Cold-rolled strip steel moment arm coefficient prediction method and device based on gene expression programming

Abstract

The embodiment of the invention discloses a cold-rolled strip steel moment arm coefficient prediction method and device based on gene expression programming. The method comprises the steps that equipment parameter sets of multiple sets of target cold rolling equipment and corresponding labeled moment arm coefficients are obtained; and screening at least one equipment parameter related to the force arm coefficient from the equipment parameter set as a related equipment parameter. And parameter fitting is carried out according to the related equipment parameters in the multiple groups of equipment parameter sets and the corresponding labeling force arm coefficients, and candidate prediction coefficients corresponding to each related equipment parameter are obtained. And executing a gene expression programming algorithm by taking the candidate prediction coefficient as an initial population to obtain a force arm coefficient prediction function. And further determining a predicted force arm coefficient according to the force arm coefficient prediction function. The moment arm coefficient prediction function capable of visualizing the relation between the parameters and the moment arm coefficient is accurately determined through the gene expression change algorithm, then accurate moment arm coefficient prediction is achieved through the function, and the prediction result has interpretability.

Description

Technical Field

[0001] This application relates to the field of cold-rolled strip steel, and includes, but is not limited to, a method and apparatus for predicting the lever arm coefficient of cold-rolled strip steel based on gene expression programming. Background Technology

[0002] Rolling, as a core process in the plastic forming of metals, relies heavily on the accurate prediction of its force parameters (rolling force, rolling torque, etc.) for process optimization, equipment design, and production stability. Among these, the lever arm coefficient, a key proportional parameter between rolling torque and rolling force, directly impacts the calculation of rolling energy consumption, roll strength verification, and shape control accuracy. Since the mid-20th century, scholars both domestically and internationally have conducted extensive research on the theoretical modeling and experimental measurement of the lever arm coefficient. However, due to the strong nonlinearity and multi-field coupling characteristics of the rolling process, its prediction accuracy and the interpretability of its mechanism still face significant challenges. Summary of the Invention

[0003] In view of this, the method and apparatus for predicting the lever arm coefficient of cold-rolled strip steel based on gene expression programming provided in the embodiments of this application can accurately predict the lever arm coefficient and ensure the interpretability of the lever arm coefficient prediction process.

[0004] The method and apparatus for predicting the lever arm coefficient of cold-rolled strip steel based on gene expression programming provided in this application are implemented as follows: One aspect of this application provides a method for predicting the lever arm coefficient of cold-rolled strip steel based on gene expression programming, the method comprising: Obtain a set of equipment parameters for multiple target cold rolling equipment, each set of equipment parameters having a corresponding labeled lever arm coefficient; Select at least one equipment parameter related to the lever arm coefficient from the set of equipment parameters as the relevant equipment parameter; Based on the relevant equipment parameters and corresponding labeled lever arm coefficients in multiple sets of equipment parameters, parameter fitting is performed to obtain the candidate prediction coefficients for each relevant equipment parameter; The gene expression programming algorithm is executed with at least one candidate prediction coefficient as the initial population to obtain the lever arm coefficient prediction function with at least one candidate prediction coefficient as the variable. When it is necessary to predict the lever arm coefficient of the target cold rolling equipment, the predicted lever arm coefficient is determined based on the lever arm coefficient prediction function.

[0005] In one possible implementation, a set of equipment parameters for multiple target cold rolling mills is obtained, including: Obtain multiple sets of candidate parameters for target cold rolling equipment, each set of candidate parameters including the corresponding labeled lever arm coefficient; Perform data normalization on each candidate parameter set; The candidate parameter sets are filtered based on the labeled lever arm coefficients corresponding to each candidate parameter set to obtain the equipment parameter set.

[0006] In one possible implementation, the candidate parameter set is filtered based on the labeled lever arm coefficient corresponding to each candidate parameter set to obtain the equipment parameter set, including: Calculate the mean and standard deviation of the labeled lever arm coefficients for each candidate parameter set; Calculate the difference between the labeled lever arm coefficient and the average value for each candidate parameter set; If the absolute value of the coefficient difference is greater than three times the standard deviation, the candidate parameter set is deleted, and the candidate parameter set that passes the screening is determined as the equipment parameter set.

[0007] In one possible implementation, at least one equipment parameter related to the lever arm coefficient is selected from the set of equipment parameters as the relevant equipment parameter, including: The correlation values ​​between each equipment parameter and the lever arm coefficient in the equipment parameter set are calculated using the maximum mutual information coefficient method. Device parameters whose relevant values ​​are greater than the relevant threshold are identified as relevant device parameters.

[0008] In one possible implementation, parameter fitting is performed based on relevant equipment parameters and corresponding labeled lever arm coefficients from multiple sets of equipment parameter sets to obtain candidate prediction coefficients for each relevant equipment parameter, including: Multiple regression analysis is performed on the relevant equipment parameters and corresponding labeled lever arm coefficients in multiple sets of equipment parameters to obtain the multiple linear regression equation of lever arm coefficients and relevant equipment parameters. The multiple linear regression equation includes a constant term and an unknown term composed of each labeled lever arm coefficient and coefficient. Each unknown term is identified as a candidate prediction coefficient.

[0009] In one possible implementation, a gene expression programming algorithm is executed using at least one candidate prediction coefficient as the initial population to obtain a lever arm coefficient prediction function with at least one candidate prediction coefficient as a variable, including: At least one candidate prediction coefficient is used as the initial population to perform multiple population evolutions using the gene expression programming algorithm to obtain the corresponding candidate coefficient prediction function. After each population evolution, the candidate lever arm coefficient is calculated using the currently determined candidate coefficient prediction function, and the mean absolute percentage error is calculated based on the candidate lever arm coefficient and the labeled lever arm coefficient to obtain the loss value. If the loss value meets the preset constraints, the current candidate coefficient prediction function is determined as the lever arm coefficient prediction function.

[0010] In one possible implementation, when it is necessary to predict the lever arm coefficient of the target cold rolling mill, the predicted lever arm coefficient is determined based on the lever arm coefficient prediction function, including: When it is necessary to predict the lever arm coefficient of the current target cold rolling equipment, obtain the current relevant equipment parameters of the target cold rolling equipment; The relevant equipment parameters obtained at the moment are substituted into the lever arm coefficient prediction function to calculate the corresponding predicted lever arm coefficient.

[0011] Another aspect of this application provides a device for predicting the lever arm coefficient of cold-rolled strip steel based on gene expression programming. The device includes: The parameter acquisition module is used to acquire a set of equipment parameters for multiple target cold rolling equipment. The set of equipment parameters has a corresponding labeled lever arm coefficient. The parameter filtering module is used to filter at least one equipment parameter related to the lever arm coefficient from the equipment parameter set as the relevant equipment parameter; The coefficient determination module is used to perform parameter fitting based on the relevant equipment parameters and the corresponding labeled lever arm coefficients in multiple sets of equipment parameter sets, and to obtain the candidate prediction coefficients for each relevant equipment parameter. The function determination module is used to execute a gene expression programming algorithm with at least one candidate prediction coefficient as the initial population to obtain a lever arm coefficient prediction function with at least one candidate prediction coefficient as the variable. The coefficient prediction module is used to determine the predicted lever arm coefficient based on the lever arm coefficient prediction function when it is necessary to predict the lever arm coefficient of the target cold rolling equipment.

[0012] In one possible implementation, the parameter acquisition module is further used for: Obtain multiple sets of candidate parameters for target cold rolling equipment, each set of candidate parameters including the corresponding labeled lever arm coefficient; Perform data normalization on each candidate parameter set; The candidate parameter sets are filtered based on the labeled lever arm coefficients corresponding to each candidate parameter set to obtain the equipment parameter set.

[0013] In one possible implementation, the parameter acquisition module is further used for: Calculate the mean and standard deviation of the labeled lever arm coefficients for each candidate parameter set; Calculate the difference between the labeled lever arm coefficient and the average value for each candidate parameter set; If the absolute value of the coefficient difference is greater than three times the standard deviation, the candidate parameter set is deleted, and the candidate parameter set that passes the screening is determined as the equipment parameter set.

[0014] In one possible implementation, the parameter filtering module is further used for: The correlation values ​​between each equipment parameter and the lever arm coefficient in the equipment parameter set are calculated using the maximum mutual information coefficient method. Device parameters whose relevant values ​​are greater than the relevant threshold are identified as relevant device parameters.

[0015] In one possible implementation, the coefficient determination module is further used for: Multiple regression analysis is performed on the relevant equipment parameters and corresponding labeled lever arm coefficients in multiple sets of equipment parameters to obtain the multiple linear regression equation of lever arm coefficients and relevant equipment parameters. The multiple linear regression equation includes a constant term and an unknown term composed of each labeled lever arm coefficient and coefficient. Each unknown term is identified as a candidate prediction coefficient.

[0016] In one possible implementation, the function-determining module is further used for: At least one candidate prediction coefficient is used as the initial population to perform multiple population evolutions using the gene expression programming algorithm to obtain the corresponding candidate coefficient prediction function. After each population evolution, the candidate lever arm coefficient is calculated using the currently determined candidate coefficient prediction function, and the mean absolute percentage error is calculated based on the candidate lever arm coefficient and the labeled lever arm coefficient to obtain the loss value. If the loss value meets the preset constraints, the current candidate coefficient prediction function is determined as the lever arm coefficient prediction function.

[0017] In one possible implementation, the coefficient prediction module is further used for: When it is necessary to predict the lever arm coefficient of the target cold rolling equipment, obtain the current relevant equipment parameters of the target cold rolling equipment; The relevant equipment parameters obtained at the moment are substituted into the lever arm coefficient prediction function to calculate the corresponding predicted lever arm coefficient.

[0018] The electronic device provided in this application includes a memory and a processor. The memory stores a computer program that can run on the processor. When the processor executes the program, it implements the method described in this application.

[0019] The computer-readable storage medium provided in this application embodiment stores a computer program thereon, which, when executed by a processor, implements the method provided in this application embodiment.

[0020] In this embodiment, the method obtains multiple sets of equipment parameters for target cold rolling equipment and corresponding labeled lever arm coefficients. At least one equipment parameter related to the lever arm coefficient is selected from the equipment parameter sets as a relevant equipment parameter. Parameter fitting is performed based on the relevant equipment parameters and corresponding labeled lever arm coefficients in the multiple sets of equipment parameters to obtain candidate prediction coefficients for each relevant equipment parameter. The candidate prediction coefficients are used as the initial population to execute a gene expression programming algorithm to obtain a lever arm coefficient prediction function. The predicted lever arm coefficient is further determined based on the lever arm coefficient prediction function. This application accurately determines a lever arm coefficient prediction function that can visualize the relationship between parameters and lever arm coefficients through a gene expression programming algorithm, thereby achieving accurate lever arm coefficient prediction through this function, and the prediction results are interpretable. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 A flowchart illustrating a method for predicting the lever arm coefficient of cold-rolled strip steel based on gene expression programming according to an embodiment of this application is shown. Figure 2 A schematic diagram of a target cold rolling apparatus according to an embodiment of this application is shown; Figure 3 This diagram illustrates a device parameter corresponding to a relevant value according to an embodiment of this application. Figure 4 A flowchart illustrating an algorithm for executing gene expression programming according to an embodiment of this application is shown; Figure 5 This diagram illustrates a gene decoding process using a gene expression programming algorithm according to an embodiment of this application. Figure 6 A schematic diagram illustrating the effect of lever arm coefficient prediction according to an embodiment of this application is shown. Figure 7 A flowchart illustrating another lever arm coefficient prediction effect according to an embodiment of this application; Figure 8 A schematic diagram of a cold-rolled strip lever arm coefficient prediction device based on gene expression programming according to an embodiment of this application is shown. Figure 9 A schematic diagram of an electronic device according to an embodiment of this application is shown. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the specific technical solutions of this application will be further described in detail below with reference to the accompanying drawings of the embodiments of this application. The following embodiments are used to illustrate this application, but are not intended to limit the scope of this application.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0025] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is 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.

[0026] It should be noted that the terms "first, second, third" used in the embodiments of this application are used to distinguish similar or different objects and do not represent a specific order of objects. It can be understood that "first, second, third" can be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0027] The method for predicting the lever arm coefficient of cold-rolled strip steel based on gene expression programming in this application can be executed by any electronic device, including but not limited to mobile phones, wearable devices (such as smartwatches, smart bracelets, smart glasses, etc.), tablet computers, laptops, vehicle terminals, PCs (Personal Computers), etc. The functions implemented by this method can be achieved by a processor in the electronic device calling program code. Of course, the program code can be stored in a computer storage medium. Therefore, the electronic device includes at least a processor and a storage medium.

[0028] The cold-rolled strip lever arm coefficient prediction method based on gene expression programming in this application can be used for any application scenario requiring prediction of the cold-rolled strip lever arm coefficient. For example, this application can be applied to application scenarios involving overload protection and strip breakage prevention for target cold rolling equipment. Alternatively, it can be applied to application scenarios involving computer-controlled settings for the strip production process of target cold rolling equipment.

[0029] In the above application scenarios, the relevant technologies typically employ the following methods to predict the lever arm coefficient: Early related studies established a theoretical framework for the lever arm coefficient based on the plane section assumption and slip line field theory. Tselikov first proposed an empirical relationship between the contact arc projection length and the lever arm coefficient, laying the foundation for subsequent research. Ford and Sims derived an expression for the lever arm coefficient considering friction effects by analyzing the balance equation of rolling force and torque, but their model did not consider the influence of temperature on the material's deformation resistance. Siebel introduced a correction formula for the geometric factor of the deformation zone and found that the width-to-thickness ratio has a significant regulatory effect on the lever arm coefficient. Nikitin et al., through rolling experiments on 15 kinds of heat-resistant steels, first proposed a nonlinear relationship between the lever arm coefficient and the reduction rate, finding that it fluctuates in the range of 0.18-0.3. Pietrzyk established an exponential fitting model of the lever arm coefficient and the contact arc drop using measured data from a hot strip mill, but its applicability is limited to specific steel grades. Klarin and Lundberg, through computer-aided slip line field analysis, revealed the differential influence of front and rear tensions on the lever arm coefficient and proposed a calculation method based on stress state partitioning. Moo et al., combining finite element simulation and rolling experiments, verified the synergistic effect of the contact arc length and average thickness ratio on the lever arm coefficient, and constructed a corrected model considering the elastic flattening of the rolls. Research in the field of cold rolling is relatively lagging. Safia, through cold rolling experiments on low-carbon strip steel, measured the average lever arm coefficient to be between 0.19 and 0.24, indicating that the influence of roll elastic deformation on the lever arm coefficient in cold rolling is significantly higher than that in hot rolling. Zhao Zhiye analyzed the control mechanism of cold rolling reduction rate and tension distribution on the lever arm coefficient, and proposed an empirical formula based on the geometric parameters of the deformation zone. Sun J et al., based on a three-dimensional elastoplastic finite element model, revealed the coupling law between the friction coefficient and the front and rear tensions during cold rolling, but their model calculation error was still as high as 8%.

[0030] With the development of computational mechanics, the finite element method (FEM) and multi-field coupling technology have provided a new paradigm for the study of lever arm coefficient. Jiao Zhijie studied the influence of the geometric conditions of the rolling deformation zone on the lever arm coefficient, determining the shape factor and reduction rate of the deformation zone as key influencing parameters. Through regression analysis of actual data, the optimal model form and model coefficients were obtained, and the average error of the rolling torque calculated using the improved lever arm coefficient model was within 5%. Zhang et al. proposed an analytical model of rolling force based on a cosine velocity field, improving the accuracy of contact stress calculation by introducing an equal area criterion, indirectly optimizing the prediction of the lever arm coefficient. Jie and Liu used a hyperbolic sinusoidal velocity field to analyze the cold continuous rolling process, constructing a correlation model between the lever arm coefficient and the strain rate gradient. At the microscopic mechanism level, molecular dynamics simulations (MD) have begun to reveal the influence of interfacial dislocation evolution on the lever arm coefficient. For example, simulations of FeCrNi / Fe bimetallic rolling show that the shearing effect of interfacial atoms can reduce the matrix yield strength but inhibit dislocation propagation, leading to an increase in the lever arm coefficient.

[0031] Currently, most constructed lever arm coefficient models are theoretical mathematical models. While these models can intuitively demonstrate the influence of various variables on the lever arm coefficient and facilitate understanding of the mechanisms of action of each parameter, they often involve simplifications and assumptions during their construction, leading to reduced model accuracy. In recent years, the development of machine learning has provided a new reference for shifting lever arm coefficient prediction from mechanism-driven to data-driven approaches. Zhang et al. used LSTM networks to predict the rolling force of hot-rolled strip steel, combining time-series data characteristics, achieving a 30% reduction in error compared to traditional models. Zhang Shunhu et al. utilized error spacing compensation to organically integrate neural network models with existing theoretical models, ultimately obtaining an integrated model of rolling force with a model determination coefficient of 0.97. Chen et al. combined XGBoost and Bayesian optimization to achieve collaborative prediction of rolling force in multi-stand cold continuous rolling mills, improving computational efficiency by 50%. Klarin et al. used SVR to predict the rolling torque of hot-rolled wide and thick plates, reducing the error by 40% compared to linear regression. The advantage of intelligent algorithms lies in their ability to handle complex nonlinear relationships, typically outperforming theoretical models in prediction accuracy. However, they also have drawbacks such as opaque internal working mechanisms and difficulty in explaining the influence of individual parameters on the lever arm coefficient.

[0032] Therefore, the technical problem solved by the embodiments of this application is how to accurately, reliably and interpretably predict the lever arm coefficient of cold-rolled strip steel.

[0033] The following section, in conjunction with the accompanying drawings, describes the prediction of the lever arm coefficient for cold-rolled strip steel according to embodiments of this application. The plan is explained in detail.

[0034] Figure 1 A flowchart illustrating a method for predicting the lever arm coefficient of cold-rolled strip steel based on gene expression programming, according to an embodiment of this application, is shown. Figure 1 As shown, the cold-rolled strip lever arm coefficient prediction method based on gene expression programming in this application embodiment may include the following steps S10-S50.

[0035] For ease of description, the method for predicting the lever arm coefficient of cold-rolled strip steel based on gene expression programming in this application embodiment is described using an electronic device as the execution subject. It should be understood that the execution subject in this application embodiment can also be a processor or chip in an electronic device, and this application embodiment does not impose any limitations.

[0036] Step S10: Obtain a set of equipment parameters for multiple target cold rolling equipment.

[0037] In one possible implementation, electronic equipment acquires a set of equipment parameters for a target cold rolling mill to predict the lever arm of cold-rolled strip. The target cold rolling mill, used for producing strip, includes multiple stands for rolling strip. In this embodiment, multiple sensors for acquiring different equipment parameters can be installed on the target cold rolling mill to collect various parameters during each operation of the target cold rolling mill to determine a corresponding set of equipment parameters. Optionally, the set of equipment parameters has a corresponding labeled lever arm coefficient, used to characterize the lever arm coefficient of the strip produced by the target cold rolling mill when this set of equipment parameters is acquired.

[0038] Figure 2 A schematic diagram of a target cold rolling apparatus according to an embodiment of this application is shown. Figure 2 As shown, Figure 2 This is a target cold rolling mill comprising five stands. Sensors such as thickness gauges, laser velocimeters, and tension meters can be installed between each stand. During each operation, these sensors measure a preset number (e.g., 13) of equipment parameters across the five stands, including speed, rolling force, tension between stands, work roll bending force, and reduction. The set of equipment parameters can then be determined based on these preset number of parameters acquired each time. Simultaneously, the electronic equipment can use the cold-rolled strip lever arm coefficient from each operation as the corresponding standard lever arm coefficient for each equipment parameter.

[0039] In some embodiments, the electronic device can directly determine the set of device parameters based on the multiple device parameters acquired each time, or it can first treat the multiple acquired device parameters as a set of candidate parameters and determine that each set of candidate parameters includes the corresponding labeled lever arm coefficient. Then, data normalization processing is performed on each set of candidate parameters, and the set of candidate parameters is filtered according to the labeled lever arm coefficient corresponding to each set of candidate parameters to obtain the set of candidate parameters that passes the filtering as the set of device parameters.

[0040] For example, data normalization is used to eliminate the influence of dimensional differences between multidimensional equipment parameters in the measured data of the rolling production process on the modeling. By performing a linear transformation on the original data, the equipment parameter values ​​are converted into dimensionless values, so that they fall within the [0,1] interval after transformation. The calculation formula can be... In the formula, For normalized data, These are the original target device parameter values. This represents the minimum value of the target device parameter among different candidate parameter sets. This represents the maximum value of the target device parameter among different candidate parameter sets. Electronic devices can normalize each dimension of the device parameter as a target device parameter.

[0041] In other embodiments, during actual steel rolling production, due to factors such as equipment detection errors, environmental noise interference, abnormal operating condition fluctuations, and human operational deviations, the collected industrial data samples often contain significant outliers. The presence of outliers can adversely affect the accuracy of the lever arm coefficient prediction function obtained from modeling. To eliminate the negative impact of outliers on the accuracy of the lever arm coefficient prediction function, electronic equipment can identify outlier data in the candidate parameter set through the corresponding labeled lever arm coefficient, and filter out the normal equipment parameter set. This filtering method involves calculating the average and standard deviation of the labeled lever arm coefficients corresponding to each candidate parameter set, and calculating the coefficient difference between the labeled lever arm coefficient and the average value for each candidate parameter set. Candidate parameter sets are deleted if the absolute value of the coefficient difference is greater than three times the standard deviation, and the filtered candidate parameter set is determined as the equipment parameter set.

[0042] For example, electronic devices can be implemented using formulas Identify anomalous labeled lever arm coefficients and determine the candidate parameter set corresponding to these anomalous labeled lever arm coefficients as anomalous data. Among these, , a i Let be the i-th labeled lever arm coefficient, ā and σ be the mean and standard deviation of all labeled lever arm coefficients, respectively, and n be the sample size.

[0043] Through experiments, after acquiring 336 sets of candidate parameters from the electronic device, 16 sets of outliers were removed using the aforementioned outlier removal method. The remaining 320 sets of candidate parameters were used as the device parameter set for establishing the lever arm coefficient prediction function model in the later stages. Optionally, the composition of this device parameter set can be represented by the following table:

[0044] Step S20: Select at least one equipment parameter related to the lever arm coefficient from the set of equipment parameters as the relevant equipment parameter.

[0045] In one possible implementation, the set of device parameters determined by the electronic device may contain too many types of device parameters. To ensure the scientific validity of the independent variable selection in the lever arm coefficient prediction function obtained from mathematical modeling, this embodiment of the application can screen at least one relevant device parameter related to the lever arm coefficient from the set of device parameters. Optionally, the screening of relevant device parameters can be performed using the maximum mutual information coefficient (MIC) method to verify the correlation between the device parameters and the lever arm coefficient. The advantage of this method is that it can analyze linear or nonlinear functional relationships between variables, and it has low complexity and high robustness. That is, the electronic device can calculate the correlation value between each device parameter in the set of device parameters and the lever arm coefficient using the maximum mutual information coefficient method, and then determine the device parameters whose correlation value is greater than the correlation threshold as relevant device parameters.

[0046] In some embodiments, the specific process for MIC verification of relevance is as follows: The mutual information between any device parameter X and lever arm coefficient Y can be defined as:

[0047] Where p(x,y) is the joint probability distribution function between X and Y; p(x) and p(y) are the marginal probability distribution functions of X and Y, respectively. A grid is plotted on the scatter plot of the data composed of X and Y, and the mutual information between each grid cell is calculated. The maximum mutual information is selected using different grid division criteria, and the calculation formula is as follows:

[0048] Wherein, MIC(X,Y) represents the maximum mutual information coefficient between X and Y. The MIC value increases as the correlation between X and Y strengthens. When X and Y are independent, the MIC value is 0. a and b are the number of grids divided in the x and y directions, respectively, and B is the maximum grid size, typically taken as 0.6 times the sample size.

[0049] Optionally, the electronic device can repeat the above steps until the MIC value I(X;Y) between all device parameters and lever arm coefficients is obtained as the relevant value.

[0050] Figure 3 This diagram illustrates a device parameter corresponding to a relevant value according to an embodiment of this application. For example... Figure 3 As shown, the correlation value MIC between each equipment parameter and the lever arm coefficient is obtained during a calculation. The electronic device can have a pre-set correlation threshold of 0.7. When MIC > 0.7, it is considered that there is a strong correlation between the equipment parameter and the lever arm coefficient, and the equipment parameter can be identified as a relevant equipment parameter for the lever arm coefficient. According to the figure, after filtering according to the above rules, the final relevant equipment parameters of the electronic device are inlet thickness, outlet thickness, pre-tension stress, post-tension stress, and flattening radius.

[0051] The relevant equipment parameters in some of the equipment parameter sets can be shown in the table below:

[0052] Step S30: Perform parameter fitting based on the relevant equipment parameters and corresponding labeled lever arm coefficients in multiple sets of equipment parameters to obtain the candidate prediction coefficients corresponding to each relevant equipment parameter.

[0053] In one possible implementation, after determining at least one relevant device parameter, the electronic device can perform parameter fitting based on the relevant device parameters and corresponding labeled lever arm coefficients from multiple sets of acquired device parameter sets to obtain candidate prediction coefficients for each relevant device parameter. These candidate prediction coefficients are used in a gene expression programming algorithm to model and obtain a lever arm coefficient prediction function.

[0054] Optionally, the electronic device can first perform multiple regression analysis based on the relevant equipment parameters and corresponding labeled lever arm coefficients from multiple sets of equipment parameter sets, obtaining a multiple linear regression equation between the lever arm coefficients and the relevant equipment parameters. The multiple linear regression equation includes a constant term and an unknown term consisting of each labeled lever arm coefficient and its coefficient. Then, each unknown term is determined as a candidate predictive coefficient.

[0055] For example, an electronic device can perform multiple regression analysis by inputting relevant device parameters and corresponding labeled lever arm coefficients from multiple sets of device parameter sets into a preset application, thereby obtaining a multiple linear regression equation between the lever arm coefficients and their corresponding relevant device parameters. To further improve prediction accuracy, the electronic equipment multiplies the coefficient before each relevant device parameter in the multiple linear regression equation with the corresponding relevant device parameter to form new candidate prediction coefficients. , used in gene expression programming algorithms to model and obtain the lever arm coefficient prediction function. "?" represents a constant randomly selected from the set of constants, denoted by the additional field Dc.

[0056] The candidate prediction coefficients can be used as the lever arm coefficients in the gene expression programming algorithm, as shown in the table below.

[0057]

[0058] Step S40: Using at least one of the candidate prediction coefficients as the initial population, execute the gene expression programming algorithm to obtain a lever arm coefficient prediction function with at least one of the candidate prediction coefficients as variables.

[0059] In one possible implementation, after determining at least one candidate prediction coefficient, the electronic device can use this candidate prediction coefficient as an initial population to execute a gene expression programming algorithm, thereby obtaining a lever arm coefficient prediction function with the at least one candidate prediction coefficient as a variable. That is, the at least one candidate prediction coefficient can be used as the initial population to perform multiple population evolutions using the gene expression programming algorithm to obtain the corresponding candidate coefficient prediction function. After each population evolution, the currently determined candidate coefficient prediction function calculates the candidate lever arm coefficient, and calculates the mean absolute percentage error based on the candidate lever arm coefficient and the labeled lever arm coefficient to obtain the loss value. If the loss value meets preset constraints, the current candidate coefficient prediction function is determined as the lever arm coefficient prediction function.

[0060] In some embodiments, Gene Expression Programming (GEP) is an algorithm proposed by Ferreiral in 1999 that draws inspiration from biological gene expression. This algorithm, while inheriting the dynamic chromosome length characteristic of Genetic Algorithms (GA), integrates the tree-like hierarchical structure of Genetic Programming (GP), achieving efficient modeling of complex problems through a decoupling mechanism between gene encoding and phenotype. In the GEP framework, each chromosome consists of a fixed-length linear symbol sequence. Genetic information is hierarchically encoded into a head and tail through a multi-gene structure. This design ensures both the efficiency of genetic operations and the ability to express complex functional relationships. These chromosomes are continuously iterated and updated through genetic methods such as inheritance, mutation, recombination, and insertion until the optimal fitness is reached, at which point the optimal chromosome represents the optimal solution to the problem.

[0061] Figure 4 A flowchart illustrating an algorithm for executing gene expression programming according to an embodiment of this application is shown. Figure 4 As shown, the electronic device can use at least one candidate prediction coefficient as input population during the mathematical modeling process, and iteratively perform decoding, fitness evaluation, optimal individual selection, and new population generation to evolve the population. When the loss value meets preset constraints, the evolution termination condition is determined, and the current candidate coefficient prediction function is taken as the lever arm coefficient prediction function.

[0062] Each chromosome consists of one or more genes. A gene has a head and a tail. The symbol for the head can come from either a function symbol set or a terminal symbol set, while the symbol for the tail can only come from the terminal symbol set. The length of the gene head generally depends on the problem, while the length of the tail is determined by the formula... The calculation yielded the following result. This represents the number of parameters in the function that requires the most variables. For gene head length, This represents the length of the gene tail. Genes are represented as expression trees, constructed from genotypes according to grammatical rules and hierarchical order. Genotypes arrange the variables of the expression tree from top to bottom and left to right. Each gene defines one expression tree, and a chromosome is composed of multiple expression trees interconnected by function symbols. For example, suppose the set of function symbols is... (q represents the SQRT function), the terminal symbol set is {x, y}, and we take h=11, n=2, then Therefore, the total gene length is 11 + 12 = 23.

[0063] Figure 5 This diagram illustrates gene decoding using a gene expression programming algorithm according to an embodiment of this application. Figure 5 As shown, the starting position (-) of the gene corresponds to the root of the expression tree, and each subsequent function operator... For each function, one or more parameters are represented by a leaf node or another subtree, corresponding to a subtree. The specific decoding rules include: the first character in the gene corresponds to the root node of the expression tree, i.e., the unique node on the first level of the tree. The number of subtrees (branches) corresponding to the next level below each function operator is the same as the number of parameters the function operator has. The branch nodes corresponding to the function nodes are filled in sequentially from left to right by the successor elements in the gene. The above steps are repeated until a new level containing only leaf nodes is formed. The resulting expression tree is then combined from bottom to top, combining leaf nodes with function nodes sequentially until the root node is reached.

[0064] Optionally, during the process of modeling the lever arm coefficient prediction function by performing the GEP algorithm on at least one candidate prediction coefficient as input population, the required parameters are shown in the table below:

[0065] Where Q represents S represents C represents E represents P represents L represents F represents G represents .

[0066] Optionally, in this embodiment, the electronic device can calculate the candidate coefficient prediction function loss obtained from population evolution by determining it through a fitness function. That is, the electronic device uses the fitness function as a quality evaluation standard for individual solutions, and its design directly affects the algorithm's convergence efficiency. This embodiment addresses the need for explicit functional relationship modeling of the lever arm coefficient by constructing the fitness function as a global error quantification index between predicted and experimental observations. After comparing various error functions such as mean square error and relative error, the robust mean absolute percentage error (MAPE) is ultimately adopted to construct the fitness function, and its mathematical expression is: ,in For fitness value, For the sample size, For the first The sample lever arm coefficient for each sample; The first candidate coefficient prediction function One predictive lever arm coefficient.

[0067] In some embodiments, the electronic device can divide the set of 320 measured device parameters into a training set and a test set in a 7:3 ratio. Based on the GEP algorithm parameter settings described above, a GEP prediction model for the lever arm coefficient is established using MATLAB. After repeated training, the optimal chromosome lever arm coefficients corresponding to the models are obtained respectively.

[0068] Using the decoding concept of GEP chromosome genes, the electronic device decodes the chromosome to obtain the lever arm coefficient prediction function corresponding to the optimal chromosome in the lever arm coefficient model:

[0069] After obtaining the lever arm coefficient prediction function, the electronic device can substitute the actual relevant equipment parameters corresponding to the previously acquired test set into the above function expression to calculate the corresponding predicted lever arm coefficient. The expanded result of the above formula is as follows:

[0070] Figure 6 A schematic diagram illustrating the effect of lever arm coefficient prediction according to an embodiment of this application is shown. Figure 6 As shown, after the electronic device predicts the predicted lever arm coefficient using the lever arm coefficient prediction function determined above, it determines the correlation between the predicted lever arm coefficient and the actual labeled lever arm coefficient. The mean absolute percentage error (MAPE) of the lever arm coefficient prediction function is 4.31%, and the coefficient of determination (R²) is 0.9355.

[0071] Figure 7 A flowchart illustrating another lever arm coefficient prediction effect according to an embodiment of this application is shown. Figure 7 As shown, after the electronic device predicts the predicted lever arm coefficients using the aforementioned lever arm coefficient prediction function, it determines the statistical error between the predicted lever arm coefficients and the actual labeled lever arm coefficients. In the prediction results statistics, 91.2% of the lever arm coefficient prediction results have a relative error of less than 10%. Among them, 67.8% of the lever arm coefficient prediction results have a relative error of less than 5%.

[0072] according to Figure 6 and Figure 7 The results show that the lever arm coefficient prediction model determined by the GEP model can achieve stable and accurate prediction of the lever arm coefficient, and has a certain generalization ability.

[0073] Furthermore, this application embodiment also compares the effects of mainstream lever arm coefficient prediction methods with the lever arm coefficient prediction method of this application embodiment. The error comparison of the lever arm coefficient prediction model is shown in the table below:

[0074] As shown in the table above, for the lever arm coefficient, the GEP model's mean absolute percentage error is 1.48% lower than that of the BP neural network, 3.55% lower than that of the SVR support vector regression, and 0.29% lower than that of the RF random forest. Overall, the prediction accuracy and generalization ability of the GEP model are roughly equivalent to mainstream "black box" prediction models. However, there is currently no definitive optimal method for the number of hidden layers and nodes in the BP neural network model, making it difficult to find the optimal parameter combination. Meanwhile, the SVR support vector regression and RF random forest models are prone to overfitting when dealing with noisy datasets. Furthermore, these three machine learning algorithms are all "black box" models, unable to further control the model's intrinsic parameters, thus failing to clearly describe the relationship between output and input variables. In contrast, the GEP prediction model can obtain a definite expression for the function, facilitating calculation and providing a new method for lever arm coefficient prediction.

[0075] Step S50: When it is necessary to predict the lever arm coefficient of the target cold rolling equipment, determine the predicted lever arm coefficient according to the lever arm coefficient prediction function.

[0076] In some possible implementations, after acquiring the lever arm coefficient prediction function, the electronic device can predict the lever arm coefficient of the strip corresponding to the target cold rolling mill in real time based on this function. Specifically, when it is necessary to predict the lever arm coefficient of the target cold rolling mill, the electronic device can acquire the current relevant equipment parameters of the target cold rolling mill through sensors. Then, the acquired relevant equipment parameters are input into the lever arm coefficient prediction function to calculate the corresponding predicted lever arm coefficient.

[0077] Based on the above technical features, the embodiments of this application accurately determine the lever arm coefficient prediction function that can visualize the relationship between parameters and lever arm coefficient through gene expression transformation algorithm, and then achieve accurate lever arm coefficient prediction through the function, and the prediction result is interpretable.

[0078] It should be understood that although the steps in the above flowcharts are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the above flowcharts may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0079] Based on the foregoing embodiments, this application provides a cold-rolled strip steel lever arm coefficient prediction device based on gene expression programming. The device includes various modules and units included in each module, which can be implemented by a processor; of course, it can also be implemented by specific logic circuits. In the implementation process, the processor can be a central processing unit (CPU), microprocessor (MPU), digital signal processor (DSP), or field programmable gate array (FPGA), etc.

[0080] Figure 8 This diagram illustrates a device for predicting the lever arm coefficient of cold-rolled strip steel based on gene expression programming, according to an embodiment of this application. Figure 8 As shown, the cold-rolled strip steel lever arm coefficient prediction device based on gene expression programming in this application embodiment includes: The parameter acquisition module 80 is used to acquire multiple sets of equipment parameters for target cold rolling equipment. The equipment parameter sets have corresponding labeled lever arm coefficients. The parameter filtering module 81 is used to filter at least one equipment parameter related to the lever arm coefficient from the equipment parameter set as the relevant equipment parameter; The coefficient determination module 82 is used to perform parameter fitting based on the relevant equipment parameters and the corresponding labeled lever arm coefficients in multiple sets of equipment parameter sets, and to obtain the candidate prediction coefficients corresponding to each relevant equipment parameter. The function determination module 83 is used to execute a gene expression programming algorithm with at least one candidate prediction coefficient as the initial population to obtain a lever arm coefficient prediction function with at least one candidate prediction coefficient as the variable. The coefficient prediction module 84 is used to determine the predicted lever arm coefficient based on the lever arm coefficient prediction function when it is necessary to predict the lever arm coefficient of the target cold rolling equipment.

[0081] In one possible implementation, the parameter acquisition module 80 is further used for: Obtain multiple sets of candidate parameters for target cold rolling equipment, each set of candidate parameters including the corresponding labeled lever arm coefficient; Perform data normalization on each candidate parameter set; The candidate parameter sets are filtered based on the labeled lever arm coefficients corresponding to each candidate parameter set to obtain the equipment parameter set.

[0082] In one possible implementation, the parameter acquisition module 80 is further used for: Calculate the mean and standard deviation of the labeled lever arm coefficients for each candidate parameter set; Calculate the difference between the labeled lever arm coefficient and the average value for each candidate parameter set; If the absolute value of the coefficient difference is greater than three times the standard deviation, the candidate parameter set is deleted, and the candidate parameter set that passes the screening is determined as the equipment parameter set.

[0083] In one possible implementation, the parameter filtering module 81 is further used for: The correlation values ​​between each equipment parameter and the lever arm coefficient in the equipment parameter set are calculated using the maximum mutual information coefficient method. Device parameters whose relevant values ​​are greater than the relevant threshold are identified as relevant device parameters.

[0084] In one possible implementation, the coefficient determination module 82 is further used for: Multiple regression analysis is performed on the relevant equipment parameters and corresponding labeled lever arm coefficients in multiple sets of equipment parameters to obtain the multiple linear regression equation of lever arm coefficients and relevant equipment parameters. The multiple linear regression equation includes a constant term and an unknown term composed of each labeled lever arm coefficient and coefficient. Each unknown term is identified as a candidate prediction coefficient.

[0085] In one possible implementation, the function determination module 83 is further used for: At least one candidate prediction coefficient is used as the initial population to perform multiple population evolutions using the gene expression programming algorithm to obtain the corresponding candidate coefficient prediction function. After each population evolution, the candidate lever arm coefficient is calculated using the currently determined candidate coefficient prediction function, and the mean absolute percentage error is calculated based on the candidate lever arm coefficient and the labeled lever arm coefficient to obtain the loss value. If the loss value meets the preset constraints, the current candidate coefficient prediction function is determined as the lever arm coefficient prediction function.

[0086] In one possible implementation, the coefficient prediction module 84 is further used for: When it is necessary to predict the lever arm coefficient of the target cold rolling equipment, obtain the current relevant equipment parameters of the target cold rolling equipment; The relevant equipment parameters obtained at the moment are substituted into the lever arm coefficient prediction function to calculate the corresponding predicted lever arm coefficient.

[0087] The descriptions of the above device embodiments are similar to those of the above method embodiments, and have similar beneficial effects. For technical details not disclosed in the device embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0088] It should be noted that, in the embodiments of this application... Figure 8 The module division of the cold-rolled strip steel lever arm coefficient prediction device based on gene expression programming shown is illustrative and only represents one logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, exist as separate physical units, or be integrated into one unit with two or more units. The integrated units can be implemented in hardware, as software functional units, or a combination of both.

[0089] It should be noted that, in the embodiments of this application, if the above-described methods are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, or the parts that contribute to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause an electronic device to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), magnetic disks, or optical disks. Thus, the embodiments of this application are not limited to any specific hardware and software combination.

[0090] Figure 9 A schematic diagram of an electronic device according to an embodiment of this application is shown. For example... Figure 9 As shown in the figure, this application provides an electronic device, which can be a server, and its internal structure diagram can be as follows. Figure 9As shown, the electronic device includes a processor 920, a memory, and a transceiver 940 connected via a system bus 910. The processor 920 provides computing and control capabilities. The memory includes a non-volatile storage medium 931 and internal memory 932. The non-volatile storage medium 931 stores an operating system, computer programs, and a database. The internal memory 932 provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium 931. The database stores data. The transceiver 940 communicates with an external terminal via a network connection. The computer program is executed by the processor 920 to implement the aforementioned methods.

[0091] This application provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor 920, implements the steps of the method provided in the above embodiments.

[0092] This application provides a computer program product containing instructions that, when run on a computer, cause the computer to perform the steps in the method provided in the above-described method embodiments.

[0093] Those skilled in the art will understand that Figure 9 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the electronic device to which the present application is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0094] In one possible implementation, the shooting prompting device provided in this application can be implemented as a computer program, which can be configured as follows: Figure 9 The device operates on the electronic device shown. The memory of the electronic device can store various program modules that make up the above-described apparatus. The computer program composed of the various program modules causes the processor 920 to execute the steps of the methods in the various embodiments of this application described in this specification.

[0095] It should be noted that the descriptions of the storage medium and device embodiments above are similar to the descriptions of the method embodiments above, and have similar beneficial effects. For technical details not disclosed in the storage medium, storage medium, and device embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0096] It should be understood that the phrases "one embodiment," "an embodiment," or "some embodiments" mentioned throughout the specification mean that a specific feature, structure, or characteristic related to an embodiment is included in at least one embodiment of this application. Therefore, phrases such as "in one possible implementation," "in one embodiment," or "in some embodiments" appearing throughout the specification do not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the above-described embodiments are merely for descriptive purposes and do not represent the superiority or inferiority of the embodiments. The descriptions of the various embodiments above tend to emphasize the differences between the various embodiments; their similarities or commonalities can be referred to mutually, and for the sake of brevity, they will not be repeated here.

[0097] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that there can be three kinds of relationships. For example, object A and / or object B can represent three situations: object A exists alone, object A and object B exist simultaneously, and object B exists alone.

[0098] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0099] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The embodiments described above are merely illustrative. For example, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple modules or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or modules can be electrical, mechanical, or other forms.

[0100] The modules described above as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules. They may be located in one place or distributed across multiple network units. Some or all of the modules may be selected to achieve the purpose of this embodiment according to actual needs.

[0101] In addition, each functional module in the various embodiments of this application can be integrated into one processing unit, or each module can be a separate unit, or two or more modules can be integrated into one unit; the integrated modules can be implemented in hardware or in the form of hardware plus software functional units.

[0102] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as mobile storage devices, read-only memory (ROM), magnetic disks, or optical disks.

[0103] Alternatively, if the integrated units described above are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, or the parts that contribute to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause an electronic device to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0104] The methods disclosed in the several method embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments.

[0105] The features disclosed in the several product embodiments provided in this application can be arbitrarily combined without conflict to obtain new product embodiments.

[0106] The features disclosed in the several method or device embodiments provided in this application can be arbitrarily combined without conflict to obtain new method or device embodiments.

[0107] The above description is merely an embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for predicting the lever arm coefficient of cold-rolled strip steel based on gene expression programming, characterized in that, The method includes: Obtain a set of equipment parameters for multiple target cold rolling equipment, wherein the set of equipment parameters has a corresponding labeled lever arm coefficient; Select at least one equipment parameter related to the lever arm coefficient from the set of equipment parameters as the relevant equipment parameter; Based on the relevant equipment parameters and corresponding labeled lever arm coefficients in multiple sets of equipment parameters, parameter fitting is performed to obtain the candidate prediction coefficients for each relevant equipment parameter; The gene expression programming algorithm is executed using at least one of the candidate prediction coefficients as the initial population to obtain a lever arm coefficient prediction function with at least one of the candidate prediction coefficients as variables. When it is necessary to predict the lever arm coefficient of the target cold rolling equipment, the predicted lever arm coefficient is determined according to the lever arm coefficient prediction function.

2. The method according to claim 1, characterized in that, The acquisition of multiple sets of equipment parameter sets for target cold rolling equipment includes: Obtain multiple sets of candidate parameters for target cold rolling equipment, each set of candidate parameters including a corresponding labeled lever arm coefficient; Data normalization processing is performed on each of the candidate parameter sets; The candidate parameter sets are filtered according to the labeled lever arm coefficients corresponding to each candidate parameter set to obtain the equipment parameter set.

3. The method according to claim 2, characterized in that, The step of filtering the candidate parameter sets according to the labeled lever arm coefficients corresponding to each candidate parameter set to obtain the equipment parameter set includes: Calculate the mean and standard deviation of the labeled lever arm coefficients corresponding to each of the candidate parameter sets; Calculate the coefficient difference between the labeled lever arm coefficient and the average value for each candidate parameter set; If the absolute value of the coefficient difference is greater than three times the standard deviation, the candidate parameter set is deleted, and the candidate parameter set that passes the screening is determined as the equipment parameter set.

4. The method according to claim 1, characterized in that, The step of selecting at least one equipment parameter related to the lever arm coefficient from the set of equipment parameters as relevant equipment parameters includes: The correlation values ​​between each of the equipment parameters in the set of equipment parameters and the lever arm coefficient are calculated using the maximum mutual information coefficient method. Device parameters whose relevant values ​​are greater than the relevant threshold are identified as relevant device parameters.

5. The method according to claim 1, characterized in that, The step involves fitting parameters based on relevant equipment parameters and corresponding labeled lever arm coefficients from multiple sets of equipment parameter sets to obtain candidate prediction coefficients for each relevant equipment parameter, including: Multiple regression analysis is performed based on the relevant equipment parameters and corresponding labeled lever arm coefficients in multiple sets of equipment parameters to obtain a multiple linear regression equation between the lever arm coefficients and the relevant equipment parameters. The multiple linear regression equation includes a constant term and an unknown term composed of each labeled lever arm coefficient and a coefficient. Each of the aforementioned unknowns is determined as a candidate prediction coefficient.

6. The method according to claim 1, characterized in that, The step of using at least one of the candidate prediction coefficients as the initial population to execute a gene expression programming algorithm to obtain a lever arm coefficient prediction function with at least one of the candidate prediction coefficients as variables includes: Using at least one of the candidate prediction coefficients as the initial population, the gene expression programming algorithm is used to perform multiple population evolutions to obtain the corresponding candidate coefficient prediction function. After each population evolution, the currently determined candidate coefficient prediction function calculates the candidate lever arm coefficient, and the average absolute percentage error is calculated based on the candidate lever arm coefficient and the labeled lever arm coefficient to obtain the loss value; If the loss value satisfies the preset constraints, the current candidate coefficient prediction function is determined as the lever arm coefficient prediction function.

7. The method according to claim 1, characterized in that, The step of determining the predicted lever arm coefficient based on the lever arm coefficient prediction function when it is necessary to predict the lever arm coefficient of the target cold rolling equipment includes: When it is necessary to predict the lever arm coefficient of the target cold rolling equipment, obtain the current relevant equipment parameters of the target cold rolling equipment; The relevant equipment parameters obtained at the moment are substituted into the lever arm coefficient prediction function to calculate the corresponding predicted lever arm coefficient.

8. A device for predicting the lever arm coefficient of cold-rolled strip steel based on gene expression programming, characterized in that, The device includes: The parameter acquisition module is used to acquire multiple sets of equipment parameters for target cold rolling equipment, wherein the equipment parameter sets have corresponding labeled lever arm coefficients. The parameter filtering module is used to filter at least one equipment parameter related to the lever arm coefficient from the set of equipment parameters as the relevant equipment parameter; The coefficient determination module is used to perform parameter fitting based on the relevant equipment parameters and the corresponding labeled lever arm coefficients in multiple sets of equipment parameter sets, and to obtain the candidate prediction coefficients for each relevant equipment parameter. The function determination module is used to execute a gene expression programming algorithm with at least one of the candidate prediction coefficients as the initial population to obtain a lever arm coefficient prediction function with at least one of the candidate prediction coefficients as variables. The coefficient prediction module is used to determine the predicted lever arm coefficient based on the lever arm coefficient prediction function when it is necessary to predict the lever arm coefficient of the target cold rolling equipment.

9. An electronic device comprising a memory and a processor, the memory storing a computer program executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 7.

Images