Prediction method and device for pearlite transformation curve of roller
By constructing a machine learning model and combining it with SHAP value interpretation analysis, the problem of predicting the pearlite transformation curve of rolls was solved, achieving fast and accurate prediction of the pearlite transformation curve of rolls. It is applicable to the prediction of phase transformation critical points under different cooling rates and reduces costs.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-08
- Publication Date
- 2026-03-31
AI Technical Summary
Existing technologies lack effective methods to predict the continuous cooling transformation curve of supercooled austenite in rolls, especially the nonlinear relationship in the pearlite region. Furthermore, traditional methods are time-consuming and costly, making it difficult to meet the needs of the rapid development of new materials.
A combined model was constructed using machine learning methods. After screening alloy composition data and standardizing it, it was divided into training and testing sets. The model was optimized using algorithms such as random forest and extreme gradient boosting. Combined with SHAP value interpretation and analysis, the pearlite onset and end temperatures were accurately predicted, and the pearlite transformation curve of the roll was plotted.
It enables rapid and accurate prediction of pearlite transformation curves in rolls, improves model accuracy and efficiency, reduces costs, and is applicable to phase transformation critical point prediction under different cooling rates.
Smart Images
Figure CN121768541A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of heat treatment technology for metallic materials, and in particular to a method and apparatus for predicting the pearlite transformation curve of a rolling mill. Background Technology
[0002] The continuous cooling transformation curve (CCT curve) of supercooled austenite is a core tool for achieving integrated design and control of materials, processes, microstructure, and properties. The CCT curve of a roll can be divided into two parts: the pearlite transformation curve and the bainite transformation curve. Existing literature lacks prediction methods for roll CCT curves, and the accuracy of both the CCT curves and the trained models needs improvement. The dilatation method is generally used to determine the CCT curve, but this method is time-consuming, difficult to operate, and costly, which does not meet the expectations of rapid development in new materials. Thermodynamic model calculations also rely on expensive commercial thermodynamic calculation software and corresponding databases. For the pearlite region of the CCT curve, there is a nonlinear relationship between the chemical composition of steel and the phase transformation temperature. Machine learning excels at handling this nonlinear relationship and can make rapid and effective predictions. However, existing technologies using machine learning to predict the pearlite region portion of the continuous cooling transformation curve of supercooled austenite in steel are lacking. Summary of the Invention
[0003] To address the technical problem of the lack of a specific method for predicting the CCT curve of a rolling mill roll in existing technologies, this invention provides a method and apparatus for predicting the pearlite transformation curve of a rolling mill roll. The technical solution is as follows: On the one hand, a method for predicting the pearlite transformation curve of a rolling mill is provided. This method is implemented by a device for predicting the pearlite transformation curve of a rolling mill, and includes: S1: Select steel alloy composition data through preset screening conditions, calculate steel alloy composition, and obtain alloy composition dataset. The alloy composition data is used to plot the continuous cooling transformation curve of supercooled austenite. S2: The alloy composition dataset is standardized to obtain a standardized dataset. The dataset is divided into a training set and a test set according to a preset ratio. The standardization process includes making the data conform to a standard normal distribution. S3: Construct a prediction model for the pearlite transformation curve of the roll. The prediction model for the pearlite transformation curve of the roll is a combined model, which is composed of a machine learning model. The output of the prediction model for the pearlite transformation curve of the roll includes the pearlite initiation temperature Ps and the pearlite transformation termination temperature Pf. S4: Based on the training set and the test set, the prediction model of the pearlite transformation curve of the roll is trained and tested to obtain a pre-trained prediction model of the pearlite transformation curve of the roll. S5: Based on the prediction model of the pre-trained pearlite transformation curve of the roll, a complete pearlite transformation curve of the roll is obtained. The complete pearlite transformation curve of the roll divides the pearlite into an upper part and a lower part. The upper part is formed by the combination of the pearlite initiation temperature Ps and the pearlite transformation termination temperature Pf to form a C-shape. S6: The SHAP value is used to interpret and analyze the prediction model of the pre-trained pearlite transformation curve of the roll, and the nonlinear relationship and importance ranking of the transformation temperature of the alloying elements and the phase are obtained. The nonlinear relationship between the alloying elements and the transformation temperature of the phase includes a negative correlation between the alloying elements Mn and Ni and a positive correlation between the alloying elements and Cr. The negative correlation indicates that the alloying elements cause the pearlite to move downward and the positive correlation indicates that the alloying elements cause the pearlite to move upward.
[0004] Preferably, in step S1, data on the alloy composition of steel are selected based on preset screening conditions, and the alloy composition of the steel is calculated to obtain an alloy composition dataset, including: S11: Select data related to the composition of alloying elements from the original data after using preset screening conditions. The preset screening conditions include element content and phase transformation behavior. The alloying elements include Fe, C, Mn, Si, Cr, Ni, Mo, V, Al, W, Cu, Nb and N. S12: Calculate and organize the alloy element composition of these data; S13: Finally, obtain the alloy composition dataset.
[0005] Preferably, the alloy composition data is used to plot the continuous cooling transformation curve of supercooled austenite, including:
[0006] Machine learning methods are used to predict the pearlite region portion of the continuous cooling transformation curve of supercooled austenite in steel. Multi-dimensional alloy element composition data is used as input feature vectors, and the model is trained in a data-driven manner to learn the influence mechanism of each element on phase transformation dynamics. Based on the model, key critical points such as the phase transformation start temperature and end temperature under different cooling rates in the pearlite transformation region are accurately predicted, and the continuous cooling transformation curve of supercooled austenite is plotted.
[0007] Preferably, in step S3, a prediction model for the pearlite transformation curve of the roll is constructed. This prediction model is a combined model, composed of machine learning models. The output of the prediction model includes the pearlite initiation temperature Ps and the pearlite transformation termination temperature Pf, comprising: S31: Based on multiple machine learning methods, a model is constructed to predict the pearlite initiation temperature Ps and the pearlite transformation end temperature Pf. The machine learning method is a combined model. The model for predicting the pearlite initiation temperature Ps and the pearlite transformation end temperature Pf is used to plot the pearlite transformation curve of the roll. S32: The combined model consists of multiple machine learning models, including random forest, extreme gradient boosting, lightweight gradient boosting machine algorithm, gradient boosting decision tree, support vector machine and K-nearest neighbors; S33: Use a genetic algorithm to optimize the parameters of the combined model to obtain a prediction model for the pearlite transformation curve of the roll.
[0008] Preferably, step S4, based on the training set and the test set, trains and tests the prediction model for the pearlite transition curve of the roll to obtain a pre-trained prediction model for the pearlite transition curve of the roll, including: S41: Determine the input and output of the prediction model for the pearlite transformation curve of the roll. The input includes alloying elements, austenitizing temperature, and cooling rate. The output includes Ps temperature and Pf temperature. The austenitizing temperature is used to replace the austenite grain size for modeling. S42: Extract variables from the training set according to the input and output, and fit and train the prediction model of the pearlite transformation curve of the roll to obtain the trained prediction model of the pearlite transformation curve of the roll. S43: Based on the test set, test the prediction model of the pearlite transformation curve of the rolled mill after training to verify the effectiveness of the model and obtain the pre-trained prediction model of the pearlite transformation curve of the rolled mill.
[0009] Preferably, the prediction model based on the pre-trained pearlite transformation curve of the roll in S5 obtains a complete pearlite transformation curve of the roll. This complete pearlite transformation curve divides the pearlite into an upper part and a lower part. The upper part is formed by the combination of the pearlite initiation temperature Ps and the pearlite transformation termination temperature Pf, creating a C-shape, including: S51: Using alloying elements, austenitizing temperature, and cooling rate as inputs, the model inference is performed based on the pre-trained prediction model of the pearlite transformation curve of the roll to obtain the predicted pearlite initiation temperature Ps and pearlite transformation termination temperature Pf. S52: Based on the predicted pearlite initiation temperature Ps and pearlite transformation end temperature Pf, the complete pearlite transformation curve of the roll is obtained. S53: By collecting information from the upper and lower parts of the complete pearlite transformation curve of the roll, the C-shaped feature formed by the combination of Ps and Pf in the upper part and the distinction between the lower part are obtained.
[0010] Preferably, in step S6, the SHAP value is used to interpret and analyze the pre-trained prediction model of the pearlite transformation curve of the roll, obtaining the nonlinear relationship and importance ranking of the transformation temperature of the alloying elements and the phase. The nonlinear relationship between the alloying elements and the transformation temperature of the phase includes a negative correlation between the alloying elements Mn and Ni, and a positive correlation between the alloying elements and Cr. The negative correlation indicates that the alloying elements cause the pearlite to shift downward, and the positive correlation indicates that the alloying elements cause the pearlite to shift upward. S61: Based on the test set, calculate the SHAP value of each alloy element characteristic and draw a SHAP dependency graph. The SHAP dependency graph uses the actual value of a specific alloy element on the horizontal axis and its corresponding SHAP value on the vertical axis. S62: Based on the SHAP values of each alloying element, calculate the average value of the absolute values of the SHAP values of each alloying element. The larger the average value, the more significant the contribution of the element in the model prediction. S63: Observe the distribution trend of points in the SHAP dependency graph and identify the complex nonlinear relationship between the element and the pearlite transformation curve; S64: Based on the average absolute value of the SHAP values of various alloying elements, the importance ranking of the influence of each alloying element on the pearlite transformation curve is obtained.
[0011] On the other hand, a device for predicting the pearlite transition curve of a roll is provided. This device is applied to a method for predicting the pearlite transition curve of a roll, and the device includes: Data module: used to select steel alloy composition data through preset screening conditions, calculate the steel alloy composition, and obtain alloy composition dataset. The alloy composition data is used to plot the continuous cooling transformation curve of supercooled austenite. Preprocessing module: used to standardize the alloy composition dataset to obtain a standardized dataset, and divide the dataset into a training set and a test set according to a preset ratio. The standardization process includes making the data conform to a standard normal distribution. Prediction Model Module: Used to construct a prediction model for the pearlite transformation curve of the roll. The prediction model for the pearlite transformation curve of the roll is a combined model, which is composed of a machine learning model. The output of the prediction model for the pearlite transformation curve of the roll includes the pearlite initiation temperature Ps and the pearlite transformation termination temperature Pf. Training module: used to train and test the prediction model of the pearlite transformation curve of the roll based on the training set and the test set, so as to obtain a pre-trained prediction model of the pearlite transformation curve of the roll. Transformation curve module: Used to obtain a complete pearlite transformation curve based on a pre-trained prediction model of the pearlite transformation curve of the roll. The complete pearlite transformation curve divides the pearlite into an upper part and a lower part. The upper part is formed by the combination of the pearlite initiation temperature Ps and the pearlite transformation termination temperature Pf to form a C-shape. SHAP module: Used to interpret and analyze the prediction model of the pre-trained pearlite transformation curve of the roll using SHAP values, and obtain the nonlinear relationship and importance ranking of the transformation temperature of the alloying elements and the phase. The nonlinear relationship between the alloying elements and the transformation temperature of the phase includes a negative correlation between the alloying elements Mn and Ni, and a positive correlation between the alloying elements and Cr. The negative correlation indicates that the alloying elements cause the pearlite to move downward, and the positive correlation indicates that the alloying elements cause the pearlite to move upward.
[0012] On the other hand, a device for predicting the pearlite transition curve of a roll is provided. The device for predicting the pearlite transition curve of a roll includes: a processor; a memory storing computer-readable instructions, which, when executed by the processor, implement the method described in any of the above-described methods for predicting the pearlite transition curve of a roll.
[0013] On the other hand, a computer-readable storage medium is provided, characterized in that the computer-readable storage medium stores program code, which can be invoked by a processor to execute the method as described in any one of claims 1 to 7.
[0014] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: Attached Figure Description
[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0016] Figure 1 This is a flowchart of a method for predicting the pearlite transformation curve of a rolling mill, provided in an embodiment of the present invention. Figure 2 This is a pearlite transformation curve and SHAP plots of four models, Ps and Pf, provided in an embodiment of the present invention; Figure 3 This is a block diagram of a device for predicting the pearlite transformation curve of a rolling mill, provided in an embodiment of the present invention. Figure 4 This is a schematic diagram of the structure of a device for predicting the pearlite transformation curve of a rolling mill, provided in an embodiment of the present invention. Detailed Implementation
[0017] The technical solution of the present invention will now be described with reference to the accompanying drawings.
[0018] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.
[0019] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.
[0020] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.
[0021] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.
[0022] This invention provides a method for predicting the pearlite transition curve of a rolling mill roll. This method can be implemented using a device for predicting the pearlite transition curve of a rolling mill roll, which can be a terminal or a server. For example... Figure 1 The flowchart shown is a method for predicting the pearlite transformation curve of a rolling mill. The processing flow of this method may include the following steps: Preferably, data on the alloy composition of steel are selected after passing preset screening conditions, the alloy composition of the steel is calculated, and an alloy composition dataset is obtained, including: After using preset screening conditions, data related to the composition of alloying elements are selected from the raw data. The preset screening conditions include element content and phase transformation behavior. The alloying elements include Fe, C, Mn, Si, Cr, Ni, Mo, V, Al, W, Cu, Nb, and N. Calculate and organize the alloy element composition of these data; Finally, the alloy composition dataset was obtained.
[0023] In some embodiments, iron (Fe) is the matrix element of most steel alloys, and is the core component that constitutes the basic phases such as ferrite and austenite, providing the basic mechanical structure of the alloy.
[0024] Carbon (C) is the core element that regulates the properties of steel. When the content is low, it forms ferrite and pearlite with Fe, which increases the strength of steel. When the content is high, it forms cementite, which significantly improves the hardness and wear resistance of steel.
[0025] Manganese (Mn) is an excellent deoxidizer and desulfurizer. It forms a solid solution with Fe, which strengthens the steel by solid solution and enhances its strength and hardness. It can also stabilize the austenitic structure and significantly improve the hardenability of the steel.
[0026] Silicon (Si) can dissolve in ferrite to produce significant solid solution strengthening, which improves the elastic limit and yield strength of steel, making it a key element for spring steel.
[0027] Chromium (Cr) can significantly improve the hardenability and oxidation resistance of steel, and enhance the surface wear resistance of steel.
[0028] Nickel (Ni) can strengthen ferrite and refine pearlite, thereby increasing the strength of steel without significantly reducing its plasticity.
[0029] Molybdenum (Mo) can improve the hardenability and hot strength of steel and prevent temper brittleness.
[0030] Vanadium (V) can refine alloy grains and reduce overheating sensitivity.
[0031] Aluminum (Al) is a powerful deoxidizer and grain refiner in steel production. It can inhibit the aging phenomenon of low-carbon steel and enhance its low-temperature toughness.
[0032] Tungsten (W) partially forms refractory carbides and partially dissolves in iron, which can increase the tempering stability, red hardness and wear resistance of steel. It is mainly used in high-speed steel and hot forging die steel, which can delay carbide aggregation and maintain high-temperature strength.
[0033] Copper (Cu) can significantly improve the atmospheric corrosion resistance of steel.
[0034] Niobium (Nb) can partially dissolve in solid solutions to strengthen the steel, and partially form carbides to refine the grains. It also has a secondary hardening effect and can prevent intergranular corrosion in austenitic steel, thus improving the high-temperature performance of heat-resistant steel.
[0035] Nitrogen (N) can solid solution strengthen steel and slightly improve hardenability; after nitriding the steel surface, the hardness, wear resistance and corrosion resistance will be enhanced.
[0036] Preferably, the alloy composition dataset is standardized to obtain a standardized dataset, and the dataset is divided into a training set and a test set according to a preset ratio. The standardization process includes making the data conform to a standard normal distribution. It should be further explained that the dataset is divided into training and test sets in a 7:3 ratio. The training set is used to fit the model, and the test set is used to validate the model's effectiveness. Standardization transforms the data into uniform standard values, ensuring that the standardized data conforms to a standard normal distribution, i.e., the mean is 0 and the standard deviation is 1.
[0037] Preferably, the alloy composition data is used to plot the continuous cooling transformation curve of supercooled austenite, including: Machine learning methods are used to predict the pearlite region portion of the continuous cooling transformation curve of supercooled austenite in steel. Multi-dimensional alloy element composition data is used as input feature vectors, and the model is trained in a data-driven manner to learn the influence mechanism of each element on phase transformation dynamics. Based on the model, key critical points such as the phase transformation start temperature and end temperature under different cooling rates in the pearlite transformation region are accurately predicted, and the continuous cooling transformation curve of supercooled austenite is plotted.
[0038] In some embodiments, the Continuous Cooling Transformation Curve (CCT curve) is a kinetic graph reflecting the transformation of supercooled austenite in steel under continuous cooling conditions. It plots temperature on the ordinate and the logarithm of time on the abscissa, and requires the labeling of key thermodynamic parameters of the alloy (such as the equilibrium critical temperature). Critical heating temperature A (etc.) and the characteristics of different phase transition regions. During the plotting process, it is necessary to focus on determining key critical points such as the phase transition start temperature and end temperature under different cooling rates within the pearlite transformation region. The pearlite transformation region consists of a transformation start line, a transformation end line, and a transformation termination line—the curve on the left is the pearlite transformation start line, the curve on the right is the transformation end line, and the curve at the bottom is the transformation termination line. When the cooling curve intersects with the termination line, the pearlite transformation stops.
[0039] Preferably, a predictive model for the pearlite transformation curve of the roll is constructed. This predictive model is a combined model, composed of machine learning models. The output of the predictive model includes the pearlite initiation temperature Ps and the pearlite transformation termination temperature Pf, comprising: Based on multiple machine learning methods, a model is constructed to predict the pearlite initiation temperature Ps and the pearlite transformation end temperature Pf. The machine learning methods are combined models. The model for predicting the pearlite initiation temperature Ps and the pearlite transformation end temperature Pf is used to plot the pearlite transformation curve of the roll. The combined model consists of multiple machine learning models, including random forest, extreme gradient boosting, lightweight gradient boosting machine algorithm, gradient boosting decision tree, support vector machine and K-nearest neighbors; The parameters of the combined model are optimized using a genetic algorithm to obtain a predictive model for the pearlite transformation curve of the roll.
[0040] In some embodiments, the parameters of the machine learning are continuously adjusted so that the error between the outputs of the temperature prediction model and the feature parameters in the test set is within an allowable range, ensuring the accuracy of the temperature prediction model. The parameters of each machine learning method are shown in Table 1. The MAE, RMSE, and R-values of the training and test sets are then used to evaluate these parameters. 2 The evaluation determined the optimal model: GBDT predicts Ps, and LGBM predicts Pf. GBDT's performance on the training set was evaluated. MAE It was 3.16℃. RMSE It is 5.72℃, R 2 The value is 0.99 on the test set. MAE It was 4.84℃. RMSE It was 9.72℃. R 2 The value is 0.98, which is the performance of LGBM on the training set. MAE It was 4.98℃. RMSE It is 11.9℃, R 2 The value is 0.96 on the test set. MAE It was 6.13℃. RMSE It was 14.16℃. R 2 It is 0.95.
[0041] Table 1
[0042] It should be noted that: n_estimators: the number of decision trees, used to control the model's fitting ability and generalization; learning_rate: the step size parameter of gradient boosting models, balancing training convergence speed and overfitting risk; max_depth: the maximum depth of a single tree in decision tree models, limiting model complexity to avoid overfitting; C: the regularization penalty coefficient for Support Vector Regression (SVR); gamma: the bandwidth parameter of the RBF kernel function in the SVR model; num_leaves: the maximum number of leaf nodes in a single tree in Lightweight Gradient Boosting Machine (LGBM), affecting model fitting accuracy and training efficiency; n_neighbors: the number of nearest neighbor samples participating in prediction in the K-Nearest Neighbors (KNN) model, determining the range of local feature extraction; leaf_size: the size of the leaf nodes used to build the KD tree in the KNN model, balancing search efficiency and prediction accuracy.
[0043] It should be further noted that gradient boosting classes (XGBoost, LGBM, GBDT) are suitable for capturing complex nonlinear relationships, with XGBoost and LGBM being superior in both speed and accuracy.
[0044] Preferably, based on the training set and the test set, the prediction model for the pearlite transition curve of the roll is trained and tested to obtain a pre-trained prediction model for the pearlite transition curve of the roll, including: The inputs and outputs of the prediction model for the pearlite transformation curve of the roll are determined. The inputs include alloying elements, austenitizing temperature, and cooling rate. The outputs include Ps temperature and Pf temperature. The austenitizing temperature is used to replace the austenite grain size for modeling. Variables are extracted from the training set according to the input and output, and the prediction model of the pearlite transformation curve of the roll is fitted and trained to obtain the trained prediction model of the pearlite transformation curve of the roll. Based on the test set, the prediction model of the pearlite transition curve of the rolled mill is tested to verify the effectiveness of the model and obtain the pre-trained prediction model of the pearlite transition curve of the rolled mill.
[0045] Preferably, based on the pre-trained prediction model of the pearlite transformation curve of the roll, the complete pearlite transformation curve of the roll is obtained, such as... Figure 2-1 As shown, the complete pearlite transformation curve of the roll divides the pearlite into an upper part and a lower part. The upper part is formed by the combination of the pearlite initiation transformation temperature Ps and the pearlite transformation termination temperature Pf, forming a C-shape, including: Using alloying elements, austenitizing temperature, and cooling rate as inputs, and based on a pre-trained prediction model of the pearlite transformation curve of the roll, model inference is performed to obtain the predicted pearlite initiation temperature Ps and pearlite transformation termination temperature Pf. Based on the predicted pearlite initiation temperature Ps and pearlite transformation end temperature Pf, the complete pearlite transformation curve of the roll is obtained. By collecting information from the upper and lower parts of the complete pearlite transformation curve of the roll, the C-shaped feature formed by the combination of Ps and Pf in the upper part and the distinction between the lower part can be obtained.
[0046] Preferably, the SHAP value is used to interpret and analyze the pre-trained prediction model of the pearlite transformation curve of the roll, obtaining the nonlinear relationship and importance ranking of the transformation temperature of the alloying elements and the phase. The nonlinear relationship between the alloying elements and the transformation temperature of the phase includes a negative correlation between the alloying elements Mn and Ni, and a positive correlation between the alloying elements and Cr. The negative correlation indicates that the alloying elements cause the pearlite to shift downward, and the positive correlation indicates that the alloying elements cause the pearlite to shift upward. Based on the test set, the SHAP values of each alloy element characteristic are calculated, and a SHAP dependency graph is plotted. The SHAP dependency graph uses the actual value of a specific alloy element on the horizontal axis and its corresponding SHAP value on the vertical axis. Based on the SHAP values of each alloying element, the average value of the absolute values of the SHAP values of each alloying element is calculated. The larger the average value, the more significant the contribution of the element in the model prediction. like Figure 2-2 As shown, by observing the distribution trend of points in the SHAP dependency graph, the complex nonlinear relationship between this element and the pearlite transformation curve can be identified. Based on the average absolute value of the SHAP values of various alloying elements, the importance ranking of the influence of each alloying element on the pearlite transformation curve is obtained.
[0047] In some embodiments, the SHAP value is used to interpret the effect of the predictive model's display features on the pearlite transition curve. (Reference) Figure 2-2 The SHAP plots for the four models Ps and Pf are shown, illustrating the nonlinear relationship between alloying elements and the pearlite transformation curve. Each point in the plot represents a sample, and the color of the point corresponds to an eigenvalue. A redder color corresponds to a larger eigenvalue, and a bluer color corresponds to a smaller eigenvalue. SHAP correlation analysis of the input variables shows that the addition of alloying elements alters the austenite-pearlite equilibrium temperature. For Ps and Pf, the alloying elements Mn and Ni show a negative correlation, while Cr shows a positive correlation. To further verify the accuracy of the models, the Ps and Pf models were combined into a single CCT plot showing the pearlite region. Five pearlite transformation curves were used for verification and compared with the prediction results from JMatPro software. The compositions of these five steels are shown in Table 2. Figure 2-1 To illustrate the prediction of pearlite transformation curves using a combined machine learning model, the experimental data and the predicted data in the figure show roughly the same trend, which can be used for formulating heat treatment processes in actual production. A comparison was made between the pearlite transformation curves predicted by the machine learning model and JMatPro, and the experimental pearlite transformation curves. It can be seen that the transformation temperature predicted by the machine learning model is closer to the actual value than that predicted by JMatPro. For these five types of steel, JMatPro software showed significant differences in its prediction of Pf temperature. This is because the alloy element content of roll steel is high, and elements such as Cr, Mo, and V in high-alloy steel significantly affect carbon activity and diffusion coefficient. However, the thermodynamic database may simplify the processing of interactions in multi-component systems, leading to error accumulation. For the prediction of Ps in #1, #2, #3, and #5 steels, the error of the machine learning model is smaller than that of JMatPro. This machine learning model can accurately predict the pearlite transformation curve.
[0048] Table 2
[0049] The above is an introduction to the method embodiments. The following describes the solution described in this application through device embodiments.
[0050] Figure 3 This is a block diagram illustrating a device for predicting the pearlite transition curve of a roll according to an exemplary embodiment. The device is used in a method for predicting the pearlite transition curve of a roll. (Refer to...) Figure 3 The device includes a data module, a preprocessing module, a prediction model module, a training module, a transformation curve module, and a SHAP module.
[0051] Data module: used to select steel alloy composition data through preset screening conditions, calculate the steel alloy composition, and obtain alloy composition dataset. The alloy composition data is used to plot the continuous cooling transformation curve of supercooled austenite. Preprocessing module: used to standardize the alloy composition dataset to obtain a standardized dataset, and divide the dataset into a training set and a test set according to a preset ratio. The standardization process includes making the data conform to a standard normal distribution. Prediction Model Module: Used to construct a prediction model for the pearlite transformation curve of the roll. The prediction model for the pearlite transformation curve of the roll is a combined model, which is composed of a machine learning model. The output of the prediction model for the pearlite transformation curve of the roll includes the pearlite initiation temperature Ps and the pearlite transformation termination temperature Pf. Training module: used to train and test the prediction model of the pearlite transformation curve of the roll based on the training set and the test set, so as to obtain a pre-trained prediction model of the pearlite transformation curve of the roll. Transformation curve module: Used to obtain a complete pearlite transformation curve based on a pre-trained prediction model of the pearlite transformation curve of the roll. The complete pearlite transformation curve divides the pearlite into an upper part and a lower part. The upper part is formed by the combination of the pearlite initiation temperature Ps and the pearlite transformation termination temperature Pf to form a C-shape. SHAP module: Used to interpret and analyze the prediction model of the pre-trained pearlite transformation curve of the roll using SHAP values, and obtain the nonlinear relationship and importance ranking of the alloying elements and the pearlite transformation curve. The nonlinear relationship between the alloying elements and the pearlite transformation curve includes a negative correlation between the alloying elements Mn and Ni, and a positive correlation between the alloying elements and Cr. The negative correlation indicates that the alloying elements cause the pearlite to move downward, and the positive correlation indicates that the alloying elements cause the pearlite to move upward.
[0052] A device for predicting the pearlite transition curve of a rolling mill, the device comprising: a processor; and a memory storing computer-readable instructions, wherein when the computer-readable instructions are executed by the processor, the method described in any one of the above-described methods for predicting the pearlite transition curve of a rolling mill is implemented.
[0053] A computer-readable storage medium, characterized in that the computer-readable storage medium stores program code, the program code being invoked by a processor to execute the method as described in any one of claims 1 to 7.
[0054] Figure 4 This is a schematic diagram of the structure of a predictive device for the pearlite transformation curve of a rolling mill provided in an embodiment of the present invention, as shown below. Figure 4 As shown, the device for predicting the pearlite transformation curve of the roll can include the above-mentioned... Figure 3 The illustrated device for predicting the pearlite transition curve of a roll. Optionally, the device 410 for predicting the pearlite transition curve of a roll may include a first processor 2001.
[0055] Optionally, the prediction device 410 for the pearlite transformation curve of the roll may also include a memory 2002 and a transceiver 2003.
[0056] The first processor 2001, memory 2002, and transceiver 2003 can be connected via a communication bus.
[0057] The following is combined Figure 4 A detailed description of each component of the device 410 for predicting the pearlite transformation curve of a rolling mill is provided below: The first processor 2001 is the control center of the prediction device 410 for the pearlite transformation curve of the roll. It can be a single processor or a collective term for multiple processing elements. For example, the first processor 2001 can be one or more central processing units (CPUs), application-specific integrated circuits (ASICs), or one or more integrated circuits configured to implement embodiments of the present invention, such as one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs).
[0058] Optionally, the first processor 2001 can perform various functions of the roll pearlite transformation curve prediction device 410 by running or executing software programs stored in the memory 2002 and calling data stored in the memory 2002.
[0059] In a specific implementation, as one example, the first processor 2001 may include one or more CPUs, for example... Figure 4 CPU0 and CPU1 are shown in the diagram.
[0060] In a specific implementation, as one example, the prediction device 410 for the pearlite transformation curve of the roll may also include multiple processors, such as... Figure 4 The first processor 2001 and the second processor 2004 are shown in the diagram. Each of these processors can be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). Here, a processor can refer to one or more devices, circuits, and / or processing cores used to process data (e.g., computer program instructions).
[0061] The memory 2002 is used to store the software program that executes the present invention, and is controlled by the first processor 2001 to execute it. The specific implementation method can be referred to the above method embodiment, and will not be repeated here.
[0062] Optionally, the memory 2002 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory 2002 may be integrated with the first processor 2001 or may exist independently, and may be connected via the interface circuit of the predictor device 410 for the pearlite transformation curve of the rolling mill. Figure 4 (Not shown in the image) is coupled to the first processor 2001, and this embodiment of the invention does not specifically limit this.
[0063] The transceiver 2003 is used to communicate with network devices or with terminal devices.
[0064] Alternatively, transceiver 2003 may include a receiver and a transmitter. Figure 4 (Not shown separately). The receiver is used to implement the receiving function, and the transmitter is used to implement the transmitting function.
[0065] Optionally, the transceiver 2003 can be integrated with the first processor 2001, or it can exist independently and be connected to the interface circuit of the prediction device 410 for the pearlite transformation curve of the roll. Figure 4 (Not shown in the image) is coupled to the first processor 2001, and this embodiment of the invention does not specifically limit this.
[0066] It should be noted that, Figure 4 The structure of the prediction device 410 for the pearlite transformation curve of the roll shown in the figure does not constitute a limitation on the router. The actual knowledge structure identification device may include more or fewer components than shown, or combine certain components, or have different component arrangements.
[0067] Furthermore, the technical effect of the prediction device 410 for the pearlite transformation curve of the roll can be referred to the technical effect of the prediction method for the pearlite transformation curve of the roll described in the above method embodiments, and will not be repeated here.
[0068] It should be understood that the first processor 2001 in the embodiments of the present invention may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.
[0069] It should also be understood that the memory in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).
[0070] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.
[0071] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.
[0072] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.
[0073] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.
[0074] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.
[0075] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.
[0076] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.
[0077] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0078] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.
[0079] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0080] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.
Claims
1. A method for predicting the pearlite transformation curve of a rolling mill, characterized in that, The method includes: S1: Select steel alloy composition data through preset screening conditions, calculate steel alloy composition, and obtain alloy composition dataset. The alloy composition data is used to plot the continuous cooling transformation curve of supercooled austenite. S2: The alloy composition dataset is standardized to obtain a standardized dataset. The dataset is divided into a training set and a test set according to a preset ratio. The standardization process includes making the data conform to a standard normal distribution. S3: Construct a prediction model for the pearlite transformation curve of the roll. The prediction model for the pearlite transformation curve of the roll is a combined model, which is composed of a machine learning model. The output of the prediction model for the pearlite transformation curve of the roll includes the pearlite initiation temperature Ps and the pearlite transformation termination temperature Pf. S4: Based on the training set and the test set, the prediction model of the pearlite transformation curve of the roll is trained and tested to obtain a pre-trained prediction model of the pearlite transformation curve of the roll. S5: Based on the prediction model of the pre-trained pearlite transformation curve of the roll, a complete pearlite transformation curve of the roll is obtained. The complete pearlite transformation curve of the roll divides the pearlite into an upper part and a lower part. The upper part is formed by the combination of the pearlite initiation temperature Ps and the pearlite transformation termination temperature Pf to form a C-shape. S6: The SHAP value is used to interpret and analyze the prediction model of the pre-trained pearlite transformation curve of the roll, and the nonlinear relationship and importance ranking of the transformation temperature of the alloying elements and the phase are obtained. The nonlinear relationship between the alloying elements and the transformation temperature of the phase includes a negative correlation between the alloying elements Mn and Ni and a positive correlation between the alloying elements and Cr. The negative correlation indicates that the alloying elements cause the pearlite to move downward and the positive correlation indicates that the alloying elements cause the pearlite to move upward.
2. The method for predicting the pearlite transformation curve of a rolling mill according to claim 1, characterized in that, S1 selects steel alloy composition data through preset screening conditions, calculates the steel alloy composition, and obtains an alloy composition dataset, including: S11: Select data related to the composition of alloying elements from the original data after using preset screening conditions. The preset screening conditions include element content and phase transformation behavior. The alloying elements include Fe, C, Mn, Si, Cr, Ni, Mo, V, Al, W, Cu, Nb and N. S12: Calculate and organize the alloy element composition of these data; S13: Finally, obtain the alloy composition dataset.
3. The method for predicting the pearlite transformation curve of a rolling mill according to claim 1, characterized in that, The alloy composition data is used to plot the continuous cooling transformation curve of supercooled austenite, including: Machine learning methods are used to predict the pearlite region portion of the continuous cooling transformation curve of supercooled austenite in steel. Multi-dimensional alloy element composition data is used as input feature vectors, and the model is trained in a data-driven manner to learn the influence mechanism of each element on phase transformation dynamics. Based on the model, key critical points such as the phase transformation start temperature and end temperature under different cooling rates in the pearlite transformation region are accurately predicted, and the continuous cooling transformation curve of supercooled austenite is plotted.
4. The method for predicting the pearlite transformation curve of a rolling mill according to claim 1, characterized in that, The S3 construction of the prediction model for the pearlite transformation curve of the roll is a combined model composed of machine learning models. The output of the prediction model for the pearlite transformation curve of the roll includes the pearlite initiation temperature Ps and the pearlite transformation termination temperature Pf, including: S31: Determine the input and output of the prediction model for the pearlite transformation curve of the roll. The input includes alloying elements, austenitizing temperature, and cooling rate. The output includes Ps temperature and Pf temperature. The austenitizing temperature is used to replace the austenite grain size for modeling. S32: Extract variables from the training set according to the input and output, and fit and train the prediction model of the pearlite transformation curve of the roll to obtain the trained prediction model of the pearlite transformation curve of the roll. S33: Based on the test set, test the prediction model of the pearlite transformation curve of the rolled mill after training, verify the effectiveness of the model, and obtain the pre-trained prediction model of the pearlite transformation curve of the rolled mill.
5. The method for predicting the pearlite transformation curve of a rolling mill according to claim 1, characterized in that, S4, based on the training set and the test set, trains and tests the prediction model for the pearlite transition curve of the roll to obtain a pre-trained prediction model for the pearlite transition curve of the roll, including: S41: Determine the input and output of the prediction model for the pearlite transformation curve of the roll. The input includes alloying elements, austenitizing temperature, and cooling rate. The output includes Ps temperature and Pf temperature. The austenitizing temperature is used to replace the austenite grain size for modeling. S42: Extract variables from the training set according to the input and output, and fit and train the prediction model of the pearlite transformation curve of the roll to obtain the trained prediction model of the pearlite transformation curve of the roll. S43: Based on the test set, test the prediction model of the pearlite transformation curve of the rolled mill after training to verify the effectiveness of the model and obtain the pre-trained prediction model of the pearlite transformation curve of the rolled mill.
6. The method for predicting the pearlite transformation curve of a rolling mill according to claim 1, characterized in that, The prediction model of the pre-trained pearlite transformation curve of the roll in S5 yields a complete pearlite transformation curve. This complete curve divides the pearlite into an upper and lower part. The upper part is a C-shape formed by the combination of the pearlite initiation temperature Ps and the pearlite transformation termination temperature Pf, including: S51: Using alloying elements, austenitizing temperature, and cooling rate as inputs, the model inference is performed based on the pre-trained prediction model of the pearlite transformation curve of the roll to obtain the predicted pearlite initiation temperature Ps and pearlite transformation termination temperature Pf. S52: Based on the predicted pearlite initiation temperature Ps and pearlite transformation end temperature Pf, the complete pearlite transformation curve of the roll is obtained. S53: By collecting information from the upper and lower parts of the complete pearlite transformation curve of the roll, the C-shaped feature formed by the combination of Ps and Pf in the upper part and the distinction between the lower part are obtained.
7. The method for predicting the pearlite transformation curve of a rolling mill according to claim 1, characterized in that, The S6 step uses the SHAP value to interpret and analyze the pre-trained prediction model of the pearlite transformation curve of the roll, obtaining the nonlinear relationship and importance ranking of the transformation temperature of alloying elements and phases. The nonlinear relationship between the transformation temperature of alloying elements and phases includes a negative correlation between alloying elements Mn and Ni, and a positive correlation between alloying elements Cr and Ni. The negative correlation indicates that the alloying elements cause the pearlite to shift downwards, and the positive correlation indicates that the alloying elements cause the pearlite to shift upwards. S61: Based on the test set, calculate the SHAP value of each alloy element characteristic and draw a SHAP dependency graph. The SHAP dependency graph uses the actual value of a specific alloy element on the horizontal axis and its corresponding SHAP value on the vertical axis. S62: Based on the SHAP values of each alloying element, calculate the average value of the absolute values of the SHAP values of each alloying element. The larger the average value, the more significant the contribution of the element in the model prediction. S63: Observe the distribution trend of points in the SHAP dependency graph and identify the complex nonlinear relationship between the element and the pearlite transformation curve; S64: Based on the average absolute value of the SHAP values of various alloying elements, the importance ranking of the influence of each alloying element on the pearlite transformation curve is obtained.
8. A device for predicting the pearlite transition curve of a rolling mill, the device being used to implement the method for predicting the pearlite transition curve of a rolling mill as described in any one of claims 1-7, characterized in that, The device includes: Data module: used to select steel alloy composition data through preset screening conditions, calculate the steel alloy composition, and obtain alloy composition dataset. The alloy composition data is used to plot the continuous cooling transformation curve of supercooled austenite. Preprocessing module: used to standardize the alloy composition dataset to obtain a standardized dataset, and divide the dataset into a training set and a test set according to a preset ratio. The standardization process includes making the data conform to a standard normal distribution. Prediction Model Module: Used to construct a prediction model for the pearlite transformation curve of the roll. The prediction model for the pearlite transformation curve of the roll is a combined model, which is composed of a machine learning model. The output of the prediction model for the pearlite transformation curve of the roll includes the pearlite initiation temperature Ps and the pearlite transformation termination temperature Pf. Training module: used to train and test the prediction model of the pearlite transformation curve of the roll based on the training set and the test set, so as to obtain a pre-trained prediction model of the pearlite transformation curve of the roll. Transformation curve module: Used to obtain a complete pearlite transformation curve based on a pre-trained prediction model of the pearlite transformation curve of the roll. The complete pearlite transformation curve divides the pearlite into an upper part and a lower part. The upper part is formed by the combination of the pearlite initiation temperature Ps and the pearlite transformation termination temperature Pf to form a C-shape. SHAP module: Used to interpret and analyze the prediction model of the pre-trained pearlite transformation curve of the roll using SHAP values, and obtain the nonlinear relationship and importance ranking of the transformation temperature of the alloying elements and the phase. The nonlinear relationship between the alloying elements and the transformation temperature of the phase includes a negative correlation between the alloying elements Mn and Ni, and a positive correlation between the alloying elements and Cr. The negative correlation indicates that the alloying elements cause the pearlite to move downward, and the positive correlation indicates that the alloying elements cause the pearlite to move upward.
9. A device for predicting the pearlite transformation curve of a rolling mill, characterized in that, The predictive processor for the pearlite transformation curve of the roll; a memory storing computer-readable instructions that, when executed by the processor, implement the method as described in any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium contains program code that can be invoked by a processor to execute the method as described in any one of claims 1 to 7.