A method, apparatus, electronic device and storage medium for optimizing steelmaking process parameters

Abstract

The present application provides a method, device, electronic device and storage medium for optimizing steel process parameters. The method comprises obtaining an original data set in a steel industry scenario, wherein the original data set comprises an index to be optimized and multiple process parameters, determining an optimization target for the index to be optimized, clustering the data of each process parameter in the original data set to obtain a discrete data set of multiple process parameters, scanning the discrete data set row by row, and screening to obtain multiple frequent process parameter combinations based on the scanning results, extracting a sub-data set corresponding to each frequent process parameter combination from the original data set, calculating an optimization score for each sub-data set based on the optimization target and the index to be optimized, and selecting a target process parameter combination from multiple frequent process parameter combinations based on the optimization score. The method solves the technical problems of poor process parameter optimization effect and long time consumption in related technologies.

Description

Technical Field

[0001] The present application relates to the technical field of steel data processing, and in particular to a steel process parameter optimization method, device, electronic equipment and storage medium. Background Art

[0002] In the steel industry, there are many indicators that measure the quality of steel intermediates or final products. For example, the mechanical properties of steel produced by ESP production lines or hot rolling production lines (yield strength, tensile strength, etc.), the drum index that measures the mechanical strength of sintered ore during the sintering process, and the iron loss that measures the performance of silicon steel in the cold rolling production line. These indicators are affected by process operating parameters such as temperature adjustment, pulling speed, and stress conditions in the steel production process. Therefore, workers adjust process parameters based on process experience to achieve the purpose of improving the quality of steel products.

[0003] However, the adjustment of process parameters based solely on process experience is rough, with only a general adjustment direction but no way to determine the adjustment size, resulting in an insignificant improvement effect. At the same time, it is impossible to flexibly adjust the process parameters for specific situations, and it cannot adapt to the ever-changing working conditions in the steel scene. To address these problems, most of the related technologies are based on deep learning methods to build accurate indicator prediction models, which have a strong dependence on data quality and data volume. In the steel industry scenario, due to the long production process, complex process parameters, and difficulties in process parameter monitoring and performance testing, there is an impact on both data volume and data quality. If the inaccurate indicator prediction model continues to be used in the subsequent optimization algorithm, it will cause errors to accumulate, resulting in poor process optimization results. In addition, both the prediction model construction and the optimization algorithm optimization steps are time-consuming, and become more serious with the increase in data volume. It cannot be applied to real-time production processes that require rapid process parameter recommendation adjustments. Summary of the invention

[0004] In view of the shortcomings of the related technologies mentioned above, the present application provides a steel process parameter optimization method, device, electronic device and storage medium to solve the technical problems of poor process parameter optimization effect and long time consumption in the related technologies.

[0005] The present application provides a steel process parameter optimization method, which includes: obtaining an original data set in a steel industry scenario, the original data set including an indicator to be optimized and multiple process parameters, determining an optimization target for the indicator to be optimized, the optimization target including fixed value optimization, interval range optimization, and maximum and minimum value optimization; performing a clustering operation on the data of each process parameter in the original data set to obtain a discrete data set of multiple process parameters, scanning the discrete data set row by row, and constructing a process parameter search tree based on the scanning results, and screening multiple frequent process parameter combinations from the discrete data set based on a preset support threshold and the process parameter search tree; extracting a sub-data set corresponding to each frequent process parameter combination from the original data set, and calculating the optimization score of each of the sub-data set based on the optimization target and the indicator to be optimized, and selecting a target process parameter combination from the multiple frequent process parameter combinations based on the optimization score.

[0006] In one embodiment of the present application, before clustering the data of each process parameter in the original data set, it also includes: calculating the correlation coefficient between each process parameter and the correlation coefficient between each process parameter and the indicator to be optimized; if the correlation coefficient between two process parameters exceeds a preset correlation threshold, the process parameter with a lower correlation coefficient between the two process parameters and the indicator to be optimized is eliminated.

[0007] In one embodiment of the present application, a clustering operation is performed on the data of each process parameter in the original data set to obtain a discrete data set of multiple process parameters, including: calculating local density and relative distance based on the data of the process parameters; performing density sorting on all the data of the process parameters according to the local density, and selecting potential cluster center points from all the data based on the density sorting; taking the potential cluster center points whose relative distances are greater than a preset distance threshold as cluster centers, and performing data clustering based on the cluster centers to obtain a cluster data set corresponding to the process parameters; and obtaining a discrete data set of the multiple process parameters based on the cluster data set of the multiple process parameters, wherein each column of the discrete data set represents the same type of process parameter.

[0008] In one embodiment of the present application, the discrete data set is scanned row by row, and a process parameter search tree is constructed based on the scanning results, and multiple frequent process parameter combinations are obtained from the discrete data set based on a preset support threshold and the process parameter search tree, including: scanning the discrete data set row by row, sorting each item of all rows according to the frequency of occurrence, and filtering out the items whose occurrence frequency is lower than the preset support threshold, and constructing the process parameter search tree according to the sorted results; and obtaining multiple frequent process parameter combinations from the discrete data set based on the process parameter search tree, wherein the frequent process parameters are items whose occurrence frequency is greater than or equal to the preset support threshold.

[0009] In one embodiment of the present application, if the optimization target is fixed value optimization, the optimization score of each of the sub-data sets is calculated based on the optimization target and the indicator to be optimized, and the target process parameter combination is selected from the multiple frequent process parameter combinations based on the optimization score, including: calculating the difference between the mean of the optimization indicator in each of the sub-data sets and a preset fixed value; using the difference as the optimization score corresponding to each of the sub-data sets; and using the frequent process parameter combination corresponding to the sub-data set with the lowest optimization score as the target process parameter combination.

[0010] In one embodiment of the present application, if the optimization target is interval range optimization, the optimization score of each of the sub-data sets is calculated based on the optimization target and the indicator to be optimized, and the target process parameter combination is selected from the multiple frequent process parameter combinations based on the optimization score, including: if there is an optimization indicator in the sub-data set within the preset interval, the optimization score of the sub-data set is the preset maximum score; if there is an optimization indicator in the sub-data set outside the preset interval, the optimization score of the sub-data set is determined according to the difference between the optimization indicator and the maximum value or minimum value of the preset interval, the larger the difference, the lower the optimization score; based on all the optimization scores of each sub-data set, the optimization score corresponding to each sub-data set is obtained; the frequent process parameter combination corresponding to the sub-data set with the highest optimization score is used as the target process parameter combination.

[0011] In one embodiment of the present application, if the optimization goal is maximum and minimum value optimization, the optimization score of each of the sub-data sets is calculated based on the optimization goal and the indicator to be optimized, and the target process parameter combination is selected from the multiple frequent process parameter combinations based on the optimization score, including: calculating the mean of the optimization indicators in each of the sub-data sets; using the mean as the optimization score corresponding to each of the sub-data sets; and using the frequent process parameter combination corresponding to the sub-data set with the highest optimization score or the lowest optimization score as the target process parameter combination.

[0012] Beneficial effects of the present application: The embodiments of the present application provide a method, device, electronic device and storage medium for optimizing steel process parameters. The method obtains an original data set in a steel industry scenario, wherein the original data set includes an indicator to be optimized and multiple process parameters, determines the optimization target of the indicator to be optimized, clusters the data of each process parameter in the original data set to obtain a discrete data set of multiple process parameters, scans the discrete data set line by line, and obtains multiple frequent process parameter combinations based on the scanning results, extracts a sub-data set corresponding to each frequent process parameter combination from the original data set, calculates the optimization score of each sub-data set based on the optimization target and the indicator to be optimized, and selects a target process parameter combination from multiple frequent process parameter combinations based on the optimization score. By clustering and screening the process parameters, a target process parameter combination with high credibility and strong operability can be selected, thereby improving the optimization effect of the process parameters on the optimization indicators and reducing the optimization time. In addition, the present application includes multiple optimization targets, so that the optimization method of the present application can be applied to different process parameters and industrial scenarios, and has a wide applicability.

[0013] 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

[0014] Figure 1 is a schematic diagram of an implementation environment of a steel process parameter optimization method shown in an exemplary embodiment of the present application;

[0015] Figure 2 is a flow chart of a method for optimizing steel process parameters shown in an exemplary embodiment of the present application;

[0016] Figure 3 is a steel process parameter mining flow chart shown in an exemplary embodiment of the present application;

[0017] Figure 4 is a block diagram of a steel process parameter optimization device shown in an exemplary embodiment of the present application;

[0018] Figure 5 It is a structural schematic diagram of an electronic device shown in an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0019] The following describes the embodiments of the present application through specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. The present application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0020] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present application, and thus the drawings only show components related to the present application rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed at will, and the component layout may also be more complicated.

[0021] It should be noted that in this application, "first", "second", etc. are only used to distinguish similar objects, and are not used to limit the order or precedence of similar objects. The variations of "including", "having", etc. described above indicate that the scope covered by the subject of the word is not exclusive except for the examples shown by the word.

[0022] It is understood that the various numbers, step numbers, and other reference numerals recorded in this application are distinguished for the convenience of description and are not intended to limit the scope of this application. The size of the reference numerals in this application does not mean the order of execution. The execution order of each process should be determined by its function and internal logic.

[0023] In the following description, a large number of details are discussed to provide a more thorough explanation of the embodiments of the present application. However, it is obvious to those skilled in the art that the embodiments of the present application can be implemented without these specific details. In other embodiments, well-known structures and devices are shown in the form of block diagrams rather than in detail to avoid making the embodiments of the present application difficult to understand.

[0024] Most of the related technologies for optimizing process parameters are based on deep learning methods to build accurate indicator prediction models, which are highly dependent on data quality and data volume. In the steel industry scenario, due to the long production process, complex process parameters, and difficult process parameter monitoring and performance testing, both data volume and data quality are affected. If the inaccurate indicator prediction model continues to be used in the subsequent optimization algorithm, it will cause errors to accumulate, resulting in poor process optimization results. These prediction models based on deep learning methods generally have uncertainty and low generalization, and the optimization algorithm only optimizes the model prediction effect, which may cause the optimized process parameters to exceed the upper and lower limits of the process regulations on the production line, or the coupling relationship between the process parameters is disordered (for example, the IH outlet temperature in the ESP production line must be greater than the IH inlet temperature, etc.), resulting in the invalidity of the process parameter optimization scheme. In addition, the two steps of prediction model construction and optimization algorithm optimization involved in these process optimization algorithms are time-consuming, and become more serious with the increase of data volume, and cannot be applied to real-time production processes that require rapid process parameter recommendation and adjustment.

[0025] In view of the deficiencies of the above-mentioned related technologies, the embodiments of the present application respectively propose a steel process parameter optimization method, a steel process parameter optimization device, an electronic device, a computer-readable storage medium and a computer program product, which interactively analyze multiple processes and various steel industry scenarios such as the process cost of steel rolling, blast furnace conditions, and steel rolling quality, intelligently match process parameters, and assist process technicians in finding the optimal parameter range in historical production control. These embodiments will be described in detail below.

[0026] See also Figure 1 , Figure 1 It is a schematic diagram of an implementation environment of a steel process parameter optimization method shown as an exemplary embodiment of the present application.

[0027] like Figure 1As shown, the implementation environment may include a data acquisition device 101 and a computer device 102, wherein the computer device 102 may be at least one of a microcomputer, an embedded computer, a neural network computer, etc., and the data acquisition device 101 includes industrial data acquisition devices in various steel industry scenarios, including various sensors, such as temperature sensors, etc. The computer device 102 obtains the original data set in the steel industry scenario through the data acquisition device 101, the original data set includes the index to be optimized and multiple process parameters, determines the optimization target of the index to be optimized, performs clustering operation on the data of each process parameter in the original data set, obtains a discrete data set of multiple process parameters, scans the discrete data set line by line, and obtains multiple frequent process parameter combinations based on the scanning results, extracts a sub-data set corresponding to each frequent process parameter combination from the original data set, calculates the optimization score of each sub-data set based on the optimization target and the index to be optimized, and selects the target process parameter combination from the multiple frequent process parameter combinations based on the optimization score. After obtaining the target process parameter combination, the numerical value of the target process parameter can be adjusted to achieve accurate and rapid optimization of the index to be optimized.

[0028] See also Figure 2 , Figure 2 is a flowchart of a steel process parameter optimization method shown in an exemplary embodiment of the present application. The method can be applied to Figure 1 The implementation environment shown is as follows. The method may also be applicable to other exemplary implementation environments and be specifically executed by devices in other implementation environments. This embodiment does not limit the implementation environment to which the method is applicable.

[0029] like Figure 2 As shown, in an exemplary embodiment, the steel process parameter optimization method includes at least steps S210 to S240, which are described in detail as follows:

[0030] Step S210, obtaining an original data set in a steel industry scenario, the original data set including indicators to be optimized and multiple process parameters, determining optimization targets for the indicators to be optimized, the optimization targets including fixed value optimization, interval range optimization, and maximum and minimum value optimization.

[0031] In one embodiment of the present application, the data in the original data set is firstly subjected to outlier identification and screening, null value processing, and data cleaning assisted by process knowledge to ensure that the original data has a high quality for subsequent operations.

[0032] In one embodiment of the present application, data cleaning is performed on the original data set, including combining steel process knowledge and common sense to set the illegal values ​​in the original data set to null, such as the induction heating temperature is a negative number, and filtering out abnormal values ​​according to the coupling relationship between multiple process parameters, such as the induction heating outlet temperature needs to be greater than the inlet temperature, if it is less than, it is filtered out as an abnormal value. Then, after the null values ​​in the data are mean-filled, the DBSCAN algorithm is used to identify and remove outliers in the data. After the data cleaning operation, irrelevant noise or outliers in the data are avoided, and the data quality is improved.

[0033] In one embodiment of the present application, the indicators to be optimized are determined according to different steel industry scenarios. For example, in the ESP or hot-rolled steel industry scenario, the target to be optimized may be yield strength or tensile strength, and in the cold-rolled steel industry scenario, the target to be optimized may be iron loss, etc.

[0034] In one embodiment of the present application, after determining the indicator to be optimized, it is also necessary to determine the optimization target. If it is expected that the indicator to be optimized is as close to a certain value as possible after the process parameters are optimized, the optimization target is determined as fixed value optimization. If it is expected that the indicator to be optimized is as high as possible after the process parameters are optimized, or as low as possible, the optimization target is determined as maximum and minimum value optimization. If it is expected that the indicator to be optimized falls within a certain interval after the process parameters are optimized, the optimization target is determined as interval range optimization.

[0035] Step S220 , performing a clustering operation on the data of each process parameter in the original data set to obtain a plurality of discrete data sets of process parameters.

[0036] In one embodiment of the present application, a clustering operation is performed on the data of each process parameter in the original data set to obtain a discrete data set of multiple process parameters, including: calculating local density and relative distance based on the data of the process parameters; sorting all the data of the process parameters according to the local density, and selecting potential cluster centers from all the data based on the density sorting; taking the potential cluster centers whose relative distances are greater than a preset distance threshold as cluster centers, and clustering the data based on the cluster centers to obtain a cluster data set corresponding to the process parameters; and obtaining a discrete data set of multiple process parameters based on the cluster data set of multiple process parameters, wherein each column of the discrete data set represents the same type of process parameter.

[0037] In one embodiment of the present application, a density-based clustering algorithm (Density Peaks Clustering, DPC) clusters each process parameter to learn the setting mode of each process parameter. Compared with traditional clustering methods such as K-means and KNN, this clustering method does not require a specified number, and the algorithm adaptively identifies various clusters, which is more in line with actual production scenarios. Some process parameters have multiple set values, and some process parameters have only one or two set values. In the DPC clustering algorithm, two important values ​​need to be calculated: local density and relative distance. Local density measures the number of points within a certain distance around each data point, and relative distance measures the distance to the higher density point closest to the current point. Suppose the data of a process parameter in the screened process parameter set is S = {s1, s2, ..., s n}, n is the number of data for a process parameter. For each process parameter s i , first calculate its local density:

[0038]

[0039] In formula (1), ρ(i) represents the local density of the i-th data point, n represents the number of data, and d ij Represents the distance between the i-th data point and the j-th data point, d c is the cutoff distance.

[0040] Secondly, calculate its relative distance:

[0041]

[0042] In formula (3), δ(i) represents the local density of the i-th data point, d ij represents the distance between the i-th data point and the j-th data point, ρ(i) represents the local density of the i-th data point, and ρ(j) represents the local density of the j-th data point.

[0043] In one embodiment of the present application, for the cluster center point, it is expected that its local density and relative distance are both large. Therefore, all data points under the process parameters are sorted according to the local density, and the top 10% of the largest data points are selected as potential cluster centers. Among these potential cluster centers, a threshold of relative distance is set, and data points greater than the threshold are selected as cluster centers. The remaining data points are assigned to these cluster centers.

[0044] In one embodiment of the present application, assuming that a certain process parameter S i are clustered into three clusters, denoted as {S i1 , S i2 , S i3}. If S i1The maximum and minimum values ​​of all data points in the cluster are S i1_max , S i1_min , then S i1 It represents the process parameter S i In the interval [S i1_min , S i1_max ] range. Perform clustering operations on each process parameter separately, and finally the original data set can be transformed into a discrete data set corresponding to each process parameter, which is convenient for subsequent process parameter interval optimization.

[0045] Step S230 , scanning the discrete data set row by row, and constructing a process parameter search tree based on the scanning result, and screening multiple frequent process parameter combinations from the discrete data set based on a preset support threshold and the process parameter search tree.

[0046] In one embodiment of the present application, before clustering the data of each process parameter in the original data set, it also includes: calculating the correlation coefficient between each process parameter and the correlation coefficient between each process parameter and the indicator to be optimized; if the correlation coefficient between two process parameters exceeds a preset correlation threshold, the process parameter with a lower correlation coefficient with the indicator to be optimized is eliminated.

[0047] In one embodiment of the present application, the correlation coefficients between all process parameters are calculated first. If the correlation coefficients of two process parameters exceed 0.9, it proves that the information provided by the two process parameters is highly overlapping, so the one with the greater correlation with the indicator to be optimized is retained. In addition, the process parameters most relevant to the selected indicator can be selected by low variance screening, low contribution screening, Fisher score-based methods, etc., to preliminarily screen the process parameters and improve the efficiency and generalization ability of subsequent tasks.

[0048] In one embodiment of the present application, a discrete data set is scanned row by row, and a process parameter search tree is constructed based on the scanning results, and multiple frequent process parameter combinations are obtained from the discrete data set based on a preset support threshold and the process parameter search tree, including: scanning the discrete data set row by row, sorting each item of all rows according to the frequency of occurrence, and filtering out items with an occurrence frequency lower than the preset support threshold, and constructing the process parameter search tree according to the sorted results; and obtaining multiple frequent process parameter combinations from the discrete data set based on the process parameter search tree, wherein frequent process parameters are items with an occurrence frequency greater than or equal to the preset support threshold.

[0049] In one embodiment of the present application, n rows of data of a discrete data set can be expressed in the following form:

[0050] {S 11 , S23 , ..., S k3},

[0051] {S 12 , S 25 , ..., S k1},

[0052] …

[0053] {S 13 , S 22 , ..., S k2},

[0054] Where k represents the number of process parameter types after process parameter screening, each column represents the same process parameter, and n represents the number of data rows. For example, S in the first row of data k3 Indicates that the kth process parameter in the first row of data belongs to the third cluster. 13 It means that the first process parameter in the nth row belongs to the third cluster.

[0055] In one embodiment of the present application, the FP-growth (Frequent Pattern growth) algorithm is introduced to mine frequent item sets. The algorithm only needs to scan the data twice to find all frequent item sets, that is, frequent process parameter combinations. First, a conditional threshold, namely the support threshold, needs to be set. This threshold is used to help limit the frequent item sets mined to be supported by a certain amount of data, which is a credible rule rather than an accidental phenomenon. In implementation, it can be set to 5% of the amount of data in the data set. Each row in the discrete data set is regarded as a transaction. When the data is scanned for the first time, the algorithm will count the frequency of occurrence of each item in all transactions, sort them by frequency, and filter out items that do not meet the support threshold. Then, an FP tree is constructed based on the sorted results, and each transaction is converted into a path and added to the tree, and the same items share the path in the same tree. In the second scan, the algorithm creates a conditional FP tree for each frequent item, and mines out all frequent item sets containing the item. Similarly, in this step, it is still necessary to ensure that the frequent item sets meet the support threshold. Finally, all the mined frequent item sets are sorted in descending order of support as the mined frequent process parameter combinations.

[0056] Step S240, extracting a sub-dataset corresponding to each frequent process parameter combination from the original data set, calculating an optimization score for each sub-dataset based on the optimization target and the indicator to be optimized, and selecting a target process parameter combination from multiple frequent process parameter combinations based on the optimization score.

[0057] In one embodiment of the present application, it is assumed that the frequent process parameter combinations mined are {(S 23)、(S 23 , S 35 ),...,(S 23 , S 35 , S k2 )}, for each frequent process parameter combination, the corresponding sub-dataset is obtained. 23 , S 35 , S k2 ) as an example, the method for obtaining the sub-dataset corresponding to the frequent process parameter combination is as follows:

[0058] Find S 23 The corresponding interval [S 23 _ min , S 23 _ max ], and then select S2 process parameters in the original data set [S 23 _ min , S 23 _ max ] all data rows;

[0059] Based on the previous step, find S 35 The corresponding interval [S 35 _ min , S 35 _ max ], and then select S3 process parameters in [S 35 _ min , S 35 _ max ] all data rows;

[0060] Based on the previous step, find S k2 The corresponding interval [S k2 _ min , S k2 _ max ], and then filter S in the original data set k The process parameters are [S k2 _ min , S k2 _ max ] all data rows;

[0061] For each frequent process parameter combination, the corresponding sub-dataset is obtained according to the above steps. For each sub-dataset, the corresponding optimization score is calculated to indicate the optimization effect of the indicator to be optimized.

[0062] In one embodiment of the present application, if the optimization goal is fixed value optimization, the optimization score of each sub-data set is calculated based on the optimization goal and the indicator to be optimized, and the target process parameter combination is selected from the multiple frequent process parameter combinations based on the optimization score, including: calculating the difference between the mean of the optimization indicator in each sub-data set and the preset fixed value; using the difference as the optimization score corresponding to each sub-data set; and using the frequent process parameter combination corresponding to the sub-data set with the lowest optimization score as the target process parameter combination.

[0063] In one embodiment of the present application, if the optimization target is interval range optimization, the optimization score of each sub-data set is calculated based on the optimized target and the indicator to be optimized, and the target process parameter combination is selected from the multiple frequent process parameter combinations based on the optimization score, including: if there is an optimization indicator in the sub-data set within a preset interval, the optimization score of the sub-data set is a preset maximum score; if there is an optimization indicator in the sub-data set outside the preset interval, the optimization score of the sub-data set is determined according to the difference between the optimization indicator and the maximum or minimum value of the preset interval, the larger the difference, the lower the optimization score; based on all the optimization scores of each sub-data set, the optimization score corresponding to each sub-data set is obtained; the frequent process parameter combination corresponding to the sub-data set with the highest optimization score is used as the target process parameter combination.

[0064] In one embodiment of the present application, if the optimization goal is maximum and minimum value optimization, the optimization score of each sub-data set is calculated based on the optimization goal and the indicator to be optimized, and the target process parameter combination is selected from the multiple frequent process parameter combinations based on the optimization score, including: calculating the mean of the optimization indicator in each sub-data set; taking the mean as the optimization score corresponding to each sub-data set; and taking the frequent process parameter combination corresponding to the sub-data set with the highest optimization score or the lowest optimization score as the target process parameter combination.

[0065] In one embodiment of the present application, the optimization objectives include at least three types, namely fixed value optimization, interval range optimization and maximum and minimum value optimization. For each optimization objective, the calculation method is different. Assume that the list of indicators to be optimized for a sub-data set is T = {t1, t2, ..., t m}, the detailed calculation methods of the three optimization objectives are as follows:

[0066] In one embodiment of the present application, the fixed value optimization is to assume that the specified optimized fixed value (preset fixed value) is T select , then the optimization score of each sub-dataset is the difference between the mean of the indicator in the sub-dataset and T:

[0067] score=|mean(T)-T select | Formula (4)

[0068] In formula (4), mean(T) is the mean of the indicator to be optimized in the sub-dataset, T select is a preset fixed value, and score is the optimization score.

[0069] In one embodiment of the present application, interval range optimization assumes that the specified optimization target is the range [T min , T max ], then the optimization score of each sub-dataset is a comprehensive measure of the distance between the indicator to be optimized and the target range in the sub-dataset:

[0070]

[0071] In formula (5) and formula (6), m represents the amount of data in the sub-dataset, condition (t i , [T min , T max ]) represents the score of the i-th sample in the sub-dataset within the optimization target range. When the value of the index to be optimized of the i-th sample is within this range, the optimization score is maximized to 1 (preset maximum score); when the index to be optimized of the i-th sample is greater than T max or less than T min When the distance T max or T min The farther away, the lower the score, and the closer to the range, the closer the optimization score is to 1.

[0072] In one embodiment of the present application, the maximum and minimum value optimization is that when the indicator to be optimized is as large as possible or as small as possible, the optimization score of each sub-data set is the average value of the indicator in the sub-data set:

[0073] score=mean(T) Formula (7)

[0074] In formula (7), score represents the optimization score, and mean(T) represents the mean of the indicator to be optimized in the sub-dataset.

[0075] In one embodiment of the present application, after obtaining the optimization scores of the sub-data sets corresponding to each frequent process parameter combination, they can be sorted to obtain the final result. According to the definition of the above calculation method, it can be summarized that: when the optimization target is fixed value optimization, the frequent process parameter combination with the lowest optimization score is taken as the target process parameter combination as the final result output; when the optimization target is interval range optimization, the frequent process parameter combination with the highest optimization score is taken as the target process parameter combination as the final result output; when the optimization target is maximum and minimum value optimization, the frequent process parameter combinations with the highest and lowest optimization scores are taken as the target process parameter combination as the final result output.

[0076] See also Figure 3 , Figure 3 is a steel process parameter mining flow chart shown in an exemplary embodiment of the present application, according to Figure 3 As shown in the figure, a subject data set (original data set) related to the steel industry scenario is first prepared, and data cleaning tasks are performed through outlier identification and screening, null value processing, and process knowledge assistance to ensure that the subject data (original data set) has high quality for subsequent operations. Then determine the optimization target of the indicator to be optimized. The target has multiple modes such as fixed value optimization, interval range optimization, maximum and minimum value optimization, etc., and is personalized according to the specific needs of the indicator to be optimized. Through the specified optimization target, the process parameters most relevant to the target to be optimized are screened out by Fisher score, Pearson correlation coefficient sorting, etc. According to the screened process parameters, each process parameter is discretized according to the intelligent clustering of historical patterns to obtain a discrete data set corresponding to each process parameter, and the different control modes of each process parameter in the historical production process are simulated. According to the discrete data set, a process parameter search tree is constructed as the process parameter optimization space. On this process parameter search tree, various frequent process parameter combinations are searched, and the optimization effects (optimization scores) of the indicators to be optimized under various frequent process parameter combinations are calculated respectively. A set of frequent process parameter combinations with the best effect is selected as the optimal process parameter interval (target process parameter combination) and output.

[0077] In the embodiments of the present application, the process parameter discretization method based on DPC: Compared with equal spacing, equal frequency, traditional KNN and other clustering methods, the process parameter discretization method based on DPC is more suitable for the discretization task of process parameters. Because the adjustment of each process parameter in historical production has a process procedure setting, the use of equal spacing and equal frequency methods will lose these process procedure information. Traditional clustering algorithms such as KNN require the number of clusters to be specified in advance, but each process parameter has an indefinite number of process procedures. Therefore, the process parameter discretization method based on DPC can adaptively identify the number of process procedure information for each process parameter while retaining the process procedure information.

[0078] In an embodiment of the present application, a method for mining frequent process parameter combinations is performed based on a process parameter search tree based on FP-growth: in order to cope with the current situation of huge data volume in the steel industry, and the high computational cost of process optimization methods based on deep learning, a method is introduced that can quickly calculate and mine process parameter setting modes that frequently appear in historical production data, thereby greatly improving the operating efficiency of the model; at the same time, the process parameter combination results are directly mined from historical data, which not only ensures that the process parameters will not exceed the process regulations, but also conforms to the coupling relationship between the process parameters, with high credibility and strong operability.

[0079] In the embodiments of this application, the index optimization target is customized: three methods of customizing the index optimization target and detailed calculation methods are proposed, so that the algorithm can be widely adapted to different demand scenarios. When it is necessary to improve product quality, you can choose maximum and minimum value optimization; when it is necessary to avoid unqualified products, you can choose interval range optimization; when it is necessary to stabilize product quality, you can choose fixed value optimization.

[0080] See also Figure 4 , Figure 4 is a block diagram of a steel process parameter optimization device shown in an exemplary embodiment of the present application. The device can be applied to Figure 1 The implementation environment shown is as follows. The device may also be applicable to other exemplary implementation environments and specifically configured in other devices. This embodiment does not limit the implementation environment to which the device is applicable.

[0081] like Figure 4 As shown, the exemplary steel process parameter optimization device includes:

[0082] The parameter input module 401 is used to obtain the original data set in the steel industry scenario, the original data set includes the index to be optimized and multiple process parameters, and determine the optimization target of the index to be optimized, the optimization target includes fixed value optimization, interval range optimization and maximum and minimum value optimization;

[0083] The parameter clustering module 402 is used to perform a clustering operation on the data of each process parameter in the original data set to obtain a plurality of discrete data sets of process parameters.

[0084] The parameter screening module 403 is used to scan the discrete data set line by line, and build a process parameter search tree based on the scanning results, and screen multiple frequent process parameter combinations from the discrete data set based on a preset support threshold and the process parameter search tree;

[0085] The parameter optimization module 404 is used to extract the sub-dataset corresponding to each frequent process parameter combination in the original data set, calculate the optimization score of each sub-dataset based on the optimization target and the indicator to be optimized, and select the target process parameter combination from multiple frequent process parameter combinations based on the optimization score.

[0086] Figure 5 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 5 The computer system 500 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.

[0087] like Figure 5As shown, the computer system 500 includes a central processing unit (CPU) 501, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 502 or the program loaded from the storage part 508 to the random access memory (RAM) 503, such as executing the method described in the above embodiment. In the RAM 503, various programs and data required for system operation are also stored. The CPU 501, the ROM 502 and the RAM 503 are connected to each other through the bus 504. The input / output (I / O) interface 505 is also connected to the bus 504.

[0088] The following components are connected to the I / O interface 505: an input section 506 including a keyboard, a mouse, etc.; an output section 507 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the I / O interface 505 as needed. A removable medium 511, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 510 as needed so that a computer program read therefrom is installed into the storage section 508 as needed.

[0089] 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 509, and / or installed from a removable medium 511. When the computer program is executed by a central processing unit (CPU) 501, various functions defined in the system of the present application are executed.

[0090] 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.

[0091] 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.

[0092] 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.

[0093] Another aspect of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor of a computer, enables the computer to execute the steel process parameter optimization method as described above. The computer-readable storage medium may be included in the electronic device described in the above embodiment, or may exist independently without being assembled into the electronic device.

[0094] Another aspect of the present application also provides a computer program product or a computer program, the computer program product or the computer program includes computer instructions, the computer instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the steel process parameter optimization method provided in each of the above embodiments.

[0095] The above embodiments are merely illustrative of the principles and effects of the present application, and are not intended to limit the present application. Anyone familiar with the technology may modify or change the above embodiments without violating the spirit and scope of the present application. Therefore, all equivalent modifications or changes made by a person of ordinary skill in the art without departing from the spirit and technical ideas disclosed in the present application shall still be covered by the claims of the present application.

Claims

1. A method for optimizing steel process parameters, characterized in that: The steel process parameter optimization method comprises: Acquire an original data set in a steel industry scenario, the original data set including an indicator to be optimized and multiple process parameters, and determine an optimization target for the indicator to be optimized, the optimization target including fixed value optimization, interval range optimization, and maximum and minimum value optimization; Performing a clustering operation on the data of each process parameter in the original data set to obtain a plurality of discrete data sets of process parameters, Scanning the discrete data set row by row, and constructing a process parameter search tree based on the scanning result, and screening a plurality of frequent process parameter combinations from the discrete data set based on a preset support threshold and the process parameter search tree; A sub-dataset corresponding to each frequent process parameter combination is extracted from the original data set, and an optimization score of each sub-dataset is calculated based on the optimization target and the indicator to be optimized, and a target process parameter combination is selected from the multiple frequent process parameter combinations based on the optimization score.

2. The method for optimizing steel process parameters according to claim 1, characterized in that: Before performing a clustering operation on the data of each process parameter in the original data set, the method further includes: Calculating the correlation coefficients between each of the process parameters and the correlation coefficients between each of the process parameters and the index to be optimized; If the correlation coefficient between two process parameters exceeds a preset correlation threshold, the process parameter with a lower correlation coefficient between the two process parameters and the indicator to be optimized is eliminated.

3. The method for optimizing steel process parameters according to claim 1, characterized in that: A clustering operation is performed on the data of each process parameter in the original data set to obtain a plurality of discrete data sets of process parameters including: Calculating local density and relative distance based on the data of the process parameters; Sorting all the data of the process parameters according to the local density, and selecting a potential cluster center point from all the data based on the density sorting; Taking the potential cluster center point whose relative distance is greater than a preset distance threshold as the cluster center point, and performing data clustering based on the cluster center point to obtain a cluster data set corresponding to the process parameters; A discrete data set of the plurality of process parameters is obtained based on the clustered data set of the plurality of process parameters, wherein each column of the discrete data set represents the same type of process parameter.

4. The method for optimizing steel process parameters according to any one of claims 1 to 3, characterized in that: Scanning the discrete data set line by line, and constructing a process parameter search tree based on the scanning results, and screening multiple frequent process parameter combinations from the discrete data set based on a preset support threshold and the process parameter search tree, including: Scanning the discrete data set row by row, sorting each item of all rows according to the frequency of occurrence, filtering out the items whose frequency of occurrence is lower than the preset support threshold, and constructing the process parameter search tree according to the sorted results; Based on the process parameter search tree, a plurality of frequent process parameter combinations are screened from the discrete data set, where the frequent process parameters are items whose occurrence frequency is greater than or equal to the preset support threshold.

5. The method for optimizing steel process parameters according to any one of claims 1 to 3, characterized in that: If the optimization target is fixed value optimization, calculating the optimization score of each of the sub-data sets based on the optimization target and the index to be optimized, and selecting a target process parameter combination from the multiple frequent process parameter combinations based on the optimization score includes: Calculate the difference between the mean of the optimization index in each of the sub-data sets and a preset fixed value; Using the difference as the optimization score corresponding to each of the sub-data sets; The frequent process parameter combination corresponding to the sub-data set with the lowest optimization score is used as the target process parameter combination.

6. The method for optimizing steel process parameters according to any one of claims 1 to 3, characterized in that: If the optimization target is interval range optimization, calculating the optimization score of each of the sub-data sets based on the optimization target and the index to be optimized, and selecting a target process parameter combination from the multiple frequent process parameter combinations based on the optimization score includes: If there is an optimization index in the sub-data set within the preset interval, the optimization score of the sub-data set is within the preset maximum score; If there is an optimization index in the sub-dataset that is outside the preset interval, the optimization score of the sub-dataset is determined according to the difference between the optimization index and the maximum value or the minimum value of the preset interval, and the larger the difference, the lower the optimization score; Obtaining an optimization score corresponding to each of the sub-datasets based on all optimization scores of each of the sub-datasets; The frequent process parameter combination corresponding to the sub-data set with the highest optimization score is used as the target process parameter combination.

7. The method for optimizing steel process parameters according to any one of claims 1 to 3, characterized in that: If the optimization objective is maximum and minimum value optimization, calculating the optimization score of each of the sub-data sets based on the optimization objective and the index to be optimized, and selecting a target process parameter combination from the multiple frequent process parameter combinations based on the optimization score includes: Calculating the mean of the optimization index in each of the sub-data sets; Taking the mean as the optimization score corresponding to each of the sub-data sets; The frequent process parameter combination corresponding to the sub-data set with the highest optimization score or the lowest optimization score is taken as the target process parameter combination.

8. A steel process parameter optimization device, characterized in that: The steel process parameter optimization device comprises: A parameter input module is used to obtain an original data set in a steel industry scenario, wherein the original data set includes an indicator to be optimized and multiple process parameters, and determine an optimization target for the indicator to be optimized, wherein the optimization target includes fixed value optimization, interval range optimization, and maximum and minimum value optimization; A parameter clustering module is used to perform a clustering operation on the data of each process parameter in the original data set to obtain a discrete data set of multiple process parameters. A parameter screening module, configured to scan the discrete data set line by line, and construct a process parameter search tree based on the scanning result, and screen a plurality of frequent process parameter combinations from the discrete data set based on a preset support threshold and the process parameter search tree; A parameter optimization module is used to extract a sub-dataset corresponding to each frequent process parameter combination from the original data set, calculate the optimization score of each sub-dataset based on the optimization target and the indicator to be optimized, and select a target process parameter combination from the multiple frequent process parameter combinations based on the optimization score.

9. An electronic device, characterized in that: The electronic device comprises: 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 steel process parameter optimization method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor of a computer, the computer is enabled to execute the steel process parameter optimization method according to any one of claims 1 to 7.

Images