Steel pipe cold drawing process parameter optimization method and system

Abstract

The invention discloses a steel pipe cold-drawing process parameter optimization method and system, and relates to the related technical field of steel pipe cold-drawing The method comprises the following steps: receiving and analyzing a cold-drawing task, and establishing cold-drawing task demand characteristics; historical cold drawing process parameters are obtained, fitness evaluation under multiple targets is carried out, and non-dominated solution distribution is constructed; crowding degree calculation is carried out to generate crowding degree distribution; population optimization iteration is carried out, and an optimized population is constructed; and selecting a solution with the maximum fusion fitness in the optimized population as an optimal process parameter, and controlling cold-drawing equipment to perform cold-drawing process control on the steel pipe raw material. The technical problems that in the prior art, cold-drawing process parameters cannot be precisely matched with steel pipe raw materials and cold-drawing equipment, multi-target complex requirements are difficult to balance, and then the steel pipe size precision is not high, the product quality is not stable and raw materials are wasted in industrial control are solved, precise control over steel pipe cold drawing is achieved, and the production efficiency is improved. The technical effects of improving the product quality and the size precision and reducing the raw material loss are achieved.

Description

Technical Field

[0001] The present application relates to the technical field related to cold drawing of steel pipes, and specifically to a method and system for optimizing process parameters of cold drawing of steel pipes. Background Art

[0002] In the field of modern steel pipe manufacturing, the cold drawing process is a key link in improving the quality and performance of steel pipes. With the continuous advancement of industrial technology, various industries have put forward increasingly stringent requirements on the dimensional accuracy, surface quality and mechanical properties of steel pipes. The traditional cold drawing process parameters of steel pipes are often set based on experience, and it is difficult to achieve the optimal balance in multiple target dimensions, resulting in problems such as waste of raw materials, low production efficiency and unstable product quality in industrial production control. On the one hand, different steel pipe raw materials have different characteristics and respond very differently to the cold drawing process. If the process parameters are not accurately adapted to the raw materials, it is easy to cause internal defects in the steel pipe; on the other hand, the performance of cold drawing equipment will gradually change during long-term use in industrial production control, and the existing process parameters cannot be dynamically adjusted according to the characteristics of the equipment in real time. At the same time, the diversity of target steel pipe sizes also brings challenges to the cold drawing process.

[0003] Therefore, in the current relevant technologies, there are technical problems such as the inability of cold drawing process parameters to accurately adapt to the steel pipe raw materials and cold drawing equipment, and difficulty in balancing the complex needs of multiple objectives, which leads to low dimensional accuracy of steel pipes in industrial control, unstable product quality and waste of raw materials. Summary of the Invention

[0004] This application provides a method and system for optimizing the cold drawing process parameters of steel pipes, thereby solving the technical problems in the prior art that the cold drawing process parameters cannot be accurately adapted to the steel pipe raw materials and cold drawing equipment, and it is difficult to balance the complex requirements of multiple objectives, which in turn leads to low dimensional accuracy of steel pipes, unstable product quality and waste of raw materials in industrial control. It realizes precise control of cold drawing of steel pipes, and achieves the technical effect of improving product quality and dimensional accuracy and reducing raw material loss.

[0005] The present application provides a method for optimizing process parameters of cold drawing of steel pipes, the method comprising: in an interactive target cold drawing scenario, receiving and analyzing a cold drawing task, determining characteristics of the steel pipe raw material, cold drawing equipment, and target steel pipe size, and establishing a cold drawing task requirement characteristic; obtaining historical cold drawing process parameters based on the cold drawing task requirement characteristic, and performing fitness evaluation under multiple objectives, performing non-dominated sorting according to the fitness evaluation result, and constructing a non-dominated solution distribution; performing congestion calculation on each solution in the non-dominated solution distribution to generate a congestion distribution; screening seed solutions based on the non-dominated solution distribution and the congestion distribution, performing population optimization iterations until convergence conditions are met, and constructing an optimized population; selecting the solution with the largest fusion fitness in the optimized population as the optimal process parameter, and controlling the cold drawing equipment to perform cold drawing process control on the steel pipe raw material.

[0006] In a possible implementation, the steel tube cold drawing process parameter optimization method further performs the following processing: using the cold drawing task requirement characteristics as retrieval elements, performing historical data mining, and constructing an original cold drawing process parameter set, wherein any original cold drawing process parameter has a cold drawing quality inspection result mark; obtaining the steel tube cold drawing quality index, and constructing multiple objective functions with the steel tube cold drawing quality index; based on the cold drawing quality inspection result mark, calling the multiple objective functions to perform fitness evaluation on any original cold drawing process parameter in the original cold drawing process parameter set, and generating an original fitness information set; calling a non-dominated sorting mechanism, and performing non-dominated sorting on all individual solutions in the original cold drawing process parameter set based on the original fitness information set, and generating the non-dominated solution distribution, wherein the non-dominated sorting mechanism includes a first fitness comparison constraint and a second fitness comparison constraint, and the first fitness comparison constraint and the second fitness comparison constraint need to be satisfied at the same time.

[0007] In a possible implementation, the steel pipe cold drawing process parameter optimization method further performs the following processing: the first fitness comparison constraint is a first fitness optimization value based on the multiple objective functions, and the second fitness comparison constraint is a second fitness optimization value based on one objective function among the multiple objective functions.

[0008] In a possible implementation, the steel pipe cold drawing process parameter optimization method also performs the following processing: step a: randomly selecting the first original solution and the second original solution in the original cold drawing process parameter set, and matching multiple first fitnesses and multiple second fitnesses in the original fitness information set; step b: aligning the multiple first fitnesses with the multiple second fitnesses according to the objective function type, and performing fitness difference calculation of the same type to generate multiple first fitness differences; step c: based on the multiple first fitness differences, performing a dominance relationship analysis on the first original solution and the second original solution according to the first fitness comparison constraint and the second fitness comparison constraint to generate a first dominance relationship; step d: hierarchically stratifying the first original solution and the second original solution based on the first dominance relationship to generate a first stratification result; step e: and repeating steps a to d until any two original solutions in the original cold drawing process parameter set are traversed, and the non-dominated solution distribution is constructed with all generated stratification results.

[0009] In a possible implementation, the steel tube cold drawing process parameter optimization method further performs the following processing: if the multiple first fitness differences cannot simultaneously satisfy the first fitness comparison constraint and the second fitness comparison constraint, jump to step e.

[0010] In a possible implementation, the steel pipe cold drawing process parameter optimization method further performs the following processing: for each solution in the non-dominated solution distribution, calculating the neighborhood solution distribution density within a preset neighborhood range; normalizing and assigning the neighborhood solution distribution density, generating the congestion degree of each solution, and constructing the congestion degree distribution.

[0011] In a possible implementation, the steel tube cold drawing process parameter optimization method further performs the following processing: based on the non-dominated solution distribution, screening solutions with a distribution level higher than a preset excellent level to generate a first seed solution space; based on the congestion distribution, screening solutions with a congestion level less than a preset congestion level to generate a second seed solution space; merging the first seed solution space and the second seed solution space and deleting duplicate solutions to construct a first-generation population; based on the first-generation population, performing crossover and mutation processing on the solutions to generate a second-generation population; continuing to analyze the non-dominated solution distribution and congestion distribution of the second-generation population to screen the third-generation population, and so on, iterating multiple times until the convergence condition is met, and constructing the optimized population with the population obtained at convergence.

[0012] In a possible implementation, the steel pipe cold drawing process parameter optimization method further performs the following processing: the convergence condition is that the congestion degree of the solutions with a distribution level higher than a preset excellent level in the iteratively updated non-dominated solution distribution of the population is greater than or equal to the preset congestion degree.

[0013] In a possible implementation, the steel tube cold drawing process parameter optimization method further performs the following processing: extracting multiple fitnesses corresponding to each solution in the optimization population and performing weighted fusion to generate a fusion fitness corresponding to each solution; based on the fusion fitness corresponding to each solution, selecting the solution with the largest fusion fitness as the optimal process parameters.

[0014] The present application also provides a steel pipe cold drawing process parameter optimization system, which includes: a cold drawing task requirement feature building module, which is used to interact with the target cold drawing scenario, receive and analyze the cold drawing task, determine the steel pipe raw material, cold drawing equipment characteristics and target steel pipe size, and build the cold drawing task requirement feature; a non-dominated solution distribution building module, which is used to obtain historical cold drawing process parameters based on the cold drawing task requirement feature, and perform fitness evaluation under multiple objectives, perform non-dominated sorting according to the fitness evaluation results, and construct a non-dominated solution distribution; a congestion distribution generation module, which is used to calculate the congestion of each solution in the non-dominated solution distribution and generate a congestion distribution; an optimized population building module, which is used to combine the non-dominated solution distribution and the congestion distribution to screen seed solutions, perform population optimization iterations until the convergence conditions are met, and construct an optimized population; a cold drawing process control module, which is used to select the solution with the largest fusion fitness in the optimized population as the optimal process parameter, and control the cold drawing equipment to perform cold drawing process control on the steel pipe raw material.

[0015] The steel tube cold drawing process parameter optimization method and system proposed in this application is intended to receive and analyze cold drawing tasks, establish cold drawing task requirement characteristics; obtain historical cold drawing process parameters, and conduct fitness evaluation under multiple objectives to construct a non-dominated solution distribution; perform congestion calculation to generate a congestion distribution; perform population optimization iteration to construct an optimized population; select the solution with the largest fusion fitness in the optimized population as the optimal process parameter, and control the cold drawing equipment to perform cold drawing process control on the steel tube raw material. This solves the technical problems in the existing technology that the cold drawing process parameters cannot accurately adapt to the steel tube raw material and the cold drawing equipment, and it is difficult to balance the complex requirements of multiple objectives, which leads to low dimensional accuracy of steel tubes, unstable product quality, and waste of raw materials in industrial control. It realizes precise control of steel tube cold drawing, and achieves the technical effect of improving product quality and dimensional accuracy and reducing raw material loss. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] To more clearly illustrate the technical solutions of the embodiments of the present disclosure, the accompanying drawings of the embodiments of the present disclosure are briefly introduced below. Flowcharts are used in this application to illustrate the operations performed by the systems according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in precise order. Instead, various steps may be processed in reverse order or simultaneously as needed. Furthermore, other operations may be added to these processes, or one or more operations may be removed from these processes.

[0017] Figure 1 Schematic diagram of the process flow of the steel tube cold drawing process parameter optimization method provided in the embodiment of the present application.

[0018] Figure 2 Schematic diagram of the steel pipe cold drawing process parameter optimization system structure provided in an embodiment of the present application.

[0019] Explanation of the accompanying symbols: cold drawing task requirement feature building module 10, non-dominated solution distribution building module 20, congestion distribution generation module 30, optimized population building module 40, cold drawing process control module 50. DETAILED DESCRIPTION

[0020] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0021] In order to make the purpose, technical solutions and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limiting this application. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0022] In the following description, reference is made to “some embodiments”, which describes a subset of all possible embodiments, but it will be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict, and the terms “first\second” involved are merely used to distinguish similar objects and do not represent a specific ordering of the objects. The terms “including” and “having” and any variations are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or modules that are not clearly listed or that are inherent to these processes, methods, products or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application belongs. The terms used herein are for the purpose of describing the embodiments of this application only.

[0023] The present application provides a method for optimizing the cold drawing process parameters of a steel pipe. Figure 1 As shown, the method includes: Step S100, interactive target cold drawing scenario, receiving and analyzing the cold drawing task, determining the steel pipe raw material, cold drawing equipment characteristics and target steel pipe size, and establishing the cold drawing task requirement characteristics.

[0024] Preferably, there are different cold drawing production scenarios in actual production, such as different production workshop environments, different order requirements, etc. By interacting with the target cold drawing scenario, receiving and parsing the cold drawing task, obtaining relevant information in the cold drawing scenario, and understanding the specific situation and constraints of the current cold drawing production, specifically, performing a detailed analysis of the received cold drawing task to extract key information, such as clarifying the type of steel pipe to be cold drawn, production process requirements, etc. from the task description, and then determining the steel pipe raw material, cold drawing equipment characteristics and target steel pipe size, wherein the steel pipe raw material includes material (such as carbon steel, alloy steel, etc.), initial size (outer diameter, wall thickness, length), hardness, toughness, etc. These characteristics directly influence the selection of cold drawing process parameters. For example, steel pipes of different materials have different deformation capabilities during the cold drawing process, requiring different pulling forces and die sizes. The characteristics of cold drawing equipment may include the equipment's maximum pulling force, stretching speed range, die accuracy, and dimensional specifications. Different cold drawing equipment has differences in structure and performance, which will have a significant impact on the cold drawing process. The target steel pipe size is the final dimensional requirement that the steel pipe needs to achieve after the cold drawing process, including the tolerance range of outer diameter, wall thickness, and length. The optimization of process parameters is to ensure that the final product can meet these dimensional requirements. This demand feature information is combined to form a set of cold drawing task demand features.

[0025] Step S200 , obtaining historical cold drawing process parameters based on the cold drawing task requirement characteristics, performing fitness evaluation under multiple objectives, performing non-dominated sorting according to the fitness evaluation results, and constructing a non-dominated solution distribution.

[0026] Preferably, a search and match is performed in the historical steel pipe production data according to the requirements of the cold drawing task to obtain the historical cold drawing process parameters, including searching for cold drawing task records of the past production with the same raw material as the current steel pipe, the same or similar cold drawing equipment model, and the target steel pipe size. The cold drawing process parameters used at that time, such as drawing speed, die size, lubrication method, tensile force, etc., are obtained. The initial process parameters are provided by historical data to improve the efficiency and accuracy of the cold drawing process parameter determination; then a fitness evaluation is performed under multiple objectives. For example, the cold drawing process optimization goal is to improve the dimensional accuracy of the steel pipe, improve the surface quality, reduce the production cost, and improve the production efficiency. According to these multiple objectives, the pros and cons of the historical cold drawing process parameters under the current cold drawing task requirements are measured. Specifically, for each objective, a corresponding evaluation function is established to perform fitness evaluation, and then the fitness of each objective is combined to comprehensively evaluate the adaptability of each historical process parameter combination to the current cold drawing task.

[0027] Preferably, a non-dominated sorting is performed based on the fitness evaluation results, that is, the pros and cons of different individuals in the multi-objective optimization problem are compared, that is, each historical cold drawing process parameter combination is regarded as a whole. For example, for any two individuals (process parameter combinations) A and B, if the performance of individual A is better than or equal to that of individual B in all objective functions, and is strictly better than that of individual B in at least one objective function, then individual A is said to dominate individual B; if the process parameter combination A is not inferior to the combination B in ensuring the dimensional accuracy and surface quality of the steel pipe, and is significantly lower than the combination B in terms of production cost, then A dominates B; and then All historical process parameter combinations are compared pairwise and divided into different non-dominated layers. The individuals in the first layer are those that are not dominated by any other individuals and represent relatively optimal process parameter combinations under multi-objective optimization. After completing the non-dominated sorting, the process parameter combinations in different non-dominated layers are sorted to form a non-dominated solution distribution. This shows the distribution of different non-dominated solutions (i.e., better process parameter combinations) in the multi-objective optimization space, allowing an intuitive understanding of which process parameter combinations achieve a good balance between multiple objectives and the performance characteristics of these combinations for different objectives. For example, through the non-dominated solution distribution, it may be found that some process parameter combinations perform well in improving dimensional accuracy and surface quality, but have relatively low production efficiency; while other combinations have advantages in production efficiency and cost control, but slightly lack dimensional accuracy.

[0028] Furthermore, step S200 also includes step S210, using the cold drawing task requirement characteristics as retrieval elements, performing historical data mining, and constructing an original cold drawing process parameter set, wherein any original cold drawing process parameter has a cold drawing quality inspection result mark; step S220, obtaining the steel pipe cold drawing quality index, and constructing multiple objective functions with the steel pipe cold drawing quality index; step S230, based on the cold drawing quality inspection result mark, calling the multiple objective functions to perform fitness evaluation on any original cold drawing process parameter in the original cold drawing process parameter set, and generating an original fitness information set; step S240, calling the non-dominated sorting mechanism, performing non-dominated sorting on all individual solutions in the original cold drawing process parameter set based on the original fitness information set, and generating the non-dominated solution distribution, wherein the non-dominated sorting mechanism includes a first fitness comparison constraint and a second fitness comparison constraint, and the first fitness comparison constraint and the second fitness comparison constraint need to be satisfied at the same time.

[0029] Step S240 further includes: the first fitness comparison constraint is a first fitness optimization value based on the multiple objective functions; the second fitness comparison constraint is a second fitness optimization value based on one objective function among the multiple objective functions.

[0030] Preferably, a search is performed in a historical production data repository based on the cold drawing task requirement characteristics to identify historical cold drawing task records that are similar or matching to the current cold drawing task requirement characteristics, extract the cold drawing process parameters used in these historical tasks, and establish an original cold drawing process parameter set. Each original cold drawing process parameter has a cold drawing quality inspection result mark, which records the quality inspection status of the steel pipe obtained after cold drawing production using the process parameter, such as whether it is qualified, specific quality index values, etc.; the cold drawing quality index of the steel pipe is obtained, which may include dimensional accuracy (deviation between actual size and target size), surface roughness, mechanical properties (such as strength, toughness, etc.), etc. For each quality index, a corresponding objective function is constructed, that is, the numerator is the actual quality of the steel pipe (such as specific quality-related data such as the dimensional accuracy value and surface roughness value obtained by actual measurement), and the denominator is the qualified quality (the pre-set qualified standard value of the quality index) plus a smaller real number to prevent the denominator from being zero. The actual quality of the steel pipe is compared with the qualified quality through the objective function, and the influence of the process parameters on the quality index is evaluated.

[0031] Preferably, the cold drawing quality inspection result mark of each original cold drawing process parameter is used to obtain the actual quality data of the steel pipe corresponding to the process parameter, and then the actual quality data is substituted into the objective function corresponding to each quality index for calculation to obtain the fitness value of the process parameter under the current quality index. Since there are multiple quality indicators, there are also multiple objective functions, so each process parameter will obtain multiple fitness values. Then, this operation is performed on all process parameters in the original cold drawing process parameter set, and all the obtained fitness values are integrated to form an original fitness information set, wherein each group of process parameters corresponds to multiple fitness values, which respectively reflect the performance of the process parameters under different quality indicators.

[0032] Preferably, the non-dominated sorting mechanism is used to compare the pros and cons of different process parameter combinations (each process parameter combination is regarded as an individual solution) in the multi-objective optimization problem, including the first fitness comparison constraint and the second fitness comparison constraint, and these two constraints need to be satisfied at the same time, and then all individual solutions in the original cold drawing process parameter set are non-dominated sorted based on the original fitness information set. Specifically, for any two individual solutions (process parameter combinations) A and B in the original cold drawing process parameter set, the first fitness comparison constraint is the first fitness optimization value based on multiple objective functions, and the performance of individual A on all objective functions must be better than or equal to Individual B, i.e., A, must have a fitness value calculated for each objective function corresponding to each quality indicator that is no less than B's fitness value under the corresponding objective function. The second fitness comparison constraint is based on the second fitness optimization value of one of the multiple objective functions. Individual A must strictly outperform individual B on at least one objective function. In other words, A's fitness value under the objective function corresponding to a certain quality indicator must be greater than B's fitness value under that objective function. When A satisfies both constraints, it is said that individual A dominates individual B. The optimization value of A relative to B is the fitness difference between A and B. As long as it is greater than or equal to 0, individual A dominates individual B. By comparing all individual solutions pairwise and dividing them into non-dominated layers based on dominance relationships, the first layer represents solutions that are not dominated by any other individual solution and represent relatively optimal process parameter combinations under multi-objective optimization. Solutions in the second layer are dominated by some solutions in the first layer but not by solutions in other layers. This process continues in this order, ultimately forming a distribution of non-dominated solutions, demonstrating the distribution of different non-dominated solutions (i.e., optimal process parameter combinations) in the multi-objective optimization space.

[0033] Furthermore, step S240 also includes step a: randomly selecting the first original solution and the second original solution in the original cold drawing process parameter set, and matching multiple first fitnesses and multiple second fitnesses in the original fitness information set; step b: performing fitness difference calculation of the same type on the multiple first fitnesses and the multiple second fitnesses according to the objective function type to generate multiple first fitness differences; step c: based on the multiple first fitness differences, performing a dominance relationship analysis on the first original solution and the second original solution according to the first fitness comparison constraint and the second fitness comparison constraint to generate a first dominance relationship; step d: hierarchically stratifying the first original solution and the second original solution based on the first dominance relationship to generate a first stratification result; step e: and repeating steps a to d until any two original solutions in the original cold drawing process parameter set are traversed, and the non-dominated solution distribution is constructed with all generated stratification results.

[0034] Preferably, two sets of process parameter combinations are randomly selected from the original cold drawing process parameter set and named as the first original solution and the second original solution, respectively, representing different cold drawing process parameter setting schemes. Since each original solution corresponds to multiple fitness values in the original fitness information set (each fitness value corresponds to the calculation result of an objective function, reflecting the performance of the original solution under a certain quality indicator), multiple first fitness values and multiple second fitness values that match the two original solutions are found in the original fitness information set. For example, if there are three objective functions (such as objective functions corresponding to dimensional accuracy, surface quality, and production efficiency), then the first original solution has three corresponding first fitness values, and the second original solution also has three corresponding second fitness values. The multiple fitness values of the first original solution and the multiple fitness values of the second original solution are matched (aligned) one-to-one according to the type of the objective function. For example, the fitness value of the first original solution corresponding to the dimensional accuracy objective function is matched with the fitness value of the second original solution corresponding to the dimensional accuracy objective function. The fitness values of other objective functions are processed similarly. Then, the difference between each set of corresponding fitness values is calculated to obtain multiple first fitness differences.

[0035] Preferably, the first fitness comparison constraint requires that the performance of the first original solution on all objective functions is better than or equal to the second original solution (that is, all first fitness differences are greater than or equal to 0), and the second fitness comparison constraint requires that the performance of the first original solution on at least one objective function is strictly better than the second original solution (that is, at least one first fitness difference is greater than a preset strictly better threshold, such as 20%); according to the fitness comparison constraint, based on multiple first fitness differences, the first original solution and the second original solution are analyzed to determine the dominance relationship between them. If the first original solution meets the above two constraints, then the first original solution dominates the second original solution; conversely, if the second original solution meets the conditions, then the second original solution dominates the first original solution; if neither meets the conditions, then there is no dominance relationship, and finally the first dominance relationship between the first original solution and the second original solution is determined and generated.

[0036] Preferably, the first original solution and the second original solution are graded according to the determined first dominating relationship. Specifically, if the first original solution dominates the second original solution, the grade of the first original solution is relatively higher; if the second original solution dominates the first original solution, the grade of the second original solution is relatively higher; if there is no dominating relationship between the two, they may be at the same grade; then the grade division result is used as the first stratification result to clarify the relative positions of the two original solutions in the non-dominated sorting; two different original solutions in the original cold drawing process parameter set are repeatedly randomly selected to perform fitness difference calculation, dominating relationship analysis and grade stratification until the processing of any two original solutions in the original cold drawing process parameter set is completed; finally, all stratification results are integrated and a non-dominated solution distribution is constructed, wherein original solutions of different grades are clearly divided, and original solutions at a higher grade (such as the first grade) represent relatively better cold drawing process parameter combinations in multi-objective optimization, which are used for subsequent selection of optimal process parameters, thereby ensuring the quality of steel pipe products.

[0037] Furthermore, step c also includes step c1: if the multiple first fitness differences cannot simultaneously satisfy the first fitness comparison constraint and the second fitness comparison constraint, jump to step e.

[0038] Preferably, after calculating multiple first fitness differences between the first original solution and the second original solution, if there is one or some first fitness difference values less than 0, or all first fitness difference values are greater than or equal to 0 but none of the differences is greater than a preset strictly superior threshold value (such as 20%), it means that multiple first fitness differences cannot simultaneously satisfy the first fitness comparison constraint and the second fitness comparison constraint, and it is determined that there is no dominance relationship between them that satisfies the conditions, that is, no further in-depth analysis of the dominance relationship of this set of first original solutions and the second original solutions is performed, and operations are directly started on the other two original solutions in the original cold drawing process parameter set that have not been compared, that is, two original solutions are randomly selected again, the fitness difference is calculated, and dominance relationship analysis and hierarchical stratification are performed until any two original solutions in the original cold drawing process parameter set are traversed, and finally a non-dominated solution distribution is constructed.

[0039] Step S300 : calculating the congestion degree of each solution in the non-dominated solution distribution to generate a congestion degree distribution.

[0040] Step S300 further includes step S310, calculating the neighborhood solution distribution density within a preset neighborhood range for each solution in the non-dominated solution distribution; step S320, normalizing the neighborhood solution distribution density, generating the congestion degree of each solution, and constructing the congestion degree distribution.

[0041] Preferably, in the non-dominated solution distribution, each solution represents a set of cold drawing process parameter combinations (relatively optimal combinations under multi-objective optimization). For each solution, a preset neighborhood range is defined according to actual conditions. For example, with the solution as the center, a certain distance range is set in the objective function space, and then the number of other solutions within this preset neighborhood range is counted, and this number is used as a measure of the neighborhood solution distribution density of the solution. For example, if there are 10 other solutions within the neighborhood range centered on a certain solution, then the neighborhood solution distribution density of this solution can be expressed as 10 (that is, the number of solutions in the neighborhood is counted). By performing this operation on each solution in the non-dominated solution distribution, the neighborhood solution distribution density corresponding to each solution is obtained. Since the numerical range of the neighborhood solution distribution density of different solutions may be different, these density values are normalized, that is, all density values are mapped to a unified range (usually between 0 and 1). Specifically, the maximum and minimum normalization method is used to calculate the crowding of each solution, and the higher the crowding, the more solutions there are around it, indicating that the density of solutions in this area is greater in the multi-objective optimization space. Then, the crowding of all solutions is integrated to form a crowding distribution, which shows the distribution density of non-dominated solutions in the multi-objective optimization space. Then, the crowding can be used to select relatively uniformly distributed and better solutions as seed solutions to avoid the concentration of solutions in certain local areas, thereby better exploring the entire multi-objective optimization space.

[0042] Step S400 , screening seed solutions based on the non-dominated solution distribution and the crowding degree distribution, performing population optimization iterations until convergence conditions are met, and constructing an optimized population.

[0043] Step S400 further includes step S410, screening solutions with distribution levels higher than a preset excellent level based on the non-dominated solution distribution to generate a first seed solution space; step S420, screening solutions with congestion levels less than a preset congestion level based on the congestion distribution to generate a second seed solution space; step S430, merging the first seed solution space and the second seed solution space and deleting duplicate solutions to construct a first-generation population; step S440, performing crossover and mutation processing on the solution based on the first-generation population to generate a second-generation population; step S450, continuing to analyze the non-dominated solution distribution and congestion distribution of the second-generation population, screening the third-generation population, and so on, iterating multiple times until the convergence condition is met, and constructing the optimized population with the population obtained at convergence.

[0044] Preferably, each solution in the non-dominated solution distribution is divided into different levels (according to the results of non-dominated sorting), and a preset excellent level (such as the first layer, the second layer, etc.) is set according to the actual situation and optimization requirements. Then, solutions with a distribution level higher than the preset excellent level are selected from the non-dominated solution distribution to form a first seed solution space. Since these solutions are at a higher non-dominated level, they are relatively better in multi-objective optimization and have better comprehensive performance on various objective functions. Each solution in the congestion distribution has a corresponding congestion value, which reflects the density of the solution in the multi-objective optimization space. A preset congestion value (such as 0.5) is set according to the actual steel pipe processing requirements. Solutions with a congestion less than the preset congestion are screened out from all solutions to form a second seed solution space. The areas where these solutions are located are relatively sparse in the multi-objective optimization space. Selecting solutions with a small congestion helps to maintain the diversity of the population during the optimization process and avoid the concentration of solutions in certain local areas.

[0045] Preferably, the solutions in the first seed solution space and the second seed solution space are merged to check whether there are duplicate solutions (i.e., solutions with exactly the same process parameter combination). If so, the duplicate solutions are deleted. The resulting solution set is called the first generation population, which contains relatively good and diverse solutions selected from the two perspectives of non-dominated solution distribution and congestion distribution. Then, based on the first generation population, crossover and mutation processing of the solutions is performed. Specifically, two solutions (which can be understood as two sets of cold drawing process parameter combinations) are randomly selected from the first generation population, and a crossover operation is performed according to a certain crossover probability (e.g., 0.8, i.e., there is an 80% probability). The method uses the crossover and mutation operations to exchange some parameters of the two solutions to generate a new solution. For each solution in the first-generation population, some parameters in the solution are randomly changed according to a certain mutation probability (for example, 0.01, which means there is a 1% probability of mutation operation) to generate a mutated solution. For example, the drawing speed parameter in a certain solution is randomly increased or decreased. The new solution obtained by crossover and mutation operations is merged with the solution in the first-generation population to form a second-generation population. The second-generation population not only contains the solutions of the first-generation population, but also introduces new solutions, which increases the diversity of the population and the possibility of exploring new combinations of process parameters.

[0046] Preferably, the analysis and construction of the non-dominated solution distribution and the crowding distribution are repeated for the second-generation population, that is, a multi-objective fitness evaluation is performed on each solution in the second-generation population (if a process parameter that has not been used in history is encountered, a deep learning model is constructed based on historical data to perform fitness evaluation of each objective function), and then non-dominated sorting is performed to obtain the non-dominated solution distribution, and the crowding of each solution is calculated to obtain the crowding distribution. Then, based on the non-dominated solution distribution and the crowding distribution, relatively better and more diverse solutions are screened out to form the third-generation population; the analysis and construction and solution screening are repeated continuously for multiple iterations. When the optimal solution of several consecutive generations of populations does not change significantly (such as the change in the fitness value of the optimal solution is less than a certain threshold), or the diversity of the population reaches a certain level, it is considered that the convergence condition is met. The population obtained at this time is the optimized population, which contains the relatively optimal cold drawing process parameter combination obtained through multiple iterative screening under multi-objective optimization, and is used to guide the actual steel pipe cold drawing production process and ensure the accuracy and quality stability after the steel pipe cold drawing process.

[0047] Furthermore, step S450 also includes that the convergence condition is that the congestion degree of the solutions with distribution levels higher than a preset excellence level in the iteratively updated non-dominated solution distribution of the population is greater than or equal to a preset congestion degree.

[0048] Preferably, operations such as crossover (exchanging some parameters of different process parameter combinations) and mutation (randomly changing some parameters in a certain process parameter combination) of the population solutions in the steel tube cold drawing process parameter optimization generate population updates. For each solution in the updated population (i.e., each cold drawing process parameter combination), fitness evaluation is performed according to multiple objective functions (such as steel tube dimensional accuracy, surface quality, production efficiency, etc.), and the population solutions are divided into non-dominated layers to form a non-dominated solution distribution. For each solution in the non-dominated solution distribution, the neighborhood solution distribution density within a preset neighborhood is calculated (i.e., the number of other solutions in the neighborhood is counted), and normalization is performed to obtain the crowding degree and crowding degree distribution. If the distribution level of the relatively better solution is higher than the preset excellent level, and the crowding degree is greater than or equal to the preset crowding degree, it is considered that the convergence condition has been met, indicating that the area where the relatively better solutions are located is sufficiently dense under the multi-objective optimization, and further iteration may not significantly improve the quality or diversity of the solutions, or the desired optimization effect has been achieved, and then the iteration is stopped, and the current population is used as the final optimized population for actual steel tube cold drawing process parameter setting.

[0049] Step S500 , selecting the solution with the largest fusion fitness in the optimization population as the optimal process parameters, and controlling the cold drawing equipment to perform cold drawing process control on the steel pipe raw material.

[0050] Preferably, the fusion fitness is the result of evaluating each solution by comprehensively considering multiple objective functions (such as dimensional accuracy, strength, surface quality, etc. of the steel pipe), reflecting the comprehensive performance of the process parameter combination on multiple objectives. Selecting the solution with the largest fusion fitness from the optimization population is to select the parameter combination that can optimally meet the requirements of multiple objectives among all possible process parameter combinations, and use it as the control parameters for the operation of the cold drawing equipment. For example, the parameters may include the magnitude of the cold drawing force, the size and shape of the mold, the drawing speed, the lubrication conditions, etc. The cold drawing equipment cold-draws the steel pipe raw material according to these optimal process parameters, which can enable the produced steel pipe to achieve the best comprehensive performance in terms of dimensional accuracy, strength, surface quality, etc., thereby realizing the optimized control of the cold drawing process and improving production efficiency and product quality.

[0051] Furthermore, step S500 also includes step S510, extracting multiple fitness corresponding to each solution in the optimization population and performing weighted fusion to generate a fusion fitness corresponding to each solution; step S520, based on the fusion fitness corresponding to each solution, selecting the solution with the largest fusion fitness as the optimal process parameter.

[0052] Preferably, the fitness of each solution in the optimization population is subjected to weighted fusion processing, that is, a weight is assigned to the fitness value corresponding to each objective function (the weight is determined according to the importance of each objective in actual production), and then the multiple fitness values of each solution are calculated according to the weighted formula to obtain the fusion fitness corresponding to each solution, which represents the comprehensive performance of the solution on multiple objectives. The larger the value, the better the effect in multi-objective optimization. After calculating the fusion fitness of each solution in the optimization population, these fusion fitness values are compared to find the solution with the largest fusion fitness, which can achieve a better balance between multiple objectives. Finally, a set of cold drawing process parameters corresponding to this solution is determined as the optimal process parameters, which are used to control the cold drawing equipment to perform cold drawing process operations on steel pipe raw materials, so as to obtain steel pipe products that are relatively ideal in terms of dimensional accuracy, surface quality, production efficiency and other aspects.

[0053] In the above, refer to Figure 1 The steel pipe cold drawing process parameter optimization method according to the embodiment of the present invention is described in detail. Figure 2 A steel pipe cold drawing process parameter optimization system according to an embodiment of the present invention is described.

[0054] The steel pipe cold drawing process parameter optimization system according to the embodiment of the present invention is used to solve the technical problems existing in the prior art, such as the inability of cold drawing process parameters to accurately adapt to steel pipe raw materials and cold drawing equipment, and the difficulty in balancing the complex needs of multiple objectives, which in turn leads to low steel pipe dimensional accuracy, unstable product quality, and waste of raw materials in industrial control. It realizes precise control of steel pipe cold drawing, achieving the technical effect of improving product quality and dimensional accuracy and reducing raw material loss. Figure 2 As shown, the steel pipe cold drawing process parameter optimization system includes: a cold drawing task requirement feature building module 10, a non-dominated solution distribution building module 20, a congestion distribution generating module 30, an optimized population building module 40, and a cold drawing process control module 50.

[0055] The cold drawing task requirement feature building module 10 is used to interact with the target cold drawing scenario, receive the cold drawing task and analyze it, determine the steel pipe raw material, cold drawing equipment characteristics and target steel pipe size, and build the cold drawing task requirement features; the non-dominated solution distribution building module 20 is used to obtain historical cold drawing process parameters based on the cold drawing task requirement features, and perform fitness evaluation under multiple objectives, perform non-dominated sorting according to the fitness evaluation results, and construct a non-dominated solution distribution; the congestion distribution generation module 30 is used to calculate the congestion of each solution in the non-dominated solution distribution and generate a congestion distribution; the optimized population building module 40 is used to combine the non-dominated solution distribution and the congestion distribution to screen seed solutions, perform population optimization iterations until the convergence conditions are met, and construct an optimized population; the cold drawing process control module 50 is used to select the solution with the largest fusion fitness in the optimized population as the optimal process parameter, and control the cold drawing equipment to perform cold drawing process control on the steel pipe raw material.

[0056] The specific configuration of the non-dominated solution distribution construction module 20 will be described in detail below. The non-dominated solution distribution construction module 20 further includes: using the cold drawing task requirement feature as a retrieval element, performing historical data mining, and constructing an original cold drawing process parameter set, wherein any original cold drawing process parameter has a cold drawing quality inspection result mark; obtaining a steel pipe cold drawing quality index, and constructing multiple objective functions based on the steel pipe cold drawing quality index; based on the cold drawing quality inspection result mark, calling the multiple objective functions to perform fitness evaluation on any original cold drawing process parameter in the original cold drawing process parameter set to generate an original fitness information set; calling a non-dominated sorting mechanism, and performing non-dominated sorting on all individual solutions in the original cold drawing process parameter set based on the original fitness information set to generate the non-dominated solution distribution, wherein the non-dominated sorting mechanism includes a first fitness comparison constraint and a second fitness comparison constraint, and the first fitness comparison constraint and the second fitness comparison constraint need to be satisfied simultaneously.

[0057] The specific configuration of the non-dominated solution distribution construction module 20 will be described in detail below. The non-dominated solution distribution construction module 20 further includes: the first fitness comparison constraint is a first fitness optimization value based on the multiple objective functions, and the second fitness comparison constraint is a second fitness optimization value based on one of the multiple objective functions.

[0058] The specific configuration of the non-dominated solution distribution construction module 20 will be described in detail below. The non-dominated solution distribution construction module 20 further includes: step a: randomly selecting a first original solution and a second original solution from the original cold drawing process parameter set, and matching multiple first fitnesses and multiple second fitnesses from the original fitness information set; step b: performing fitness difference calculations of the same type on the multiple first fitnesses and the multiple second fitnesses after aligning them according to the objective function type, and generating multiple first fitness differences; step c: based on the multiple first fitness differences, performing a dominance relationship analysis on the first original solution and the second original solution according to the first fitness comparison constraint and the second fitness comparison constraint, and generating a first dominance relationship; step d: hierarchically stratifying the first original solution and the second original solution based on the first dominance relationship, and generating a first stratification result; step e: and repeating steps a to d until any two original solutions in the original cold drawing process parameter set are traversed, and the non-dominated solution distribution is constructed with all generated stratification results.

[0059] The specific configuration of the non-dominated solution distribution construction module 20 will be described in detail below. The non-dominated solution distribution construction module 20 further includes: if the multiple first fitness differences cannot simultaneously satisfy the first fitness comparison constraint and the second fitness comparison constraint, skip to step e.

[0060] The specific configuration of the congestion distribution generation module 30 will be described in detail below. The congestion distribution generation module 30 further includes: calculating, for each solution in the non-dominated solution distribution, a neighborhood solution distribution density within a preset neighborhood range; normalizing the neighborhood solution distribution density, generating a congestion degree for each solution, and constructing the congestion distribution.

[0061] The specific configuration of the optimized population construction module 40 will be described in detail below. The optimized population construction module 40 further includes: screening solutions with a distribution level higher than a preset excellent level based on the non-dominated solution distribution to generate a first seed solution space; screening solutions with a congestion level less than a preset congestion level based on the congestion distribution to generate a second seed solution space; merging the first seed solution space and the second seed solution space and deleting duplicate solutions to construct a first-generation population; performing crossover and mutation processing on the solutions based on the first-generation population to generate a second-generation population; continuing to analyze the non-dominated solution distribution and congestion distribution of the second-generation population to screen a third-generation population, and so on, iterating multiple times until the convergence condition is met, and constructing the optimized population with the population obtained at convergence.

[0062] The following will further describe the specific configuration of the optimized population construction module 40. The optimized population construction module 40 further includes: the convergence condition is that the congestion degree of the solutions with distribution levels higher than the preset excellence level in the iteratively updated non-dominated solution distribution of the population is greater than or equal to the preset congestion degree.

[0063] The specific configuration of the cold drawing process control module 50 will be described in detail below. The cold drawing process control module 50 further includes: extracting multiple fitness values corresponding to each solution in the optimization population and performing weighted fusion to generate a fused fitness value corresponding to each solution; and based on the fused fitness values corresponding to each solution, selecting the solution with the largest fused fitness value as the optimal process parameters.

[0064] The steel pipe cold drawing process parameter optimization system provided in the embodiment of the present invention can execute the steel pipe cold drawing process parameter optimization method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0065] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other and are not used to limit the scope of protection of the present invention.

[0066] The above specific embodiments do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application shall be included within the scope of protection of this application.

Claims

1. A method for optimizing process parameters of cold drawing of steel pipes, characterized in that: The method comprises: Interactive target cold drawing scenario, receiving and analyzing cold drawing tasks, determining the steel pipe raw material, cold drawing equipment characteristics and target steel pipe size, and establishing the cold drawing task requirement characteristics; Based on the cold drawing task requirement characteristics, historical cold drawing process parameters are obtained, and fitness evaluation under multiple objectives is performed. Non-dominated sorting is performed according to the fitness evaluation results, and a non-dominated solution distribution is constructed. Calculating the congestion degree of each solution in the non-dominated solution distribution to generate a congestion degree distribution; Combining the non-dominated solution distribution and the crowding degree distribution to select seed solutions, performing population optimization iterations until convergence conditions are met, and constructing an optimized population; The solution with the largest fusion fitness is selected from the optimization population as the optimal process parameters, and the cold drawing equipment is controlled to perform cold drawing process control on the steel pipe raw material.

2. The method for optimizing the cold drawing process parameters of a steel pipe according to claim 1, wherein: Based on the cold drawing task requirement characteristics, historical cold drawing process parameters are obtained, and fitness evaluation under multiple objectives is performed. Non-dominated sorting is performed according to the fitness evaluation results, and a non-dominated solution distribution is constructed, including: Using the cold drawing task requirement characteristics as retrieval elements, historical data mining is performed to construct a set of original cold drawing process parameters, wherein any original cold drawing process parameter has a cold drawing quality inspection result mark; Acquire a steel pipe cold-drawing quality index, and construct a plurality of objective functions based on the steel pipe cold-drawing quality index; Based on the cold drawing quality inspection result mark, calling the multiple objective functions to perform fitness evaluation on any original cold drawing process parameter in the original cold drawing process parameter set to generate an original fitness information set; A non-dominated sorting mechanism is called to perform non-dominated sorting on all individual solutions in the original cold drawing process parameter set based on the original fitness information set to generate the non-dominated solution distribution, wherein the non-dominated sorting mechanism includes a first fitness comparison constraint and a second fitness comparison constraint, and the first fitness comparison constraint and the second fitness comparison constraint need to be satisfied at the same time.

3. The method for optimizing the cold drawing process parameters of a steel pipe according to claim 2, wherein: The first fitness comparison constraint is a first fitness optimization value based on the multiple objective functions, and the second fitness comparison constraint is a second fitness optimization value based on one objective function among the multiple objective functions.

4. The method for optimizing the cold drawing process parameters of a steel pipe according to claim 3, wherein: Invoking a non-dominated sorting mechanism to perform non-dominated sorting on all individual solutions in the original cold drawing process parameter set based on the original fitness information set to generate the non-dominated solution distribution, including: Step a: randomly selecting a first original solution and a second original solution from the original cold drawing process parameter set, and matching a plurality of first fitnesses and a plurality of second fitnesses in the original fitness information set; Step b: performing fitness difference calculation of the same type on the multiple first fitnesses and the multiple second fitnesses according to the objective function type to generate multiple first fitness differences; Step c: Based on the multiple first fitness differences, performing a dominance relationship analysis on the first original solution and the second original solution according to the first fitness comparison constraint and the second fitness comparison constraint to generate a first dominance relationship; Step d: hierarchically stratifying the first original solution and the second original solution based on the first dominance relationship to generate a first stratification result; Step e: Repeat steps a to d until any two original solutions in the original cold drawing process parameter set are traversed, and the non-dominated solution distribution is constructed with all generated hierarchical results.

5. The method for optimizing the cold drawing process parameters of a steel pipe according to claim 4, wherein: Performing a dominance relationship analysis on the first original solution and the second original solution according to the first fitness comparison constraint and the second fitness comparison constraint further includes: If the multiple first fitness differences cannot simultaneously satisfy the first fitness comparison constraint and the second fitness comparison constraint, jump to step e.

6. The method for optimizing process parameters of cold drawing of steel pipe according to claim 1, characterized in that: Calculating the congestion degree of each solution in the non-dominated solution distribution to generate a congestion degree distribution includes: For each solution in the non-dominated solution distribution, calculating a neighborhood solution distribution density within a preset neighborhood range; Normalizing and assigning values to the neighborhood solution distribution density, generating the congestion degree of each solution, and constructing the congestion degree distribution.

7. The method for optimizing cold drawing process parameters of a steel pipe according to claim 1, wherein: The seed solution is selected by combining the non-dominated solution distribution and the crowding degree distribution, and the population optimization iteration is performed until the convergence condition is met to construct the optimized population, including: Screening solutions with distribution levels higher than a preset excellence level based on the non-dominated solution distribution to generate a first seed solution space; Based on the congestion distribution, solutions with a congestion degree less than a preset congestion degree are screened to generate a second seed solution space; Merging the first seed solution space and the second seed solution space and deleting duplicate solutions to construct a first generation population; Based on the first generation population, crossover and mutation processing of the solution are performed to generate the second generation population; The non-dominated solution distribution and the congestion distribution of the second generation population are continuously analyzed to select the third generation population, and so on, iterating multiple times until the convergence condition is met, and constructing the optimized population with the population obtained at the time of convergence.

8. The method for optimizing process parameters of cold drawing of steel pipe according to claim 1, wherein: The convergence condition is that the congestion degree of the solutions with distribution levels higher than a preset excellence level in the iteratively updated non-dominated solution distribution of the population is greater than or equal to a preset congestion degree.

9. The method for optimizing process parameters of cold drawing of steel pipe according to claim 1, characterized in that: Selecting the solution with the largest fusion fitness in the optimization population as the optimal process parameters includes: Extracting multiple fitnesses corresponding to each solution in the optimized population and performing weighted fusion to generate a fusion fitness corresponding to each solution; Based on the fusion fitness corresponding to each solution, the solution with the largest fusion fitness is selected as the optimal process parameters.

10. Steel pipe cold drawing process parameter optimization system, characterized by: The system is used to implement the steel pipe cold drawing process parameter optimization method according to any one of claims 1 to 9, and the system comprises: The cold drawing task requirement feature building module is used to interact with the target cold drawing scenario, receive and analyze the cold drawing task, determine the steel pipe raw material, cold drawing equipment characteristics and target steel pipe size, and build the cold drawing task requirement features; a non-dominated solution distribution construction module, configured to obtain historical cold drawing process parameters based on the cold drawing task requirement characteristics, perform fitness evaluation under multiple objectives, perform non-dominated sorting according to the fitness evaluation results, and construct a non-dominated solution distribution; a congestion degree distribution generating module, configured to calculate the congestion degree of each solution in the non-dominated solution distribution to generate a congestion degree distribution; An optimized population construction module is used to select seed solutions by combining the non-dominated solution distribution and the crowding degree distribution, perform population optimization iterations until convergence conditions are met, and construct an optimized population; The cold drawing process control module is used to select the solution with the largest fusion fitness in the optimization population as the optimal process parameters, and control the cold drawing equipment to perform cold drawing process control on the steel pipe raw material.

Images