data packets containing the values of the set of parameters, such as the pulling speed of 12 meters per minute. The control system receives and updates the operating parameters, enabling continuous production adjustment. In the extrusion of metal wire batches, this sending back ensures seamless transition, avoiding production interruptions.

[0099] Preferably, the sending back mechanism refers to the process by which the system sends the adjustment parameters back to the device, and this mechanism includes a verification step to confirm the validity of the set of parameters, where the set of parameters refers to a set of motor speed adjustment parameters.

[0100] For example, after the system receives the set of parameters, a checksum is performed and verified, i.e. the sum S of the data of the set of parameters is calculated and compared with a preset value, and if it passes, it is applied to the extruder drive, i.e. the device that controls the extruder, to adjust the motor speed. This verification enhances the reliability of the adjustment, which is suitable for large production lines in the field of raw material processing.

[0101] For example, in the context of textile fiber raw material extrusion, the adjustment after sending back can monitor the batch output (yield and quality indicators of each batch of fibers) in real time, and if the set of parameters leads to improvement (yield increase of more than 10% or defect rate reduction of 5%), the system will store the set of parameters in the database as a benchmark for future optimization.

[0102] It should be noted that in the closed-loop control of the extrusion process, the personal information processing module is used for operator identity verification and parameter optimization to ensure system security and adapt to the environment. Through these steps, precise management of continuous production and data privacy protection are achieved.

[0103] If the technical solutions of the present application involve the collection, storage, use, processing, transmission, provision, disclosure or deletion of personal information, the products applying the technical solutions of the present application have clearly and understandably informed the personal information processing rules before processing the personal information, and have legally obtained the individual's independent consent. If the technical solutions of the present application involve sensitive personal information (such as biometric identification, religious beliefs, specific identity, medical health, financial accounts, whereabouts, etc.), the products applying the present application have obtained the individual's separate consent before processing sensitive personal information, and at the same time meet the requirement of "explicit consent", ensuring that individuals make authorization decisions voluntarily on the basis of full knowledge.

[0104] The specific embodiments include but are not limited to the following cases: at the camera, sensor and other personal information collection devices, set up clear, obvious mark, inform the relevant personnel has entered the personal information collection range, will be collected and processed, if the individual in the know still voluntarily enter the collection range, that is considered to agree to collect personal information; or in the terminal device or system interface of personal information processing, using obvious icons, text instructions and other ways to inform the personal information processing rules, and through the pop-up prompt, check the confirmation box or please upload their personal information and other interactive ways to obtain the explicit authorization of the individual.

[0105] The above-mentioned personal information processing rules should include but not limited to: the name and contact information of the personal information processor, the specific purpose of the personal information processing, the processing method, the type of personal information processed, the retention period and the way and procedure of the individual exercising the relevant rights and other information.

[0106] The above-mentioned personal information processing rules should include but not limited to: the name and contact information of the personal information processor, the specific purpose of the personal information processing, the processing method, the type of personal information processed, the retention period and the way and procedure of the individual exercising the relevant rights and other information.

[0106] The above-mentioned personal information processing rules should include but not limited to: the name and contact information of the personal information processor, the specific purpose of the personal information processing, the processing method, the type of personal information processed, the retention period and the way and procedure of the individual exercising the relevant rights and other information.

Claims

1. An adaptive control method for quality fluctuations in a wire extrusion process, characterized in that, include: Data on insulation layer thickness, sheath layer thickness, outer diameter, and surface quality at the extruder outlet are collected and processed using digital signal processing to obtain a standardized thickness uniformity index. The standardized thickness uniformity index is then categorized, and any deviations from a preset threshold are marked as quality fluctuation events. The fluctuation type is determined based on these events. A fuzzy logic control algorithm is used to adjust relevant parameters of the extruder temperature, screw speed, and traction speed according to the fluctuation type, resulting in an optimized temperature setpoint. The extruder control system parameters are updated based on the optimized temperature setpoint, and the operating status is fed back. The effectiveness of the adjustment is confirmed by comparing the operating status with the initial standardized thickness uniformity index. Once the adjustment is confirmed to be effective, raw material batch diameter and ambient temperature and humidity data are collected and fused to obtain a comprehensive environmental impact factor. Based on the comprehensive environmental impact factor, it is determined whether a secondary adjustment to the traction speed is needed, resulting in a final stable set of extrusion process parameters, which is then transmitted back to the control system.

2. The adaptive control method for quality fluctuation in wire extrusion process as described in claim 1, characterized in that, The process of collecting insulation layer thickness, sheath layer thickness, outer diameter, and surface quality data at the extruder outlet and processing them through digital signal processing to obtain a standardized thickness uniformity index includes: acquiring insulation layer thickness data, sheath layer thickness data, outer diameter data, and raw surface quality data through a sensor array; filtering the raw data to obtain filtered thickness sequence data, outer diameter sequence data, and surface quality sequence data; amplifying the filtered thickness sequence data and outer diameter sequence data to obtain enhanced thickness variation sequence and outer diameter variation sequence; marking defect locations and recording defect types in the filtered surface quality sequence data to obtain a surface defect recording sequence; extracting thickness deviation values, outer diameter deviation values, and corresponding defect marker information from the enhanced thickness variation sequence, outer diameter variation sequence, and surface defect recording sequence, and correcting the thickness deviation values ​​according to a pre-set weighting relationship to obtain a defect-corrected thickness deviation sequence; and normalizing the defect-corrected thickness deviation sequence to calculate a standardized thickness uniformity index.

3. The adaptive control method for quality fluctuation in wire extrusion process as described in claim 1, characterized in that, The classification judgment based on the standardized thickness uniformity index includes: acquiring temperature influence data at the extruder outlet and filtering it to obtain filtered temperature sequence data; correcting the deviation value sequence of the filtered temperature sequence data to obtain a corrected temperature deviation sequence; calculating a normalized index value by weighted averaging the corrected temperature deviation sequence with a preset thickness uniformity index; inputting the normalized index value into a support vector machine model for classification analysis to obtain a classification result; if the classification result shows a deviation from a preset threshold, it is marked as a quality fluctuation event and triggers the subsequent fluctuation type determination process.

4. The adaptive control method for quality fluctuation in wire extrusion process as described in claim 1, characterized in that, The step of adjusting extruder temperature, screw speed, and traction speed parameters using a fuzzy logic control algorithm after determining the fluctuation type based on the quality fluctuation event includes: acquiring a historical data sequence corresponding to the fluctuation type; calculating the deviation of the historical data sequence to obtain a temperature deviation correction sequence; weighting and fusing the temperature deviation correction sequence with historical screw speed adjustment values ​​to obtain traction speed optimization parameters; using a fuzzy logic control algorithm to perform fuzzification transformation on the traction speed optimization parameters and reasoning through a preset rule base to obtain a preliminary set value; and fusing the preliminary set value with historical extruder parameter monitoring data to obtain an optimized temperature set value.

5. The adaptive control method for quality fluctuation in wire extrusion process as described in claim 1, characterized in that, The step of updating the extruder control system parameters and feeding back the operating status based on the optimized temperature setpoint includes: updating the temperature parameters in the extruder control system according to the optimized temperature setpoint to obtain an updated temperature parameter sequence; obtaining the traction speed linkage data corresponding to the updated temperature parameter sequence to determine the operating status index; extracting real-time sensing features from the operating status index and performing matching judgment; fusing the matched features to form an updated extrusion process operating status; and feeding back the updated extrusion process operating status to the real-time sensing database for storage.

6. The adaptive control method for quality fluctuation in wire extrusion process as described in claim 1, characterized in that, The step of confirming the effectiveness of the adjustment by comparing the operating status with the initial standardized thickness uniformity index includes: extracting the updated extrusion process operating status and the initial standardized thickness uniformity index from the real-time sensing database to form comparison data; obtaining a linkage traction speed verification sequence based on the speed-related part of the comparison data; determining the quality fluctuation trend based on the verification sequence; if the quality fluctuation trend shows a decrease in fluctuation, then fusing the thickness uniformity index and the fluctuation trend data to obtain a parameter adjustment verification sequence; comparing and judging the indicators based on the parameter adjustment verification sequence to obtain the adjustment confirmation basis; and generating an adjustment confirmation signal based on the adjustment confirmation basis.

7. The adaptive control method for quality fluctuation in wire extrusion process as described in claim 1, characterized in that, The step of collecting raw material batch diameter and environmental temperature and humidity data and fusing them to obtain a comprehensive environmental impact factor after confirming the adjustment is effective includes: activating the environmental factor monitoring module according to the adjustment confirmation signal; obtaining raw material batch diameter data and environmental temperature and humidity data from the environmental factor monitoring module; fusing the raw material batch diameter data and environmental temperature and humidity data using data fusion technology to obtain a preliminary environmental dataset; performing batch variation assessment on the preliminary environmental dataset and fusing the adjustment confirmation signal to obtain a comprehensive environmental impact sequence; and extracting the fusion result from the comprehensive environmental impact sequence to obtain the comprehensive environmental impact factor.

8. The adaptive control method for quality fluctuation in wire extrusion process as described in claim 1, characterized in that, The step of determining whether a secondary adjustment to the traction speed is needed based on the comprehensive environmental impact factor includes: comparing the comprehensive environmental impact factor with a preset threshold to obtain a judgment result; if the judgment result shows that it exceeds the preset threshold, then using a fuzzy logic control algorithm to perform a secondary fine-tuning of the current traction speed value to obtain an adjusted speed value; and then fusing the adjusted speed value with the raw material batch diameter data and environmental temperature and humidity data to determine the final stable extrusion process parameter set.

9. The adaptive control method for quality fluctuation in wire extrusion process as described in claim 1, characterized in that, The step of obtaining the final stable extrusion process parameter set and transmitting it back to the control system includes: transmitting the final stable extrusion process parameter set back to the extruder control system; updating the current operating parameters of the extruder according to the final stable extrusion process parameter set; and enabling the extruder to continuously operate according to the updated parameter set to achieve continuous production adjustment.

10. The adaptive control method for quality fluctuation in a wire extrusion process as described in claim 1, characterized in that, The standardized thickness uniformity index, the quality fluctuation event, the fluctuation type, the optimized temperature setpoint, the comprehensive environmental impact factor, and the final stable extrusion process parameter set are all used cyclically in the closed-loop control process to achieve continuous adaptive adjustment of the extrusion process quality.

Images