Oriented silicon steel production parameter selection method, device, equipment and medium

Abstract

The invention provides an oriented silicon steel production parameter selection method, device, equipment and medium, and the oriented silicon steel production parameter selection method comprises the following steps: obtaining process production parameters and iron loss values of oriented silicon steel in a historical production process, screening the process production parameters, and obtaining target production parameters; training a preset prediction model by taking the target production parameters and the iron loss values as training samples to obtain a final prediction model, and screening the iron loss values to obtain a target iron loss value; clustering target production parameters associated with the target iron loss values to obtain a plurality of parameter data sets, calculating to obtain a plurality of iron loss predicted values based on each parameter data set and the final prediction model, and taking the process production parameter corresponding to the minimum iron loss predicted value as a selected production parameter; according to the method, selection of better process production parameters can be achieved, and when the selected process parameters are used for the subsequent production process of oriented silicon steel, the production quality and the production process stability of oriented silicon steel products are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of oriented silicon steel production, and in particular to a method, device, equipment and medium for selecting production parameters of oriented silicon steel. Background Art

[0002] Grain-oriented silicon steel is widely used in the electrical field, and its performance optimization has become a hot topic of research. Iron loss is a key indicator for evaluating the performance of grain-oriented silicon steel. The lower the iron loss, the higher the grade and the higher the quality. The iron loss value of grain-oriented silicon steel is affected by a variety of process production parameters (such as cold rolling temperature, tension, etc.) during its preparation. Therefore, the control of these process production parameters directly affects the quality of the product.

[0003] At present, although the industrial preparation process of oriented silicon steel is clear, the relationship between its process production parameters and performance is not yet fully understood due to the long and complex preparation process of oriented silicon steel. In addition, some mechanistic issues are still controversial. These factors lead to the current production method that relies on manual experience to select production parameters, which results in unstable production and large fluctuations in product quality.

[0004] Therefore, it is necessary to improve the current method of selecting production parameters of oriented silicon steel. Summary of the invention

[0005] In view of the above-mentioned shortcomings of the prior art, the present application provides a method, device, equipment and medium for selecting production parameters of oriented silicon steel to solve the above-mentioned technical problems.

[0006] According to one aspect of an embodiment of the present application, a method for selecting production parameters of oriented silicon steel is provided, comprising: obtaining process production parameters and iron loss values ​​of oriented silicon steel in a historical production process; calculating the variance of the process production parameters, the correlation between the iron loss value and the process production parameters, and screening the process production parameters in combination with the variance of the process production parameters and the correlation between the iron loss value and the process production parameters to obtain target production parameters; training a preset prediction model using the target production parameters and the iron loss values ​​as training samples to obtain a final prediction model; comparing the iron loss values ​​with a preset iron loss threshold, and screening the iron loss values ​​based on the comparison result to obtain a target iron loss value; clustering the target production parameters associated with the target iron loss value according to a preset number of categories to obtain multiple parameter data sets; calculating multiple iron loss prediction values ​​based on each of the parameter data sets and the final prediction model, and using the process production parameter corresponding to the minimum iron loss prediction value as the selected production parameter.

[0007] In one embodiment of the present application, if the process production parameters have multiple categories, then the process of screening the process production parameters in combination with the variance of the process production parameters and the correlation between the iron loss value and the process production parameters includes: selecting a parameter category as a target category; if the process production parameters of the target category meet the parameter screening conditions, then deleting the process production parameters of the target category from the process production parameter set to obtain an updated production parameter set; the parameter screening conditions include that the variance of the process production parameters of the target category is less than a preset variance threshold, or that the correlation between the iron loss value and the process production parameters of the target category is less than a preset correlation threshold. relevance threshold; the process production parameter set consists of the process production parameters in the historical production process; the correlation between the iron loss value and the process production parameters of the target category is characterized by the Pearson correlation coefficient between the iron loss value and the process production parameters of the target category, and the Spearman correlation coefficient between the iron loss value and the process production parameters of the target category; continue to select another parameter category as the update category; if the process production parameters of the update category meet the parameter screening condition, the process production parameters of the update category are deleted from the update production parameter set until all process production parameters are screened and the target production parameters are obtained.

[0008] In one embodiment of the present application, the preset prediction model is trained by the training sample to obtain the final prediction model, and the process includes: selecting a group of hyperparameters from the preset prediction model hyperparameter set as the target group hyperparameters, the preset prediction model hyperparameter set is composed of multiple groups of preset prediction model hyperparameters, the preset prediction model hyperparameters are composed of preset ridge parameters and preset random forest model hyperparameters, and the preset random forest model hyperparameters include the number of trees, the depth of the tree, and the number of samples per leaf node; constructing a preset ridge regression model with the preset ridge parameters in the target group hyperparameters, and constructing a preset random forest model with the preset random forest model hyperparameters in the target group hyperparameters; dividing the training sample into a training data set and a validation data set, training the preset ridge regression model with the training data set to obtain a trained ridge regression model; inputting the target production parameters in the training data set into the trained ridge regression model to obtain the model prediction value, and inputting the training data set into the trained ridge regression model to obtain the model prediction value, and The difference between the iron loss value in the training data set and the predicted value of the model is taken as the residual; the target production parameter of the training data set and the residual are used as residual samples to train the preset random forest model to obtain the trained random forest model; the sum of the trained ridge regression model and the trained random forest model is used as the iron loss prediction model; the verification data set is input into the iron loss prediction model to obtain the prediction error corresponding to the target group hyperparameter; other groups of hyperparameters are continuously selected from the preset prediction model hyperparameter set as the update group hyperparameters, and the iron loss prediction model is obtained again based on the update group hyperparameters and the training data set; the verification data set is input into the re-obtained iron loss prediction model to obtain the prediction error corresponding to the update group hyperparameter, until the prediction errors corresponding to all groups of hyperparameters are obtained; the minimum prediction error among the prediction errors corresponding to all groups of hyperparameters is selected, and the iron loss prediction model with the minimum prediction error calculated is used as the final prediction model.

[0009] In one embodiment of the present application, if the historical production process is characterized by different furnaces, the iron loss value is compared with a preset iron loss threshold value, and according to the comparison result, the process of screening the iron loss value includes: selecting a furnace as a target furnace; if the iron loss value of the target furnace is less than the preset iron loss threshold value, the iron loss value is used as a screening iron loss value; continue to select another furnace as an updated furnace; if the iron loss value of the updated furnace is less than the preset iron loss threshold value, the iron loss value is used as a newly added screening iron loss value, until the iron loss value screening of all furnaces is completed, and all screening iron loss values ​​are used as the target iron loss values.

[0010] In one embodiment of the present application, based on each of the parameter data sets and the final prediction model, the process of calculating multiple iron loss prediction values ​​includes: obtaining the parameter interval of each of the parameter data sets; and randomly generating a parameter population within each of the parameter intervals to obtain multiple parameter populations, each of the parameter populations including a first preset number of parameter vectors; selecting a parameter population as a target parameter population, inputting each parameter vector in the target parameter population into the final prediction model, and obtaining multiple fitness function values; and recording the number of population iterations; if the number of population iterations is greater than or equal to the preset number of iterations, stopping the iterative calculation process of the target parameter population, and continuing to select other parameter populations, and performing iterative calculations on other parameter populations until all parameter populations are calculated. The iterative operation process is completed, and all the obtained fitness function values ​​are used as the multiple iron loss prediction values; if the number of population iterations is less than the preset number of iterations, then by performing mutation, crossover and selection operations on each parameter vector in the target parameter population, an updated parameter population and a newly added fitness function value are obtained, and the number of population iterations is updated until the updated number of population iterations is greater than or equal to the preset number of iterations threshold; if the updated number of population iterations is greater than or equal to the preset number of iterations, the iterative operation process of the target parameter population is stopped, and other parameter populations are continuously selected and iterative operations are performed on other parameter populations until the iterative operation process of all parameter populations is completed, and all the obtained fitness function values ​​are used as the multiple iron loss prediction values.

[0011] In one embodiment of the present application, the process of performing mutation, crossover and selection operations on each parameter vector in the target parameter population includes: selecting a parameter vector as the target parameter vector; randomly selecting a second preset number of parameter vectors from the target parameter population, the second preset number of parameter vectors being different from the target parameter vector; performing a mutation operation on the second preset number of parameter vectors to obtain a mutation vector; the second preset number is less than the first preset number; using the target parameter vector as the initial vector, and performing a crossover operation on the components of the initial vector and the components of the mutation vector to obtain an operation vector; inputting the operation vector into the final prediction model to obtain a first fitness function value; and inputting the target parameter vector into the final prediction model to obtain a second fitness function value; if the first fitness function value is less than the second fitness function value, then The operation vector replaces the target parameter vector to obtain the parameter population for the initial update; and selects other parameter vectors in the parameter population for the initial update as the update parameter vector, and performs mutation, crossover and selection operations on the update parameter vector until the mutation, crossover and selection operations on all parameter vectors in the target parameter population are completed, and the parameter population finally updated is used as the update parameter population, and all first fitness function values ​​are used as the newly added fitness function values; if the first fitness function value is greater than or equal to the second fitness function value, the target parameter vector is not updated, and other parameter vectors in the target parameter population are selected as the update parameter vector, and mutation, crossover and selection operations are performed on the update parameter vector until the mutation, crossover and selection operations on all parameter vectors are completed to obtain the newly added fitness function value and the update parameter population.

[0012] In one embodiment of the present application, the process production parameters at least include: process production parameters of a normalizing pickling unit, process production parameters of a rolling mill, and process production parameters of a continuous annealing unit.

[0013] According to one aspect of an embodiment of the present application, a device for selecting production parameters of oriented silicon steel is provided, comprising: a parameter acquisition module, used to acquire process production parameters and iron loss values ​​of oriented silicon steel in a historical production process; a parameter screening module, used to calculate the variance of the process production parameters, the correlation between the iron loss value and the process production parameters, and screen the process production parameters in combination with the variance of the process production parameters and the correlation between the iron loss value and the process production parameters to obtain target production parameters; a model training module, used to train a preset prediction model using the target production parameters and the iron loss values ​​as training samples to obtain a final prediction model; an iron loss value screening module, used to compare the iron loss value with a preset iron loss threshold, and based on the comparison result, screen the iron loss value to obtain the target iron loss value; based on a preset number of categories, clustering the target production parameters associated with the target iron loss value to obtain multiple parameter data sets; a parameter selection module, used to calculate multiple iron loss prediction values ​​based on each of the parameter data sets and the final prediction model, and use the process production parameter corresponding to the minimum iron loss prediction value as the selected production parameter.

[0014] According to one aspect of an embodiment of the present application, an electronic device is provided, comprising: one or more processors; a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the electronic device implements the method for selecting production parameters of oriented silicon steel as described above.

[0015] According to one aspect of an embodiment of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor of a computer, the computer executes the above-mentioned method for selecting production parameters of oriented silicon steel.

[0016] Beneficial effects of the present invention: The present invention obtains process production parameters and iron loss values ​​of oriented silicon steel in the historical production process, calculates the variance of the process production parameters, the iron loss value and the correlation between the process production parameters, and screens the process production parameters in combination with the variance of the process production parameters, the iron loss value and the correlation between the process production parameters to obtain target production parameters, uses the target production parameters and the iron loss value as training samples to train a preset prediction model to obtain a final prediction model, compares the iron loss value with a preset iron loss threshold, and screens the iron loss value according to the comparison result to obtain the target iron loss value; clusters the target production parameters associated with the target iron loss value according to the preset number of categories to obtain multiple parameter data sets, and calculates multiple parameter data sets based on each parameter data set and the final prediction model. An iron loss prediction value is obtained, and the process production parameter corresponding to the minimum iron loss prediction value is used as the selected production parameter. In the above process, the process production parameters with less influence on the iron loss value are eliminated by screening the process production parameters, and the smaller iron loss value is obtained by screening the iron loss value. On the basis of the smaller iron loss value, the target production parameters associated with the target iron loss value are clustered to obtain multiple parameter data sets. Based on each parameter data set and the final prediction model, multiple iron loss prediction values ​​are calculated, and the process production parameter corresponding to the minimum iron loss prediction value is used as the selected production parameter, thereby realizing the selection of better process production parameters. When the selected process parameters are applied to the subsequent production process of oriented silicon steel, the production quality of oriented silicon steel products and the stability of the production process can be improved.

[0017] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings herein are incorporated into the specification and constitute a part of the specification, showing embodiments consistent with the present application, and together with the specification, are used to explain the principles of the present application. Obviously, the drawings described below are only some embodiments of the present application, and for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0019] Figure 1 is a schematic diagram of an exemplary system architecture shown in an exemplary embodiment of the present application;

[0020] Figure 2 is a flow chart of a method for selecting production parameters of oriented silicon steel shown in an exemplary embodiment of the present application;

[0021] Figure 3 is a flow chart of a method for selecting production parameters of oriented silicon steel shown in another exemplary embodiment of the present application;

[0022] Figure 4 is a flowchart of searching for optimal production parameters based on a differential evolution algorithm, shown in an exemplary embodiment of the present application;

[0023] Figure 5 is a block diagram of a system for selecting production parameters of oriented silicon steel shown in an exemplary embodiment of the present application;

[0024] Figure 6 A block diagram of a device for selecting production parameters of oriented silicon steel suitable for implementing an embodiment of the present application is shown;

[0025] Figure 7 A schematic diagram of the structure of a computer system suitable for implementing an electronic device of an embodiment of the present application is shown. DETAILED DESCRIPTION

[0026] Here, exemplary embodiments will be described in detail, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are only examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims.

[0027] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0028] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0029] The term "multiple" as used in this application refers to two or more than two. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the related objects are in an "or" relationship.

[0030] The technical solution of the embodiment of the present application involves related technologies such as welding of steel structure parts, which is specifically described by the following embodiments:

[0031] Figure 1 It is a schematic diagram of an exemplary system architecture shown in an exemplary embodiment of the present application.

[0032] Reference Figure 1 As shown, the system architecture may include a data storage device 101 and a computer device 102. The computer device 102 may be at least one of a desktop graphics processing unit (GPU) computer, a GPU computing cluster, a neural network computer, etc. Relevant technicians may use the computer device 102 to obtain the process production parameters and iron loss values ​​of oriented silicon steel in the historical production process, calculate the variance of the process production parameters, the iron loss value and the correlation between the process production parameters, and screen the process production parameters in combination with the variance of the process production parameters, the iron loss value and the correlation between the process production parameters to obtain the target production parameters, train the preset prediction model using the target production parameters and the iron loss value as training samples to obtain the final prediction model, compare the iron loss value with the preset iron loss threshold, and screen the iron loss value according to the comparison result to obtain the target iron loss value; cluster the target production parameters associated with the target iron loss value according to the preset number of categories to obtain multiple parameter data sets, calculate multiple iron loss prediction values ​​based on each parameter data set and the final prediction model, and use the process production parameter corresponding to the minimum iron loss prediction value as the selected production parameter. The data storage device 101 is used to store the process production parameters and iron loss values ​​of the oriented silicon steel in the historical production process. In this embodiment, the data storage device 101 uses a random access memory (RAM) to store the process production parameters and iron loss values ​​of the oriented silicon steel in the historical production process, and provides them to the computer device 102 for processing.

[0033] Schematically, after obtaining the process production parameters and iron loss values ​​of the oriented silicon steel of the data storage device 101 in the historical production process, the computer device 102 calculates the variance of the process production parameters, the iron loss value and the correlation between the process production parameters, and screens the process production parameters in combination with the variance of the process production parameters, the iron loss value and the correlation between the process production parameters to obtain the target production parameters, uses the target production parameters and the iron loss value as training samples to train the preset prediction model to obtain the final prediction model, compares the iron loss value with the preset iron loss threshold, and screens the iron loss value based on the comparison result to obtain the target iron loss value; clusters the target production parameters associated with the target iron loss value according to the preset number of categories to obtain multiple parameter data sets, and based on each parameter data set and the final prediction model, a clustering operation is performed. model, multiple iron loss prediction values ​​are calculated, and the process production parameters corresponding to the minimum iron loss prediction value are used as the selected production parameters. In the above process, the process production parameters with less influence on the iron loss value are eliminated by screening the process production parameters, and the smaller iron loss value is obtained by screening the iron loss value, and on the basis of the smaller iron loss value, the target production parameters associated with the target iron loss value are clustered to obtain multiple parameter data sets, and based on each parameter data set and the final prediction model, multiple iron loss prediction values ​​are calculated, and the process production parameters corresponding to the minimum iron loss prediction value are used as the selected production parameters, thereby realizing the selection of better process production parameters, and when the selected process parameters are applied to the subsequent production process of oriented silicon steel, the production quality of oriented silicon steel products and the stability of the production process can be improved.

[0034] It should be noted that the method for selecting production parameters of oriented silicon steel provided in the embodiment of the present application is generally executed by the computer device 102 , and accordingly, the device for selecting production parameters of oriented silicon steel is generally arranged in the computer device 102 .

[0035] The implementation details of the technical solution of the embodiment of the present application are described in detail below:

[0036] Figure 2 is a flowchart of a method for selecting production parameters of oriented silicon steel according to an exemplary embodiment of the present application. The method for selecting production parameters of oriented silicon steel can be executed by a computing and processing device. The computing and processing device can be Figure 1 The computer device 102 shown in FIG. Figure 2 As shown, the method for selecting production parameters of oriented silicon steel includes at least steps S210 to S250, which are described in detail as follows:

[0037] In step S210, process production parameters and iron loss values ​​of the oriented silicon steel in the historical production process are obtained.

[0038] In one embodiment of the present application, the historical production process is characterized by different furnaces, the process production parameters in the historical production process include the process production parameters of different furnaces, the iron loss values ​​in the historical production process include the iron loss values ​​of different furnaces, the process production parameters include the production parameters of the normalizing and pickling processes during the production of oriented silicon steel, the production parameters of the cold rolling process during the production of oriented silicon steel, and the production parameters of the annealing process during the production of oriented silicon steel, etc.

[0039] In this embodiment, the process production parameters of all furnaces may be stored in the form of a matrix, and the size of the matrix is ​​n*J, wherein n represents the number of historical furnaces, and J represents the total number of categories of the process production parameters.

[0040] In step S220, the variance of the process production parameters, the iron loss value and the correlation between the process production parameters are calculated, and the process production parameters are screened based on the variance of the process production parameters, the iron loss value and the correlation between the process production parameters to obtain the target production parameters.

[0041] In this embodiment, taking the j-th category process production parameter as an example, the variance calculation formula of the j-th category process production parameter is as follows:

[0042]

[0043] Among them, D j represents the variance of the j-th category process production parameter, x ij represents the process production parameters of the jth category in the i-th heat, represents the mean value of the process production parameter of the jth category, n represents the number of historical furnaces, i = 1, 2, ..., n, j = 1, 2, ..., J, and J represents the total number of categories of process production parameters.

[0044] In this embodiment, the mean expression of the j-th type process production parameter is as follows:

[0045]

[0046] in, represents the mean value of the j-th type of process production parameters, x ij represents the process production parameters of the jth type in the ith heat, and n represents the number of historical heats.

[0047] In this embodiment, the process production parameters are screened based on the variance of the process production parameters and the correlation between the iron loss value and the process production parameters, and the process production parameters with little influence on the iron loss value are eliminated.

[0048] In step S230, the preset prediction model is trained using the target production parameters and iron loss values ​​as training samples to obtain a final prediction model.

[0049] In this embodiment, the preset prediction model includes a preset ridge regression model and a preset random forest model. In the process of training the preset prediction model through training samples, first, the regression parameters, the attributes of the splitting of each node of each tree, and the training samples that finally fall into the leaf nodes are adjusted, and then the preset ridge parameters and the preset hyperparameters are adjusted to obtain the final prediction model, thereby improving the accuracy of the final prediction model in predicting the iron loss value.

[0050] In step S240, the iron loss value is compared with a preset iron loss threshold value, and the iron loss value is screened according to the comparison result to obtain a target iron loss value; the target production parameters associated with the target iron loss value are clustered according to a preset number of categories to obtain multiple parameter data sets.

[0051] In this embodiment, the iron loss values ​​are screened to obtain a smaller iron loss value, and based on the smaller iron loss value, the target production parameters associated with the target iron loss value are clustered to obtain multiple parameter data sets. The number of preset categories can be set according to actual conditions, and will not be described in detail here.

[0052] In this embodiment, the target iron loss value includes iron loss values ​​of multiple furnaces, and the target production parameters associated with the target iron loss value are clustered to obtain multiple parameter data sets. The process includes: determining the furnace corresponding to the target iron loss value, and using the furnace corresponding to the target iron loss value as the screening furnace, extracting the process production parameters of the screening furnace from the target production parameters, and clustering the process production parameters of the screening furnace to obtain multiple parameter data sets.

[0053] In this embodiment, the iron loss values ​​and process production parameters belonging to the same heat have an associated relationship or a corresponding relationship. The method for clustering the process production parameters of the target heat adopts a kmeans clustering algorithm or the like.

[0054] In step S250, a plurality of iron loss prediction values ​​are calculated based on each parameter data set and the final prediction model, and the process production parameter corresponding to the minimum iron loss prediction value is used as the selected production parameter.

[0055] In this embodiment, based on each parameter data set and the final prediction model, multiple iron loss prediction values ​​are calculated, and the process production parameters corresponding to the minimum iron loss prediction value are used as the selected production parameters, thereby realizing the selection of better process production parameters. When the selected process parameters are applied to the subsequent production process of oriented silicon steel, the production quality of oriented silicon steel products and the stability of the production process can be improved.

[0056] In one embodiment of the present application, if the process production parameters have multiple categories, the process of screening the process production parameters in combination with the variance of the process production parameters, the iron loss value and the correlation between the process production parameters includes:

[0057] Select a parameter category as the target category.

[0058] In this embodiment, the method for selecting parameter categories may be selected in the order of the parameter categories, or may be selected randomly, and no specific limitation is made here.

[0059] If the process production parameters of the target category meet the parameter screening condition, the process production parameters of the target category are deleted from the process production parameter set to obtain an updated production parameter set.

[0060] In this embodiment, the parameter screening conditions include that the variance of the process production parameters of the target category is less than a preset variance threshold, or the correlation between the iron loss value and the process production parameters of the target category is less than a preset correlation threshold; the process production parameter set is composed of the process production parameters in the historical production process; the correlation between the iron loss value and the process production parameters of the target category is characterized by the Pearson correlation coefficient between the iron loss value and the process production parameters of the target category, and the Spearman correlation coefficient between the iron loss value and the process production parameters of the target category.

[0061] In this embodiment, the preset variance threshold can be set according to actual conditions, for example, set to 0.005, and the preset correlation threshold can be set according to actual conditions, for example, set to 0.5.

[0062] In this embodiment, the judgment process of whether the correlation between the iron loss value and the process production parameters of the target category is less than the preset correlation threshold includes: if the Pearson correlation coefficient between the iron loss value and the process production parameters of the target category is less than the preset correlation threshold, and the Spearman correlation coefficient between the iron loss value and the process production parameters of the target category is less than the preset correlation threshold, then it is determined that the correlation between the iron loss value and the process production parameters of the target category is less than the preset correlation threshold.

[0063] In this embodiment, taking the target category as the jth category as an example, the Pearson correlation coefficient calculation formula between the iron loss value and the process production parameter of the jth category is as follows:

[0064]

[0065] in, represents the Pearson correlation coefficient between the production process parameters and the iron loss value of the jth category, x ij represents the process production parameters of the jth category in the i-th heat, represents the mean value of the process production parameter of the jth category, n represents the number of historical furnaces, i = 1, 2, ..., n, j = 1, 2, ..., J, J represents the total number of categories of process production parameters, Y i represents the iron loss value of the ith heat, It represents the mean value of iron loss in all heats.

[0066] In this embodiment, the mean value expression of the iron loss value in all heats is as follows:

[0067]

[0068] in, It represents the mean value of iron loss in all heats, Y i represents the iron loss value of the ith heat, and n represents the number of historical heats.

[0069] In this embodiment, taking the target category as the jth category as an example, the Spearman correlation coefficient calculation formula between the iron loss value and the process production parameter of the jth category is as follows:

[0070]

[0071] in, represents the Spearman correlation coefficient between the production process parameters and the iron loss value of the jth category, d i It represents the difference between the descending sorted number of the process production parameter of the jth category and the descending sorted number of the iron loss value, i represents the furnace, and n represents the number of historical furnaces.

[0072] In this embodiment, the calculation formula of the difference between the descending sorting sequence number of the process production parameter of the jth category and the descending sorting sequence number of the iron loss value is as follows:

[0073] d i =R(x ij )-R(Y i ) Formula (6)

[0074] Among them, d i represents the difference between the descending order of the process production parameters of the jth category and the descending order of the iron loss value, R(x ij ) represents the descending order of the jth process production parameter in the i-th heat among the process production parameters of the same category, R(Y i ) represents the descending sorting number of the iron loss value in the ith heat among all the iron loss values, and n represents the number of historical heats.

[0075] Continue to select another parameter category as the update category.

[0076] In this embodiment, the method for selecting another parameter category may be selected according to the order of the parameter categories, or may be randomly selected, which is not specifically limited herein.

[0077] If the process production parameters of the updated category meet the parameter screening conditions, the process production parameters of the updated category will be deleted from the updated production parameter set until all process production parameters are screened and the target production parameters are obtained.

[0078] In this embodiment, by screening all process production parameters, the target production parameters obtained are the process production parameters that have a greater impact on the iron loss value, and the preset prediction model is trained by the target production parameters and the iron loss value to improve the accuracy of the final prediction model.

[0079] In one embodiment of the present application, the process of training the preset prediction model through training samples to obtain the final prediction model includes:

[0080] A set of hyperparameters is selected from a preset prediction model hyperparameter set as a target group hyperparameter.

[0081] In this embodiment, the preset prediction model hyperparameter set consists of multiple groups of preset prediction model hyperparameters, the preset prediction model hyperparameters consist of preset ridge parameters and preset random forest model hyperparameters, and the preset random forest model hyperparameters include the number of trees, the depth of the tree, and the number of samples for each leaf node.

[0082] In this embodiment, the target group of hyperparameters may be selected from the preset prediction model hyperparameter set by random selection or by selecting in the order of each group of hyperparameters in the preset prediction model hyperparameter set.

[0083] The preset ridge regression model is constructed with the preset ridge parameters in the target group hyperparameters, and the preset random forest model is constructed with the preset random forest model hyperparameters in the target group hyperparameters.

[0084] In this embodiment, the process of constructing a preset ridge regression model with the preset ridge parameters in the target group hyperparameters can be implemented by referring to the process of constructing a preset ridge regression model with the preset ridge parameters in the relevant materials, and the process of constructing a preset random forest model with the preset random forest model hyperparameters in the target group hyperparameters can be implemented by referring to the process of constructing a preset random forest model with the preset random forest model hyperparameters in the relevant materials, and no specific limitation is made here.

[0085] The training samples are divided into a training data set and a validation data set. The preset ridge regression model is trained with the training data set to obtain a trained ridge regression model.

[0086] In this embodiment, the training data set and the verification data set are divided according to the ratio of required training data and required verification data, the training data set includes a first part of target production parameters and a first part of iron loss values, the verification data set includes a second part of target production parameters and a second part of iron loss values, the first part of target production parameters and the second part of target production parameters constitute all target production parameters, and the first part of iron loss values ​​and the second part of iron loss values ​​constitute all iron loss values.

[0087] The target production parameters in the training data set are input into the trained ridge regression model to obtain the model prediction value, and the difference between the iron loss value in the training data set and the model prediction value is taken as the residual.

[0088] In this embodiment, taking the ith heat as an example, the residual calculation formula for the ith heat is as follows:

[0089]

[0090] in, represents the residual of the ith heat, Y i represents the iron loss value of the i-th batch in the training data set, represents the model prediction value of the ith heat, X i represents the target production parameters of the i-th batch in the training data set, represents the regression parameter in the trained ridge regression model, and λ represents the preset ridge parameter.

[0091] The target production parameters and residuals of the training data set are used as residual samples to train the preset random forest model to obtain a trained random forest model.

[0092] In this embodiment, given the preset random forest model hyperparameters, the preset random forest model is trained through residual samples to adjust the classification attributes of the leaf nodes and the training samples falling into the leaf nodes.

[0093] The sum of the trained ridge regression model and the trained random forest model is used as the iron loss prediction model; the validation data set is input into the iron loss prediction model to obtain the prediction error corresponding to the target group hyperparameters.

[0094] In this embodiment, the expression of the iron loss prediction model is as follows:

[0095]

[0096] in, It represents the prediction result obtained when the process production parameter X is input. represents the trained ridge regression model, T A (X) represents the trained random forest model, represents the regression parameters in the trained ridge regression model, λ represents the preset ridge parameters, and A represents the preset random forest model hyperparameters.

[0097] Continue to select other groups of hyperparameters from the preset prediction model hyperparameter set as the update group hyperparameters, and re-obtain the iron loss prediction model based on the update group hyperparameters and the training data set; input the verification data set into the re-obtained iron loss prediction model to obtain the error corresponding to the update group hyperparameters, until the prediction errors corresponding to all groups of hyperparameters are obtained.

[0098] In this embodiment, based on the update group hyperparameters and the training data set, the process of re-obtaining the iron loss prediction model includes: reconstructing the preset ridge regression model with the preset ridge parameters in the update group hyperparameters, and reconstructing the preset random forest model with the preset random forest model hyperparameters in the update group hyperparameters; training the reconstructed preset ridge regression model with the training data set to obtain a retrained ridge regression model; inputting the target production parameters in the training data set into the retrained ridge regression model to obtain the newly added model prediction value, and taking the difference between the iron loss value in the training data set and the newly added model prediction value as the newly added residual; taking the target production parameters of the training data set and the newly added residual as the newly added residual samples, training the reconstructed preset random forest model to obtain the retrained random forest model; and taking the sum of the retrained ridge regression model and the retrained random forest model as the re-obtained iron loss prediction model.

[0099] The minimum prediction error among the prediction errors corresponding to all groups of hyperparameters is selected, and the iron loss prediction model with the minimum prediction error is calculated as the final prediction model.

[0100] In this embodiment, the iron loss prediction model with the minimum prediction error is calculated as the final prediction model, so as to find the optimal ridge parameter and the optimal random forest model hyperparameter in the preset prediction model hyperparameter set, thereby improving the prediction accuracy of the final prediction model.

[0101] In one embodiment of the present application, if the historical production process is characterized by different heats, the iron loss value is compared with a preset iron loss threshold, and the process of screening the iron loss value according to the comparison result includes:

[0102] Select a heat as the target heat.

[0103] In this embodiment, the target heat can be selected randomly, and different heats can be sorted and the target heat can be selected according to the sorting order.

[0104] If the iron loss value of the target heat is less than the preset iron loss threshold, the iron loss value is used as the screening iron loss value.

[0105] In this embodiment, the preset iron loss threshold is set according to actual conditions. For example, the preset iron loss threshold is set to 1.

[0106] Continue to select another heat as the update heat.

[0107] In this embodiment, the selection method for updating the furnace can be a random selection method, and different furnaces can be sorted and the updated furnace can be selected according to the sorting order.

[0108] If the iron loss value of the updated furnace is less than the preset iron loss threshold, the iron loss value will be used as the newly added screening iron loss value until the iron loss values ​​of all furnaces are screened and all the screened iron loss values ​​are used as the target iron loss values.

[0109] In this embodiment, all screened iron loss values ​​include iron loss values ​​in all heats that are less than a preset iron loss threshold.

[0110] In one embodiment of the present application, based on each parameter data set and the final prediction model, the process of calculating multiple iron loss prediction values ​​includes:

[0111] A parameter interval of each parameter data set is obtained; and a parameter population is randomly generated within each parameter interval to obtain multiple parameter populations.

[0112] In this embodiment, each parameter population includes a first preset number of parameter vectors, the upper limit of the parameter interval is the maximum value of the parameter data in the parameter data set, and the lower limit of the parameter interval is the minimum value of the parameter data in the parameter data set. The first preset number can be set according to actual conditions.

[0113] A parameter population is selected as the target parameter population, and each parameter vector in the target parameter population is input into the final prediction model to obtain multiple fitness function values; and the number of population iterations is recorded.

[0114] In this embodiment, each parameter vector is input into the final prediction model to obtain a fitness function value. After each parameter vector in the target parameter population is input into the final prediction model to obtain multiple fitness function values, the number of population iterations is recorded as 1.

[0115] If the number of population iterations is greater than or equal to the preset number of iterations, the iterative calculation process of the target parameter population is stopped, and other parameter populations are continuously selected and iteratively calculated for the other parameter populations until the iterative calculation process for all parameter populations is completed, and all the fitness function values ​​obtained are used as multiple iron loss prediction values.

[0116] In this embodiment, the iterative calculation process of other parameter populations is the same as the iterative calculation process of the target parameter population. The preset number of iterations can be set according to actual conditions and is not specifically limited here.

[0117] If the number of population iterations is less than the preset number of iterations, the updated parameter population and the newly added fitness function value are obtained by performing mutation, crossover and selection operations on each parameter vector in the target parameter population, and the number of population iterations is updated until the updated number of population iterations is greater than or equal to the preset iteration threshold.

[0118] In this embodiment, after obtaining the updated parameter population and the newly added fitness function value, mutation, crossover and selection operations are continued to be performed on each parameter vector in the updated parameter population until the number of iterations of the updated population is greater than or equal to the preset iteration number threshold. By updating the parameter population, newly added fitness function values ​​are continuously obtained, thereby increasing the number of fitness function values.

[0119] If the number of iterations of the updated population is greater than or equal to the preset number of iterations, the iterative calculation process of the target parameter population is stopped, and other parameter populations are continuously selected and iteratively calculated for the other parameter populations until the iterative calculation process of all parameter populations is completed, and all the fitness function values ​​obtained are used as multiple iron loss prediction values.

[0120] In this embodiment, by updating and iterating all populations, not only the number of fitness function values ​​is increased, but also the diversity of parameter populations is enriched, so that the number of obtained iron loss prediction values ​​is richer and more diverse.

[0121] In one embodiment of the present application, the process of performing mutation, crossover and selection operations on each parameter vector in the target parameter population includes:

[0122] A parameter vector is selected as a target parameter vector; and a second preset number of parameter vectors are randomly selected from the target parameter population.

[0123] In this embodiment, the second preset number of parameter vectors is different from the target parameter vectors, the second preset number is smaller than the first preset number, and the second preset number can be set according to actual conditions. For example, the second preset number is set to 3.

[0124] A mutation operation is performed on a second preset number of parameter vectors to obtain a mutation vector.

[0125] In this embodiment, the expression of the mutation vector is as follows:

[0126] V=X c +F(X a -X b ) Formula (9)

[0127] Among them, V represents the mutation vector, X a , X b and Xc They represent three randomly selected parameter vectors, and F represents the variation factor.

[0128] The target parameter vector is used as the initial vector, and the components of the initial vector and the components of the mutation vector are cross-operated to obtain the operation vector.

[0129] In this embodiment, the process of performing crossover operations on the components of the initial vector and the components of the mutation vector includes: randomly extracting a component of the initial vector, assuming that the sth component of the initial vector is extracted, setting the sth component of the initial vector as the sth component of the mutation vector, generating a uniformly distributed random number between 0 and 1 for each component at other positions of the initial vector, if the random number at a certain position is less than a preset crossover probability factor, setting the component value of the initial vector at that position to the component value of the mutation vector at that position, repeating the above steps until the component values ​​at all positions of the initial vector are set to obtain the operation vector.

[0130] The operation vector is input into the final prediction model to obtain a first fitness function value; and the target parameter vector is input into the final prediction model to obtain a second fitness function value.

[0131] In this embodiment, after the operation vector is obtained, the first fitness function value is calculated by the operation vector, thereby increasing the number and diversity of fitness function values.

[0132] If the first fitness function value is less than the second fitness function value, the target parameter vector is replaced by the operation vector to obtain the parameter population for the first update; and other parameter vectors in the parameter population for the first update are selected as update parameter vectors, and the update parameter vectors are mutated, crossed over and selected until the mutation, crossover and selection operations are completed for all parameter vectors in the target parameter population, and the parameter population finally updated is used as the update parameter population, and all the first fitness function values ​​are used as newly added fitness function values.

[0133] In this embodiment, when the first fitness function value is less than the second fitness function value, it means that the first fitness function value is a more ideal iron loss prediction value than the second fitness function value. Therefore, the first fitness function value is used as a newly added fitness function value. At the same time, the target parameter vector is updated, and other parameter vectors in the initially updated parameter population are continuously selected as update parameter vectors, and the update parameter vectors are mutated, crossed over, and selected until the mutation, crossover, and selection operations on all parameter vectors in the target parameter population are completed, and the target parameter population completes an iterative operation.

[0134] If the first fitness function value is greater than or equal to the second fitness function value, the target parameter vector is not updated, and other parameter vectors in the target parameter population are selected as update parameter vectors, and the update parameter vectors are mutated, crossed over, and selected until all parameter vectors are mutated, crossed over, and selected to obtain newly added fitness function values ​​and update parameter populations.

[0135] In this embodiment, when the first fitness function value is greater than or equal to the second fitness function value, it means that the second fitness function value is a more ideal iron loss prediction value compared with the first fitness function value. In this case, the target parameter vector is not updated, and other parameter vectors in the target parameter population are selected as update parameter vectors, and the update parameter vectors are mutated, crossed over and selected until the mutation, crossover and selection operations are completed for all parameter vectors, and the target parameter population completes one iterative operation.

[0136] In one embodiment of the present application, the process production parameters at least include: process production parameters of a normalizing pickling unit, process production parameters of a rolling mill, and process production parameters of a continuous annealing unit.

[0137] Figure 3 is a flow chart of a method for selecting production parameters of oriented silicon steel shown in another exemplary embodiment of the present application, such as Figure 3 As shown, the method for selecting production parameters of oriented silicon steel includes: (1) obtaining historical production data of oriented silicon steel, the historical production data including process production parameters and iron loss values ​​of all processes such as normalization and pickling process, cold rolling process and return process of oriented silicon steel production; (2) calculating the variance of the process production parameters, the Pearson correlation coefficient between the iron loss value and the process production parameters, and the Spearman correlation coefficient between the iron loss value and the process production parameters. If the process production parameters meet the parameter screening conditions, the process production parameters are removed from the historical production data to obtain the target production parameters; (3) selecting the iron loss value that is less than the preset iron loss threshold As the target iron loss value, determine the furnace corresponding to the target iron loss value, and use the furnace corresponding to the target iron loss value as the screening furnace, extract the process production parameters of the screening furnace from the target production parameters, cluster the process production parameters of the screening furnace by the kmeans clustering algorithm, and obtain the parameter interval of each type of parameter data set. The upper limit value of the parameter interval is the maximum value of the parameter data in the parameter data set, and the lower limit value of the parameter interval is the minimum value of the parameter data in the parameter data set; (4) train the preset prediction model through the training samples to obtain the final prediction model; (5) use the differential evolution algorithm to search for the optimal production parameters.

[0138] Figure 4 is a flowchart of searching for optimal production parameters based on a differential evolution algorithm, as shown in an exemplary embodiment of the present application. Figure 4As shown in FIG. 1 , the process of searching for optimal production parameters based on the differential evolution algorithm includes: (1) selecting a parameter data set, generating population individuals X1, X2, …, X2 of the target parameter population within the parameter interval of the parameter data set, and NP , where NP represents the size of the target parameter population. Each population individual is brought into the final prediction model, the fitness function value of each population individual is obtained and the number of population iterations is recorded; each population individual is a parameter vector, and each parameter vector includes a second preset number of parameter components; (2) It is determined that the number of population iterations is greater than or equal to the preset number of iterations. If the number of population iterations is greater than or equal to the preset number of iterations, the iterative operation process of the target parameter population is stopped, and other parameter populations are continuously selected and iterative operations are performed on other parameter populations until the iterative operation process of all parameter populations is completed. All the fitness function values ​​obtained are used as multiple iron loss prediction values, and the minimum iron loss prediction value corresponding to the output is output. process production parameters; if the number of population iterations is less than the preset number of iterations, mutation, crossover and selection operations are performed on each parameter vector in the target parameter population to determine whether the mutation, crossover and selection operations on each parameter vector in the target parameter population are completed; if the mutation, crossover and selection operations on each parameter vector in the target parameter population are completed, then it is determined that the number of population iterations of the target parameter population is greater than or equal to the preset number of iterations; if the mutation, crossover and selection operations on each parameter vector in the target parameter population are not completed, then the mutation, crossover and selection operations on the parameter vectors in the target parameter population are continued until the mutation, crossover and selection operations on each parameter vector in the target parameter population are completed.

[0139] In this embodiment, the process of performing mutation, crossover and selection operations on each parameter vector in the target parameter population includes: selecting a parameter vector from the target parameter population as the target parameter vector X i , and randomly select 3 different parameter vectors X a , X b and X c , the three parameter vectors are different from the target parameter vector; perform mutation operation on the three parameter vectors to obtain the mutation vector V; replace the target parameter vector X iAs the initial vector U, a component of the initial vector U is randomly extracted. Assuming that the s-th component of the initial vector U is extracted, the s-th component of the initial vector U is set as the s-th component of the mutation vector V. For each component at other positions of the initial vector U, a uniformly distributed random number between 0 and 1 is generated. If the random number at a certain position is less than the preset crossover probability factor, the component value of the initial vector U at this position is set to the component value of the mutation vector V at this position. The above steps are repeated until the component values ​​at all positions of the initial vector U are set to obtain an operation vector; the operation vector is input into the final prediction model to obtain a first fitness function value; and the target parameter vector is input into the final prediction model to obtain a second fitness function value. If the first fitness function value is less than the second fitness function value, the target parameter vector is replaced by the operation vector, thereby realizing the update of the target parameter vector and obtaining the parameter population for the first update.

[0140] On the basis of the parameter population updated for the first time, other parameter vectors in the parameter population updated for the first time are selected as update parameter vectors, and mutation, crossover and selection operations are performed on the update parameter vectors until the mutation, crossover and selection operations on all parameter vectors in the target parameter population are completed, and the parameter population finally updated is used as the update parameter population, and all the first fitness function values ​​are used as newly added fitness function values.

[0141] Figure 5 is a block diagram of a system for selecting production parameters of oriented silicon steel as shown in an exemplary embodiment of the present application. Figure 5 As shown, the oriented silicon steel production parameter selection system includes: a data acquisition module, a feature selection module, a clustering module, an iron loss prediction model construction module and an optimal production parameter solution module, the data acquisition module is used to obtain the historical production data of oriented silicon steel; the feature selection module is used to screen the target production parameters and the target iron loss value; the clustering module is used to determine the furnace corresponding to the target iron loss value, and the furnace corresponding to the target iron loss value is used as the screening furnace, and the process production parameters of the screening furnace are clustered to obtain the parameter interval of each type of parameter data set; the iron loss prediction model construction module is used to train the preset prediction model based on the training sample to obtain the final prediction model; the optimal production parameter solution module is used to use the differential evolution algorithm to calculate the iron loss prediction value in the parameter interval of each type of parameter data set, and the process production parameter corresponding to the minimum iron loss prediction value is used as the selected production parameter.

[0142] The following describes an apparatus embodiment of the present application, which can be used to execute the method for selecting production parameters of oriented silicon steel in the above-mentioned embodiment of the present application. For details not disclosed in the apparatus embodiment of the present application, please refer to the embodiment of the method for selecting production parameters of oriented silicon steel in the above-mentioned embodiment of the present application.

[0143] Figure 6is a block diagram of a device for selecting production parameters of oriented silicon steel, shown in an exemplary embodiment of the present application. The device can be applied to Figure 1 The implementation environment shown in the figure is specifically configured in the computer device 102. The device can also be applied to other exemplary implementation environments and specifically configured in other devices. This embodiment does not limit the implementation environment to which the device is applied.

[0144] like Figure 6 As shown, the exemplary device for selecting production parameters of oriented silicon steel comprises:

[0145] The parameter acquisition module 601 is used to acquire the process production parameters and iron loss values ​​of the oriented silicon steel in the historical production process.

[0146] The parameter screening module 602 is used to calculate the variance of the process production parameters, the iron loss value and the correlation between the process production parameters, and screen the process production parameters based on the variance of the process production parameters, the iron loss value and the correlation between the process production parameters to obtain the target production parameters.

[0147] The model training module 603 is used to train the preset prediction model using the target production parameters and iron loss values ​​as training samples to obtain the final prediction model.

[0148] The iron loss value screening module 604 is used to compare the iron loss value with the preset iron loss threshold, and screen the iron loss value according to the comparison result to obtain the target iron loss value; according to the preset number of categories, the target production parameters associated with the target iron loss value are clustered to obtain multiple parameter data sets.

[0149] The parameter selection module 605 is used to calculate multiple iron loss prediction values ​​based on each parameter data set and the final prediction model, and use the process production parameter corresponding to the minimum iron loss prediction value as the selected production parameter.

[0150] In one embodiment of the present application, the historical production process is characterized by different furnaces, the process production parameters in the historical production process include the process production parameters of different furnaces, the iron loss values ​​in the historical production process include the iron loss values ​​of different furnaces, the process production parameters include the production parameters of the normalizing and pickling processes during the production of oriented silicon steel, the production parameters of the cold rolling process during the production of oriented silicon steel, and the production parameters of the annealing process during the production of oriented silicon steel, etc.

[0151] In this embodiment, the process production parameters of all furnaces may be stored in the form of a matrix, and the size of the matrix is ​​n*J, wherein n represents the number of historical furnaces, and J represents the total number of categories of the process production parameters.

[0152] In this embodiment, taking the j-th category process production parameter as an example, the variance calculation formula of the j-th category process production parameter is shown in formula (1), and no specific limitation is given here.

[0153] In this embodiment, the process production parameters are screened based on the variance of the process production parameters and the correlation between the iron loss value and the process production parameters, and the process production parameters with little influence on the iron loss value are eliminated.

[0154] In this embodiment, the preset prediction model includes a preset ridge regression model and a preset random forest model. In the process of training the preset prediction model through training samples, first, the regression parameters, the attributes of the splitting of each node of each tree, and the training samples that finally fall into the leaf nodes are adjusted, and then the preset ridge parameters and the preset hyperparameters are adjusted to obtain the final prediction model, thereby improving the accuracy of the final prediction model in predicting the iron loss value.

[0155] In this embodiment, the iron loss values ​​are screened to obtain a smaller iron loss value, and based on the smaller iron loss value, the target production parameters associated with the target iron loss value are clustered to obtain multiple parameter data sets. The number of preset categories can be set according to actual conditions, and will not be described in detail here.

[0156] In this embodiment, the target iron loss value includes iron loss values ​​of multiple furnaces, and the target production parameters associated with the target iron loss value are clustered to obtain multiple parameter data sets. The process includes: determining the furnace corresponding to the target iron loss value, and using the furnace corresponding to the target iron loss value as the screening furnace, extracting the process production parameters of the screening furnace from the target production parameters, and clustering the process production parameters of the screening furnace to obtain multiple parameter data sets.

[0157] In this embodiment, the iron loss values ​​and process production parameters belonging to the same heat have an associated relationship or a corresponding relationship. The method for clustering the process production parameters of the target heat adopts a kmeans clustering algorithm or the like.

[0158] In this embodiment, based on each parameter data set and the final prediction model, multiple iron loss prediction values ​​are calculated, and the process production parameters corresponding to the minimum iron loss prediction value are used as the selected production parameters, thereby realizing the selection of better process production parameters. When the selected process parameters are applied to the subsequent production process of oriented silicon steel, the production quality of oriented silicon steel products and the stability of the production process can be improved.

[0159] It should be noted that the device for selecting production parameters of oriented silicon steel provided in the above embodiment and the method for selecting production parameters of oriented silicon steel provided in the above embodiment belong to the same concept, wherein the specific manner in which each module and unit performs the operation has been described in detail in the method embodiment and will not be repeated here. In actual application, the device for selecting production parameters of oriented silicon steel provided in the above embodiment can allocate the above functions to different functional modules as needed, that is, divide the internal structure of the device into different functional modules to complete all or part of the functions described above, and this is not limited here.

[0160] An embodiment of the present application also provides an electronic device, comprising: one or more processors; a storage device for storing one or more programs, and when the one or more programs are executed by one or more processors, the electronic device implements the method for selecting production parameters of oriented silicon steel provided in the above-mentioned embodiments.

[0161] Figure 7 The structure diagram of the computer system suitable for implementing the electronic device of the embodiment of the present application is shown. It should be noted that: Figure 7 The computer system 700 of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0162] like Figure 7 As shown, the computer system 700 includes a central processing unit (CPU) 701, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 702 or the program loaded from the storage part 708 to the random access memory (RAM) 703, such as executing the method in the above embodiment. In the RAM 703, various programs and data required for system operation are also stored. The CPU 701, ROM 702 and RAM 703 are connected to each other through a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0163] The following components are connected to the I / O interface 705: an input section 706 including a keyboard, a mouse, etc.; an output section 707 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker; a storage section 708 including a hard disk, etc.; and a communication section 709 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as needed. A removable medium 711, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 710 as needed so that a computer program read therefrom is installed into the storage section 708 as needed.

[0164] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication section 709, and / or installed from a removable medium 711. When the computer program is executed by a central processing unit (CPU) 701, various functions defined in the system of the present application are executed.

[0165] It should be noted that the computer-readable medium shown in the embodiment of the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, wherein a computer-readable computer program is carried. This propagated data signal can take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. A computer program contained on a computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0166] The flowchart and block diagram in the accompanying drawings illustrate the possible architecture, functions and operations of the system, method and computer program product according to various embodiments of the present application. Wherein, each box in the flowchart or block diagram can represent a module, a program segment, or a part of the code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0167] The units involved in the embodiments described in this application may be implemented by software or hardware, and the units described may also be set in a processor. The names of these units do not, in some cases, constitute limitations on the units themselves.

[0168] Another aspect of the present application provides a computer-readable storage medium on which computer-readable instructions are stored. When the computer-readable instructions are executed by a processor of a computer, the computer is caused to execute the method for selecting production parameters of oriented silicon steel provided in the above embodiments. The computer-readable storage medium may be included in the electronic device described in the above embodiments, or may exist independently without being assembled into the electronic device.

[0169] It should be noted that, although several modules or units of the equipment for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present application, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into being embodied by multiple modules or units.

[0170] Through the description of the above implementation methods, it is easy for those skilled in the art to understand that the example implementation methods described here can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution according to the implementation methods of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a touch terminal, or a network device, etc.) to execute the method according to the implementation methods of the present application.

[0171] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the embodiments disclosed herein. The present application is intended to cover any variations, uses or adaptations of the present application, which follow the general principles of the present application and include common knowledge or customary technical means in the art that are not disclosed in the present application.

[0172] It should be understood that the above content is only a preferred exemplary embodiment of the present application and is not intended to limit the implementation scheme of the present application. Ordinary technicians in this field can easily make corresponding changes or modifications based on the main concept and spirit of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection required by the claims.

Claims

1. A method for selecting production parameters of oriented silicon steel, characterized in that: include: Obtain the process production parameters and iron loss values ​​of oriented silicon steel in the historical production process; Calculating the variance of the process production parameters, the correlation between the iron loss value and the process production parameters, and screening the process production parameters based on the variance of the process production parameters, the correlation between the iron loss value and the process production parameters to obtain target production parameters; Using the target production parameters and the iron loss value as training samples to train a preset prediction model to obtain a final prediction model; Comparing the iron loss value with a preset iron loss threshold, and screening the iron loss value according to the comparison result to obtain a target iron loss value; Clustering target production parameters associated with the target iron loss value according to a preset number of categories to obtain multiple parameter data sets; Based on each of the parameter data sets and the final prediction model, multiple iron loss prediction values ​​are calculated, and the process production parameters corresponding to the minimum iron loss prediction value are used as the selected production parameters.

2. The method for selecting production parameters of oriented silicon steel according to claim 1, characterized in that: If the process production parameter has multiple categories, the process of screening the process production parameter in combination with the variance of the process production parameter and the correlation between the iron loss value and the process production parameter includes: Select a parameter class as the target class; If the process production parameters of the target category meet the parameter screening conditions, the process production parameters of the target category are deleted from the process production parameter set to obtain an updated production parameter set; the parameter screening conditions include that the variance of the process production parameters of the target category is less than a preset variance threshold, or that the correlation between the iron loss value and the process production parameters of the target category is less than a preset correlation threshold; the process production parameter set is composed of the process production parameters in the historical production process; the correlation between the iron loss value and the process production parameters of the target category is characterized by the Pearson correlation coefficient between the iron loss value and the process production parameters of the target category and the Spearman correlation coefficient between the iron loss value and the process production parameters of the target category; Continue to select another parameter category as the update category; If the process production parameters of the updated category meet the parameter screening condition, the process production parameters of the updated category are deleted from the updated production parameter set until all process production parameters are screened to obtain the target production parameters.

3. The method for selecting production parameters of oriented silicon steel according to claim 1 or 2, characterized in that: The process of training the preset prediction model by using the training samples to obtain the final prediction model includes: Selecting a set of hyperparameters from a preset prediction model hyperparameter set as a target set of hyperparameters, wherein the preset prediction model hyperparameter set is composed of multiple sets of preset prediction model hyperparameters, wherein the preset prediction model hyperparameters are composed of preset ridge parameters and preset random forest model hyperparameters, and wherein the preset random forest model hyperparameters include the number of trees, the depth of the tree, and the number of samples per leaf node; Constructing a preset ridge regression model with the preset ridge parameters in the target group hyperparameters, and constructing a preset random forest model with the preset random forest model hyperparameters in the target group hyperparameters; Dividing the training samples into a training data set and a validation data set, training the preset ridge regression model using the training data set to obtain a trained ridge regression model; Inputting the target production parameter in the training data set into the trained ridge regression model to obtain a model prediction value, and taking the difference between the iron loss value in the training data set and the model prediction value as a residual; Using the target production parameters of the training data set and the residual as residual samples, training the preset random forest model to obtain a trained random forest model; The sum of the trained ridge regression model and the trained random forest model is used as an iron loss prediction model; the validation data set is input into the iron loss prediction model to obtain a prediction error corresponding to the target group hyperparameter; Continue to select other groups of hyperparameters from the preset prediction model hyperparameter set as update group hyperparameters, and re-obtain the iron loss prediction model based on the update group hyperparameters and the training data set; input the verification data set into the re-obtained iron loss prediction model to obtain the prediction error corresponding to the update group hyperparameter, until the prediction error corresponding to all groups of hyperparameters is obtained; The minimum prediction error among the prediction errors corresponding to all groups of hyperparameters is selected, and the iron loss prediction model that calculates the minimum prediction error is used as the final prediction model.

4. The method for selecting production parameters of oriented silicon steel according to claim 1 or 2, characterized in that: If the historical production process is characterized by different heats, the iron loss value is compared with a preset iron loss threshold, and the process of screening the iron loss value according to the comparison result includes: Select a heat as the target heat; If the iron loss value of the target heat is less than the preset iron loss threshold, the iron loss value is used as the screening iron loss value; Continue to select another heat as the updated heat; If the iron loss value of the updated furnace is less than the preset iron loss threshold, the iron loss value is used as a newly added screening iron loss value until the iron loss values ​​of all furnaces are screened, and all the screened iron loss values ​​are used as the target iron loss values.

5. The method for selecting production parameters of oriented silicon steel according to claim 1 or 2, characterized in that: Based on each of the parameter data sets and the final prediction model, the process of calculating a plurality of iron loss prediction values ​​comprises: Obtaining a parameter interval of each parameter data set; and randomly generating a parameter population within each parameter interval to obtain a plurality of parameter populations, each of which includes a first preset number of parameter vectors; Selecting a parameter population as a target parameter population, inputting each parameter vector in the target parameter population into the final prediction model to obtain a plurality of fitness function values; and recording the number of population iterations; If the number of iterations of the population is greater than or equal to the preset number of iterations, the iterative operation process of the target parameter population is stopped, and other parameter populations are continuously selected and iterative operations are performed on the other parameter populations until the iterative operation process of all parameter populations is completed, and all the obtained fitness function values ​​are used as the multiple iron loss prediction values; If the population iteration number is less than the preset iteration number, then by performing mutation, crossover and selection operations on each parameter vector in the target parameter population, an updated parameter population and a newly added fitness function value are obtained, and the population iteration number is updated until the updated population iteration number is greater than or equal to the preset iteration number threshold; If the number of iterations of the updated population is greater than or equal to the preset number of iterations, the iterative calculation process of the target parameter population is stopped, and other parameter populations are continuously selected and iterative calculations are performed on the other parameter populations until the iterative calculation process of all parameter populations is completed, and all the fitness function values ​​obtained are used as multiple iron loss prediction values.

6. The method for selecting production parameters of oriented silicon steel according to claim 5, characterized in that: The process of performing mutation, crossover and selection operations on each parameter vector in the target parameter population includes: Selecting a parameter vector as a target parameter vector; and randomly selecting a second preset number of parameter vectors from the target parameter population, wherein the second preset number of parameter vectors is different from the target parameter vector; Performing a mutation operation on the parameter vectors of the second preset number to obtain a mutation vector; wherein the second preset number is smaller than the first preset number; The target parameter vector is used as an initial vector, and a cross operation is performed on components of the initial vector and components of the variation vector to obtain an operation vector; Inputting the operation vector into the final prediction model to obtain a first fitness function value; and inputting the target parameter vector into the final prediction model to obtain a second fitness function value; If the first fitness function value is less than the second fitness function value, the target parameter vector is replaced by the operation vector to obtain the parameter population for the initial update; and other parameter vectors in the parameter population for the initial update are selected as update parameter vectors, and the update parameter vectors are mutated, crossed and selected until the mutation, crossover and selection operations are completed for all parameter vectors in the target parameter population, and the parameter population finally updated is used as the update parameter population, and all the first fitness function values ​​are used as the newly added fitness function values; If the first fitness function value is greater than or equal to the second fitness function value, the target parameter vector is not updated, and other parameter vectors in the target parameter population are selected as update parameter vectors, and the update parameter vectors are mutated, crossover and selected until the mutation, crossover and selection operations on all parameter vectors are completed to obtain the newly added fitness function value and the update parameter population.

7. The method for selecting production parameters of oriented silicon steel according to claim 1 or 2, characterized in that: The process production parameters at least include: process production parameters of the normalizing pickling unit, process production parameters of the rolling mill and process production parameters of the continuous annealing unit.

8. A device for selecting production parameters of oriented silicon steel, characterized in that: include: A parameter acquisition module is used to obtain the process production parameters and iron loss values ​​of oriented silicon steel in the historical production process; A parameter screening module, used for calculating the variance of the process production parameters, the correlation between the iron loss value and the process production parameters, and screening the process production parameters based on the variance of the process production parameters, the correlation between the iron loss value and the process production parameters to obtain target production parameters; A model training module, used to train a preset prediction model using the target production parameter and the iron loss value as training samples to obtain a final prediction model; An iron loss value screening module is used to compare the iron loss value with a preset iron loss threshold value, and screen the iron loss value according to the comparison result to obtain a target iron loss value; Clustering target production parameters associated with the target iron loss value according to a preset number of categories to obtain multiple parameter data sets; A parameter selection module is used to calculate multiple iron loss prediction values ​​based on each of the parameter data sets and the final prediction model, and use the process production parameters corresponding to the minimum iron loss prediction value as the selected production parameters.

9. An electronic device, characterized in that: include: one or more processors; A storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, enables the electronic device to implement the method for selecting production parameters of oriented silicon steel as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: Computer-readable instructions are stored thereon, and when the computer-readable instructions are executed by a processor of a computer, the computer is enabled to execute the method for selecting production parameters of oriented silicon steel according to any one of claims 1 to 7.

Images