Cable sheath die change control method and system

Abstract

The invention belongs to the technical field of cable sheath production, and discloses a cable sheath die change control method and system.The method comprises the steps that A1, according to the specification of a target sheath, initial process parameters are determined through a historical reference model to start extrusion equipment; a2, in the operation process of extrusion equipment, collecting real-time process parameters and corresponding real-time sheath product quality data; a3, calculating the parameter compensation adjustment amount between the current working condition and the historical reference working condition according to the real-time process parameters and the corresponding real-time sheath product quality data; a4, combining the initial process parameter with the parameter compensation adjustment amount to obtain a process parameter adjustment value aiming at the current working condition, and sending the process parameter adjustment value to a control system of the extrusion equipment so as to enable the extrusion equipment to operate according to the process parameter adjustment value; a5, repeating the steps A2 to A4 until the quality of the sheath product meets the target requirement; therefore, the efficiency and precision of process parameter adjustment after die replacement can be improved, the dependence on the experience of operators is reduced, and the rejection rate is reduced.

Description

Technical Field

[0001] This application relates to the technical field of cable sheath production, and in particular, to a die change control method and system for a cable sheath. Background Art

[0002] The production of cable sheaths is a key link in the cable manufacturing process, and the extrusion process is usually adopted. In this process, the polymer material is heated and plasticized in an extruder, and then extruded through a die with a specific shape, and the final product is obtained after cooling, curing and traction. In order to meet the requirements of different cable specifications or materials, the extrusion die often needs to be replaced during the production process.

[0003] After changing the die, the process parameters of the extrusion equipment, such as extrusion temperature, screw speed, traction speed, etc., need to be reset and adjusted to ensure that the cable sheath produced meets the quality requirements (such as outer diameter, wall thickness, roundness, etc.). The traditional parameter adjustment method highly depends on the experience of the operator for trial-and-error adjustment. This trial-and-error adjustment method is inefficient, time-consuming, and prone to producing many unqualified products, resulting in waste of raw materials. At the same time, the efficiency and accuracy of the adjustment are greatly affected by the skill level of the operator, and it is difficult to ensure the stability and consistency of production.

[0004] In order to improve the efficiency and accuracy of parameter adjustment, the industry has tried to train machine learning models using historical production data in order to learn the relationship between process parameters and product quality, so as to recommend initial process parameters after die change. However, the cable sheath production line usually runs continuously for a long time, and the key components of the equipment will experience normal wear or aging, such as wear of the screw, barrel and die. In addition, even for raw materials of the same brand, there may be slight characteristic differences between different batches. These changes in equipment status and raw material characteristics are hidden factors that are difficult to directly and accurately measure in real time.

[0005] When using the model trained based on historical production data to recommend initial process parameters, the challenge is that the historical data was collected under specific equipment status and raw material batches, and the actual working conditions during current production may deviate from the historical working conditions. This deviation in working conditions caused by hidden factors that are difficult to directly perceive makes it possible that even if production is started according to the initial parameters recommended by the model, the actual product quality may not fully meet the target requirements, or it may take a long time to reach a stable and qualified state.

[0006] In this case, the operator still needs to make manual fine-tuning according to the actual product quality deviation. However, due to the dynamic and difficult-to-quantify impacts of equipment status and material properties, and the complex relationship between process parameters and product quality, it is difficult for the operator to accurately judge the cause of the deviation and determine the appropriate adjustment direction and adjustment amount. This results in an inefficient adjustment process, prolongs the die change time, and generates waste products. More importantly, equipment wear and raw material batch changes occur continuously, and the production conditions are always changing slowly. The accuracy of the recommended parameters of the model trained based on fixed historical data will gradually decline. How to enable the parameter adjustment process to adapt to the current specific working conditions and provide effective adjustment guidance, reducing the dependence on the operator's experience, becomes a key technical problem in the presence of these difficult-to-directly-perceive hidden factors. The existing technologies lack a method and system that can effectively sense or adapt to the working condition deviation caused by difficult-to-quantify hidden factors and perform accurate, rapid, and adaptive adjustment of process parameters accordingly.

[0007] In view of the above problems, there is an urgent need for improvement in the existing technologies. Summary of the Invention

[0008] The purpose of this application is to provide a method and system for controlling die change of cable sheaths, which can improve the efficiency and accuracy of process parameter adjustment after die change, reduce the dependence on the operator's experience, and lower the scrap rate.

[0009] In a first aspect, this application provides a method for controlling die change of cable sheaths, which is used to adjust and control the process parameters of an extrusion device after die change of the cable sheath extrusion device. The steps of this method include: A1. According to the target sheath specification, use the historical benchmark model to determine the initial process parameters for starting the extrusion device; A2. During the operation of the extrusion device, collect the real-time process parameters and the corresponding real-time sheath product quality data; A3. According to the real-time process parameters and the corresponding real-time sheath product quality data, calculate the parameter compensation adjustment amount between the current working condition and the historical benchmark working condition; A4. Combine the initial process parameters with the parameter compensation adjustment amount to obtain the process parameter adjustment value for the current working condition, and send it to the control system of the extrusion device, so that the extrusion device operates according to the process parameter adjustment value; A5. Repeat steps A2 - A4 until the sheath product quality meets the target requirements.

[0010] In a second aspect, this application provides a control system for die change of cable sheaths, which is used to adjust and control the process parameters of an extrusion device after die change of the cable sheath extrusion device. The system includes: An initial parameter recommendation module, which is used to determine initial process parameters according to the target sheath specification by using a historical benchmark model to start the extrusion equipment; A real-time data acquisition module, which acquires real-time process parameters and corresponding real-time sheath product quality data during the operation of the extrusion equipment; A working condition compensation calculation module, which is used to calculate the parameter compensation adjustment amount between the current working condition and the historical benchmark working condition according to the real-time process parameters and the corresponding real-time sheath product quality data; A parameter comprehensive adjustment module, which is used to combine the initial process parameters with the parameter compensation adjustment amount to obtain a process parameter adjustment value for the current working condition and send it to the control system of the extrusion equipment, so that the extrusion equipment operates according to the process parameter adjustment value; An iterative control module, which is used to trigger the real-time data acquisition module, the working condition compensation calculation module and the parameter comprehensive adjustment module to run repeatedly until the sheath product quality meets the target requirements.

[0011] Beneficial effects: A cable sheath die change control method and system provided by the present application effectively solve the problems of low efficiency of parameter adjustment after die change, dependence on the experience of operators and difficulty in adapting to equipment aging and material batch changes in the prior art by determining initial parameters using a historical benchmark model and performing iterative compensation adjustment in combination with real-time data. It has the advantages of improving the efficiency and accuracy of process parameter adjustment after die change, reducing the dependence on the experience of operators and reducing the scrap rate. Description of the Drawings

[0012] Figure 1 It is a flowchart of the cable sheath die change control method provided by the embodiment of the present application.

[0013] Figure 2 It is a structural schematic diagram of the cable sheath die change control system provided by the embodiment of the present application.

[0014] Label description: 1. Initial parameter recommendation module; 2. Real-time data acquisition module; 3. Working condition compensation calculation module; 4. Parameter comprehensive adjustment module; 5. Iterative control module. Detailed Embodiments

[0015] The technical solutions in the present application will be clearly and completely described below with reference to the accompanying drawings in the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the present application usually described and illustrated in the accompanying drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the present application claimed, but only represents the selected embodiments of the present application. All other embodiments obtained by those skilled in the art based on the embodiments of the present application without creative efforts belong to the scope of protection of the present application.

[0016] It should be noted that: similar reference numerals and letters denote similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. At the same time, in the description of the present application, the terms "first", "second", etc. are only used for distinguishing descriptions and cannot be understood as indicating or implying relative importance.

[0017] Referring to Figure 1 , the present application proposes a method for controlling die change of a cable sheath, which is used to adjust and control the process parameters of an extrusion device after the die of the cable sheath extrusion device is changed. The steps of the method include: A1. According to the target sheath specification, use the historical benchmark model to determine the initial process parameters for starting the extrusion device; A2. During the operation of the extrusion device, collect the real-time process parameters and the corresponding real-time sheath product quality data; A3. Calculate the parameter compensation adjustment amount between the current working condition and the historical benchmark working condition according to the real-time process parameters and the corresponding real-time sheath product quality data; A4. Combine the initial process parameters with the parameter compensation adjustment amount to obtain the process parameter adjustment value for the current working condition, and send it to the control system of the extrusion device so that the extrusion device operates according to the process parameter adjustment value; A5. Repeat steps A2 - A4 until the sheath product quality meets the target requirements.

[0018] Among them, the historical benchmark model refers to a mathematical model trained based on historical production data and capable of reflecting the relationship between process parameters and sheath product quality. It can be implemented by using machine learning algorithms, such as neural network models, support vector machine models, decision tree models. It is mainly used to provide a reasonable starting point for the initial process parameters for the production of new specification products.

[0019] Among them, the real-time process parameters refer to the real-time values of extrusion temperature, screw speed, traction speed, etc. collected by sensors during the actual operation of the extrusion equipment. It can be achieved by using the sensors and data acquisition modules of the industrial automation control system, such as temperature sensors, speed sensors, and velocity sensors, which are mainly used to reflect the operating state of the current equipment.

[0020] Among them, the real-time sheath product quality data refers to the quality index values obtained by measuring the actually produced sheath products during the operation of the extrusion equipment. It can be achieved by using on-line or off-line measuring equipment, such as outer diameter measuring instruments, wall thickness measuring instruments, and roundness measuring instruments, which are mainly used to evaluate whether the quality of the current product meets the requirements.

[0021] Among them, the parameter compensation adjustment amount refers to the adjustment value calculated according to the current real-time data and used to correct the initial process parameters to adapt to the current actual production conditions. It can be calculated by using methods based on model prediction error or sensitivity analysis, such as reverse derivation based on local sensitivity coefficients, which is mainly used to make up for the deviation between the historical benchmark model and the current actual conditions.

[0022] Among them, the process parameter adjustment value refers to the final parameter setting value used to control the extrusion equipment after combining the initial process parameters with the calculated parameter compensation adjustment amount. It can be obtained by using simple addition or more complex weighted fusion methods, which is mainly used to guide the extrusion equipment to operate according to parameters more suitable for the current working conditions.

[0023] The core innovation point of this application is that by combining the initial process parameters determined based on the historical benchmark model with the parameter compensation adjustment amount calculated according to the real-time data, the adaptive adjustment of the process parameters of the extrusion equipment is realized, achieving the effect of quickly and accurately making the quality of the sheath product meet the target requirements.

[0024] The solution of this application first uses a historical benchmark model to provide a reasonable initial startup parameter for the extrusion equipment, avoiding the need to rely entirely on manual experience for exploration. After the equipment starts and runs, the solution continuously collects real-time process parameters and product quality data. These real-time data reflect the current actual production conditions, including the influence of hidden factors such as equipment status and material characteristics that are difficult to directly quantify. Based on these real-time data, the solution calculates the deviation between the current working condition and the historical benchmark working condition, and converts it into a parameter compensation adjustment amount (instead of converting the deviation between the actual quality index and the target sheath specification into a parameter compensation adjustment amount. By calculating the deviation between the current working condition and the historical benchmark working condition, this solution can more comprehensively reflect the influence of these hidden factors, thereby making more targeted parameter compensation adjustments). This compensation adjustment amount is a correction to the initial parameter, making it more suitable for the current actual production environment (this means that the adjusted parameter can not only make the product quality approach the target specification, but also better adapt to the current equipment status and material characteristics, thereby improving the efficiency and accuracy of the adjustment and ensuring the stability of the production process). Subsequently, the initial parameter and the compensation amount are combined to form a new process parameter adjustment value and sent to the equipment control system for execution. The entire process forms a closed loop. By continuously collecting real-time data, calculating compensation, and adjusting parameters until the product quality meets the target requirements. This iterative adjustment method enables the system to dynamically adapt to the changes in the working conditions during the production process, gradually optimize the parameter settings, and finally stably produce qualified products.

[0025] Through the above technical solution, this application can effectively solve the problems of low production efficiency and unstable product quality caused by relying on manual experience to adjust process parameters after die change of the cable sheath extrusion equipment. Using the historical benchmark model to provide initial parameters shortens the startup and debugging time. By collecting real-time data and calculating the parameter compensation adjustment amount, the system can adapt to the current actual production conditions and overcome the influence of hidden factors such as equipment aging and material batch differences. Based on the closed-loop iterative control of compensation adjustment, rapid and accurate optimization of process parameters is achieved, reducing the generation of waste products during the trial-and-error process. Ultimately, the production efficiency and product quality stability after die change are improved, and the dependence on the experience of operators is reduced.

[0026] In some embodiments, the historical benchmark model includes multiple parameter optimization sub-models, and each parameter optimization sub-model corresponds to a quality index; the parameter optimization sub-model takes process parameters as input and the corresponding quality index as output; the quality indexes include outer diameter, wall thickness, and roundness; the process parameters include extrusion temperature, screw speed, and traction speed; Step A1 includes: A101. Extract multiple quality indexes from the target sheath specification, including outer diameter, wall thickness, and roundness; A102. For each quality indicator, call the corresponding parameter optimization sub-model respectively; A103. Based on the preset weight allocation strategy, allocate weights to each parameter optimization sub-model; this weight reflects the importance of the corresponding quality indicator in the overall quality assessment; A104. Adopt a multi-objective optimization algorithm, aiming to minimize the deviation between the weighted output of all parameter optimization sub-models and the corresponding quality indicators in the target sheath specification, solve to obtain a set of initial process parameters, and use this set of initial process parameters as the startup parameters of the extrusion equipment.

[0027] Among them, the parameter optimization sub-model refers to the model unit in the historical benchmark model that performs parameter optimization or prediction for a specific quality indicator. Each sub-model focuses on describing the relationship between process parameters and a specific quality indicator, and can be implemented by an independent prediction function. Its purpose is to decompose the complex parameter-multi-quality indicator relationship.

[0028] Among them, the quality indicator refers to the key parameters for measuring the quality of the cable sheath product, including outer diameter, wall thickness, and roundness. These indicators directly reflect the geometric dimensions and shape accuracy of the product, and their purpose is to quantify whether the product meets the specification requirements.

[0029] Among them, the process parameter refers to the key control variables that affect the cable sheath extrusion process and product quality, including extrusion temperature, screw speed, and traction speed. These parameters are the inputs that the control system can adjust, and their purpose is to control the product quality by adjusting these parameters.

[0030] Among them, the target sheath specification refers to the quality standard of the cable sheath product required for this production task, including the specific numerical requirements for outer diameter, wall thickness, and roundness. Its purpose is to provide a target for parameter setting and quality control.

[0031] Among them, the weight allocation strategy refers to the method for determining the importance of different quality indicators in the overall optimization goal, which can be set based on product type or importance. Its purpose is to reflect the relative importance of different quality indicators in multi-objective optimization.

[0032] Among them, the multi-objective optimization algorithm refers to a class of algorithms used to solve optimization problems with multiple objective functions, and can be implemented by a genetic algorithm. Its purpose is to find a set of solutions that can optimize multiple objectives simultaneously.

[0033] Among them, the weighted output refers to the result obtained by weighted combination of the outputs of multiple parameter optimization sub-models according to the preset weights. Its purpose is to integrate the predicted values of multiple quality indicators into a comprehensive evaluation value.

[0034] Among them, deviation minimization refers to optimizing the objective function of the algorithm, aiming to minimize the difference between the weighted output of the parameter optimization sub-model and the weighted target value of the corresponding quality index in the target sheath specification, with the purpose of finding a combination of process parameters that makes the predicted quality closest to the target quality.

[0035] The solution of this application refines the historical benchmark model into multiple parameter optimization sub-models, with each sub-model focusing on one quality index, so as to be able to more accurately capture the relationship between process parameters and each quality index. When determining the initial process parameters, first extract the quality indexes in the target sheath specification and call the prediction results of the corresponding parameter optimization sub-models. By introducing a weight assignment strategy, the importance of different quality indexes can be adjusted according to actual needs, enabling the optimization process to preferentially meet the requirements of key indexes. Subsequently, a multi-objective optimization algorithm is used to solve for a set of initial process parameters with the goal of minimizing the weighted deviation between the weighted prediction values of all parameter optimization sub-models and the weighted target specification. This method can consider multiple quality indexes such as outer diameter, wall thickness, and roundness simultaneously and find a balance among them, avoiding the problem that a single model is difficult to take into account all indexes. Using this set of initial parameters that balance multiple quality indexes to start the extrusion equipment can make the product closer to the target specification at the start-up stage, providing a better starting state for subsequent real-time adjustment. This helps to reduce the adjustment time and waste generation in the start-up stage and improves the die change efficiency.

[0036] Through the above technical solution, refining the historical benchmark model and combining multi-objective optimization can better balance multiple quality indexes such as outer diameter, wall thickness, and roundness when determining the initial process parameters, enabling the extrusion equipment to obtain product quality closer to the target specification at the start-up stage, reducing the difficulty and time of subsequent adjustment, improving the die change efficiency, and reducing the scrap rate.

[0037] Preferably, step A103 may include: For the target sheath specification, query the preset mapping relationship table of product specifications and weight assignment strategies to determine the weight assignment strategy matching the current target sheath specification; If a matching weight assignment strategy is queried, directly assign weights to each parameter optimization sub-model according to the queried strategy; if no matching weight assignment strategy is queried, assign weights to each parameter optimization sub-model according to the preset default weight assignment strategy or according to the operator's custom settings. For the aging degree of the extrusion equipment, according to the preset mapping relationship table between the equipment aging degree and the weight adjustment coefficient, determine the weight adjustment coefficient that matches the current aging degree of the extrusion equipment, and adjust the allocated weight based on this weight adjustment coefficient; wherein, the weight adjustment coefficient is used to adjust the relative importance between different quality indicators to adapt to the impact brought by equipment aging.

[0038] Among them, the preset mapping relationship table between the product specification and the weight allocation strategy refers to a data structure used to store the association relationship between different target sheath specifications and the corresponding weight allocation strategies. Specifically, it can be a lookup table or a database, and its purpose is to quickly obtain the corresponding weight configuration according to the target sheath specification currently being produced.

[0039] Among them, the weight allocation strategy refers to a set of numerical values, and each numerical value corresponds to a parameter optimization sub-model, indicating the relative importance of the quality indicator corresponding to this parameter optimization sub-model when determining the initial process parameters. Specifically, it can be a set of normalized weight numerical values, and its purpose is to quantify the priority of different quality indicators in the overall optimization goal.

[0040] Among them, the preset default weight allocation strategy refers to the basic weight configuration adopted when no strategy matching the current target sheath specification is found in the preset mapping relationship table between the product specification and the weight allocation strategy. Its purpose is to provide a general and fallback weight allocation scheme.

[0041] Among them, the operator's custom setting means that the operator is allowed to manually input or modify the weight allocation value according to actual production experience or specific requirements. Its purpose is to increase the flexibility of the system and the ability of manual intervention.

[0042] Among them, the preset mapping relationship table between the equipment aging degree and the weight adjustment coefficient refers to another data structure used to store the association relationship between different extrusion equipment aging degrees and the corresponding weight adjustment coefficients. Specifically, it can be a lookup table or a database, and its purpose is to obtain the adjustment factor for correcting the weight according to the current aging state of the equipment.

[0043] Among them, the weight adjustment coefficient refers to the numerical value used to correct the allocated weight. Specifically, it can be a multiplication factor or an addition amount, and its purpose is to dynamically adjust the relative importance of different quality indicators to better adapt to the impact of equipment aging on product quality.

[0044] The solution of this application dynamically adjusts the weight allocation of the parameter optimization sub-model by combining the target sheath specifications and the aging degree of the extrusion equipment. First, according to the current target sheath specifications, the system queries the preset mapping relation table to obtain the weight allocation strategy matching this specification. This weight allocation based on product specifications enables the system to prioritize the indicators that have a greater impact on the quality of different types and sizes of sheath products. For example, for thin-wall sheaths, wall thickness uniformity may be more concerned, while for large outer diameter sheaths, roundness may be more concerned. If there is no preset strategy for a specific specification, the default strategy is adopted or the operator is allowed to set it manually, ensuring the universality and flexibility of the solution. On this basis, the aging degree of the extrusion equipment is further considered. Equipment aging will affect the sensitivity of some process parameters to specific quality indicators. For example, screw wear may cause the wall thickness to be less sensitive to the change of screw speed. The aging degree can be pre-evaluated by integrating factors such as the cumulative running time of the equipment, the change of equipment energy consumption, and the vibration of equipment operation (such as using an evaluation model or expert experience evaluation). By querying the mapping relation table between the equipment aging degree and the weight adjustment coefficient, the system obtains the adjustment coefficient corresponding to the current equipment aging state. These adjustment coefficients are applied to the weights determined according to the product specifications before to perform a secondary correction on the relative importance of different quality indicators. For example, if equipment aging causes the wall thickness to be more prone to deviation, the weight corresponding to the wall thickness may be increased by the adjustment coefficient, so as to give a higher priority to the wall thickness index during parameter optimization.

[0045] Thus, through the dual consideration and dynamic adjustment of product specifications and equipment aging degree, the weight allocation for the multi-objective optimization algorithm is finally determined. This mechanism of dynamically adjusting weights enables the parameter optimization process based on the historical benchmark model to more accurately reflect the requirements and challenges of the current actual production conditions, improving the recommendation accuracy of the initial process parameters. Compared with the solution that only relies on fixed weights or single-factor adjustment, it can more effectively cope with the uncertainties brought by product diversity and equipment state changes in actual production, thereby providing a more accurate starting point for subsequent process parameter adjustment, reducing the number of trial-and-error times and adjustment time, and improving the production efficiency and product quality stability after die change.

[0046] Preferably, step A104 may include: Determine the constraint conditions of the multi-objective optimization algorithm; the constraint conditions include the upper and lower limits of the extrusion temperature, the upper and lower limits of the screw speed, and the upper and lower limits of the traction speed; Initialize the parameters of the multi-objective optimization algorithm; the initialized parameters include the population size, the crossover probability, and the mutation probability; Adopt the genetic algorithm as the multi-objective optimization algorithm, and take the minimization of the deviation between the weighted output of all parameter optimization sub-models and the corresponding quality indicators in the target sheath specifications as the objective function; Under the described constraints, based on the objective function, through iterative optimization using the genetic algorithm, a Pareto optimal solution set is obtained; each solution in the Pareto optimal solution set corresponds to a set of initial process parameters. From the Pareto optimal solution set, according to the preset evaluation index, a set of initial process parameters is selected as the starting parameters of the extrusion equipment; the evaluation index comprehensively considers the deviation degree between each quality index and the target sheath specification as well as the stability of the process parameters.

[0047] Among them, the constraint condition refers to the restriction imposed on the value range of the optimization variable or the relationship between variables, and it can be processed by the penalty function method, the feasibility rule method or the special operator method.

[0048] Among them, initializing the parameters of the multi-objective optimization algorithm means setting the key configuration items that affect the behavior and performance of the algorithm before it starts running, and it can be determined by the random initialization method, the initialization method based on prior knowledge or the adaptive initialization method.

[0049] Among them, the genetic algorithm refers to a search algorithm that simulates natural selection and genetic mechanisms, and iteratively improves the solution set through operations such as selection, crossover and mutation. It can be implemented by the standard genetic algorithm, the elitist retention genetic algorithm or the genetic algorithm with constraint handling. The objective function refers to the mathematical expression that needs to be minimized or maximized in the optimization problem, and is used to measure the quality of the solution. It can be constructed as the weighted sum of each sub-objective function, the Chebyshev function or the function based on the ideal point.

[0050] Among them, the Pareto optimal solution set refers to the set composed of all Pareto optimal solutions in the multi-objective optimization problem, where the Pareto optimal solution refers to the solution that cannot improve any one objective without deteriorating at least one objective. The process of obtaining the Pareto optimal solution set through iterative optimization using the genetic algorithm is the prior art and will not be elaborated here.

[0051] Among them, the evaluation index refers to the standard used to measure the quality or applicability of each solution in the Pareto optimal solution set, and it can include dimensions such as the degree of closeness to the target, the robustness of the solution, the difficulty of implementation, and the adaptability to the equipment state.

[0052] In the process of solving the initial process parameters using a multi-objective optimization algorithm in the solution of this application, the constraint conditions of the multi-objective optimization algorithm are first clearly determined. These constraint conditions directly correspond to the physical limitations and process requirements during the actual operation of the extrusion equipment, such as the upper and lower limits of the extrusion temperature, screw speed, and traction speed. Due to the introduction of these constraint conditions, the search space of the optimization algorithm is limited within the actual feasible range, thus avoiding obtaining parameter combinations that do not conform to the actual production, and ensuring the effectiveness and safety of the solution results. At the same time, by initializing the parameters of the multi-objective optimization algorithm, a basis is set for the subsequent iterative optimization process. Reasonable parameter settings help improve the convergence efficiency and solution quality of the algorithm. Further, the genetic algorithm is used as the specific multi-objective optimization algorithm, and the objective function is to minimize the deviation between the weighted output of all parameter optimization sub-models and the corresponding quality indicators in the target sheath specification, enabling the algorithm to effectively search in the multi-objective space and find a series of Pareto optimal solutions that achieve a balance between different quality indicators. Each solution in these Pareto optimal solution sets represents a set of potential initial process parameters. However, simply obtaining the Pareto optimal solution set is not sufficient to be directly used in production because the performance of different solutions may vary in actual production. Therefore, this application further selects a set of initial process parameters from the Pareto optimal solution set according to a preset evaluation index. This evaluation index comprehensively considers the deviation degree of each quality indicator from the target sheath specification and the stability of the process parameters, which means that the selected parameters should not only make the product quality close to the target but also ensure the smooth operation of the equipment.

[0053] Due to this comprehensive evaluation and selection mechanism, the finally determined initial process parameters can not only guide the product quality to quickly approach the target specification but also reduce the production risks caused by parameter fluctuations or instability. Applying such a set of initial process parameters obtained through constrained optimization and comprehensive evaluation screening to the startup of the extrusion equipment can provide more reliable and safer startup parameters compared with only relying on the historical benchmark model or the optimization method that does not consider the actual constraints and parameter stability, thus shortening the debugging time after mold change, reducing waste generation, and improving the mold change efficiency and production stability. Combining this method with the method based on the historical benchmark model and parameter optimization sub-models further optimizes the determination process of the initial parameters on the basis of utilizing the prediction ability of historical data by introducing actual constraints and stability considerations, making the entire mold change control method more robust and practical.

[0054] Further, the step of selecting a set of initial process parameters from the Pareto optimal solution set according to a preset evaluation index as the startup parameters of the extrusion equipment may include: B1. For each solution in the Pareto optimal solution set, calculate the deviation degree between its corresponding quality indicators and the target sheath specifications. The deviation degree includes absolute deviation and relative deviation, and normalize the absolute deviation and relative deviation to obtain the normalized deviation degree. B2. For each solution in the Pareto optimal solution set, evaluate the stability of its corresponding process parameters. The stability of the process parameters is determined by calculating the fluctuation range of each process parameter in the historical production data. The smaller the fluctuation range, the higher the stability. B3. According to the aging degree of the extrusion equipment, perform aging correction on the stability to obtain the parameter stability after aging correction. B4. According to the target sheath specifications and the aging degree of the extrusion equipment, determine the weight factors of the normalized deviation degree and the parameter stability after aging correction. B5. Using the method of weighted summation, according to the determined weight factors, comprehensively score the normalized deviation degree and the parameter stability after aging correction to obtain the comprehensive score value of each solution. B6. Select the solution with the highest comprehensive score value as the starting parameters of the extrusion equipment.

[0055] Among them, normalization refers to converting data with different dimensions or ranges to a unified scale, which can be achieved by methods such as min-max normalization or Z-score normalization. Its purpose is to eliminate the influence of the dimensions of different quality indicator deviations and parameter stabilities, so that they can be effectively compared and weighted.

[0056] Among them, the stability of the process parameters is determined by calculating the fluctuation range of each process parameter in the historical production data. The fluctuation range can be measured by statistical quantities such as standard deviation, range (the difference between the maximum value and the minimum value), or interquartile range. The reciprocal of the fluctuation range can be normalized to obtain the evaluation result of the stability of the corresponding process parameter.

[0057] Among them, aging correction refers to the adjustment of the parameter stability evaluation result according to the aging degree of the extrusion equipment. It can be achieved by looking up the mapping relationship table based on the aging degree of the equipment and the correction coefficient, or by establishing an equipment aging model to calculate the correction coefficient, etc. Its purpose is to compensate for the influence of equipment aging on parameter stability, so that parameter selection is more in line with the current equipment state. The numerical range of the parameter stability after aging correction is [0,1], and the larger the value, the more stable.

[0058] Among them, the weight factor refers to a value used to measure the relative importance of the normalized deviation degree and the parameter stability after aging correction in the comprehensive score. It can be determined by methods such as setting based on expert experience, determining based on historical data analysis, or reasoning based on fuzzy rules. Its purpose is to balance the quality requirements and parameter stability requirements in a targeted manner according to the target sheath specification and the equipment aging degree.

[0059] The solution of this application conducts multi-dimensional evaluation on each candidate solution in the Pareto optimal solution set, so as to select the initial process parameters most suitable for the current working conditions. First, calculate the deviation degree between the quality index corresponding to each solution and the target specification, and perform normalization processing, which is a quantitative evaluation of the product quality compliance. At the same time, evaluate the stability of the process parameters corresponding to each solution in historical production. The stability reflects the reliability of the parameters in actual applications. The key lies in performing aging correction on the parameter stability according to the aging degree of the extrusion equipment, because equipment aging will affect the actual fluctuation and effect of the parameters. Through correction, the stability evaluation is made closer to the true state of the current equipment. Then, according to the target sheath specification and the equipment aging degree, dynamically determine the weight factors of the quality deviation and the stability after aging correction, which enables the parameter selection to take into account the different emphases on quality and stability of different products and the influence of the equipment state. Finally, use weighted summation to comprehensively score the normalized deviation degree and the parameter stability after aging correction. A solution with a high score means better performance in meeting quality requirements, ensuring parameter stability, and adapting to equipment aging. Selecting the solution with the highest comprehensive score as the starting parameter is precisely based on this comprehensive consideration, thereby improving the accuracy of the initial parameters and reducing the need for subsequent adjustments. It is precisely because on the basis of the Pareto optimal solution set, the correction of the parameter stability by the equipment aging degree and the dynamic determination of the weight factor are further introduced, so that the initial parameters selected from the solution set can better adapt to the actual working conditions of the current equipment, thus solving the problem that only considering the quality deviation and the original stability is not sufficient to cope with the challenges brought by equipment aging, and improving the die change efficiency and the stability of product quality.

[0060] Through the above technical solution, when selecting the initial process parameters from the Pareto optimal solution set, not only the deviation degree between the quality index and the target specification and the stability of the process parameters are considered, but also the parameter stability is corrected according to the aging degree of the extrusion equipment, and the weight of the evaluation index is dynamically determined according to the target sheath specification and the equipment aging degree. This parameter selection method that comprehensively considers the actual state of the equipment enables the selected initial process parameters to better adapt to the current production working conditions, improves the accuracy and applicability of the initial parameters, reduces the need and time for manual parameter adjustment after die change, thereby improving the die change efficiency of cable sheath production and enhancing the stability of product quality.

[0061] In some embodiments, step A2 includes: respectively disturbing each process parameter to obtain the disturbed process parameter, and collecting the real-time sheath product quality data corresponding to the disturbed process parameter; the process parameters include extrusion temperature, screw speed, and traction speed, and the real-time sheath product quality data includes three quality indicators: outer diameter, wall thickness, and roundness; Step A3 includes: A301. Based on the process parameters before and after the disturbance and the corresponding real-time sheath product quality data, calculate the local sensitivity coefficient of each process parameter to each quality indicator; the local sensitivity coefficient characterizes the influence degree of the change of the process parameter on the quality indicator under the current working condition; A302. Based on the historical benchmark model, calculate the predicted values of each quality indicator under the current process parameters; A303. For each quality indicator, calculate the deviation between the indicator value in the current real-time sheath product quality data and the corresponding predicted value, denoted as the prediction deviation; A304. According to the local sensitivity coefficient, inversely deduce the amount of adjustment required for each process parameter, denoted as the parameter adjustment amount, to reduce the prediction deviation; A305. Integrate the parameter adjustment amounts corresponding to the prediction deviations of each quality indicator, and use the weighted average method to calculate the final parameter compensation adjustment amount.

[0062] Among them, the disturbance refers to a small-scale, planned change to the selected process parameter during the normal operation of the extrusion equipment without significantly affecting the product qualification, which can be achieved by means of pseudo-random sequence, step signal, pulse signal, etc.

[0063] Among them, the local sensitivity coefficient refers to the change amount caused by a unit change of a certain process parameter to each quality indicator under the current production working condition, which can be calculated by methods such as linear regression, non-linear fitting, or difference calculation based on the disturbance data, and its purpose is to quantify the influence degree of different process parameters on each quality indicator near the current working point.

[0064] Among them, the inverse deduction refers to calculating the change amount of the process parameter required to eliminate or reduce the deviation according to the prediction deviation of the quality indicator and the local sensitivity relationship between the process parameter and the quality indicator, which can be achieved by methods such as the inverse operation of the sensitivity matrix, gradient descent method, or optimization solution, and its purpose is to determine how to adjust the process parameter for the quality deviation; Among them, weighted average means that the parameter adjustment amounts calculated for different quality indicators are given different weights according to the importance or deviation degree of each quality indicator, and then linearly combined or non-linearly combined to obtain a comprehensive parameter adjustment amount. It can be achieved by methods such as based on preset weights, dynamically adjusting weights based on deviation magnitudes, or allocating weights based on fuzzy rules. The purpose is to balance the adjustment requirements of different quality indicators and obtain an overall optimized or requirement-satisfied parameter adjustment plan.

[0065] In the solution of this application, during the operation of the extrusion equipment, planned perturbations are made to various process parameters, and the process parameter values before and after the perturbations and the corresponding real-time product quality data are synchronously collected, so as to actively obtain the actual influence information of the change of process parameters on product quality under the current working conditions. Based on the collected perturbation data, the local sensitivity coefficients of each process parameter to each quality indicator are calculated. These coefficients reflect the influence degree of the small change of the process parameter on the product quality indicator under the current specific production conditions. At the same time, based on the historical benchmark model, the theoretical values of each quality indicator are predicted according to the current process parameters. The real-time product quality data is compared with the predicted values, and the prediction deviation of each quality indicator is calculated. Using the calculated local sensitivity coefficients and combining with the prediction deviations of each quality indicator, the adjustment amounts that each process parameter theoretically needs to make to reduce these deviations are inversely deduced. Since there may be multiple quality indicators that need to be adjusted and different adjustment directions may conflict, a weighted average method is adopted to comprehensively consider the adjustment requirements of each quality indicator and calculate a final parameter adjustment amount for compensating the deviation of the current working condition. This adjustment amount is then combined with the initial process parameters to form a new process parameter adjustment value and applied to the extrusion equipment. This compensation calculation mechanism based on real-time perturbation and local sensitivity analysis can accurately perceive the influence of the current working condition on the relationship between process parameters and quality, and calculate the parameter adjustment amount that meets the actual requirements, thus overcoming the limitations of only relying on historical models or current parameters for adjustment. Applying this precise compensation calculation method to the iterative adjustment framework of the cable sheath die change control method enables the system to quickly converge to the target quality under the influence of hidden factors such as equipment wear and material batch differences, improves the accuracy and efficiency of parameter adjustment, and reduces the number of trial-and-error times and waste product generation.

[0066] Through the above technical solution, by disturbing the process parameters and collecting data, the present application can obtain real-time information on the relationship between process parameters and quality indicators under the current working conditions, calculate the local sensitivity coefficient, and accurately quantify this relationship. Combining the deviation between the predicted value of the historical model and the real-time quality data, and using the local sensitivity coefficient to inversely deduce the parameter adjustment amount, it is possible to accurately determine the direction and amplitude of adjustment. Considering the adjustment requirements of different quality indicators comprehensively, methods such as weighted average are used to obtain the final compensation amount, making the adjustment plan effective. This adjustment method based on real-time working condition information and local sensitivity analysis can effectively cope with the working condition changes caused by hidden factors such as equipment wear and material batch differences that are difficult to directly perceive, improve the calculation accuracy of the parameter compensation adjustment amount, and thus can adjust the product quality to the target requirements faster and more accurately, reducing the number of trial-and-error times and waste product generation.

[0067] Further, in step A2, each real-time process parameter is disturbed respectively to obtain the disturbed process parameters, and the real-time sheath product quality data corresponding to the disturbed process parameters is collected, including: C1. Taking the extrusion temperature, screw speed, and traction speed as the process parameters to be disturbed respectively, and generating corresponding pseudo-random disturbance sequences; wherein, the disturbance amplitude of the pseudo-random disturbance sequence is determined based on the fluctuation range of each process parameter in the historical production data, and the disturbance frequency is determined according to the dynamic response characteristics of the extrusion process; C2. According to the generated pseudo-random disturbance sequences, each process parameter to be disturbed is disturbed in sequence to obtain the disturbed process parameters, and the corresponding real-time sheath product quality data is collected in real time; Step A301 includes: For each process parameter to be disturbed, according to the numerical change before and after its disturbance and the change amount of each quality indicator caused, the local sensitivity coefficient of this process parameter to each quality indicator is fitted by the least squares method.

[0068] Among them, the pseudo-random disturbance sequence refers to a sequence with similar statistical characteristics of random signals but generated by a deterministic algorithm, which can be generated by methods such as linear feedback shift register (LFSR) or M sequence. Its purpose is to provide a wide-band and repeatable disturbance signal for system identification. Among them, the disturbance amplitude refers to the maximum offset of the disturbance signal relative to the reference value of the process parameter, which is determined based on the fluctuation range of each process parameter in the historical production data. Its purpose is to ensure that the disturbance is within the range allowed by normal production and avoid adverse effects on the production process.

[0069] Among them, the disturbance frequency refers to the frequency components and their distributions contained in the disturbance signal, which is determined according to the dynamic response characteristics of the extrusion process. The purpose is to enable the frequency components of the disturbance signal to fully cover the main dynamic response frequency range of the extrusion process, thereby effectively exciting the dynamic behavior of the system and obtaining effective data for identification.

[0070] Among them, the dynamic response characteristics refer to the response behavior of the system to changes in the input signal, including at least one of response speed, delay, time constant, etc. It can be obtained by performing step response, impulse response, or frequency response tests on the extrusion process. The purpose is to guide the design of the disturbance frequency and ensure that the disturbance can fully reflect the dynamic characteristics of the system.

[0071] Among them, the least squares method refers to a mathematical optimization technique that estimates model parameters by minimizing the sum of the squares of the errors between the observed data and the model prediction values. It can be used to fit linear or nonlinear models. The purpose is to accurately estimate the local linear relationship between the process parameters and the quality indicators, that is, the local sensitivity coefficient, from the disturbance data containing measurement noise.

[0072] The solution of this application provides a more accurate and effective method for calculating process parameter disturbance and local sensitivity coefficient by introducing a pseudo-random disturbance sequence and the least squares method. Specifically, in step C1, the extrusion temperature, screw speed, and traction speed are respectively used as the process parameters to be disturbed, and a pseudo-random disturbance sequence is generated for each parameter. The disturbance amplitude of the pseudo-random disturbance sequence is determined based on the fluctuation range of each process parameter in the historical production data, which means that the disturbance will not exceed the normal production range and avoids excessive interference with the production process. The disturbance frequency is determined according to the dynamic response characteristics of the extrusion process, ensuring that the disturbance can effectively excite the system response and thus obtain useful data. In this way, it can be ensured that the disturbance can provide sufficient information for parameter identification without causing significant negative impacts on the production process. In step C2, according to the generated pseudo-random disturbance sequence, each process parameter to be disturbed is disturbed in turn, and the corresponding real-time sheath product quality data is collected in real time. This way of disturbing one by one can avoid the coupling effect between parameters and improve the interpretability of the data. In step A301, for each process parameter to be disturbed, according to the numerical change before and after its disturbance and the change amount of each quality indicator caused, the least squares method is used to fit the local sensitivity coefficient of this process parameter to each quality indicator. The least squares method is a commonly used data fitting method that can effectively extract the relationship between parameters from the noisy data.

[0073] In this way, the influence degree of process parameters on quality indicators can be estimated more accurately, providing a more reliable basis for subsequent parameter adjustment. This more precise calculation method of sensitivity coefficient enables the steps of reverse derivation of parameter adjustment amount and comprehensive calculation of parameter compensation adjustment amount based on the sensitivity coefficient to obtain more accurate inputs, thereby improving the accuracy and effect of overall parameter compensation adjustment.

[0074] In some preferred embodiments, the specific implementation is as follows. For the extrusion temperature, screw speed, and traction speed, corresponding pseudo-random perturbation sequences are generated respectively. The pseudo-random perturbation sequence can use the linear feedback shift register (LFSR) algorithm to generate a binary sequence, and then obtain a perturbation sequence suitable for process parameters through appropriate scaling and offset transformations. The determination of the perturbation amplitude can be based on the statistical analysis of historical production data. For example, if the historical data shows that the normal fluctuation range of the extrusion temperature is ±5°C, the perturbation amplitude of the extrusion temperature can be set within the range of ±2°C; similarly, the perturbation amplitudes of the screw speed and traction speed are determined according to historical data. The determination of the perturbation frequency can first analyze its dynamic response characteristics by performing a step response test on the extrusion process. For example, if the response time constants for the outer diameter and wall thickness are observed to be 10 seconds and 15 seconds respectively, a suitable perturbation frequency can be selected, such as 0.1 Hz (period 10 seconds), to fully stimulate these dynamic responses. During the actual perturbation process, according to the generated pseudo-random perturbation sequence, first perturb the extrusion temperature for a period of time while collecting the corresponding outer diameter, wall thickness, and roundness data in real time; then restore the extrusion temperature to the reference value, perturb the screw speed and collect data; finally, restore the screw speed and perturb the traction speed and collect data. After collecting the instantaneous values of each process parameter and the instantaneous values of the corresponding quality indicators during the perturbation period, for each process parameter to be perturbed and each quality indicator, the least squares method is used for fitting. For example, to calculate the local sensitivity coefficient of the extrusion temperature to the outer diameter, a simple linear model can be constructed: Δ outer diameter = S_temperature * Δ extrusion temperature + error, where Δ outer diameter is the change in the outer diameter, S_temperature is the local sensitivity coefficient of the extrusion temperature to the outer diameter, and Δ extrusion temperature is the change in the extrusion temperature. Using the collected multiple groups of (Δ extrusion temperature, Δ outer diameter) data pairs, the coefficient S_temperature is solved by the least squares method. Similarly, the local sensitivity coefficients of other process parameters to other quality indicators are calculated.

[0075] Through the above technical solution, the design of the pseudo-random perturbation sequence takes into account the actual production and system characteristics, can effectively stimulate the system dynamic response, obtain data rich in information, and at the same time ensure that the perturbation is within a safe range without affecting normal production; sequential perturbation avoids parameter coupling and improves data purity; the least squares method can accurately fit the local sensitivity coefficient from the noisy data, improving the calculation accuracy and reliability of the local sensitivity coefficient, providing a more accurate basis for subsequent parameter compensation adjustment, thereby improving the efficiency of parameter adjustment after die change and the stability of product quality.

[0076] Preferably, step A304 may include: D1. For each quality index, according to the local sensitivity coefficient and in combination with the preset parameter adjustment constraint conditions, construct an optimization model with the goal of minimizing the prediction deviation of this quality index; wherein, the parameter adjustment constraint conditions include: the upper and lower limits of the adjustment range of each process parameter, the upper and lower limits of the adjustment rate, and the stability constraint during the adjustment process, and the stability constraint is achieved by restricting the second derivative of the process parameter adjustment (that is, the stability constraint includes the upper and lower limits of the second derivative of the process parameter adjustment); D2. For each quality index, use the sequential quadratic programming algorithm to solve the optimization model to obtain the optimal parameter adjustment amount that satisfies the parameter adjustment constraint conditions, forming a set of optimal parameter adjustment amounts; D3. According to the preset quality index priority and in combination with the fuzzy logic reasoning method, screen and fuse the set of optimal parameter adjustment amounts to obtain the final parameter compensation adjustment amount.

[0077] Among them, in step D1, for each key quality index in the production process of the cable sheath, such as outer diameter, wall thickness, roundness, etc., use the calculated local sensitivity coefficient and in combination with the pre-set process parameter adjustment limit conditions to construct a mathematical optimization model. The objective function of this optimization model is to minimize the prediction deviation of the current quality index, that is, the difference between the actual measured value and the predicted value of the historical benchmark model.

[0078] Among them, the parameter adjustment constraint conditions refer to the limit conditions that must be satisfied when adjusting process parameters, which may include the maximum and minimum values allowed for each process parameter to be adjusted, as well as the maximum change amount allowed for each process parameter to be adjusted per unit time; the stability constraint refers to the limit imposed to avoid drastic fluctuations in the process state during parameter adjustment, which can be achieved by restricting the change rate of the parameter change rate. These constraint conditions can be preset according to specific equipment characteristics, material properties, and production experience.

[0079] Among them, the sequential quadratic programming algorithm is an iterative optimization algorithm used to solve nonlinear programming problems, especially suitable for dealing with optimization problems with constraints. The specific calculation process of the sequential quadratic programming algorithm is prior art and will not be elaborated here. The optimal parameter adjustment amount refers to the process parameter adjustment amount that can minimize the prediction deviation of a specific quality index under the condition of satisfying all parameter adjustment constraint conditions. The set of optimal parameter adjustment amounts refers to a set of optimal parameter adjustment amounts obtained by separately solving the optimization model for each quality index.

[0080] Among them, the quality index priority refers to when there are multiple quality indices, different importance levels are assigned to different quality indices according to product requirements or production needs, which can be determined by expert experience setting, historical data analysis, user definition, etc. The fuzzy logic reasoning method is a reasoning method based on fuzzy set theory and fuzzy rules, which can handle uncertain and fuzzy information for decision-making. Screening and fusion refer to selecting or synthesizing a final parameter adjustment amount from the set of optimal parameter adjustment amounts according to the quality index priority and the fuzzy logic reasoning method. The final parameter compensation adjustment amount refers to the process parameter adjustment amount obtained after screening and fusion, which is used to compensate for the deviation of the current working condition.

[0081] The solution of this application constructs an optimization model with the goal of minimizing the prediction deviation of each quality index by combining the local sensitivity coefficient and the preset parameter adjustment constraint conditions for each quality index, thereby transforming the parameter adjustment problem into a constrained optimization problem. The parameter adjustment constraint conditions include the upper and lower limits of the adjustment range of each process parameter, the upper and lower limits of the adjustment rate, and the stability constraint during the adjustment process. The stability constraint is achieved by restricting the second derivative of the process parameter adjustment, which ensures that the obtained parameter adjustment amount is within the feasible range of the actual process and the adjustment process is stable, avoiding affecting product quality due to excessive or unstable adjustment. Based on this, for each quality index, the sequential quadratic programming algorithm is used to solve the optimization model to obtain the optimal parameter adjustment amount that satisfies the parameter adjustment constraint conditions, forming a set of optimal parameter adjustment amounts. The sequential quadratic programming algorithm can effectively find the optimal solution of the constrained nonlinear optimization problem, ensuring the reliability of the solution result. Since different quality indices may interact or even conflict with each other, directly applying a certain solution in the set of optimal parameter adjustment amounts may cause other indices to deteriorate. Therefore, according to the preset quality index priority and combined with the fuzzy logic reasoning method, the set of optimal parameter adjustment amounts is screened and fused. The fuzzy logic reasoning method can handle the trade-off problem between multiple objectives, comprehensively evaluate the advantages and disadvantages of different adjustment schemes according to the deviation degree and priority of each quality index, and finally obtain a parameter compensation adjustment amount that comprehensively considers all quality indices.

[0082] This method utilizes the current working condition information provided by the local sensitivity coefficient, combines the actual process constraints and the multi-objective decision-making mechanism, overcomes the problems of infeasible parameters, unstable adjustment, and multi-objective conflicts that may be caused by simply performing reverse derivation based on the sensitivity coefficient alone, improves the accuracy, stability, and robustness of parameter adjustment, thereby more effectively reducing the prediction deviation and enabling the quality of the sheath product to quickly meet the target requirements.

[0083] Further, step D3 may include: D301. Divide each quality index into a core index and a secondary index according to the preset quality index priority; D302. For each set of optimal parameter adjustment amounts in the optimal parameter adjustment amount set, calculate its influence degree on the core index, and screen out the optimal parameter adjustment amounts that can make the core index reach the preset quality threshold range to form a core parameter adjustment quantum set; D303. For each set of parameter adjustment amounts in the core parameter adjustment quantum set, adopt the fuzzy logic reasoning method, calculate the membership degree according to the deviation degree of the secondary index and the adjustment amplitude of the process parameters, and perform fuzzy reasoning according to the membership degree to obtain a comprehensive evaluation value; D304. Select the parameter adjustment amount with the highest comprehensive evaluation value as the final parameter compensation adjustment amount.

[0084] Among them, the core index and the secondary index refer to dividing all quality indexes into core indexes that must strictly meet the requirements and secondary indexes with relatively lower importance according to the quality index priority, and they can be divided by setting a priority threshold or directly specifying.

[0085] Among them, in step D302, it can be realized by applying each set of optimal parameter adjustment amounts to the current process parameters, then using the historical benchmark model to predict the adjusted quality index, and comparing it with the preset quality threshold range. The preset quality threshold range refers to the acceptable product quality range set for the core index, and it can be determined by means such as process specifications, product standards, or user requirements.

[0086] Among them, step D303 can be realized by constructing a fuzzy inference system. This system takes the current deviation of the secondary index (such as the prediction deviation of roundness) and the change amplitude of the process parameters caused by this set of parameter adjustment amounts (such as the adjustment amounts of extrusion temperature, screw speed, and traction speed) as inputs, and outputs a quantified comprehensive evaluation value through the processes of fuzzification, fuzzy rule reasoning, and defuzzification. For example, fuzzy sets such as "small deviation of secondary index", "large deviation of secondary index", "small adjustment amplitude", and "large adjustment amplitude" can be defined, and fuzzy rules such as "if the deviation of the secondary index is small and the adjustment amplitude is small, then the comprehensive evaluation value is high" can be established.

[0087] The solution of this application divides each quality index into a core index and a secondary index according to the preset quality index priority, so as to distinguish the importance of different quality indexes and provide a basis for screening the parameter adjustment amount in the subsequent steps. Then, for each set of optimal parameter adjustment amounts in the optimal parameter adjustment amount set, calculate its influence degree on the core index, and screen out the optimal parameter adjustment amounts that can make the core index reach the preset quality threshold range to form a core parameter adjustment quantum set. Through this step, it can be ensured that the adjusted process parameters can meet the requirements of the core quality index and avoid the situation of out-of-tolerance core index. Next, for each set of parameter adjustment amounts in the core parameter adjustment quantum set, use the fuzzy logic inference method to calculate the membership degree according to the deviation degree of the secondary index and the adjustment amplitude of the process parameters, and perform fuzzy inference according to the membership degree to obtain a comprehensive evaluation value. The fuzzy logic inference method can comprehensively consider the influence of multiple factors, avoid the excessive influence of a single factor on the selection of the parameter adjustment amount, and improve the rationality of the parameter adjustment. At the same time, considering the deviation degree of the secondary index and the adjustment amplitude of the process parameters can, on the premise of ensuring that the core index meets the requirements, give full consideration to the secondary index as much as possible and avoid too large an adjustment amplitude of the process parameters to ensure the stability of the production process. Finally, select the parameter adjustment amount with the highest comprehensive evaluation value as the final parameter compensation adjustment amount. By selecting the parameter adjustment amount with the highest comprehensive evaluation value, it is possible to give full consideration to the secondary index as much as possible on the premise of meeting the requirements of the core index, and ensure the rationality and stability of the process parameter adjustment, so as to obtain the optimal parameter compensation adjustment amount.

[0088] Based on the obtained optimal parameter adjustment amount set, this solution further introduces the division of quality index priority and fuzzy logic inference, and conducts targeted screening and fusion on this set, so that the finally selected adjustment amount can not only reduce the prediction deviation, but also meet the different quality index requirements in actual production, especially giving priority to ensuring the core index, while taking into account the secondary index and adjustment stability, making the entire parameter adjustment process closer to the actual production needs and improving the effectiveness and robustness of the adjustment.

[0089] Reference Figure 2 , this application provides a cable sheath die change control system for adjusting and controlling the process parameters of an extrusion device after the die of the cable sheath extrusion device is changed. The system includes: An initial parameter recommendation module 1 for determining the initial process parameters according to the target sheath specification by using the historical benchmark model to start the extrusion device (the specific process refers to step A1 in the previous text); A real-time data acquisition module 2 for collecting real-time process parameters and corresponding real-time sheath product quality data during the operation of the extrusion device (the specific process refers to step A2 in the previous text); The working condition compensation calculation module 3 is used to calculate the parameter compensation adjustment amount between the current working condition and the historical reference working condition according to the real-time process parameters and the corresponding real-time sheath product quality data (for the specific process, refer to step A3 in the previous text); The parameter comprehensive adjustment module 4 is used to combine the initial process parameters with the parameter compensation adjustment amount to obtain the process parameter adjustment value for the current working condition, and send it to the control system of the extrusion equipment, so that the extrusion equipment operates according to the process parameter adjustment value (for the specific process, refer to step A4 in the previous text); The iterative control module 5 is used to trigger the real-time data acquisition module, the working condition compensation calculation module and the parameter comprehensive adjustment module to run repeatedly until the quality of the sheath product meets the target requirements (for the specific process, refer to step A5 in the previous text).

[0090] The above are only the embodiments of the present application and are not used to limit the protection scope of the present application. For those skilled in the art, various changes and modifications can be made to the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for controlling die change of a cable sheath, which is used to adjust and control the process parameters of an extrusion device after die change of the cable sheath extrusion device, is characterized in that The steps of this method include: A1. According to the target sheath specifications, determine the initial process parameters using the historical benchmark model to start the extrusion equipment; A2. During the operation of the extrusion equipment, collect the real-time process parameters and the corresponding real-time sheath product quality data; A3. Calculate the parameter compensation adjustment amount between the current working condition and the historical benchmark working condition according to the real-time process parameters and the corresponding real-time sheath product quality data; A4. Combine the initial process parameters with the parameter compensation adjustment amount to obtain the process parameter adjustment value for the current working condition, and send it to the control system of the extrusion equipment so that the extrusion equipment operates according to the process parameter adjustment value; A5. Repeat steps A2 - A4 until the sheath product quality meets the target requirements.

2. The cable sheath die change control method according to claim 1, wherein, The historical benchmark model includes multiple parameter optimization sub-models, and each parameter optimization sub-model corresponds to a quality index; the parameter optimization sub-model takes the process parameters as the input and the corresponding quality index as the output; the quality indexes include outer diameter, wall thickness, and roundness; The process parameters include extrusion temperature, screw speed, and traction speed; Step A1 includes: A101. Extract multiple quality indexes from the target sheath specifications, including outer diameter, wall thickness, and roundness; A102. For each quality index, call the corresponding parameter optimization sub-model respectively; A103. Based on the preset weight assignment strategy, assign weights to each parameter optimization sub-model; this weight reflects the importance of the corresponding quality index in the overall quality assessment; A104. Use a multi-objective optimization algorithm to solve a set of initial process parameters with the goal of minimizing the deviation between the weighted outputs of all parameter optimization sub-models and the corresponding quality indexes in the target sheath specifications, and use this set of initial process parameters as the start parameters of the extrusion equipment.

3. A method for controlling die change of a cable sheath according to claim 2, characterized in that, Step A103 includes: For the target sheath specifications, query the preset mapping relationship table of product specifications and weight assignment strategies to determine the weight assignment strategy that matches the current target sheath specifications; If a matching weight assignment strategy is found, directly assign weights to each parameter optimization sub-model according to the found strategy; if no matching weight assignment strategy is found, assign weights to each parameter optimization sub-model according to the preset default weight assignment strategy, or perform weight assignment according to the operator's custom settings; For the aging degree of the extrusion equipment, according to the preset mapping relationship table of equipment aging degree and weight adjustment coefficient, determine the weight adjustment coefficient that matches the current aging degree of the extrusion equipment, and adjust the already assigned weights based on this weight adjustment coefficient; where the weight adjustment coefficient is used to adjust the relative importance between different quality indexes to adapt to the influence brought by equipment aging.

4. The cable sheath die change control method according to claim 2, characterized in that Step A104 includes: Determine the constraint conditions of the multi-objective optimization algorithm; the constraint conditions include the upper and lower limits of the extrusion temperature, the upper and lower limits of the screw speed, and the upper and lower limits of the traction speed; Initialize the parameters of the multi-objective optimization algorithm; the initialized parameters include population size, crossover probability, and mutation probability; The genetic algorithm is used as the multi-objective optimization algorithm, and the objective function is to minimize the deviation between the weighted output of all parameter optimization sub-models and the corresponding quality indicators in the target sheath specification; Under the above constraints, based on the objective function, through iterative optimization of the genetic algorithm, a Pareto optimal solution set is obtained; each solution in the Pareto optimal solution set corresponds to a set of initial process parameters; From the Pareto optimal solution set, according to the preset evaluation index, a set of initial process parameters is selected as the starting parameters of the extrusion equipment; the evaluation index comprehensively considers the deviation degree of each quality index from the target sheath specification and the stability of the process parameters.

5. A method for controlling die change of a cable sheath according to claim 4, characterized in that, The step of selecting a set of initial process parameters from the Pareto optimal solution set as the starting parameters of the extrusion equipment according to the preset evaluation index includes: B1. For each solution in the Pareto optimal solution set, calculate the deviation degree of the corresponding quality index from the target sheath specification. The deviation degree includes absolute deviation and relative deviation, and the absolute deviation and relative deviation are normalized to obtain the normalized deviation degree; B2. For each solution in the Pareto optimal solution set, evaluate the stability of the corresponding process parameters; the stability of the process parameters is determined by calculating the fluctuation range of each process parameter in the historical production data. The smaller the fluctuation range, the higher the stability; B3. According to the aging degree of the extrusion equipment, perform aging correction on the stability to obtain the parameter stability after aging correction; B4. According to the target sheath specification and the aging degree of the extrusion equipment, determine the weight factors of the normalized deviation degree and the parameter stability after aging correction; B5. Using the method of weighted summation, according to the determined weight factors, comprehensively evaluate the normalized deviation degree and the parameter stability after aging correction to obtain the comprehensive score value of each solution; B6. Select the solution with the highest comprehensive score value as the starting parameters of the extrusion equipment.

6. A cable sheath die change control method according to claim 1, characterized in that, Step A2 includes: respectively perturbing each process parameter to obtain the perturbed process parameters, and collecting the real-time sheath product quality data corresponding to the perturbed process parameters; the process parameters include extrusion temperature, screw speed and traction speed, and the real-time sheath product quality data includes three quality indicators of outer diameter, wall thickness and roundness; Step A3 includes: A301. Based on the process parameters before and after perturbation and the corresponding real-time sheath product quality data, calculate the local sensitivity coefficient of each process parameter to each quality indicator; the local sensitivity coefficient characterizes the influence degree of the change of the process parameter on the quality indicator under the current working condition; A302. Based on the historical benchmark model, calculate the predicted values of each quality indicator under the current process parameters; A303. For each quality indicator, calculate the deviation between the indicator value in the current real-time sheath product quality data and the corresponding predicted value, which is recorded as the prediction deviation; A304. According to the local sensitivity coefficient, inversely deduce the amount of adjustment required for each process parameter, which is recorded as the parameter adjustment amount, to reduce the prediction deviation; A305. Combine the parameter adjustment amounts corresponding to the prediction deviations of each quality index, and use the weighted average method to calculate the final parameter compensation adjustment amount.

7. A cable sheath die change control method according to claim 6, characterized in that, In step A2, each real-time process parameter is perturbed separately to obtain the perturbed process parameters, and the real-time sheath product quality data corresponding to the perturbed process parameters is collected, including: C1. Respectively use the extrusion temperature, screw speed, and traction speed as the process parameters to be perturbed, and generate corresponding pseudo-random perturbation sequences; among them, the perturbation amplitude of the pseudo-random perturbation sequence is determined based on the fluctuation range of each process parameter in the historical production data, and the perturbation frequency is determined according to the dynamic response characteristics of the extrusion process; C2. According to the generated pseudo-random perturbation sequence, sequentially perturb each process parameter to be perturbed to obtain the perturbed process parameters, and collect the corresponding real-time sheath product quality data in real time; Step A301 includes: For each process parameter to be perturbed, according to the numerical change before and after its perturbation and the change amount of each quality index caused, use the least squares method to fit the local sensitivity coefficient of this process parameter to each quality index.

8. A cable sheath die change control method according to claim 6, characterized in that, Step A304 includes: D1. For each quality index, according to the local sensitivity coefficient and combined with the preset parameter adjustment constraint conditions, construct an optimization model with the goal of minimizing the prediction deviation of this quality index; among them, the parameter adjustment constraint conditions include: the upper and lower limits of the adjustment range of each process parameter, the upper and lower limits of the adjustment rate, and the stability constraint during the adjustment process, and the stability constraint is realized by restricting the second derivative of the process parameter adjustment; D2. For each quality index, use the sequential quadratic programming algorithm to solve the optimization model to obtain the optimal parameter adjustment amount that meets the parameter adjustment constraint conditions, and form a set of optimal parameter adjustment amounts; D3. According to the preset quality index priority and combined with the fuzzy logic reasoning method, screen and fuse the set of optimal parameter adjustment amounts to obtain the final parameter compensation adjustment amount.

9. A method for controlling die change of a cable sheath, according to claim 8, characterized in that, Step D3 includes: D301. According to the preset quality index priority, divide each quality index into core indicators and secondary indicators; D302. For each set of optimal parameter adjustment amounts in the set of optimal parameter adjustment amounts, calculate its influence degree on the core indicators, and screen out the optimal parameter adjustment amounts that can make the core indicators reach the preset quality threshold range to form a subset of core parameter adjustment amounts; D303. For each set of parameter adjustment amounts in the subset of core parameter adjustment amounts, use the fuzzy logic reasoning method to calculate the membership degree according to the deviation degree of the secondary indicators and the adjustment amplitude of the process parameters, and perform fuzzy reasoning according to the membership degree to obtain a comprehensive evaluation value; D304. Select the parameter adjustment amount with the highest comprehensive evaluation value as the final parameter compensation adjustment amount.

10. A cable sheath die change control system is used to adjust and control the process parameters of an extrusion device after the die of the cable sheath extrusion device is changed. It is characterized in that, The system includes: An initial parameter recommendation module, which is used to determine the initial process parameters according to the target sheath specification by using the historical benchmark model to start the extrusion equipment; A real-time data acquisition module, which collects real-time process parameters and corresponding real-time sheath product quality data during the operation of the extrusion equipment; The working condition compensation calculation module is used to calculate the parameter compensation adjustment amount between the current working condition and the historical reference working condition according to the real-time process parameters and the corresponding real-time sheath product quality data; The parameter comprehensive adjustment module is used to combine the initial process parameters with the parameter compensation adjustment amount to obtain the process parameter adjustment value for the current working condition, and send it to the control system of the extrusion equipment, so that the extrusion equipment operates according to the process parameter adjustment value; The iterative control module is used to trigger the real-time data acquisition module, the working condition compensation calculation module and the parameter comprehensive adjustment module to run repeatedly until the quality of the sheath product meets the target requirements.

Images