Steel billet production parameter adaptive adjustment method and system

Abstract

The present invention relates to the technical field of steel billet production, and in particular to a method and system for adaptively adjusting steel billet production parameters. The method comprises: setting production data observation points to obtain production process parameters, organizing the production process parameters to generate primitive parameter groups; calculating initial pheromone values ​​for each primitive parameter group, and obtaining heuristic factors based on the initial pheromone values ​​and the production process parameters; aggregating the initial pheromone values ​​to obtain pheromone concentrations, and using the pheromone concentrations and the heuristic factors as adaptive conditions to establish a production parameter adaptive model; and obtaining a final solution combination of production parameters based on the production parameter adaptive model. The present invention effectively solves the problems of low parameter optimization efficiency, insufficient real-time adjustment capabilities, and low utilization of intelligent experience in steel billet production.

Description

Technical Field

[0001] The present invention relates to the technical field of steel billet production, and in particular to a method and system for adaptively adjusting steel billet production parameters. Background Art

[0002] During the steel billet production process, process parameters (such as temperature, casting speed, and cooling rate) have a significant impact on product quality, production efficiency, and energy consumption. However, traditional parameter adjustment methods often rely on manual experience or static rules, which are difficult to adapt to the complexity and dynamic nature of production conditions. This leads to low adjustment efficiency, delayed response, and limited optimization results. Furthermore, when faced with multi-parameter, high-dimensional search spaces, traditional optimization methods suffer from low optimization efficiency, prone to local optimality, lack of dynamic adjustment capabilities, and inability to effectively leverage historical production experience. These factors make it difficult to meet the demands of modern steel manufacturing for efficient and intelligent production.

[0003] The information disclosed in this background technology section is only intended to deepen the understanding of the overall background technology of the present disclosure and should not be regarded as an admission or any form of suggestion that the information constitutes the prior art known to those skilled in the art. Summary of the Invention

[0004] The present invention provides a method and system for adaptively adjusting production parameters of steel billets, which can effectively solve the problems in the background technology.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is:

[0006] A method for adaptively adjusting production parameters of steel billets, the method comprising:

[0007] Setting production data observation points to obtain production process parameters, and arranging the production process parameters to generate an original parameter group;

[0008] Calculating an initial pheromone value of each of the original parameter groups, and obtaining a heuristic factor according to the initial pheromone value and the production process parameters;

[0009] Aggregating initial pheromone values ​​to obtain pheromone concentrations, and using the pheromone concentrations and the heuristic factors as adaptive conditions to establish a production parameter adaptive model;

[0010] According to the production parameter adaptive model, a final solution combination of production parameters is obtained.

[0011] Furthermore, the final solution combination of production parameters is obtained, including:

[0012] Establishing parameter combination candidate solutions based on the initial pheromone values ​​and heuristic factors of each of the original parameter groups;

[0013] Each parameter combination candidate solution determines a production parameter combination path according to the corresponding pheromone concentration and heuristic factor;

[0014] Calculating the fitness value for each of the production parameter combination paths, and preferentially exploring the production parameter combination path with the highest fitness value at the same level;

[0015] Iteratively executing the selection of the production parameter combination path and the calculation of the fitness value until a fitness value convergence condition is met;

[0016] At the end of the iteration, the parameter combination candidate solution included in the production parameter combination path with the highest fitness value is selected to generate the production parameter final solution combination.

[0017] Furthermore, iteratively performing the selection of the production parameter combination path includes:

[0018] updating the pheromone concentration on the production parameter combination path according to the fitness value;

[0019] Comparing the fitness of each of the production parameter combination paths, increasing the pheromone concentration corresponding to a high fitness value, and gradually evaporating the pheromone concentration corresponding to a low fitness value;

[0020] Correcting the pheromone concentration result of each production parameter combination path according to the heuristic factor reflecting the change trend of the production process parameters;

[0021] Detecting whether a local optimal solution appears, and if so, performing a perturbation operation on the production parameter combination path with similar and continuous changes in pheromone concentration;

[0022] After each iteration, the change of the fitness value of each production parameter combination path is monitored. When the change rate is lower than the fitness value convergence condition, the iteration process is terminated and the final solution combination of the production parameters is selected.

[0023] Furthermore, performing a perturbation operation on the production parameter combination path with similar and continuous changes in pheromone concentrations includes:

[0024] Setting a disturbance factor for each of the production parameter combination paths with similar and continuous changes in pheromone concentration, wherein the disturbance factor is set according to the stability of the changes in pheromone concentration;

[0025] Based on the disturbance factor, the production parameter combination path is offset and disturbed to generate a disturbance combination path;

[0026] adjusting the disturbance intensity of the disturbance combination path according to the fitness value, and reducing the disturbance intensity for the disturbance combination path with an improved fitness value;

[0027] The fitness value of the disturbance combination path is evaluated, and when the evaluation result is higher than the fitness value of the corresponding production parameter combination path, the path is replaced and the pheromone concentration is updated.

[0028] Furthermore, obtaining a heuristic factor according to the initial pheromone value and the production process parameter includes:

[0029] Analyzing the initial pheromone of the original parameter group under each production target, wherein the initial pheromone is captured and determined based on the impact of the original parameter group on the production target;

[0030] quantifying the initial pheromone to generate the initial pheromone value, and combining the original parameter group to generate an initial heuristic factor;

[0031] Obtaining a historical parameter group, and assigning corresponding influence weights to the initial heuristic factors based on similarities between the original parameter group and the historical parameter group;

[0032] Clustering the initial heuristic factors into groups, combining the initial heuristic factors with similar influence weights, and analyzing the similarity between the groups;

[0033] The initial heuristic factors are weighted according to the inter-group similarities, and the heuristic factors are obtained by integration based on the weighted correction results.

[0034] Furthermore, the initial heuristic factor is weighted according to the inter-group similarity, including:

[0035] For the initial heuristic factors within each cluster grouping, a similarity measure within the group is calculated, and the data distribution between the groups is adjusted according to the similarity measure;

[0036] Analyzing the difference in data distribution between different cluster groups, assigning comprehensive correction weights to the cluster groups with similarity matching according to the difference in similarity between the groups, and screening out cluster groups with low similarity;

[0037] Based on the difference results of the similarities between the groups and the comprehensive correction weight, a global weight correction is performed on the initial heuristic factor of each cluster grouping;

[0038] The global weight correction is performed by weighted combination of the comprehensive correction weight and the influence weight of each cluster group to calculate the final correction weight of each initial heuristic factor;

[0039] The final modified weight normalization results corresponding to the initial heuristic factors are integrated to obtain the heuristic factors.

[0040] Furthermore, the pheromone concentration and the heuristic factor are used as adaptive conditions to establish a production parameter adaptive model, including:

[0041] Based on the pheromone concentration and the heuristic factor, establishing an initial state of the production parameter adaptive model, and setting multiple parameter adjustment paths for the production parameter adaptive model;

[0042] Setting a pheromone concentration threshold, using a change in pheromone concentration as a self-adaptive trigger condition for the model, and adjusting the heuristic factor when the pheromone concentration reaches the pheromone concentration threshold;

[0043] According to the change of the heuristic factor, a plurality of the parameter adjustment paths are used to perform an adaptive search for the production process parameters, and calculate an optimized production path;

[0044] According to the optimized production path, the production parameter adaptive model is updated, and the pheromone concentration and the heuristic factor are adjusted synchronously;

[0045] Feedback verification is performed on the production parameter adaptive model, the pheromone concentration and the heuristic factor are recalculated based on the verification result, and the final production parameter adaptive model is generated after the system is stabilized.

[0046] Furthermore, feedback verification is performed on the production parameter adaptive model, including:

[0047] Collecting production effect data, generating a production effect data set, and establishing a production experience database based on the production effect data set;

[0048] Comparing and analyzing the production effect data set with the expected output of the production parameter adaptive model, and calculating the expected deviation;

[0049] According to the expected deviation, the corresponding heuristic factor is judged and the pheromone concentration is corrected to generate an expected correction result;

[0050] The production effect data is repeatedly collected to verify the stability of the production parameter adaptive model, and the parameter combination of the production parameter adaptive model that has passed the verification and the expected correction result are stored in the production experience database.

[0051] A steel billet production parameter adaptive adjustment system, the system comprising:

[0052] Parameter acquisition module, sets production data observation points to obtain production process parameters, and organizes the production process parameters to generate original parameter groups;

[0053] The information calculation module calculates the initial pheromone value of each original parameter group and obtains the heuristic factor based on the initial pheromone value and the production process parameters;

[0054] The model generation module aggregates the initial pheromone values ​​to obtain the pheromone concentration, and uses the pheromone concentration and heuristic factors as adaptive conditions to establish a production parameter adaptive model;

[0055] The parameter combination module obtains the final solution combination of production parameters based on the production parameter adaptive model.

[0056] Furthermore, the parameter combination module includes:

[0057] The parameter candidate unit establishes parameter combination candidate solutions based on the initial pheromone values ​​and heuristic factors of each original parameter group;

[0058] Path generation unit, each parameter combination candidate solution determines a production parameter combination path according to the corresponding pheromone concentration and heuristic factor;

[0059] The adaptation calculation unit calculates the fitness value for each production parameter combination path and prioritizes the production parameter combination path with the highest fitness value at the same level;

[0060] Iterative selection unit, iteratively executing the selection of production parameter combination path and the calculation of fitness value until the fitness value convergence condition is met;

[0061] The final solution generating unit selects the parameter combination candidate solution contained in the production parameter combination path with the highest fitness value to generate the final solution combination of production parameters at the end of the iteration.

[0062] The technical solution of the present invention can achieve the following technical effects:

[0063] It effectively solves the problems of low parameter optimization efficiency, insufficient real-time adjustment capabilities and low utilization of intelligent experience in steel billet production, accelerates the optimization process, and thus improves the quality, efficiency and cost control level of steel billet production.

[0064] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0066] Figure 1 A schematic flow chart of a method for adaptively adjusting production parameters of steel billets;

[0067] Figure 2 A schematic diagram of the structure for generating the final solution combination of production parameters;

[0068] Figure 3 Schematic diagram of the structure for generating heuristic factors;

[0069] Figure 4 Schematic diagram of the process flow generated for the production parameter adaptation model. DETAILED DESCRIPTION

[0070] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0071] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification are for the purpose of describing specific embodiments only and are not intended to limit the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0072] Example 1

[0073] like Figure 1 As shown, the present application provides a method for adaptively adjusting production parameters of steel billets, the method comprising:

[0074] S10: Setting production data observation points to obtain production process parameters, and arranging the production process parameters to generate an original parameter group;

[0075] S20: calculating the initial pheromone value of each original parameter group, and obtaining the heuristic factor according to the initial pheromone value and the production process parameters;

[0076] S30: Aggregating the initial pheromone values ​​to obtain the pheromone concentration, and using the pheromone concentration and the heuristic factor as adaptive conditions to establish a production parameter adaptive model;

[0077] S40: Obtaining a final solution combination of production parameters according to the production parameter adaptive model.

[0078] Specifically, production data observation points are set up, and sensors are arranged at key process nodes (such as casting, rolling, and cooling stages) in steel billet production to collect production process parameters, including temperature, drawing speed, cooling rate and other process parameters. The collected production process parameters are organized into original parameter groups, which are generally a multidimensional vector set. According to the performance of each original parameter group in historical production data, its contribution to the target production index (such as product quality, energy consumption, and efficiency) is calculated as the initial pheromone value. According to the initial pheromone value of the original parameter group, combined with the production conditions, a heuristic factor is generated. According to the similarity between the current production conditions and the historical best parameter combination, the weight of the initial pheromone value is adjusted. The parameter combinations with similar initial pheromone values ​​are grouped through a clustering algorithm, and the comprehensive analysis of each group is carried out. The heuristic factors are generated based on the characteristics, and the heuristic factors are normalized to ensure that the weights of different parameters are comparable; the initial pheromone values ​​of each original parameter group are aggregated to generate the overall pheromone concentration to represent the global adaptability of the parameter combination under the current production environment; the pheromone concentration and heuristic factors are used as input to establish a production parameter adaptation model, and the production parameter adaptation model is optimized based on the ant colony algorithm; through multiple iterative optimizations, under the condition of fitness convergence, the parameter combination with the highest fitness value is selected as the final solution combination of production parameters; the obtained final solution combination is applied in actual production, the production effect is monitored in real time, and the pheromone concentration and heuristic factors are dynamically adjusted according to the deviation of production data to ensure the robustness and adaptability of the production parameter model under different conditions.

[0079] The technical solution of the present invention effectively solves the problems of low parameter optimization efficiency, insufficient real-time adjustment capability and low utilization rate of intelligent experience in steel billet production, accelerates the optimization process, and thus improves the quality, efficiency and cost control level of steel billet production.

[0080] Further, if Figure 2 As shown, the final solution combination of production parameters is obtained, including:

[0081] Based on the initial pheromone values ​​and heuristic factors of each original parameter group, a candidate solution of parameter combination is established;

[0082] Each parameter combination candidate solution determines a production parameter combination path according to the corresponding pheromone concentration and heuristic factor;

[0083] Calculate the fitness value for each production parameter combination path, and prioritize exploring the production parameter combination path with the highest fitness value at the same level;

[0084] Iteratively execute the selection of production parameter combination paths and the calculation of fitness values ​​until the fitness value convergence condition is met;

[0085] At the end of the iteration, the parameter combination candidate solution contained in the production parameter combination path with the highest fitness value is selected to generate the final production parameter solution combination.

[0086] As a preferred embodiment of the above, based on the original parameter group, combined with the initial pheromone value and the heuristic factor, multiple parameter combination candidate solutions are generated. Each parameter combination candidate solution is composed of different production parameter combinations, specifically including process parameters such as temperature range, drawing speed range, cooling speed range, etc. Through the random initialization method, several initial parameter combinations are generated as candidate solutions in the parameter search space. The initial pheromone values ​​of the candidate solutions are assigned based on historical data and current production conditions. Each parameter combination candidate solution determines its path selection in the production process based on the corresponding pheromone concentration and heuristic factor. The probability of the path is determined by the following formula:

[0087]

[0088] in P ij Represents a slave node i Move to Node j The probability of τ ij Indicates the path ij The pheromone concentration on η ij Indicates the path ij The heuristic factor, α 、 β are the weight factors of pheromone concentration and heuristic factor respectively; for each production parameter combination path, the fitness value is calculated based on the fitness function, which is used to measure the performance of the path under the target production indicators (such as quality, efficiency, and energy consumption), and the production parameter combination path with the highest fitness value among the paths of the same level is explored first, and the path selection is guided by the high fitness value of the path; iterative optimization is performed, and the selection of production parameter combination paths and the calculation of fitness values ​​are repeated until the convergence conditions of the fitness values ​​are met. In each iteration, the pheromone concentration on the path is updated, so that the pheromone concentration of the high fitness path increases and the pheromone concentration of the low fitness path gradually evaporates, thereby strengthening the ant's exploration tendency for the optimal path; at the end of the iteration, the parameter combination candidate solution contained in the production parameter combination path with the highest fitness value is selected as the final production parameter final solution combination. The production parameter final solution combination includes the optimal process parameter settings, such as optimal temperature, pulling speed, cooling speed, etc.

[0089] More specifically, the selection of the production parameter combination path is performed iteratively, including:

[0090] Update the pheromone concentration on the production parameter combination path according to the fitness value;

[0091] The fitness of each production parameter combination path is compared, and the pheromone concentration corresponding to the high fitness value increases, while the pheromone concentration corresponding to the low fitness value gradually evaporates;

[0092] The pheromone concentration results of each production parameter combination path are modified according to the heuristic factors reflecting the changing trend of production process parameters;

[0093] Detect whether a local optimal solution appears. If so, perform a perturbation operation on the production parameter combination path with similar and continuous changes in pheromone concentrations.

[0094] After each iteration, the changes in the fitness values ​​of the production parameter combination paths are monitored. When the rate of change is lower than the fitness value convergence condition, the iteration process is terminated and the final solution combination of production parameters is selected.

[0095] As a preferred embodiment of the above embodiment, in each iteration, the pheromone concentration on the production parameter combination path is updated according to the fitness value of the production parameter combination path. For the production parameter combination path with a higher fitness value, the pheromone concentration is increased, thereby increasing the probability of the path being preferentially selected and strengthening the exploration of these paths. For the production parameter combination path with a lower fitness value, the pheromone concentration is gradually reduced to reduce the selection probability of these paths, thereby guiding more resources to be concentrated on potential high-quality paths; the heuristic factor is used to evaluate the closeness of the current path to the target optimization by analyzing the difference between the current production parameters and the historical optimal parameters. If the parameter combination of a path is similar to the optimal parameter combination in the historical data, its weight is increased and its pheromone concentration is further increased. Otherwise, the weight is reduced. Through this process, the pheromone concentration update result is corrected in combination with historical experience and real-time data. The results make the path selection more in line with production reality; during the iteration process, it is detected whether the changes in pheromone concentration tend to be continuous and similar (that is, the concentration changes of multiple paths are very small) to determine whether a local optimal solution appears. If the concentration changes of multiple paths remain slightly different, it means that the algorithm may fall into a local optimal solution. To avoid this problem, a perturbation operation is performed on the paths with continuous and similar changes in pheromone concentration, that is, a small range adjustment (offset) is made to the path parameter combination to generate a new path, and a small amount of pheromone is added to some low-concentration paths to break the current centralized trend, redistribute the exploration direction, and improve the global search capability; after each iteration, the fitness value changes of all paths are monitored to determine whether the convergence conditions are met. If the fitness value of the path changes very little in multiple iterations, it means that the path has stabilized and reached a convergence state. At this time, the iteration process is stopped, and the production parameter combination contained in the path with the highest fitness value is selected as the final optimization result, and this path is defined as the final solution combination of production parameters.

[0096] Furthermore, the perturbation operation is performed on the production parameter combination path with similar and continuous changes in pheromone concentration, including:

[0097] A disturbance factor is set for each production parameter combination path with similar and continuous changes in pheromone concentration. The disturbance factor is set according to the stability of the pheromone concentration changes.

[0098] Based on the disturbance factor, the production parameter combination path is offset and disturbed to generate a disturbance combination path;

[0099] Adjust the perturbation intensity of the perturbation combination path according to the fitness value. For the perturbation combination path with improved fitness value, reduce the perturbation intensity.

[0100] The fitness value of the disturbance combination path is evaluated. When the evaluation result is higher than the fitness value of the corresponding production parameter combination path, the path is replaced and the pheromone concentration is updated.

[0101] As a preferred embodiment of the above embodiment, a disturbance factor is set for each production parameter combination path with similar and continuous changes in pheromone concentration, so as to break the similarity of the production parameter combination paths. When the concentration changes of multiple production parameter combination paths tend to be consistent and the difference in fitness values ​​is small (that is, the path performance tends to be stable but has not reached the global optimum), the disturbance operation is triggered and the disturbance factor is set. The disturbance factor is dynamically set according to the stability of the pheromone concentration change. The higher the stability, the greater the intensity of the disturbance factor. Based on the disturbance factor, the production parameter combination paths with similar and continuous pheromone concentrations are offset and disturbed to generate a new disturbance combination path. On the basis of the original production parameter combination path, some parameter values ​​are randomly adjusted in a small range (such as increasing or decreasing the temperature by a small range, and the pulling speed fluctuating by a certain proportion), to generate a new production parameter combination path after disturbance. The random disturbance factor is introduced based on the offset of the production parameter combination path. The fitness value adjusts the disturbance intensity to optimize the disturbance efficiency; when the fitness value of the production parameter combination path after disturbance is higher than the fitness value of the original production parameter combination path, the disturbance intensity is reduced and the convergence is gradually achieved to avoid excessive random disturbance affecting the accuracy of the final solution; when the fitness value of the production parameter combination path after disturbance is lower than the fitness value of the original path, the disturbance intensity is increased, the search range is further expanded, and potential high-quality paths are explored; the fitness value of the generated disturbance combination path is evaluated and compared with the fitness value of the original path. Based on the production objectives (such as product quality, efficiency and energy consumption), it is evaluated whether the disturbance path can show better results in the objective function; if the fitness value of the disturbance path is higher than the corresponding original production parameter combination path, the original production parameter combination path is replaced with the disturbance combination path, and the pheromone concentration of the disturbance combination path is updated. The updated pheromone concentration will reflect the new fitness value to guide subsequent iterations to explore better paths.

[0102] Further, if Figure 3 As shown, the heuristic factors are obtained according to the initial pheromone values ​​and production process parameters, including:

[0103] Analyze the initial pheromone of the original parameter group under each production target. The initial pheromone is determined based on the impact of the original parameter group on the production target.

[0104] The initial pheromone is quantified to generate the initial pheromone value, and the initial heuristic factor is generated by combining the original parameter group;

[0105] Obtain the historical parameter group, and assign corresponding influence weights to the initial heuristic factors based on the similarity between the original parameter group and the historical parameter group;

[0106] Cluster the initial heuristic factors, combine the initial heuristic factors with similar influence weights, and analyze the similarity between groups;

[0107] The initial heuristic factors were weighted and modified according to the similarity between groups, and the heuristic factors were obtained by integrating the weighted modification results.

[0108] As a preferred embodiment of the above embodiment, by analyzing the degree of influence of the original parameter group (such as temperature, drawing speed, cooling speed, etc.) on different production targets (such as quality, efficiency, energy consumption), the initial pheromone of each original parameter group is determined, and the contribution of each parameter group to the production target is classified and captured. The more significant the impact of the parameter, the higher its initial pheromone value; conversely, the parameter with less impact corresponds to a lower pheromone value; the captured initial pheromone is quantified to generate an initial pheromone value in a unified format. During the quantification process, the positive and negative effects of the parameters on the production target need to be considered (for example, increasing the temperature may increase energy consumption but improve product quality). The initial heuristic factor of each parameter combination is calculated by combining the quantized initial pheromone value and the original parameter group. The heuristic factor is used to reflect the influence trend of the production parameter on the target optimization direction; the historical parameter group similar to the current production conditions is extracted. The historical parameter group records the parameter settings in the production process and their influence on the target. The initial parameter group is matched with the historical parameter group for similarity. The higher the matching degree, the more significant the impact of the historical parameters on the initial heuristic factors. According to the matching results, the initial heuristic factors are assigned corresponding influence weights to make them more suitable for the current production goals. The initial heuristic factors are clustered and the factors with similar influence weights are divided into several groups. Each group represents a parameter combination with similar contributions in the optimization of production goals. The similarities between different cluster groups are further analyzed, and the degree of difference between groups is calculated to identify factor groups with potential global optimization capabilities. Groups with high inter-group similarity are assigned higher comprehensive weights to highlight their importance in target optimization. For groups with low inter-group similarity, their weights are reduced to reduce their interference in the generation of the final factors. The initial heuristic factors after weight correction are normalized and integrated to generate the final heuristic factors. The heuristic factors combine the best matching results of current production conditions and historical experience and have dynamic adaptability.

[0109] Furthermore, the initial heuristic factors are weighted and modified according to the similarity between groups, including:

[0110] For the initial heuristic factors within each cluster grouping, the similarity measure within the group is calculated, and the data distribution between groups is adjusted according to the similarity measure;

[0111] Analyze the differences in data distribution between different cluster groups, assign comprehensive correction weights to cluster groups with similarity matching according to the differences in inter-group similarity, and screen out cluster groups with low similarity;

[0112] Based on the difference results of inter-group similarity and the comprehensive correction weight, the initial heuristic factors of each cluster grouping are globally modified;

[0113] The global weight correction is performed by combining the comprehensive correction weight and influence weight of each cluster group to calculate the final correction weight of each initial heuristic factor;

[0114] The final modified weight normalization results corresponding to the initial heuristic factors are integrated to obtain the heuristic factors.

[0115] As a preferred embodiment of the above, for the initial heuristic factors in each cluster grouping, the consistency of the factor characteristics in the grouping is measured by calculating the similarity measure between the factors in the grouping. The grouping with a higher similarity measure indicates that its factor characteristic concentration is higher and is suitable for further optimization. The grouping with a lower similarity measure may affect the accuracy of the final result. According to the similarity measure within the group, the data distribution between the groups is adjusted, the weight of the factors with lower similarity within the group is weakened, and the contribution ratio of the factors with high similarity is increased, so that the data distribution within the group is more concentrated; the inter-group similarity between different cluster groups is compared, and the differences in group characteristics under global production conditions are analyzed, such as judging which groups have more significant contributions to the target optimization and which groups have larger deviations. According to the inter-group similarity, the cluster groups with higher similarity matching degrees are given comprehensive correction weights, and these groups are retained first and their influence in subsequent calculations is strengthened, while the groups with similarity to other groups are screened out. The groups with low similarity and weaker contribution are selected to reduce the interference of invalid groups on the optimization process; the comprehensive correction weight is calculated for each group, and the comprehensive correction weight is combined with the matching results of the similarities between groups and the influence contribution of the group on the production target as a quantification of the overall optimization ability of the group. The comprehensive correction weight is combined with the influence weight of the group to make a global correction to the initial heuristic factor in each group. Specifically, according to the balance between the importance of the group and the characteristics of the factors within the group, the weight of each factor is adjusted to make it more in line with the global optimization goal; the final correction weight of each initial heuristic factor is calculated by weighted combination of the comprehensive correction weight and the influence weight. The final correction weight reflects the importance and contribution of each factor in the optimization of the production target. The final correction weight is normalized to ensure that the sum of all correction weights remains consistent, so that the weights of different factors are comparable. The normalized final correction weight is integrated with the corresponding initial heuristic factor to generate a heuristic factor.

[0116] Further, if Figure 4 As shown in the figure, the pheromone concentration and heuristic factor are used as adaptive conditions to establish a production parameter adaptive model, including:

[0117] Based on the pheromone concentration and heuristic factors, the initial state of the production parameter adaptive model is established, and multiple parameter adjustment paths are set for the production parameter adaptive model;

[0118] Set the pheromone concentration threshold and use the change of pheromone concentration as the adaptive trigger condition of the model. When the pheromone concentration reaches the pheromone concentration threshold, adjust the heuristic factor.

[0119] According to the changes of heuristic factors, multiple parameter adjustment paths are used to adaptively search the production process parameters and calculate the optimized production path;

[0120] Based on the optimized production path, the production parameter adaptive model is updated, and the pheromone concentration and heuristic factor are adjusted synchronously;

[0121] Feedback verification is performed on the production parameter adaptive model, the pheromone concentration and heuristic factor are recalculated based on the verification results, and the final production parameter adaptive model is generated after the system is stable.

[0122] As a preferred embodiment of the above embodiment, a heuristic factor and pheromone concentration are used as input parameters to construct a production parameter adaptive model. The pheromone concentration reflects the global optimization direction, and the heuristic factor reflects the real-time change trend of the production parameters. Multiple parameter adjustment paths are defined in the production parameter adaptive model. Each path corresponds to a different production parameter adjustment direction and adjustment strategy (such as a temperature change path, a casting speed change path, etc.). These parameter adjustment paths will serve as the search space in the adaptive model. Based on historical production experience and target production indicators (such as quality, efficiency, and energy consumption), a pheromone concentration threshold range is set. The pheromone concentration threshold is used to determine whether the current production parameters need adaptive adjustment. When the monitored pheromone concentration reaches the set pheromone concentration threshold, the production parameter adaptive model will activate the adaptive adjustment mechanism and dynamically adjust the heuristic factor to reflect changes in current production conditions. For example, when the pheromone concentration tends to be too high or too low, the weight of the heuristic factor is adjusted to rebalance the model. Multiple parameter adjustment paths are used to adaptively search for production parameters. The production parameter adaptive model gradually adjusts the parameter values ​​along the path, evaluates the impact of each adjustment on the target production indicator, and calculates the optimization degree of each path based on the search results to optimize production. A path is a path that improves the target production indicator (such as improved product quality or reduced energy consumption). The model prioritizes the path with the most significant optimization effect for the next update. Based on the search results of the optimized production path, the internal state of the production parameter adaptive model is updated, including adjusting the pheromone concentration and recalculating the heuristic factor. The update process ensures that the model can reflect the current production conditions in real time and avoids optimization failure caused by parameter lag. The pheromone concentration update strengthens the global search capability, and the heuristic factor correction enhances real-time adaptability. Through the simultaneous adjustment of the two, the model's dynamic response capability to complex production environments is improved. The updated production parameter adaptive model is applied in actual production, and production performance data (such as product quality, efficiency, and energy consumption) is collected. This data is compared and analyzed with the expected results of the production parameter adaptive model to verify whether the model accurately reflects production conditions and optimizes the target production indicators. If the production performance data deviates from the expected, the pheromone concentration and heuristic factor are recalculated based on the verification results, and the model parameters are further optimized. When the feedback verification results indicate that the production performance has reached a stable state, the parameters of the production parameter adaptive model are fixed to generate the final production parameter adaptive model, which is then applied to subsequent production processes.

[0123] Furthermore, feedback verification of the production parameter adaptive model includes:

[0124] Collect production effect data, generate a production effect data set, and establish a production experience database based on the production effect data set;

[0125] Compare and analyze the production effect data set with the expected output of the production parameter adaptive model and calculate the expected deviation;

[0126] According to the expected deviation, the corresponding heuristic factors are judged and the pheromone concentration is corrected to generate the expected correction result;

[0127] Repeatedly collect production effect data to verify the stability of the production parameter adaptive model, and store the parameter combination and expected correction results of the verified production parameter adaptive model in the production experience database.

[0128] As a preferred embodiment of the above embodiment, during the production process of steel billets, production effect data, including product quality (such as yield rate, surface defect rate), production efficiency (such as production rhythm) and energy consumption indicators (such as energy consumption), are collected in real time. The collected real-time data are sorted and labeled to generate a structured production effect data set to facilitate subsequent analysis and model optimization. The production effect data set should cover the correspondence between production parameters and target indicators to support model verification; the sorted production effect data set is stored in a production experience database, which records parameter settings, production effects and optimization results under different production conditions, providing a reference basis for subsequent similar working conditions. The production experience database is continuously updated as production conditions and optimization results change, ensuring that it always contains the latest and optimal experience data; based on the production parameter adaptive model, the expected production effect under current conditions is calculated as the output reference value of the model, the actual production effect in the production effect data set is compared with the expected output of the model item by item, and the deviation is analyzed. Based on the comparison results, the expected deviation value between the actual production effect and the expected production effect is calculated to judge the accuracy of the model optimization. The smaller the expected deviation value, the closer the model is to the actual production situation; based on the calculated The expected deviation value is used to judge the heuristic factors in the model. If the deviation value exceeds the set threshold, it means that the model fails to accurately adapt to the current production conditions and needs to be adjusted. For the production parameters with deviations, the corresponding heuristic factor weights are adjusted to make them more in line with the current production reality. The pheromone concentration is recalculated based on the corrected heuristic factors to guide the model to prioritize the adjusted parameter path in subsequent optimization. The corrected heuristic factors and pheromone concentrations are used as expected correction results and fed back to the production parameter adaptive model. During the production process, new production effect data is continuously collected to observe the performance of the model parameters after adjustment. The collected production effect data is verified to determine whether the model can stably achieve the target production effect after adjustment. If the model performance is stable, it means that the optimization is complete. If deviations still exist, corrections and verifications are continued. For the production parameter adaptive model that passes the verification, its optimized parameter combination is stored in the production experience database for direct call under similar production conditions in the future. The detailed process of deviation correction, including deviation analysis, heuristic factor adjustment and pheromone concentration update results, is stored in the production experience database to form a complete optimization record, providing data support for subsequent model improvement.

[0129] Example 2

[0130] Based on the same inventive concept as the method for adaptively adjusting production parameters of steel billets in the aforementioned embodiment, the present invention further provides a system for adaptively adjusting production parameters of steel billets, the system comprising:

[0131] Parameter acquisition module, sets production data observation points to obtain production process parameters, and organizes the production process parameters to generate original parameter groups;

[0132] The information calculation module calculates the initial pheromone value of each original parameter group and obtains the heuristic factor based on the initial pheromone value and the production process parameters;

[0133] The model generation module aggregates the initial pheromone values ​​to obtain the pheromone concentration, and uses the pheromone concentration and heuristic factors as adaptive conditions to establish a production parameter adaptive model;

[0134] The parameter combination module obtains the final solution combination of production parameters based on the production parameter adaptive model.

[0135] The above-mentioned adjustment system in the present invention can effectively realize a method for adaptively adjusting the production parameters of steel billets. The technical effects that can be achieved are as described in the above-mentioned embodiments and will not be repeated here.

[0136] More specifically, the parameter combination module includes:

[0137] The parameter candidate unit establishes parameter combination candidate solutions based on the initial pheromone values ​​and heuristic factors of each original parameter group;

[0138] Path generation unit, each parameter combination candidate solution determines a production parameter combination path according to the corresponding pheromone concentration and heuristic factor;

[0139] The adaptation calculation unit calculates the fitness value for each production parameter combination path and prioritizes the production parameter combination path with the highest fitness value at the same level;

[0140] Iterative selection unit, iteratively executing the selection of production parameter combination path and the calculation of fitness value until the fitness value convergence condition is met;

[0141] The final solution generating unit selects the parameter combination candidate solution contained in the production parameter combination path with the highest fitness value to generate the final solution combination of production parameters at the end of the iteration.

[0142] Similarly, the above-mentioned optimization schemes for the system can also respectively achieve the corresponding optimization effects of the method in Example 1, which will not be repeated here.

[0143] Although the present application has been described with reference to specific features and embodiments thereof, it is apparent that various modifications and combinations may be made thereto without departing from the spirit and scope of the present application. Accordingly, this specification and drawings are merely illustrative of the present application as defined herein and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, those skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, the present application is intended to include such modifications and variations as fall within the scope of the present application and its equivalents.

Claims

1. A method for adaptively adjusting production parameters of steel billets, characterized in that: The method comprises: Setting production data observation points to obtain production process parameters, and arranging the production process parameters to generate an original parameter group; Calculating the initial pheromone value of each of the original parameter groups, and obtaining a heuristic factor according to the initial pheromone value and the production process parameters, including: Analyzing the initial pheromone of the original parameter group under each production target, wherein the initial pheromone is captured and determined based on the impact of the original parameter group on the production target; quantifying the initial pheromone to generate the initial pheromone value, and combining the original parameter group to generate an initial heuristic factor; Obtaining a historical parameter group, and assigning corresponding influence weights to the initial heuristic factors based on similarities between the original parameter group and the historical parameter group; Clustering the initial heuristic factors into groups, combining the initial heuristic factors with similar influence weights, and analyzing the similarity between the groups; The initial heuristic factors are weighted according to the inter-group similarity, and the heuristic factors are obtained by integration according to the weighted correction results, including: For the initial heuristic factors within each cluster grouping, a similarity measure within the group is calculated, and the data distribution between the groups is adjusted according to the similarity measure; Analyzing the difference in data distribution between different cluster groups, assigning comprehensive correction weights to the cluster groups with similarity matching according to the difference in similarity between the groups, and screening out cluster groups with low similarity; Based on the difference results of the similarities between the groups and the comprehensive correction weight, a global weight correction is performed on the initial heuristic factor of each cluster grouping; The global weight correction is performed by weighted combination of the comprehensive correction weight and the influence weight of each cluster group to calculate the final correction weight of each initial heuristic factor; Integrating the final modified weight normalization results corresponding to the initial heuristic factors to obtain the heuristic factors; Aggregating initial pheromone values ​​to obtain pheromone concentrations, and using the pheromone concentrations and the heuristic factors as adaptive conditions to establish a production parameter adaptive model; According to the production parameter adaptive model, a final solution combination of production parameters is obtained.

2. The method for adaptively adjusting steel billet production parameters according to claim 1, characterized in that: Obtain the final solution combination of production parameters, including: Establishing parameter combination candidate solutions based on the initial pheromone values ​​and heuristic factors of each of the original parameter groups; Each parameter combination candidate solution determines a production parameter combination path according to the corresponding pheromone concentration and heuristic factor; Calculating the fitness value for each of the production parameter combination paths, and preferentially exploring the production parameter combination path with the highest fitness value at the same level; Iteratively executing the selection of the production parameter combination path and the calculation of the fitness value until a fitness value convergence condition is met; At the end of the iteration, the parameter combination candidate solution included in the production parameter combination path with the highest fitness value is selected to generate the production parameter final solution combination.

3. The method for adaptively adjusting steel billet production parameters according to claim 2, characterized in that: Iteratively performing selection of the production parameter combination path includes: updating the pheromone concentration on the production parameter combination path according to the fitness value; Comparing the fitness of each of the production parameter combination paths, increasing the pheromone concentration corresponding to a high fitness value, and gradually evaporating the pheromone concentration corresponding to a low fitness value; Correcting the pheromone concentration result of each production parameter combination path according to the heuristic factor reflecting the change trend of the production process parameters; Detecting whether a local optimal solution appears, and if so, performing a perturbation operation on the production parameter combination path with similar and continuous changes in pheromone concentration; After each iteration, the change of the fitness value of each production parameter combination path is monitored. When the change rate is lower than the fitness value convergence condition, the iteration process is terminated and the final solution combination of the production parameters is selected.

4. The method for adaptively adjusting steel billet production parameters according to claim 3, characterized in that: Performing a perturbation operation on the production parameter combination path with continuous and similar changes in pheromone concentrations, including: Setting a disturbance factor for each of the production parameter combination paths with similar and continuous changes in pheromone concentration, wherein the disturbance factor is set according to the stability of the changes in pheromone concentration; Based on the disturbance factor, the production parameter combination path is offset and disturbed to generate a disturbance combination path; adjusting the disturbance intensity of the disturbance combination path according to the fitness value, and reducing the disturbance intensity for the disturbance combination path with an improved fitness value; The fitness value of the disturbance combination path is evaluated, and when the evaluation result is higher than the fitness value of the corresponding production parameter combination path, the path is replaced and the pheromone concentration is updated.

5. The method for adaptively adjusting steel billet production parameters according to claim 1, characterized in that: The pheromone concentration and the heuristic factor are used as adaptive conditions to establish a production parameter adaptive model, including: Based on the pheromone concentration and the heuristic factor, establishing an initial state of the production parameter adaptive model, and setting multiple parameter adjustment paths for the production parameter adaptive model; Setting a pheromone concentration threshold, using a change in pheromone concentration as a self-adaptive trigger condition for the model, and adjusting the heuristic factor when the pheromone concentration reaches the pheromone concentration threshold; According to the change of the heuristic factor, a plurality of the parameter adjustment paths are used to perform an adaptive search for the production process parameters, and calculate an optimized production path; According to the optimized production path, the production parameter adaptive model is updated, and the pheromone concentration and the heuristic factor are adjusted synchronously; Feedback verification is performed on the production parameter adaptive model, the pheromone concentration and the heuristic factor are recalculated based on the verification result, and the final production parameter adaptive model is generated after the system is stabilized.

6. The method for adaptively adjusting steel billet production parameters according to claim 5, characterized in that: Feedback verification of the production parameter adaptive model includes: Collecting production effect data, generating a production effect data set, and establishing a production experience database based on the production effect data set; Comparing and analyzing the production effect data set with the expected output of the production parameter adaptive model, and calculating the expected deviation; According to the expected deviation, the corresponding heuristic factor is judged and the pheromone concentration is corrected to generate an expected correction result; The production effect data is repeatedly collected to verify the stability of the production parameter adaptive model, and the parameter combination of the production parameter adaptive model that has passed the verification and the expected correction result are stored in the production experience database.

7. A steel billet production parameter adaptive adjustment system, characterized in that: The system comprises: A parameter acquisition module sets production data observation points to acquire production process parameters, and organizes the production process parameters to generate an original parameter group; An information calculation module calculates the initial pheromone value of each of the original parameter groups and obtains a heuristic factor based on the initial pheromone value and the production process parameters, including: Analyzing the initial pheromone of the original parameter group under each production target, wherein the initial pheromone is captured and determined based on the impact of the original parameter group on the production target; quantifying the initial pheromone to generate the initial pheromone value, and combining the original parameter group to generate an initial heuristic factor; Obtaining a historical parameter group, and assigning corresponding influence weights to the initial heuristic factors based on similarities between the original parameter group and the historical parameter group; Clustering the initial heuristic factors into groups, combining the initial heuristic factors with similar influence weights, and analyzing the similarity between the groups; The initial heuristic factors are weighted according to the inter-group similarity, and the heuristic factors are obtained by integration according to the weighted correction results, including: For the initial heuristic factors within each cluster grouping, a similarity measure within the group is calculated, and the data distribution between the groups is adjusted according to the similarity measure; Analyzing the difference in data distribution between different cluster groups, assigning comprehensive correction weights to the cluster groups with similarity matching according to the difference in similarity between the groups, and screening out cluster groups with low similarity; Based on the difference results of the similarities between the groups and the comprehensive correction weight, a global weight correction is performed on the initial heuristic factor of each cluster grouping; The global weight correction is performed by weighted combination of the comprehensive correction weight and the influence weight of each cluster group to calculate the final correction weight of each initial heuristic factor; Integrating the final modified weight normalization results corresponding to the initial heuristic factors to obtain the heuristic factors; A model generation module aggregates initial pheromone values ​​to obtain pheromone concentrations, and uses the pheromone concentrations and the heuristic factors as adaptive conditions to establish a production parameter adaptive model; The parameter combination module obtains the final solution combination of production parameters according to the production parameter adaptive model.

8. The steel billet production parameter adaptive adjustment system according to claim 7, characterized in that: The parameter combination module includes: The parameter candidate unit establishes parameter combination candidate solutions based on the initial pheromone values ​​and heuristic factors of each original parameter group; Path generation unit, each parameter combination candidate solution determines a production parameter combination path according to the corresponding pheromone concentration and heuristic factor; The adaptation calculation unit calculates the fitness value for each production parameter combination path and prioritizes the production parameter combination path with the highest fitness value at the same level; Iterative selection unit, iteratively executing the selection of production parameter combination path and the calculation of fitness value until the fitness value convergence condition is met; The final solution generating unit selects the parameter combination candidate solution contained in the production parameter combination path with the highest fitness value to generate the final solution combination of production parameters at the end of the iteration.

Images