Rolling parameter optimization method and device, equipment, storage medium and product

Abstract

The invention provides a rolling parameter optimization method and device, equipment, a storage medium and a product. The method comprises the steps that strip steel specification information of strip steel currently produced by a cold continuous rolling unit is obtained; according to the strip steel specification information, whether the corresponding strip steel needs to be subjected to production optimization or not is determined through a preset support vector machine model; under the condition that the strip steel needs to be subjected to production optimization, rolling parameter constraint conditions of the strip steel are determined; and according to the rolling parameter constraint condition, an optimization result is obtained through a preset evolutionary algorithm model, and the optimization result comprises optimized rolling parameters.

Description

Technical Field

[0001] This application relates to the technical field of metal processing, and in particular, to a rolling parameter optimization method, device, equipment, storage medium, and product. Background Art

[0002] In modern steel production, the tandem cold rolling mill is an important part of strip production. The existing rolling force calculation methods cannot accurately reflect the rolling force in actual production. At present, there is a large difference between the calculated value and the actual value of the rolling force in the tandem cold rolling mill, and there are also problems with unreasonable grading of some steel grades, resulting in a large length of out-of-tolerance thickness at the head and tail of the strip and a high cutting loss, which seriously affects the production efficiency and brings varying degrees of economic losses.

[0003] In the prior art, related optimization algorithms, such as optimizing the reduction schedule to reduce rolling force fluctuations or improving the prediction accuracy of rolling force, mainly focus on optimizing the rolling force and fail to effectively increase the rolling speed and improve the production efficiency of rolling. Summary of the Invention

[0004] Embodiments of this application provide a rolling parameter optimization method, device, equipment, storage medium, and product, which can improve the production efficiency of rolling.

[0005] In a first aspect, this application provides a rolling parameter optimization method, and the method includes:

[0006] Obtain the strip specification information of the strip currently produced by the tandem cold rolling mill;

[0007] According to the strip specification information, determine whether the corresponding strip needs production optimization through a preset support vector machine model;

[0008] In the case where the strip needs production optimization, determine the rolling parameter constraint conditions of the strip;

[0009] According to the rolling parameter constraint conditions, obtain an optimization result through a preset evolutionary algorithm model, and the optimization result includes optimized rolling parameters.

[0010] In some possible implementation manners, the obtaining an optimization result through a preset evolutionary algorithm model according to the rolling parameter constraint conditions, where the optimization result includes optimized rolling parameters, includes:

[0011] According to the rolling parameter constraint conditions and the optimization objective corresponding to the preset evolutionary algorithm model, determine an initial population and the corresponding fitness evaluation criteria, and the initial population includes multiple initial individuals representing rolling parameter combinations;

[0012] Perform an evolutionary operation on the initial population according to the preset evolutionary algorithm model and the fitness evaluation criteria to determine the optimal individual in the initial population;

[0013] Obtain an optimization result according to the optimal individual.

[0014] In some possible implementation manners, the performing an evolutionary operation on the initial population according to the preset evolutionary algorithm model and the fitness evaluation criteria to determine the optimal individual in the initial population includes:

[0015] Use a preset selection algorithm to select two individuals from the initial population, and determine the individual with a higher fitness value according to the fitness evaluation criteria among the two individuals as the parent individual;

[0016] Use a simulated binary crossover algorithm to perform a crossover operation on the parent individuals to generate new offspring individuals;

[0017] Use a polynomial mutation algorithm to perform a mutation operation on the offspring individuals to update the initial population;

[0018] In the case where the initial population does not meet the preset stop condition, return to the step of using a preset selection algorithm to select two individuals from the initial population and determining the individual with a higher fitness value according to the fitness evaluation criteria among the two individuals as the parent individual;

[0019] In the case where the initial population meets the preset stop condition, determine the optimal individual in the initial population according to the fitness evaluation criteria.

[0020] In some possible implementation manners, before determining whether the corresponding strip steel needs to be optimized for production according to the strip steel specification information through a preset support vector machine model, the method further includes:

[0021] Obtain the historical production data of the tandem cold rolling mill;

[0022] Construct a training data set according to the historical production data, where the training data set includes multiple training samples;

[0023] Use the training data set to train an initial support vector machine model to obtain a preset support vector machine model.

[0024] In some possible implementation manners, the obtaining the historical production data of the tandem cold rolling mill includes:

[0025] Obtain the data, steel grade, and corresponding strip steel physical parameters, strip steel rolling parameters, and strip steel inlet and outlet speeds in the historical production of the tandem cold rolling mill;

[0026] Classify the data according to the steel grade number to obtain historical production data.

[0027] In some possible implementation manners, constructing a training data set according to the historical production data includes:

[0028] Determine a low-speed threshold and a high-speed threshold corresponding to strip steel production according to the production conditions of the tandem cold rolling mill;

[0029] Screen out the data in the historical production data that meet the low-speed threshold and the high-speed threshold to obtain a first data set;

[0030] Delete the data containing missing values in the first data set to obtain a second data set;

[0031] Normalize the second data set to construct a training data set.

[0032] In some possible implementation manners, training an initial support vector machine model by using the training data set to obtain a preset support vector machine model includes:

[0033] Determine an initial support vector machine model and a corresponding kernel function;

[0034] Set an initial support vector machine model and corresponding hyperparameters;

[0035] Train the initial support vector machine model by using each training sample in the training data set to obtain a preset support vector machine model.

[0036] In a second aspect, the present application provides a rolling parameter optimization device, where the device includes:

[0037] An acquisition module, configured to acquire strip steel specification information of the strip steel currently produced by the tandem cold rolling mill;

[0038] A determination module, configured to determine whether the corresponding strip steel needs production optimization through a preset support vector machine model according to the strip steel specification information;

[0039] The determination module is further configured to determine rolling parameter constraint conditions of the strip steel in the case where the strip steel needs production optimization;

[0040] An optimization module, configured to obtain an optimization result according to the rolling parameter constraint conditions through a preset evolutionary algorithm model, where the optimization result includes optimized rolling parameters.

[0041] In a third aspect, the present application provides a rolling parameter optimization device, where the device includes: a processor and a memory storing computer program instructions; the processor reads and executes the computer program instructions to implement the rolling parameter optimization method described above.

[0042] Fourthly, the present application provides a computer-readable storage medium, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the rolling parameter optimization method described above is implemented.

[0043] Fifthly, the present application provides a computer program product. When the instructions in the computer program product are executed by a processor of an electronic device, the electronic device is caused to execute the rolling parameter optimization method described above.

[0044] The rolling parameter optimization method, device, equipment, storage medium and product provided by the embodiments of the present application apply machine learning and evolutionary algorithms to the optimization of cold tandem rolling production, realizing the intelligentization and adaptive optimization of the production process. The support vector machine algorithm can autonomously learn and predict based on historical production data, and the evolutionary algorithm model can autonomously search and optimize rolling parameters, thereby improving the production efficiency of rolling. Description of the Drawings

[0045] The present application can be better understood from the following description of the specific embodiments in conjunction with the drawings, where:

[0046] By reading the following detailed description of non-limiting embodiments with reference to the drawings, other features, objects and advantages of the present application will become more obvious, where the same or similar reference numerals represent the same or similar features.

[0047] Figure 1 is a flowchart of the rolling parameter optimization method provided by an embodiment of the present application;

[0048] Figure 2 is a flowchart of the rolling parameter optimization method provided by another embodiment of the present application;

[0049] Figure 3 is a flowchart of the rolling parameter optimization method provided by yet another embodiment of the present application;

[0050] Figure 4 is a schematic structural diagram of the rolling parameter optimization device provided by an embodiment of the present application;

[0051] Figure 5 is a schematic hardware structure diagram of the rolling parameter optimization equipment provided by the embodiments of the present application. Detailed Embodiments

[0052] The features and exemplary embodiments of various aspects of the present application will be described in detail below. To make the objectives, technical solutions, and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application and not to limit the present application. For those skilled in the art, the present application can be implemented without some of these specific details. The following description of the embodiments is only intended to provide a better understanding of the present application by showing examples of the present application.

[0053] It should be noted that, in this document, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, such that a process, method, article or device including a series of elements not only includes those elements but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, elements defined by the statement "including..." do not exclude the presence of additional identical elements in the process, method, article or device including the said elements.

[0054] To solve the problems of the prior art, embodiments of the present application provide a rolling parameter optimization method, device, equipment, storage medium and product. First, the rolling parameter optimization method provided by the embodiments of the present application will be introduced below.

[0055] Figure 1 The flowchart of the rolling parameter optimization method provided by an embodiment of the present application is shown. As Figure 1 shown, the method includes the following steps: S101 to S104.

[0056] S101: Obtain the strip specification information of the strip currently produced by the tandem cold rolling mill.

[0057] In a specific implementation, connect to the tandem cold rolling mill equipment through an industrial control system or a production information management system to obtain the real-time production data of the current strip. The data includes specification information such as the width, thickness, material, and alloy composition of the strip. The received raw data usually needs to be formatted and cleaned. Common processes include removing noise data, filling in missing data, and converting to a unified unit.

[0058] S102: Determine whether the corresponding strip needs production optimization through a preset support vector machine model according to the above strip specification information.

[0059] In specific implementation, beforehand, it is necessary to train a support vector machine model using the specification information of the strip steel and the annotation of whether to perform production optimization based on historical data and production experience. The training data includes the strip steel specifications and the corresponding optimization requirements. During the real-time production process, the strip steel specification information obtained from step S101 is used as input data and passed to the SVM model. The data may be preprocessed, such as normalized, standardized, etc., to ensure its consistency with the training data. The SVM model will classify based on the input strip steel specification information and output the prediction result of whether production optimization is required. If the model predicts optimization is needed, the system will enter step S103 to determine the rolling parameters for optimization.

[0060] S103: In the case where the above strip steel needs to be optimized in production, determine the rolling parameter constraint conditions for the above strip steel.

[0061] In specific implementation, in the case where the above strip steel needs to be optimized in production, according to the strip steel specifications and material properties, combined with process specifications, determine the initial rolling parameter range. These process specifications usually come from the standardized production process or are obtained from production experience. During the rolling process, parameters such as temperature, pressure, rolling speed, and tension will all affect the final strip steel quality. The system analyzes which parameters need special attention based on the strip steel specification information. For example, strip steel with a larger thickness may require a lower rolling speed and a higher reduction amount to ensure no defects occur. Set the upper and lower limits of the rolling parameters according to the system's process model and empirical rules. For example, the maximum value of the rolling temperature, the minimum value of the reduction amount, the range of the rolling speed, etc. These parameters should not only meet the quality requirements but also consider the operating range and production capacity of the equipment. Determine the rolling parameters that need to be adjusted during the optimization process, as well as their priorities and constraint conditions.

[0062] S104: According to the above rolling parameter constraint conditions, obtain the optimization result through a preset evolutionary algorithm model. The above optimization result includes the optimized rolling parameters.

[0063] In a specific implementation, based on the set rolling parameter constraints, an evolutionary algorithm model is initialized. For each individual, the evolutionary algorithm evaluates its fitness. Fitness is usually calculated through an objective function that may consider various factors such as the final quality of the strip steel, production efficiency, and equipment load. In each generation of the algorithm, individuals with higher fitness are selected for crossover and mutation operations to generate new combinations of rolling parameters. The crossover operation simulates the recombination of species genes, and the mutation operation introduces diversity by randomly adjusting some parameters of the individual to avoid getting trapped in local optimal solutions. After several generations of iteration, the algorithm gradually converges to find the optimal rolling parameters. Each generation is screened according to the constraint conditions to ensure that the final optimized parameters meet all production requirements. When the algorithm reaches the convergence condition, the optimized rolling parameters are output to obtain the optimization result.

[0064] To ensure that the final combination of rolling parameters achieves the optimal production effect, in some embodiments, the above S104 may include the following steps: S1041 to S1043.

[0065] S1041: According to the above rolling parameter constraints and the optimization objectives corresponding to the preset evolutionary algorithm model, determine the initial population and the corresponding fitness evaluation criteria. The above initial population includes multiple initial individuals representing combinations of rolling parameters.

[0066] In a specific implementation, parameters related to the rolling process are determined, such as temperature, reduction, rolling speed, and tension. Each individual represents a combination of rolling parameters. Different individuals will have different parameter values, forming a diverse initial population. The initial population consists of multiple individuals, that is, different combinations of rolling parameters. The gene values of these individuals are randomly generated within a defined range. The population size is a hyperparameter and is usually determined according to the complexity of the problem and the computing power. A larger population can increase the search space and improve the probability of finding the optimal solution. The fitness function defines the "goodness" of each individual under a specific objective. If there are multiple optimization objectives, different weights may need to be assigned to each objective. For example, in some cases, quality is more important than efficiency, so the quality objective may have a higher weight. Each individual obtains a fitness value by calculating the objective function, indicating the quality of the combination. Individuals with high fitness values represent good optimization effects, while low ones may indicate unsuitable rolling parameters.

[0067] S1042: Through the preset evolutionary algorithm model, according to the above fitness evaluation criteria, perform evolutionary operations on the above initial population to determine the optimal individual in the above initial population.

[0068] In a specific implementation, the selection operation selects which individuals will reproduce based on fitness values. During this process, individuals with higher fitness are more likely to be selected, while those with lower fitness may be eliminated. Common selection methods include roulette wheel selection and tournament selection. Subsequently, the crossover operation is performed. The crossover operation simulates gene recombination by combining the genes of two individuals into a new individual. Common ways of the crossover operation include single-point crossover, multi-point crossover, etc. For example, the rolling parameters of two parental individuals are divided into two segments, and some parameters are exchanged to generate two new offspring individuals. Moreover, the mutation operation is also carried out. The mutation operation randomly modifies some genes of an individual within a small range to introduce diversity and prevent the algorithm from falling into a local optimal solution. For the new population after selection, crossover, and mutation, the system will re-evaluate the fitness of each individual. This process will continue for multiple generations. Each generation generates new individuals through evolutionary operations (selection, crossover, mutation) and continuously optimizes the solution space. When the change in population fitness becomes tiny or reaches the preset maximum number of generations, the evolutionary process will stop.

[0069] S1043: Obtain the optimization result based on the above optimal individual.

[0070] In a specific implementation, after multiple generations of evolution, the system will screen out the individual with the highest fitness. This individual represents the optimal combination of rolling parameters. Individuals with high fitness scores will be selected as the optimal solution. The gene values (rolling parameters) of the optimal individual are the final optimization result.

[0071] Through the above implementation manner of the embodiment of the present application, according to the above rolling parameter constraint conditions and the optimization objective corresponding to the preset evolutionary algorithm model, the initial population and the corresponding fitness evaluation criteria are determined. Then, through the preset evolutionary algorithm model, according to the above fitness evaluation criteria, evolutionary operations are performed on the above initial population to determine the optimal individual in the above initial population. Based on the above optimal individual, the optimization result is obtained. The initial population is randomly generated and evolved through operations such as crossover and mutation, and finally the optimal individual is selected as the production parameter to ensure that the final rolling parameter combination achieves the optimal production effect.

[0072] In order to obtain the optimal individual in the population, in some implementation manners, the above S1042 may include the following steps: S10421 to S10425.

[0073] S10421: Use a preset selection algorithm to select two individuals from the above initial population, and determine the individual with a higher fitness value according to the above fitness evaluation criteria among the above two individuals as the parental individual.

[0074] In a specific implementation, in a genetic algorithm, an initial population consists of multiple individuals, and each individual represents a possible solution. Here, a preset selection algorithm is used. Common selection algorithms include roulette wheel selection, tournament selection, ranking selection, etc. In this step, two individuals are selected as parents. The fitness evaluation criterion is defined according to the specific requirements of the problem and is usually a function that gives a numerical value based on the performance of an individual. The larger the value, the better the performance of the individual. From the two individuals selected by the selection algorithm, the individual with a higher fitness value is selected as the parent. The purpose of this step is to ensure that individuals with stronger fitness have a higher probability of participating in the crossover process and thus may produce better offspring.

[0075] S10422: Use the simulated binary crossover algorithm to perform a crossover operation on the above parent individuals to generate new offspring individuals.

[0076] In a specific implementation, simulated binary crossover is a crossover operation for continuous spaces. It generates new offspring individuals by simulating the crossover of two parent individuals. The key lies in generating a new solution that not only retains the characteristics of the parent individuals but also may introduce new characteristics through crossover. Assuming that the gene encoding of the parent individuals is a binary string or a floating-point number, the core of the crossover operation is to exchange part of the gene information of the parent individuals. After the crossover operation, a pair of new individuals is generated. The gene information of these offspring individuals comes from two parents and may perform better than the parent individuals in some aspects.

[0077] S10423: Use the polynomial mutation algorithm to perform a mutation operation on the above offspring individuals to update the above initial population.

[0078] In a specific implementation, according to the set mutation probability, it is determined whether to mutate an individual. Each gene is operated on, and a mutation value is generated using a polynomial distribution. This makes the mutated gene closer to the gene of the current individual but with a slight change, thus introducing a new solution. Through the mutation operation, the genes of the offspring individuals will change, and the new individuals will be added to the population, and the original population will also be updated.

[0079] S10424: In the case where the above initial population does not meet the preset stop condition, return to the step of using the preset selection algorithm to select two individuals from the above initial population and determining the individual with a higher fitness value among the two individuals as the parent individual according to the above fitness evaluation criterion.

[0080] In specific implementation, the preset stopping conditions generally include the maximum number of iterations, the fitness threshold reaching a certain set value, or the diversity of the population becoming too low, etc. The judgment of the stopping conditions is an important basis for whether the algorithm can stop iterating. If the current population has not met the stopping conditions, it means that the algorithm needs to continue to perform the evolution operation. At this time, return to the step of selecting parental individuals (S10421), and continue to perform operations such as crossover and mutation. The process of continuous iteration helps to optimize the solution, while maintaining the diversity of the population and preventing premature convergence. This process continuously improves the overall fitness of the population.

[0081] S10425: In the case where the above initial population meets the preset stopping conditions, according to the above fitness evaluation criteria, determine the optimal individual in the above initial population.

[0082] In specific implementation, after the stopping conditions are met, the system will evaluate the fitness of the entire population, usually by calculating the fitness values of all individuals to select the optimal individual. According to the fitness evaluation criteria, select the individual with the strongest fitness value as the final output. This optimal individual is the result of the genetic algorithm, which represents the optimal solution or approximate optimal solution to solve the problem.

[0083] The above implementation manner of the embodiments of the present application enables the individuals in the population to gradually evolve into solutions that are more suitable for the problem through continuous operations of selection, crossover, and mutation. The evolution of each generation is based on the parental individuals of the previous generation, and new offspring individuals are generated through crossover and mutation, thereby improving the fitness of the population in each generation. Until the preset stopping conditions are reached, and then the optimal individual in the population is obtained.

[0084] In order to train a high-performance prediction model, in some embodiments, refer to Figure 2 , before the above S102, the above method may further include the following steps: S201 to S203.

[0085] S201: Obtain the historical production data of the tandem cold rolling mill.

[0086] In specific implementation, the tandem cold rolling mill is usually equipped with a variety of sensors, and these sensors continuously collect real-time data during the production process. The data generated during the production process is often stored in a local database or a cloud database. The historical production data may contain some missing values. Usually, interpolation methods, mean filling methods, or deletion methods are used to process the missing data. By setting thresholds or using statistical methods to detect and process outliers, abnormal data is avoided from affecting model training. Different data sources may have different data formats, and it is necessary to uniformly format the data.

[0087] S202: According to the above historical production data, construct a training data set, and the above training data set includes multiple training samples.

[0088] In a specific implementation, each training sample usually includes input features and label values. The input features are used to describe the working state of the tandem cold rolling mill unit, while the label values represent the production results under this state. According to business requirements and actual production conditions, features closely related to production performance (such as temperature, speed, pressure, time, etc.) are selected. These features may come from different sensors, devices, or operating parameters. If the original data is relatively complex (such as multi-dimensional time series data), feature extraction is required. Feature extraction techniques include time window sliding, Fourier transform, wavelet transform, etc., which are used to convert complex original data into features that can be used in machine learning models.

[0089] The label set represents the targets to be predicted by the model (such as product quality, equipment failure, etc.). The labels can be numerical values in a regression task (such as coil thickness, production speed, etc.), or categories in a classification task (such as whether a failure occurs, whether the product quality standard is met, etc.). Label data is usually obtained through annotation of on-site production data or manual annotation. In some cases, the software may need to use specific algorithms or rules to automatically calculate the labels.

[0090] To verify the performance of the model, the data set is usually divided into a training set and a validation set. The training set is used to train the model, and the validation set is used to verify the accuracy of the model during the training process to ensure that the model is not overfitted. Since the historical data is time series data, special attention needs to be paid to the time series characteristics when constructing the training set to avoid data leakage. The common method is to divide the training set and the validation set in chronological order to ensure that the training set only contains past data and the validation set contains the data that can be seen when the model makes predictions. Through the above steps, a data set containing multiple training samples is constructed.

[0091] S203: Use the above training data set to train the initial support vector machine model to obtain a preset support vector machine model.

[0092] To verify the performance of the model, during the model training stage, the training data set is input into the SVM model, and an optimization algorithm is used to find the optimal hyperplane or regression curve, so as to minimize the classification or regression error and obtain a preset support vector machine model.

[0093] The above implementation manner of the embodiment of the present application constructs a suitable training data set by obtaining the historical production data of the tandem cold rolling mill unit, and uses the support vector machine model to train the data to ensure that a high-performance prediction model is trained.

[0094] To achieve in-depth analysis of the production data of the tandem cold rolling mill unit, in some implementation manners, the above S201 may include the following steps: S2011 to S2012.

[0095] S2011: Obtain the data, steel grade, and corresponding strip physical parameters, strip rolling parameters, and strip inlet and outlet speeds in the historical production of the tandem cold rolling mill.

[0096] In specific implementation, first, a connection needs to be established with the historical data storage system. Historical data storage usually has a set of timestamps and data structures to save the detailed information in each production process. The data to be obtained includes: steel grade, physical parameters, rolling parameters, and strip inlet and outlet speeds.

[0097] The steel grade is the steel type and its corresponding grade information for each production; the physical parameters include the thickness, width, surface quality, etc. of the strip; the rolling parameters include temperature, pressure, speed, load, etc.

[0098] S2012: Classify the above data according to the above steel grade to obtain historical production data.

[0099] In specific implementation, extract the steel grade by parsing the records of historical data. Once the steel grade data is obtained, classify the historical data according to the steel grade. The classified data can be stored as multiple data sets, and each data set corresponds to a steel grade. These data sets can be stored in different database tables or stored in memory in different variable or file formats.

[0100] The above implementation manner of the embodiment of the present application realizes in-depth analysis and optimization of the production data of the tandem cold rolling mill by obtaining historical data from different data sources, ensuring data quality, classifying the data according to the steel grade, and facilitating subsequent analysis.

[0101] In order to obtain data sets with a unified scale, in some implementation manners, the above S202 may include the following steps: S2021 to S2024.

[0102] S2021: Determine the low-speed threshold and high-speed threshold corresponding to the strip production according to the production conditions of the above tandem cold rolling mill.

[0103] In specific implementation, in the tandem cold rolling mill, the production rate of the strip is usually determined by multiple parameters (such as the thickness of the strip, material type, mill equipment configuration, temperature, etc.). According to the process requirements of the tandem cold rolling mill and the working characteristics of the machine, the "low-speed threshold" and "high-speed threshold" can be determined. These thresholds are usually set through process analysis, statistics of historical data, or standards of equipment performance.

[0104] S2022: Screen out the data in the above historical production data that meets the above low-speed threshold and the above high-speed threshold to obtain the first data set.

[0105] In a specific implementation, once the low-speed and high-speed thresholds are determined, the system needs to screen the historical production data according to these thresholds. The screened data set contains all records that meet the production speed requirements, which is the first data set.

[0106] S2023: Delete the data containing missing values in the above first data set to obtain a second data set.

[0107] In a specific implementation, in the historical production data, there may be some records missing certain fields, such as production speed, temperature, rolling force, etc. These missing values may affect subsequent data analysis and model training, so they need to be deleted. After screening out the first data set, check whether each piece of data contains missing values. If a key field (such as speed, temperature, etc.) of a piece of data is missing, then delete that piece of record. The data set after deleting the missing values is the second data set. All records in this data set are complete and contain all required fields.

[0108] S2024: Normalize the above second data set to construct a training data set.

[0109] In a specific implementation, the strip production data may contain multiple features such as production speed, temperature, and pressure, and the value ranges of these features vary greatly. To avoid some features dominating the model training process due to excessive values, normalization processing is usually required. The data after normalization is the input data (feature data) of the model. After normalization, these data can be split into a training set and a test set to construct a training data set for machine learning models to train and evaluate.

[0110] The above implementation manner of the embodiment of the present application finally obtains a clean and uniformly scaled data set suitable for the training process of machine learning algorithms through threshold screening, missing value processing, and normalization operations on the production speed.

[0111] To obtain a support vector machine model for effective prediction, in some implementation manners, refer to Figure 3 , the above S203 may include the following steps: S2031 to S2033.

[0112] S2031: Determine an initial support vector machine model and the corresponding kernel function.

[0113] In a specific implementation, determine the initial support vector machine model and decide the corresponding kernel function according to the data characteristics of the problem.

[0114] S2032: Set the initial support vector machine model and the corresponding hyperparameters.

[0115] In specific implementation, when initializing the model, reasonable default values are set, and corresponding hyperparameters are tuned through methods such as cross-validation.

[0116] S2033: Use each training sample in the above training dataset to train the initial support vector machine model to obtain a preset support vector machine model.

[0117] In specific implementation, a method is used to train the support vector machine model. The training process is based on the features and target labels in the training dataset. The model calculates the support vectors of each training sample and searches for a hyperplane in the high-dimensional feature space to separate the positive and negative samples as much as possible and maximize the interval between them, thus obtaining a preset support vector machine model.

[0118] The above implementation manner of the embodiments of the present application obtains a support vector machine model that can effectively predict new data by reasonably setting the kernel function and hyperparameters and using the training set to train the model.

[0119] In an embodiment of the present application, first, the production historical data of the tandem cold rolling mill unit is obtained and classified by steel grade. Then, for steel grades of the same brand, according to the production speed threshold, a strip specification dataset that needs production optimization is screened out and trained using the support vector machine algorithm. For steel grades that need production optimization, a decomposition-based multi-objective evolutionary algorithm is used to optimize the relevant rolling parameters and the strip exit speed. Finally, in actual tandem cold rolling production, the above algorithm is applied. For strips of different specifications, first, it is judged whether production optimization is needed, and then the production capacity of the strips that need optimization is increased.

[0120] For steel grades of the same brand, according to the production speed threshold, a strip specification dataset that needs production optimization is screened out. In the specific process of using the support vector machine algorithm for training, for steel grades of the same brand, according to the production historical data, the low production speed threshold v low and the high production speed threshold v high are determined. According to the determined thresholds, a strip specification dataset that needs production optimization is screened out. The screened dataset is divided into a training set according to a certain ratio, and the rest is used as a test set.

[0121] Data preprocessing is performed on the training set and the test set to handle missing values. For missing data, the samples containing missing values are deleted. Then, data normalization is performed to normalize the data to the same scale range to eliminate the influence of the dimension of different features.

[0122] Feature selection is performed, specifically by selecting strip specifications as the input features of the model for subsequent model training and prediction. Here, the strip specifications include the strip width B and the strip thickness H.

[0123] During the model training process, the kernel function f0 of the support vector machine is selected, and then the hyperparameters of the support vector machine are set, namely the penalty coefficient C and the kernel function parameter γ. Furthermore, the support vector machine model is trained using the training set data, including: inputting the input features and corresponding target variables of the training set into the support vector machine model; determining the support vectors and model parameters by optimizing the objective function; and recording important information during the training process, including the number of iterations and the convergence situation.

[0124] After the model is trained, model evaluation is carried out. Specifically, the trained support vector machine model is evaluated using the test set data. The input features of the test set are input into the model to obtain the prediction effect, and the prediction results are compared with the true target variables of the test set and the evaluation indexes are calculated.

[0125] Furthermore, for the steel grades that need to be optimized in production, the decomposition-based multi-objective evolutionary algorithm is used to optimize the relevant rolling parameters and the strip exit speed. The optimization objectives are defined as maximizing the strip exit speed and minimizing the energy consumption during the rolling process, and the corresponding optimization objective function is established. The relevant rolling parameters are used as decision variables to form multi-dimensional decision variables. Among them, the rolling parameters include: the deformation resistance parameters l, m, n, and the friction coefficient μ. These rolling parameters are combined into a four-dimensional decision vector x = [l, m, n, μ]. Then, according to the requirements of the cold tandem rolling production process and equipment limitations, the value ranges of each parameter are determined as the constraint conditions of the decision variables.

[0126] Then, the population is initialized and the individual fitness is evaluated, specifically including: initializing the population, setting the population size as N0, randomly generating 50 initial solutions, each solution corresponding to a set of value combinations of the rolling parameters. These initial solutions are used as the individuals of the population to form the initial population. For each individual in the population, the corresponding rolling parameter values are substituted into the optimization objective function to calculate the fitness value. The fitness values of multiple objective functions are combined into a comprehensive fitness value using the weighted summation method, and the weight coefficients are reasonably set.

[0127] After decomposing the optimization objectives, the evolutionary operation is performed and the population is updated until the evolutionary generation is satisfied or the convergence condition is met. Specifically, first, the optimization objectives are decomposed. Using the decomposition-based multi-objective evolutionary algorithm, the original multi-objective optimization problem is decomposed into 50 single-objective optimization sub-problems. Each sub-problem corresponds to a weight vector, and the value range of the weight vector is [0, 1], and the sum of each weight is 1. For each sub-problem, the following three evolutionary operations are performed, including selection, crossover, and mutation.

[0128] During the selection process, the binary tournament selection algorithm is used to randomly select two individuals from the population, and the individual with the higher fitness value is taken as the parent. During the crossover process, the simulated binary crossover (SBX) algorithm is used to perform crossover operations on the selected parent individuals to generate new offspring individuals. The crossover probability is set to 0.9, and the distribution index is set to 20. During the mutation process, the polynomial mutation algorithm is used to perform mutation operations on the offspring individuals to introduce new gene information. The mutation probability is set to 0.1, and the distribution index is set to 20. The generated offspring individuals are merged with the original population to form a new population.

[0129] For each sub-problem, according to its corresponding weight vector, calculate the fitness value of each individual in the population on this sub-problem; select the individual with the highest fitness value as the optimal solution for this sub-problem and update the population. Obtain the optimal solution set and select a satisfactory solution as the final optimization scheme.

[0130] In the actual tandem cold rolling production, the above algorithm is applied. For strip steels of different specifications, first judge whether production optimization is needed, and then improve the production capacity for the strip steels that need to be optimized. Obtain the strip steel specification information currently being produced from the production control system of the tandem cold rolling mill, including parameters such as steel grade, strip width, and strip thickness; match the obtained strip steel specification information with the previous production history data classified by steel type to determine the steel grade of the currently produced strip steel. According to the established support vector machine model, input the characteristic parameters of the current strip steel specification into the support vector machine model to judge whether the currently produced strip steel needs production optimization. For the strip steels that need production optimization, take the deformation resistance parameters l, m, n, and the friction coefficient μ as the inputs of the optimization problem, and optimize the rolling energy consumption and strip steel exit speed of the current strip steel according to the established multi-objective evolutionary algorithm model based on decomposition; select a solution from the optimization results as the final optimization scheme, and transfer the optimized rolling parameters to the secondary control system of the tandem cold rolling mill.

[0131] By obtaining and classifying the production history data of the tandem cold rolling mill; for steel types of the same grade, screen out the strip steel specification data sets that need production optimization according to the production speed threshold; for the steel types that need production optimization, use the multi-objective evolutionary algorithm based on decomposition to optimize the relevant rolling parameters and the strip steel exit speed; thus achieving the purpose of improving production efficiency and increasing production capacity.

[0132] Based on the rolling parameter optimization method provided in the above embodiments, correspondingly, the present application also provides a specific implementation manner of a rolling parameter optimization device. Please refer to the following embodiments.

[0133] First, refer to Figure 4 , the rolling parameter optimization device 400 provided in the embodiments of the present application includes the following modules:

[0134] An acquisition module 401, configured to acquire strip specification information of the strip currently produced by the tandem cold rolling mill.

[0135] A determination module 402, configured to determine whether the corresponding strip needs production optimization through a preset support vector machine model according to the strip specification information.

[0136] The determination module 402 is further configured to determine the rolling parameter constraint conditions of the strip in the case that the strip needs production optimization.

[0137] An optimization module 403, configured to obtain an optimization result according to the rolling parameter constraint conditions through a preset evolutionary algorithm model, where the optimization result includes optimized rolling parameters.

[0138] As an implementation manner of this application, the rolling parameter optimization device 400 includes:

[0139] A determination module, configured to determine an initial population and a corresponding fitness evaluation criterion according to the rolling parameter constraint conditions and the optimization objective corresponding to the preset evolutionary algorithm model, where the initial population includes multiple initial individuals representing rolling parameter combinations.

[0140] The determination module is further configured to perform an evolution operation on the initial population according to the fitness evaluation criterion through the preset evolutionary algorithm model to determine the optimal individual in the initial population.

[0141] The determination module is further configured to obtain an optimization result according to the optimal individual.

[0142] As an implementation manner of this application, the rolling parameter optimization device 400 includes:

[0143] A determination module, configured to use a preset selection algorithm to select two individuals from the initial population, and determine the individual with a higher fitness value according to the fitness evaluation criterion among the two individuals as the parent individual.

[0144] A crossover module, configured to perform a crossover operation on the parent individuals by using a simulated binary crossover algorithm to generate new offspring individuals.

[0145] A mutation module, configured to perform a mutation operation on the offspring individuals by using a polynomial mutation algorithm to update the initial population.

[0146] A return module, configured to, in the case that the initial population does not meet the preset stop condition, return to the step of using the preset selection algorithm to select two individuals from the initial population and determining the individual with a higher fitness value according to the fitness evaluation criterion among the two individuals as the parent individual.

[0147] The determining module is further configured to determine the optimal individual in the initial population according to the fitness evaluation criterion when the initial population meets the preset stop condition.

[0148] As an implementation manner of the present application, the rolling parameter optimization device 400 includes:

[0149] An obtaining module, configured to obtain the historical production data of the tandem cold rolling mill.

[0150] A constructing module, configured to construct a training data set according to the historical production data, where the training data set includes a plurality of training samples.

[0151] A training module, configured to train an initial support vector machine model by using the training data set to obtain a preset support vector machine model.

[0152] As an implementation manner of the present application, the rolling parameter optimization device 400 includes:

[0153] An obtaining module, configured to obtain the data, steel grade, and corresponding strip physical parameters, strip rolling parameters, and strip inlet and outlet speeds in the historical production of the tandem cold rolling mill.

[0154] A classification module, configured to classify the data according to the steel grade to obtain historical production data.

[0155] As an implementation manner of the present application, the rolling parameter optimization device 400 includes:

[0156] A determining module, configured to determine the low-speed threshold and high-speed threshold for the production of the corresponding strip according to the production conditions of the tandem cold rolling mill.

[0157] The determining module is further configured to filter out the data in the historical production data that meet the low-speed threshold and the high-speed threshold to obtain a first data set.

[0158] The determining module is further configured to delete the data containing missing values in the first data set to obtain a second data set.

[0159] A constructing module, configured to normalize the second data set to construct a training data set.

[0160] As an implementation manner of the present application, the rolling parameter optimization device 400 includes:

[0161] A determining module, configured to determine an initial support vector machine model and a corresponding kernel function.

[0162] A setting module, configured to set the initial support vector machine model and corresponding hyperparameters.

[0163] A training module, configured to train an initial support vector machine model by using each training sample in the above-mentioned training dataset to obtain a preset support vector machine model.

[0164] Each module in the rolling parameter optimization device provided in the embodiments of the present application can implement each step in the above-mentioned rolling parameter optimization method and achieve corresponding effects. For the sake of brevity, details are not described herein again.

[0165] Figure 5 FIG. shows a schematic structural diagram of the rolling parameter optimization hardware provided in the embodiments of the present application.

[0166] The rolling parameter optimization device may include a processor 501 and a memory 502 storing computer program instructions.

[0167] Specifically, the above-mentioned processor 501 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0168] The memory 502 may include a mass storage for data or instructions. By way of example and not limitation, the memory 502 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disc, a magneto-optical disc, a magnetic tape, or a universal serial bus (USB) drive or a combination of two or more of these. In a suitable case, the memory 502 may include a removable or non-removable (or fixed) medium. In a suitable case, the memory 502 may be internal or external to the integrated gateway disaster recovery device. In a specific embodiment, the memory 502 is a non-volatile solid state memory.

[0169] The memory may include a read only memory (ROM), a random access memory (RAM), a magnetic disk storage media device, an optical storage media device, a flash memory device, an electrical, optical, or other physical / tangible memory storage device. Thus, in general, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., a memory device) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the rolling parameter optimization method according to any one of the embodiments of the present disclosure.

[0170] The processor 501 reads and executes the computer program instructions stored in the memory 502 to implement any one of the rolling parameter optimization methods in the above embodiments.

[0171] In one example, the rolling parameter optimization device may further include a communication interface 503 and a bus 510. Among them, as Figure 5 shown, the processor 501, the memory 502, and the communication interface 503 are connected through the bus 510 and complete communication with each other.

[0172] The communication interface 503 is mainly used to implement communication between various modules, devices, units, and / or devices in the embodiments of the present application.

[0173] The bus 510 includes hardware, software, or both, and couples the components of the online data flow metering device to each other. By way of example and not limitation, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a MicroChannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses or a combination of two or more of these. In a suitable case, the bus 510 may include one or more buses. Although the embodiments of the present application describe and illustrate specific buses, the present application contemplates any suitable bus or interconnect.

[0174] In addition, in combination with the method for optimizing rolling parameters in the above embodiments, the embodiments of the present application may provide a computer storage medium to implement. Computer program instructions are stored on the computer storage medium; when the computer program instructions are executed by a processor, any one of the methods for optimizing rolling parameters in the above embodiments is implemented.

[0175] The embodiments of the present application also provide a computer program product, including a computer program, and when the computer program is executed by a processor, any one of the methods for optimizing rolling parameters in the above embodiments is implemented.

[0176] It should be clear that the present application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present application is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications, and additions, or change the order between steps after understanding the spirit of the present application.

[0177] The functional blocks shown in the above-described structural block diagrams can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, it can be, for example, an electronic circuit, an application-specific integrated circuit (ASIC), appropriate firmware, a plug-in, a functional card, and so on. When implemented in software, the elements of the present application are programs or code segments for performing the required tasks. The program or code segment can be stored in a machine-readable medium or transmitted via a data signal carried in a carrier wave over a transmission medium or a communication link. "Machine-readable medium" can include any medium that can store or transmit information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical discs, hard disks, fiber optic media, radio frequency (RF) links, and so on. The code segment can be downloaded via a computer network such as the Internet, an intranet, and so on.

[0178] It should also be noted that in the exemplary embodiments mentioned in the present application, some methods or systems are described based on a series of steps or devices. However, the present application is not limited to the order of the above steps, that is, the steps can be executed in the order mentioned in the embodiments, or can be different from the order in the embodiments, or several steps can be executed simultaneously.

[0179] Aspects of the present disclosure have been described above with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present disclosure. It should be understood that each block in the flowcharts and / or block diagrams, and the combinations of blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device to produce a machine, such that the instructions executed by the processor of the computer or other programmable data processing device enable the implementation of the functions / actions specified in one or more blocks of the flowcharts and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and the combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by dedicated hardware for performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0180] The above are only specific embodiments of the present application. Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, modules, and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein. It should be understood that the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of various equivalent modifications or substitutions, and these modifications or substitutions should be covered within the protection scope of the present application.

Claims

1. A method for optimizing rolling parameters, characterized in that, The method includes: Obtaining the strip specification information of the strip currently produced by the tandem cold rolling mill; According to the strip specification information, determining whether the corresponding strip needs production optimization through a preset support vector machine model; When the strip needs production optimization, determining the rolling parameter constraint conditions of the strip; According to the rolling parameter constraint conditions, obtaining an optimization result through a preset evolutionary algorithm model, where the optimization result includes optimized rolling parameters.

2. The rolling parameter optimization method according to claim 1, characterized in that The step of obtaining an optimization result through a preset evolutionary algorithm model according to the rolling parameter constraint conditions, where the optimization result includes optimized rolling parameters, includes: According to the rolling parameter constraint conditions and the optimization objective corresponding to the preset evolutionary algorithm model, determining an initial population and the corresponding fitness evaluation criteria, where the initial population includes multiple initial individuals representing rolling parameter combinations; Through the preset evolutionary algorithm model, performing an evolutionary operation on the initial population according to the fitness evaluation criteria to determine the optimal individual in the initial population; Obtaining an optimization result according to the optimal individual.

3. The rolling parameter optimization method according to claim 2, characterized in that The step of performing an evolutionary operation on the initial population through the preset evolutionary algorithm model according to the fitness evaluation criteria to determine the optimal individual in the initial population includes: Using a preset selection algorithm to select two individuals from the initial population, and determining the individual with a higher fitness value according to the fitness evaluation criteria among the two individuals as the parent individual; Using a simulated binary crossover algorithm to perform a crossover operation on the parent individuals to generate new offspring individuals; Using a polynomial mutation algorithm to perform a mutation operation on the offspring individuals to update the initial population; When the initial population does not meet the preset stop condition, returning to the step of using a preset selection algorithm to select two individuals from the initial population and determining the individual with a higher fitness value according to the fitness evaluation criteria among the two individuals as the parent individual; When the initial population meets the preset stop condition, determining the optimal individual in the initial population according to the fitness evaluation criteria.

4. The rolling parameter optimization method according to claim 1, characterized in that Before determining whether the corresponding strip needs production optimization through a preset support vector machine model according to the strip specification information, the method further includes: Obtaining the historical production data of the tandem cold rolling mill; According to the historical production data, constructing a training data set, where the training data set includes multiple training samples; Using the training data set to train an initial support vector machine model to obtain a preset support vector machine model.

5. The rolling parameter optimization method according to claim 4, characterized in that, The step of obtaining the historical production data of the tandem cold rolling mill includes: Obtaining the data, steel grade, and corresponding strip physical parameters, strip rolling parameters, and strip inlet and outlet speeds in the historical production of the tandem cold rolling mill; Classifying the data according to the steel grade to obtain historical production data.

6. The rolling parameter optimization method according to claim 4, characterized in that The step of constructing a training data set according to the historical production data includes: According to the production conditions of the tandem cold rolling mill, determining the low-speed threshold and high-speed threshold for the corresponding strip production; Filter out the data in the historical production data that meet the low-speed threshold and the high-speed threshold to obtain a first data set; Delete the data containing missing values in the first data set to obtain a second data set; Normalize the second data set to construct a training data set.

7. The rolling parameter optimization method according to claim 4, characterized in that The training of the initial support vector machine model using the training data set to obtain a preset support vector machine model includes: Determine the initial support vector machine model and the corresponding kernel function; Set the initial support vector machine model and the corresponding hyperparameters; Use each training sample in the training data set to train the initial support vector machine model to obtain a preset support vector machine model.

8. A rolling parameter optimization device, characterized in that, The device includes: An acquisition module for acquiring strip specification information of the currently produced strip of the tandem cold rolling mill; A determination module for determining whether the corresponding strip needs production optimization through a preset support vector machine model according to the strip specification information; The determination module is further configured to determine the rolling parameter constraint conditions of the strip when the strip needs production optimization; An optimization module for obtaining an optimization result according to the rolling parameter constraint conditions through a preset evolutionary algorithm model, where the optimization result includes optimized rolling parameters.

9. A rolling parameter optimization device, characterized in that, The device includes: a processor and a memory storing computer program instructions; the processor reads and executes the computer program instructions to implement the rolling parameter optimization method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, Computer program instructions are stored on the computer-readable storage medium, and when the computer program instructions are executed by a processor, the rolling parameter optimization method according to any one of claims 1-7 is implemented.

11. A computer program product, characterized in that, When the instructions in the computer program product are executed by the processor of the electronic device, the electronic device is caused to execute the rolling parameter optimization method according to any one of claims 1-7.

Images